Fuzzy Logic Applications to Power Electronics and Drives zyxw - An Overview zy Gilbert0 C. D. Sousa Departamento de Engenharia Eletrica Universidade Federal do Espirito Santo c. P. 01-9011 Vitoria, ES 29060-970 - Brazil Abstract - Applications of fuzzy logic (FL) to power electronics and drives are on the rise. The paper discusses some representative applications of FL in the area, preceded by an interpretative review of fuzzy logic controller (FLC) theory. A discussion on design and implementation aspects is presented, that also considers the interaction of neural networks and fuzzy logic techniques. Finally, strengths and limitations of FLC are considered, including possible applications in the area. I - INTRODUCTION Fuzzy logic (FL), the logic of approximate reasoning, continues to grow in importance, as its application to a number of practical problems further demonstrates its usefulness. In fact, fuzzy logic has been used in areas such as process control, estimation, identification, diagnostics, agriculture, medicine, stock market, etc. However, process control is by far its most important and visible application. Lofti Zadeh, father of fuzzy logic, has classified computing as hard computing and soft computing. The computations based on Boolean algebra and other crispy numerical computations are defined as hard computing, whereas fuzzy logic, neural network and probabilistic reasoning techniques, such as genetic algorithm, chaos theory, belief networks and parts of learning theory are categorized as soft computing. Soft computing differs from conventional (hard) computing in that, unlike hard computing, it is tolerant of imprecision, uncertainty and partial truth. Soft computing is more analogous to thinking of human mind. Fuzzy Logic applications to the control of power electronics and drive (PE&D) systems have been increasing exponentially in the past few years. Power converter and drive systems possess inherent characteristics, such as non-linearities, unavailability of a precise model or its excessive complexity, that make them well suited for FL control. The fuzzy logic controller (FLC) of a given process is capable of embedding, in the control strategy, the qualitative knowledge and experience of an operator or field engineer about the process. Therefore, fuzzy logic plays the role of a suitable ”user interface”, in the task of translating designer’s insight about the system into the control law, resulting in an inherently nonlinear adaptive controller, capable of Bimal K. Bose Electrical and Computer Engineering Knoxville, TN 37996 - USA The University of Tennessee zyx 0-7803-3026-9/95 $4.00 zyxwvutsrq 0 1995 IEEE 57 outperforming other control techniques, such and MRAC and sliding mode controllers. In spite of its practical success, fuzzy logic has been criticized for its limitations, such as the lack of a formal design methodology and the difficulty in predicting stability and robustness of FL controlled systems. In reality, these aspects have evolved considerably in the past few years, as the heuristic approaches, commonly used in FLC design have gradually been replaced by more formal methods. In the PE&D field, FL has been applied to various problems, such as robust feedback control of a phase-controlled converter dc drive system [ 11, efficiency optimization of drive systems [3], hierarchical control of circulating-current converters [5], slip gain tuning of vector-controlled induction motor drive system [6], servo drives [7]-[ lo], induction machine direct self control drives [ 111, estimation of power electronic waveforms [ 121, wind generation systems [4], ultrasonic motor control [9], etc. This paper presents an overview of some relevant work in the area, preceded by an insightful review of FLCs. Salient design and implementation aspects of FLCs are also considered, along with an analysis of FLC merits and Iimitations. 11-FUZZY LOGIC CONTROLLERS REVISITED Several interpretations of fuzzy logic systems have been proposed. One can view them as special type expert systems (ES), where the symbolic processing is not as rigorous as in the conventional ES. On the other extreme, they might be seen as pure numeric input/output mapping, quite similar to that of an artificial neural network (ANN). In most applications, however, both aspects are simultaneously present, and this ability of providing a mathematical interpretation from a qualitative description is its major strength. Some of these aspects will be reviewed here. zyx A. zyxwvutsr Fuzzy logic controllers as approximators In many situations FL constitutes an efficient way to represent a non-analytical mapping, such that an output value y can be infcrrcd from a given input x. In order to construct the fuzzy model, the input/output pattern must be known, as illustrated in Fig. 1.