Alternative configurations for induction-generator based geared wind turbine systems for reliability and availability improvement

Transcription

1 Proceedings of the 14th International Middle East Power Systems Conference (MEPCON 10), Cairo University, Egypt, December 19-21, 2010, Paper ID 228. Alternative configurations for induction-generator based geared wind turbine systems for reliability and availability improvement M. EL-Shimy Electric Power and Machines Department, Faculty of Engineering, Ain Shams University, 11517, Cairo, Egypt Mobile phone: Abstract - The main objectives of this paper are to study and improve the reliability and the structural availability of WTG systems. Due to limitations on the availability and accuracy of failure and repair data, the scope of the study is limited to the main items comprising the electrical subsystems of the induction generator based WTG systems. However, induction generator based WTG systems are the most widely used systems in wind power generation and the mechanical subsystems of such generating systems are almost identical. Previous studies show that the electronic subassemblies in WTG systems are among the main causes of the reduction of the overall system availability. Hence, the proposed alternative configurations are based on either redundancy in the converter subassemblies (active and standby redundancy) or converter bypass during converter failure and repair times. Operational limitations of the proposed configurations as well as some previously proposed configurations are discussed. Suitability of the proposed configurations for offshore applications is considered. It is found that the squirrel-cage induction generator (SCIG) with full-scale converter (FSC) and active-redundant converter configuration is the optimal WTG system for offshore applications. However, attempts should be made to improve the maintainability of such a configuration. Index Terms -Wind power; Reliability; DFIG; SCIG; BDFIG I. INTRODUCTION Referring to the rotational speed, wind turbine (WT) concepts can be classified into fixed speed, limited variable speed and variable speed. For variable speed wind turbines, based on the rating of the power converter related to the generator capacity, they can be further classified into wind generator systems with a partial-scale and a full-scale power electronic converter. In addition, considering the drive train components, the wind turbine concepts can be classified into geared-drive and direct-drive wind turbines. In geared-drive wind turbines, one conventional configuration is a multiplestage gear with a high-speed generator; the other one is the multibrid concept that has a single-stage gear and a low-speed generator. Extended details about wind turbine concepts and their comparison can be found in [1-5]. The multiple-stage geared drive DFIG concept is still dominant in the current market. Additionally, the market shows interest in the direct-drive or geared-drive concepts with a full-scale power electronic converter. Current developments of wind turbine concepts are mostly related to offshore wind energy; variable speed concepts with power electronics will continue to dominate and be very promising technologies for large wind farms [1]. Geared wind turbine systems with induction generators have been shown to be the most common configurations (more than 55%) used for large wind turbines [2] where DFIG based system is the more common configuration among them [3]. Compared with the DFIG system, the Brush-less Doubly- Fed Induction Generator (BDFIG) does not require slip rings; however, it requires double stator windings, with a different number of poles in both stator layers. The second stator layer generally has lower copper mass, because only a part of the generator nominal current flows in the second winding. This second stator winding is connected through a power electronic converter, which is rated at only a fraction of the wind turbine rating [1]. One of the main reasons for lower reliability of the system with the DFIG, in comparison to SCIG based fixed speed systems, is the presence of brushes in the configuration. With the advent of BDFIG technology, this drawback could be overcome in future years [2]. Test results from prototype of BDFIG indicate that it is a valid alternative to the DFIG for future wind turbines; however, the machine operation principle and its assembly are relatively complex [1, 2]. To understand WT reliability, we need to break down the WT system into subsystems and in turn, subsystems are divided into subassemblies [2-6]. A subsystem of WT system could for example be the drive train, consisting of rotor hub, shaft, bearing, gearbox, couplings, and generator. Components that constitute a subsystem are subassemblies such as the gearbox. Fig. 1 illustrates a typical configuration and main components of horizontal axis geared wind turbine system. It is depicted from Fig. 1 that a wind turbine system consists of several components. A component can be considered as a subsystem if it is divided into its constituting items. For example, the converter of the DFIG system can be considered as a subsystem consisting of four subassemblies namely, the rotor side converter (RSC), the grid side converter (GSC), the DC link, and the control unit (CU) [2]. Reliability is the probability of a subassembly to perform its purpose adequately, under the operating conditions encountered, for the intended period. Analytical methods are available for evaluating reliability, depending on the data available, the depth of study, and the expected accuracy of the model [8, 9]. A reliability model can only provide correct conclusions if accurate data are used [2]. Operational data will verify correctness of the predicted system lifetime. Statistical data analysis may result in a component redesign or a changed maintenance schedule [5]. The control unit inside the turbine regularly collects operational statistics from wind power plants. Today, most turbines are fitted with equipment that makes it possible to collect the data remotely via modems or internet [5]. The basis for developing and establishing a database for collecting reliability and reliability-related data, for assessing the 538

2 reliability of wind turbine components and subsystems and wind turbines as a whole, as well as for assessing wind turbine availability while ranking the contributions at both the component and system levels is presented in [4, 9, 10]. Fig. 1: A typical configuration and main components of horizontal axis geared wind turbine system [4] Collecting accurate wind-turbine reliability data is considered a challenging task [2, 5]. This was for several reasons, e.g.,: no statistical data were collected, wind turbine manufacturers refused to reveal data, data from different designs could not be compared, or data retrieval was too expensive to access [5]. Even if it is available, the field failure data are usually tainted, incomplete, lack sufficient detail, or do not satisfy the assumptions of a model selected for analysis [7]. In order to consider such an incompletion and obtain a more accurate reliability growth of wind turbines, a general three-parameter Weibull failure rate function is presented in reference [7] to depict the reliability growth. The parameters of this function are estimated by two techniques, maximum likelihood and least squares. Similar results have been achieved by the two techniques. Despite the deficiencies of this data, reliability-growth analysis methods allow the extraction of reliability trends over an observed period [11]. The analysis can also differentiate between subassemblies in a system subject to human-driven reliability improvement and mature technology, and subassemblies that are deteriorating, and characterized by increasing failure intensity. The main literature findings from the investigations of the failure statistics of WT systems indicate the following [2-6]: The gearbox is critical to the availability of the wind turbine. Most of the gearbox failures are caused by wear on the mechanical parts. Direct drive WT systems are not necessarily more reliable than geared WT systems. Aggregate failure intensities of generators and converters in direct drive WT systems are greater than the aggregate failure rate of gearboxes, generators, and converters in geared WT systems. The gears and the drive train are the components that demand the longest downtime per failure. Since drive train and gearboxes seldom fail, one reason for the long downtime could be that spare parts need to be ordered, which could prolong the time for repair. Power electronic converters of direct and geared drive WT system exhibit higher failure intensities throughout their operation than converters in other industries. Although the fixed-speed wind turbine is less aerodynamically efficient, its availability is higher, when its reliability is taken into account, at least in its electrical subassemblies. If the wind power is to be competitive, the downtime needs to be shortened and visits to the turbine should be kept to a minimum [5]. This can be achieved through improvement in WT system design, fault detection and monitoring, and maintenance procedures. Better reliability of small wind turbines could be achieved with grid-connected configurations that require minimal power electronics [12-13]. The main objectives of this paper are to study and improve the reliability and the structural availability of WTG systems. Due to limitations on the availability and accuracy of failure and repair data, the scope of the study is limited to the main items comprising the electrical subsystems of the induction generator based WTG systems. However, induction generator based WTG systems are the most widely used systems in wind power generation and the mechanical subsystems of such generating systems are almost identical. Previous studies show that the electronic subassemblies in WTG systems are among the main causes of the reduction of the overall system availability. Hence, the proposed alternative configurations are based on either redundancy in the converter subassemblies (active and standby redundancy) or converter bypass during converter failure and repair times. Operational limitations of the proposed configurations as well as some previously proposed configurations are discussed. Suitability of the proposed configurations for offshore applications is considered. II. WTG SUBASSEMBLIES AND RELIABILITY MODELLING Three configurations are considered in this study, all of them follow the variable speed WTG concept as shown in Fig. 2. The first configuration, shown in Fig. 2(a), is based on DFIG with a partial-scale converter. The second configuration, shown in Fig. 2(b), is based on BDFIG with a partial-scale converter. The third configuration, shown in Fig. 2(c), is based on SCIG with a full-scale power converter. The considered electrical subassemblies for each configuration, which are the generator s subassemblies, and the converter s subassemblies (the rotor or machine side converter (RSC or MSC), the grid-side converter (GSC), the DC link, and the control unit (CU)) are shown in Fig. 2. Failure and repair data for each of 539

3 the considered subassemblies in each configuration are based on a recent survey in the Manjil wind farm in Iran and are obtained from [2]. For standby redundant equal items system, shown in Fig. 4(c), (4) (5) (6) Fig. 2: Configurations and subassemblies of the considered WTGs For its simplicity and suitability to the considered problem, reliability block diagram (RBD) modeling technique is used to model the considered configurations. From reliability point of view, the considered subassemblies of each of the base configurations shown in Fig. 2 are connected in series. Fig. 3 shows the RBD for the generator and converter subsystems. Fig. 4: Basic RBD connections. (a) Series, (b) Active redundancy, (c) Standby redundancy The failure and repair rates are the reciprocal of the meantime between failures (MTBF or m) and the mean-time to repair (MTTR or r). The reliability and maintainability are usually demonstrated by the values of failure and repair rates respectively. The availability is calculated by (7) Fig. 3: RBD for various subsystems Based on the reliability theory [8, 14] the following formulae apply under steady state analysis regardless of the distributions of failure and repair except for the case of standby redundancy where each block must have exponentially distributed active failure and repair times and passive and switching failure rates assumed to be zero. Fig. 4 shows the basic RBD connections. The failure ( ) and repair ( ) rates for basic RBD configurations are calculated as follows. For series connected items, shown in Fig. 4(a), For active redundant items, shown in Fig. 4(b), (1) (2) (3) III. ANALYSIS OF WTG RELIABILITY DATA Based on the failure and repair data [2], that are plotted in Fig. 5, for configuration (a) of Fig. 2, it is depicted that the subassemblies characterized by high failure rates (low reliability) as in comparison to the rest of the considered subassemblies, in descending order, are the RSC, the GSC, and the brush gear. From availability point of view, it is depicted from Fig. 5(c) that both the RSC and the GSC are characterized by lower availability in comparison with the rest of the subassemblies. The high maintainability characteristic of the brush gear excluded it from being characterized by low availability. Configurations (b) and (c) of Fig. 2 do not include brush gears and the characteristics of the subassemblies of configuration (a) of Fig. 2 are applied to the subassemblies of these configurations. Higher failure rate of the stator of the BDFIG with respect to stators in other configurations is assumed because of its double stator winding design. The stator of the BDFIG is assumed, from reliability point of view, to have failure and repair rates of two series connected stators of the DFIG type. 540

4 power system. Therefore, improved alternative configurations that are not requiring transition to fixed-speed operation are favorable. Despite these difficulties, the converter bypass option is considered herein assuming that a negligible possible trip and actuation-transition times, in comparison with the converter downtime, are required to allow successful transition to fixed-speed operation of the variable-speed WTG system. Moreover, it is assumed that the bypass system is 100% reliable as in [2]. Fig. 5: Subassemblies characteristics of the DFIG with a partial-scale converter configuration. (a) Failure rate (b) Repair rate (c) Availability IV. ALTERNATIVE CONFIGURATION AND RESULTS Configurations shown in Fig. 2 are referred to as the base configurations. Alternative configurations are modified versions of the base configurations. The considered modifications are based on either redundancy in the converter subassemblies (active and standby redundancy) or converter bypass during converter failure and repair times. The converter bypass alternative configuration presented in [2] allows the WTG system in the case of converter failure to continue running and deliver power to the grid in a temporarily fixed-speed operation mode instead of the default variable-speed operation. However, the instantaneous bypass of the converter may be practically impossible. From transient response point of view, the converter bypass should not be done simultaneously with the converter failure because of the unpredictable response of the DFIG following blocking of the RSC; this phenomena along with the factor affecting successful restarting of the converter is fully covered in [15]. A bypass logic-control mechanism that considers the operating point (the generator speed and powers) of the WTG system at the instant of failure of the converter is required to reduce the impact of the bypass on both the WTG and the A. DFIG based WTG system Five alternative configurations are considered for the DFIG based WTG systems shown in Fig. 2(a), these alternative configurations are listed in Table 1. It is depicted from Table 1 that all alternatives exhibits higher availability than the base-case configuration. The alternative configuration (6-b), with a converter bypass system and delta connected rotor winding, is characterized by the lowest failure rate (highest reliability) and the highest availability making it the optimal configuration for offshore installations where minimum site visits are required. However, such a configuration is not favorable because of the switching and transition risks that are previously mentioned. Therefore, attempts should be made to reduce or eliminate such risks, for example, through an appropriate trip time to allow successful transition from viable-speed to fixed speed operation. Among variable-speed alternative configurations, the active-redundant configuration is characterized by the lowest failure rate and highest availability. This may be the best choice for offshore applications. However, either the active- or standby- redundant RSC configurations may be economically suitable for land-based installations. Although, the standbyredundant RSC configuration exhibits higher repair rate (maintainability), the reliability and availability of the activeredundant RSC configuration are much better. The highest repair rate (maintainability) is obtained with the standby-redundant converter configuration. Both standbyredundant converter and RSC configurations have the same reliability. However, the standby-redundant converter configuration exhibits higher availability and maintainability. B. BDFIG based WTG system Compared with the DFIG system, the BDFIG system does not require slip rings; however, it requires double stator windings, with a different number of poles in both stator layers [1]. Therefore, it assumed from reliability point of view that the stator of the BDFIG is consisted of two series connected stators each having the same failure and repair rates as the stator winding of the DFIG. Five alternative configurations are considered for the BDFIG based WTG systems shown in Fig. 2(b), these alternative configurations are listed in Table 1. Apart from the converter-bypass alternative which has lowest failure rate and availability among all alternatives, the failure rates of all configurations of the BDFIG-based WTG are lower (higher reliability) than that for the DFIG-based WTG. This is because 541

5 of the absence of the brush gear subassembly in the BDFIG machine. However, inspecting Table 1 shows that the availability of both systems is comparable. The comments about the characteristics and the applications of the proposed configurations of the BDFIG-based WTG system are similar to those for the DFIG-based WTG configurations. C. SCIG with a Full-Scale Converter (FSC) based WTG system Unlike the partial-scale converter based configurations, where the RSC play the major role that facilitate variable speed operation, the MSC and GSC are equally important in configurations with full-scale converter. In addition, the failure rate of the RSC in DFIG based configuration is twice that for the GSC. However, both the MSC and the GSC in SCIG with a full-scale converter configuration are of equal failure rates [2]. Therefore, Configurations with redundant MSC are not considered. Three alternative configurations are considered for the SCIG with a full-scale converter based WTG systems shown in Fig. 2(c), these alternative configurations are listed in Table 1. The following are comments about the results shown in Table 1. Table 1: Alternative configurations for the DFIG and BDFIG based WTG systems DFIG based system BDFIG based system SCIG and full-scale conv. s.n Configuration yr yr A yr yr A yr yr A Variable-speed alternative configurations 1 Base Case Active-Redundant Converter Active-Redundant RSC Standby-Redundant Converter Standby-Redundant RSC Fixed-speed alternative configurations 6 Conv. bypass 6-a b Y-connected rotor winding 2. -connected rotor winding Although the base-case configuration of the SCIG with FSC system is characterized by highest failure rate and lowest availability in comparison with the base case of all other systems, it is characterized by lowest failure rate and equal availability when there is an active redundancy in the converter subsystem relative to similar converter arrangement in the other configurations. This suggests that the SCIG with FSC and active-redundant converter configuration is the optimal WTG system for offshore applications. However, attempts should be made to improve its maintainability. V. CONCLUSIONS This paper presents study, analysis, and improvement of the reliability and availability of the most widely used systems in wind power generation, which are the induction generator based WTG systems. The availability and accuracy of failure and repair data of the subassemblies of WTG systems limit the study to the electrical subsystems. However, the main outcomes are independent on this limitation because the mechanical subsystems of the considered WTG configurations are almost identical. Due to their major effect on the availability of WTG systems, alternative configurations for the converter subassemblies are proposed in order to improve the system s reliability and availability. The effect of the alteration on the systems' configurations from the points of view of maintainability and operational limitations are considered. It is clarified that the converter-bypass technique results on highest reliability and availability among the considered alternatives; however, the probable instability of the WTG system due to sudden blocking of the converter subsystem hinder the practical implementation of such technique, unless an appropriate trip time is considered to allow successful transition from viable-speed to fixed speed operation. By implementing proper switching logic, the converter bypass alternative may be the optimal choice of offshore applications. However, other configurations based on redundancy of the converter subassemblies show comparable reliability and availability levels without hindering the variable-speed operation. Several alternative configurations are demonstrated along with numerical demonstration of their reliability, availability, and maintainability. It is found that the SCIG with FSC and active-redundant converter configuration is the optimal WTG system for offshore applications. However, attempts should be made to improve the maintainability of such a configuration. REFERENCES [1] H. Li, and Z. Chen, Overview of different wind generator systems and their comparisons, IET Renew. Power Gener., vol. 2, no. 2, pp , [2] H. Arabian-Hoseynabadi, H. Oraee, and P.J. Tavner, Wind turbine productivity considering electrical subassembly reliability, Renewable Energy, vol. 35, pp , [3] F. Spinato, P.J. Tavner, G.J.W Bussel, E. Koutoulakos, Reliability of wind turbine subassemblies, IET Renew. Power Gener., vol. 3, no. 4, pp , [4] C.C. Ciang, J.R Lee Jung-Ryul, and H.J Bang, Structural health 542

6 monitoring for a wind turbine system: a review of damage detection methods, doi: / /19/12/122001, Meas. Sci. Technol., vol.19, pp. 1-20, [5] J. Ribrant, and L.M. Bertling, Survey of Failures in Wind Power Systems With Focus on Swedish Wind Power Plants During , IEEE Trans. on Energy Conversion, vol. 22, no. 1, pp , [6] P.J Tavner, F. Spinato, G.J.W Bussel, and E. Koutoulakos, Reliability of different wind turbine concepts with relevance to offshore application. European Wind Energy Conf., EWEC2008, Brussels, Belgium, April [7] H. Guo, S. Watson, P. Tavner, and J.Xiang, Reliability analysis for wind turbines with incomplete failure data collected from after the date of initial installation, Reliability Engineering and System Safety, vol. 94, pp , [8] R. Billinton, and R.N Allan, Reliability evaluation of engineering systems: concepts and techniques, 2nd ed., Plenum Press, ISBN , [9] I. Kozine, P. Christensen, and M. Winther-Jensen, Failure Database and Tools for Wind Turbine Availability and Reliability Analyses, Risø National Laboratory, Roskilde, ISBN , January [10] W. Yang, P.J Tavner, and M. Wilkinson, Wind Turbine Condition Monitoring and Fault Diagnosis Using both Mechanical and Electrical Signatures, Proceedings of the 2008 IEEE/ASME International Conference on Advanced Intelligent Mechatronics, Xi'an, China, July 2-5, 2008, pp [11] F. Spinato, and P. Tavner, Reliability-Growth Analysis of Wind Turbines from Fleet Field Data, ARTS Conf., Loughborough, April [12] Md. Arifujjama, M.T Iqbal, and J.E Quaicoe, Reliability analysis of grid connected small wind turbine power electronics, Applied Energy, vol. 86, pp , [13] Md. Arifujjaman, M.T Iqbal, and J.E Quaicoe, A comparative study of the reliability of the power electronics in grid connected small wind turbine systems, 2009 IEEE Conf., pp , [14] Applied R&M Manual, for Defence Systems (GR-77 Issue 2009), available at accessed Sept [15] M. Kayikçi, and J.V. Milanovic, Assessing Transient Response of DFIG-Based Wind Plants The Influence of Model Simplifications and Parameters, IEEE Trans. On Power Systems, vol. 23, no. 2, pp , M. EL-Shimy was born in Cairo in the Arab Republic of Egypt. He completed his Electrical Engineering B.Sc, M.Sc, and PhD degrees from Faculty of Engineering Ain Shams University, Egypt, in 1997, 2001, and 2004 respectively. He is now an associate professor in Department of Electrical Power and Machines -Faculty of Engineering Ain Shams University. He is a consultant and trainer and a member of many renewable energy associations. He teaches several undergraduates, graduate, and training courses in Egypt Universities and outside. His fields of interest include power system stability, power system equivalents, load aggregation, load signature, electric power distribution, optimal power flow studies, flexible ac transmission systems (FACTS), power system optimization, new energy resources, and power system reliability. For more details, please visit: 543