CHAPTER 2 PREVENTIVE MAINTENANCE STRATEGIES

Transcription

1 CHAPTER 2 PREVENTIVE MAINTENANCE STRATEGIES 2.1 Intrductin Mst f the industries are having systems fr prductin f gds and having systems fr delivery f services. These systems are subjected t deteriratin with usage and age. System deteriratin is ften reflected in higher prductin csts and lwer prduct quality. Preventive maintenance (PM) ften perfrmed a majr rle fr up keep f prductin csts dwn while maintaining gd quality [8]. In 1951 PM was intrduced in Japan. Hwever the cncept f PM was taken frm USA [2]. Nippndens f the Tyta grup was the first cmpany wh intrduced plant-wide PM in At that perid peratrs prduced gds by using machinery and maintenance aspect was handled by anther team f wrkers wh were specialised in that field. In the 1970s maintenance became a prblem due t requirement f many maintenance persnnel fr maintenance [8]. As a result new cncepts were develped t maximise plant and equipment effectiveness t achieve ptimum life cycle cst f prductin equipment, such as crrective maintenance, autnmus maintenance, predictive maintenance, TPM etc. When the financial situatin is cnsidered, maintenance plays a key rle in mst industries. Ppular examples are pinted ut as fllws. [15] Cited that the amunt f mney spent cmpany-wide maintenance by du Pnt in 1991 was rughly equal t its net incme. In many cases it is difficult t estimate the ttal indirect maintenance csts, such as lss f incme due t breakdwns, pr quality, lss f custmers and market shares. Fr Swedish industry in 1998 [16] estimated average direct and indirect csts f annual maintenance arund SEK billin. Accrding t that study the direct maintenance cst cnsisted f apprximately half frm the ttal maintenance csts. Accrding t a study [17], 15% t 40% (average 28%) f ttal prductin cst is attributed t maintenance activities in the factry. Anther study [18] reprted that the estimated cst f maintenance fr a selected grup f cmpanies increased frm $200 billin in 1979 t $600 billin in By cnsidering the financial impact peple realised that the maintenance is playing a key rle specially in manufacturing industries. As a result f this mre research were 14

2 dne n several studies related t PM because it is helpful in increasing perfrmance and imprving the availability, prduct quality, safety requirements, and plant cst effectiveness levels. When intrducing a PM prgram fr a plant and machinery it shuld cntain a suitable PM strategy and plant maintenance schedule in rder t achieve the equipment reliability, failure avidance, defect eliminatin, least perating csts, risk reductin and maximum prductin. If selected strategy fails t achieve ne f the abve, the effect and the utcme f the maintenance will suffer. Mdern PM fcuses n maximising perating prfit by prviding reliability grwth thrughut the equipment lifetime. Research surveys n different aspects f PM have been carried ut [19]. As a result the researchers have been able t identify different facts f PM prgrams: Time based- eg: tw shts f lubricant per week Cnditin based- eg: if belt is squeaking, then tight it Failure finding- eg: verify functin f a high level prbe OEM directed- eg: recmmended in the equipment manual Hand me dwn experience- eg: we have always dne it that way Regulatry- eg: required fr cmpliance with EPA, OSHA, etc. Risk- eg: this is really expensive equipment, s let s d this take mre ften Based n abve mentin principles, strategies have been develped by different researchers and rest f the space in this chapter is devted t their discussin. 2.2 Preventive Maintenance Methdlgies Optimal maintenance mdels fr systems subject t failure During the 1965 and 1967 perid a great deal f research was dne by McCall and Barzilvich [20] in the field f ptimal maintenance mdelling, invlving the aspects f ptimal preventive and preparedness maintenance plicies. Accrding t the research they categrised PM as Deterministic Mdel, Stchastic Mdel under Risk and Stchastic Mdels under Uncertainty. 15

3 Stchastic mdels under risk Simple system preventive maintenance mdel (peridic, sequential) Simple system preparedness maintenance mdel (peridic, sequential) Cmplex system preventive maintenance mdel (peridic, sequential,..pprtunistic) Cmplex system preparedness maintenance mdel (peridic, sequential,..pprtunistic) Stchastic mdels under uncertainty Preventive maintenance mdel fr simple and cmplex systems Simple (cmplex) system preparedness maintenance mdel T develp thse mdels they have used several ptimizatin techniques such as linear prgramming, nnlinear prgramming, dynamic prgramming, pntryagin maximum principle, mixed-integer prgramming, decisin thery, search techniques, heuristic appraches, etc. [20]. Beside that fr each and every mdel they have used several assumptins. The main advantage f this strategy is the readiness in develping a mdel accurately if machine cnditin is prperly identified. T meet slight variatins in machine cnditins different mdels have t be develped accrding t the situatin and it will becme a serius issue during the implementatin and handling stages Optimum preventive maintenance plicies Richard Barlw and Larry Hunter develped different maintenance strategies t suit the requirements f simple equipment and anther in respect f large and mre cmplex systems in late 1950 [21]. Fr less cmplex equipment, repair at the time f failure (r replacement) may actually crrespnd t general verhaul. PM is scheduled fr mre cmplex systems when taken int cnsideratin the number f peratinal hurs and the general maintenance invlved, is knwn as repair after the failure. PM strategy fr simple systems is defined as fllws (Plicy I): Perfrm PM after t 0 hurs f cntinuing peratin withut failure. We will allw the pssibility that t 0 be infinite, in which case n PM is scheduled. If the system fails befre t 0 hurs have elapsed, perfrm maintenance at the time f failure. PM is then rescheduled. Fr this 16

4 plicy, we assume that the system is as gd as new after any type f maintenance (r replacement) is perfrmed [21]. PM strategy fr cmplex systems is defined as fllws (Plicy II): - Perfrm PM n the system after it has been perating a ttal f t* hurs regardless f the number f intervening failures. Again we allw the pssibility that t* be infinite. We assume that after each failure nly minimal repair is made. The additinal assumptin is made that the system failure rate is nt disturbed after perfrming minimal repair. Fr instance, after replacing a single tube in a televisin set, the set as a whle will be abut as prne t failure after the replacement as befre the tube failure. This will be due t the aging f the ther cmpnents. Again, the assumptin is made that the system is as gd as new after PM is perfrmed. If the system is actually replaced, this assumptin is clearly n restrictin [21]. Accrding t the abve definitins mathematical mdels have been develped fr the simple as well as cmplex systems. Using this mathematical mdel they have develped a simple plicy cmparisn graph which is shwn in Figure 2.1. Figure 2.1 Plicy cmparisn chart T m = expected time t perfrm minimal repair T e = expected time t perfrm emergency maintenance T s = expected time t perfrm scheduled maintenance 17

5 Accrding t the graph n Figure 2.1 plicy cmparisn chart; if T e = 1 hur, then plicy I can be selected, If T m > 19 minutes r therwise plicy II can be selected. Hence if T m = 15 minutes plicy II shuld be used. This strategy plays a vital rle t decide the mst ecnmical type f maintenance. But the ptimal t and t*, fr plicies I and II respectively, depends nly n the ratis T e /T s and T m /T s respectively [21]. The peple wh are in pssessin f have specialised knwledge n statistical functins are required fr evaluate this strategy. Therefre in the implementatin stages f this strategy it is likely t encunter serius issues. Hwever cmparisn chart is prviding sme easy guidelines t categrise and help t select suitable PM strategy fr simple as well as cmplex equipment. Therefre in develping a PM strategy fr aluminium extrusin industry it is useful t absrb imprtant cncepts frm this and they will be helpful in srt ut practical prblems arising in the implementatin stages Preventive maintenance prgram at Malaysia palm il mill Research was dne at palm il mill in Malaysia [22] t evaluate the mathematical analysis methds t measure the machine perfrmance and evaluate their level f availability by using five year f data. The mathematical analyses were dne based n the fllwing indicatrs. [2.1] [2.2] [2.3] Ttal reakdwn time hurs um er f failures [2.4] 18

6 When taken in t cnsideratin the abve indicatrs it is very much easy t priritise critical equipment. S PM can be fcused t critical equipment and it will help t detect deviatins in its early stage t avid further ecnmic lsses. By applying these mathematical indicatrs directly t critical equipment they were able t maximize the plant availability and reduce the maintenance cst drastically. Therefre abve mentined mathematical indicatrs are very vital t aluminium extrusin industry because it invlves critical equipment. As such the evaluatin f the impact f the mathematical indicatrs further facilitates t develp an accurate maintenance plan fr the right equipment Vibratin based maintenance Vibratin based maintenance (VBM) [23] is mre widespread analysis methd fr PM specially when high dwn time cst incidents are invlved. Nt nly this strategy, research dne by Cllactt [24], Al-Najjar [25] and DeBttn et al. [26] highlights the tremendus pssibilities available fr VBM, ne f which is the ability t receive indicatins f changes f the cnditin f a machine at an early stage. This study based n Life Cycle Cst (LCC), will identify the cst factrs by describing their behaviur during equipment life. LCC is the ttal cst f wnership f an item, taking int accunt with all the csts f acquisitin, persnnel training, peratin, maintenance, mdificatin and dispsal [4]. In rder t place maintenance in a prper fcus amng its plant activities this strategy attempts t intrduce a new perspective. This strategy deals with an integrated apprach fr the jint ptimizatin f prductininventry cntrl and PM plicy fr the manufacturing cell.as the first step f this strategy cst factrs are identified t determine relevant technical and ecnmic input data. The next step is t intrduce these input data in the accuntancy system by making use f suitable frmulas which is used t assess the fllwing mdel s utput, as shwn in Figure 2.2 [23]. 19

7 Figure 2.2 Maintenance cst, savings and prfits mdel LCC factrs can be used fr mnitring parameters t prvide the required infrmatin fr decisin-making, t ensure the cst-effective actins and enhance cntinuus imprvement. And als t detect deviatins at early stages, thereby avid further ecnmic lsses, applicatins f relevant perfrmance measurement is helpful. Finally a recessin f investment fr maintenance shall be bn rather than reducing its budget, because investment f maintenance will return nine times f the invested capital ver the depreciatin perid. In additin t the benefits mentin abve, this mdel als prvides infrmatin based n invested capital as t where, why and hw much f it shuld be used. Nt nly that this mdel assess the effectiveness f the cst. Further this mdel utilizes the perfrmance measurement analysis n maintenance. This mdel facilitates t identify the prblem areas and recgnizes the investment pprtunities. Therefre cntrl f better data cverage is required t maintain quality and it is pssible in saving direct maintenance csts and further ensuring prfits in maintenance. This gist f this strategy discusses several factrs beynd the traditinal PM and it is very much useful t incrprate these facts fr develping PM strategy fr aluminium extrusin industry. 20

8 2.2.5 Prprtinal hazards mdel fr preventive maintenance T develp practical mdel t suit the PM has been described as the mst difficult part in the field f maintenance. The traditinal methd fr mdelling the PM schedule is t give pririty t d the maintenance based n time interval (L) spent rather than the urgency f the real requirement [28]. The develped mathematical mdel based n the abve descriptin is as fllws. [27] [2.5] [2.6] C(L) - lng run expected cst per unit time A(L) - lng run expected availability L - time interval C c - average cst f crrective peratin C pm - PM interventin t pm - expected dwn time fr PM t c - expected dwn time fr crrective peratin N(L) - expected number f failures in L CO - crrective peratin Statinary renewal cnditin is assumed in abve mdels. S ageing factr f the equipment r pssible imprvement due t design changes is nt cnsidered. In rder t avid deviatin frm actual requirement Ascher and Kbbacy [29] have prpsed the preventative actin by making use f a Nn-Hmgeneus Pissn Prcess (NHPP) t mdel PM situatins with increasing Rate f Occurrence f Failures (ROCOF). Ascher, Kbbacy and Percy [30] identified that several factrs affected the system perfrmance in additin t the PM schedule, t develped analytical apprach t the PM mdel. Identified factrs are included in the nature f the system itself, the envirnmental stresses seen by the system, the detailed list f tasks perfrmed during PM, the training, cmpetence and mtivatin f the persnnel perfrming PM and management attitude. 21

9 After dne several studies Kbbacy and his team define C(L) and A(L) as fllws [27]. [2.7] [2.8] i - cst f ne PM added t the cst f the number f failures, k - simulatin runs In the mdel presented in this study, researchers have cnsidered nly the next PM interval. During this interval the hazard functin f CO is updated after each failure. This leads t change in the intensity f failure during the PM interval, unlike the statinary prcess which has cnstant intensity. This apprach can be applied t repairable systems and that is the majr advantage f this mdel and that des nt require any restrictive assumptins regarding the quality f crrective wrk r planned maintenance, e.g. renewal. Accrding t the gathered data by a case study in a lcal aluminium extrusin plant cnstant frequent failures are nt reprted frequently. Therefre PM intervals are t be adjusted depending n the failure rates. S the cncept f this strategy is very much applicable fr develping the PM strategy t suit the needs f aluminium extrusin industry Wrld class maintenance management Terry Wireman has develped a lengthy PM strategy which cntains 25 steps and the business reasns fr the implementatin as given as fllws [18]. Pr Return On Investment (ROI) fr the ttal plant r facility valuatin Pr thrughput fr the design f the plant Inability t meet prductin demand High cst f ccupancy fr a facility Excessive dwntime Prductin inefficiencies T eliminate abve mismatches Terry Wireman has develped the strategy which is shwn as in the fllwing flw diagrams [18]. 22

10 Develp PM prgram Imprve PM prgram Develp MRO prcesses Imprve MRO prcesses Starting imprvement Des a PM prgram exist? Is the PM prgram effective Des MRO prcess exist? Are the MRO prcess effective? Imprve wrk management prcess Develp wrk management prcess Cntinued Is the wrk management prcess effective? Des a wrk management prcess exit? Figure 2.3 Terry Wireman Maintenance strategy series part 1 23

11 Develp Planning & Scheduling Practices Imprve Planning & Scheduling Select & Implement a CMMS/EAM System Imprve CMMS/EAM Utilizatin Frm part 1 Is Planning & Scheduling Utilized? Is Planning & Scheduling Effective? Is the CMMS/EAM System Utilized? Is the CMMS/EAM System Effective? Imprve Maintenance Training Prgram Develp Maintenance Skill Training Prgram Cntinued Are the Maintenance Training Prgram Effective? Des Maintenance Skill Training Prgram Exit? Figure 2.4 Terry Wireman Maintenance strategy series part 2 24

12 Re-evaluate Basic Maintenance Tasks Re-evaluate Operatr Perfrmed Maintenance Tasks Develp Predictive Maintenance Tasks Imprve Utilizatin f Predictive Tls N N N N Frm part 2 Are Operatrs invlved Maintenance Activities? Are peratr Perfrmed Maintenance Tasks Effective? Are Predictive Techniques Utilized? Are PDM tasks Effective? Imprve Utilizatin f Relia ility Technique Develp Relia ility Techniques N N Cntinued Are Relia ility Techniques Effective? Are Relia ility Techniques Being Utilized? Figure 2.5 Terry Wireman Maintenance strategy series part 3 25

13 Develp TPM/OEE Methdlgy Imprve OEE Utilizatin Develp Ttal Cst Measurements Imprve Ttal Cst Measurements Tls & Techniques Frm part 3 Are TPM/OEE Methdlgy eing Utilized? Is OEE Being Effectively Utilized? Des Ttal Cst Management Utilized? Is Ttal Cst Management Utilized? Imprve Effectiveness f CI Tls & Techniques fr Develp CI Tls & Techniques fr Maintenance/Asset Management Strive fr Cntinuus Imprvement in all Aspects f Maintenance/Asset Management Are CI Tls & Techniques Effective? Are CI techniques Utilized fr Mntnce/Asset Magt? Figure 2.6 Terry Wireman Maintenance strategy series part 4 26

14 T develp cntinual imprving strategy t aluminium extrusin industry is pssible by using abve strategies. The main advantages f this PM system are fcusing each and every pint in maintenance, easy t evaluate, mre infrmative, etc. It is very much difficult t implement abve type f strategy in practically because f its length. Attentin cannt be fcused n each and every device separately which is cnsidered as the main disadvantage f the system. S it is useful t develp a practical scenari fr aluminium industry by cmpacting this system f several steps in t ne Opprtunistic plicy fr ptimal preventive maintenance Prductin lss is ften very large when unexpected shutdwn ccurs n cntinuus perating units. Implementatin f suitable maintenance plicy implies the ecnmic prfitability which culd increase the availability and reduce the perating csts. The cmmn planning appraches used fr multi-cmpnent manufacturing systems include the grup/blck replacement mdels and the pprtunistic maintenance mdels. In the blck maintenance plicy, an entire grup f units r cmpnents are replaced at peridic intervals. The time perid is decided based n time/ cst r bth [32]. Peridical maintenance is cnsidered as a cstly ptin because maintenance activities are carried ut even thugh cmpnents are in gd cnditin. S several techniques and appraches are there dealing with multicmpnent systems. Based n that research fr slutin t cntinuus perating units, an algrithm was develped which is illustrated in Figure 2.7 [31]. 27

15 Figure 2.7 Flwchart f the algrithm The prpsed algrithm is a numerical prcedure, where the life cycles are simulated and the ptimal slutin is numerically searched. The mst imprtant facts related in this study are, Effectiveness f the pprtunistic plicy in cst saving fr multi-cmpnent systems and the capacity f the Mnte Carl simulatins in slving these cmplex prblems, Evident gruping cnfiguratin, which is smetimes adpted by maintenance managers, is nt necessary cst-effective; therefre, an ptimizatin prcedure must be cnsidered. 28

16 2.3 Summary PM is playing a key rle in almst every maintenance strategies. As a result mst f the research develped maintenance strategies are based n the PM. Several types f PM strategies are discussed in this chapter. Main fcus f the chapter is gathering methdlgies t develp PM strategy t aluminium extrusin industry. Strategy ne is categrising machinery based n deterministic mdel and stchastic mdel t be in accrdance with PM. The accuracy f the strategy depends n the strength f the histrical data. Secnd strategy is based n the mathematical mdel. But t find an actual type f maintenance fr the machinery, cmparisn chart is very much imprtant. Third strategy is mainly fcused n find ut the perfrmance level f maintenance. It is imprtant fr priritise the equipment cnsidering number f lss hurs. This methd is mre apprpriate t develp PM strategy where critical equipment is largely engaged. Frth strategy is cnsidered fr vibratin based maintenance. Vibratin is an early stage indicatr in mst f the machinery befre the breakdwn. Mst f the machinery is run under heavy stress lads therefre this methd is very much imprtant fr aluminium extrusin industry. The methdlgy t find ut frequency f the PM schedule is the fifth strategy. Time interval f the PM prgram is changing accrding t the failure rate based n mathematical mdel. Sixth strategy is a lengthy ne, which absrbs each and every aspects f cnsideratin fr the PM. The gd qualities f this system when cncentrated culd be useful t develp a suitable PM strategy fr aluminium extrusin industry. PM strategy fr cntinuus peratin machinery is described as a last strategy thrugh this chapter. It is very much imprtant t intrduce this cncept t aluminium extrusin industry because mst f the machinery runs cntinuusly in high temperature cnditins. Strategies discussed n PM systems in this chapter is further elabrated n chapter 3, based n the develpments tk place in PM in respect f aluminium extrusin industry. 29