2 Condition Based Maintenance

Condition Based Maintenance

Dr. Ir. M. Sabri, MT

Definisi dan Tujuan CBM Perawatan Berbasis Kondisi (Condition Based Maintenance, CBM) adalah seperangkat tindakan perawatan berdasarkan real-time atau penilaian waktu terdekat-sebenarnya dari kondisi peralatan yang dapat diperoleh melalui sensor tertanam dan / atau tes eksternal & pengukuran yang dilakukan dengan alat portabel. “Tujuan dari strategi CBM adalah untuk menjalankan perawatan hanya jika terdapat bukti obyektif dari kebutuhan, sambil memastikan keamanan, keandalan peralatan dan pengurangan biaya total kepemilikan.” kepemilikan. ”

What is CBM? Proses condition based maintenance (CBM) memerlukan teknologi, keterampilan, dan komunikasi untuk mengintegrasikan semua kondisi data peralatan yang tersedia, seperti: data diagnostik dan kinerja; sejarah pemeliharaan; log operator, dan data desain, untuk membuat keputusan yang tepat waktu tentang persyaratan pemeliharaan pemeliharaan utama/peralatan penting. Metodologi baru ini telah dikembangkan bertahun-tahun, yang sebenarnya, telah berkembang dari metode pemeliharaan terdahulu selama tiga dekade terakhir.

CBM Phillosophies • CBM menganggap bahwa semua peralatan akan menurun kondisinya dan bahwa akan terjadi hilangnya sebagian atau seluruh fungsi • CBM monitors the condition or performance of plant equipment through various technologies technologies • The data is collected, analyzed, trended, and used to project equipment failures • Once the timing of equipment failure is known, action can be taken to prevent or delay failure • Condition based maintenance uses various process parameters (e.g. pressure, temperature, vibration, flow) and material samples (e.g. oil and air) to monitor conditions

CBM Goals Tujuan dari perawatan berbasis kondisi adalah untuk mengoptimalkan keandalan dan ketersediaan dengan menentukan kebutuhan untuk kegiatan pemeliharaan berdasarkan kondisi peralatan. Menggunakan "teknik prediksi", teknologi, pemantauan kondisi, dan pengamatan dapat digunakan untuk memproyeksikan ke depan dalam upaya untuk menetapkan waktu yang paling mungkin dari kegagalan dan tindakan ini untuk meningkatkan kemampuan pabrik untuk merencanakan dan bertindak secara proaktif. PDM / CBM mengasumsikan bahwa peralatan memiliki indikator yang dapat dipantau dan dianalisis untuk menentukan kebutuhan kondisi diarahkan kegiatan pemeliharaan. Pemeliharaan berbasis kondisi memungkinkan biaya terendah dan program pemeliharaan yang paling efektif dengan menentukan aktivitas yang benar pada waktu yang tepat.

CBM Benefits • To reduce or eliminate unnecessary repairs, prevent catastrophic machine failures and reduce the negative impact of the maintenance operation on the profitability of manufacturing and production plants • Condition based maintenance has the ability to reduce the actual time required to repair or rebuild plant equipment (MTTR) • A condition-based condition-based predictive maintenance program prevents serious damage to machinery and other plant systems (MTBF) • Provides the means to verify the purchased condition of new equipment or offsite offsite rebuilt before acceptance • Predictive data can provide the information required to plan the specific repairs and other activities during the

CBM Advantages • A principal advantage of CBM is the capability it offers the user to perform inspections while the equipment is operating. • Although the technical knowledge required for CBM inspections is usually higher than those for PM, the inspection time required per asset is much less. • When comparing cost advantages of CBM over PM, consider customer downtime costs, maintenance labor costs, maintenance materials costs, and the cost of holding spare parts in inventory.

CBM Objectives • Improve availability i. Reduced forced outages ii. Improve reliability reliabili ty • Enhance Equipment Life i. Re Redu duce ce wear wear from from frequ frequen entt rebu rebuild ilding ing ii. Minimiz Minimize e potential potential for for problems problems in disassem disassembly bly and reassembly iii. Detect problems problems as they occur occur • Save Maintenance Costs i. Reduc educed ed repai epairr cos costs ts ii. Reduc educed ed over overti time me iii. Reduced parts inventory inventory requiremen requirements ts

CBM Targets Kondisi Berbasis Perawatan bukanlah pengganti metode manajemen pemeliharaan yang lebih tradisional. bagaimanapun, Hal ini, jadi tambahan yang berharga untuk keseluruhan program pemeliharaan pabrik yang komprehensif. Dimana program manajemen pemeliharaan tradisional mengandalkan servis rutin semua mesin dan respon yang cepat terhadap kegagalan tak terduga, kondisi berdasarkan jadwal program tugas pemeliharaan khusus yang benar-benar dibutuhkan oleh peralatan pabrik. CBM tidak bisa menghilangkan keseluruhan kebutuhan program pemeliharaan tradisional, yaitu mendekati-kegagalan dan preventif, pemeliharaan prediktif dapat mengurangi jumlah kegagalan tak terduga dan menyediakan alat penjadwalan yang lebih dapat diandalkan untuk tugas-tugas pemeliharaan preventif rutin.

Maintenance Strategies

TDM versus CDM • Time-directed maintenance (TDM) Attempts to avoid failures by retiring, replacing or overhauling components components at a specific age (TBO)

• Condition-directed maintenance (CDM) Attempts to avoid failures by monitoring component condition to detect potential failures before they become functional failures

• CDM is always more efficient than TDM TDM should be used only when CDM is infeasible

When to Use CBM ■ Consider the variety of problems (defects) that develop in your equipment. ■ Use the predictive method if a predictive tool is adequate for detecting the variety of maintenance problems you normally experience. One or a combination of several CBM methods may be required. ■ Use PM if it is apparent that CBM tools do not adequately apply. Inspection tasks must be developed that reveal the defects not adequately covered by preventive maintenance. ■ After you have decided the combination of inspection methods, determine the frequency at which the particular inspection tasks must be applied.

Decisions Tree

CBM Demands • Real-time application • High reliability • At early stage, alert when fault is impending, so that maintenance can be planned when asset is not being used • Identification of the fault and tell where the fault is located • Classify faults in different categories, when a fatal fault occurs automated shut down should be a possibility • The alerts should be easy to understand • The system should be connected to a superior computer

The CM data collected is used in one of the following ways to determine the condition of the equipment and to identify the precursors of failure: • Trend Analysis. Reviewing data to see if a machine is on an obvious and immediate “downward slide” slide” toward failure. For trending purposes, a minimum of three monitoring points before failure may reasonably be expected are recommended. Three data points allow one to determine whether equipment condition depreciates linearly. • Pattern Recognition. Looking at the data and realizing the causal relationship between certain events and machine failure. For example, noticing that after machine x is used in a certain production run, component ax fails due to stresses unique to that run. • Tests against Limits and Ranges. Setting alarm limits (based on professional intuition) and seeing if they are exceeded. • Statistical Process Analysis. If published failure data on a certain machine/ component component exists, comparing failure data collected on site

CBM Implementation Process Conditions necessary for a successful implementation process are typically ones of culture change and change management. They require substantial efforts by all site personnel and management. C o m m i t m e n t - The staff must have commitment to the process and its new technologies as well as their use. Staff has to trust the training and the technology. Management must have the commitment to procure adequate equipment and training for the Staff. Participation - All groups must participate in the program. The organization’s support organization’s support for condition based maintenance must be 100% to achieve success. Management has to reinforce this expectation. H o l i s t i c a p p r o a c h - It applies to all systems throughout the plant. No exceptions. Sus tainability - The program, its staff, its equipment has to be maintained through time to ensure the long term benefits of the process. As people move into and out of organizations, the needed resources must be available. This includes the management support and attitude to trust and maintain it.

CBM Implementation Process




CBM Technologies

Condition Monitoring Technologies

Vibration Analysis Today, electronic instrumentation is available that goes far beyond the human limitations with which the old time craftsperson had to contend when trying to interpret vibration

signals

with

a

screwdriver-handle-to-the-ear

method. Today’s instruments Today’s instruments can detect with accuracy and repeatability, extremely low amplitude vibration signals. They can assign a numerical dimension to the amplitude of vibration and can isolate the frequency at which the vibration is occurring. When measurements of both amplitude and frequency are available, diagnostic methods can be used to determine the magnitude of a problem and its probable cause.

Vibration Analysis

Early detection of mechanical fatigue and breakdown

Vibration Analysis

Vibration Analysis issues that can be found – EARLY! EARLY!

Vibration Analysis

Sample Platforms

Vibration Analysis Vibration Analysis will find defects …

Avoiding disassembly and Avoiding averting unplanned downtime

Vibration Analysis

Vibration Analysis When you use electronic instruments in organized and methodical programs of vibration analysis you are able to: ■ Detect asset problems long before the onslaught of failure ■ Isolate conditions causing accelerated wear ■ Make conclusions concerning the nature of defects causing asset problems ■ Execute advance planning and scheduling of corrective repair so that catastrophic failure may be avoided ■ Execute repair at a time which has minimum impact on operations

Thermography Infrared thermography has grown by leaps and bounds in the past 10 years. Equipment is easier than ever to use and more effective. The real power of thermography is that it allows quick location and monitoring of problems. It presents critical decision-making information in visual form making it easy for management to understand. Infrared imaging systems, as they are generally called, produce a picture, either black or white or color, of the invisible thermal patterns of a component or process. These thermal patterns, when understood, can be used to monitor actual operating conditions of equipment or processes. processes.

Thermography A widely used tool in all facets of industry to measure anywhere a fault can be predicted by a temperature differential.

Non-destructive tool in the analysis and evaluation of electrical distribution Non-destructive equipment.

Reference point of equipment temperature under normal operating conditions.

Thermography Using infrared thermography. Thermography can be used to quickly locate equipment and process problems and in preventing the recurrence of the following problems: ■ Catastrophic electrical failures ■ Unscheduled electrical outages or shutdowns ■ Chronic electrical problems in a piece of equipment or process ■ Excessive steam usage ■ Frozen or plugged product transport lines ■ An inability to predict failures accurately Inefficient use of downtime maintenance opportunities

Thermography ■ Friction failures in rotating equipment ■ Poor product quality due to uneven hearing or cooling or moisture content ■ A fire in a wall or enclosed space ■ Inability to locate or verify a level in a tank ■ Replacement of refractory in a boiler, furnace, or kiln ■ A leaking flat roof roo f ■ Uneven room temperatures affecting product quality or employee productivity ■ Trouble locating underground water, steam, or sewer

Thermography What can Thermography find?

What you see

What thermography thermography sees s ees

Overheating electrical connection indicates a serious fire hazard.

Thermography

Indication of bearing overheating, will eventually cause failure

Thermography Can you afford not to have a Thermography Thermograph y Survey?

Fire damaged electrical system

Cruise Ship Fire Damage

Oil Analysis The spectrometric oil analysis process is a laboratory technique which uses various instruments to analyze a used oil sample from a machine.

A spectrometer spectrometer is used to show when a significant wear mode is underway. Some varieties of these instruments can isolate up to perhaps 80 different types of metals in a sample. The spectrometric result is compared to a baseline level of metal found to be typically

suspended in the oil under normal operating conditions. An important feature of the spectroscopic method is that it not only determines the amount of metal in the sample but also the type. The analysis not only permits the discovery of severe wear, but also analysis of the possible

location of severe wear in a machine. With this information you can take timely action to prevent further deterioration, either by oil purification, oil replacement, or some other means appropriate to the problem.

Oil Analysis Oil Analysis is a non-destructiv non-destructive e test used to assess the condition of lubricants and determine the type and amount of contamination present. present.

Criticality of lubrication to most industrial equipment, oil analysis trending over time is one of the most powerful predictive tools for identifying potential failures.

3 basic categor categories ies of elements affecting the lubrication effectiveness: effectiveness: wear metals, contaminants, contam inants, and additives.

Oil Analysis In addition to the spectrometric analysis, the oil laboratories also check the oil using common oil analysis techniques. For example, the oil is usually checked for viscosity. If the viscosity of the oil has changed 5 percent from new oil of the same type, it is probably time for an oil change due to contamination. Typical of these types of analysis are total acid number, percent moisture, particle count (for hydraulic systems), total solids, and percent silicon (representing dirt from the atmosphere in the form of silicon dioxide or perhaps just from an additive). With prior agreement, special tests are performed at additional cost.

Oil Analysis Benefits

 Improve

oil sampling methods with emerging technologies  Improve machine condition and reliability with oil analysis  Increase the remaining useful life of your lubricant  Reduce maintenance costs associated with unplanned

Oil Analysis Consequences

Oil Analysis Early detection with oil analysis can allow for corrective corrective action such as repairing an air intake leak before major damage occurs.

One of the major advantages of an oil analysis program is being able to anticipate problems and schedule repair work to avoid downtime during a critical time of use.

Oil Analysis The relatively low costs of spectrometric oil analysis make it a very valuable and commonly used CBM method. It is practical for asset systems and subsystems that have a reasonable inventory of oil and are provided with “reuse” “reuse” methods of oil application such as circulating, bath, splash, flood, and ringoiling designs. In addition to providing advance warning when severe wear occurs, spectrometric oil analysis can give important assistance in machine lubrication programs. Conventional oil change frequencies are arrived at according to the traditional PM approach based upon time, hours, or miles operated. A best estimate (or guess) is made of the time or calendar period over which the oil in the system will degenerate and require changing under current operating conditions. Spectrometric oil analysis enables you to change oil only when the actual condition of the oil requires.

Condition Monitoring Standard

Challenges in Implementing CBM • Accurate and Full Utilization of Data • Variable ariable Operating Operating Patterns • Complexities of Real-life Systems • Documentation • Customer Customer Specific User Profiles

Opportunities in Implementing CBM • Establish Root Root Causes of Failures and the Consequences

• Reduce the Life Cycle Cost of Systems in the Field

CBM Reporting • Action Report Containing: Containing: - Asset Asset ID, Date, Current Alarm Alarm Status, Fault, Action, Priority • Missed Measurement Report listing: - Asset Asset ID, Date and reason measurements missed • Typical Action/Advisory Priorities:-

Typical Action Report

CBM Assessment

CBM Design

Framework for CBM Design Condition Based Maintenance Design

Selection of a Unit to monitor

Selection of the Condition Indicator(s)

Selecting a Prognostic Modeling approach

Determination of the Maintenance Policy

• Selection of a Unit to monitor - this selection process should be based on the potential benefits of a CBM program for a specific unit, and the impact of the failure modes of this unit on the overall condition of a system. • Selection of the Condition Indicator(s) - to obtain insights into the relation between the failure mode(s) of a specific unit and the related deterioration parameter(s). • Selecting a Prognostic Modeling approach - these models are divided into Knowledge based models, life expectancy models, Artificial Neural Network (ANN) models, and physical models. • Determination of the Maintenance Policy - determination of the maintenance policy, policy, the overall CBM program for the specific unit can be evaluated on profitability.






As the asset base of a company can be large and systems complex, criteria need to be formulated such that precious resources, time and money are spent on the most important assets of a company. Over different assets, the impact of a CBM program on the operational and technical costs can depend on numerous factors. factors.

Selection of a Unit to monitor • System Decomposition Decomposition of the system into lower level units (i.e. components, items) for which the failure modes and interactions are comprehensible is key at the start of a CBM program. By gaining insights into these failure modes of the specific units, gradually, an overall condition parameter for the system can be developed. Depending on the level of knowledge of these failure modes, analysis on the item, component, or system level is possible. For example, the wheels of a car can be considered as a unit, but also the entire steering mechanism (i.e. all components enabling the function of steering) can be considered to be the unit for analysis. Depending on the way of defining the unit, different failure modes will have different effects on the unit and will have different causes depending on the level of decomposition.

Selection of a Unit to monitor • Criteria Analysis As the field of maintenance management is concerned with production, finance and quality within organizations, multiple criteria can be involved in maintenance decision making. Maintenance Maintenance personnel personnel can consider the unit with the highest frequency of failures to be the most critical unit while employees of the production department can consider the unit with the longest downtime to be the most critical one. Also, from a financial point of view, the spare parts consumption can be considered to be a criteria for defining the criticality of a unit.

Selection of a Unit to monitor • Multi Criteria Decision Analysis Support Tools Tools Different MCDA methods exist depending on the decision making situation, and the information available. This also includes different techniques for determining the criteria weights. The overall structural elements of a MCDA are criteria to compare different alternatives with, alternatives to decide upon, stakeholders and decision makers involved in the process, uncertainty in the decision making process which can be caused by external factors or a lack of knowledge on parameter influences, and the environment (i.e. time and place) at which the decision is made

Selection of a Unit to monitor • Determining the Critical Failure Modes Some commonly used tools in the field of Root Cause Analysis (RCA) are:i. FMEA - supports the process of selecting the most critical failure modes of a specific unit by evaluating the consequential damage related to it. ii. Pareto An Analysis - aims at selecting the actions that have the biggest impact on the overall overall result. iii. iii. Ishi Ishik kawa Diagr iagram am - by providing the relationships between variables of influence on a specific event, these diagrams can give insights into specific process behavior.

Selection of a Unit to monitor • Determining Determining the Critical Failure Modes FMEA

Selection of a Unit to monitor • Determining Determining the Critical Failure Modes Pareto Analysis

Bearing Failures Study

Selection of a Unit to monitor • Determining Determining the Critical Failure Modes Ishikawa Diagram






After discussing the selection of a unit to monitor, it is important to obtain insights into (1) the failure modes of the specific unit and relating these failure modes to a single or multiple deterioration parameters, (2) gather data on the deterioration parameters, (3) gain insights into the deterioration process of a specific parameter, and (4) define a failure threshold level to specify at what condition level the unit is considered to be unable to perform according to functional specifications (i.e. the unit is considered to be in the failed condition).

Selection of the t he Condition Indicator(s) • Relation Between Monitored Parameter and Deterioration In order to make an accurate prediction of the deterioration condition of a system, selection of correlated quantitative variables to the deterioration deterioration is key key. Insights into the failure modes and possible poss ible effects of the system is crucial in this parameter selection process. Also, the failure dependencies can play an important role in the parameter selection process. While some deterioration parameters correlate to a single failure mode, some failure modes will need multiple deterioration measures to provide an accurate indication of the deterioration deterioration condition. It should be noted noted that a situation might arise in which a deterioration parameter cannot be determined, the deterioration parameter cannot be measured, or where it is too costly to measure the deterioration parameter. Hence, an imminent mode of failure cannot be predicted. In these cases, implementation of a CBM program might not be efficient, depending on how critical this parameter parameter is with respect to predicting a system system failure.

Selection of the t he Condition Indicator(s) • Data Acquisition The necessary data collected can be categorized in two types; event data and condition monitoring data. The event data consists of information extracted from information systems (e.g. CMMS, MIS). The collection of event data done by manual data entry into the information system could result in the erroneous entry of data in the system because of the involvement involvement of human input. The condition monitoring monitoring data data can fall into three types of different data categories - Value type data is collected data that is registered as a single value variable. Examples can be oil particles analysis data, temperature, pressure and humidity data. - Waveform type data is data collected as a time series, which is referred to as a waveform. Examples of these type of data are vibration data and acoustic data. - Multidimensional type data is data that is of a multidimensional character. Most of the time this is some sort of image data like infrared thermographs, X-ray images or visual images.

Selection of the t he Condition Indicator(s) • Trend Evaluation For the purpose of predictability, the trend characteristics of the deterioration signal are key. Knowledge on the deterioration behavior over time should be obtained. Depending on this behavior, it is determined what the best model fit for the specific unit is. This can for example be a linear or exponential model. Most important is to observe a specific pattern over time. Often, a phase of normal operation can be defined in the early lifetime of the unit.

Selection of the t he Condition Indicator(s) • Trend Evaluation The Potential Failure – Failure Failure (P-F) Interval

Selection of the t he Condition Indicator(s) • Trend Evaluation Lifetime Characteristics

Selection of the t he Condition Indicator(s) • Determining a Threshold Level Depicts the condition level of the item at which a functional failure is likely to occur. This can happen because of actual failure of the unit, or because the unit fails to meet a functional requirement as specified by the customer or company. Relevant information can be acquired from (1) historical data, (2) from expert opinion, or in case no data is available for determining the threshold level, (3) a tentative threshold level can be selected in order to gain more insights and obtain more experience with the deterioration of the unit.






As the condition parameters defined and deterioration data are gathered, the selection of the prognostic model is considered to be not straightforward within companies as this requires both knowledge on how the model should be used and understanding of the mathematical background of the model. This chapter provides an overview into the different models available for the purpose of prognostic modeling and the relevant selection criteria. As depicted in figure 13, the area of prognostic models can be generally divided into; (1) knowledge based models, (2) life expectancy models, (3) Artificial Neural Network (ANN) models, and (4)

Selecting a Prognostic Prognostic Modeling approach • Knowledge Knowledge Based Models Within the knowledge-based models, the RUL is determined using historical data. These events are compared to a database of failure events based on historical data and expert knowledge. The knowledge-based models can be divided into a class of:-

Fixed Rule Systems The knowledge from experts is gathered and used to define rules within a software program. This can be done by formulating ifthen statements for specific problem situations. Fuzzy Systems Use simple if-then rules based on empirical data to solve problems. However, the rules are defined based on specific conditions. (e.g. if(friction is high) and (temperature builds quickly) then (cool down))

Selecting a Prognostic Prognostic Modeling approach • Life Expectancy Models The life expectancy models determine the Remaining Useful Life (RUL) of a unit by assessing the risk of degradation under known operating conditions. The life expectancy models can be classified into the group of:(1) stochastic models, and (2) statistical statistical models.

Selecting a Prognostic Prognostic Modeling approach • Artificial Neural Networks (ANN) ANN‟s predict the RUL of a unit by using a mathematical representation of the unit which is built from historical observations. Therefore, this approach lacks understanding of the underlying physical failure processes. It is typically used for modeling complex non-linear units by using input data such as process variables, condition monitoring data, unit characteristics and maintenance history data to give the desired maintenance action or the RUL as output of the network model

Selecting a Prognostic Prognostic Modeling approach • Physical Models Physical models assumes that physical laws can be used to quantitatively model the behavior of a unit. This is done by using scientific or empirical knowledge and transforming this into a deterministic equation to estimate the RUL. To enable this, unit specific parameters need to be identified to build this equation. Sensory measurements are used to compare the model with the real-life unit behavior. A drawback of this method is that the failure behavior of a unit is influenced by a lot of different variables.






The goal of CBM prognostics is to gain insights and support into making appropriate maintenance decisions. To enable this, a maintenance policy should be incorporated into the prognostic implementation process. Less mathematical models are applicable to the situation of CBM compared to conventional maintenance policies.

Determination of the Maintenance Policy Based on a minimizing cost, maximizing availability, or maximizing profit objective function, a maintenance policy can be determined. The decision structure to come to an appropriate maintenance policy is illustrated in below figure.

Determination of the Maintenance Policy • RUL (Remaining Useful Life) The output of the selected prognostic model is the RUL. This RUL is considered as the input for operations research decision making. The RUL is the time interval between the actual point in time between obtaining the last data point and the point in time where a failure occurs. This failure time differs for a particular unit. Therefore, it is a random variable following a probability distribution.

Determination of the Maintenance Policy • Condition Monitoring Interval The information on the monitoring interval can be used to update (fixed) intervals which are established in the design phase of a unit, or to use the prognostic information for determining the time until the next maintenance action. Condition monitoring can be done on continuous and on periodic basis. In continuous monitoring, unit data is collected on continuous basis and an alarm signal is triggered when a unit is operating outside of the normal operating operating specifications.

Determination of the Maintenance Policy • Delay Time Within delay time modeling (DTM) it is assumed that failures occur according to a multi-phase deterioration process. These phases can be classified as representing the state in which the unit is operating according to specifications (i.e. “good” state), the state in which a deterioration can be observed (i.e. the “degraded” state), and the unit condition at which the unit is considered to be no more functional (i.e. the “failed” state).

Determination of the Maintenance Policy • Number of Renewals Renewal theory is used to determine the average cost of a maintenance policy where components are assumed to be restored into an “as good as new” state. Within maintenance modeling, replacing an item by a preventive action and repairing an item by a corrective action are assumed to do so

Determination of the Maintenance Policy • Expected Total Relevant Maintenance Mai ntenance Cost The goal of an economic objective function is to minimize the expected total relevant maintenance costs (E[TRMC]) per unit time. Under a CBM policy costs are built up from the following cost factors (1) Costs for PM, (2) Costs for CM (i.e. emergency costs), (3) Downtime costs, (4) inspection costs.

Determination of the Maintenance Policy • Maintenance Objective In order for companies to cope with a variety of risk factors, it is necessary to assign resources, time and money in an optimal way. This means optimization of the maintenance policies. Quantitative maintenance modeling is concerned with mathematical optimization of processes with respect to cost or reliability criterion.

Determination of the Maintenance Policy • Determine Policy Based on the insights in the long term cost per unit time for replacing an item with respect to a certain maintenance objective, the time of maintenance can be determined for the unit. Based on the long term cost per unit time, the CBM policy can be compared to a conventional maintenance policy using fixed replacement intervals. Application of a CBM program can result in savings due to extending the maintenance intervals and by preventing unexpected failures from ever occurring due to the unit condition information. An appropriate way of determining the savings of implementing a CBM program would be determining the remaining service life of the component at the moment of replacement.

CBM Key Step 1

• Cost/Benefit analysis

2

• Carry out equipment audit

3

• Reliability & criticality audit

4

• Select monitoring method

5

• Select monitoring method

6

• Data acquisition and analysis

7

• Determine maintenance action

8

• Review & measure effectiveness

Key Step to Implement Condition Based Maintenance

Cost Benefit Analysis

•

This will highlight where Condition Based Maintenance will reduce costs – Do – Do we have life cycle cost info? – What – What is cost of failures? – What – What is cost benefit of avoiding failure?

Carry out equipment audit

•

Make sure we have accurate databases with assets clearly identified and labeled – The importance of this step is often overlooked – Without clear identification identification of assets most activities activities are compromised

Reliability & criticality audit

•

Carrying out this means we can target the most important

•

What Availability & Reliability does the business need?

•

MTTR – MTTR – Mean Mean Time to Repair MTBF – MTBF – Mean Mean Time Between Failure FMEA – Failure – Failure Modes and Effect Analysis

Select Appropriate Maintenance Strategy

•

Selecting the appropriate combination of maintenance tasks depends on the failure history – Initially the results from Step 3, – Later on from feedback and analysis from Steps 6 – 8 – 8

Select monitoring method

• We need to identify the best parameters to be measured to detect faults • Fault and failure characteristics c haracteristics linked to measurable parameters and symptoms allow us to do this (Output from FMEA in step 3) • We can then select the best measurement technique and then select the appropriate transducers and condition monitoring system

Data acquisition and analysis

• Quality of Measurements OK? • Possible errors come from - poor readings, transducer faults or adjacent machines. • If confidence in readings low: - take more readings or apply other types of CM • Review symptoms, rules etc

Determine maintenance action

• The key output of a CM program is recommended maintenance maintenance actions • Poor feedback allows valuable information to leak away

Measuring Effectiveness

• Managing CBM is a continuous process • Technology & Techniques change • Periodic reviewing the process is an important step

Thank You

2 Condition Based Maintenance

Recommend Documents