Expert system, fuzzy logic, and neural network applications in power electronics and motion control

Abstract

Artificial intelligence (AI) tools, such as expert system, fuzzy logic, and neural network are expected to usher a new era in power electronics and motion control in the coming decades. Although these technologies have advanced significantly in recent years and have found wide applications, they have hardly touched the power electronics and mackine drives area. The paper describes theseAi tools and their application in the area of power electronics and motion control. The body of the paper is subdivided into three sections which describe, respectively, the principles and applications of expert system, fuzzy logic, and neural network. The theoretical portion of each topic is of direct relevance to the application of power electronics. The example applications in the paper are taken from the published literature. Hopefully, the readers will be able to formulate new applications from these examples.

Adaptive Fuzzy Logic Control for High Performance PM Synchronous Drives

“…T he hysteresis property allows the actual value of i a to exceed or to be less than the reference value by D i; the logic is shown in Table 2. Similar logic is applied to the other two phases [2,3].…”

“…T he degrees of memberships are then determined for each crisp value in relation with all fuzzy sets [3]. T he linguistic control rules are developed based on human's experience and knowledge about the system that has to be controlled.…”

A situation risk awareness approach for process systems safety

“…This property allows fuzzy logic to deal with uncertain situations in a fairly natural way. It may be mentioned that although fuzzy logic deals with imprecise information, it is based on sound quantitative mathematical theory (Bose, 1994). Fuzzy theory was first proposed by Zadeh in 1965.…”

ANN Based Flux and Torque Estimator for Vector Controlled Induction Drive with Dynamic Loading

“…Most control techniques of squirrel cage Induction Motor (IM) require speed feedback signal from the shaft encoder and these devices have various disadvantage and are undesirable in many applications [17], [18]. In order to control an IM requiring high dynamic performance, an accurate knowledge of the load torque is necessary.…”

“…The instantaneous magnitude and position of rotor flux and load torque estimation using single ANN structure has been tried out in most of the previous works [17], [18]. This paper proposes two ANN structure, one ANN structure estimates stator d-axis and q-axis flux using (5) and (6), whereas second ANN structure estimates the instantaneous magnitude of rotor q-axis flux, d-axis flux, rotor flux, rotor position and load torque using equations (9), (10), (11), (13), and (14) respectively.…”