Whittle Simultaneous Op Tim is at i on Tutorial

Copyright © 2012 Gemcom Software International Inc. (Gemcom)

All All rights rese reserved. rved. Gemcom Gemcom publishe publishess this documentation documentation for for the sole use of Gemcom Gemcom license licensees. es. Without writt written en permission, permission, you may not sell, sell, reproduce, reproduce, store in a retrie retrieval val system, or transmit any part of this documentation. documentation. For such permission, permission, or to obtain extra extra copies copies please please contact your local Gemcom office, office, or visit www www.gemcomso .gemcomsoftware ftware.com. .com. This software software and documentation is proprietary to Gemcom Gemcom and, except except where expressly provided otherwise otherwise,, does not form part part of any contract. contract. Changes may be made in products or services services at at any time without without notice. While While eve every ry precauti precaution on has been been taken in in the preparation preparation of this manual, manual, neither neither the authors nor Gemcom Gemcom assumes assumes responsibil responsibility ity for errors errors or omissions. Neithe Neitherr will will be hel held d liable liable for any damages damages caused or allege alleged d to be caused caused from from the use of the information information contained contained herein. herein. Gemcom Gemcom Software Software International International Inc., Gemcom, the Gemcom Gemcom logo, combinations thereof thereof,, and GEMS, Surpac, Minex, MineSched, Whittle, PCBC, Gemcom InSite, and Gemcom Hub are trademarks trademarks of Gemcom Gemcom Software Software International International Inc. or its whollywholly-owne owned d subsidiaries. subsidiaries. Product

Gemcom Gemcom Whittle™ Whittle™ 4.5 Last modifie modified: d: Wednesday, 24 October 2012 2012

Table of Contents Introduction

4

Prerequisites

4

Projects folder

4

Processing time

4

Simultaneous optimisation theory

5

Net present value and discounting

5

Costs

5

How simultaneous optimisation works

5

When can you benefit from SIMO

6

Bulk blend

7

Running the existing bulk blend scenario

7

Creating and running a simultaneous optimisation node

7

Examining the schedule results

11

Creating a mining sequence file

13

Extractive blend

14

Running simultaneous optimisation for an extractive blend scenario

14

Viewing the sequence and grades in the 3D viewer

16

Operational scenario

17

Examining a comparative node

17

Setting up simultaneous optimisation for an operational scenario

18

Examining how the NPV result is calculated

20

Examining the average grade input to MILL

22

Introduction

Prerequisites

Introduction In this tutorial, you will add a simultaneous optimisation (SIMO) node, and a schedule report graph node, to an operational scenario node, a bulk blend node and an extractive bulk blend node. You will compare the net present value (NPV) generated by simultaneous optimisation with the NPV generated from the parent nodes.

Prerequisites l

l

You must have a basic understanding of Whittle concepts. If you are new to Whittle, complete one of the two introductory tutorials before trying this Simultaneous Optimisation tutorial. Ideally you will have completed an introductory Whittle training course so that you understand the concepts and procedure of pit optimisation and how to determine optimum pushbacks.

Projects folder For this tutorial you use the bulkBlend, marvin and marvinBlend data sets that are installed with Whittle. These data sets are in the default locations as in the following table, unless you selected a different folder when you installed Whittle. Operating System

Windows 7 or Windows Vista Windows XP

Projects Folder C:\Users\Public\Gemcom\Whittle\ \projects C:\Documents and Settings\All Users\Gemcom\Whittle\ \projects

Processing time Because of the amount of processing required for simultaneous optimisation, it can take several hours to run. In this tutorial, you will change some optimisation settings to speed up the processing so that the example data runs in approximately 15 minutes or less. As a result, the NPV returned is reduced compared to the NPV you can achieve with standard optimisation settings. SIMO runs faster on machines with a faster CPU and more available RAM.

Gemcom Whittle™ 4.5

Page 4 of25

Simultaneous Optimisation tutorial


Net present value and discounting

Simultaneous optimisation theory Net present value and discounting A dollar that you receive today is more valuable to you than a dollar that you expect to receive next year. You can reinvest money you ho ld today, and earn benefits from it. This concept is referred to as the time value of money. When estimating the value of a project, it is common to reduce the value of expected future cash flows by a certain percentage per year, to allow for the time value of money and other factors like risk. This process is called discounting. The sum of all expected discounted cash flows is called the net present value (NPV).

Costs When you are thinking about NPV and discounting, it is helpful to remember the following points about costs: l

l

Same currency. In Whittle, you must enter all costs and prices in the same currency. If your costs are in the local currency and prices in US dollars, convert them to the same currency before you enter them in Whittle. Period variation. If you forecast costs or prices to change in different periods, you can use the Period Variation button in Whittle to specify what cost or price for each period.

The Period Variation button is also useful when you forecast a change in exchange rate, or monetary devaluation, in a currency that is relevant to the mining project. For more information see the Period Variation help topic. l

Initial CapCosts. Initial capital costs are incurred in period 0. Therefore they are not discounted. Conversely, future cash flows are discounted.

How simultaneous optimisation works Simultaneous optimisation (SIMO) is designed to maximise the net present value (NPV) of your mine. Before running SIMO, you must determine: l l

the final pit shell the intermediate pit shells to mine to – that is the pushbacks

You can use the existing Whittle features, such as the Lerchs Grossman pit optimisation, the Milawa algorithm and the NPV practical pushbacks node to determine the final pit and the pushbacks. SIMO concurrently optimises: l l l l

schedule cut-off stockpiles blend


Page 5 of25



When can you benefit from SIMO

When can you benefit from SIMO You gain most benefit from SIMO when: l

l

it is valuable to your operation to process high-grade material at early stages in the mine life. However, even if a higher grade (beyond a prescribed cut-off) does not affect your profits, such as in many iron ore or gypsum operations, SIMO can still improve the NPV because it also optimises the blend. In a schedule where the blend is optimised, material just below the cutoff is saved (blended with higher grade material and either processed or sent to stockpile) rather than sent to the waste dump. your operation is flexible enough to deploy the people and equipment to mine at a higher rate in specific periods. One of the reasons that SIMO can return a high NPV is because it uses cut-off optimisation. In cut-off optimisation, the processing limit stays constant, and the cut-off is raised in specific periods – usually the early periods. To uncover and mine enough material to meet the processing limit you must increase the mining rate (and send the material that is below the cut-off either to the waste dump or to a stockpile). To mine at that rate, you must deploy more people and equipment in the specific periods when they are needed.


Page 6 of25


Bulk blend

Running the existing bulk blend scenario

Bulk blend Running the existing bulk blend scenario 1. Start Whittle. 2. Open bulkBlend1.fxp. 3. Run all the nodes in the project. You will compare the NPV that the simultaneous optimisation (SIMO) node returns with the NPV that you know you can already obtain with this data. Specifically you will use the bulk blend node b) Single blend with blend bins + stockpile as the comparative node. 4. Under b) Single blend with blend bins + stockpile , select Lead 4:phases 3 4 5 6. 5. Select the Summary tab.

This node returns an NPV of approximately $987.6 M.

Creating and running a simultaneous optimisation node You can create a SIMO node underneath the following types of node: l l l

bulk blend scenario extractive blend scenario operational scenario

Before creating the SIMO node under a bulk blend scenario node, you should complete pit optimisation and determine the optimum pushbacks. The pushbacks have already been determined for the Lead 4:phases 3 4 5 6 node, and for this tutorial, you can use the same settings. 1. On the Lead 4 : phases 3 4 5 6 node, Schedule tab, notice that pit shells 3, 4, 5 and 6 have been chosen as pushbacks. 2. Right-click on b) Single blend with blend bins + stockpile, and choose Add > Simultaneous Optimisation. Blend bins, blend constraints and limits have already been defined on the ‘parent’ bulk blend scenario node. These bins, constraints, and limits are ‘inherited’ by the SIMO node and, to compare ‘like with like’ do not change them on the SIMO node. 3. On the Optimisation tab, enter a Period horizon of 15. At this time, you do not know how many years the mining operation will ideally run. Therefore you will initially enter a high number of periods, based on the mine life from a comparative, nonSIMO node. After you have run SIMO once, the Report tab will show you the number of periods that have mining activity according to the SIMO schedule. For optimum performance on reruns,


Page 7 of25


Bulk blend


you want the last period to have no activity at all to show that no more optimisation is possible. However, to minimise processing time, you want only one empty period. 4. On the Optimisation tab, select Advanced optimisation controls. The standard settings for simultaneous optimisation return close to the highest possible NPV. These settings can take several hours to run. For this tutorial, however, you will make it run faster and possibly accept a small reduction in NPV. 5. Set the Minimum Start Points to 5 and the Number of Best Solutions to 5, and click OK. Note: More information on Advanced Optimisation Controls is available in the Whittle help.

6. Click Accept and click the Save icon on the toolbar. 7. Restart Whittle for these settings to take effect. 8. On the New Simultaneous Optimisation node, click the Blend Bins tab. For this project, the blend bins are copied automatically from the parent node. Examine the bins. You can see that blend bins have been created for the L1, L2, and L3 rock types (because these rock types are the only ones included in the product blend). You can also see that the Ash % is used to define the different bins, because Ash% is the most significant element in the blending process for this project. 9. Click the Blend tab.


Page 8 of25


Bulk blend


You can specify blend constraints on this tab. For this tutorial, there is no need to change the blend settings – they are the same as the blend settings specified on the comparative node. However, you will need to add the stockpiles into the blend. 10. At the Stockpiles field, click Edit. The Sources dialog box is displayed. 11. Click All and click OK to add all the stockpiles.

12. On the Schedule tab, enter the same details as you used on the Lead 4:phases 3 4 5 6 node. That is, use the largest pit as the final pit and add pushback for pits 3, 4, 5, and 6.

13. Click Accept in the lower right corner. 14. Run the simultaneous optimisation node. When SIMO is processing, the status bar shows you the number of the solution that SIMO is processing and the NPV of the best solution processed so far. For more information about solutions, see the Whittle help.

The previous image shows SIMO processing solution 23 which has an NPV of approximately $1.04 billion. When processing is complete, a green check mark appears beside the SIMO node.


Page 9 of25


Bulk blend


15. Select the Report tab, and scroll near the end to the list of grand totals.

Underneath Grand.Total.DiscountedCashflow, you can see the total run time — that is, the amount of time the SIMO node took to run.

This SIMO node in the sample bulk blend data, took 5 minutes to run on one computer. In some situations, SIMO can take much longer to process, and if there is the potential to increase the NPV of a mine by millions of dollars, the time is well spent. Higher up on the Report tab, the report shows you the periods in which there was mining activity. For example, period 15 had no mining activity.


Page10 of25


Bulk blend


If you scroll up, you see more periods where there was no mining activity. Which was the last period that does have mining activity? Tip: If you are going to run this SIMO node again, you can make SIMO run slightly faster by adjusting the Period Horizon. On the Optimisation tab in the Period Horizon box, enter one more than the number of the last period with mining activity.

Examining the schedule results There are two types of report you can use to examine the schedule. Report

Description

Schedule Report

Shows details of material mined in each period .

Bench Report

Shows details of material mined in each bench in each period .

To examine the results of the SIMO schedule in a schedule report node, do the following. 1. Right-click the New Simultaneous Optimisation node and choose Add > Schedule Report. 2. Click the Definition tab of the Schedule Graph node. The Definition tab lists a default set of variables that you can display using the Graph tab. You can use the Add button to display the Data Selector to add a different variable, but there is no need to do so at this time in this tutorial. 3. Run the Schedule Report Graph. 4. Click the Output tab. The Output tab reports the values per period for each of the variables on the Definition tab.


Page11 of25


Bulk blend


Tip: To output this data to a spreadsheet, click the Open in Spreadsheet button toolbar.

on the

5. Click the Summary tab. A formatted summary of the output is displayed.

In this case, the SIMO NPV is $1041 M which compares with $987 M on the lead 4 node. To see more information about which benches are being mined, add a bench report. 1. Right-click the New Simultaneous Optimisation node, and choose Add > Bench Report. 2. Select the New Bench Report node and run it.


Page12 of25


Bulk blend

Creating a mining sequence file

3. Click the Output tab to view the results.

Creating a mining sequence file By adding a schedule report graph node you create a long-term schedule that you can export as a mining sequence (.msq) file for use in a detailed scheduling program. 1. Right-click the Schedule Report graph node, and choose Other > Bulk Export . 2. Select the checkbox for an msq file, and click Run. 3. When you see the prompt that the export has completed successfully, click OK. The .msq file is created in your project folder.


Page13 of25


Extractive blend


Extractive blend To run simultaneous optimisation (SIMO) on an extractive blend scenario, the procedure is the same as for a bulk blend scenario, and the same procedure as for a standard operational scenario.

Running simultaneous optimisation for an extractive blend scenario 1. Open marvinBlend1.fxp. 2. Run all the nodes in the project. 3. Under the Add Stockpiles node, which is an extractive blend scenario node, select the Maintain soluble copper < 25% + stockpile node and select the Summary tab.

This node returns an NPV of $384M. 4. Add a simultaneous optimisation node under Add Stockpiles. 5. Set it to have the following settings: In the Optimisation tab; Period horizon: 13 Advanced Optimisation Settings, Minimum Start points: 5, Maximum Start Points: 10, Number of Best Solutions: 5 (to make it run faster at the possible expense of NPV). In the Blend tab; add all the stockpiles to the blend. In the Schedule tab; Final Pit: 35 l

l l

l l

l


Page14 of25


Extractive blend


l

Pushbacks: 1, 6, 14, 19, 28, 31, 35

6. Click Accept to save the changes. Before you run this node, be aware that it can take approximately 11 minutes to run. 7. Add a Schedule Report node under the New Simultaneous Optimisation node. 8. Run to the Schedule Report graph node. The node returns an NPV of $460M compared to the $384M of the Maintain soluble copper < 25% + stockpile node.


Page15 of25


Extractive blend

Viewing the sequence and grades in the 3D viewer

Much of this additional value comes from mining higher grade material in the early periods. You will examine this in more detail when you add a SIMO node under an operational scenario node.

Viewing the sequence and grades in the 3D viewer 1. Select the Schedule Report graph node beneath the SIMO node. 2. Click on the toolbar to start the Three-D Viewer. The Select data to display form is displayed. 3. Click OK. 4. To see which areas are mined in each period, select Show Period, and change the period number to cycle through the periods. 5. Select Au in the box under Show Data, and fill in the form as shown in the following image. You can use the Three-D Viewer to see when and where the higher-grade blocks are mined. Note: Click Invert if you want to give the background a white background. A white background is useful when you want to paste an image of the Three-D Viewer into a document and produce a better image when it is printed.


Page16 of25



Examining a comparative node

Operational scenario You can use a simultaneous optimisation (SIMO) node beneath an operational scenario node. First, you will examine the net present value (NPV) you can achieve using only three stockpiles, one for each rock type: OX, MX and PM. Then this tutorial will show how you can improve NPV using simultaneous optimisation.

Examining a comparative node Suppose that you have completed pit optimisation and decided that the best pushbacks to use are 11 12 14 16 20 25 and 31. 1. Open marvin1.fxp. 2. Run all nodes. 3. Under the Buffer stockpile example node, examine the P/B 11 12 14 16 20 25 31 – Lead 4 node, Schedule tab, and Summary tab.

The NPV returned using a Fixed Lead 4 algorithm is approximately $440M.


Page17 of25




Before you try simultaneous optimisation, you might wonder whether you can achieve a higher NPV just by using a different algorithm. The Milawa algorithm typically returns a result with a high NPV, so try that. 4. Select the P/B 11 12 14 16 20 25 31 – Milawa node, and view the Summary tab.

The Milawa algorithm returns $445M.

Setting up simultaneous optimisation for an operational scenario Because you do not know how many years of mining activity there will be, you will set the period value to 15 years, which is the mine life in the comparative nodes. In the following procedure, the nodes you do not need have been deleted to make it easier to see the data of interest. 1. Add a SIMO node beneath the Buffer Stockpile Example node.


Page18 of25




2. Set Period Horizon to 15. 3. Set the following advanced optimisation settings: Minimum Start Points: 5, Maximum Start Points: 5, Number of Best Solutions: 5 (to make it run faster at the possible expense of NPV). 4. Generate a suitable number of blend bins. a. On the Blend Bins tab, select Use equiv. metal grade, and click Generate. b. Set the number of blend bins for each element to 10.

With ten blend bins for each element, you get a good balance between speed of processing and NPV returned for the Marvin data set. In a final optimisation, you could see whether a larger number of blend bins returns a higher NPV. 5. 6. 7. 8. 9.

c. Click OK. On the Schedule tab, set the Final Pit to 31 (the same pit number as used in the comparative pushbacks node). Add pushbacks 11 12 14 16 20 25 and 31. On the Description tab, change the description of the node to SIMO 11 12 14 16 20 25 31 so you can easily see which pushbacks it uses. Add a Schedule Report node beneath the SIMO node. Click Accept. It might take approximately 9 minutes to run this node.

10. Run all nodes.


Page19 of25




The Schedule Report graph node reports an NPV of $469M, as compared to $440M from the Lead 4 comparative node and $445M from the Milawa comparative node.

Examining how the NPV result is calculated You can validate and examine the results that simultaneous optimisation returns using the Definition tab of the Schedule Report graph node, and the Data Selector. For example, here is how you can find out which components make up the final NPV amount. 1. On the Schedule Report graph node, select the Definition tab. Some variables are already listed on this tab. If you wish, you can leave them there. 2. Click Add and use the Data Selector to add the following variables: User Specified > Open Pit > Capital > Initial capital expenditure Calculated > Open Pit > Output > Discounted revenue and cash flow > Discounted revenue from MILL Calculated > Open Pit > Output > Discounted costs > Discounted selling cost of product from MILL Calculated > Open Pit > Output > Discounted costs > Discounted replacement capital expenditure Calculated > Open Pit > Material Mined > Discounted costs > Discounted cost of mining all rock


Page20 of25




Calculated > Open Pit > Material processed > Discounted costs > Discounted cost of processing for MILL Alternatively, you could add variables based on element, such as Discounted Revenue from Au and Discounted Revenue from Cu. However, because there is only one processing method (MILL) and two elements, the set of variables listed in the previous step is easier to work with.

3. Run the node. 4. Select the Output tab. 5.

Click Open in Spreadsheet on the toolbar to send the output data to Microsoft Excel. 6. To remove the columns you are not using, delete columns B to M from the spreadsheet. 7. Use Autosum in Excel to calculate the total for each column. 8. In the spreadsheet, calculate the total cost and subtract it from the revenue.


Page21 of25




Look back at the NPV reported on the Schedule Report graph node, Summary tab, which is also approximately $469.7M. The currency settings on the Formats tab of the block model node determine the format in which currency values are reported. In this tutorial, currency values are displayed in thousands of dollars. Therefore, currency values can be slightly higher or lower than you expect because of rounding variations.

Examining the average grade input to MILL Simultaneous optimisation can deliver a higher NPV than other solutions, partly because it allows the economic cut-off to change in each period rather than to have the same cut-off for all periods. With this flexibility, SIMO schedules the mining of more of the higher grade material early in the mine life. You can see this effect by comparing the grades input to the MILL. 1. Add a Schedule Graph node underneath Buffer stockpile example. 2. On the Schedule tab, specify the same final pit and pushback details that you have used in the other nodes.


Page22 of25




3. In Specified Case Scheduling Algorithm select Milawa NPV , and click Accept. 4. Run the node. 5.

Select the Output tab, and click Open in Spreadsheet on the toolbar. The grade information is displayed in a spreadsheet. The other columns are of no interest for this exercise.

6. Now select the Simultaneous Optimisation > Schedule Report graph node, and output the Output tab to a spreadsheet similarly. Note: Alternatively, you can get the grade data by adding and running a Bench Report node under the SIMO node. 7. Now copy the Grade input to MILL columns from one spreadsheet to another to compare the grades.


Page23 of25




8. To see this data in a chart, use the charting tools in Excel to make a line chart that shows both sets of grades.


Page24 of25




The charts show that, when you use the SIMO solution rather than the Buffer Stockpile > New Schedule Graph solution, higher grades of gold and copper are processed through the mill in the first nine periods.


Page25 of25


Whittle Simultaneous Op Tim is at i on Tutorial

Recommend Documents