Reliability Analysis of a Diesel Engine Driven Electric Power Unit E.C. NASIOULAS 1, G.J. TSEKOURAS 1, F.D. KANELLOS 2

Transcription

1 Reliability Aalysis of a Diesel Egie Drive Electric Power Uit E.C. NASIOULAS, G.J. TSEKOURAS, F.D. KANELLOS 2 Departmet of Electrical & Computer 2 Productio Egieerig & Maagemet Sciece, Helleic Naval Academy Departmet, Techical Uiversity of Crete Terma Hatzikyriakou, 8539, Piraeus Techical Uiversity Campus, Chaia, 7300 GREECE asioulas37@hotmail.com, tsekouras_george_j@yahoo.gr, kaellos@mail.tua.gr Abstract: - Thermal-electric power uits costitute oe of the mai parts of a power system affectig decisively its reliability. Usually the reliability of a thermal power uit is quatified by the Forced Outage Rate (F.O.R.) idicator, which expresses the uexpected ad uforesee shutdow rate of the thermal uits. It ca be calculated from data records of faults ad their duratio durig uit s lifetime. However, the thermal electric power uit is a complex system with several compoets. I case of diesel egie drive electric power uit the respective subsystems are the diesel egie, the electric geerator, the frequecy regulator, ad circuits of compressed air, oil, fuel, water coolig, air coolig etc. The aim of this paper is to carry out a reliability study for idividual sub-systems through fixed expected failure rate of each sub-system s compoet, so as the most frequetly malfuctioig parts of the system be idetified ad proactive measures be take. The disadvatage of this solutio is that the operatio of each subsystem ad its compoets should be tagible by the thermal uit s operator completely. Key-Words: - Reliability, thermal electric power uit Itroductio I specialized applicatios of small autoomous power systems the reliability of the power uits is a crucial factor affectig eve the survivability of their users i.e. helicopter s power system. The uavailability of a power uit is quatified by the Forced Outage Rate (FOR), which is the umber of hours the uit is o forced outage over the total umber of hours i a year. The use of differet power uits affects the uavailability rate of the power system which is expressed by two idicators: () the expected loss of load power (LOLP), which is the equivalet time per year for which the load demad is ot covered by the geeratio power system completely, (2) the expected loss of load eergy (LOLE), which is the total active eergy ot supplied i a year year. They are practical measures of a black-out probability. Ideally, these idicators should be equal to zero; practically they should be as small as possible. The respective idicators are calculated by classical reliability assessmet methods [-2]. The kowledge of the FOR of each power uit is satisfactory for the desig ad the operatio of the power system. But the thermal electric power uit is a complex system of several sub-systems ad compoets. The operator of the power uit eeds more iformatio about the reliability of each sub-system ad its compoets, so as the most frequetly malfuctioig parts of the system are idetified. Afterwards, techoecoomical study should be carried out comparig the possible measures, which ca be take, i.e. replacemet of a compoet, use of parallel elemet i case of a sesitive compoet, etc. The respective bottom-up reliability aalysis is usually avoided, as the theoretical aalysis is combied with the real model of the power uit ad the ecessary data cocetratio is a time-cosumig process. I this paper the reliability study of some subsystems of a thermal uit is carried out, so as the ecessary process ad the respective difficulties be examied. This study is a ecessary step of the bottom-up reliability aalysis of ay thermal power uit. I our case a diesel egie drive electric power geerator has bee chose, as it is usually used as mai power uit i autoomous small power systems ad as emergecy uit i hospitals, etc. I sectio 2 the ecessary reliability defiitios are itroduced. I sectio 3 the formatio ad the subsystems of a typical diesel thermal electric power uit are preseted ad the respective theoretical reliability aalysis is carried out. I sectio 4 the practical applicatio of the methodology is preseted with the respective sesitivity aalysis, while i sectio 5 the major coclusios are discussed. 2 Short Itroductio to Reliability Aalysis System s reliability is the probability of ormal system operatio durig its lifetime operatio uder proper quality ad operatio coditios. It is ISBN:

2 coected with the repairability, which is the probability of a system with failure to come back to ormal operatio coditios i a predefied time iterval. The major defiitios related with reliability are listed ext. The expected failure rate λ is the ratio of the mea value of the observed K / to the respective operatio time T / durig the observatio time period: / K % λ =,, / () T h year 000h The lifetime variatio of the expected failure rate of a compoet is preseted i Fig. ad it cosists of three separate parts: () iitial, (2) casual ad (3) decay, which occur durig the respective life time periods of iitial operatio, mai life ad decay. It is obvious from Fig. that the expected failure rate is practically costat durig the mai operatio life. Failure rate λ Iitial operatio life Iitial Mai operatio life Total Casual Decay operatio life Decay Life time Fig.. Failure rate λ with respect to life time. The reliability fuctio R(t) of a elemet is the probability of the elemet ot failig for the time period [0 t]. Applyig Bayes theorem, R(t) is obtaied by: t R() t = exp λ () t dt 0 (2) I case of costat failure rate λ the reliability fuctio R(t) follows the expoetial (Poisso) probability distributio: R() t = exp( λ t) (3) The mea time to failure, m (MTTF), is the mea value of the time itervals betwee sequetial durig the observatio time period: / T m = / [ h K ] (4) I case of costat failure rate λ the mea time to failure m is equal to the iversed ratio of the costat expected failure rate λ (m=/λ). The uavailability U or Forced Outage Rate (FOR) is the ratio of the mea time to repair r to the sum of the mea time to repair r ad the mea time to failure m: r r λ U = = = (5) m r T μ λ Where λ is the costat expected failure rate, μ the costat expected repair rate, m the mea time to failure, r the mea time to repair (=/μ), T the cycle time or the mea time betwee (=mr). I case of a complex system with compoets coected i series the total reliability fuctio R series (t) is give by the multiplicatio of the respective reliability fuctios R i (t) of each i-th compoet: R () t = R() t (6) series If the failure rates of all compoets are costat, the the respective total failure rate λ series is give by: λ series i = λ (7) I case of a complex system with compoets coected i parallel ad the operatio of the system demads the operatio of oly oe compoet the the total reliability fuctio R parallel (t) is give by the followig equatio with the assumptio of very small repair rate of each compoet: R parallel ( i () t = R i () t (8) If the failure rates of all compoets are costat the the respective total failure rate λ parallel is give by:... λ λ2 λ λ λ2 λ λ3 λi λ j λ (9) parallel λ λ2 λ3 λi λj λ k ( ) λi Eq. (9) is modified for two compoets as followig: = (0) λparallel 2 λ λ2 λ λ2 I case of a system with series ad parallel compoets, the respective calculatios ca be realized simplifyig the examied system step-bystep ito subsystems. ) ISBN:

3 3 Sub-systems Reliability Study of A Diesel Electric Power Uit 3. Geeral A diesel egie drive electric power uit is a simple power uit compared with power uits employig steam, gas turbies or uclear power uits. The required iitial ivestmet capital is smaller tha the those of other classical power uits but fuel cost is larger. Nomial power usually rages betwee 20 kw ad MW ad it is used either as the mai power uit i autoomous small power systems, such as islads, ships, etc., or as emergecy uit i hospitals, hotels, worksites, supermarkets, etc. Here, the typical costructio of a diesel electric power uit for marie applicatios is aalyzed. Specifically, it cosists of the followig sub-systems ad major compoets with the respective expected failure rates i brackets: diesel egie (λ DIES ), electric geerator (λ GEN ), frequecy regulator (λ FR ), automatic voltage regulator (λ AVR ), compressed air circuit (λ AIR ), lubricatig oil circuit (λ OIL ), fuel circuit (λ FUEL ), treated water coolig circuit (λ TW ), sea water coolig circuit (λ SW ). If the aforemetioed rates are kow, the total failure rate of the system is give by: λdies λgen λfr λavr λtot () λair λoil λfuel λtw λsw Eq. () is the last step of the bottom-up reliability aalysis of this thermal electric power uit ad it assumes that if oe system fails the etire uit fails. However, the determiatio of the expected failure rate of each sub-system requires kowledge regardig the costructio ad operatio of each sub-system. Two major circuits are aalyzed ext; amely fuel ad sea water coolig circuits. It is assumed that all failure rates are costat. 3.2 Fuel circuit Fuel circuit is preseted i Fig. 2. Next, fuel circuit elemets are listed with the respective idicative umbers of Fig. 2 give i brackets: fuel tak (), pipes (2), soleoid valve (3), gear fuel pump (4), relief valve (5), fuel filters (6), (two are used i parallel coectio), remote pressure gauge (7), pressure switch - boss (8), slide valve (9), fuel ijectio pump (0), fuel pipes coectig ijectio pump with burers (), (eight fuel pipes, as may as the fuel burers, are used ), 8 cylider fuel burers (A- B4) (2), drip tak (3), had-pump (4), o-retur valve (5). The operatio of the fuel circuit is the followig: Gear fuel pump (4) drive by the egie pumps fuel from daily tak () ad delivers it to the duplex filters (6) at a specific pressure (usually ot exceedig,5 bar). The relief valve (5) discharges, at higher pressures, the excessive fuel flow. pipe (2) soleoid valve (3) had pump (4) pipe (2) pressure switchfuel tak () boss (8) remote o-retur gauge (7) Α A2 A3 A4 valve (5) filter (6) slide valve (9) relief valve (5) gear fuel pump (4) fuel ijectio pump (0) Ijector leak detector pipe (2) B B2 B3 B4 Fuel drais drip tak (3) Fig. 2. Fuel circuit of a typical diesel thermal electric power uit for ship. ISBN:

4 At the outlet of the filter (6), the fuel is supplied to slide valve (9) by a duct. This duct is fitted with two bosses for coectig a remote pressure gauge (7) ad a pressure switch - boss (8) - cotrollig a audible ad visible alarm ad the egie if pressure falls uder a specific level (usually 0,5 bar). Slide valve (9) (at rest positio) feeds with fuel the ijectio pump (0) which delivers it to ijectors through two sheathed pipes sets. Both discharges coected to barrels are joiig for direct retur to daily tak of excess fuel (). The drai coected to the drip tak (3) allows to remove booster pump (4) leaks, duplex filter (6) draiig, leaks at ijectio pipe coupligs, ijector leaks collected at cylider-head outlets, slidevalve pisto leaks ad overflows. Oe had-pump (4), attached to the egie at the air-itercooler level, draws dowstream fuel booster pump (4) ad delivers through a o-retur valve (5) before filter (6). It allows pressurizig ad draiig o the egie fuel circuit. For the purposes of the reliability study it is oted that: The expected failure rate of the i-th elemet is deoted with λ FUEL-i. The filters (6) are coected i parallel, as oe filter is eeded for the ormal operatio of the fuel circuit. Assumig that the two filters have the same expected falure rates λ FUEL-6, the respective equivalet failure rate of the duplex filters λ FUEL-p6 is obtaied by eq. (0): = λ λ λ λ λ FUEL p6 FUEL 6 FUEL 6 FUEL 6 FUEL 6 2 λfuel p6 = λfuel 6 3 (2) The eight burers (2) are coected i series, because if oe burer fails, the diesel egie will be shut dow immediately for safety reasos. The same happes also i case of fuel pipes coectig ijectio pump with burers (). The drip tak (3) ad the had pump (5) are ot cosidered as crucial elemets for the thermal uit s operatio. The isolatio valves, the o-retur valve (5) ad the pipes of the circuit are referred to as pipe (2). Hece, the equivalet expected failure rate λ FUEL of the fuel circuit is give by: λfuel λfuel 2 λfuel 3 λfuel 4 2 λfuel 5 λfuel 6 λfuel 7 λfuel 3 (3) λfuel 8 λfuel 9 λfuel 0 8 λfuel 8 λfuel Sea water coolig circuit Sea water coolig circuit is preseted i Fig. 3. Next, water coolig circuit elemets are listed with the respective idicative umbers of Fig. 3 give i brackets: sea water ilet (), pipes from poit A to poit B (2), filter (3), high pressure cetrifugal sea water pump (4), pipes from poit B to poit C (5), air-itercooler (6), lubricatio oil/sea water exchager (7), geerator air-cooler (8), thermostatic valve (9), treated fresh water/sea water exchager (0), pilot-valve () with its sesig ad cotrollig equipmet ( / ), sea water outlet (2), pipes from poit D to poit B (3), o-retur valve (4), fire fightig sea water circuit (pipes & pumps) (5). Through flow type tube exchagers mouted i series, the sea water cools successively the charge-air delivered by the turbo-charger, the lubricatig oil for the egie, the coolig air loop for the alterator ad the coolig treated water for the egie. Sea-water flow is provided by oe high pressure cetrifugal pump (4) drive by the diesel egie. Draw trough filter (3) sea water is at first delivered by the pump to the air-itercooler (6) ad the to the lubricatio oil/sea water exchager (7), where the sea water chills the lubricatio oil. The it flows through the thermostatic valve (9) distributed i oe part flowig to the alterator geerator air-cooler (8), ad aother oe through pipe (7). Accordig to the temperature measured by the sesor of the thermostatic valve (9) the flow will be directed i the air-cooler (8) ad by pass pipe (7). The alterator air-cooler outlet ad the by-pass pipe are coected i order to deliver the etire flow ito treated fresh water / sea water exchager (0) outside the uit. Treated fresh water / sea water heat exchager (0) outlet is coected to the pilot-valve () to prevet from ay temperature variatios of the sea water eterig the circuit by re-circulatig a part of the heated sea water while the remaiig part is removed outside the ship hull. The re-circulatig pipe is coected to the pump suctio upstream equipmet ( / ). This equipmet seses ad cotrols the temperature of the cold sea water mixture i the egie. ISBN:

5 A sea water ilet () pilot valve sea water () outlet (2) C pipes (2) pipes (5) lubricatio oil/sea water exchager pipes (5) pipes (5) sea water filter (3) pipes (2) Airitercooler (6) Α A2 A3 A4 pipes (2) HP cetrifugal pump (4). B pitot valve equipmet ( / ) B pipes (3) fire pumps (5) D o-retur valve (4) B2 B3 B4 Fig. 3. Sea water coolig circuit of a typical diesel thermal electric power uit for ship. (7) treated fresh water /sea water exchager (0) by-pass pipe (7) thermostatic valve (9) pipes (5) geerator (8) air-cooler If ecessary, the electric pump of the fire circuit may be used istead of the drive pump (4) by meas of the o-retur valve (4) upstream of the tappig (poit D) provided for coectig this emergecy circuit. It is oted that for this reliability study: The expected failure rate of the i-th elemet is deoted with λ SW-i. The isolatio valves ad the pipes of each sectio of the circuit are referred to as pipe ((2), (5) ad (3)). The elemets () to (4) betwee poits A ad B are coected i series. If oe of them fails the the respective sectio of the sea water circuit fails. Applyig eq. (7) the equivalet failure rate is obtaied as followig: λsw A B = λfw λfw 2 λfw 3 λfw 4 (4) The elemets (5) to (2) betwee poits B ad C are coected i series. If oe of them fails the the diesel egie will be shut dow immediately for safety reasos. Applyig eq. (7) the equivalet failure rate is equal to: λfw 5 λfw 6 λfw 7 λfw 8 λsw B C (5) λfw 9 λfw 0 λfw λfw 2 I a emergecy coditio of the sea water pump (4) the fire fightig sea water circuit ca be used. The respective elemets, (3) to (5), betwee poits D ad B are i series coectio. If oe of them fails the the respective sectio of the sea water circuit will be out of operatio. Applyig eq. (7) the equivalet failure rate is obtaied as: λ = λ λ λ (6) SW D B FW 3 FW 4 FW 5 Circuit sectio betwee poits A ad B ad the sectio betwee poits D ad B are i parallel coectio, as oly oe of them is eeded for the ormal operatio of the sea water coolig circuit. Based o the equivalet expected failure rates of eq. (4) ad (6), the respective equivalet failure rate λ SW-pB is obtaied with use of eq. (0): λsw A B λsw D B (7) λsw pb λsw A B λsw D B So the equivalet expected failure rate λ SW of the sea water coolig circuit is give by: λ = λ λ (8) SW SW pb SW B C 4 Case Study 4. Geeral I this case study all expected failure rates are assumed costat. The target of the respective aalysis is to study the sesitivity of each circuit failure rate for differet failure rates of its compoets. 4.2 Fuel circuit The most vulurable elemets of the fuel circuit are the fuel filters (6) ad the fuel pipes from ijectio pump to burers () with typical expected failure rates of /000 [/h] ad /5000 [/h] respectively. The failure rate of each oe of the rest compoets is /00000 [/h]. The followig scearios have bee examied: Sceario (basic): The typical failure rates of ISBN:

6 all elemets are used. Sceario 2: Oe filter is used. Sceario 3: The failure rate of the filter is miimized to /0000 [/h]. Sceario 4: The failure rate of the fuel pipes from ijectio pump to burers is miimized to /0000 [/h]. Sceario 5: The failure rate of the rest elemets is icreased at /50000 [/h]. The respective results of the reliability aalysis are registered to Table. Accordig to the obtaied results the additio of oe filter improves sigificatly the overall reliability of fuel circuit. The use of a filter with smaller failure rate does ot improve sigificatly the total failure rate, as 0 times smaller filter failure rate improves the total failure rate oly by 24.6%. The failure rate of the fuel pipes from ijectio pump to burers is more crucial, as a 2 times smaller failure rate improves the total failure rate by 32.8%. The failure rates of the rest elemets are ot sigificat, as log as they are sigificatly smaller tha the respective oes of the crucial elemets. TABLE FUEL CIRCUIT TOTAL FAILURE RATE FOR 5 DIFFERENT SCENARIOS No. of sceario Fuel circuit total failure rate [/h] Improvemet comparig to () % % % % 4.3 Sea water coolig circuit The most valurabe elemets of the this circuit are the sea water filter (3), the air-itercooler (6), the lubricatio oil/sea water exchager (7), the geerator air-cooler (8) ad the treated fresh water/sea water exchager (0). The typical expected failure rates are /000, /5000, /5000, /5000 ad /5000 [/h] respectively while the failure rate of each oe of the rest compoets is /00000 [/h]. Oly the fire fightig circuit has its ow failure rate with a typical value of /0000 [/h]. The followig scearios have bee examied: Sceario (basic): The typical failure rates of all elemets are used. Sceario 2: The failure rate of the filter is miimized to /2000 [/h]. Sceario 3: The failure rate of the lubricatio oil/sea water exchager (or ay other elemet from (6), (8) or (0)) is miimized to /0000 [/h]. Sceario 4: The failure rate of the rest elemets except from the firefightig circuit is icreased to /50000 [/h]. The respective results are registered i Table 2. Accordig to the obtaied results the use of a filter with smaller failure rate does ot improve practically the total failure rate, as a 2 times smaller filter failure rate of improves the total rate oly by 0,4%. The failure rate of the lubricatio oil/sea water exchager (or of the air-itercooler, the geerator air-cooler ad of the treated fresh water/sea water exchager) is more crucial, as a 2 times smaller failure rate of the exchager improves the total rate by 0.4%. The failure rates of the rest elemets are ot sigificat, as log as they are sigificatly smaller tha the respective oes of the crucial elemets. TABLE 2 SEA WATER COOLING CIRCUIT TOTAL FAILURE RATE FOR FOR 4 DIFFERENT SCENARIOS No. of sceario Sea water coolig circuit total failure rate [/h] Improvemet comparig to () % % % 5 Coclusios This paper presets study of specific idividual subsystems of a diesel electric power uit i terms of their reliability based o costat failure rate assumptios for each of their compoet. The target is to idetify which elemets affect more the total failure rate of the circuit they belog. I future this bottom-up reliability study will be upgraded to take ito accout all circuits of a diesel egie power uit. The results will be combied to obtai the FOR of the power uit. Ackowledgemets The authors would like to thak Dr. N. Mastorakis, Professor at Helleic Naval Academy ad expresidet of WSEAS for his overall support ad scietific guidace. Refereces: [] E.N. Dialyas. Power System Reliability. Athes, 996 (i greek). [2] R. Billito, R. N. Alla. Reliability Evaluatio of Power Systems. Pleum Press, New York ad Lodo, 996. ISBN: