System Health Prognostics - Research Paper Example

Running Head: System health prognostics System Health Prognostics of the of the Project proposal System health prognostics are a set of actions performed on a system to preserve it in operable condition. Prognostics may be limited to the examination of current system states, with maintenance and repair activities driven by these examinations. This paper will discuss how to improve a current method or develop a new one. The paper will divided into different section with headlines and sub-headlines for the better understanding of the subject. The Introduction will provide a general overview of the topic and discuss the factors behind the rapid change in the manufacturing environment around the globe. The following section, Maintenance Management, will be divided into four subsections which will discuss classification of maintenance theories and their cost analysis. The section three, Literature Review, will look into the works of different writers. This will tell how literature has played an important role in the development of maintenance theories. Standards in Maintenance, the fourth section, will highlight the role of organizations in the implementation of maintenance strategies. The following section, Health Prognostics — Methods and Techniques, will help in understanding the methods and techniques of maintenance. This section will also discuss the role of traditions in the maintenance strategies. System Monitoring in Industrial Applications, the sixth section, will highlight the role of production operators in detecting the equipment failure. The last section, Conclusion, will sum up the points discussed in the paper. The sources which will be used in the paper will be referenced in alphabetical order at the end of the paper. System health prognostics are being expanded with forecast of future operating states and analytical finding of future failure states. This kind of analytical finding or system health prognostics are motivated by the requirement for manufacturers and other operators of complex systems for improving performance of equipment and reducing costs of maintenance and surprise failure of equipment. Health system prognostics are difficult tasks requiring accurate, adaptive and instinctive models to forecast future machine health states. A number of modelling procedures have been suggested in the literature and put into practice. This paper will discuss how to apply available methods to a practical problem. The paper will also review the ideas and procedures that focus on improving dependability and reducing sudden failure by monitoring and forecasting machine health. The Introduction of the paper will give general view of maintenance strategies. The other sections will discuss tools and techniques of maintenance strategies. A vital part of the examination is analyzing effective ways of assessing the output of system health prognostics because at present there are no unanimously accepted techniques. It is explained that the model outputs can be used to forecast effectively whether the system health prognostics was employed within a given implementation of the simulation model. ‘Fix it when it is broken’ is an old and most common maintenance and repair strategy. This approach is prevalent among firms, costing them potentially millions of dollar every year. According to this approach, no analysis or planning is required for the maintenance. The surprise failure of equipment at times that may be inconvenient is often the result of this approach. The troubles with this approach include preventing achievement of committed production schedules. The approach will be discussed in detail in the paper. Unscheduled downtime can have very negative results in applications such as aircraft engines. These kinds of problems highlight the importance of maintenance and repair. The situation give reasons why maintenance and repair should be performed before the surprise problems arise. According to this approach, maintenance and repair should be performed at pre-established intervals. This approach can provide comparatively high equipment reliability, but it tends to do so at too much cost. Another difficulty with time-based approaches is that breakdowns are assumed to take place at specific intervals. The finding of this paper will explain that the only way to reduce costs of both maintenance and repair and possibility of failure is to perform ongoing evaluation of machine health and ongoing forecast of future failures on the basis of current health and operating and maintenance record. System health prognostics have incentives for firms in the shape of reducing costs of repair and maintenance and associated operational disruptions. The system also helps in minimizing the risk of surprise failure of equipment. Repair and maintenance of a system usually starts with the objective to minimize the chances of unscheduled downtime. The system health prognostics help in minimizing the cases of catastrophic failures that paralyze the system. Although it is easy to implement time-based maintenance or breakdown maintenance, condition-based maintenance is increasingly popular among firms because of its practical approach. Condition-based maintenance is a comprehensive investigative process that requires in some cases detailed instrumentation and in most cases complex modelling techniques. Therefore, it is essential to perform a requirement analysis before implementing such an endeavour. System health prognostics — Project Report 1. Introduction The last 15 years saw the change in the manufacturing environment which moved from a mass production economy to a customer-driven economy. The factors behind the rapid change were international aggressive competition, high fluctuations in market demand and the fast arrival of new technologies. Koren and fellows (1999) say that for survival in the fast changing competitive market manufacturers should learn how to respond to new environment with rapid and cost effective measures. With the changing environment the focus of manufacturers has shifted from cost reduction and quality enhancement to cost-effective responsiveness while maintaining the maximum level of product quality. The tool for applying this new philosophy is system health prognostics. According to Saranga (2000), the relationship between strong maintenance management practices and major improvements in efficiency and profitability has been well documented. While the output of investment is highly dependent on the specific manufacturing unit and the equipment involved, Rao’s study says that an investment in monitoring of between $10,000 and $20,000 dollars results in savings of $500,000 a year (Rao 1996). Ben-Daya and fellows (2000) say that in the prevailing competitive marketplace, maintenance administration and machine health monitoring play an increasingly significant role in combating competition by reducing the surprise failure of equipment and associated costs and scheduling disruptions. Across a number manufacturing units, 15–40 per cent of production costs are usually attributable to maintenance activities. 2. Maintenance Management The advancement in system health prognostics has an interesting chronological outlook. Health prognostics and the connected functions has been the centre of attention of research and implementation for the past few years. Through all this time they have considerably evolved in terms of governing philosophy, execution, and enabling arrival of new technologies, modelling techniques, and emerging or redefined requirements. Maintenance theories can be generally classified as reactive and proactive. 2.1 Reactive maintenance For the reactive maintenance philosophy this term is very much true, ‘fix it when it is broken’. Reactive or unintended maintenance is old practice according to which maintenance should be performed only after the appearance of the fault, breakdown, or stoppage. In this practice it is not necessary to plan maintenance. According to Eisenmann (1997), this practice is suitable in facilities where the installed machinery has not a big size and the industry is not totally dependent on the reliability of any individual machine. It might also be suitable when the breakdown rate is negligible and failure does not result in severe cost setbacks or safety consequence. Corrective maintenance and emergency maintenance are parts of reactive maintenance which can be understand in the following two categories. i) According to Williams (1994) and Blanchard (1995), corrective maintenance is explained as the action carried out after a breakdown has occurred and is planned to fix a machine to a situation in which it can perform its necessary function. It is also important in this strategy that the other functions of the business are not affected by the failure of the machine. ii) Emergency maintenance is explained as the maintenance action that is essential to complete immediately to avoid severe consequences. 2.2 Proactive maintenance Proactive maintenance is an approach for ensuring the reliability of equipment. It is also called planned maintenance and can further be categorized as preventive and predictive maintenance. Under this strategy, it is wrong to wait for a breakdown. Under this strategy it is necessary to plan maintenance before it is required. Mobley (1990) is of the view that proactive approach does not wait for the equipment to fail before beginning of the maintenance activities. In a number of cases, better utilization of resources is seen compared to reactive approaches. Preventive or planned maintenance is the approach designed to perform maintenance at prearranged intervals to diminish the chance of failure or performance degradation. Proactive maintenance can further be classified into three following categories. i) Constant interval maintenance — this kind of maintenance is performed at fixed intervals. This is done even an additional maintenance that prompted by failure has already been performed. Timing is very important in this strategy and the maintenance should always be performed at its scheduled time. ii) Age-based maintenance — this strategy suggests protective maintenance at fixed intervals. Under this strategy maintenance is performed after a certain time period. iii) Imperfect maintenance activities can be performed to extend the lifetime of equipment. This strategy concerns about the current state of the system and plan future actions. 2.3 Predictive maintenance As the name suggests predictive maintenance involves forecasts about the health of equipment. The prearranged interval is estimated from the breakdown rate distribution that is designed from past data taken from the system or given by the supplier of individual sections in the system. Predictive and preventive maintenance are different in the arrangement of maintenance. In the predictive maintenance, the action is done on a fixed schedule while in the preventive maintenance the action is adaptively determined. Predictive maintenance has further classification. i) Condition-based maintenance — this is the strategy aimed at extending equipment life, improving productivity, and reducing daily operating costs. In this strategy the condition of system and its components become the reason of the maintenance. Decision-making is important in this strategy and after observing the condition of the system it is decided that the maintenance is become a necessity. The current state of a system is enumerated by parameters that are constantly monitored and are system or application specific. For example, in the case of rotary systems a vibration characteristic or index is a suitable option. In a manufacturing unit, the system is open to haphazard disturbances which cause divergences in the operational characteristics. Therefore, it is always a good strategy to monitor the condition of system and take right decision at right time by observing the condition of the system. The strategy of condition-based maintenance has a number of benefits for maintaining good production environment. One of the advantages of this strategy is obvious as the decision is taken on depictive and corroborative data that in fact reflects the condition of the system and its component. It is extremely presumptive to think that the condition of equipment would always follow the same operational curve, which is the fundamental assumption in preventive maintenance. Prior alarm for the approaching breakdown and increased accuracy in failure forecast are also the benefits of the strategy of condition-based maintenance. The strategy gives support in diagnostic procedures as it is comparatively easy to associate the breakdown to specific components through the monitored factors. The drawback of the condition-based maintenance strategy is the requirement to install and use monitoring the system and to make some level of modelling or decision-making plan. ii) Reliability-cantered maintenance — the reliability approach which is basically developed in the aircraft industry is to make the use of reliability estimates of the system to plan a cost-effective schedule for maintenance. For every safety-related application, cost-efficiency is balanced with safety and availability with the objective of reducing amount of costs and number of breakdown but eliminating the possibility of a downtime. Two important goals are included in reliability-cantered maintenance, one of which is to examine and classify failure modes based on the consequences of the failure on the system and the other is to evaluate the impact of maintenance schedules on reliability. The failure investigation begins with the recognition of all the failure modes and proceeds with classification of these failure modes on the basis of the effects of each failure. According to Rao (1996), the effects of failure are generally operational, environmental/safety, or economic. As soon as the consequences are recognized, the decision logic algorithms make priority of the consequences. These algorithms have a tendency to be industry specific as the restrictions and requirements of each industry differ considerably. Kumar (1990) and Sandtorv (1991) who presented excellent understanding for reliability strategy in their studies were of the view that although maintenance under the reliability-cantered maintenance intervals were decided in the same way to planned or scheduled maintenance, condition monitoring practices are increasingly being applied to decide the optimum interval. Therefore, though originally a preventive maintenance technique, reliability-cantered maintenance is turning into predictive maintenance (Moubray 1997; Jones 1995). 2.4 Cost Analysis While estimating the construction cost of mechanical-electrical systems, an excellent rule of thumb is to assume that 40% of the construction cost will be for these services. For example, if a firm was planning to build a 200,000-sq-ft facility and the construction cost was budgeted at $150/sq ft, the project cost would equate to roughly $30 million. The estimated trades cost would then equate to 40% of this, or $12 million. In addition, this facility will most likely have an annual cost of $5/sq ft -- $1 million -- to operate and maintain. Another excellent rule of thumb is to assume that most mechanical-electrical systems have an approximate service life of 20 years. By applying simple math and omitting inflation factors, operating cost over the next 20 years can estimated at approximately $20 million. If the facility owner is not cautious and does not implement an operation and maintenance action plan, then equipment and systems are going to fail earlier than anticipated. Consider equipment that prematurely fails after 12 to 15 years of service. It can be assumed the following premiums will be incurred: — Reduction in building occupant work performance due to poor or marginal system performance. Hvac equipment that does not operate efficiently can and will affect work-force performance. In turn, space comfort may also be compromised, along with indoor air quality (IAQ). Employee discomfort, lost time, excessive sick time, and reduction in productivity will occur. These hidden costs can be difficult to measure. Assume that a $0.50/sq-ft premium is incurred for reduction in work performance each year for 20 years (the anticipated hvac service life); this equates to $2 million. — Unscheduled breakdown of equipment due to an equipment failure. Unexpected shutdowns can have a direct impact on the cost of doing business. A stoppage in manufacturing, disruption in a research project, and/or sending employees home early due to equipment breakdown can be calculated into lost revenue. Estimate a $0.20/sq-ft premium for unscheduled breakdowns each year for 20 years; this equates to $800,000. — Wasted energy due to the system not being properly tuned up. Like an automobile, an hvac system not running smoothly will waste fuel energy. Estimate a $1/sq-ft premium energy waste each year for 20 years; this equates to $4 million. These cost premiums are only the tip of the iceberg for infrastructure that is not properly maintained and efficiently operated. The premature re-investment in equipment and/or systems to replace the infrastructure within this 200,000-sq-ft facility example can be budget-estimated at $10/sq ft, for a $2 million total. All of these added expenditures contribute to overruns in operating costs that result from poor facility management. 3. Literature Review Literature has planed important role in the advancement of maintenance strategies. According to Abdulnour (1995), the vital inspiration for improved maintenance management is its natural objective to increase machine availability. It has a direct impact on the performance of a manufacturing unit. Because of ever increasing customer demands and fast arrival of new technologies, management strategies such as just-in-time (JIT) and material resource planning (MRP) become necessary. JIT and MRP are among those activities which improve efficiency of firms by eradicating wasteful production activities. Surprise failures of equipment or frequent breakdowns create a major obstacle in implementing such techniques (Abdulnour 1995). These problems also result in high variance in production activities and therefore increasing the onus on the other organizational functions such as scheduling. Albino (1992) and Sun (1994) provide detailed studies about the effects of maintenance policies on manufacturing systems. Another convincing but less addressed reason of maintenance is safety and environmental preservation. Environmental preservation is one of the most important issues of today’s organizations. With the increase in severity of safety and environmental legislation, proactive maintenance assumes an essential role. Rao (1996) was of the view that accidents and operational hazards often become a big problem for organizations because these cases lead to huge legal expenses. Improvement in quality is increasingly considered as a motivation for improved maintenance management. However, there is no surprise in this fact that the connection between quality and maintenance has not received enough research attention because the relation between them is not immediately apparent. As improvement in quality is going to become practical by merging it with methods like process control and productivity improvement, the impact of equipment maintenance on quality is being exposed (Abdulnour 1995). The studies of Antero (1999) Makis (1998) provide similar analysis. Improvement in quality has its impact on all the functions of the business. Manufacturing and quality give motivating outlooks to maintenance in the form of objectives to give a boost to availability and reduce the chances of defective outputs. In some cases manufacturing and quality help in maximizing process capability. According to Nakajima (1988), theses processes are the conjoined goals of total productive maintenance the main aim of which is to increase equipment efficiency. Husband (1978) says terotechnology comes with the same goal but in a much broader sense with addition to the supplier and all the concerned engineering users. Because of these approaches, there has been a significant shift in perceptions governing maintenance practices in manufacturing industry. Behind the rapidly changes techniques in manufacturing industry is computer. Similarly new theoretical ideas have provided critical new maintenance management capabilities. All these methods are often interdisciplinary, beginning in quite disparate fields. 4. Standards in Maintenance The past 10 years saw remarkable academic and organizational contributions to the design and execution of maintenance strategies. A number of studies highlighted the importance of maintenance and organizations played important role in its implementation. The developments are boosted by diverse needs and therefore the maintenance strategies and algorithms have been refined to meet the challenges of the environment of today’s manufacturing units. Not only individual efforts, but a number of organizations came together to modernize the current and future developments in the maintenance. All these efforts resulted in creation of specific goals and ideas which introduced a lot of tools for maintenance. Open system alliance CBM, popularly known as OSA-CBM, is one of such development. It is being supported in an attempt to integrate various maintenance efforts and to widen the scope of maintenance by achieving a faultless integration with the different functional arms of a manufacturing unit. At the beginning, OSA-CBM was used for developing open and interchangeable system that could provide the ground for maintenance, particularly condition-based maintenance system. The purpose is to identify a standard on these numerous systems so that the future developments have a tendency toward multi-purpose, swappable components. OSA-CBM has made a seven-layered design that covers the characteristic stages in the development, deployment, and integration of maintenance strategies under the CBM techniques. There is another excellent system that enables enterprise asset optimization (EAO) coming from the productized mixing of building, plant and machine data into and with enterprise business information. It is called Machinery Information Management Open Systems Alliance and popularly known as MIMOSA. OSA-CBM and MIMOSA adjust their research and actions to put together individual goals. The main goal of MIMOSA is to project maintenance management as a business activity that operates on business objects with well described properties, techniques and information interfaces. 5. Health Prognostics — Methods and Techniques Understanding methods and techniques is very important in system health prognosis. Maintaining the health of a system is a difficult work and it involves in-depth examination of the target system, principles concerned, and their applicability and implementation procedures. There are a number of processes, investigations and modelling tools, and methods to provide data for modelling and examination. Each application in the maintenance strategy has a lot of tools and techniques. There is also a close connection between reliability-based maintenance and statistical maintenance methods. Traditions play important role in selecting maintenance strategies in organizations and a well-known approach among organizations to the maintenance of complex systems is through estimating its reliability. Conventionally, organization estimate reliability from the time-to-failure distributions of equipment and the biggest disadvantage of such a strategy is that multiple failure mechanisms most of the time interact with each other in perhaps mysterious ways. The actions have an effect on the degradation rate of the system, causing it to deviate significantly from the calculated failure distribution. The study of Knezevic (1987) suggested a different approach very similar to condition-based maintenance which is known as the relevant-condition-parameter-based (RCPbased) approach. Knezevic’s suggested approach is on the basis of the identification of RCPs that calculate or reflect a particular breakdown mechanism. Another important maintenance strategy is model-based approach to failure detection, isolation and identification (FDI). This approach is applied on the basis of investigative or functional redundancy which means that different signals are compared and assessed to find out the existing faults in equipment. The assessment under the model-based approach is between the measured signal and the estimated values developed by the mathematical model of the equipment. Model-based approach heavily depends on residual general, but methods involved in model-based diagnosis are different in the generation and definition of a residual. There are a number of techniques used for residual generation, including observer-based approaches, parameter estimation technique and party space approach. According to Simani and fellows (2003), observer-based strategies depend on estimating the outputs from either Luenberger observers or Kalman filters. Observer-based strategies are cantered on the philosophy that the state estimation error is zero in a setup where there is no doubt about fault and it is not so otherwise. Observer-based approaches can further be categorized into dedicated observer, fault detection filters, and output observers. The philosophy of the parameter estimation methods is that faults have effects on the production through the system parameters. Therefore, the idea is cantered on developing online estimates of the parameters and examining the changes in the estimates. The equation error techniques examined the parameters directly. In this system least square estimation most of time is used. The standard of parity space connections is to test for parity of the measurements from the procedure, developing a residual by comparing the model and the performance of procedure. Patton and Chen (1994) say that this idea has close association with the observer-based techniques. A perfect model is never obtained due to nuances of the pragmatic world such as noise, etc., and to be effectual the model-based FDI should learn to distinguish between these reservations and the changes due to breakdown. According to the studies of Chen & Chen (1999) and Struss & Dressler (1989), one more problem is to recognize not just the existing faults but the developing faults which at present do not have considerable effect on equipment. The centre of attention in the signal-based FDI methods is on identifying the changes or variations in a signal and then analyzing the change. A number of studies have very comprehensively investigated the change detention in a system. There are several strong techniques that have integrated different ideas from parametric modelling principles with signal-based principles such as spectral examination. A number of methods are created around model-based techniques, for example, generation of residuals and diagnosis of the residuals. According to the study of Staszewski and Tomlinson (1994), data compression and feature extraction are used as two vital features of wavelets to FDI. Data compression is related to encoding the data in a compressed form. Feature extraction categorizes features inside these encoded signals that would help find out the faults in the monitored systems. As soon as wavelet transform is applied to the signal output, the coefficients are examined for any variation from the normal signal. The detection of coefficients that would confirm a failure is a thorough process. These wavelets are mainly used for FDI in gears, as vibration examination is quite effectual for these domains (McFadden 1994; Paya 1997). Another strong tool for vibration examination is time-frequency analysis by the use of Wigner-Ville distribution (WVD). Time-frequency analysis is considered quite successful in conditions where neither the time domain nor frequency domains can generate vital patterns (Staszewski 1997). The contour plots developed by WVD are visually checked for the failure features that point out its progression and existence. A number of occasions these plots are examined with the support of classification algorithms ranging from parametric to soft computing. Bearing condition diagnosis is one more largely accepted option for signal-based FDI. In this method signals vary enormously because of variable loads. Shiroshi and fellows (1997) say that these applications characteristically necessitate signal enhancement via filtering followed by a condition parameter examination. For FDI, another method is detection signal techniques. In this process a detection signal works as an input to the system for a precise period of time and the diagnosis is performed on the basis of the way of working of the system during this period. Two more very popular methods for FDI are Bayesian statistics and Bayesian parameter estimation. These methods are used a number of applications. Another significant aspect is to recognize the detection intervals, optimization of cost and replacement decision-making. According to the studies of Wang & Sheung (2003) and Al-Hassan & fellows (2000), the use of Markov chains are going to be increased for optimizing maintenance processes. The proportional hazards modelling (PHM) is well known for reliability estimation and estimation of effects on failure rate. 6. System Monitoring in Industrial Applications Production operators play an important role in detecting failure of equipment. Production operators constantly check operation of their system and examine sensor date developed by the system. They examine important trends in the operational states of the system and operate the system by connecting trend patterns with failure modes. The problem in this approach comes when the system have new operators in the case of retirement of old guards and change of employment. On many occasions, the gap between levels of expertise encourages substantial operational variability by individual operators which is harmful to the system’s permanence. Another problem can be a large volume of sensor and operational data which creates a cognitive overload for the operator, increases the possibility of the operator reaching a wrong decision. Another problem is that direct examination of operations and accessible sensor data may not be enough to detect incipient failure. Advanced modelling techniques are mainly helpful in these cases, making possible detection and forecasting of breakdowns by finding ways. Different applications have different approaches. They are different in knowledge acquisition, but common in gathering of rules from domain experts and discovery of relationships in data. The study of Villanueva and Lamba (1997) provides excellent understanding for guidance system for operators in an industrial plant. They used the knowledge-based system (KBS) method to breakdown diagnosis for coal processing machine. They applied in their study the standards advocated by Hickman and fellows (1989) in KADS methodology to implement their KBS. Goal tree – success tree (GTST) and Fault-cause network are the two parts of Villanueva and Lamba’s knowledge component of control model (AshMod). The GTST knowledge is very helpful in understanding calculated aspects of a system. It summarizes the plant’s planned aspects through a problem reduction technique and is modelled by a tree structure of objectives that must be fulfilled for the correct operation of the system. Trend examination on the GTST tree recognizes an unusual trend and then AshMod applies a mixture of backward and forward reasoning to examine the causes from the fault cause network. Alarms are important in detecting that when the system needs maintenance. Missed and false alarms are main troubles in analytical systems, because they weaken the reliability of the monitoring system and reduce the motivation to continue using the monitoring system. It is true that KBS is an effective method, but it would be wrong to say that it is free of all weaknesses. 7. Conclusion Organizations face a number of challenges in maintaining the system. It is true that there are numerous problems in process control and fault analysis, but artificial intelligence techniques have been shown to be successful for both the time-based maintenance and the breakdown maintenance. Data collection is very important for examining the health of system and the modelling of the breakdown mechanism or process control begins with this collection. Similarly, data cleansing is enormously significant, especially in the case of adaptive control. This is general practice that machine models are defined while keeping in mind these two applications, and are validated and improved to preserve accuracy and lessen the chances of false alarms and missed hits. This paper has discussed almost all the issues organizations face in the development and maintenance of prognostic systems. The paper has also discussed how prognostic systems help play their role in smooth functioning in a manufacturing unit. Development and maintenance of prognostic systems comprise the selection of knowledge acquirement and modelling technologies, with deliberations including available types of knowledge and techniques to attain and maintain accuracy of the models and knowledge bases. The paper has discussed advancement in the approaches of maintenance. References Abdulnour G, Dudek RA, Smith ML (1995) Effect of maintenance policies on the just-in-time production system. IJPR 33:565–583 Albino V, Carella G, Okogbaa OG (1992) Maintenance policies in just-in-time manufacturing lines. IJPR 30:369–382 Al-Hassan K, Swailes DC, Chan JFL, Metcalfe AV (2000) Markov models for promoting total productive maintenance. Proc Ind Sta Action 1:1–12 Antero O, Markku M (1999) Maintenance has a role in quality. TQM Mag 11(1):17–21 Ben-Daya M, Duffua SO, Abdul R (2000) Overview of maintenance modeling areas. Maintenance, modeling and optimization. Kluwer, Dordrecht, MA, pp 3–35 Blanchard B, Verma D, Peterson EL (1995) Maintanability: a key to effective serviceability and maintenance management. Wiley, New York Chen J, Patton RJ (1999) Robust model-based fault diagnosis for dynamic systems. Kluwer, New York Eisenmann R Sr, Eisenmann R Jr (1997) Applied condition monitoring. Machinery malfunction diagnosis and correction 13:703–741 Hickman et al. (1989) Analysis for knowledge based systems: a practical guide to the KADS methodology. Prentice-Hall, Chinchester, UK Husband TM (1978) Maintenance management and terotechnology. Gower, Aldershot, UK Jones RB (1995) Risk based management: a reliability centered approach. Gulf Publishers, Houston Knezevic J (1987) Condition parameter based approach to calculation of reliability characteristics. Reliab Eng, 19(1):29–39 Koren, Y., Heisel, U., Jovane, F., Moriwaki, T., Pritschow, G., Ulsoy, G., VanBrussel, H. (1999), Reconfigurable manufacturing systems., Annals of the CIRP, Vol. Kumar U, Granholm S (1990) Reliability centered maintenance – a tool for higher profitability. Maintenance 5(3):23–26 Makis V (1998) Optimal lot sizing and inspection policy for an EMQ model with imperfect inspections. Naval Res Logist 45:165–186 McFadden PD (1994) Application of the wavelet transform to early detection of gear failure by vibration analysis. Proc International Conference of Condition Monitoring 1:172–183 Mobley RK (1990) An introduction to preventive maintenance: plant engineering series. Van Nostrand, New York Moubray J (1997) Reliability centered maintenance. Industrial, New York Nakajima S (1988) Total productive maintenance. Productivity, Cambridge, MA Paya BA, Esat II, Badi MNM (1997) Artificial neural network based fault diagnostics of rotating machinery using wavelet transforms as a preprocessor. Mech Syst Signal Process 11(5):751–765 Rao BKN (1996) The need for condition monitoring and maintenance management in industries. Handbook of condition monitoring. Elsevier, Amsterdam, pp 1–36 Sandtorv H (1991) RCM – closing the loop between design, reliability and operational reliability. Maintenance 6(1):13–21 Saranga H, Knezevic J (2000) Reliability prediction for condition based maintained systems. Reliab Eng Syst Saf 71:219–224 Shiroshi J, Li Y, Liang S, Kurfess T, Danyluk S (1997) Bearing condition diagnostics via vibration and acoustic emission measurements. Mech Syst Signal Process 11(5):693–705 Simani S, Fantuzii C, Patoon R (2003) Model-based fault diagnosis in dynamic system using identification techniques. Springer, Berlin Heidelberg New York Staszewski WJ, Tomlinson GR (1994) Application of wavelet transform to fault detection in a spur gear. Mech Syst Signal Process 8:319–356 Staszewski WJ, Worden K, Tomlinson GR (1997) Time-frequency analysis in gearbox fault detection using the wigner-ville distribution and pattern recognition. Mech Syst Signal Process 11(5):673–692 Struss P (1989) Diagnosis as a process. In: Hamscher et al (eds) Readings in model based diagnosis. Morgan Kauffman, San Mateo, pp 408–418 Sun Y (1994) Simulation for maintenance of an FMS: an integrated system of maintenance and decision making. Int J Adv Manuf Technol 9:35–39 Villanueva, H, Lamba H (1997) Operator guidance system for industrial plant supervision. Expert Syst Appl 12(4):441–454 Wang CH, Sheung SH (2003) Determining the optimal productioninspection intervals with inspection errors: using a Markov chain. Comput Oper Res 30:1–17 Williams JH, Davies A, Drake PR (1994) Condition based maintenance and machine diagnostics. Chapman & Hall, London, pp 1–18 Read More