Transitioning to Physics-of-Failure as a Reliability Driver in Power Electronics

IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 2, NO. 1, MARCH 2014 97

Transitioning to Physics-of-Failure as a ReliabilityDriver in Power Electronics

Huai Wang, Member, IEEE, Marco Liserre, Fellow, IEEE, Frede Blaabjerg, Fellow, IEEE,Peter de Place Rimmen, John B. Jacobsen, Thorkild Kvisgaard, and Jrn Landkildehus

Abstract Power electronics has progressively gained animportant status in power generation, distribution, and consump-tion. With more than 70% of electricity processed through powerelectronics, recent research endeavors to improve the reliabilityof power electronic systems to comply with more stringentconstraints on cost, safety, and availability in various applications.This paper serves to give an overview of the major aspectsof reliability in power electronics and to address the futuretrends in this multidisciplinary research direction. The ongoingparadigm shift in reliability research is presented first. Then,the three major aspects of power electronics reliability arediscussed, respectively, which cover physics-of-failure analysisof critical power electronic components, state-of-the-art designfor reliability process and robustness validation, and intelligentcontrol and condition monitoring to achieve improved reliabilityunder operation. Finally, the challenges and opportunities forachieving more reliable power electronic systems in the futureare discussed.

Index Terms Capacitors, design for reliability (DFR),insulated-gate bipolar transistor (IGBT) modules, physics-of-failure (PoF), power electronics, robustness validation.

I. INTRODUCTION

POWER electronics enables efficient conversion andflexible control of electric energy by taking advantageof the innovative solutions in active and passive components,circuit topologies, control strategies, sensors, digital signalprocessors, and system integrations. While targets concerningefficiency of power electronic systems are within reach, theincreasing reliability requirements create new challenges dueto the following factors:

1) mission profiles critical applications (e.g., aerospace,military, more electrical aircrafts, railway tractions, auto-motive, data center, and medical electronics);

2) emerging applications under harsh environment and longoperation hours [e.g., onshore and offshore wind tur-

Manuscript received April 22, 2013; revised August 21, 2013 andOctober 12, 2013; accepted November 1, 2013. Date of publicationNovember 11, 2013; date of current version January 29, 2014. Recommendedfor publication by Associate Editor Philip T. Krein.

H. Wang and F. Blaabjerg are with the Department of Energy Technology,Aalborg University, Aalborg DK-9220, Denmark (e-mail: [email protected];[email protected]).

M. Liserre is with the Department of Faculty of Engineering, Christian-Albrechts-University of Kiel, Kiel 24143, Germany (e-mail: [email protected]).

P. de Place Rimmen and J. Landkildehus are with the Research andDevelopment Design Center, Danfoss Power Electronics A/S, GrstenDK-6300, Denmark (e-mail: [email protected]; [email protected]).

J. B. Jacobsen and T. Kvisgaard are with GRUNDFOS HoldingA/S, Bjerringbro DK-8850, Denmark (e-mail: [email protected];[email protected]).

Color versions of one or more of the figures in this paper are availableonline at http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/JESTPE.2013.2290282

bines (WTs), photovoltaic (PV) systems, air conditions,and pump systems];

3) more stringent cost constraints, reliability requirements,and safety compliances, e.g., demand for parts permillion (ppm) level failure rates in future products;

4) continuous need for higher power density in power con-verters and higher level integration of power electronicsystems, which may invoke new failure mechanisms andthermal issues;

5) uncertainty of reliability performance for new materialsand packaging technologies (e.g., SiC and GaN devices);

6) increasing complexity of electronic systems in termsof functions, number of components, and control algo-rithms;

7) resource constraints (e.g., time and cost) for reliabilitytesting and robustness validation due to time-to-marketpressure and financial pressure.

Table I lists the industrial challenges in a reliability perspec-tive of yesterday, today, and tomorrow. To meet the futureapplication trends and customer expectations for ppm levelfailure rate per year, it is essential to have better understandingof failure mechanisms of power electronic components and toexplore innovative Research and Development approaches tobuild reliability in power electronic circuits and systems.

From this perspective, opportunities exist for powerelectronics to expand its role in dealing with efficient andreliable power processing in different kinds of applications.Nearly four decades ago, the scope of power electronicswas defined by William E. Newell as three of the majordisciplines of electrical engineering shown in Fig. 1(a) [1].Likewise, the future reliability research in power electronicsinvolves multidisciplinary knowledge as defined here, shownin Fig. 1(b). It covers the following three major aspects:1) analytical analysis to understand the nature of why andhow power electronic products fail; 2) design for reliability(DFR) and robustness validation process to build in reliabilityand sufficient robustness in power electronic products duringeach development process; and 3) intelligent control andcondition monitoring to ensure reliable field operation underspecific mission profiles. Robustness validation is a processthat is widely accepted and implemented in the automotivesector, which is to demonstrate that a product performs itsintended functions with sufficient margin under a definedmission profile within its specified lifetime [2]. Mission pro-file is a representation of all of the relevant operation andenvironmental conditions throughout the full life cycle [2]through production process, test, shipping, and service to

2168-6777 2013 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

98 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 2, NO. 1, MARCH 2014

TABLE IRELIABILITY CHALLENGE IN INDUSTRY SEEN BEFORE, TODAY, AND IN THE FUTURE

Fig. 1. Defined scope in (a) power electronics by William E. Newell in the1970s [1] and (b) power electronic reliability research needs seen from today.

end of life. The robustness validation process involves theactivities of verification, legal validation, and producer riskmargin validation.

This paper gives an overview of power electronics reliabilityfrom the three respective aspects defined in Fig. 1(b) andaddresses the future trends in this multidisciplinary researchdirection. Section II lists the ongoing paradigm shift inreliability research in power electronics. Section III presentsthe physics-of-failure (PoF) analysis of reliability-criticalcomponents to provide a basis for system level design.Section IV discusses the state-of-the-art DFR and robustnessvalidation process to build-in reliability through design.Section V presents the control and monitoring (including fault-tolerant strategies) methods to improve reliability of powerelectronic systems under field operation. Finally, the futurechallenges and opportunities in reliability of power electronicsare discussed.

II. ONGOING PARADIGM SHIFT IN RELIABILITYRESEARCH IN POWER ELECTRONICS

Reliability is defined as the ability of an item to performthe required function under stated conditions for a certainperiod of time [3], which is often measured by probabilityof survival and failure rate. It is relevant to the durability(i.e., lifetime) and availability of the item. The essence ofreliability engineering is to prevent the creation of failures. Thedeficiencies in the design phase have effect on all produceditems and the cost to correct them is progressively increasedas the development proceeds.

TABLE IITYPICAL LIFETIME TARGET IN DIFFERENT POWER ELECTRONIC

APPLICATIONS

A. Reliability in Typical Power Electronic ApplicationsThe performance requirements of power electronic products

are increasingly demanding in terms of cost, efficiency,reliability, environmental sustainable materials, size, andpower density. Of which, the reliability performance hasinfluences on the safety, service quality, lifetime, availability,and life cycle cost of the specific applications. Table II liststhe typical design target of lifetime in different applications.To meet those requirements, paradigm shift is going on in thearea of automotive electronics, more electrical aircrafts, andrailway tractions by introducing new reliability design toolsand robustness validation methods [2], [4], [5].

With the increasing penetration of renewable energy sourcesand the increasing adoption of more efficient variable-speedmotor drives [6][8], the failure of power electronic convertersin WTs, PV systems, and motor drives are becoming an issue.Field experiences in renewables reveal that power electronicconverters are usually one of the most critical assembliesin terms of failure level, lifetime, and maintenance cost [9].For example, it shows that frequency converters cause 13% ofthe failure level and 18.4% of the downtime of 350 onshoreWTs in a recent study associated with 35 000 downtime events[10]. Another representative survey in [11] concludes thatPV inverters are responsible for 37% of the unscheduledmaintenance and 59% of the associated cost during five yearsof operation of a 3.5-MW PV plant. It should be noted thatsuch statistics always look backward as those designs are morethan 10 years old. The present technology could have differentfigures.

B. Ongoing Paradigm Shift in Reliability ResearchThe reliability engineering has emerged as an identified

discipline since the 1950s with the demands to address

WANG et al.: TRANSITIONING TO PHYSICS-OF-FAILURE AS A RELIABILITY DRIVER 99

the reliability issues in electronic products for militaryapplications [12]. Since then, much pioneer work has beendevoted to various reliability topics. One of the main streamsis the quantitative reliability prediction based on empiricaldata and various handbooks released by military and industry[13]. Another stream of the discipline focuses on identifyingand modeling of the physical causes of component failures,which was the initial concept of the PoF presented in 1962[14]. However, until the 1980s, the handbook based constantfailure rate models (e.g., Military-Handbook-217 series [15])have been dominantly applied for describing the useful life ofelectronic components. Since the 1990s, with the increasedcomplexity of electronic systems and especially the applicationof integrated circuits (ICs), more and more evidences weresuggesting that constant failure rate models are inadequate[16]. The Military-Handbook-217F is therefore officiallycancelled in 1995. The PoF approach has started to gain itsmore and more important role in reliability engineering.

In recent years, the initiatives to update the Military-Handbook-217F have turned to a hybrid approach, whichis proposed for the planned version of Military-Handbook-217H [17]. During the stage of programs acquisition-supplierselection activities, updated empirical models will be used forcomparing different solutions. During the actual system designand development stage, scientific-based reliability modelingtogether with probabilistic methods will be applied. IntensivePoF research has been continuously conducted since 1990s inmicroelectronics and the state-of-the-art results are presentedin [16] and [18]. With the transition from pure empirical-basedmethods to more scientific-based approaches, the paradigmshift in reliability research is going on from the followingaspects.

1) From Components to Failure Mechanisms: The PoFapproach is a methodology based on root-cause failure mecha-nism analysis and the impact of materials, defects, and stresseson product reliability [19]. It changes the analysis of systemfrom a box of components to a box of failure mechanisms.The traditional handbook-based reliability prediction providesfailure rate models for various components. The PoF approachanalyzes and models each failure mechanism induced byenvironmental and usage stresses. For a given component,there could be multiple failure mechanisms, which should beidentified individually. In addition, failure mechanisms are notlimited to the component level. As discussed in the standardANSI/VITA 51.2 [18], there are various failure mechanismsin component level (i.e., single transistor level), package level,and printed circuit board (PCB) level. From this prospective,it is challengeable to apply the PoF to a complex systemof which limited number of models and their associatedparameters are available [18]. Therefore, it is important toidentify and to focus on the critical failure mechanisms inspecific applications.

2) From Constant Failure Rate to MCF Curve: The con-ventional reliability metrics constant failure rate (defined as) and the corresponding mean-time-between-failures (MTBF)(defined as 1/) are found to be inappropriate to most prac-tical cases as discussed in [9], [13], and [20]. Therefore, itdiscouraged the indiscriminate use of these metrics.

Fig. 2. Example of MCF or M(t) curve for explaining and measuringreliability.

The failure rate over operational time is not constant.An alternate technique to present the failure level and time isthe mean cumulative function (MCF) curve [21]. When analyz-ing repairable systems, it graphs the number of failures versustime (i.e., since installation). It is also possible to representthe behavior of the group of systems by an average number offailures versus time, which is known as the MCF. As shownin Fig. 2, it can be broken down to the main functions, whichcan be described as following parameters: Zero time failuresoccur from lack of robustness to transportation or installation,indicated by the red line. Early failures come from lack ofproduction capabilities. Some few products are slipped throughthe control parameters in the production, which are marked asthe green Weibull curve ( 1, where is defined as theshape parameter in Weibull distribution [22]). If the product isnot robust to catastrophic stress, the product might fail. Thisweakness is designed in and the time when failure occurshas nothing to do with the age of the product. The onlyway such accident can be shoved as random in the operationtime ( = 1), marked in orange curve. This has nothing withfailure rate values or MTBF to do. The last dominant curve(i.e., the blue one) is the lack of lifetime. This is the accumu-lated degradation for all parameters, which are able to degradeas a function of operational time. The customer will be theperson who sees the accumulated failure level of all theseweaknesses in the purple curve shown in Fig. 2. This figure isalso an integration of the bathtub model, but here it is possibleto operate with quantitative figures, which can be broken downin budgets (e.g., the degradation budget).

3) From Reliability Prediction to Robustness Validation:Conventional empirical methodologies mainly attempt todetermine the feasibility in fulfilling certain reliability goalsand to predict the warranty-costs and maintenance-supportrequirements [12]. They provide limited insights in the designof the systems themselves to eliminate failures within targetedservice life. Compared with them, the concept of the PoF isto identify the root causes of different types of failure underenvironmental and operation stress conditions. Therefore, ithelps to locate the weak links and formulate the correspondingguidelines on robustness design, process control, validationtesting, and filed operation. Take the design as an example;Fig. 3 shows how the designer shall understand degradation.Products should be designed by considering the degraded


Fig. 3. Concept for robustness design.

parameters at the end-of-life with certain level of designmargins. It is also shown that it makes only sense to measureperformance inside the design specifications. The optimalsituation is that the design is so good that no weakness canbe found inside the customer or design specifications. Allvalidations to reliability, lifetime, and robustness have to bedemonstrated outside or at design specifications under relativehigh stress levels.

4) From Microelectronics to Also Power Electronics: ThePoF approach has been extensively applied to microelectronicsystems in the last two decades. Different failure mechanisms,lifetime models, and equivalent damaged circuit simulationmodels of electron devices are well presented in [16].More and more new models are under development. SeveralPoF-based industry standards or guidelines have been released(see [2] and [18]). One of the common driven factors fromindustry, academia, and military behind this is the demandingfor more reliable commercial-of-the-shell devices and systems.

In power electronic applications, reliability has been andwill continue to be one of the important performance aspects inmany applications as discussed in Section II-A. To address thechallenges discussed in Section I, power electronic engineersand scientists have started to apply various reliability tools forreliability prediction and reliability-oriented design of powerelectronic converters or systems. Several literature reviewson field experiences [23], strategies to improve reliability ofpower electronic systems [24], and DFR for power electronicsystems [9] have been presented in the last two years. Respec-tive research in different applications is also discussed invarious literatures, such as three-phase converters for aircrafts[25], power inverters for railway tractions [26], invertersfor hybrid electric vehicles [27], high power variable-speedmotor drives [28], and pulsed power converters for industrialprocess control [29]. Besides these applications, last decadealso saw much pioneering work on the reliability of power

converters for WTs [30][32] and inverters for PV systems[33][50]. It reveals that, unlike the case in microelectronics,conventional handbook methods are still dominantly appliednowadays for the reliability prediction in those studies.

While the pace of power electronics toward PoF approachis relatively slower than that of microelectronics, the need forthis paradigm shift has been well recognized in automotiveindustry [2] and then in other sectors. Especially, muchinteresting work from the semiconductor side investigatesthe failure mechanisms of insulated-gate bipolar transistor(IGBT) modules [51] and physical-based lifetime models [52].More realistic thermal stress analysis of Si- and SiC-baseddevices under long-term mission profile are also studied in [49]and [50], respectively. The level of technology and scientificunderstanding is still highly evolving. The research in micro-electronics could provide a very important foundation for theongoing and future work in power electronics, especially fromthe methodologies point of view. Nevertheless, it should benoted that most of the physical-based models are not scalablefor power electronic components. System level reliabilityproblems (e.g., active thermal stresses, interconnections amongcomponents, and interaction of different components) are stillof interest to be investigated. Therefore, the following threesections intend to provide a basic framework of the futurereliability research in power electronics relevant to the ongoingparadigm shift.

III. PoF ANALYSIS OF RELIABILITY CRITICALCOMPONENTS IN POWER ELECTRONICS

As shown in Fig. 1(b), understanding of the reliabilityphysics of components applied in power electronics is themost fundamental aspect. The PoF approach is based onanalyzing and modeling each failure mechanism under variousenvironmental and usage stresses. In practice, the PoF analysisfocuses on critical components under critical stress conditions.


TABLE IIIFPM IN RELIABILITY OF POWER ELECTRONIC COMPONENTS

Among other components, switching devices and capacitorsare two of the most vulnerable components in terms of failurelevel and time as analyzed in [23], [34], [35], and [53].They are considered as the reliability critical componentsin power electronic converters, especially the IGBT modulesin medium to high-power applications and capacitors fordc-link applications. Therefore, in the following parts, thecritical stressors for different power electronic components arediscussed first. Then, the PoF analysis of IGBT modules anddc-link capacitors is given.

A. Critical Stressors for Different Power ElectronicComponents

Focus point matrix (FPM), as suggested in [2], is a use-ful way to analyze the critical stressors that will kill thecomponents. Based on the accumulated industrial experiencesand future research needs, Table III shows the critical stres-sors for different components in power electronic systems.It can be noted that steady-state temperature, temperatureswings, humidity, voltage, and vibrations have different levelof impact on semiconductor devices, capacitors, inductors,and low power control boards. It provides the information ondetermining the critical failure mechanisms. The interactionsamong different stressors are also of interest to be explored.

B. PoF Analysis of IGBT ModulesFig. 4 shows a typical structure of the IGBT modules [54].

There are three dominant wear-out failure mechanisms for theIGBT modules due to cyclic thermal stress: 1) baseplate solderjoints cracking; 2) chip solder joint cracking; and 3) the wirebonds liftoff. The cyclic thermal stress is a response to theconverter line and loading variations as well as periodically

Fig. 4. Structural details of an IGBT module (connections that are relevantto module lifetime are marked red) [54].

Fig. 5. Typical stress-strain ( ) curve for a material [55].

commutation of power switching devices. It will induce ther-mal cycling on different layers of materials used for fabricationof power electronic components.


Fig. 6. Typical catastrophic failure of IGBT modules [60].

TABLE IVOVERVIEW OF FAILURE MODES, CRITICAL FAILURE MECHANISMS AND CRITICAL STRESSORS OF THE THREE MAIN TYPES OF DC-LINK CAPACITORS

(WITH EMPHASIS ON THE ONES RELEVANT TO DESIGN AND OPERATION OF POWER CONVERTERS)

Thermal cycling is found to be one of the main drivers forfailure of the IGBT modules. The effect of the temperaturecycling can be explained by the typical stressstrain curve [55]shown in Fig. 5. The is defined as the cyclic stress (e.g.,temperature cycling) and is defined as the deformation. Witha low cyclic stress below yield, no damage occurs and thematerial is in the elastic region. When the stress is increasedabove yield, an irreversible deformation is induced and thematerial enters into the plastic region. The coefficients ofthermal expansion of different materials in the IGBT modulesare different, leading to stress formation in the packaging andcontinuous degradation with each cycle until the material fails.As derived in [9], the number of cycles to failure under thermalcycling can be obtained as

N = k (T T0)m (1)where k and m are empirically determined constants and Nis the number of cycles to failure. The T is the thermal

cycling range and T0 is the portion of T that in the elasticstrain range. If T0 is negligible compared with T , it canbe dropped out from the above equation, which then becomesthe CoffinManson model discussed in [56][58].

The model shown in (1) considers the influence of thermalcycling only. It does not take into account the effect of steady-state temperature, thermal cycle time, and geometry. In [59],an empirical model is developed for bond wire fatigue ofIGBT modules, which tends to treat all the above factorsas well as failure of the diodes in parallel with the IGBTswitches. In [52], a physical-based model for wire bond fatiguehas been developed which could analyze the cycle-to-failureunder different steady-state temperature and thermal cycletime. Although it may be difficult to obtain some of theparameters required by the model, it is a promising modelin its kind for the PoF analysis.

Besides wear-out failure that was discussed, different typesof catastrophic failure could also occur triggered by single-


TABLE VWEAR-OUT FAILURE CRITERION AND PARAMETERS FOR CONDITION

MONITORING OF CAPACITORS

Fig. 7. 2L-BTB converter for a 2.3-MW WT using a permanent magnetsynchronous generator [9].

TABLE VICONVERTER PARAMETERS FOR THE CASE STUDY [9]

event overstress. Unlike the wear-out failure, the catastrophicfailure is difficult to be predicted and thus may lead toserious consequence to the power electronic converters. Fig. 6classifies the IGBT catastrophic failure into open-circuit andshort-circuit modes induced by different failure mechanisms.It should be noted that both wear-out and catastrophic failuresmay have the same failure mechanisms (e.g., bond wire liftoff)but the former one is due to long-term degradation (see theblue curve in Fig. 2) and the latter one is due to single-eventoverstress within short-time duration (see the orange curve inFig. 2).

C. PoF Analysis of DC-Link CapacitorsThe dc-link capacitors contribute to cost, size, and failure

of power electronic converters on a considerable scale. Toaddress the issue, research efforts can be divided into two

TABLE VIILIFETIME PREDICTION RESULTS OF THE SELECTED IGBT MODULES

directions: 1) advance the capacitor technology with improvedand predetermined reliability built in and 2) a proper andoptimal dc-link design based on the commercially availablecapacitors to ensure reliable field operation. The latter oneis more relevant from the perspective of power electronicdesigners, which is discussed here.

Three main types of capacitors are available for dc-linkapplications, which are the aluminum electrolytic capacitors(Al-Caps), metallized polypropylene film capacitors, (MPPF-Caps) and high capacitance multilayer ceramic capacitors(MLC-Caps). The dc-link design requires the matching ofavailable capacitor characteristics and parameters to the partic-ular application needs under specific environmental, electrical,and mechanical stresses. Table IV lists the failure modes,critical failure mechanisms, and corresponding stressors. Moredetailed discussions on them have been given in [61]. Table Vlists the wear-out failure criterion and typical electrical para-meters for the condition monitoring of capacitors.

Lifetime prediction of capacitors is mainly based on empiri-cal models as physical-based models are still not available. Themost widely used empirical model for capacitors is shown in(2), which describes the influence of temperature and voltagestress

L = L0 (

VV0

)n exp

[(EaK B

)(1T

1T0

)](2)

where L and L0 are the lifetime under the use conditionand testing condition, respectively. V and V0 are the voltageat use and test conditions, respectively. T and T0 are thetemperature in Kelvin at use and test conditions, respectively.Ea is the activation energy, K B is Boltzmanns constant(8.62 105 eV/K), and n is the voltage stress exponent.Therefore, the values of Ea and n are the key parameters tobe determined in the above model.

However, the voltage dependence of lifetime for Al-Capsquite depends on the voltage stress level. In [62], instead of apower law relationship, a linear equation is found to be moresuitable to model the impact of voltage stress. In order toobtain the physical explanations of the lifetime model variantsfrom different capacitor manufacturers, a generic model isderived in [9] as shown in (3), as shown at the next top of thepage, where a0 and a1 are constants describing the voltage andtemperature dependence of Ea . Ea0 is the activation energyunder test. It can be noted that the influence of voltage stressis modeled as linear, power law, and exponential relationship,respectively, for low-, medium-, and high-voltage stresses.Another important observation is that the activation energy Ea


LL0

=

(V0V

) exp

[(EaK B

) (1T 1T0

) ]( low voltage stress)

(VV0

)n exp [( EaK B) (

1T 1T0

)]( medium voltage stress)

exp [a1 (V0 V )] exp[

Ea0a0 VK B T

Ea0a0 V0K B T0

](high voltage stress)

(3)

Fig. 8. Single-phase grid-connected PV inverter. (a) Simplified structure.(b) Instantaneous power flow.

is varying with voltage and temperature, especially under high-voltage stress condition. It is still a challenge to determine thevalue of the parameters and the boundaries of low-, medium-,and high-voltage stresses in (3).

D. Case Studies on the Application of IGBTs andCapacitors in Power Converters

1) IGBT Modules in a 2.3-MW Grid-Side Wind PowerConverter: A case study on the lifetime prediction of IGBTmodules in a 2.3-MW wind power converter has been studiedin [9]. A two-level back-to-back (2L-BTB) converter is appliedin the paper as shown in Fig. 7. The technical advantage ofthe 2L-BTB topology is the relatively simple structure andfew components, which contributes to a well-proven robustand reliable performance. Table VI lists the specifications andselections of the IGBT modules. By following the predictionprocedure from wind speed profile analysis, case tempera-ture and junction temperature estimation, cycling counting oftemperature swings to parameter estimation of lifetime models,the lifetime of two condidates of IGBT modules for the grid-side converter is predicted in [9]. The lifetime prediction isbased on each of the three critical failure mechanisms relatedto thermal cycling discussed in Section III-A. The results areshown in Table VII. It should be noted that other failuremechanisms induced by thermal stress or other types ofstresses need also to be considered, besides those listed inTable VII.

2) DC-Link Capacitors in a 1-kW PV Inverter: Electrolyticcapacitors that widely used in PV inverters are considered asthe weakest link with respect to the semiconductor devices[34], [35]. Therefore, the case study for dc-link capacitors isperformed on a 1-kW 400 V dc-link PV inverter.

Fig. 9. Simulation results of different capacitors under various ambienttemperatures [(a)(c) are with 1-kW output power and (d) is with minimumlifetime of 20 years]. (a) Capacitor power losses. (b) Capacitor hotspottemperatures. (c) Capacitor predicted lifetime. (d) Output power deratingcurves.

Fig. 8(a) shows a simplified structure of the inverter. Theinput power of the PV inverter is assumed constant within


TABLE VIIITHREE KINDS OF CAPACITORS FROM DIFFERENT MANUFACTURES CONSIDERED FOR THE PV INVERTER DESIGN

one cycle of the grid voltage. Fig. 8(b) shows the instan-taneous power balancing function of the input capacitor C .The nominal input voltage of the inverter is 400 V with amaximum voltage ripple of 5% and a maximum input voltageof 600 V. The calculated minimum required capacitance is398 F and ripple current stress is 1.8 A. A reliability-orienteddesign guideline proposed in [63] is applied for the selectionof the input capacitor to fulfill 20 years of lifetime. Accordingto the electrical stress analysis, preliminary choices of thecapacitors are determined as shown in Table VIII. Then, thethermal stresses of those capacitors are estimated based ontheir specific thermal models. The lifetime of the selectedcapacitors is therefore can be estimated based on the missionprofile, operation mode, and specific lifetime model. Theapplied empirical lifetime models from the respective capacitormanufacturers are consistent with the generic lifetime modelshown in (3). Finally, the optimal capacitors can be chosen bycomparing different options.

Fig. 9(a)(c) shows the comparison of the power loss,hotspot temperature, and lifetime prediction with 1-kW outputpower. Fig. 9(d) shows the power derating curve to fulfillthe lifetime requirement. If 100 000 h of lifetime (equivalentto 12 h/day operation in 20 years with a design margin of12.4%) is required, it can be noted from Fig. 9(c) and (d) thatonly the Case 3 can fulfill the requirement in a wide ambienttemperature range. The selection of Case 2 can only have 20years of lifetime when the ambient temperature is


Fig. 10. State-of-the-art reliability design procedure for power electronics.

fails when the applied load exceeds the design strength. Theload refers to a kind of stress (e.g., voltage, cyclic load,temperature, etc.) and the strength refers to any resistingphysical property (e.g., harness, melting point, adhesion, etc.)[64]. Load and strength of power electronic components areallocated within a certain interval, which can be described by aspecific probability density function (e.g., normal distribution).In addition, the strength of a material or device could bedegraded with time. Theoretically, the probability of failurecan be obtained by analyzing the overlap area between theload and strength distributions. From another prospective, itimplies that failure could be reduced or eliminated withinservice life by either design with an increased strength (i.e.,an increased design margin), or with a reduced load by control(i.e., stress control or load management), or both. Practically,the exact distributions of load and strength are very oftennot available, Monte Carlo simulation [64] can be applied torandomly select samples from each distribution, compare themand thus roughly estimate the probability of failure.

B. Reliability Prediction ToolboxReliability prediction (not based on constant failure rate )

is an important tool to quantify the lifetime, failure level,

and design robustness based on various source of data andprediction models. Fig. 11 shows a generic prediction toolboxbased on the PoF approach. The toolbox includes statisticalmodels and lifetime models and various sources of avail-able data (e.g., manufacturer testing data, simulation data,field data, etc.) for the reliability prediction of individualcomponents and the overall system. The statistical modelsare well presented in [64], whereas the number of physicalbased lifetime models available for power electronic com-ponents is still limited. Research efforts to both acceleratedtesting and advanced multidisciplinary simulations will bebeneficial to obtaining those lifetime models. A more detailedstep-by-step procedure for lifetime prediction is presentedin [70].

To map the reliability from component level to the systemlevel [3], reliability block diagram (RBD), fault-tree analysis,(FTA) and state-space analysis (e.g., Markov analysis) arewidely applied as summarized in Table IX.

It should be noted that the tabulated three methods areconventionally applicable to constant failure rate cases, whichare corresponding to the handbook-based reliability predic-tion methods. The PoF-based system level reliability pre-diction is still an open research topic even in microelec-tronics [16], [18]. Interactions among different failure mech-


Fig. 11. Reliability prediction toolbox for power electronic systems.

TABLE IXSUMMARY OF SYSTEM LEVEL RELIABILITY PREDICTION METHODS

anisms will bring additional complexity for the analysis.Therefore, it was argued that the PoF approach is not prac-tical for assessing an entire system in [12]. In addition,it should be noted that the system reliability depends not

only on components, but also on packaging, interconnects,manufacturing process, and human errors. The latters needalso to be treated properly for a more accurate reliabilityassessment.


V. INTELLIGENT CONTROL AND MONITORING OF POWERELECTRONIC SYSTEMS

After power electronic systems have been designed, theirreliability could be further improved through control andcondition monitoring. This is the third important aspect shownin Fig. 1(b). Among many options, three main actions can betaken to increase the reliability of power electronic systems:1) prognostics and health management; 2) active thermalcontrol for reducing temperature and temperature swing thatare the main killing factors of power devices modules; and 3)fault-tolerant operation to continue operate the system even incase of failures. The last can be considered as an alternativemeasure with respect to the first two or like the last attemptto make the system operating if it was not possible to predictfailures or to avoid them. Of course, all these actions entailimportant investments in terms of devices, sensors and controlactions, and even request redundancies. All of them shall beevaluated in terms of cost respect to the specific application.

A. Prognostics and Health ManagementThe Electronic Prognostics and Health Management

Research Center at the University of Maryland has categorizedthe main approaches as: 1) use of fuses and canary devices;2) built-in-test (BIT); 3) monitoring and reasoning of failureprecursors; and 4) modeling accumulated damage based onmeasured life-cycle loads [71].

Apart from the first category that relies on the presence ofdevices that fail before the main one (like the canary in amine full of hazardous gas), the three other categories need:1) sensors to check the functionality of a device or circuit,like in case of BIT, to monitor precursors of the failureor to measure cycles whose number can be correlated tofailure and 2) data logging to store the data coming fromsensors. Accordingly, Fig. 12 shows the configurations for theprognostics and health management. The last three categoriescan in part overlap with the concept of condition monitoringthat refers to online monitoring of the device. In case of powersemiconductor, this can be done by modifying the gate drivercircuit.

Different sensors can be employed and can be classifiedin two categories: 1) ambient sensors (temperature, humidity,and pollution) and 2) internal sensors (module temperature,vibration, and electrical parameters). Since the main failurecause for power module is the junction temperature swing,measuring it or estimating by means of thermo-sensitiveelectrical parameters (TSEPs) is one of the most interestingchallenges [72]. In fact, sensing the junction temperature dur-ing converter operation is notoriously difficultdirect accessto chips is prevented by module packaging and dielectric gel,which therefore limits the optical and physical contact methodssuch as the use of infrared cameras or optical fibers.

Electrical methods allow the measurement of temperaturewithout any physical alteration to a device. However, one dis-advantage compared with optical and physical contact methodsis that the latter can be used to measure the temperatureat specific points in the die or module. Electrical methodsgenerally give an average temperature across the die. For

Fig. 12. Prognostics and health management of power electronic systems.

instance, the voltage drop across a p-n-junction is knownto vary with temperaturea measurement of this voltagecan therefore be used to derive a temperature solely for thejunction. However, this perhaps does not give a reliable enoughestimation of temperature elsewhere in the devices, such as inbond wires, solder joints, and so forth.

Data acquisition is very important since once determined theimportant quantities to be sensed or estimated, there are issuesrelated to the amount of data that is possible to store and howto use those data. Hence, developing a data-acquisition systemtaking into account both the ambient and internal quantitiescan be a challenge [73].

The goal is to have the real-time operating characteristicsand the health conditions of the components (particularly ofthe power modules and capacitors) and of the overall powerconverter [74].

This information can be used for two main goals:1) implement a proactive maintenance plan (i.e., prognosticmaintenance) and 2) provide information for proactive controlschemes that can be a simple load management (i.e., reductionor sharing of the load among different units) or the moreadvanced active thermal control, briefly introduced in thefollowing.

B. Active Thermal ControlThe thermal analysis of power converters, especially in case

of more complex structures, such as multilevel or multicellones, reveals that some of the power semiconductor devicescan be more stressed with respect to others and this differencecan be even more evident in some particular conditions likethose caused by system faults [75]. Hence, the possibility tomodify the modulation and control of the power converterusing as a feedback the junction temperature of the moststressed device is an appealing possibility. The more straight-forward approach is manipulating the switching frequencyand the current limit to regulate the losses and prevent over-temperature or to reduce temperature swing [76]. In view ofcontrolling the junction temperature, estimating it by meansof TSEP, as already mentioned, or using an observer based


Fig. 13. Active thermal control of the power semiconductor junctiontemperature Tj by means of y (switching frequency, reactive power, or anyother quantity that can modify the power semiconductor losses). Tj is obtainedusing an estimator based on TSEP or an observer using measured voltagesand currents.

on FEM modeling of the device [76] are two interestingalternatives to the more expensive ones using of integratedsensors in the chip [77].

Fig. 13 shows the general block diagram for active ther-mal control of the power semiconductors once the junctiontemperature is measured or estimated. The easiest way isto change the switching frequency [78], [79] to control thejunction temperature. In case of parallel power converters thatpresent some redundancies, it is also possible to share theload among the different units also in view of controlling thetemperature swing [80]. Another alternative is to circulatereactive power among the different power converters con-nected in parallel in a high power converter or in a windor PV park to reduce the temperature swing in the moststressed power semiconductor devices [81]. The idea can beapplied only in case of power converters, such as the neutral-point-clamped inverters, where there is an uneven distributionof power losses and as a consequence of the temperatureof the power semiconductor devices, being this differenceeven bigger in some particular stressing circumstances likein the case of sudden power changes (e.g., for a wind gust),grid faults, or variable atmospheric conditions. The maindrawbacks are higher losses and higher mean temperatureof the most stressed device but also of the other devices.However, one risk is to move the stress from bond wires tosolder joints.

C. Fault-Tolerant OperationWorking outside the safe operating area leads power

semiconductors to damage. The main failure causes are:1) fault currents either over-current; 2) short-circuit current orearth fault current; 3) over-voltage; 4) over-temperature; and5) cosmic radiation [54]. Other problems may arise becauseof the driver of the power semiconductor: malfunctioningof the driver board, auxiliary power supply failure or dv/dtdisturbance. As a consequence, five main types of faults can beidentified: 1) single switch short-circuit (power semiconductoris desaturated working as current source or it is a physicalshort-circuit; 2) phase-leg short-circuit; 3) single switch open-circuit; 4) single-phase open-circuit; and 5) intermittent gate-misfiring [82]. Fig. 14 shows the first four types of faults.

Fig. 14. Inverter faults considered. (a) Single switch short-circuit.(b) Phase-leg short-circuit. (c) Single switch open-circuit. (d) Single-phaseopen-circuit [83].

Fig. 15. Switch redundant topology for fault tolerant control [83].

Three levels of protections can prevent failures or limit theireffects: 1) fast (in the switch, 10 ns); 2) slow (outside theswitch); and 3) very slow (system level). Several diagnosticmethods can be used to detect failures and they can be mainlyclassified in those used for open- or closed-switch failuresand in software one or hardware one. Generally, softwaremethods are more suitable in case of open-switch failure,whereas closed-switch failures need a fast detection becausethey can lead to destructive failure of the overall system,the desaturation one is the most famous [82]. Once a faultis detected and isolated or online repair is implemented,the system can continue to operate safely and fault-tolerantoperation can be implemented. There are already severalsimple solutions implemented in industry also without anyredundancy if operation with high harmonic content and lowerpower level can be accepted [83] as shown in Fig. 15.Otherwise, redundancies based on paralleling or connecting inseries power semiconductor devices [84] as shown in Fig. 16are the simplest and most adopted solution in industry anduse of devices that can continue to operate in short-circuitsuch as press-pack IGCT can help. The next step in usingredundancy to improve fault-tolerant operation is in addingextra legs to the power converter. The procedure consists ofthe following steps: 1) detection of the faulty leg; 2) stopthe control signal for the two switching drivers of the faultyleg; 3) trigger the bidirectional switch connecting the new


Fig. 16. Redundancies by connecting IGBTs in parallel or in series [84].

Fig. 17. Fault-tolerant voltage source inverter by adding extra leg for moreelectrical aircraft application [86].

Fig. 18. Redundancy achieved by connecting converters in parallel [89].

leg; and 4) use the control signals of the faulty leg for theredundant one. The use of fast digital computational devices[85] and the integration in the power module of the extra legand of a thyristor that can survive high energy pulses [86] arethe enabling technologies to isolate a faulty leg and onlinesubstitute it with a healthy one. The latter has been developedwithin an EU project for more electric aircraft as shown inFig. 17, and this is an interesting sign of the importance offault operation capability solutions that should be designed toguarantee high reliability in safety critical applications.

The last two solutions to guarantee fault-tolerant operationentail larger investments at system level and in general a sig-nificant shift in the power converter design: multiphase powerconverters and machines [87], [88] and use of parallel or seriesconnection of power converters [89], [90]. Figs. 18 and 19

Fig. 19. Common duty ratio, automatic master-slave control scheme withidentical and independent modules for input-series, output-parallel connec-tion [90].

show the redundancy achieved by connecting converters inparallel and in input-series output-parallel, respectively. Bothof the solutions have been proposed and in some casesimplemented in the aforementioned more electric aircraft [88],[89]. Particularly, connection in series or in parallel power con-verters or both is already widely adopted in dc/dc convertersand requires the power modules to be identical and capableof working independently as the case shown in Fig. 19. Inaddition, the faulted module(s) must be quickly isolated fromthe system, which is not always easy during operation of thepower converter especially taking into account the associatedtransients that should be minimized to avoid damage of thehealthy modules while replacing the faulty ones.

VI. CONCLUSIONReliability is an important performance index of power

electronic systems. The status and future trends of the DFRin power electronics are presented in this paper. A paradigmshift in reliability research on power electronics has leftsimple handbook based on constant failure rate for the PoFapproach and DFR process. Accordingly, three major aspectsof power electronics reliability are discussed: 1) the PoFanalysis of reliability critical components (e.g., IGBT modulesand dc-link capacitors) and two associated study cases; 2)the state-of-the-art DFR process and robustness validationfor power electronic systems; and 3) the prognostics andhealth management, active thermal control, and fault-tolerantstrategies for reliable field operation.

Joint efforts from engineers and scientists in the multipledisciplines are required to fulfill the research needs andpromote the paradigm shift in reliability research. The majorchallenges and opportunities in the research on reliability forpower electronic systems are addressed as follows.

A. Challenges1) Pervasive and fast implementation of power electronics

in a large variation of applications with all kind ofenvironmental exposures.


2) Outdated paradigms and lack of understanding in theDFR process in power electronics.

3) Uncertainties in mission profiles and variations instrength of components.

4) Increasing electrical/electronic content and complexity.5) Lack of understanding in failure mechanisms and failure

modes of reliability critical components.6) Traditional system level reliability prediction methods

are based on constant failure rates. However, physics-of-failure-based component level reliability predictionresults in varying failure level with time.

7) Resource-consuming testing for reliability predictionand robustness validation from components to entiresystems.

8) End up with ppm level return rates for mass-manufactured power electronic products.

9) Higher operating temperature (e.g., with wide bandgapdevices) that challenges the overall reliability and life-time.

10) Software reliability becomes an issue with more andmore digital controllers are introduced in power elec-tronic systems, which should be treated adequately.

B. Opportunities1) The research in microelectronics provides an important

foundation for the ongoing and future work in powerelectronics, especially from the methodologies point ofview.

2) More and more mission profiles and online monitoringdata from the field are available and accessible.

3) PoF approach provides insights to avoid failures inpower electronic components, circuits, and systems.

4) Active thermal control by controlling the power flow inpower electronic circuits.

5) Component level and system level smart derating oper-ations.

6) Condition monitoring and fault-tolerant design thatallow extended lifetime and reduced failure rate.

7) Emerging semiconductor and capacitor technologiesenable more reliable power electronic components andsystems.

8) Computer-aided automated design software to save timeand cost in the development process.

9) Trends for modular design of power converters and stan-dardized power electronic components and packagingtechnologies.

10) With better understanding of failure mechanisms inpower electronics, more failure mechanism specificaccelerated testing could be designed, leading toimproved reliability predictions for targeted applications.

REFERENCES

[1] W. E. Newell, Power electronicsEmerging from limbo, IEEE Trans.Ind. Appl., vol. 10, no. 1, pp. 711, Jan. 1974.

[2] Handbook for Robustness Validation of Automotive Electrical/ElectronicModules, ZVEL, Frankfurt am Main, Germany, Jun. 2013.

[3] IEEE Standard Framework for the Reliability Prediction of Hardware,IEEE Standard 1413, 2009.

[4] T. Jomier. (2009, Dec.). Final Public MOET Technical Report, EUFP6 Project on More Open Electrical Technologies [Online]. Available:http://www.eurtd.com/moet/

[5] X. Perpinya, Reliability and Safety in Railway. Rijeka, Croatia: InTech,2012, ch. 7.

[6] F. Blaabjerg, Z. Chen, and S. B. Kjaer, Power electronics as efficientinterface in dispersed power generation systems, IEEE Trans. PowerElectron., vol. 19, no. 4, pp. 11841194, Sep. 2004.

[7] F. Blaabjerg, M. Liserre, and K. Ma, Power electronics convertersfor wind turbine systems, IEEE Trans. Ind. Appl., vol. 48, no. 2,pp. 708719, Mar./Apr. 2012.

[8] S. B. Kjaer, J. K. Pedersen, and F. Blaabjerg, A review of single-phasegrid connected inverters for photovoltaic modules, IEEE Trans. Ind.Appl., vol. 41, no. 5, pp. 12921306, Sep./Oct. 2005.

[9] H. Wang, K. Ma, and F. Blaabjerg, Design for reliability of powerelectronic systems, in Proc. 38th Annu. Conf. IEEE Ind. Electron. Soc.,Oct. 2012, pp. 3344.

[10] (2011). Report on Wind Turbine Reliability ProfilesField DataReliability Analysis [Online]. Available: http://www.reliawind.eu/files/file-inline/110502_Reliawind Deliverable_D.1.3ReliabilityProfilesResults.pdf

[11] L. M. Moore and H. N. Post, Five years of operating experience at alarge, utility-scale photovoltaic generating plant, J. Progr. Photovolt.,Res. Appl., vol. 16, no. 3, pp. 249259, May 2008.

[12] W. Denson, The history of reliability prediction, IEEE Trans. Rel.,vol. 47, no. 3, pp. SP321SP328, Sep. 1998.

[13] M. G. Pecht and F. R. Nash, Predicting the reliability of electronicequipment, Proc. IEEE, vol. 82, no. 7, pp. 9921004, Jul. 1994.

[14] K. Chatterjee, M. Modarres, and J. B. Bernstein, Fifty years of physicsof failure, J. Rel. Inf. Anal. Center, vol. 20, no. 1, pp. 15, Jan. 2012.

[15] Military Handbook: Reliability Prediction of Electronic Equipment,Standard MIL-HDBK-217F, Dec. 1991.

[16] M. White and J. B. Bernstein, Microelectronics Reliability: Physics-of-Failure Based Modeling and Life Evaluation. Pasadena, CA, USA:Jet Propulsion Lab., 2008.

[17] J. G. Mecleish, Enhancing MIL-HDBK-217 reliability predictions withphysics of failure methods, in Proc. IEEE Annu. RAMS, Jan. 2010,pp. 16.

[18] Physics of Failure Reliability Predictions, VMEbus International TradeAssociation, Standard ANSI/VITA 51.2, 2011.

[19] M. Pecht and A. Dasgupta, Physics-of-failure: An approach to reliableproduct development, in Proc. Int. Integr. Rel. Workshop, Oct. 1995,pp. 14.

[20] M. Krasich, How to estimate and use MTTF/MTBF would the realMTBF please stand up? in Proc. IEEE Annu. RAMS, Jan. 2009,pp. 353359.

[21] W. B. Nelson, Recurrent-Events Data Analysis for Repairs, DiseaseEpisodes, and Other Applications (Asa-Siam Series on Statistics andApplied Probability). Philadelphia, PA, USA: SIAM, 2003.

[22] W. Weibull, Statistical distribution function of wide applicability,ASME J. Appl. Mech., vol. 18, no. 3, pp. 293297, Sep. 1951.

[23] S. Yang, A. T. Bryant, P. A. Mawby, D. Xiang, L. Ran, and P. Tavner,An industry-based survey of reliability in power electronic converters,IEEE Trans. Ind. Appl., vol. 47, no. 3, pp. 14411451, May/Jun. 2011.

[24] Y. Song and B. Wang, Survey on reliability of power electronicsystems, IEEE Trans. Power Electron., vol. 28, no. 1, pp. 591604,Jan. 2013.

[25] R. Burgos, C. Gang, F. Wang, D. Boroyevich, W. G. Odendaal, andJ. D. V. Wyk, Reliability-oriented design of three-phase power con-verters for aircraft applications, IEEE Trans. Aerosp. Electron. Syst.,vol. 48, no. 2, pp. 12491263, Apr. 2012.

[26] X. Perpi, X. Jord, M. Vellvehi, J. Rebollo, andM. Mermet-Guyennet, Long-term reliability of railway power inverterscooled by heat-pipe-based systems, IEEE Trans. Ind. Electron., vol. 58,no. 7, pp. 26622672, Jan. 2011.

[27] D. Hirschmann, D. Tissen, S. Schroder, and R. W. De Doncker,Reliability prediction for inverters in hybrid electrical vehicles, IEEETrans. Power Electron., vol. 22, no. 6, pp. 25112517, Nov. 2007.

[28] P. Wikstrom, L. A. Terens, and H. Kobi, Reliability, availability,and maintainability of high-power variable-speed drive systems, IEEETrans. Ind. Appl., vol. 36, no. 1, pp. 231241, Jan. 2000.

[29] F. Carastro, A. Castellazzi, J. Clare, and P. Wheeler, High-efficiencyhigh-reliability pulsed power converters for industrial processes, IEEETrans. Power Electron., vol. 27, no. 1, pp. 3745, Jan. 2012.

[30] M. Boettcher and F. W. Fuchs, Power electronic converters in windenergy systemsConsiderations of reliability and strategies for increas-ing availability, in Proc. 14th Eur. Conf. Power Electron. Appl.,Sep. 2011, pp. 110.


[31] S. S. Smater and A. D. Dominguez-Garcia, A framework for reliabilityand performance assessment of wind energy conversion systems, IEEETrans. Power Syst., vol. 26, no. 4, pp. 22352245, Nov. 2011.

[32] K. Fischer, F. Besnard, and L. Bertling, Reliability-centered main-tenance for wind turbines based on statistical analysis and practicalexperience, IEEE Trans. Energy Convers., vol. 27, no. 1, pp. 184195,Mar. 2012.

[33] H. Laukamp, Reliability study of grid connected PV systemsFieldexperience and recommended design practice, IEA Int. Agency, Paris,France, Tech. Rep. IEA-PVPS T7-08: 2002, Mar. 2002.

[34] D. Ton and W. Bower, Summary Report on the DOE High-Tech InverterWorkshop. Washington, DC, USA: U.S. Dept. Energy, Jan. 2005.

[35] Navigant Consulting Inc., A review of PVinverter technology cost and performance projection,National Renewable Energy Lab., Golden, CO, USA,Tech. Rep. NREL-SR-620-38771, Jan. 2006.

[36] S. Atcitty, J. E. Granata, M. A. Quintana, and C. A. Tasca, Utility-scalegrid-tied PV inverter reliability workshop summary report, Sandia Nat.Lab., Albuquerque, NM, USA, Tech. Rep. SAND2011-4778, Jul. 2011.

[37] E. L. Meyer and E. E. Van Dyk, Assessing the reliability and degrada-tion of photovoltaic module performance parameters, IEEE Trans. Rel.,vol. 53, no. 1, pp. 8392, Mar. 2004.

[38] C. Rodriguez and G. Amaratunga, Long-lifetime power inverter forphotovoltaic AC modules, IEEE Trans. Ind. Electron., vol. 55, no. 7,pp. 25932601, Jul. 2008.

[39] A. Ristow, M. Begovic, A. Pregelj, and A. Rohatgi, Development ofa methodology for improving photovoltaic inverter reliability, IEEETrans. Ind. Electron., vol. 55, no. 7, pp. 25812592, Jul. 2008.

[40] G. Petrone, G. Spagnuolo, R. Teodorescu, M. Veerachary, and M. Vitelli,Reliability issues in photovoltaic power processing systems, IEEETrans. Ind. Electron., vol. 55, no. 7, pp. 25692580, Jul. 2008.

[41] F. Chan and H. Calleja, Design strategy to optimize the reliability ofgrid-connected PV systems, IEEE Trans. Ind. Electron., vol. 56, no. 11,pp. 44654472, Nov. 2009.

[42] R. S. Balog, Y. Kuai, and G. Uhrhan, A photovoltaic module thermalmodel using observed insolation and meteorological data to support along life, highly reliable module-integrated inverter design by predictingexpected operating temperature, in Proc. IEEE ECCE, Sep. 2009,pp. 33433349.

[43] F. Chan and H. Calleja, Reliability estimation of three single-phasetopologies in grid-connected PV systems, IEEE Trans. Ind. Electron.,vol. 58, no. 7, pp. 26832689, Jul. 2011.

[44] S. V. Dhople, A. Davoudi, A. D. Dominguez-Garci, and P.L. Chapman, A unified approach to reliability assessment ofmultiphase DCDC converters in photovoltaic energy conversionsystems, IEEE Trans. Power Electron., vol. 27, no. 2, pp. 739751,Feb. 2012.

[45] P. Zhang, Y. Wang, W. Xiao, and W. Li, Reliability evaluation of grid-connected photovoltaic power systems, IEEE Trans. Sustain. Energy,vol. 3, no. 3, pp. 379389, Jul. 2012.

[46] S. Harb and R. S. Balog, Reliability of candidate photovoltaic module-integrated-inverter (PV-MII) topologiesA usage model approach,IEEE Trans. Power Electron., vol. 28, no. 6, pp. 30193027, Jun.2013.

[47] E. Koutroulis and F. Blaabjerg, Design optimization of transformer-lessgrid-connected PV inverters including reliability, IEEE Trans. PowerElectron., vol. 28, no. 1, pp. 325335, Jan. 2013.

[48] S. Saridakis, E. Koutroulis, and F. Blaabjerg, Optimal design of moderntransformerless PV inverter topologies, IEEE Trans. Energy Convers.,vol. 28, no. 2, pp. 394404, Jun. 2013.

[49] Y. Yang, H. Wang, F. Blaabjerg, and K. Ma, Mission profile basedmulti-disciplinary analysis of power modules in single-phase trans-formerless photovoltaic inverters, in Proc. 15th Eur. Conf. PowerElectron. Appl., Sep. 2013, pp. 110.

[50] N. Sintamarean, F. Blaabjerg, and H. Wang, Real field mission profileoriented design of a SiC-based PV-inverter application, in Proc. IEEEECCE, Sep. 2013, pp. 940947.

[51] M. Ciappa, Selected failure mechanisms of modern power modules,J. Microelectron. Rel., vol. 42, nos. 45, pp. 653667, Apr./May 2002.

[52] L. Yang, P. A. Agyakwa, and C. M. Johnson, Physics-of-failure lifetimeprediction models for wire bond interconnects in power electronicmodules, IEEE Trans. Device Mater. Rel., vol. 13, no. 1, pp. 917,Mar. 2013.

[53] E. Wolfgang, Examples for failures in power electronics systems,in Proc. Tuts. Rel. Power Electron. Syst., Apr. 2007, pp. 15.

[54] A. Wintrich, U. Nicolai, W. Tursky, and T. Reimann, Application ManualPower Semiconductors. Nuremberg, Germany, Semikron, 2011.

[55] J. W. McPherson, Reliability Physics and Engineering: Time-to-FailureModeling. New York, NY, USA: Springer-Verlag, 2010, ch. 8.

[56] L. F. Coffin, A study of the effects of cyclic thermal stresses on aductile metal, Trans. ASME, vol. 76, no. 6, pp. 931950, 1954.

[57] S. S. Manson, Behaviour of materials under conditions of thermalstress, Nat. Advisory Committee Aeronautics, Washington, DC, USA,Tech. Rep. 1170, 1954.

[58] C. Busca, R. Teodorescu, F. Blaabjerg, S. Munk-Nielsen, L. Helle,T. Abeyasekera, et al., An overview of the reliability predictionrelated aspects of high power IGBTs in wind power applications, J.Microelectron. Rel., vol. 51, nos. 911, pp. 19031907, 2011.

[59] U. Scheuermann, Pragmatic bond wire model, in Proc. ECPE Work-shop Lifetime Model. Simul., Jul. 2013, pp. 19.

[60] R. Wu, F. Blaabjerg, H. Wang, M. Liserre, and F. Iannuzzo,Catastrophic failure and fault-tolerant design of IGBT power electronicconvertersAn overview, in Proc. IEEE Ind. Electron. Soc. Annu.Conf., Nov. 2013, pp. 505511.

[61] H. Wang and F. Blaabjerg, Reliability of capacitors for DC-linkapplicationsAn overview, in Proc. IEEE Energy Convers. Congr.Exposit., Sep. 2013, pp. 18661873.

[62] C. Dubilier. (2013). Application GuideAluminum Elec-trolytic Capacitors [Online]. Available: http://www.cde.com/catalogs/AEappGUIDE.pdf

[63] H. Wang, Y. Yang, and F. Blaabjerg, Reliability-oriented design andanalysis of input capacitors in single-phase transformer-less PV invert-ers, in Proc. IEEE Appl. Power Electron. Conf. Exposit., Mar. 2013,pp. 29292933.

[64] P. OConnor and A. Kleyner, Practical Reliability Engineering, 5th ed.New York, NY, USA: Wiley, 2012, ch. 2.

[65] A. Mettas, Design for reliability: Overview of the process and applica-ble techniques, Int. J. Performabil. Eng., vol. 6, no. 6, pp. 577586,Nov. 2010.

[66] V. Crk. (2011). Reliability Engineering Challenges in AerospaceIndustry, Presentation [Online]. Available: http://ieee.rackoneup.net/rrvs/11/Reliability%20Engineering%20Challenges%20in%20Aerospace%20Industry.pdf

[67] P. V. Varde, Role of statistical VIS--VIS physics-of-failure failuremethods in reliability engineering, J. Rel. Statist. Stud., vol. 2, no. 1,pp. 4151, Jun. 2009.

[68] P. L. Hall and J. E. Strutt, Probabilistic physics-of-failure modelsfor component reliabilities using Monte Carlo simulation and Weibullanalysis: A parametric study, J. Rel. Eng. Syst. Safety, vol. 80, no. 3,pp. 233242, Jun. 2003.

[69] H. Janssen, Monte-Carlo based uncertainty analysis: Sampling effi-ciency and sampling convergence, J. Rel. Eng. Syst. Safety, vol. 109,no. 1, pp. 123132, Jan. 2013.

[70] How to Measure Lifetime for Robustness Validation-Step by Step, ZVEL,Elea, Greece, Nov. 2012.

[71] N. M. Vichare and M. G. Pecht, Prognostics and health managementof electronics, IEEE Trans. Compon. Packag. Technol., vol. 29, no. 1,pp. 222229, Mar. 2006.

[72] Y. Avenas, L. Dupont, and Z. Khatir, Temperature measure-ment of power semiconductor devices by thermo-sensitive electricalparametersA review, IEEE Trans. Power Electron., vol. 27, no. 6,pp. 30813092, Jun. 2012.

[73] E. Koutroulis and K. Kalaitzakis, Development of an integrated data-acquisition system for renewable energy sources systems monitoring,Renew. Energy, vol. 28, no. 1, pp. 139152, Jan. 2003.

[74] S. Yang, D. Xiang, A. Bryant, P. Mawby, L. Ran, and P. Tavner, Con-dition monitoring for device reliability in power electronic converters: Areview, IEEE Trans. Power Electron., vol. 25, no. 11, pp. 27342752,Nov. 2010.

[75] K. Ma, F. Blaabjerg, and M. Liserre, Thermal analysis of multilevelgrid side converters for 10 MW wind turbines under low voltageride through, IEEE Trans. Ind. Appl., vol. 49, no. 2, pp. 909921,Mar./Apr. 2013.

[76] D. Murdock, J. Torres, J. Connors, and R. Lorenz, Active thermalcontrol of power electronic modules, IEEE Trans. Ind. Appl., vol. 42,no. 2, pp. 552558, Mar./Apr. 2006.

[77] E. R. Motto and J. F. Donlon, IGBT module with user accessible on-chip current and temperature sensors, in Proc. 27th Annu. IEEE APECExposit., Feb. 2012, pp. 176181.

[78] M. Weckert and J. Roth-Stielow, Lifetime as a control variable inpower electronic systems, in Proc. Emobil. Electr. Power Train,Nov. 2010, pp. 16.

[79] L. Wei, J. McGuire, and R. A. Lukaszewski, Analysis of PWMfrequency control to improve the lifetime of PWM inverter, IEEETrans. Ind. Appl., vol. 47, no. 2, pp. 922929, Mar./Apr. 2011.


[80] M. N. Marwali, J. W. Jung, and A. Keyhani, Control of distributedgeneration systemsPart II: Load sharing control, IEEE Trans. PowerElectron., vol. 19, no. 6, pp. 15511561, Nov. 2004.

[81] K. Ma, M. Liserre, and F. Blaabjerg, Reactive power influence on thethermal cycling of multi-MW wind power inverter, IEEE Trans. Ind.Appl., vol. 49, no. 2, pp. 922930, Mar./Apr. 2013.

[82] B. Lu and S. Sharma, A literature review of IGBT fault diagnosticand protection methods for power inverters, IEEE Trans. Ind. Appl.,vol. 45, no. 5, pp. 17701777, Sep./Oct. 2009.

[83] B. A. Welchko, T. A. Lipo, T. M. Jahns, and S. E. Schulz, Fault tolerantthree-phase AC motor drive topologies: A comparison of features,cost, and limitations, IEEE Trans. Power Electron., vol. 19, no. 4,pp. 11081116, Jul. 2004.

[84] F. Richardeau and L. Pham, Reliability calculation of multilevelconvertersTheory and applications, IEEE Trans. Ind. Electron.,vol. 60, no. 10, pp. 42254233, Oct. 2013.

[85] S. Karimi, P. Poure, and S. Saadate, Fast power switch failure detectionfor fault tolerant voltage source inverters using FPGA, IET Trans.Power Electron., vol. 2, no. 4, pp. 346354, Jul. 2009.

[86] K. Kriegel, A. Melkonyan, M. Galek, and J. Rackles, Power modulewith solid state circuit breakers for fault-tolerant applications, in Proc.6th CIPS, Mar. 2010, pp. 16.

[87] E. Levi, Multiphase electric machines for variable-speed applications,IEEE Trans. Ind. Electron., vol. 55, no. 5, pp. 18931909, May 2008.

[88] R. Bojoi, M. G. Neacsu, and A. Tenconi, Analysis and survey ofmulti-phase power electronic converter topologies for the more electricaircraft applications, in Proc. Int. Symp. Power Electron., Electr.Drives, Autom. Motion, Jun. 2012, pp. 440445.

[89] R. De Maglie, G. Osvald, J. Engstler, A. Engler, and J. P. Carayon, Opti-mized 70 kW power inverter dedicated to future aircraft applications,in Proc. 13th Eur. Conf. Power Electron. Appl., Sep. 2009, pp. 110.

[90] V. Choudhary, E. Ledezma, R. Ayyanar, and R. M. Button, Faulttolerant circuit topology and control method for input-series andoutput-parallel modular DCDC converters, IEEE Trans. PowerElectron., vol. 23, no. 1, pp. 402411, Jan. 2008.

Huai Wang (S07M12) received the B.Eng.degree in electrical and electronic engineering fromthe Huazhong University of Science and Technology,Wuhan, China, in 2007, and the Ph.D. degree inElectronic Engineering from the City University ofHong Kong, Kowloon, Hong Kong, in 2012.

He has been with Aalborg University, Aalborg,Denmark, since 2012, where he is currently anAssistant Professor with the Department of EnergyTechnology. He was a Visiting Scientist with theMassachusetts Institute of Technology, Cambridge,

MA, USA, in 2013. He was with the ABB Corporate Research Center, Baden,Switzerland, in 2009. He has contributed over 40 journal and conferencepapers and filed three patents. His current research interests include thereliability of DC-link capacitors, reliability of power electronic systems, high-voltage DC-DC power converters, time-domain control of converters, andpassive components reduction technologies.

Dr. Wang is a recipient of the five paper awards and project awards fromindustry, IEEE, and the Hong Kong Institution of Engineers. He serves theGuest Associate Editor of the IEEE TRANSACTIONS ON POWER ELECTRON-ICS Special Issue on Robust Design and Reliability in Power Electronics, anda Session Chair of various conferences in power electronics.

Marco Liserre (S00M02SM07F13) receivedthe M.Sc. and Ph.D. degrees in electrical engineeringfrom the Bari Polytechnic, Bari, Italy, in 1998 and2002, respectively.

He has been an Associate Professor with the BariPolytechnic and a Professor of reliable power elec-tronics with Aalborg University, Aalborg, Denmark.He is currently a Full Professor and Chair of powerelectronics with Christian-Albrechts-University ofKiel, Kiel, Germany. He has published 168 technicalpapers (44 of them in international peer-reviewed

journals), three chapters of a book, and a book titled Grid Converters forPhotovoltaic and Wind Power Systems (Wiley), which is also translated inChinese. He has more than 6000 citations. He has been a Visiting Professorwith Alcala de Henares University, Madrid, Spain.

Dr. Liserre is a member of IAS, PELS, PES, and IES. He is an AssociateEditor of the IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, theIEEE INDUSTRIAL ELECTRONICS MAGAZINE, and the IEEE TRANSAC-TIONS ON INDUSTRIAL INFORMATICS. He is currently a Co-Editor-in-Chiefof the IEEE TRANSACTIONS ON POWER ELECTRONICS and the IEEEJOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS.He has been a Founder and an Editor-in-Chief of the IEEE INDUSTRIALELECTRONICS MAGAZINE, a Founder and the Chairman of the TechnicalCommittee on Renewable Energy Systems, a Co-Chairman of the Interna-tional Symposium on Industrial Electronics in 2010, a IES Vice-Presidentresponsible of the publications. He has received the IES 2009 Early CareerAward, the IES 2011 Anthony J. Hornfeck Service Award, the 2011 IndustrialElectronics Magazine Best Paper Award, and the Third Prize Paper Award bythe Industrial Power Converter Committee at ECCE in 2012. He is SeniorMember of the IES AdCom. He has been elevated to the IEEE fellow gradewith the following citation for contributions to grid connection of renewableenergy systems and industrial drives.

Frede Blaabjerg (S86M88SM97F03) waswith ABB-Scandia, Randers, Denmark, from 1987to 1988. From 1988 to 1992, he was a Ph.D. Studentwith Aalborg University, Aalborg, Denmark. Hebecame an Assistant Professor in 1992, AssociateProfessor in 1996, and a Full Professor of powerelectronics and drives in 1998. He has been a parttime Research Leader of wind turbines with theResearch Center Risoe. From 2006 to 2010, he wasthe Dean of the Faculty of Engineering, Science,and Medicine and became a Visiting Professor with

Zhejiang University, Hangzhou, China, in 2009. His current research interestsinclude power electronics and its applications such as in wind turbines, PVsystems, reliability, harmonics, and adjustable speed drives.

Dr. Blaabjerg received the 1995 Angelos Award for his contribution inmodulation technique and the Annual Teacher Prize at Aalborg University. In1998, he received the Outstanding Young Power Electronics Engineer Awardby the IEEE Power Electronics Society. He has received 15 IEEE Prize PaperAwards and the Prize Paper Award at PELINCEC Poland in 2005. He receivedthe IEEE PELS Distinguished Service Award in 2009, the EPE-PEMC CouncilAward in 2010, and the IEEE William E. Newell Power Electronics Award2014. He has received a number of major research awards in Denmark. He wasan Editor-in-Chief of the IEEE TRANSACTIONS ON POWER ELECTRONICSfrom 2006 to 2012. He was a Distinguished Lecturer for the IEEE PowerElectronics Society from 2005 to 2007 and the IEEE Industry ApplicationsSociety from 2010 to 2011. He was a Chairman of EPE in 2007 and PEDG,Aalborg, in 2012.


Peter de Place Rimmen has been a Reliabil-ity Advisor with Danfoss Power Electronics A/S,Graasten, Denmark, since 2009. He was involvedin research on practical approach implementing reli-ability in such companies: Vestas Wind SystemResearch and Development, Aarhus, Denmark, from2004 to 2009, Grundfos Management Research andDevelopment from 1997 to 2004, and Bang &Olufsen Research and Development, Struer, Den-mark, from 1988 to 1997. He was with B&O asa Constructor, Test Engineer, Plant manager, and

Project manager. He has participated in IEC dependability group. He hastogether with Nokia trained Nokia Research and Development and VestasResearch and Development in Design for Quality and Reliability. He isparticipating in the CORPE Centre of Reliable Power Electronics, AalborgUniversity, participating in ZVEI Facts Sheets Group for Robustness Vali-dation, ECPE Course Instructor, Board Member FAST (Danish Society forApplied Statistics) and initiated in 2001, and a member of the Danish SixSigma ERFA-Group, subgroup of FAST. He holds more than one patent forVestas concern Lifetime improvement by thermal control improvements andfor Danfoss, he holds two patents on Dehumidifier for enclosures, and onepatent for monitoring device usage for stress.

John B. Jacobsen was born in Hrning, Denmark,in 1960. He received the B.Sc. degree in elec-trical engineering from Aarhus Teknikum, Aarhus,in 1985. Main work experience is from Grundfos,Denmark; seven years in hybrid technology devel-opment, seven years in hybrid production and morethan ten years as a Chief Specialist in integrationof power electronics into products, i.e. mechatronicdisciplines including thermal management, inter-connection, fixation, protection from environment(impact and humidity), reliability and cost. Opti-

mization criterion being needed performance and reliability at lowest possiblecost. Generalist in understanding value chain and all the disciplines that meetin the physical mechatronic reality, i.e. function trade-offs in design andproduce ability in production.

Thorkild Kvisgaard was born in Fabjerg, Denmark,in 1958. He received the B.Sc. degree in electri-cal engineering from Aarhus Teknikum, Aarhus, in1984, and the e-M.B.A. degree in innovation andtechnology management from Aalborg University,Aalborg, Denmark, in 2006. Main work experienceis from Scanvaegt International where he was aManager of Electronics Development creating solu-tions for the food industry as well as electronics foroffshore applications.

In 1994, he joined Grundfos and served as aProduct Development Manager in 11 years. After this Thorkild Kvisgaardchanged position in Grundfos to Global Technology Manager. In addition, hebecame a member of the Board in Center for Electrical Energy Systems in2006 and he act as a Chairman of the Center for Intelligent and Efficient PowerElectronics. In the Center Of Reliable Power Electronics Thorkild Kvisgaardhas the role as a Vice Chairman. He has several patents granted and wasnominated for the best Danish patent granted in 1992.

Jrn Landkildehus was born in Ebeltoft, Denmark,in 1969. He received the M.Sc.E.E. from AalborgUniversity, Aalborg, Denmark, in 1995.

Since 1995, he has been with Danfoss Power Elec-tronics, Graasten, Denmark, engaged with develop-ment of variable speed drives and research in powertopologies. He has been specializing in the designfor EMC and in recent years been leading reliabilityengineering department at Danfoss Power Electron-ics. His current research interests include develop-ment processes, multidisciplinary design techniques,

and design for reliability.

/ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 600 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 400 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False

/Description >>> setdistillerparams> setpagedevice

Transitioning to Physics-of-Failure as a Reliability Driver in Power Electronics

Documents

Transcript of Transitioning to Physics-of-Failure as a Reliability Driver in Power Electronics