Aspen Plus Hydrocracker_User's Guide V7.3.pdf

Aspen Plus Hydrocracker

User's Guide

Version Number: V7.3 March 2011 Copyright (c) 2003-2011 by Aspen Technology, Inc. All rights reserved. Aspen Plus HydrocrackerTM, Aspen Plus HydrotreaterTM, Aspen Plus CatCrackerTM, Aspen Plus HydrocrackerTM, Aspen Plus®, Aspen PIMSTM, aspenONE, the aspen leaf logo and Plantelligence and Enterprise Optimization are trademarks or registered trademarks of Aspen Technology, Inc., Burlington, MA. All other brand and product names are trademarks or registered trademarks of their respective companies. This document is intended as a guide to using AspenTech's software. This documentation contains AspenTech proprietary and confidential information and may not be disclosed, used, or copied without the prior consent of AspenTech or as set forth in the applicable license agreement. Users are solely responsible for the proper use of the software and the application of the results obtained. Although AspenTech has tested the software and reviewed the documentation, the sole warranty for the software may be found in the applicable license agreement between AspenTech and the user. ASPENTECH MAKES NO WARRANTY OR REPRESENTATION, EITHER EXPRESSED OR IMPLIED, WITH RESPECT TO THIS DOCUMENTATION, ITS QUALITY, PERFORMANCE, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. Aspen Technology, Inc. 200 Wheeler Road Burlington, MA 01803-5501 USA Phone: (781) 221-6400 Toll free: (888) 996-7100 Website: http://www.aspentech.com

Contents About This Document ..............................................................................................1 Who Should Read This Guide .............................................................................1 Technical Support ............................................................................................1 Introducing Aspen Plus Hydrocracker .....................................................................3 Overview.........................................................................................................3 Introduction To Aspen Plus Hydrocracker ............................................................3 The Aspen Plus Hydrocracker Engine ..................................................................4 Equation-Oriented Modeling...............................................................................4 Pressure Drop Model Example............................................................................5 Model Specifications and Degrees-of-Freedom .....................................................6 Modes and Multi-Mode Specifications ..................................................................7 Measurements and Parameters ..........................................................................8 Changing Specifications with Specification Options ...............................................9 Optimization ....................................................................................................9 1 Using Aspen Plus Hydrocracker .........................................................................11 Starting Aspen Plus Hydrocracker for the First Time ........................................... 11 Resetting the Aspen Plus Connection ................................................................ 14 Exiting Aspen Plus Hydrocracker ...................................................................... 14 General Guidelines for Using the Excel Interface ................................................ 15 Saving and Loading Data Files ......................................................................... 16 Saving Data Files ................................................................................. 16 Loading Data Files ................................................................................ 17 2 The User Interface ............................................................................................19 The Sheets of the User Interface ...................................................................... 19 Flow Diagram Sheet ............................................................................. 20 Separation Section ............................................................................... 21 Buttons on the Flow Diagram Sheet........................................................ 22 Hidden Worksheets ........................................................................................ 28 Command Line Window................................................................................... 29 Overview............................................................................................. 29 Abort Button........................................................................................ 31 No Creep Button .................................................................................. 31 Close Residuals Button.......................................................................... 31 Close Button ........................................................................................ 31 Manual Access to the Command Line Window .......................................... 31 Toolbar and Menu .......................................................................................... 32 Startup Aspen Plus Hydrocracker Submenu ............................................. 33 File Submenu....................................................................................... 36 Setup Cases Submenu .......................................................................... 38

Contents

iii

Run Cases Submenu............................................................................. 38 Tools Submenu .................................................................................... 39 Development Tools Submenu................................................................. 39 Help Submenu ..................................................................................... 40 Exit Aspen Plus Hydrocracker........................................................................... 40 3 Working With The Equation-Oriented Solver .....................................................42 Introduction to the Equation-Oriented Solver ..................................................... 42 Successive Quadratic Programming (SQP) ......................................................... 42 Changing EO Solver Parameters ....................................................................... 43 Basic EO Solver Parameters............................................................................. 44 EO Solver Output to the Command Window ....................................................... 44 EO Solver Log Files......................................................................................... 46 ATSLV File Problem Information ....................................................................... 46 ATSLV Details ................................................................................................ 47 Basic Iteration Information .................................................................... 47 Largest Unscaled Residuals.................................................................... 47 Constrained Variables ........................................................................... 47 General Iteration Information ................................................................ 48 Nonlinearity Ratios ............................................................................... 49 Usage Notes .................................................................................................. 49 Usage Notes-General ............................................................................ 49 Dealing With Infeasible Solutions ........................................................... 50 Scaling ............................................................................................... 52 Dealing With Singularities ..................................................................... 52 Notes on Variable Bounding................................................................... 54 Run-Time Intervention .......................................................................... 54 4 Model Parameterization ....................................................................................55 Introduction .................................................................................................. 55 Flow Diagram Sheet ....................................................................................... 55 Product Properties ................................................................................ 55 Model View and Specification Through the Flow........................................ 57 Model Specifications ............................................................................. 58 Running a Parameterization Case ..................................................................... 67 Reconciliation Cases ....................................................................................... 69 More Detailed Parameterization.............................................................. 69 5 Simulation .........................................................................................................71 Introduction to Simulation ............................................................................... 71 Aspen Plus Hydrocracker Simulation Strategy .................................................... 71 Commonly-Used Scripts in the EB Script Language............................................. 73 Aspen Plus Hydrocracker Variable Specifications ................................................ 73 Model CONST Specifications .................................................................. 73 Model Tuning Facts with Specifications.................................................... 77 Flowsheet Changes............................................................................... 80 Running a Simulation Case .................................................................... 82 Error Recovery - Parameterization .......................................................... 84 6 Running Multiple Cases .....................................................................................86 Overview....................................................................................................... 86

iv

Contents

Before You Start ............................................................................................ 86 7 Optimization......................................................................................................89 Optimization Basics ........................................................................................ 89 Setting Up Objective Functions ........................................................................ 90 Setting Up An Optimization.............................................................................. 95 Executing Optimization Cases .......................................................................... 98 Analyzing Optimization Solutions.................................................................... 100 8 LP Vectors .......................................................................................................102 Overview – Generating LP Vectors .................................................................. 102 Purpose of Running LP Vectors....................................................................... 102 LP Vector Generation .................................................................................... 103 9 Reaction Kinetics Details .................................................................................108 Overview..................................................................................................... 108 Component Slate ......................................................................................... 108 Kinetic Framework........................................................................................ 113 Reaction Pathways ............................................................................. 113 10 Simplified Separation Model ..........................................................................116 Simplified Separation Model........................................................................... 116 Index ..................................................................................................................119

Contents

v

About This Document

This chapter includes the following information: 

Who Should Read This Guide



Technical Support

Who Should Read This Guide This document is designed to be used by the users of Aspen Plus Hydrocracker, formerly known as Aspen Hydrocracker, in conjunction with the Aspen RxFinery family of products, including Aspen Plus Reformer, formerly known as Aspen CatRef, Aspen Plus CatCracker, formerly known as Aspen FCC, Aspen Hydrocracker, and Aspen Plus Hydrotreater, formerly known as Aspen Hydrotreater.

Technical Support AspenTech customers with a valid license and software maintenance agreement can register to access the online AspenTech Support Center at: http://support.aspentech.com This Web support site allows you to: 

Access current product documentation



Search for tech tips, solutions, and frequently asked questions (FAQs)



Search for and download service packs and product updates



Submit and track technical issues



Send suggestions



Report product defects



Review lists of known deficiencies and defects

Registered users can also subscribe to our Technical Support e-Bulletins. These are used to alert users to important technical support information such as: 

Technical advisories



Product updates and releases

About This Document

1

Customer support is also available by phone, fax, and email. The most up-todate contact information is available at the AspenTech Support Center at http://support.aspentech.com.

2

About This Document

Introducing Aspen Plus Hydrocracker

Overview Aspen Plus Hydrocracker, formerly known as Aspen Hydrocracker, is a simulation system for monitoring, planning, and optimizing hydrocracking and hydrotreating units. Aspen Plus Hydrocracker is a member of the AspenTech new generation of refinery reactor models. Aspen Plus Hydrocracker accurately predicts yields and product properties for widely different feedstocks and operating conditions. An Aspen Plus Hydrocracker flowsheet simulates all sections of the hydrocracking unit. It can include simplified or vigorous fractionation models.

Introduction To Aspen Plus Hydrocracker Aspen Plus Hydrocracker consists of a client and a server. The client, or user interface, is built from Microsoft Excel spreadsheets customized with VBA code and macros. The client and server communicate through DCOM. This communication should be transparent, and you do not have to understand how it works in order to use Aspen Plus Hydrocracker If the communication software fails, contact AspenTech. While your primary interaction with Aspen Plus Hydrocracker will be through the user interface, you need a basic understanding of how the server works in order to effectively use and troubleshoot the model. The server has several components: 

The engine (also known as the kernel or command prompt).



The solver (DMO).



The model, which is built as a custom Aspen Plus model using the PML (Process Model Library) system.


3

The Aspen Plus Hydrocracker Engine The Aspen Plus Hydrocracker engine is Aspen Plus. You do not need to be an Aspen Plus expert to use Aspen Plus Hydrocracker – this section covers the most important concepts. The first time the engine is used during an Aspen Plus Hydrocracker session is when the user interface connects to the server. This brings up a command prompt window in which you will see the invoke plant.ebs command, which tells the engine to open several data files and build the model in the computer memory. The command prompt disappears when the kernel finishes building the model. The engine is also used whenever you request a solution from the user interface. Any changes you have made to data values or model specifications (via specification options) are passed through DCOM from the client to the server. The command prompt window appears and you will see a stream of kernel commands going to the engine. These commands tell the engine: 

What mode of solution is required.



What solver settings should be used.

There are different sequences of commands for different types of solutions (parameter, simulation, optimization, reconciliation, case study, LP vector generation, and so on.). You can look at the default command sequences on the EB Script sheet on the user interface. The default command sequences are all that is necessary for running the model in any of the pre-configured solution modes, but advanced users can modify them. During a solve, you will see three buttons on the bottom of the command prompt window. These are labeled Abort, No Creep, and Close Residuals. You use them to interrupt the solver. The Abort button tells the solver to quit at the next opportunity. The engine is also used whenever case data is stored or retrieved. The user interface typically contains only the results of the most recent run of each solution type. The save/load case data options let you save the results of any number of previous runs to review or use later. This user interface option is implemented using the kernel commands read varfile from and write varfile to. You can see these commands in the command prompt while it is active, or you can recall the command prompt using the user interface menu option Aspen Plus Hydrocracker | Tools | Display Command Line to review the previous commands.

Equation-Oriented Modeling Aspen Plus Hydrocracker is based on an equation-oriented (EO) formulation, so you need to understand some EO concepts in order to use it effectively. The EO approach is also known as open-form. It can be contrasted with the closed-form or sequential-modular (SM) technique.

4


The equations in an EO model are solved simultaneously using an external solver, which iteratively manipulates the values of the model variables until all the equations are satisfied within a convergence tolerance. The solver will work for any well-posed set of variable specifications. A variable’s specification labels it as 

known (fixed) -or-



unknown (calculated)

for a given solution mode. An SM model is solved procedurally one equation at a time, and the solution procedure depends on a given specification set. For a different grouping of known and unknown variables the solution procedure will be different, since the equations will be solved in a different order.

Pressure Drop Model Example A simple example illustrates some important EO concepts. Consider this twoequation model, in which the pressure drop is correlated with the square of the mass flow of a fluid: Pressure drop correlation:

DELTAP = PRES_PARAM * MASS_FLOW^2

Define pressure drop:

DELTAP = PRES_IN – PRES_OUT

In an EO formulation, we rearrange these equations into residual format. The value of the residual indicates how close that is to being solved – at the solution the value of every residual will be zero, or at least close enough to zero to satisfy our numerical convergence tolerance. f(1) = DELTAP - PRES_PARAM * MASS_FLOW^2

(= 0 at solution)

f(2) = PRES_IN - PRES_OUT - DELTAP

(= 0 at solution)

Note that f is the name of the vector of residuals. Its length equals the number of equations. The solver prefers to work with vectors and equation index numbers, while we find it easier to use equation names. The model defines names for each residual that can be used in reports and solver debugging output. In this case, we choose the names: f(1) = ESTIMATE_DELTAP f(2) = DELTAP_DEFINITION Similarly, the five variables in this model can also be addressed as elements of a vector x having a length of 5: x(1) = DELTAP x(2) = PRES_IN x(3) = PRES_OUT x(4) = PRES_PARAM x(5) = MASS_FLOW


5

Model Specifications and Degrees-of-Freedom Once we tell the solver which variables are known (fixed) for a given solution mode, it manipulates the values of the unknown (free) variables to drive the residuals to zero. For any system of independent equations, the degrees-offreedom (DOF) is equal to the number of variables minus the number of equations minus the number of fixed variables:

DOF = #variables - #equations - #fixed variables The number of degrees-of-freedom of a system classifies it into one of three categories: Category

Degrees of Freedom

Under specified

>0

Square

0

Over specified

<0

Aspen Plus Hydrocracker modes are either under specified or square. Over specified problems are not allowed in Aspen Plus Hydrocracker. Aspen Plus Hydrocracker Mode

Category

Optimization

Under specified

Reconciliation

Under specified

Simulation

Square

Parameter

Square

Case Study

Square

LP Vector

Square

The pressure drop example has five variables and two equations, so we must fix three variables to create a square system. Furthermore, we cannot fix any arbitrary set of three variables. If all variables within one equation are either explicitly or implicitly fixed, the problem is not well posed, as the solver can no longer manipulate any variable to reduce that equation’s residual. Such an incorrect set of specifications will cause a structural singularity in the solver. However, Aspen Plus Hydrocracker is designed so that if you use the standard specification options provided in the user interface you will not create a structurally singular system.

6


Here are some specification attempts for the pressure drop example: Fix DELTAP, PRES_OUT: under specified - only acceptable for an optimization case with proper selection of independent variables. Fix DELTAP, PRES_OUT, PRES_IN, MASS_FLOW: over specified! Fix DELTAP, PRES_OUT, PRES_IN: f(1) = DELTAP - PRES_PARAM * MASS_FLOW^2 (fix)

(free)

(free)

f(2) = PRES_IN - PRES_OUT – DELTAP (fix)

(fix)

(fix)

Square, but not well posed (structurally singular) – all variables in residual 2 are fixed! If you compare this to the over specified example, you can see that over specification is not allowed since it always leads to a structurally singular system. Fix PRES_IN, PRES_PARAM, MASS_FLOW: f(1) = DELTAP - PRES_PARAM * MASS_FLOW^2 (free)

(fix)

(fix)

f(2) = PRES_IN - PRES_OUT – DELTAP (fix)

(free) (free)

Square and well posed – a valid specification set. Note that there are other valid specification sets, such as PRES_IN, PRES_OUT, and MASS_FLOW.

Modes and Multi-Mode Specifications In different situations we may want to use different sets of fixed and free variable specifications. Each set of variable specifications is a solution mode. One of the strengths of the EO approach is that the same model formulation and solver are used for all the modes. Although there are many possible modes, Aspen Plus Hydrocracker is configured for three basic modes: The Process Details, ParamData and Optimize sheets in Aspen Plus Hydrocracker correspond to those three modes. Case study and LP vector generation are also simulation modes. 

Case study is simply a series of simulations with the same specifications, but different values for key fixed variables.



LP vector generation is a simulation run followed by a sensitivity analysis. The independent and dependent variables you choose for vector generation must correspond to fixed and free variables in the simulation mode.


7

The Aspen Plus Hydrocracker user interface examines the current model specifications and lets you choose only proper independent and dependent variables. In order to label how each variable behaves in the various modes, multi-mode specifications are assigned. A variable that is fixed in every mode is called a CONST, while variables that are free in every mode are called CALC. For example, in Aspen Plus Hydrocracker, the reactor vessel diameter is usually a CONST because its value is not calculated in any mode, while the catalyst weight of each reactor bed is usually a CALC because the model calculates its value from other information.

Measurements and Parameters While many variables have CONST or CALC specifications, there are other variables whose behavior changes between modes. A MEAS variable is fixed in the parameter-fitting (tuning) mode, but free in the simulation and optimization (prediction) modes. Conversely, a PARAM variable is free in the parameter-fitting mode and fixed in the simulation and optimization modes. Usually a MEAS corresponds to a plant measurement, while a PARAM is a model tuning parameter or a bias to a measurement. Because the MEAS and PARAM variables always have opposite specifications in every mode, there are always the same number of MEAS and PARAM variables so that every mode is properly specified. Another rule of thumb is that it is possible to "swap" the specifications on a pair of related CALC and CONST variables to be MEAS and PARAM, since the number of DOF stays the same in every mode. The concepts of simulation and parameter-fitting mode and CONST/CALC/MEAS/PARAM variables can be illustrated with the pressure drop example. Assume the equipment across which the pressure drop is measured has an inlet pressure gauge, a DP cell, and a mass flowmeter. We can specify the DP measurement (variable DELTAP) to be type MEAS and the pressure drop parameter (PRES_PARAM) to be type PARAM. We can define inlet pressure (PRES_IN) and mass flowrate (MASS_FLOW) as CONST variables. The outlet pressure (PRES_OUT) is always calculated from the other variables, so it is type CALC. f(1) = DELTAP - PRES_PARAM * MASS_FLOW^2 (MEAS)

(PARAM)

(CONST)

f(2) = PRES_IN - PRES_OUT – DELTAP (CONST)

(CALC)(MEAS)

This is a valid multi-mode specification, because in the simulation mode MASS_FLOW, PRES_IN and PRES_PARAM are fixed and PRES_OUT and DELTAP can be calculated from those values. In the parameter-fitting mode:

8




DELTAP, MASS_FLOW and PRES_IN are fixed.



PRES_PARAM and PRES_OUT can be computed.

Changing Specifications with Specification Options What if the plant we are modeling has both a DP cell and an outlet pressure gauge? We have a choice as to which to use. From a mathematical standpoint, it is just as valid to declare PRES_OUT a MEAS and DELTAP a CALC as the other way around. Thus we have two possible variable specifications affecting both our simulation and parameter-fitting modes. In Aspen Plus Hydrocracker this type of spec swap is made using the Specification Options button on the Flow Diagram Sheet A specification option is a pre-set set of alternate specifications that are equally mathematically valid. The specification sets is set by the EBS script defined on the ES Scripts worksheet. One of the sets may be more appropriate for a given unit based on its configuration, control strategy, instrumentation, type of lab test, mass or volume basis for flowmeters, or a variety of other reasons. In our pressure drop example, on the param sheet we might see a Specification Option with the following options:

Use outlet pressure measurement Use pressure drop measurement These choices correspond to the following specifications: Use outlet pressure measurement

Use pressure drop measurement

DELTAP spec

CALC

MEAS

PRES_OUT spec

MEAS

CALC

Note: The scripts that are associated with the specification options are located on the EB Scripts worksheet. To view them, you must unhide the EB Scripts worksheet by clicking Format | Sheets | Unhide; then clicking EB Scripts in the Unhide dialog. The scripts start in Column L and end in Column S. You can add your own custom user scripts by inputting them into Column V through Column Z. Each line must be a valid EB script command (or comment - a line starting with //). Any blank line will be interpreted as the end of the script.

Optimization Optimization is a prediction mode, so it is similar to simulation. The main difference is that there are positive DOF in optimization mode, and the solver


9

uses those DOF to maximize or minimize an objective function within limits on certain variables. To create optimization DOF, simply change the specifications of some CONST variables to OPTIM. OPTIM variables are fixed in simulation and parameter-fitting modes and free in optimization mode and are also known as independents. The other free variables (MEAS and CALC) are known as dependents. The solver requires that the number of OPTIM variables be equal to the number of DOF, but that requirement is easy to satisfy by starting with a well-posed square set of multi-mode specifications and changing only CONST variables to OPTIM. Essentially, you must do three things: 

Define an objective function.



Specify the DOF (independents).



Put maximum and minimum limits on key independent and dependent variables.

Often a profit function contains: 

Revenue terms based on product or export utility flowrates and prices.



Cost terms based on feed or import utility flowrates and prices.

You specify the DOF by selecting independent (OPTIM) variables from a pick list. Aspen Plus Hydrocracker presents only CONST variables in this pick list in order to ensure that whatever set you choose will lead to a well-posed problem. You can put bounds on any of the independents, plus whichever dependents you select from another pick list that includes CALC and MEAS variables that you may wish to limit during the optimization run.

10


1 Using Aspen Plus Hydrocracker

Starting Aspen Plus Hydrocracker for the First Time Starting Aspen Plus Hydrocracker For the First Time: The first time you start Aspen Plus Hydrocracker, you need to: 

Load the Hydrocracker Flowsheet.



Establish a connection to the Aspen Plus Hydrocracker model, which is an Aspen Plus Flowsheet.

To Load the Hydrocracker Flowsheet: 1

From the Windows Start menu, click Programs | AspenTech | Process Modeling | Aspen Plus Based Refinery Reactors | Aspen Plus Hydrocracker to launch Excel and open the Aspen Plus Hydrocracker GUI.

2

When prompted by Excel, click the Enable Macros button.

Note: Aspen Plus Hydrocracker does not support having multiple versions of itself or Aspen Plus installed at the same time. When the Aspen Plus Hydrocracker workbook is loaded, there is no active connection to the Aspen Plus Hydrocracker model, which is an Aspen Plus flowsheet. The workbook consists of several spreadsheets where various data can be entered and retrieved. The application also creates a new menu item on the Excel menu bar called AspenPlusHYC. This menu provides access to all of the GUI’s primary functions including connecting to the model. Through the Startup Aspen Plus Hydrocracker menu command, you can load the flowsheet, modify start-up options, or reset the Aspen Plus connection. Most of the other menu commands are inactive until the flowsheet is loaded.


11

3

On the Excel menu bar click AspenPlusHYC | Startup Aspen Plus Hydrocracker | Load Hydrocracker Flowsheet.

To Establish a Connection to the Hydrocracker Flowsheet: The Connect dialog box appears.

1

In the Host field, enter the computer name using all lower case letters. The Browse button will become available If the correct computer name is entered,

Note: You can easily determine the computer name if it is not known:  Win2000: Right-click the My Computer icon on the computer desktop and select Properties from the pop-up menu. Click the Network Identification tab where the full computer name will be listed near the top.  Windows XP: Right-click the My Computer icon on the computer desktop and select Properties from the pop-up menu. Click the Computer Name tab. The computer name will be listed in the Full Computer Name field...

12


2

Click the Browse button.

3

Navigate into the Apinit directory, select the file hyc.appdf; then click the Open button.

You are returned to the Connect dialog box where the hyc.appdf file name its directory now appear in the Problem area. 4

At the bottom of the Connect dialog box, click the OK button.

Aspen Plus Hydrocracker now loads the hyc.appdf file. This loading may take up to five minutes, depending on the speed of your machine. During this time, the Excel cursor will become the hour-glass symbol, and the Excel status line will display the message Loading Aspen Plus Hydrocracker flowsheet. The cursor will return to the normal cross shape and the status line will read Ready when the process is complete. Once connected to the flowsheet, the previously inactive AspenPlusHYC menu items become active. The HydroCracker toolbar is also created. You are now ready to begin using Aspen Plus Hydrocracker. Note: Before using Aspen Plus Hydrocracker, you may want to save the computer name and hyc.appdf file location entered in the Connect dialog box.

To Do This: 

Click Saving Data Files

Starting Aspen Plus Hydrocracker After the First Time: If you saved your data file when you first opened Aspen Plus Hydrocracker, the Connect dialog box will be populated with your computer name and the name and location of the .appdf file. 1

From the Windows Start menu, click Programs | AspenTech | Process Modeling | Aspen Plus Based Refinery Reactors | Aspen Plus Hydrocracker to launch Excel and open the Aspen Plus Hydrocracker GUI.

2

When prompted by Excel, click the Enable Macros button.

The Connect dialog box appears, populated with your computer name and the name and location of the .appdf file. 3

At the bottom of the Connect dialog box, click the OK button.

Aspen Plus Hydrocracker now loads the hyc.appdf file. This loading may take up to five minutes, depending on the speed of your machine. During this time, the Excel cursor assumes the hour-glass symbol, and the Excel status line displays the message Loading flowsheet. The cursor will return to the normal cross shape and the status line will read Ready when the process is complete. Once connected to the flowsheet, the previously inactive AspenPlusHYC menu items become active. The Hydrocracker toolbar is also created. You are now ready to begin using Aspen Plus Hydrocracker.


13

Resetting the Aspen Plus Connection Occasionally, problems can occur where the AspenPlusHYC menu items and toolbar are still active, but the functions fail with various VB errors. This can be the result of loading too many applications at once, thereby causing an application conflict.

To Reset The Aspen Plus Connection: 1

On the main menu, click AspenPlusHYC | Startup Aspen Plus Hydrocracker | Reset ApMain, to reset the connection to the Aspen Plus Hydrocracker flowsheet.

The Reset Aspen Plus warning screen appears.

2

Click the OK button to reset the Aspen Plus connection (and to terminate all Aspen Plus processes).

Exiting Aspen Plus Hydrocracker The best way to exit is to use the menu item, AspenPlusHYC | Exit Aspen Plus Hydrocracker.

14


1

Click AspenPlusHYC | Exit Aspen Plus Hydrocracker.

A dialog box appears, asking for confirmation. 2

Click the OK button to proceed with exiting. You are asked whether you want to save the changes made to the Excel workbook.

3

Click the Yes button to save your changes and exit. -orClick the No button to abandon your changes and exit -orClick the Cancel button to abort the exiting operation.

If you click the Yes or No button: 

The workbook closes.



The AspenPlusHYC menu disappears.



The AspenPlusHYC toolbar is hidden.

General Guidelines for Using the Excel Interface Most of the features of Excel are available in the Aspen Plus Hydrocracker workbook. However, you should only use these features with an understanding about the overall functioning of the workbook. Here are some things to consider as you use the workbook: 

The only fields that you can make an entry into that the model will use are those colored blue.



Entries into number fields that are not colored blue are overwritten by the workbook after a case is executed.



Enter only values into blue fields. If you use a formula in a blue field, it will be overwritten after a case is executed. Therefore, enter only values in these fields.


15



If you change an option with a combo box, the color coded fields are not automatically updated. To update the fields, click the Update Fields button on the Hydrocracker menu.



If a case does not converge, the calculation engine contains a starting point that is not good for subsequent cases. Therefore, before running a suspect case, save your case data. You can then load your case data to restore a case if the problem does not converge.



The data you enter into the parameter and simulation worksheets is automatically saved by the workbook when a case is run. You can retrieve it after you restore a case to create a good starting point for the calculation engine.



The model is an equation-based model and needs a good starting point to converge. Therefore, be careful about large changes in the independent variables (color coded blue).

Saving and Loading Data Files Saving Data Files Use the AspenPlusHYC | File | Save User Data to Var File command to: 

Save the initial file when you first connect to Aspen Plus Hydrocracker. This preserves the name of your computer and the path to the appdf file you selected in the Connect dialog box.



Save a file for loading later, particularly if you suspect that a run may not converge. In this case, your saved file can provide a good starting point for other runs.

To Save A Data File: 1

On the Excel menu bar click AspenPlusHYC | File | Save User Data to Var file.

The Save User Data to File dialog appears. 2

In the File Name field, enter the name under which you want to save this file. You do not need to add the var extension.

3


The Save As dialog appears. 4

On the Save As dialog, browse to the directory in which you want to save this file.

5

On the Save As dialog, click the Save button.

The Save User Data to File dialog appears again. 6

16

On the Save User Data to File dialog, click the Save button.


Your .var file is saved.

Loading Data Files You can load .var files that you previously saved. Note: Loading a .var file will cause you to lose the data currently in the Workbook.

To Load Case Data: 1

On the main menu, click AspenPlusHYC | File | Load User Data from Var File.

The Load User Data from File dialog appears.

2


The Open dialog appears.


17

3

On the Open dialog, browse to the correct file name and directory path.

4

Left-click the file name.

The file name appears in the File name field. 5

Click the Open button.

The Load User Data from File dialog appears, with the file name and directory path you selected entered. 6

18

Click the Load button to load the file and overwrite the current workbook. -orClick the Cancel button to abort loading the .var file.


2 The User Interface

The Sheets of the User Interface When you start Aspen Plus Hydrocracker, the default spreadsheet is the Flow Diagram sheet. You can navigate to other data entry or results areas by selecting the appropriate tab at the bottom of the Excel window. You can also access many of these tabs via buttons on the Flow Diagram sheet. The Sheets of Aspen Plus Hydrocracker are: Worksheet Name

Description

Flow Diagram

Provides Flowsheet, data display entry buttons, and change the model specification

ProcessOverview

Summary of Key operation conditions

Reactor Profiles

Temperature, sulfur, nitrogen, and Aromatics profile for each reactor.

ProcessDetail

Data entry and reports for the simulation case

ParamData

Data entry and reports for the parameter and reconciliation cases

Cases

Entry and setup form for case studies

PIMS Vectors

Vectors for generating a PIMS Table

PIMS Table

PIMS Table

LP Vectors

Entry and setup form for LP vector calculations

Optimize

Entry and setup form for optimization calculations

Profit 1

Entry and setup form for profit based objective function 1

Profit 1 Report

Report form for objective function 1

Profit 2


Profit 2 Report


Profit 3


Profit 3 Report


Many of these sheets can also be accessed by buttons on the Flow Diagram sheet.


19

Flow Diagram Sheet The Flow Diagram Sheet provides a overview of the process flow for the hydrocracker. The Aspen Plus Hydrocracker/Hydrotreater model is built based on Aspen Plus. The figure below shows the process flow sheet, which is built into the Flow Diagram sheet of the Excel spreadsheet.

The Flow Diagram Sheet has three main sections: 

Feed system.



Reaction section.



Separation section.

Feed System The Feed System consists of AspenPlus blocks and customized blocks. These blocks let you build the feeds fed into the reactor. These blocks provide you flexibility in specifying the feeds.

Feed Stream Model (AFFED1-AFFED6). Six pre-specified feed stream blocks are provided based on the Aspen feed information database.

20



Light vacuum gas oil (LVGO)



Heavy vacuum gas oil (HVGO)



Light coker gas oil (LCGO)



Heavy coker gas oil (HCGO)



FCC LCO (LCO)



Hydrocracker bottoms (RECOIL)


Feed Adjuster Model (FEEDADJ). In this model, you can adjust (within reasonable constraints) all feed fingerprints by a feed adjuster model to match client-specified bulk properties: 

distillation



gravity



sulfur



total nitrogen



basic nitrogen



bromine number



refractive index (optional)



viscosity (optional)

Reaction Section The blended feed, recycle oil, and hydrogen mixed with the recycle gas through a compressor are mixed again, then sent to a furnace (modeled as a heater). The effluent from the furnace is sent to Reactor 1. Therefore, the reactor R1 inlet temperature can be adjusted by manipulating the heater outlet temperature. The whole reactor is modeled by a series of standard Aspen EO reactor blocks, which include, 

Olefins Reactor Models (OLFRXN). One extent-of-reaction block (EOXNTRXN) is used for saturating the olefins to the reactor components.



Reactor Bed Models (R1B1-R2B2). One reactor block (EORXR) is used for each catalyst bed. The Langmuir-Hinshelwood (adsorptionreaction/inhibition-desorption) mechanism is assumed. Collocated tricklebed kinetics and phase equilibria are employed. Reaction rates and phase equilibrium are recomputed at each collocation point, which provides very precise prediction of heat release. Reaction types include:  Saturation of olefins  Saturation of aromatics  Hydrodesulfurization (HDS)  Hydrodenitrogenation (HDN)  Ring opening  Ring dealkyation  Paraffin hydrocracking

Separation Section The effluent from the reactor goes through two heaters: 

E1H integrated with the feed heater E1C.



An air cooler (FinFan).

The effluent then enters the separation section.


21

The Separation Section contains a number of Aspen Plus blocks and simplified separation models: Block

Description

High Pressure Separator (HPS).

The bottom stream of HPS goes to a Low Pressure Separator (LPS). The model is an Aspen Plus flash model.

Low Pressure Separator (LPS).

The bottom of LPS is sent to the separation section (main fractionator, gas plant, and so on). The whole separation section is built as a simplified separation model.

Simplified Separation Model for Main Fractionator and Gas Plant (PRODSP)

This simplified model uses a combination of component splitters, analyzers, and calculators.

Buttons on the Flow Diagram Sheet Button

Action

R1 R2 HTR

takes you to the Process Detail sheet to display a summary of the current running conditions

Feeds

takes you to the Process Detail sheet to display feed properties of the combined feed and input sheet for each individual feed

Yields

takes you to the Process Detail sheet to display: the volume and mass yields of product the product properties of each product

Process Overview

takes you to the Process Overview sheet, on which general information for AspenPlusHYC model is presented

Reaction Profile

takes you to the Reaction Profile sheet. On this sheet a set of diagrams is set to present the temperature, sulfur, nitrogen and aromatic contents for the two beds in the two reactors

Specification Options

a dialog box pop up to let you select which mode to run: 

Temperature Control



User Scripts

Click the appropriate tab; then click the Select button. Run a Param case

22

automatically refreshes the ParamData sheet when the data is passed back to the spreadsheet from the solver


Process Overview Button On the Flow Diagram Sheet, click the Process Overview button to view the Process Overview sheet as shown below.

Reactor Profiles Button On the Flow Diagram Sheet, click the Reactor Profiles button to view the Reactor Profiles sheet as shown below.


23

Specification Options Button On the Flow Diagram Sheet, click the Specification Options button to select specification options on the Select Spec.Options dialog as shown below.

To Change A Specification: 1

Click the desired tab: Temperature Control or User Scripts.

2

Click the Specification you want to change.

3

Click the Select button.

The specification you selected is updated. For details about setting and swapping specs, click here.

Feeds Button On the Flow Diagram Sheet, click the Feeds button to view the Feeds section of the Process Detail sheet as shown below.

24


R1 Button On the Flow Diagram Sheet, click the R1 button to view the Reactor 1 section of the Process Detail sheet as shown below.

R2 Button On the Flow Diagram Sheet, click the R2 button to view the Reactor 2 section of the Process Detail sheet as shown below.


25

HTR Button On the Flow Diagram Sheet, click the Htr button to view the Heaters, Exchangers and Flashes section of the Process Detail sheet as shown below.

Save to Prior Button The Save to Prior button saves the current solution to the Prior column on these worksheets:

26



Process Details



Process Overview




Param Data

This lets you make quick comparisons between different runs. For example, you can run a base case and save the values to the Prior column. Then, for any simulate run, you can quickly compare the results in the value column to the prior column to see how things changed.

H2 Streams Button Click the H2 Streams button on the Flow Diagram Sheet to view the H2 Streams section of the Process Detail sheet as shown below.

H2 Balance Button Click the H2 Balance button on the Flow Diagram Sheet to view the H2 Balance section of the Process Detail sheet as shown below.


27

Yields Button Click the Yields button on the Flow Diagram Sheet to view the Yields section of the Process Detail sheet as shown below.

Hidden Worksheets The Aspen Plus Hydrocracker workbook contains several worksheets that are hidden by default. These are not needed for general use of the workbook, but you can view the information on them.

28


To View A Hidden Worksheet: 1

On the Excel toolbar, click Format | Sheet | Unhide.

2

Click the worksheet you want to unhide.

Note: Some of these worksheets are password protected to prevent inadvertent changes to their contents. Such changes can affect the functionality of the workbook and cause a failure to occur in this functionality. The Hidden Worksheets are listed below. Worksheet Name

Description

ProcessOverview IO

Structures the layout for the ProcessOverview worksheet

ProcessOverview Links

Contains direct cell links to the model variables available in the workbook for the ProcessOverview worksheet

ProcessDetail IO

Structures the layout for the ProcessDetail worksheet

ProcessDetail Links

Contains direct cell links to the model variables available in the workbook for the ProcessDetail worksheet

ProcessDetail UserInput

Contains a copy of your input for the ProcessDetail worksheet

ParamData IO

Structures the layout for the ParamData worksheet

ParamData Links

Contains direct cell links to the model variables available in the workbook for the ParamData worksheet

Param UserInput

Contains a copy of your input for the ParamData worksheet

ProcessOverview IO

Structures the layout for the Simulation worksheet

ProcessOverview Links

Contains direct cell links to the model variables available in the workbook for the Simulation worksheet

ProcessOverview UserInput

Contains a copy of your input for the Simulation worksheet

Feed Input EB Scripts

Contains the script for execution by the calculation engine

ReceiveVars

Contains and manages variables that are sent from the calculation engine to the workbook through DCOM

SendVars

Contains and manages variables that are sent to the calculation engine from the workbook through DCOM

Registry

Contains a collection of data and parameters for the Aspen Plus Hydrocracker workbook

PIMSin PIMSout

Command Line Window Overview The Aspen Plus Command Line window displays the output of commands sent to the Aspen Plus Hydrocracker model. It appears automatically when loading Aspen Plus Hydrocracker and when running cases.


29

After connecting to the Aspen Plus Hydrocracker flowsheet, you can also manually open this window by selecting the AspenPlusHYC | Tools | Display Command Line menu command.

When Aspen Plus Hydrocracker is loading, the Command Line window appears briefly, letting you observe the commands that are being sent to the model during the flowsheet instantiation. You cannot access any functions on the command line at this time. When a case is running, the Command Line window opens automatically. It lets you see: 

The commands that are being sent to the model.



The convergence path of a solution.

In these instances, when the command line opens automatically, the only buttons available to you are:

30



Abort



No Creep



Close Residuals


Abort Button Click the Abort button to abort the solving of a case. If you click the Abort button while a case is running, you must wait until the following messages appear in the command line window:

Error return due to an ABORT message from the user communications file DMO.MSG Problem failed to converge You can now click the Close button to close the command line window and return to the model. You should then load a data file to ensure the next case starts from a good converged solution.

No Creep Button When running a case, the default is to creep the solver (take small steps) for a few iterations to provide robust behavior. Once you have gained experience with the model and are confident that a particular case will solve well without the default number of creep steps, you can manually turn the creep steps off by clicking the No Creep button. You can click the No Creep button while a problem is converging. This causes the solver to eliminate the creep in the next iteration.

Close Residuals Button Use the Close Residuals button to have the model close the residuals without minimizing the objective function convergence. You might find this useful in cases where the objective function very nearly reaches a maximum value, but the convergence of the objective does not close.

Close Button This button closes the Command Line window and returns you to the Excel user interface. Click the Close button only: 

After a run has failed to converge.



If you aborted a case and the command line message run aborted by the user appears.



If you opened the Command Line window manually, and you have finished using it..

Manual Access to the Command Line Window After connecting to the Aspen Plus Hydrocracker flowsheet, select the AspenPlusHYC | Tools | Display Command Line menu command. The Aspen Plus Command Line window appears.


31

When you open the command line manually, some buttons are available and some are not: Button

Available

Execute

Yes

Abort

No

No Creep

No

Close Residuals

No

Close

Yes

The Abort, No Creep, and Close Residuals buttons have no effect when the command line has been opened manually unless the Execute command is invoked to run Aspen Plus. The Close button closes the command line window and returns you to the Excel spreadsheet. While the command line window is open, you cannot access the Excel spreadsheet. The command line window can be a very powerful tool in trouble-shooting problems since the commands sent to the model and the solutions of the model are stored in the buffer. You can scroll through the buffer (the top window of the command line) to see convergence paths and any error messages generated when trying to solve a problem.

Toolbar and Menu When the Aspen Plus Hydrocracker workbook is selected, Microsoft Excel is loaded and then the Hydrocracker VBA loads a drop-down menu selection to the Excel toolbar labeled AspenPlusHYC. This menu contains selections that activate macros within the Hydrocracker VBA. This section explains the options and dialogs that are available from the Aspen Plus Hydrocracker dialog. Many of the options are associated with the cases for Hydrocracker modeling.

32


Startup Aspen Plus Hydrocracker Submenu Overview When you select the AspenPlusHYC option on the toolbar, the drop-down menu appears as shown below. Some of the options are grayed out because the workbook has not yet been connected to the calculation engine through the server. The options on the Development Tools submenu are for advanced functions in the workbook and will not be covered here. The selection you should make at this time is Startup Aspen Plus Hydrocracker. With the exception of the advanced functions, this is the first selection you should make when you first activate the Aspen Plus Hydrocracker workbook. The Startup AspenPlusHYC submenu contains the commands you typically use when you first activate the Aspen Plus Hydrocracker workbook.

When you select Startup Aspen Plus Hydrocracker option, the menu shown above appears. Command

Function

Load Hydrocracker Flowsheet

Connect the workbook and load a problem file

Startup Options

Load a problem file automatically or manually

Reset ApMain

Resets the connection with the Aspen Plus server

The Load Hydrocracker Flowsheet option is normally the first command you will use. This command displays the Connect dialog box. The Reset ApMain command causes the workbook to break the connection with the server. This is necessary if you want to use the Excel File menu. If you do not close the workbook at this point, you can use the Load Hydrocracker Flowsheet command to reconnect the workbook.

Connect Dialog Box 1

On the Excel menu bar, select AspenPlusHYC | Startup Aspen Plus Hydrocracker | Load Hydrocracker Flowsheet.


33

The Connect dialog box appears.

2

In the Host box, enter the name of the host computer (normally your computer) using all lower case letters. If the correct computer name is entered, the Browse button in the Problem area will become enabled.

Note: You can easily determine the computer name if it is not known. Win2000: Right-click the My Computer icon on the computer desktop and select Properties from the pop-up menu. Click the Network Identification tab where the full computer name will be listed near the top. Windows XP: Right-click the My Computer icon on the computer desktop and select Properties from the pop-up menu. Click the Computer Name tab. The computer name will be listed in the Full Computer Name field. 3

Click the Browse button in the Problem area, navigate into the Apinit directory, select the file hyc.appdf, and then click Open.

4

At the bottom of the Connect dialog box, click OK. On a 750 MHz Pentium III PC, such as a Dell Inspiron 8000, it requires approximately two minutes to initialize the Hydrocracker flowsheet and load the data into the Excel GUI. During this time, the Excel cursor will appear as an hourglass symbol and the Excel status line will display the message Loading Aspen Plus Hydrocracker flowsheet. The cursor will return to the normal cross shape and the status line will display Ready when the process is complete. Once the connected to the flowsheet is established, the previously inactive AspenPlusHYC menu commands become active, and the Aspen Plus Hydrocracker toolbar is created.

34


5

Now save the workbook using the Excel File | Save command, to preserve the computer name and hyc.appdf file location entered in the Connect dialog box.

You are now ready to begin using Aspen Plus Hydrocracker.

Startup Options Dialog Box The Startup Options dialog box is shown below. This dialog box lets you specify a default problem solution to load into the workbook other than hyc.appdf (the base solution).

When the workbook is opened, there is by default no connection established with the Hydrocracker flowsheet. Furthermore, once the connection is established, the data loaded into the spreadsheet will be the data that comes with the generic model. You can change these default settings to improve efficiency. By modifying the startup options, you can automatically connect to the Hydrocracker spreadsheet and load a specific user data file immediately upon opening the Hydrocracker GUI. At the top of the Startup Options dialog box, you can choose to make a connection to the Hydrocracker model either manually or automatically. If you select Automatic Startup, the spreadsheet will automatically establish a connection to the model whenever it is opened. The Startup Options dialog box also has an option to load in a set of data other than the default problem data. Automatically loading data that matches your plant is more convenient.

To Set Startup Options: 1

On the Excel menu bar, select AspenPlusHYC | Startup Aspen Plus Hydrocracker | Startup Options. The Startup Options dialog box appears.


35

2

Select Manual Startup or Automatic Startup. Your choice will determine whether the connection to the Hydrocracker model is made manually or automatically.

3

If you chose the Automatic Startup option in Step 2, you can load a set of data other than the default problem data in the hyc.appdf file. To do so, select the Load User Data from File? checkbox.

4

In the File Name box, enter the name of the data file to be loaded (including the full path). Normally, this is a file that you have saved from a previous execution of the program.

5

Click OK.

File Submenu The second submenu on the AspenPlusHYC menu is File.

The File submenu contains four commands: Use this Command

To

Load Case Data

Invoke a dialog box to load a case file

Save Case Data

Invoke a dialog box to save a case file

Load User Input Sheet

Load data you previously entered on a Parameter or Simulation worksheet

Save User Input Sheet

Save data you previously entered on a Parameter or Simulation worksheet

Use the items on this menu to: 

Save and load case data.



Save and load your data entry sheets.

Load Case Data Use the Load Case Data command to load in a good starting point from saved data. Typically the file you load you have previously saved using the Save Case Data command. Do this when a solution is not achieved and the solver is left with a bad starting point.

36


Save Case Data Use the Save Case Data command to save a file which has a good starting point for possible later retrieval using the Load Case Data command. It is good practice to periodically save data, because the solver can sometimes be left with a bad starting point if a solution is not achieved. The hydrocracker model is an equation-based model that can be moved from a base solution to another base solution, if the move is not too large. Normally, as a very general rule, too large means a move of about 20% to 30% on values other than temperatures. Temperatures changes can be in the range of 10 to 20 °F.

Load User Input Sheet Use the Load User Input Sheet command to avoid retyping data if it is lost in a run that doesn’t converge. After you retrieve a case, the values in worksheets are updated. If you have entered data on the Parameter or Simulation worksheet, this data is overwritten. To retrieve this overwritten data, execute the Load User Input Sheet command.

To Load User Input Sheets: 

Click the Load User Data (

) button.

If you need to save the data you have entered, execute the Save User Input Sheet command. These can also be activated by buttons on the toolbar:

Save User Input Sheet Use the Save User Input Sheet button to save your data entry sheets for later retrieval if the data entry is lost in a run that doesn’t converge. If you need to reload a case, your data entry on Parameter or Simulation worksheets will be overwritten. To retrieve your data entry, use the Load User Input Sheet command.


37

Setup Cases Submenu The third submenu on the AspenPlusHYC drop-down menu is Setup Cases. This submenu is unavailable until you successfully connect the workbook to the calculation engine as explained (fm).

The Setup Cases submenu contains six commands: Use this command

To

Case Study

Set Up Case Studies

Optimization

Set Up Optimization Calculations

Vectors

Set Up Lp Vectors

Profit 1

Set Up Profit Function Number 1 For an Optimization Case

Profit 2


Profit 3


Run Cases Submenu The fourth submenu on the AspenPlusHYC drop-down menu is Run Cases. This submenu is unavailable until you successfully connect the workbook to the calculation engine.

38


Tools Submenu The fifth submenu on the AspenPlusHYC drop-down menu is Tools. This submenu is unavailable until you successfully connect the workbook to the calculation engine.

Development Tools Submenu The sixth submenu on the AspenPlusHYC drop-down menu is Development Tools. This submenu is unavailable until you successfully connect the workbook to the calculation engine. Development tools are reserved for expert users and their use is not covered here.


39

Help Submenu The seventh submenu on the AspenPlusHYC drop-down menu is Help.

Use the Help submenu to: 

Invoke the on-line help.



View details about this version of Aspen Plus Hydrocracker.

Exit Aspen Plus Hydrocracker The eighth submenu on the AspenPlusHYC drop-down menu is Exit Aspen Plus Hydrocracker.

40


The best way to exit is to use this menu item, 1

Click AspenPlusHYC | Exit Aspen Plus Hydrocracker.

A dialog box appears, asking for confirmation. 2

Click the OK button to proceed with exiting.

You are asked whether you want to save the changes made to the Excel workbook. 3

Click the Yes button to save your changes and exit. -orClick the No button to abandon your changes and exit -orClick the Cancel button to abort the exiting operation.

If you click either the Yes or No button: 

The workbook closes.



The AspenPlusHYC menu disappears.



The AspenPlusHYC toolbar is hidden.


41

3 Working With The Equation-Oriented Solver

Introduction to the EquationOriented Solver When you click on the solve button, Aspen Plus Hydrocracker submits the mathematical formulation of the problem to the Aspen Plus open equation model based simulation solver. If the solution is successful: 

the kernel command window closes



the results of the solution are returned to the Excel GUI



the status indicators will change to Ready and Converged

If the solver fails: 

the status indicators show Ready and Not Converged



you must perform some troubleshooting to determine the cause of the failure

The following topics provide information on the basics of the solver technology and error messages issued by the solver when certain types of errors occur.

Successive Quadratic Programming (SQP) The Aspen Plus Equation Oriented (EO) model based solver is a specific implementation of the general class of nonlinear optimization algorithms known as Successive Quadratic Programming (SQP), which perform the optimization by solving a sequence of quadratic programming subproblems. The general optimization problem that DMO solves can be expressed as follows:

Minimize f(x)

42


Subject to c(x) = 0 xmin  x  xmax Where: x  Rn

Vector of unknown variables

f(x)  R1

Objective function

c(x)  Rm

Vector of constraint equations

xmin  Rn

Vector of lower bounds on x

xmax  Rn

Vector of upper bounds on x

A simplified description of the EO model solver algorithm is outlined as follows: 1

Given an initial estimate of the solution vector, x0.

2

Set iteration counter, k = 0.

3

Evaluate derivative of the objective function, gradient, and the derivative of the constraints, Jacobian.

4

Initialize or update an approximation of the second derivative matrix, or Hessian, of the Lagrange function. The Lagrange function, f(x) +  ici, accounts for constraints through weighting factors i, often called Lagrange multipliers or shadow prices.

5

Solve a quadratic programming subproblem to determine a search direction, dk. In the quadratic programming subproblem, the objective function is replaced by a quadratic approximation, constraints are linearized, and bounds are included.

6

Check for convergence or failure. If the optimization convergence criteria are satisfied, or if the maximum number of allowed iterations, MAXITER, is reached, then end. Convergence is achieved when: 

Objective function gradient  OBJCVG



Scaled or unscaled constraint residuals  RESCVG

7

Perform a one-dimensional search to determine a search step k so that xk+kdk is a better approximation of the solution as measured by a line search or merit function. The reduction of merit function requirement is sometimes relaxed to achieve a full correction step.

8

Update iteration counter, k = k + 1, and loop back to step 3.

Changing EO Solver Parameters You can change parameters for the solver can be changed with script commands. Enter commands at the kernel command prompt or on the EB scripts sheet in the Excel GUI. The script language for a parameter change is:

DMO.parameter = value


43

The parameters are discussed in the following sections. As an example, the following commands:

DMO.MAXITER = 10 DMO.RESCVG = 1.0D-5 change the maximum number of iterations to 10 and the residual convergence tolerance to 1.0d- 5. This input would apply for all modes.

Basic EO Solver Parameters Here are the most commonly used DMO parameters for Aspen Plus Hydrocracker: MAXITER

Maximum number of SQP iterations allowed (default = 50).

MINITER

Minimum number of SQP iterations allowed (default = 0).

CREEPFLAG Flag for the creep mode. This mode makes the optimizer moves more conservative. It is very helpful when the problem diverges. No (default) Yes CREEPITER Number of iterations to perform creep mode (default = 10). CREEPSIZE Creep mode step size. This is the fraction of the full step to be taken when in creep mode (default = 0.1). RESCVG

Residual convergence tolerance (default = 1.0D-6).

OBJCVG

Objective function convergence tolerance (default = 1.0D-6).

EO Solver Output to the Command Window During each solution, the following iteration log is sent to the command window:

44




Iteration is the count of SQP iterations (QP subproblems) performed by the solver. There is one line of output for each normal iteration of the solver. Abnormal iterations may have additional lines for error or information messages.



Residual Convergence Function indicates the solver’s progress towards solution, in terms of feasibility of the residuals. The problem does not converge until this measure gets below the DMO rescvg value defined in the EB script for that solution mode.



Objective Convergence Function is a measure of the solver’s progress towards solution in terms of optimality of the objective function. This is only meaningful in modes with degrees-of-freedom, which for Aspen Plus Hydrocracker is only the optimization mode. The problem does not converge until this measure gets below the DMO objcvg value defined in the EB script for that solution mode.



Objective Function Value refers to the Jacobian of the objective function.



Nonlinearity Ratio is a measure of the nonlinearity of the problem. The closer the value is to one, the more linear the problem. A negative value indicates that the problem behaved in the opposite of what was expected. Near the solution, as the step sizes become small, this value becomes close to one. There are two nonlinearity ratios:  Overall  Model



Worst Model is the model which has the worst non-linear ratio.

The last section of the output shows the execution times for the various parts of the problem. In this example, we can see that convergence was achieved when the residual and objective convergence functions were less than their respective tolerances at iteration 3. From this output, we also see that there have been no line searches. Thus the step size for each iteration is one. When a line search is performed for an iteration, a message similar to the one below appears: ==> Step taken 3.26D-01


45

If the solver has to line search continually and the step size gets very small (less than 1.0D-2), most likely the solution is trying to move very far from the starting point or some of the specified values are nearly infeasible.

EO Solver Log Files Aspen Plus Hydrocracker outputs DMO solver information to two log files: 

ATSLV.



ATACT.

These files reside in the working directory you defined in the startup menu box (fm). The ATACT file is similar to the ATSLV file, but lists all the problem variables and independent variables, whereas the ATSLV file does not. The ATSLV file is typically more useful and is described in more detail below.

ATSLV File Problem Information At the top of the ATSLV file, a summary of the problem is printed. This shows the size of the problem and the values of some important parameters.

46


ATSLV Details Basic Iteration Information At each iteration, the following header is printed:

This shows the iteration number and the value of the objective function.

Largest Unscaled Residuals This section shows the largest unscaled residuals. A similar section shows the largest scaled residuals. This section is particularly helpful when the solver has trouble closing all the residuals because it will point to the largest.

Constrained Variables This section shows the variables that lie on their bounds. This is only meaningful in a mode with degrees of freedom (optimization for Aspen Plus Hydrocracker). The output shows the variable number, which bound is active, the variable name, the current value and the shadow price. The shadow price is also known as the Lagrange multiplier. This is the derivative of the objective function with respect to the value of the constraint and represents the cost for the constraint.

The shadow price is based on the value of the objective function that is seen by DMO. That means the shadow price is in SI units (such as $/sec) and is affected by any scaling. This is true even if you declare the units to be something other than SI (such as $/HR).


47

Consider this example. We have a tower with a composition constraint, expressed as a mole fraction of a component. The following table shows the results of two optimization runs at two different values of the composition constraint: Value

Objective

Shadow Price

0.0002

2.853

432.924

0.0003

2.893

258.664

The large change in the shadow price indicates that the effect of the composition on the objective function is very nonlinear. We can manually estimate the average shadow price in this region by a finite difference method: Price = Obj/x = ( 2.893-2.853 ) / ( 0.0003 - 0.0002 ) = 400.00 $/sec/mole fraction This value lies between the two prices. If the objective function had a scale factor of 100, we would get the following: Value

Objective

Shadow Price

0.0002

285.4

43290.7

0.0003

289.3

25860.2

We would have to remember to unscale the shadow price by dividing by 100.

General Iteration Information This section appears after the residual output:

The iteration status shows the exit condition of that iteration. Normal indicates a normal successful iteration. Warning indicates a successful iteration despite some solver difficulties. Error indicates a failure. Solved indicates the final iteration of a successfully solved problem. The degrees of freedom is the number of declared independent variables in the problem. The constrained variables are those at bounds in the QP subproblem. The current degrees of freedom is the degrees of freedom less the constrained variables. This is the true degrees of freedom for the problem. A highly constrained solution is one that has very few current degrees of freedom.

48


The number of function and Jacobian evaluations is an accumulative count and generally matches the number of iterations. The objective function convergence function is the norm of the Jacobian for the objective function. At the solution, this value should be near zero. The residual convergence function is the sum of the scaled residuals. At the solution, this value should be near zero.

Nonlinearity Ratios This section shows the nonlinearity ratio of the worst block, the objective function, and the worst equations. The criterion is the accuracy of the predicted change in the equation. If the function is linear, then the new value would match the predicted value and the nonlinearity ratio would be one. A value of the ratio other than one indicates some degree of nonlinearity. A negative value indicates that the function value moved in the opposite of the expected direction. Large negative values could indicate a discontinuity or bad derivative. This section also shows the step size for the iteration.

Usage Notes Usage Notes-General This section describes some usage notes and troubleshooting tips to improve the performance of the solver and to help diagnose common problems. The topics in this section are: 

Dealing With Infeasible Solutions



Scaling



Dealing With Singularities



Notes on Variable Bounding



Run-Time Intervention


49

Bounds Aspen Plus lets you bound every variable in the problem as shown below:

Xl < X < Xu The step bound of an independent variable defines how much the value of the variable can be changed in a single optimization run. The step bound is used along with the initial value, lower bound, and upper bound to compute the actual bounds to be used in the run:

Xl = max(X - |Xstep|, Xlower) Xu = min(X + |Xstep|, Xupper) You should define upper and lower bounds for all independent variables. You can also define the step bounds for independent variables. Most of the dependent variables in the Hydrocracker model have very wide bounds, such as –1.E20 for lower bound and 1.E20 for upper bound. However, some dependent variables have physical meaning. You should set up appropriate bounds for them to prevent the solution from getting into infeasible operating conditions. For example, there is a metallurgic limit on regenerator cyclone temperature. Hence, an upper bound should be set for this variable. Only those constrained dependent variables must be defined when setting up an optimization case in Hydrocracker model. In general, it is not recommended to heavily bound an optimization problem for reasons that are both practical and algorithmic. Bounds on independent variables are recommended in order to avoid unbounded problems. All other bounds should be used only if they are absolutely necessary. The optimization engine for Hydrocracker model is the DMO solver.

Independent Variables Independent variables are variables whose values can be changed independently, for example, the feed rate in the Hydrocracker unit. The optimizer can vary the values of independent variables to push the values of the objective function in the defined direction (maximize profit or minimize cost) until some bounds are reached. Each independent variable accounts for a degree of freedom. The number of degrees of freedom is equal to the number of independent variables in an optimization run if no independent variable is at its bound. You can impose upper and lower bounds on independent variables to prevent the final solution from deviating too far away from the starting point. You can also impose step bounds on independent variables.

Dealing With Infeasible Solutions These often occur during optimization cases where it is not possible to simultaneously solve all the equations while respecting all the variable bounds. This doesn't happen in simulation cases because DMO ignores bounds in simulation cases. If you solve a simulation case that violates a bound, then the optimization case will start at an infeasible point. In this case, the following is printed in the OUT file:

50


This says that this variable's value had to be adjusted to respect the bound. When the optimization proceeds and there is no feasible solution for the equality constraints, the screen output might look like this: Residual

Objective

Convergence Convergence Iteration

Function

Function

Objective

Overall

Model

Function

Nonlinearity

Worst

Nonlinearity

Value

Ratio

Model

Ratio

--------- ----------- ----------- ---------- ------------ ------- -----------Warning ...

QP slack variable =

2.29070D-01

Warning ...

QP slack variable =

2.29070D-01

0

9.312D-04

4.809D-03 -2.779D+00

Warning ...

QP slack variable =

1.80624D-01

Warning ...

QP slack variable =

1.80624D-01

1 Warning ... Warning ... 2

5.244D-04

4.667D-02 -2.792D+00

QP slack variable =

1.44771D-01

QP slack variable =

1.44771D-01

1.552D-02

5.479D-02 -2.922D+00

Warning ...

QP slack variable =

6.09502D-01

Warning ...

QP slack variable =

6.09502D-01

3

3.853D-02

2.379D-03 -3.083D+00

Warning ...

QP slack variable =

1.87163D-01

Warning ...

QP slack variable =

1.87163D-01

4

1.496D-02

1.040D-02 -3.075D+00

Warning ...

QP slack variable =

3.18508D-01

Warning ...

QP slack variable =

3.18508D-01

9.968D-01 C2S

-2.834D-01

2.900D-01 C2S

-1.846D+02

-7.475D-01 C2S

-1.540D+01

9.908D-01 C2S

9.914D-01

8.346D-01 C2S

6.012D-01

+---------------------- ERROR ----------------------+

Error return from [DMO] system subroutine DMOQPS because the problem has NO FEASIBLE SOLUTION.

Action : Check the bounds that are set on variables to insure consistency. Check the .ACT file for information on initial infeasibilities.

+---------------------------------------------------+

Error return, [DMO] System Status Information =


5

51

Optimization Timing Statistics

Time

Percent

================================

========

MODEL computations

1.32 secs

31.10 %

DMO computations

0.91 secs

21.28 %

Miscellaneous

2.03 secs

47.61 %

-------------------------------Total Optimization Time

--------4.26 secs

=======

------100.00 %

Updating Plex Problem failed to converge

Note the messages from the QP indicating an invalid value for a slack variable. To solve this problem, you need to be aware of the initial message indicating that the initial value of a variable violated its bound. In this case, C2S.SPC.REFL_RATIO_MASS is causing the problems. Unfortunately, the OUT file does not list this variable as constrained, since it could never solve the QP successfully.

Scaling Generally, it is not necessary to scale your equations or variables beyond what is done by default in the models. However, it may be more efficient to scale your objective function. A good rule of thumb is to scale the objective function so that its value is on the order of 10 to 1000. The scaling of the objective function plays an important role since it affects the overall convergence behavior. This is particularly important in cases where there is a large change between the original value of the objective and the expected optimum.

Dealing With Singularities Singularities often occur when the model is moved into a region where the equations are not well defined. The most common example of this is when a stream flow becomes too small. If singularities exist, they are usually detected at the start of the problem. In this case, some information is written to the OUT file and this can help locate the cause of the problem. In general, you should prevent stream flows from going near zero by placing nonzero lower bounds on the flow (e.g., 10 kg/hr). This is especially important on streams from flow splitters or feed streams whose total flow is being manipulated. In the case of a singularity the following message will be displayed:

52


The OUT file contains information on the possible cause of the singularity in the following manner:

Sometimes, singularities are simply caused by the optimization being too aggressive. This moves the models into a region where the equations are not well defined. To make the optimization more robust, DMO has a creep mode. This mode simply causes smaller steps to be taken for a specified number of iterations. To use this mode, you can enter the following script command:

DMO.CREEPFLAG = 1 This turns on the creep mode. When active, the following message is displayed at each iteration: ==> Step taken 1.00D-01

By default, this will operate for 10 iterations with a step size of 0.1. You can change these values with the commands:

DMO.CREEPITER = 20 DMO.CREEPSIZE = 0.5 In this example, we change the number of creep iterations to 20 and the step size to 0.5.


53

Notes on Variable Bounding Remember that by default DMO does not respect bounds during the solution of a SIM or PAR case. The user, however, has the capability to impose bounds in a square case by using a different line search parameter. The use of this mode is recommended only in cases where there are truly multiple solutions to a model (for example, the cubic equation) and you want to use a bound to eliminate an unwanted one. 

To use this mode, enter the following script command:

DMO.LINESEARCH = 4 In general it is not recommended to heavily bound an optimization problem for reasons that are both practical and algorithmic. Bounds on independent variables are recommended in order to avoid unbounded problems. All other bounds should be used only if they are absolutely necessary. Finally, redundant bounds should be avoided.

Run-Time Intervention During long runs, you can change the behavior of the DMO solver by clicking one of the three buttons at the bottom of the command window. Your selection takes effect at the start of the next DMO iteration. The three buttons are:

54

Button

Action

ABORT

Stops the solver

CLOSE

Fixes all the independent variables at their current values and closes the residuals

NOCREEP

Takes DMO out of creep mode


4 Model Parameterization

Introduction To provide a better understanding of the Aspen Plus Hydrocracker/Hydrotreater model, this section presents a general description and discussion of the model.

Flow Diagram Sheet Product Properties You can vary product cut-points. The table below shows a list of product properties predicted by AspenPlusHYC model. Product Stream

Properties

H2S

Mass flow, mole flow

NH3

Mass flow, mole flow

H2

Total consumption (mass and moles)

C1

C1 mass flow, mole flow, mass fraction C2 mass flow, mole flow, mass fraction

C2

C1 mass flow, mole flow, mass fraction C2 mass flow, mole flow, mass fraction C3 mass flow, mole flow, mass fraction H2S mass flow, mole flow, mass fraction H2 mass flow, mole flow, mass fraction

C3

C2 mass flow, mole flow, mass fraction C3 mass flow, mole flow, mass fraction C4 mass flow, mole flow, mass fraction

C4

C3 mass flow, mole flow, mass fraction C4 mass flow, mole flow, mass fraction C5 mass flow, mole flow, mass fraction C4 iso/normal ratio


55

Product Stream

Properties

Light naphtha

C4 mass fraction TBP distillation API gravity Specific gravity PIANO Total Sulfur Total Nitrogen RON/MON

Heavy Naphtha

TBP distillation API gravity Specific gravity PIANO Total Sulfur Total Nitrogen RON/MON

Distillate

TBP distillation API gravity Specific gravity PIANO Total Sulfur Total Nitrogen Basic Nitrogen Smoke point Pour Point Freeze point

Bottoms

TBP distillation API gravity Specific gravity PIANO Total Sulfur Total Nitrogen Basic Nitrogen Cetane Index Viscosity

For recycle hydrocracking, 

The maximum-naphtha base case recycles the 400oF-plus material.



The maximum-distillate base case recycles the 700oF-plus material.

Distillation overlap is calibrated with plant data. The following yields and product properties are provided. Recycle Gas Scrubber is simplified as a component splitter, RGSPLIT, a standard Aspen Plus component splitter block (SEP). Note: Scrubbing efficiency can be calibrated with plant data. If there is no recycle gas scrubber,no H2S is removed by this block.

56


The quench-distribution system is modeled by a Standard Aspen Plus splitter block (FSPLIT). Mixer blocks are used for quenches. You can specify: 

Heat loss.



Temperature.



Pressure.



Pressure drop.

Quench valve characteristics are not modeled, but you can specify upper and lower limits on quench flow.

Model View and Specification Through the Flow The main AspenPlusHYC model data access is provided by buttons on this sheet. A summary of the general instructions for all buttons given below: 

Buttons R1, R2, HTR. When you click one of these buttons, you are taken to the Process Detail sheet, which displays a summary of the current running conditions.



Buttons Feeds takes you to the Process Detail sheet to display feed properties of the combined feed and input sheet for each individual feed.



Button Yields takes you to the Process Detail sheet to display the volume and mass yields of product. In addition, it displays the product properties of each product.



Button Process Overview takes you to the Process Overview sheet, in which general information for AspenPlusHYC model is presented.



Button Reaction Profile takes you to the Reaction Profile spreadsheet. On this sheet a set of diagrams presents present the temperature, sulfur, nitrogen and aromatic contents the two beds in the two reactors.



Button Specification Options pops up a dialog to let you select a mode you want to run. You can select one, then click the Select button. The specification will be set properly for the case running. For each running mode, it will show later.


57

Note: All data sheets showing the value and specification are only for display. Any changes will not affect the model running in this version.

Model Specifications Strategy for Process Specification For a number of purposes of AspenPlusHYC model running, the user need provides sufficient process information to let the model can be tuned to match the process situation. The AspenPlusHYC can be tuned in a number of ways. In general, the following are required: 

Feed and product properties



Operating conditions



Selected mechanical data

All samples must be time-stamped. All samples must be taken within the same 4-8 hour period. Uncompressed, hour-average process data (all relevant tags from the DCS, if possible) must be provided for the period during which samples are taken.

Feed and Product Analysis Your specific needs determine whether the feed and product analytical requirements are simple or complex. AspenTech has a library of detailed analytical data and component distributions (fingerprints) for several feed/product combinations. Feed types include:

58




Light and heavy straight-run vacuum gas oils.



Light and heavy coker gas oils.



Light and heavy FCC cycle oils.



Pre-processed synthetic crude.

When a client’s feed resembles a feed in the AspenTech database, the Feed Adjust block in Aspen Plus Hydrocracker can obtain a good fit by skewing the distribution of components in the base feed to minimize differences between measured and calculated bulk properties. This is the usual starting point if the model is to be used offline for: 

What-if studies.



Generation of LP shift vectors.



Certain design studies.

The following bulk inspection properties are used for model calibration, Property

Required/Optional

API gravity

Required

D2887 distillation

Required

Refractive index

Optional (recommended)

Viscosity @210 F

Optional (recommended)

Bromine number

Required

Total sulfur

Required

Total and basic nitrogen

Required

The feed refractive index and viscosity are optional. If you do not provide the refractive index and viscosity, the feed adjustor model calculates them based on the API gravity and distillation. The refractive index, viscosity, API gravity, and distillation are then used to calculate: 

The CA (carbon on the aromatic rings).



The CN (naphthene ring).



CP (paraffins).

The method used for calculating CA, CN, and CP is the NDM method. To get a better result for characterization of the feed, AspenTech recommends the following methods to determine the aromatic/naphthene/paraffin breakdown for each feed, Method

Description

D1319 Fluorescent Indicator Adsorption (FIA).

provides Total Aromatics in vol%.

NMR method

provides Carbon on aromatic rings

UV method

provides wt% of MONO, DI, TRI and Tetra aromatics

HPLC

high performance liquid chromatography


59

Notes: If higher accuracy is required, for example, for an online closed loop optimization project, or if the plant frequently processes feeds for which AspenTech does not have a similar fingerprint, detailed analyses including GC/MS, HPLC and NMR may be recommended. To tune the Catalyst Deactivation block, you may be able to use historical data. However, if adequate historical data are not available, at least two sets of test-run data obtained at least three (fm) To increase model accuracy – and to be confident that feedstock effects can be differentiated from catalyst deactivation effects – individual blend stocks should be analyzed individually. Also, analyses should be obtained for typical composite feeds. The adjustment of the individual feeds creates detailed lump compositions for each individual feed. This allows a more accurate representation of the effects of individual blend stocks on hydrocracker performance. If you need detailed analytical data, the table below provides a list of the feed and product inspections needed for the creation of a new fingerprint. Aspen Hydrocracker Feed and Product Analysis - Check List

Tests (Note 1) Required for Configuration Prep Distillation into 950-,950+ (2) API Gravity Sim Dist High Temp Sim Dist NOISE HPLC PONA (FIA) HC Type, wt% C & H Content, Wt% Total Sulfur Content, wt% Total Nitrogen, ppmw Basic Nitrogen, ppmw Bromine Number H & C13 NMR Other Required Data Sim Dist C1 - C3 C6- GC (3) GC PIANO (4) Optional Data per Client Interest Metals (Ni,V,Fe,Cu,Na) ppmw Refractive Index @ 20 DegC Aniline Point Viscosity, cst @ 100 DegF Viscosity, cst @ 210 DegF Carbon Residue, wt% Cloud Point Pour Point Smoke Point Freeze Point Cetane Index ASTM Distillation RON MON RVP

60

Method

Total

Vac Dist D287 D2887 HTSD GC/MS HPLC D1319 D4124 D5291 D4294 (?) UOP269 D1159 D5292

x x

Feeds 950-

x x

x

950+

Unconverted Oil Total 950950+

x

x x

x

x

x

x x x x x

x x

x

Diesel

Kerosine

Naphtha

x x

x x

x x

x

x

x

x x

x

x

x x x x

x x x x x

x x x x

x x x x x

x x x x

x x x x

x x x x

x

x

x

x

x

x

D3710 GC/TC/FID GC/FID GC/FID ICP D1747 D611 D445 D445 D4530 D2500 D97

x x

x

x x x x x x

x x x x x x x x

x x x x

x D86 D2699 D2700 D5191

Fuel Gas

x x

x x x x

C3+C4

x

x x x x

x x x x


Note: Estimated costs for analyses may be obtained by contacting AspenTech. Heavy feeds must be separated into 950- and 950+ fractions, because each fraction requires different methods. The requirements for configuration are used to calculate feed and product compositions for 97-lump kinetics. Lighter feeds (<10% 950+) do not need this separation. They can be characterized with the 950- test set. Light feeds (950-) require SimDist NOISE and NMR. Heavy feeds (950+) require HT SimDist, HPLC and NMR. Flue gas and C6- GC are usually available from the client's lab. If multiple naphtha cuts are present, each cut must be analyzed. If multiple C3 and C4 product streams present, we need C6-GC for each stream. Sample

Amount Needed

Light feeds without prep distillation

1 quart

Heavy feeds with prep distillation

2 quarts

LCO

1 quart

Bottoms with prep distillation

2 quarts

Naphtha

1 quart minimum. (2 quarts if octanes requested. Need 1 quart for RON and 1 quart for MON engine tests.)

If the catalysts are different for Reactor 1 and Reactor 2 such as the case for hydrocracker, the reactor effluent after Reactor 1 should be analyzed to properly tune the catalyst activity for hydrotreating and hydrocracking catalyst. Some product property requirements are optional (determined by your needs). The following list provides some examples: 

C6- GCFor All Light Materials (LN, HN Light Ends) -- required



Distillation -- required



API Gravity -- required



Total sulfur and nitrogen for all liquid products -- required



RVP for the lightest naphtha product -- optional



RON/MON for all naphthas -- optional



For Distillates  Freeze Point -- optional  Smoke Point -- optional  Cloud Point -- optional  Pour Point -- optional  Sulfur/Nitrogen – required  Viscosity -- required

Material Balance and Yields Calculation Overall Plant Material Balance Sheet for Tuning Runs is required. The flow rate for all the feed and product streams should be provided. The product streams include the purge gas, LPS overhead gas, pre-stripper overhead gas, LPG and, all the liquid products. In order to do hydrogen and light gas yields


61

balance, the gas composition is required for all the product gas streams, LPG and makeup H2 rate. In order to verify the model, the gas composition for the recycle gas (make up H2 + HPS overhead vapor) is also needed. The following tables list the information needed for material balance.

Liquid Feeds and Products Feed

Bottom

Diesel

Kerosene

Naphtha

Flowrate

X

X

X

X

X

Density

X

X

X

X

X X

Liquid GC (PINA)

Gas Streams H2 Makeup Gas

HPS overhead vapor

Recycle Gas Stream

Flow Rate

X

X

X

Gas GC, H2-C5

X

X

X

Purge Gas

X

LPS overhead

Stripper or reflash overhead

X

X

X

X

LPG

X

X

Liquid GC

A good material balance is crucial before model tuning. AspenTech provides an Excel spreadsheet (Hydrocracker_ANALYZE.XLS) to analyze the plant data. The input requirements for the analyze spreadsheet include: 

Feed - flow rate, specific gravity, Molecular weight, Distillation (TBP or D2887) and Chemical composition (H, C, S and N) of all the feeds.



Flow rate and composition of Makeup H2.



Flow rate and composition of all the gas products such as purge gas, low pressure overhead gas, prefrac tower overhead gas.



Flow rate and composition of LPG



Feed - flow rate, specific gravity, Molecular weight, Distillation (TBP or D2887) and Chemical composition (H, C, S and N) of all the liquid products, such as naphtha, jet, distillate and bottoms.

The spreadsheet can estimate the molecular weight and chemical composition (H/C ratio) if the data is missing. However, its correlation for H/C ratio may not be good enough for a good hydrogen balance. The spreadsheet provides conversion between D86 distillation and D2887 distillation on the Conversion worksheet. But D86 data is usually not sharp enough for a good tuning of the simplified distillation tower. D2887 should be obtained for all the liquid products. In the Material Bal worksheet, it calculates the mass balance, sulfur balance, nitrogen balance, hydrogen balance, and carbon balance. The difference between product sulfur and feed sulfur is assumed to be converted to H2S

62


and absorbed by the H2S scrubber. The difference between product nitrogen and feed nitrogen is assumed to be converted to NH3 and dissolved in water wash. The mass balance, hydrogen balance, and carbon balance should be within 1%. The spreadsheet also calculates: 

Chemical hydrogen consumption.



C1-C5 light gas yields.



Standard cut yields (C6-221C, 221C-321C, 321-366C, 366C-510C and 510C+).



Volume and mass yields for liquid product.

This information is needed to tune the model. The data should be input into the ParamData worksheet in Section 2, Measurements. The LPS and pre-fraction tower overhead are modeled as component splitter models. The split fraction for each product stream is calculated on the Light Gas worksheet. There are three sets of split factors for streams: 

LPS overhead vapor.



Prefrac tower overhead vapor.



Prefrac tower LPG.

The split factor is input into the ParamData worksheet of the Aspen Plus Hydrocracker spreadsheet, or using the script file for variables :

PRODSP.LPS.BLK.LPS-OH_SPLIT_XXX = PRODSP.PREFRAC.BLK.OVGAS_SPLIT_XXX = PRODSP.PREFRAC.BLK.TOGASPLT_SPLIT_XXX =

Simplified Separation Model A simplified separation model is applied to model the fractionation section. Due to the variation of the products specification, operation condition, and flow sheeting, the separation status varies significantly from all specific plants. In this simplified model, there are a set of variables which can be tuned to make the products match by their quality and yields. The concept behind simplified fractionator model is that the logarithm of the ratio of a component (flow not fraction) in distillate to bottoms (Ln(Di/Bi)) when plotted over temperature (the boiling point, TBi) of that component yields a straight line (more or less a linear relationship). Figure 1 below illustrates this concept. The model is a collection of interpolation and calculator models to calculate the split fraction of a certain component based on the effective cut point and the component’s normal boiling point. The model interpolates a value of Ln(Di/Bi) for a given TBi. Usually, the model is configured with 3 points representing the two straight lines around the effective cut point (where Ln(Di/Bi) = 0, meaning a 50 % split of the component between top and bottom streams). The effective cut point is close to the 90% of the top product.


63

There are three coefficients for each cut to determine the cut point and separation efficiency: 

The effect cut point (ECP).



The slope of the upper line (SITOP).



the slope of the bottom line (SIBOT).

The three coefficients are calculated in the SEP worksheet. For example, a standard main fractionator of the hydrocracker that has four products (naphtha, jet, distillate and bottom) will have three sets of ECP, SITOP, and SIBOT. The three factors, ECP, SITOP and SIBOT, will be determined by the Aspen Plus Hydrocracker analyze spreadsheet. When the cut points of the product change, the model can recalculate the product flow rate and properties by changing the effective cut points and keep the slope constants. Figure 2 shows the effective cut points for the Jet/naphtha is 135 C. The upper and bottom slopes are 0.16. You can change the value in the highlighted cell to best fit the data points in the graph.

64


Figure 1 D

F

B

Ln (Di/Bi)

y2

SI E f f e c t iv e C u t P o in t 0

y1 x2

TBi

x0

x1

x0

x1

Di/Fi

1

0 x2

TBi

Principle of Simplified Distillation Column


65

Figure 2

Key Operating Data Aspen Plus Hydrocracker/Hydrotreater is a process flowsheet model. Each catalyst bed is treated as a separate reactor so that 

reaction kinetics



pressure drop



catalyst deactivation rates

can be configured and tuned independently. Standard equipment models are used for everything else. Therefore, to enable us to configure AspenPlusHYC for a given installation, we need the following information:

For Standard (non-Reactor) Equipment: 

Heaters – temperatures, pressure drop and constraints



Heat exchangers – temperatures and pressure drop



Compressors – temperatures, pressure profiles and performance curves (optional)



Flash drums (high pressure and low pressure separator) – pressures and temperatures



Wash water injection system – constraints



Main fractionator / gas plant – dimensions, operation conditions (rigorous fractionator models only)

For the Reactors

66



Bed dimensions (internal diameter, length)



Quench flows



Pressure drop



Inlet and outlet temperature


Catalyst Properties/Data 

Type



History (fresh, once-regenerated, and so on)



Loaded bulk density



Loading diagram for each bed (for example: 3 inches of ¼" quartz balls on top, then 3 feet of fresh 1/8 inch extrudates, then 6 feet of regenerated 1/16 inch extrudates.)



Total catalyst inventory (by type)



Deactivation data (temperature, feed properties, operating mode vs. time) from previous runs

Unit Mechanical Data (Initial tuning only) 

PFDs and P&IDs

Pricing Information (for Optimization Cases) 

Prices for all blend stocks, all products, and all utilities



Incremental value of key product properties (sulfur, octane, cetane, etc.)

Recycle Stream Data The recycle oil stream from the bottom of the main fractionator is another available specification. The recycle oil can go to: 

The inlet of Reactor 1.

-or

The inlet of Reactor 2.

The standard specification for recycle oil is the volume flow rate, which is set as 0.1 {BBL/DAY}. Do not set the recycle stream to zero.

Running a Parameterization Case Once all data has been entered, you can run a parameterization case.

To Run A Parameterization Case: 1

On the AspenPlusHYC toolbar, in the combo box, select the Parameter option.


67

2

Click the Play (

) button.

The Update Profit Reports? window appears.

3

Click the Yes button to: 

Update the initial values in these sheets:  Profit 1 Report  Profit 2 Report  Profit 3 Report



Send the new data to the model and solve to this new set of conditions.

-orClick the No button to: 

Send the new data to the model and solve to this new set of conditions without updating the initial values on the Profit sheets.

Note: You can also run a Parameter case by clicking AspenPlusHYC | Run Cases | Parameter Case on the main menu. If you choose this method, jump to Step 3 above when the Update Profit Reports? window appears.. When the parameterization has been started, the data entered into the spreadsheet is sent to the AspenPlusHYC model and the command line dialog box appears as follows:

68


The only two options on this dialog available to you are: Click this Button

to

Abort

stop execution of the solution to the model

No Creep

remove the creep size constraint on the problem

The Close Residuals button has no effect for a Parameterization run. It is relevant only when running an Optimization case. When the solution is completed: 

The command line dialog disappears.



The data is retrieved into the spreadsheet.

Typical execution time is between one and three minutes depending on the number of creep steps and how close to a solution the model starts. It is recommended that once the parameter case has been run and successfully solved, and the results have been carefully reviewed on the Analysis sheet, the data should be saved, Saving the data will allow you to load this data at any time so that the starting point for subsequent solutions will be a valid parameterization.

Reconciliation Cases You use the Reconciliation Case option to calibrate the Hydrocracker model. A reconciliation case is essentially a parameterization case with degrees of freedom. This is necessary because the Hydrocracker model has more measurements than parameters to adjust. Therefore, you input plant data (for example, yields, product properties, H2 consumption, and so on.) and then run a Reconciliation case to tune the model.

Simple Parameterization For simple parameterization (for example, just tuning to match sulfur from a Hydrocracker) follow these steps: 1

Enter all data on the ParamData page.

2

Set the desired sulfur value to MEAS (for example, bottoms sulfur or total sulfur out of the last reactor).

3

Set the tuning factor for sulfur (kinetic pathway) to PARAM.

4

Run a PARAMETER case from the menu or from the command bar.

More Detailed Parameterization 1

Enter all data on ParamData page

2

Set the desired biases and sigmas for the Recon Objective Function (Section 4 on the ParamData page).



The sigmas correspond to the weighting factors. The higher the sigma is, the less effect that variable will have on the objective function.


69



These sigmas and biases correspond to typical measurements we will want to match:  Temperature rise for each bed  Volume flow rate for each liquid product  Weight percent yield for gas yield by carbon number (C1, C2, C3, and C4)  Bottoms nitrogen and sulfur content  API gravity for each liquid product  Quench gas flow to each bed  Flow rates for LPS overhead and LPG  Purge flow  Hydrogen makeup flow  Hydrogen consumption

Note: Aspen Plus Hydrocracker comes with a pre-configured Recon Objective Function that you can use. This objective function is configured with default values for the biases and the sigmas. However, if you are an advanced user, you can set up your own objective functions, but this requires modifying the interface somewhat. 3

On the main menu, select AspenPlusHYC | Run Cases | Reconciliation Case.

Note: Tuning factors are reported in Section 3 of the ParamData page.

70


5 Simulation

Introduction to Simulation Once the model is satisfactorily tuned to match plant data, it can be used to predict how changes in the feed rates, feed types and composition, and operational conditions affect yields and product properties. Typically, you enter data into the cells highlighted in blue for each section. Note: These highlighted cells are dependent upon the options selected.

Aspen Plus Hydrocracker Simulation Strategy The Aspen Plus Hydrocracker/Hydrotreater model is an open equation-based model. The manipulated variables (set as CONST on the variable specification) can be interchanged with the calculated variables (set as CALC on the variable specification). Whether the variable is constant or calculated depends on both: 

The specification of the variables.

-and

The mode of the model.

Thus the model’s structure lets you configure the variable specifications to fix a set of specs and then have the model predict the other set of variables. You cannot change the values and/or specifications in the cells to apply to the model running. You change the model specification by: 

Invoking an EB script file.

-or

Issuing a command in the Command Line window by selecting AspenPlusHYC | Tools | Display Command Line.

Four simulation modes are provided by the Aspen Plus equation-oriented solver engine,

5 Simulation

71



Simulation mode



Parameter estimation



Reconciliation



Optimization

For instance, to run the model in a simulation mode, you would issue the following command in the Command Line window,

SET MODE = SIM Here is an example to illustrate how to do it. As in the Aspen Plus Hydrotreater model, the feed blend model lets you specify the feed flow for each feed to which it is applied. So, the flow rates for each feed are fixed, and the blended flow (total flow) is calculated with mixer models as shown below: Variable Name

Spec

FEED.AFEED1.BLK.VOLUME

CONST


CONST


CONST


CONST


CONST


CONST

And Variable Name

Spec

FEED.AFEED.BLK.VOLUME

CALC

For the model running, if you want to manipulate the total flow to fit the specification of your plant measurements, you can fix the variable FEED.AFEED.BLK.VOLUME by setting its spec as CONST and free FEED.AFEED1.BLK.VOLUME by setting it as CALC. This still keeps the model squared. For this change, you can issue the following lines to the command text box in the command line window,

FEED.AFEED1.BLK.VOLUME.SPEC = CALC FEED.AFEED.BLK.VOLUME.SPEC

= CONST

In the AspenPlus solver engine, there also are other specifications which can be specified for different models running. Some of them are: 

Constant



Calculated



Measured



Parameterized



Optimized

For the Measured and Parameterized specifications, whether the variable is a constant or is calculated depends on the running mode of the model.

72

5 Simulation

Commonly-Used Scripts in the EB Script Language Several script language commands are used with the Aspen Plus Hydrocracker model running from command line. 

Set up the mode to run

SET MODE = [SIM | PAR | REC | OPT] 

Display the current running mode. It will return the current running mode on the window’s display area.

ECHO &MODE 

Change the variable specification

(variable name}.SPEC = [ const | calc | meas | param | optim | recon ]{



Get the whole plant specification and save it to a file called "AHYC3.SCN". After saving this file, Then, you can edit the file saved as AHYC3.SCN on the disk

PRINT SPEC TO "AHYC3.SCN"

Aspen Plus Hydrocracker Variable Specifications Model CONST Specifications The Current Model Settings for the DEMO model are listed below. 

FEED Section



RXN Section - Operating



RXN Section - Kinetics



PRODSP Section

FEED Section In the feed section, most of the properties should be specified for model running. It includes, 

Individual feed volume flow rate. There are six pre-specified feeds provided in the feed section. The flow rates for each feed stream must be specified correctly. To turn off any individual stream for a specific plant, set the flow rate to zero. If you trust the total fresh feed to the reactor, the total feed flowrate can be swapped with one of the individual feeds.

Note: The calculated feed flow should be big enough to avoid a negative flow for this feed... 

5 Simulation

Base Feed fingerprint for each feed. Fingerprint data can be generated from NOISE (GC/MS) data analysis of individual feeds.

73



Feed Properties for each feed. Those properties include API, Sulfur, Total Nitrogen and basic nitrogen, Bromine number, Refractive index (20C) and Viscosity (210F)* and True boiling point or simulated distillation.

*Can be calculated from specific gravity and distillation

RXN Section - Operating The specification in the reaction section is usually presented in terms of process operating condition and reactor kinetics. With a specific running situation, some specs can be swapped with others. Therefore, the specifications are presented by CONST/CALC pairing information. 

Gas/Oil ratio to first reactor. Usually, reactor feed is performed with a given Gas/Oil ratio. So, either: 1

Free the recycle gas rate to Reactor 1.

RXN.GOR.BLK.R1_INLET_GO-RATIO.SPEC = CONST -or2

Free the H2 makeup gas rate and fix the Gas/Oil ratio.

RXN.QH2SPLIT.BLK.STRM_2_MOLES.SPEC = CALC -or-

RXN.H2MAKE1.BLK.MOLES.SPEC = CALC 

Fresh Feed inlet temperature.



Furnace outlet temp. To fix the outlet temperature of the furnace (actually the Aspen Plus heater block), you can free the duty of the block.



RXN.HTR.BLK.C_OUT_TEMP

= CONST

RXN.HTR.BLK.DUTY.SPEC

= CALC

Inlet temperature for each reactor beds. This is a general specification for reactor operation. In order to do so, the recycle gas quench flow to each reactor beds is free.

RXN.QR1B2.BLK.PROD_TEMP.SPEC = CONST RXN.QR2B1.BLK.PROD_TEMP.SPEC = CONST RXN.QR2B2.BLK.PROD_TEMP.SPEC = CONST RXN.QH2SPLIT.BLK.STRM_3_MOLES.SPEC = CALC RXN.QH2SPLIT.BLK.STRM_4_MOLES.SPEC = CALC RXN.QH2SPLIT.BLK.STRM_5_MOLES.SPEC = CALC 

High pressure receiver temperature. Put CALC on FINFAN heater block duty variable.

RXN.FINFAN.BLK.H_OUT_TEMP.SPEC = CONST RXN.FINFAN.BLK.DUTY.SPEC = CALC

74



Prod/Feed effluent heat exchanger hot side outlet temp



Recycle gas compressor outlet temperature

5 Simulation



High pressure receiver pressure. Reactor outlet pressure or recycle compressor outlet pressure is free.

RXN.HPS.BLK.PROD_PRES.SPEC = CONST RXN.R2B2.BLK.OUTLET_PRES.SPEC = CALC

or

RXN.RGCOMP.BLK.PROD_PRES.SEPC = CALC 

Pressure drop for each reactor bed, furnace, heat exchanger. These specifications allow the pressure propagation through the streams flow affecting the simulation results.



Make up Hydrogen composition.



Purge gas/total recycle gas ratio.

RXN.RGPURGE.BLK.STRM_3_RATIO.SPEC = CONST RXN.RGPURGE.BLK.STRM_3_MOLES.SPEC = CALC

RXN Section - Kinetics The Aspen Plus Hydrocracker reactor model uses the group activity factor to change the relative reaction rate of different reaction mechanism. The overall pre-exponential factor,

k total , is a combination of the three rate factors:.

k total  K global  K irgroup  K jrate Where:

K global

=

the global rate constants for each reaction bed

K irgroup

=

the reaction group activity factors

K jrate

=

the Individual rate

The tables below show the constant K-factors for each bed of the reactor model (here only R1B1 is presented)

RXN.R2B2.BLK.GLOBAL_ACTIVITY

Global activity variables for the first bed of the first reactor Activity variable

Value

Spec


CONST


CONST


CONST


CONST

Group activity variables for the first bed of the first reactor

5 Simulation

Activity variable

Value

Spec

R1B1.BLK.SAT_ACT

1

CONST

75

Activity variable

Value

Spec

R1B1.BLK.HSAT_ACT

1

CONST

R1B1.BLK.MSAT_ACT

1

CONST

R1B1.BLK.LSAT_ACT

1

CONST

R1B1.BLK.HDS_ACT

1.5141

CONST

R1B1.BLK.HHDS_ACT

1

CONST

R1B1.BLK.MHDS_ACT

0.75812

CONST

R1B1.BLK.LHDS_ACT

1

CONST

R1B1.BLK.HDN_ACT

1

CONST

R1B1.BLK.HHDN_ACT

1.6755

CONST

R1B1.BLK.LHDN_ACT

1

CONST

R1B1.BLK.PCR_ACT

0.2447

CONST

R1B1.BLK.HPCR_ACT

1

CONST

R1B1.BLK.MPCR_ACT

1

CONST

R1B1.BLK.LPCR_ACT

10

CONST

R1B1.BLK.ROP_ACT

0.1

CONST

R1B1.BLK.HROP_ACT

1

CONST

R1B1.BLK.MROP_ACT

1

CONST

R1B1.BLK.LROP_ACT

1

CONST

Constant Variables for the Individual Reaction Rate Variable Name RXN.R1B1.BLK.KREF_HDRG1

Value

Spec CONST

through RXN.R1B1.BLK.KREF_HDRG45

CONST

RXN.R1B1.BLK.KREF_PHCR1

CONST

through RXN.R1B1.BLK.KREF_PHCR12

CONST

RXN.R1B1.BLK.KREF_SHCR1

CONST


CONST


CONST

RXN.R1B1.BLK.KREF_ROP1

CONST

through RXN.R1B1.BLK.KREF_ROP39

CONST

RXN.R1B1.BLK.KREF_RDA1

CONST

through RXN.R1B1.BLK.KREF_RDA58

CONST

RXN.R1B1.BLK.KREF_HDS1

CONST

through RXN.R1B1.BLK.KREF_HDS15

CONST

RXN.R1B1.BLK.KREF_HDN1

CONST

through RXN.R1B1.BLK.KREF_HDN4

76

CONST

5 Simulation

PRODSP section CONST

CALC

PRODSP.JETTOWER.SPLTCALC.BLK.ECP

90% point of product Jet

The Distillate/Bottom cut points

PRODSP.ANJET.BLK.D86CRV_D86CRV_90

PRODSP.MFBTMS.SPLTCALC.BLK.ECP

90% of Heavy Naphtha

The heavy naphtha/Distillate cut point

PRODSP.ANHVN.BLK.D86CRV_D86CRV_90

PRODSP.MFHVN.SPLTCALC.BLK.ECP

90% of light Naphtha

The light naphtha/heavy naphtha cut points

PRODSP.ANLVN.BLK.D2887CRV_TBPCRVWT_90

The split ratio of pre-fractionation tower

LPG composition

The split ratio of LPS

Fuel gas composition

The de-lumping ratio of CMAPPER

Distillation curve of Distillate and Bottoms

Model Tuning Facts with Specifications The basic tuning facts in the reaction section of Aspen Plus Hydrocracker/Hydrotreater are provided mainly based on the types of catalysts types and their deactivation. The Aspen Plus Hydrocracker/Hydrotreater model is built so the model can be tuned consistently. This section discusses model tuning.

Reaction Rate Tuning Strategy Following table shows the group rate constants which needs to be adjusted to match the plant observation. Observation

Group Rate constants

Each Reactor Bed temperature rise

Global Rate constants for each reactor bed

Product Sulfur

HDS Group Activity factors

Product Nitrogen

HDN Group Activity factors

Product Yields

Cracking Group Activity factors

Hydrogen consumption/Total temperature rise in the reactor

Hydrogenation Group activity factors

Relative amount of paraffins/naphthene ratio in the bottoms

Ring opening group activity factors

The variables which will remain constant in the model tuning are:

5 Simulation



Equilibrium constants for Aromatic saturating



Heat of reaction for each reactions



Activation energy for each reaction



Adsorption constants for NH 3, nitrogen compounds, H2S, and Aromatics



Reaction order for H 2

77

Feed Property Tuning In the Aspen Plus Hydrocracker/Hydrotreater model, Feed Adjuster models are applied to each feed stream. For the feed specifications, the bulk properties and the distillation information for each inlet stream, such as API gravity, refractive index, and viscosity are used to calculate the carbon on the aromatics and the carbon on the naphthene in the Feed Adjuster models. Because of the degrees of freedom for the model tuning, the Feed Adjuster model does not match all properties exactly. In tuning the Feed Adjuster model with detailed feed analysis and product properties, you should calculate the bias for the gross properties. In model runs for feed property calibration, the bias of some properties should be constant and the measurement for the feed properties should be calc, which means that you need to swap the specification between measurements and biases. The table below gives the all paired variables to make sure they are swapped during the model runs. CONST

CALC

FEED.FEEDADJ.BLK.FFEED1_FEED1_RI_LAB_MEAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_RI_LAB_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_CST210_MEAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_CST210_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_TOTAL_CA

FEED.FEEDADJ.BLK.FFEED1_FEED1_TOTAL_CA_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_NDM_CN

FEED.FEEDADJ.BLK.FFEED1_FEED1_NDM_CN_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_H_CONTENT

FEED.FEEDADJ.BLK.FFEED1_FEED1_H_CONT_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_MOL_WT

FEED.FEEDADJ.BLK.FFEED1_FEED1_MOL_WT_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_LIT_LMP_WABP

FEED.FEEDADJ.BLK.FFEED1_FEED1_LIT_LMP_WABP_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_MED_LMP_WABP

FEED.FEEDADJ.BLK.FFEED1_FEED1_MED_LMP_WABP_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_HVY_LMP_WABP

FEED.FEEDADJ.BLK.FFEED1_FEED1_HVY_LMP_WABP_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_LIGHT_CUT_PT_400F

FEED.FEEDADJ.BLK.FFEED1_FEED1_LIGHT_CUT_PT_400F_BIAS

FEED.FEEDADJ.BLK.FFEED1_FEED1_HEAVY_CUT_PT_650F

FEED.FEEDADJ.BLK.FFEED1_FEED1_HEAVY_CUT_PT_650F_BIAS

Separation Model Tuning For the simplified separation model, the effective cut point (ECP) can be tuned to match your distillation with some constraints. Below is a list of ECP variables for main fractionator and jet tower separation. Variable Name

Spec


CONST


CONST


CONST

Alternatives to Model Running Mode One of the advantages of an open equation running mode is that specified variables can be swapped with others based on the way the model is built. So, the alternatives for the simulation running are configured as shown below. The tables below show the constant specifications for those eight modes: 

78

Fixed inlet temperatures. This is the demo case as specified. The constant variables:

5 Simulation

Variable Name

Spec

RXN.HTR.BLK.C_OUT_TEMP

CONST

RXN.QR1B2.BLK.PROD_TEMP

CONST


CONST


CONST

Note: The specified variables are usually set on block outlet variables. This helps to easily understand the specification. 

Variable Name

Spec

RXN.R1B1.BLK.OUTLET_TEMP

CONST


CONST


CONST


CONST



Fixed WART and T-rises

Variable Name

Spec

RXN.T-BIAS.BLK.R1B2R1B1_OUTTEMP_DELTA

CONST


CONST

RXN.WABT.BLK.R1_WART

CONST

RXN.WABT.BLK.R2_WART

CONST



Fixed sulfur in the bottom and T-rises biases

Variable Name

Spec

RXN.T-BIAS.BLK.R1B2R1B1_T-RISE_DELTA

CONST


CONST


CONST

PRODSP.ANBTMS.BLK.WSULFUR_HYCSUL

CONST



Fixed nitrogen in the bottom and T-rises biases

Variable Name

Spec


CONST


CONST


CONST

PRODSP.ANBTMS.BLK.WNITRO_HYCNITR

CONST



5 Simulation

Fixed outlet temperatures. The inlet temperature specifications are swapped to reactor bed outlet temperatures.

Fixed sulfur in the bottom and outlet T biases

Variable Name

Spec


CONST


CONST


CONST

PRODSP.ANBTMS.BLK.WSULFUR_HYCSUL

CONST

79



Fixed nitrogen in the bottom and outlet T biases

Variable Name

Spec


CONST


CONST


CONST

PRODSP.ANBTMS.BLK.WNITRO_HYCNITR

CONST



Fixed conversion, and outlet T biases

Variable Name

Spec


CONST


CONST


CONST

RXN.YIELDS.BLK.CONVERSION_WT_PCT

CONST

Flowsheet Changes By varying the flowsheet, the Aspen Plus Hydrocracker/Hydrotreater model has the flexibility to match the process measurements with certain specification changes. This provides the opportunity to use one model to fit multiple flowsheets.

Reaction Section Change In general, there is only one reactor with two to four beds in a Hydrocracker plant. To reduce redundant work on flowsheeting, you can turn off the second reactor by setting a set of specific process variables.

To Turn Off the Second Reactor: 1

Turn the quench flow to the second reactor to zero.

2

Because the reactor specifications on reactor outlet temperatures are fixed by calculating the recycle gas flow rates (quench flow), you should swap the specification on the outlet temperatures of the second reactor to the recycle gas flow.

Current Spec

Specs for R2 off

QR2B2.BLK.PROD_TEMP F007.BLK.QUENCH_TO_R2B1_PLANT QR2B1.BLK.PROD_TEMP F008.BLK.QUENCH_TO_R2B2_PLANT

Note: With different simulation schemes (given a specific set of specifications) the original constant specifications for swapping can vary. 3

Set the flow variables to zero,

F007.BLK.QUENCH_TO_R2B1_PLANT = 1.0E-6 F008.BLK.QUENCH_TO_R2B2_PLANT = 1.0E-6 4

Turn the reactor global activities for all beds to zero to avoid any reaction taking place in the reactor.

R2B1.BLK.GLOBAL_ACTIVITY.value = 0.00000001

80

5 Simulation

R2B2.BLK.GLOBAL_ACTIVITY.value = 0.00000001

Separation Section Change As the Aspen Plus Hydrocracker/Hydrotreater model is built, there are two vapor product streams: 

LPS overhead



pre-fractionation overhead gas

and five liquid streams: 

LPG



light naphtha



heavy naphtha



distillate



bottoms

If the gas flow and composition are known, to adjust the LPS overhead flow rate and composition, you can change the following split ratios: 

PRODSP.LPS.BLK.LPS-OH_SPLIT_N2



PRODSP.LPS.BLK.LPS-OH_SPLIT_NH3



PRODSP.LPS.BLK.LPS-OH_SPLIT_H2S



PRODSP.LPS.BLK.LPS-OH_SPLIT_H2



PRODSP.LPS.BLK.LPS-OH_SPLIT_C1













Similarly, to adjust the pre-fractionation overhead composition, you can change the following split ratios: 

PRODSP.PREFRAC.BLK.OVGAS_SPLIT_N2



PRODSP.PREFRAC.BLK.OVGAS_SPLIT_NH3



PRODSP.PREFRAC.BLK.OVGAS_SPLIT_H2S



PRODSP.PREFRAC.BLK.OVGAS_SPLIT_H2



PRODSP.PREFRAC.BLK.OVGAS_SPLIT_C1













To adjust the LPG flow rate and composition, you can change the following split ratios:

5 Simulation



PRODSP.PREFRAC.BLK.TOGASPLT_SPLIT_N2



PRODSP.PREFRAC.BLK.TOGASPLT_SPLIT_NH3



PRODSP.PREFRAC.BLK.TOGASPLT_SPLIT_H2S



PRODSP.PREFRAC.BLK.TOGASPLT_SPLIT_H2



PRODSP.PREFRAC.BLK.TOGASPLT_SPLIT_C1

81













The separation of the liquid products is calculated using the simplified separation model. You can change the flow rate, cut point and overlap between the liquid products by adjusting the cut point and top/bottom slopes. as shown below.

Light naphtha/heavy naphtha Cut point


Top slope

PRODSP.MFHVN.SPLTCALC.BLK.SITOP

Bottom slope

PRODSP.MFHVN.SPLTCALC.BLK.SIBOT

Heavy naphtha/distillate Cut point


Top slope

PRODSP.MFBTMS.SPLTCALC.BLK.SITOP

Bottom slope

PRODSP.MFBTMS.SPLTCALC.BLK.SIBOT

Distillate/Bottom Cut point


Top slope

PRODSP.JETTOWER.SPLTCALC.BLK.SITOP

Bottom slope

PRODSP.JETTOWER.SPLTCALC.BLK.SIBOT

To avoid negative product flows, the cut point should always be increasing from light gasoline to bottom.

Feed Specification Change As the Aspen Plus Hydrocracker/Hydrotreater model is built, there are six streams built for typical feeds fed into the reactor. So, if a feed is not applied to a specific plant, you can turn the flow rate for that stream to zero to eliminate the stream. In order to avoid model stiff and/or singularity, you can set the flow as small as you can (for example, 10E-6).

Running a Simulation Case Introduction After you have entered input data, you can solve the model with the updated data by selecting the Simulate option from the AspenPlusHYC toolbar and then selecting the play button. Alternatively, select the Run Cases option from the AspenPlusHYC menu and select the Simulation Case option from the sub-menu.

82

5 Simulation

In either case, when the solution is initiated, the model is updated with all input data and then solved. The command line dialog appears while the model is being solved. When the solver has concluded (following convergence or failure) the Simulation sheet is updated with data from the model. No other sheets are updated.

Error Recovery You should check the convergence status shown at the top of the simulation sheet after running the simulate case. The results on the simulation sheet are only meaningful if the convergence status is converged. If the status is not converged, then it is generally desirable to return the simulation sheet and model to their pre-solution states.

To Return The Simulation Sheet And Model To Their PreSolution States: 1

Restore the model to the base parameter case by clicking AspenPlusHYC | File | Load Case Data.

2

Browse for the .var file in which you saved the results of the base parameter case.

3

Restore the simulation sheet user input by clicking AspenPlusHYC | File | Load User Input Sheet.

4

Examine the input data as compared with the base parameter case. Convergence failure for the simulation case typically has one of two basic causes: 

5 Simulation

Poor or erroneous data were entered as input (blue-colored cells). For example:  Check that reasonable feed property data were entered for all feeds.  Check that reasonable catalyst property data were entered for all catalysts.

83

 

Check that the cut points entered for light naphtha and LCO are physically possible.

The solver parameters are too "aggressive" for the data entered. For example, a large change in feed rate (greater than 15%) may require more conservative solver parameters.

Error Recovery - Parameterization You should check the convergence status shown at the top of the param sheet after running the parameter case. The results on the param and analysis sheets are only meaningful if the convergence status is converged. If the status is not converged, then it is generally desirable to return the param sheet and model to their "pre-solution" states.

To Return The Simulation Sheet And Model To A Converged Parameter Case: 1

Restore the simulation sheet user input by clicking AspenPlusHYC | File | Load Case Data.

2

If this is your first attempt at running a param case, then load user_default.var or the .var file created by AspenTech for your site. If you have converged parameter cases for your unit, then load the corresponding .txt file that most closely represents the process conditions and input data for the new parameter case.

3

Restore the param sheet user input by clicking AspenPlusHYC | File | Load User Input Sheet.

4

Examine the input data as compared with the base parameter case.

Convergence failure for the parameter case typically has one of three basic causes:

84



Poor or erroneous data were entered as input (blue-colored cells). For example:  Check that physically realistic property data were entered for all feeds and products. For example, do all distillation points must increase as a function of percent distilled.  Check that physically realistic property data were entered for all catalysts. For example, the ECAT activity must always be lower than fresh activity.  Check that physically realistic mechanical data were entered. For example, the regenerator cyclone height must be greater than the bed height.



Some of the input data violate valid ranges. Such restrictions are a consequence of the equation-based manner in which the model has been formulated Observe the following guidelines when entering data:  Do not set any recycle rate to zero. For zero recycle rates, use a very small number instead (e.g. 0.1 BBL/D).  "Fraction to riser bottom": The midriser feed rate must be nonzero. If the midriser feed rate is in fact zero, set the "fraction to riser bottom" for feed 1 equal to 0.999999.  Data restrictions for light-ends analyses:

5 Simulation

 

  

5 Simulation

Compositions for any one stream must not sum to zero, including streams having a zero flow rate. For the light and heavy naphtha streams, all C5+ components must be nonzero, again including any stream having a zero flow rate. For any one component, the sum of its composition across all streams must not be zero.

Do not enter zero for any flue gas component.

The solver parameters are too "aggressive" for the data entered. For example, a large change in feed rate (greater than 15%) may require more conservative solver parameters.

85

6 Running Multiple Cases

Overview In addition to single cases, Aspen Plus Hydrocracker lets you run multiple cases at a time to retrieve the results into a single area that is easy to work with. This can be useful if you want to see how the model responds to changes in one or more variables. For example, you might want to see what the gasoline yield looks like as a function of riser overhead temperature. To do this, you would want to run multiple cases with different temperatures, and then the results reported and graphed to determine where the over-cracking peak occurs. This can easily be accomplished by running the case study option.

Before You Start Before running a case study, you must first set up which variables will be varied and which variables will be reported.

To do this:

86

1

On the AspenPlusHYC menu, select Setup Cases.

2

On the sub-menu, select Case Study as shown below.


This activates the Case Study worksheet and opens a dialog box from which you can choose the independent variables that will be varied and the variables that will be reported as shown below.

3

Select the variables and click the OK button to set up the worksheet.

4

On the new dialog box that opens, specify the first and last case to run.

5

Click the OK button.

The worksheet is updated. The independent variable names appear on the Case Study page in Column A starting with Row 9. The reported variables


87

appear on the Case Study page in Column A starting with Row 111. Column B is reserved for optional user labels. 6

Once the spreadsheet is set up, input values for the independent variables for each case to be run.

The model is now ready to execute a case study. 7

On the AspenPlusHYC toolbar, select the Case Study option; then click the play button, as shown below.

Before the case study commences, a dialog box requesting the first and last case opens. 8

Enter the correct information; then click the OK button.

The command line opens and the case study begins with the first case you specified. After the first case is solved the command line closes while the data is retrieved to the spreadsheet. The command line re-opens when each subsequent case starts. For case studies, neither the Hydrocracker Summary page nor the ParamData page is updated. After the independent variable data has been sent, the cells are highlighted in blue. Similarly, after the reported variables have been retrieved, those cells are highlighted in blue. In addition to reporting values for all of the specified report variables, a set of LP vectors is generated for each case. These LP vectors correspond to the LP vectors that have been set up on the LP Vector worksheet. These are reported in the LP vector section of the case study page starting with Row 1005. Column A lists the dependent variables and Column B lists the independent variables. The values that are returned for a case study are highlighted in blue. Note: When a new case study is set up or run, the data that are currently in the worksheet are not automatically erased. To remove any unwanted data, you must manually highlight and delete the unwanted data. Each step of the case study typically requires about one minute to solve and update the spreadsheet. A single case study is somewhat faster than a corresponding simulation because less data is sent to and retrieved from the model.

88


7 Optimization

Optimization Basics You use Optimization to maximize or minimize a specified objective function by manipulating independent variables (feed stream, block input, or other input variables). The objective function is an equation that is used by the optimization engine to determine the manner in which to manipulate the degrees of freedom (independent variables) in a problem. The optimization engine for the Hydrocracker model is DMO solver. You can have more than one objective function in a problem, but only one is used by the engine during the solution. In Optimization cases in the Hydrocracker model, the objective functions are normally economic: 

Maximize operating profit.



Minimize operating cost.



And so on.

Although Aspen Plus lets you define three different types of objective functions: 

Linear



Sum of Squares



Symbolic

the Hydrocracker model uses only objective functions that are in linear form.

Syntax of Setting Bounds: Aspen Plus lets you bound every variable in the problem as shown below:

xl < x < xu

Specifying Bounds for Independent Variables Independent variables are those variables that can be changed independently; for example, the feed rate in the Hydrocracker unit. The optimizer can vary the values of independent variables to push the values of the objective function in the defined direction (maximize profit or minimize cost) until some bounds are reached.

7 Optimization

89

Each independent variable accounts for a degree of freedom. The number of degrees of freedom is equal to the number of independent variables in an optimization run, if no independent variable is at its bound. You can impose upper and lower bounds on independent variables to prevent the solution from deviating too far from the starting point. You can also impose step bounds on independent variables.

Setting Step Bounds: The step bound of an independent variable defines how much the value of the variable can be changed in a single optimization run. Step Bound Syntax is shown below.

Xl = max(X - |Xstep|, Xlower ) Xu = min(X + |Xstep|, Xupper ) The actual bounds used in the run are computed using: 

The step bound



The initial value.



The lower bound.



The upper bound.

You must specify upper and lower bounds for all independent variables. Optionally, you can also define the step bounds for independent variables. Most dependent variables in the Hydrocracker model have very wide bounds, for example: 

1.E20 for the upper bound.



1.E20 for the lower bound.

Specifying Bounds for Dependent Variables: In general you should not heavily bound an optimization problem for reasons that are both practical and algorithmic. While AspenTech recommends bounds on independent variables to avoid unbounded problems, all other bounds should be used only when absolutely necessary. However, some dependent variables have physical meaning. For these, you should specify appropriate bounds to prevent the solution from getting into infeasible operating conditions. For example, there is a metallurgic limit on regenerator cyclone temperature. Therefore, you should set an upper bound for this variable. When setting up an Optimization case in the Hydrocracker model, you need to define only those constrained dependent variables.

Setting Up Objective Functions The first step to setting up an optimization case is to set up the objective function(s). The objective function appears in the form of profit function in the Aspen Plus Hydrocracker model. You can set up three different profit functions in EXCEL interface. The pre-configured spreadsheets (Profit1,

90

7 Optimization

Profit2, and Profit3) contained in the interface let you easily set up profit functions of the following form: Profit = F1 * (V1 + (IP1_1 –BV1_1)*IV1_1 + (IP2_1 – BV2_1)*IV2_1 + (IP3_1 –BV3_1)*IV3_1) + F2 * (V2 + (IP1_2 –BV1_2)*IV2_1 + (IP2_2 –BV2_2)*IV2_2 + (IP3_2 –BV3_2)*IV3_2) + … Where: F1 and F2 represent some principal properties to which values (V1 and V2) are attached. For example: 

Let F1 represent the product diesel flow rate



Let V1 the value of the diesel



Let F2 represent the light naphtha flow rate



Let V2 the value of the light naphtha

IP1_1, IP1_2, and IP1_3 represent incremental properties that can change the value of F1. In the example, if F1 is considered to be the diesel flow rate: 

IP1_1 might be the cetane index



IP1_2 might be the pour point



IP1_3 might be the sulfur content

BV1_1, BV1_2, and BV1_3 are the base values of cetane index, pour point and sulfur content. The difference between the calculated value and base value for each incremental property is then multiplied by the corresponding values (IV1_1, IV2_1, and IV3_1). For example, if the base value of cetane index (BV1_1) is 40: 

The calculated value (IP1_1) is 41.



The value IV1_1 is 0.25.



The value of diesel (V1) is increased by 0.25.

You can specify as many principal properties as you want. You can specify from zero to three incremental properties for each principal property. Setting up objective functions involves two steps: 

Selecting variables for the principal properties you are interested in



Selecting incremental properties for those variables



Entering data for the properties on the ProfitX worksheet

To Set Up a Specific Objective Function (For example Profit 1): 1

On the main menu, click Aspen Plus Hydrocracker | Setup Cases.

2

On the submenu, click Profit 1.

The Add Variables to Objective Functions dialog appears:

7 Optimization

91

Any variable in this list is marked Yes in Column AB (with the title Profit 1 Selected) on the ReceiveVars page. The default list includes most product streams and feed streams that can be used in a profit function. You can add variables to this list by putting Yes in Column AB on the ReceiveVars page for the desired variables. 3

Select or deselect variables to be included in the computation of the objective function. To select a variable, check the box to the left of the variable. To deselect a variable, clear the box to the left of the variable.

4

For each variable to which you want to add an incremental, click the variable.

The Add Properties to the Selected Product appears. 5

Check up to three properties to be added to the variable you selected.

6

Click the OK button to add properties to the variable (they will also appear on the Profit1 worksheet). –orClick the Cancel button to abort the adding and deleting of properties to the variable.

You are returned to the Add Variables to Objective Functions dialog. 7

When you are finished adding variables and incrementals to the objective variable: Click the OK button to add the checked (and delete the cleared) variables from the Profit1 worksheet. –orClick the Cancel button to abort the adding and deleting of variables for the objective function.

92

7 Optimization

Entering Data About Properties on the ProfitX Worksheet: After you have selected the all the principal properties and incremental properties that comprise the objective function, they will appear on the ProfitX (Profit1 in the example) worksheet. You must now enter values for each principal property and each incremental property.

Entering Data About Principal Properties: For all principal property variables: Description

Column

variable name

C

units of the variable

D

price unit

G

Each principal property occupies one row. Only one value is needed for each principal property.

You must enter the appropriate price value in Column E. In the example shown, all flow rates are in thousands of barrels per day. Therefore, the values are entered as dollars per thousands of barrels. If you select a unitless variable for the objective function, such as conversion, the value is also unitless. If the principal property is a cost (such as feed value), the value would be negative.

7 Optimization

93

Entering Data About Incremental Properties:

The incremental properties are listed to the right of the principal properties on the ProfitX (here, Profit1) worksheet. For all incremental property variables, the variable name appears in Column H. Each incremental property is in the same row of the principal property that the incremental property is associated with. In the example shown, the incremental property, Hydrocarbon of aromatic in the bottom of the main fractionator, is in Row 17, the row of its corresponding principal property, Debutanizer bottom flow. Two values are needed for each incremental property: 

Base



Price

The base value is the value at which the incremental property has no effect on the principal property. For example, the value of $10,000/MLBPH is the value of the debutanizer bottom stream with a Aromatics of 30 in the example shown. The price value is the amount of cost. In the example, each increase of 1 Aromatic raises the cost by $1. All incremental properties are either unitless or have a fixed unit type such as weight percent sulfur. Penalty or bonus is calculated by multiplying the deviation from the base value and the price of unit deviation. To identify each incremental property, you can assign a label for each incremental property. You can enter the label in Column G. In the example, the label Octane BBL is entered for RDON of light naphtha as well as for RDON of heavy naphtha. If a principal property has more than one incremental property, the second and the third one lie to further right on the same row. The entry for the second and the third incremental properties is similar to that for the first incremental property.

94

7 Optimization

To identify each incremental property, you can enter a label for each incremental property in Column G. Follow the same procedure to set up objective functions of Profit2 and Profit3. The location of marking on the ReceiveVars page and the SendVars page can be different. You should be able to find corresponding columns by looking at the column titles on those two pages. In order to save the changes made to the objective function pages, the worksheet must be saved. Any variable in this list has a name in Column AE (with title I1 name) on the ReceiveVars page. The label, base, and price of the incremental property are list in Column AF to AH on the ReceiveVars page. The default list includes most of important product properties. You can add variables to this list by editing Column AE through AH on ReceiveVars for the desired variables.

Setting Up An Optimization In addition to setting up the objective functions used to optimize, you must also set up the optimization variables and any bounds that are necessary. For example, you can choose to optimize Profit 1 by varying the feed rate of Feed 1. However, the unit may have constraints to how much wet gas can be processed, so the wet gas volume would be selected as a dependent constraint variable. The independent variables have the specification of Const if optimization has never been set up before. However, not all variables with the Const specification in the model are included in independent variable list. Only those variables that can be manipulated in the Hydrocracker unit will appear. Those variables are identified by Opt in Column R (with title Opt) on the SendVars page.

7 Optimization

95

The default list of independent variables should be able to handle most optimization runs. If you want to use other variables as independent variables, manually set up those variables on the SendVars page. After a variable is selected on the independent list, the corresponding Column V (with title Opt Select) on the SendVars page is marked YES for this variable. When solving the optimization case, the variable is sent to Command Line with Optim specification. However, the SendVars page still keeps the original Const specification for this variable. The dependent constraint variables have the specification of Meas or Calc. However, not all variables with Meas and Calc specification appear in the list. Only those variables that represent operation constraints in Hydrocracker units appear in the list. Those variables are identified by Opt in Column R (with title Opt) the SendVars page. The default list of dependent variables represents all constraints commonly met in HYCU operation. If you have a particular constraint that is not represented by any variable on the list, manually set up those variables on the SendVars page. After a variable is selected on the dependent list, the corresponding Column Y (with the title Opt Select) on the SendVars page is marked YES for this variable. Aspen Plus Hydrocracker presents only CONST variables in the pick list of independent variable and only CALC and MEAS variables in the pick list of dependent variables in order to ensure that whatever set you choose will lead to a well-posed problem.

To Set Up An Optimization: 1


2

On the submenu, select Optimization.

This activates the Optimize worksheet, and opens the Setup Optimization Case dialog from which you can select: 

The desired independent variables (extra degrees of freedom).



The dependent constraint variables.

3

Click the check box to the right of the variable name to select a variable. -orClick the check box to the right of a selected variable name to deselect a variable.

96

7 Optimization

4

When the independent variables and the dependent constraint variables have been selected, click the OK button to complete the setup. -orClick the Cancel button to close the dialog box without making any changes to the optimization problem.

The selected variables and their current values will then appear on the Optimize spreadsheet. After you set up an optimization, a message box appears to remind you to Make sure the profit function is defined before running the optimize case.

After selecting the desired independent variables and dependent constraint variables, you should then input lower and upper bounds in Columns C and F by the appropriate variables. You can also input step bounds for the independent variables in Column G. The optimization is now ready to solve. To save the changes made to the Optimize pages, the worksheet must be saved.

7 Optimization

97

Executing Optimization Cases To Solve The Optimization: 1

On the AspenPlusHYC toolbar, select the Optimize option; then click the play button.

The Select Objective Function dialog appears.

2

98

On the Select Objective Function dialog, select an active objective function. You can select only one active objective function.

7 Optimization

3

Select the direction of the optimization by selecting maximizing or minimizing function. 

If the objective function is set up as a profit function, the user should select Max.



If the objective function is set up as a cost function, the user should select Min.

4

Select the profit reports to update. Normally only the active objective function is selected.

5

Click the OK button to complete the setup. -orClick the Cancel button to close the dialog box and return to the Optimize worksheet.

If you clicked the OK button, the data from the optimization spreadsheet is sent to the model, and the command line dialog box opens.

To Change The Behavior Of The DMO Solver: You can change the behavior of the DMO solver by selecting one of the three buttons at the bottom of the command window. Your selection takes effect at the start of the next DMO iteration.

7 Optimization

Button

Action

Abort

forces the model to quit solving

No Creep

takes the DMO solver out of creep mode. This is used to expedite solving when the current run is close to the final solution, in which case both the Residual Convergence Function and the Objective Convergence Function are small and close to the convergence criteria.

Close Residuals

causes the model to close the residuals without minimizing the objective function convergence. The Close Residuals button is useful in cases where the objective function very nearly reaches a maximum value but the convergence of the objective does not close.

Close

This button is unavailable during the optimization run. It is

99

Button

Action only activated when no run is being executed. Click the Close button to close the dialog box and return the EXCEL interface.

After the model solves the optimization, the solution values are retrieved into the optimization page and the spreadsheet is updated. The corresponding report page, the Optimize page, and the Simulation page are updated to the current values in the model, but the Param page is not updated. On the Optimize page, the values after the optimized values are placed into Column E. If any upper or lower bound is reached, that value is highlighted in red. A typical optimization takes three to five minutes, but this could be higher or lower depending on how difficult it is to reach a solution.

Analyzing Optimization Solutions There is one profit report worksheet (Profit1Report, Profit2Report, Profit3Report) for each objective function.

The profit report worksheet is designed to show the change between the starting point and the optimization solution. All the principal properties and incremental properties used in the corresponding objective function are listed in Column A of the worksheet. Each property occupies one row. The incremental properties of a specific principal property occupy the rows below the row of related principal property. The rates, prices, and case flow are listed to the right of the principal properties and incremental properties in a profit report worksheet.

100

7 Optimization

On the worksheet you can see clearly how each property contributes to the objective function in terms of case flow for both the starting point and optimization solution. You can also observe how the optimizer adjusts the values of properties to maximize the profit. You can further analyze the optimization results by comparing it to the engineering knowledge of Hydrocracker operation. For example, suppose the Hydrocracker unit is operated in a Gasoline Mode. You will probably assign a high value to gasoline. You expect to see gasoline production maximized, and so the riser outlet temperature is increased from its initial value. In another case if the Hydrocracker unit is operated in Heating Oil Mode, and heating oil is assigned a higher value than gasoline, you would expect to see the Hydrocracker maximize heating oil production. In this case, the riser outlet temperature might not reach the upper limit. You can also examine the Optimize worksheet to see if any independent or dependent variable is at its limit. Normally at least one independent variable or dependent variable is at its bound in the optimization solution. For example, the riser temperature or the wet gas rate might hit its upper bound in the example of Gasoline Mode.

7 Optimization

101

8 LP Vectors

Overview – Generating LP Vectors In addition to letting you determine yields, temperatures, product properties, and so on, the Aspen Plus Hydrocracker offers the capability of generating LP (Linear Programming) vectors.

Purpose of Running LP Vectors The main purpose of generating LP vectors is to provide shift factors for an LP planning and scheduling tool, such as PIMSTM. In LP planning and scheduling tool for refining industry, every processing unit is represented by a simplified linear model. Like all linear models, a HYCU LP model uses fixed gains (base model and shift factors) to represent the relationship between operating conditions and product flow rates and properties. This table shows part of a Hydrocracker model in PIMSTM.

The model gains at three different riser temperatures are listed in Columns F, G, and H. Those gains represent the conversion of one unit feed to various light gas products.

102

8 LP Vectors

When feeds to HYCU change, you should change those conversion factors. The shift factors are listed in Columns I to L. Each shift factor represents the amount of change on conversion when a feed property changes one unit. For example, the shift factor between K factor vs. C3 is –0.0046 at the riser temperature of 970 F. This means that the C3 conversion will increase by 0.0046 for a one unit increase in feed K factor. There are also shift factors for 980 F and 990 F. Those shift factors greatly affect the accuracy of LP solution. For a highly nonlinear reaction process such as HYCU, shift factors can only be accurately estimated by a rigorous nonlinear model. The standard Aspen Plus Hydrocracker model is a rigorous nonlinear model with detailed kinetic description. It is well suited to provide accurate shift factors. The Aspen Plus Hydrocracker model has build-in function of generating LP vectors, which are equal to the shift factors in a LP model. The LP vectors can also act as the accuracy indicator of the Aspen Plus Hydrocracker model. The LP vectors can be compared to gain matrix in APC (Advanced Process Controller). The gain matrix in APC is obtained from plant step tests. It should closely represent the real relationship between independent variables and dependent variables in the process. By comparing to gain matrix in APC model, you can observe how closely the model represents the process. The comparison can also pinpoint what part of the model needs further tuning.

LP Vector Generation From the model point of view, LP vectors are the gains between a set of independent variables and a set of dependent variables. LP vectors are calculated by doing a sensitivity analysis on the model. In the Aspen Plus Hydrocracker model, LP vector generation is executed by issuing a sensitivity analysis command to the command line and retrieving the results back to the EXCEL interface. LP vector generation is run in Simulation mode. The independent and dependent variables you choose for LP vector generation must correspond to fixed and free variables in the simulation mode. The fixed variable in simulation mode has the specification of CONST or PARAM. However, the PARAM variables are normally internal to the model and have no physical meaning. Therefore, they do not appear in the set of independent variables in LP generation. All independent variables have the specification of CONST. The free variable in Simulation mode has the specification of CALC or MEAS. Therefore, a dependent variable has the specification of CALC or MEAS. Aspen Plus Hydrocracker provides you a pick list of:

8 LP Vectors



Independent variables.



Dependent variables.

103

Aspen Plus Hydrocracker presents only CONST variables in the pick list of independent variables and only CALC and MEAS variables in the pick list of dependent variables in order to ensure that whatever set you choose will lead to a well-posed problem. The two lists that Aspen Plus Hydrocracker model provide should be able to satisfy most cases. However, you can add variables to those two lists if necessary. Any variable in the independent variable list is marked LP in Column Q (with title LP) on the SendVars page. You can add variables to this list by enter Yes in Column Q on the SendVars page for the desired independent variables. Any variable in the dependent variable list is marked "LP" in Column Q (with the title LP) on the ReceiveVars page. You can add variables to this list by entering Yes in Column Q on the ReceiveVars page for the desired dependent variables.

Generating LP Vectors Generating LP vectors is a two-step process. 1

You must first specify what the independent and the dependent variables are.

2

You then run the LP vector generation command.

To Specify The LP Vectors: 1


2

On the submenu, select Vectors.

The LP Vectors spreadsheet is activated and the Setup LP Vectors dialog box appears on which you can specify the independent and dependent variables.

104

8 LP Vectors

The top list box is used to select the independent variables. You can select any or all of the variables listed. 3

Click the check box to the right of the variable name to select a variable. -orClick the check box to the right of a selected variable name to deselect a variable.

The lower list box is used to select the dependent variables. It works exactly like the independent variable list box. 4

When the independent variables and the dependent constraint variables have been selected, click the OK button to complete the setup. -orClick the Cancel button to close the dialog box without making any changes to the LP Vectors page.

If you click the OK button, the LP Vectors" page is be cleared and the independent variables appear in the seventh row and the dependent variables appear in Column C. In order to save the changes made to the LP Vectors page, the worksheet must be saved. Below is the LP Vectors Sheet after setup.

8 LP Vectors

105

5

After the independent and dependent variables have been set up, generate LP vectors by selecting LP Vectors from the AspenPlusHYC toolbar, and then selecting the play button.

The model will then calculate the Jacobian for the model and retrieve all of the desired vectors into the LP Vectors page. Below is the LP Vectors Worksheet after running a LP Vectors case.

106

8 LP Vectors

The command line dialog box will open for a short time while the Jacobian is being evaluated and while the LP vectors are being calculated. You cannot issue any commands to the command line dialog box at this time, however. Typical execution time is about 20 seconds, although it can be more or less depending on how many values are being retrieved.

8 LP Vectors

107

9 Reaction Kinetics Details

Overview The Aspen Plus Hydrocracker/Hydrotreater model is derived from a model developed by Sun Oil Company. The components and reaction networks are consistent with fundamental research conducted at the University of Delaware, the University of Utah, and other academic and industrial research institutions. This section provides details of the reaction model.

Component Slate The component slate chosen to represent the feed and the product streams of the Aspen Plus Hydrocracker/Hydrotreater plant comprises 116 components covering the full range from hydrogen to hydrocarbons, with 47 carbon components (B.P. 1300 C). The component slate varies in different sub-plants. Component mappers are used to connect adjacent sub-plants. Using different component slates helps to reduce the number of variables in the sub-plant. In the reactor model, the olefins components are assumed to be completely saturated. Table A below shows the corresponding components in the reactor model. The total number of components in the reactor model is 97. Table B below shows the corresponding olefin components in the feed but not in the reactor model. The light ends are defined using discrete components through C3. For C4 to C10 hydrocarbons, one pure component is used to represent several isomers. For example, the n-butane represents both n-butane and iso-butane. For higher boiling point components, only compounds with carbon number 14, 18, 26, and 47 are used to represent wide range of boiling point components. The components also cover different classes of hydrocarbons, which include one-ring naphthenes to 4-ring aromatics. The sulfur compounds are separated into 8 groups of 13 components which include:

108



Thiophenes.



Sulfides.




Benzothiophenes.



Tetrahydro-benzothiophenes.



Dibenzothiophenes.



Tetrahydro-dibenzothiophenes.



Naphthabenzothiophenes.



Tetrahydro-naphthabenzothiophenes.

The nitrogen compounds are represented by 10 lumps which include both basic and non-basic nitrogen compounds.

Table A- Component Slate for Hydrocracker Model in Reactor Component

Formula

Abbreviation

Nitrogen

N2

N2

Ammonia

NH3

NH3

Hydrogen Sulfide

H2S

H2S

Hydrogen

H2

H2

Class

Paraffins Methane

CH4

C1

Ethane

C2H6

C2

Propane

C3H8

C3

N-Butane

C4H10_2

C4

N-pentane

C5H12_2

C5

2,3-dimethylbutane

C6H14_2

C6P

2,3-dimethylpentane

C7H16_5

C7P

2,3-dimethylhexane

C8H18_6

C8P

2,6-dimethylheptane

C9H20_4

C9P

2,5-dimethyloctane

C10H22-1

C10P

n-tetradecane

C14H30

C14P

n-octadecane

C18H38

C18P

Tetracosane

C26H54

C26P

C47 Paraffins

C47H96

C47P

Methylcyclopentane

C6H12-2

C6N

Methylcyclohexane

C7H14-6

C7N

Cyclohexane, 1,4-dimethyl

C8H16-7

C8N

1-trans-3,5-trimethylcyclohexane

C9H18-1

C9N

C14-1-ring-cycloheaxane

C14H28

MN1Lo


C18H36

MN1Hi


C21H42

HN1


C47H94

VN1

Trans-decaline (two Ring)

C10H18-2

C10N

C14-2-ring-cyclohexane

C14H26

MN2LO


C18H34

MN2HI

CnH2n+2

Naphthenes


CnH2n

CnH2n-2

109

Component

Formula

Abbreviation


C21H40

HN2


C47H92

VN2

C14-3-ring-cyclohexane

C14H24

MN3Lo


C18H32

MN3Hi


C21H38

HN3


C47H92

VN3


C21H36

HN4


C47H88

VN4

C6H6

C6A

Class

CnH2n-4

CnH2n-6

Aromatics Benzene Toluene

CnH2n-6

C7H8

C7A

C8H10_3

C8A

2-methyl-3-ethylbenzene

C8H12-3

C9A

1,2,3,4,-tetrahydronaphthalene

C10H12

C10A

n-octylbenzene

C14H22

MA1Lo

C18-1ring-Arom

C18H30

MA1Hi

C21-1ring-Arom

C21H36

HA1

C47-1ring-Arom

C47H88

VA1

C14-tetrahydronaphthalene

C14H20

MANLo


C18H28

MANHi


C21H34

HAN


C47H86

VAN

C14-naphthalene

C14H16

MA2Lo

C18-naphthalene

C18H24

MA2Hi

C21-naphthalene

C21H30

HA2

C47-naphthalene

C47H82

VA2

C14-1

C14H18

MAN2Lo

C18H26

MAN2Hi

C21H32

HAN2

C47H32

VAN2

C14H14

MA2NLO

Para Xylene

CnH2n-8

CnH2n-12

CnH2n-10

ring-Arom-2-ring Naphthene C18-1 ring-Arom-2-ring Naphthene C21-1 ring-Arom-2-ring Naphthene C47-1 ring-Arom-2-ring Naphthene C14-2

CnH2n-14

ring-Arom-1-ring Naphthene

110


Component

Formula

Abbreviation

C18-2

C18H22

MA2NHi

C21H28

HA2N

C47H80

VA2N

C21-3ring-Arom

C21H24

HA3

C47-3ring-Arom

C47H76

VA3

Fluorene, 9-methyl

C14H12

MANALo

C18H20

MANAHi

C21H26

HANA

C47H78

VANA

C21-4ring-Arom

C21H18

HA4

C47-4ring-Arom

C47H70

VA4

C21-1

C21H30

HAN3

C47H82

VAN3

C21H24

HA2N2

C47H76

VA2N2

C4H4S

LTH

Class

ring-Arom-1-ring Naphthene C21-2 ring-Arom-1-ring Naphthene C47-2 ring-Arom-1-ring Naphthene CnH2n-18

CnH2n-16

CnH2n-24

CnH2n-12

ring-Arom-3-ring Naphthene C47-1 ring-Arom-3-ring Naphthene C21-2

CnH2n-18

ring-Arom-2-ring Naphthene C47-2 ring-Arom-2-ring Naphthene Sulfur Component Thiophene C8-Cyclo-sulfide

C8H16S

LS8

C12-Cyclo-sulfide

C12H24S

MS12

C28-Cyclo-sulfide

C28H56S

HS28

Benzothiophene

C8H6S

LTHA

Benzothiophene, dimethyl-

C10H10S

MTHA

C10-tetarhydro-benzothiophene

C10H12S

MTHN

C14-trtrahydro-dibenzothiophene

C14H16S

MTHAN

C21-trtrahydro-dibenzothiophene

C21H30S

HthAN

C14- dibenzothiophene

C14H12S

MthA2

C21- dibenzothiophene

C21H26S

HthA2


111

Component

Formula

Abbreviation

C47-tetrahydronaphthabenzothiophene

C47H84S

VthA2N

C47-naphthabenzothiophene

C47H72S2

VTHA3

Pyrrolidine (non-basic Nitrogen)

C4H9N

LBNit

Pyrrole (basic nitrogen)

C4H5N

LNNit

Quinoline,

C9H11N

MBNITN

C9H7N

MBNITA

C9H9N

MNNitA

Phenanthridine, tetrahydro-

C21H33N

HBNitAN

Phenanthridine

C21H25N

MBNitA2

Carbazole, dimethyl-

C21H27N

MNNitA2

C35H55N

VBNitA2N

C47H73N

VNNitA3

Class

Nitrogen Component

1,2,3,4-tetrahydro- (non-basic) Quinoline (basic)

Table B: Component Slate for Hydrocracker Model Only in the Feed Component

Cumene

112

Formula

Abbreviation

C6H12

C6-olef

C7H14

C7-olef

C8H16

C8_OLEF

C8H8

C8A_OLEF

C10H20

C10_OLEF

C10H16

C10N_OLE

C10H10

C10A_OLE

C14H28

C14_OLEF

C14H26

MN1Lo_OL

C14H20

MA1Lo_OL

C18H36

C18_OLEF

C18H34

MN1Hi_OL

C18H28

MA1Hi_OL

C21H40

HN1_OLEF

C21H34

HA1_OLEF

C26H52

C26_OLEF

C47H94

C47_OLEF

C47H92

VN1_OLEF

C47H86

VA1_OLEF

Class


Kinetic Framework In Aspen Plus Hydrocracker/Hydrotreater, each catalyst bed is modeled as a separate reactor. The reaction mechanism is coded in Aspen Reactors, an open-equation modeling platform in which kinetic constituents are segregated from hydraulic and heat balance relationships. This segregation permits different kinetic schemes to be implemented within the same mechanical framework. Rate equations are based on the Langmuir-Hinshelwood (adsorptionadsorption/reaction/ desorption) mechanism. H2S inhibits HDS reactions, and both NH3 and organic nitrogen inhibit acid-catalyzed reactions. Trickle-bed hydrodynamics are modeled with equations described by Satterfield. Collocated reaction rates and collocated flashes enhance the ability of the model to calculate heat release accurately.

Reaction Pathways Aspen Plus Hydrocracker and Aspen Plus Hydrotreater model the following reaction types: 

Hydrodesulfurization (HDS)



Hydrodenitrogenation (HDN)



Saturation of aromatics (Hydrogenation)



Ring opening



Ring dealkylation



Paraffin hydrocracking



Saturation of olefins

When required, hydrodemetalization (HDM) is modeled with a relatively simple extent-of-reaction block. The Aspen Plus Hydrocracker/Hydrotreater reaction scheme has the following important characteristics: 

45 reversible aromatics saturation reactions



19 irreversible olefins saturation reactions



Saturation and dealkylation of non-basic nitrogen lumps



Dealkylation and HDN for basic nitrogen lumps



Saturation and dealkylation for hindered sulfur lumps



Dealkylation and HDS for unhindered sulfur lumps

Figure A-1 shows the importance of modeling aromatics saturation reversibly. Above a certain temperature, equilibrium effects start to outweigh kinetic effects, and additional saturation becomes difficult. This temperaturedependent aromatics crossover causes the degradation of middle distillate properties – kerosene smoke point and diesel cetane – near the end of Hydrocracker catalyst cycles. Figure A-2 illustrates the importance of including both hindered and unhindered sulfur components in the reaction scheme. As discussed in recent publications (15-16), aliphatic sulfur compounds are relatively easy to remove


113

with hydroprocessing; thiophenes, benzothiophenes and dibenzopthiopenes are somewhat more difficult; and substituted benzo- and dibenzothiophenes are very hard to remove. Ref. 15 refers to the Direct Mechanism for the hydrodesulfurization of dibenzothiophene: 

Dibenzothiophene adsorbs to the catalyst surface



The catalyst abstracts sulfur



Biphenyl desorbs from the catalyst surface



Hydrogen removes sulfur from the catalyst as H2S

Alkyl substitution of dibenzothiophene at the 4-position, the 6-postion – or both – sterically hinders this pathway. Before these hindered molecules can be desulfurized, they must first be saturated (which converts a planar aromatic ring into a more flexible saturated ring) or dealkylated. As shown in Figure B below, the Aspen Plus Hydrocracker/Hydrotreater reaction scheme prohibits direct desulfurization of 4,6-alkyl dibenzothiophenes. Figure C below reflects this feature of the models. As the extent of desulfurization increases, hydrogen consumption rises geometrically, in part because the model requires alternative HDS pathways for substituted dibenzothiophenes, and in part because, at the higher required temperatures, other saturation and cracking reactions are accelerated.

Figure A: Aspen Plus Hydrocracker Case Study Showing Aromatics Crossover

114


Figure B: Reaction Pathway Illustration: Sulfur-Containing Components

Figure C: Aspen Plus Hydrocracker Case Study: H2 Consumption vs. Product Sulfur


115

10 Simplified Separation Model

Simplified Separation Model Aspen Plus Hydrocracker applies a simplified separation model to model the fractionation section. The separation status varies significantly between specific plants due to differences in product specifications, operating conditions, and flowsheeting. Therefore, this simplified model uses a set of variables that you can tune to make the products your quality and yields. The concept behind the simplified fractionator model is that the logarithm of the ratio of a component (flow not fraction) in distillate to bottoms (Ln(Di/Bi),) when plotted over temperature (the boiling point, TBi) of that component, yields a straight line; that is, a more or less linear relationship as shown in the figure below.

116


The model is a collection of interpolation and calculator models that calculate the split fraction of a certain component based on: 

The effective cut point.



The component’s normal boiling point.

The model interpolates a value of Ln(Di/Bi) for a given TBi. Usually, the model is configured with three points representing the two straight lines around the effective cut point Where:

Ln(Di/Bi) = 0


117

meaning a 50% split of the component between top and bottom streams. However, to protect from model singularities, two more points are defined at the extremes (one at each end). These points are chosen based on the test run data and remain fixed for daily operation.

118


Index

A

D

Abort button 31 Advanced Process Controller 103 APC 103 API Gravity 59 Aspen Plus 3, 4 Aspen Plus Connection Resetting 14 Aspen Plus Hydrocracker 3 Engine 4 Exiting 15 Simulation Strategy 71 Starting 13 Starting for the first time 11 AspenTech 3 ATSLV File Problem Information 46

Data About Incremental Properties 94 Data About Principal Properties 93 Data About Properties 93 Data Files 16 Loading 17 Saving 16 DCOM 3, 4 DCS 58 Degrees-of-Freedom 6, 8, 9 Detailed Parameterization 69 Development Tools submenu 39 Display Command Line 4, 71 DMO 3, 42, 50 DMO Solver Changing Behavior 99 DOF 6, 8, 9 DP 8

B Bounds 50 Specifying 89, 90 C Catalyst Properties/Data 67 Close button 31 Close Residuals button 31 Command Line window 30, 31 Manual Access 32 Command Window EO Solver Output 44 Component Slate 108 Computer Name 12, 34 Connect Dialog Box 34 Connection to Hydrocracker flowsheet Establishing 12 Constrained Variables 47

Index

E EB Scripts 9 Commonly-Used 73 ECP variables 78 EO Modeling 4 EO Solver 42 algorithm 43 changing parameters 43 log files 46 output to the command window 44 Parameters 44 Equation-oriented See EO Error Recovery 83 Parameterization 84 Excel 11, 13, 15 Excel Interface 15

119

Exit Aspen Plus Hydrocracker 14, 40 F Feed Adjuster Model 21 tuning 78 Feed Analysis 58 Feed Property Tuning 78 FEED Section 73 Feed Specification Change 82 Feed Stream Model 20 Feed System 20 Feeds Button 24 Feedstocks 3 File submenu 36 Flow Diagram Sheet 20, 23, 24, 25, 26, 27, 28 buttons 22 Flowsheet 11 Changes 80 Fractionator hydrocracker 64 G Generating LP Vectors 103 H H2 Balance Button 27 H2 Streams Button 27 HDS 113 Help submenu 40 Hidden Worksheets 29 Viewing 28 HTR Button 26 Hydrocracker Flowsheet Loading 11 Hydrodemetalization 113 I Independent Variables 50, 89 Infeasible Solutions 50 Iterations objective function value 47 status information 48

K Key Operating Data 66 Kinetic Framework 113 L Largest Unscaled Residuals 47 Load Case Data 37 Load Hydrocracker Flowsheet 33 Load User Input Sheet 37 Loading Data Files 17 LP factors 103 LP Vectors 102 generating 103 M Measurements 8 Microsoft Excel 3 Model CONST specifications 73 Model Parameterization 55 Model Running Mode Alternatives 78 Model Specifications 6 Model Tuning Facts 77 Modes 7, 8 Multi-Mode Specifications 8 N No Creep button 31 Nonlinearity Ratio 49 Notes on Variable Bounding 54 O Objective Functions setting up 90, 91 Olefins Reactor Models 21 Optimization 9, 89 analyzing solutions 100 bounds 95 executing cases 98 variables 95 Overall Plant Material Balance Sheet Tuning Runs 61 P Parameterization Case 67 Running 67

120

Index

Parameters 8 PML 3 Pressure Drop Model Example 5 Pricing Information 67 Process Details 7 Process Model Library 3 Process Overview Button 23 Process Specification 58 Strategy 58 Product Analysis 58 Product Properties 55 Q QP 48 R R1 Button 25 R2 Button 25 Reaction Pathways 113 Reaction Rate Tuning Strategy 77 Reaction Section 21 Changing 80 Reactor Bed Models 21 Reactor Profiles button 23 Reconciliation Cases 69 Recycle Stream Data 67 Reset ApMain 14 Run Cases submenu 38 Running Multiple Cases 86 Run-Time Intervention 54 RXN Section 74 S Save Case Data 37 Save to Prior Button 26 Save User Data to File dialog 16 Save User Input Sheet 37 Saving Data Files 16 Scaling 52 Scripts See EB Scripts Second Reactor 80 Turning Off 80 Select Spec.Options dialog 24 Separation Model Tuning 78 Separation Section 22 Changing 81 Sequential-Modular 4 Setup Cases submenu 38 Setup Optimization Case dialog 95 Simple Parameterization 69

Index

Simplified Separation Model 63, 116 Simulation 71 Simulation Case 82 Singularities 52 SM 5 Specification Options 9 Specification Options button 24 Specifications 9 Changing 9, 24 on Flow Diagram sheet 57 SQP 42 Starting Aspen Plus Hydrocracker 13 for the first time 11 Startup Aspen Plus Hydrocracker submenu 14, 33 Startup Options 35 Step Bounds Setting 90 Successive Quadratic Programming 42 T Tools submenu 39 U Unit Mechanical Data 67 Usage Notes 49 V VBA 3 Viewing the model 57 W Worksheets 19 Y Yields Button 28

121

Aspen Plus Hydrocracker_User's Guide V7.3.pdf

Recommend Documents