Qidi Wu

General fuzzy neural network: theory and applications

We proposed a new general fuzzy neural network (GFNN) which can process the structural self-organizing and parametric self-learning online. Based on the concept of fuzzy logic system, GFNN has a structure of a three-layer feedforward network and a wider information interface. Its neurons in the hidden layer correspond to the fuzzy rules and have some functions of information storage and inference. It is easy for GFNN to realize the novel structural self-organizing online with the presentation of three concepts: the match degree; the match degree criterion; and the minimum believable level. So it can be used to identify the time-varying nonlinear system online.

Controllability of nonisolated DC-DC converters with constant-power-load

This paper addresses the controllability of six types of nonisolated DC-DC converters (i.e. Buck, Boost, Buck-Boost, SEPIC, Cuk and Zeta) feeding the constant power load (CPL). The operation principle of SEPIC is presented so that the drift vectors and control functions are derived, which demonstrate that each switching mode is an affine nonlinear subsystem. Employing the recent findings on the controllability of switched nonlinear systems, we prove that every state on the maximal integral manifold is an interior of the convex cone generated by (G-shifted) drift vectors and control functions, thus the nonisolated DC-DC converters with CPLs are controllable. This technique could be applied to analyze other isolated converters.

The research of remote fault diagnosis system in manufacturing

Based on an analysis of the realization mode of remote fault diagnosis system in manufacturing, this paper introduced the general structure of remote fault diagnosis system in manufacturing thoroughly. The treatment is illustrated by the example of a Shanghai Volkswagen-oriented application in a Sino-German cooperative project.

Hopfield Neural Network Based Identification and Control of Induction Motor Drive System - Part I: Identification

Abstract The paper discusses a Hopfield-Neural-Network (HNN) based linear system parameters identification scheme in part I, under the assumption that HNN inputs are the detected system states delayed by sensors. Sufficient condition of correct identification is derived. The scheme is used for identification of rotation inertia (J), speed damping coefficient (Rω) and load torque (TL) of induction motor drive system. Simulation results show that even under bad working condition caused by incorrect setting to the controllers, the derived scheme guarantees correct identify results.

Neural network modeling and intelligent control of the distributed parameter greenhouse climate

Abstract In this paper, the artificial neural networks are applied to the model of the distributed parameter systems. A neural network model of the distributed parameter greenhouse climate was established. An optimal control algorithm was proposed based on the neural computation. These methods have been successfully applied to the distributed parameter greenhouse climate. It proved that these methods are effective. The expert control systems of the temperature, humidity, concentrations of carbon dioxide CO 2 , nutrient liquid supply, etc. were introduced. The results of the intelligent control for the distributed parameter greenhouse climate were given at last.

A Review Upon Concurrent Engineering

Abstract This paper makes a review study of concurrent engineering for product development. First, the essence of concurrent engineering, i.e., the concurrence in the sense of information, is analysed thoroughly. Then, meta-process mechanism and modeling, organizational structure, and related methods and tools are briefly reviewed for concurrent engineering as meta-processes. Finally, implementation aspects of concurrent engineering is briefly analysed and reviewed

A dynamic small-signal model of the constant-frequency pulse width modulator of switching converters

This paper proposes a dynamic small signal model of the constant-frequency pulse width modulator of switching converters. The existing model is represented as a simple gain block Fm, which is relatively short in both accuracy and physical meanings. Employing the methods of probability, this dynamic model provides a more accurate prediction of the system behavior in the sense of statistics, compared to its previous counterpart

Hopfield Neural Network based identification and control of induction motor drive system

Abstract The convergency of simplified Hopfield Neural Network (HNN) is proved in this paper. HNN is firstly used for identification of induction motor parameters. Then it is used for design of multi-variable system controller. The outputs of the HNN based controller are calculated according to the dynamic process of HNN for object seeking, thus make the state variables of plant converge to the control object. The derived controller has the character of good robustness. It is used in this paper for flux exciting control and speed control of induction motor. Simulation results show good speed control characteristics against great variance of some drive system parameters. For the parameter which affects control performance greatly, it is adjusted by online identify output of HNN and thus improve greatly the control performance.

Hopfield Neural Network Based Identification and Control of Induction Motor Drive System - Part II: Control

Abstract In this part, Hopfield Neural Network (HNN) is used for model reference adaptive controller design and its convergency is proved. The derived controller is used for flux and speed control of induction motor drive system. Simulation results show good robustness of the derived controller against variance of the drive system parameters (50%). The parameter which affects greatly the control performance is adjusted by online HNN identified value (identify scheme shown in part I) and the control performance is improved greatly.

Internet/intranet based digitized teleservice system

Digitized teleservice system has become a focus of international research and developing. In this paper, the basic function of Internet/intranet based digitized teleservice system was discussed and its corresponding model was proposed. Then its structure was discussed. Finally its design and realization was introduced-associated with SVW-oriented application on Sino-German Cooperation Project.