Meng-Hui Wang

Extension neural network and its applications

In this paper, a novel extension neural network (ENN) is proposed. This new neural network is a combination of extension theory and neural network. It uses an extension distance (ED) to measure the similarity between data and cluster center. The learning speed of the proposed ENN is shown to be faster than the traditional neural networks and other fuzzy classification methods. Moreover, the new scheme has been proved to have high accuracy and less memory consumption. Experimental results from two different examples verify the effectiveness and applicability of the proposed work.

Novel grey model for the prediction of trend of dissolved gases in oil-filled power apparatus

Abstract The power transformer is a major apparatus in a power system; it is of great importance to detect incipient failures in power transformers as early as possible in order to minimize system outage. In this paper, the modified grey model was applied to predict the oil-dissolved gas trend of power transformers. Then, the future fault of power transformer can directly be identified by the trend analysis, so that we can switch them off safely and improve the reliability of power systems. To verify the proposed approach, 46 sets of power transformer records in Taiwan have been tested. It is shown that the proposed method is simple and efficient.

A novel clustering algorithm based on the extension theory and genetic algorithm

This paper presents a novel clustering method this is called extension genetic algorithm (EGA). The new method is a combination of extension theory and genetic algorithm (GA). In the past, we used the extension method in some clustering problems. With the method, we had to rely on experiences to set rules on classical domain and weight, which caused to increase two tedious and complicated steps in clustering processes. In order to improve this defect, the paper uses the EGA to find the best parameter of classical domain. Through the simulations, we prove that this new method can eliminate try and error adjustment of modeling parameters and increase the accuracy of clustering problems. Experimental results from three different examples, including two benchmark data sets and one practical application, verify the effectiveness and applicability of the proposed work.

A novel analytic method of power quality using extension genetic algorithm and wavelet transform

The power quality affects the power stability of power company and customers. In order to avoid economic losses caused by the power disturbances, it is necessary to monitor power parameters. This paper aimed at power quality analyses by wavelet transform and proposed a novel algorithm called extension genetic algorithm (EGA). The paper introduced the fundamental theory of wavelet transform, current applications and the theoretical framework of EGA. Then, it described the definition of power quality problems and the characteristics of power waves. Finally, this paper compared the analysis results of EGA and other methods. As the results of simulation, this paper mentioned of methods has a very high accuracy. It can also provide an application tool on power quality and data classification for future researchers.

A novel extension neural network based partial discharge pattern recognition method for high-voltage power apparatus

This paper proposes a novel partial discharge (PD) pattern recognition method based on extension neural network (ENN) using fractal features. Five types of defect models are well-designed on the base of investigation of power apparatus failures. A PD detector is used to measure the raw three-dimension (3D) PD patterns, from which the fractal dimension, the lacunarity, and the mean discharges of phase windows are extracted as PD features. These critical features form the cluster domains of defect types. An ENN is then developed to recognize the pattern of partial discharge, which utilizes an extension distance (ED) instead of Euclidean distance to measure the similarities among the recognized data and the cluster domains. The ENN with simpler structure than traditional neural networks is capable of processing the clustering problems which have a range of feature values, supervised learning, continuous input, and descriptive output. Moreover, the ENN has the advantages of higher accuracy, shorter learning times, and noise tolerance, which are useful in recognizing the PD patterns of electrical apparatus. To demonstrate the effectiveness of the proposed method, comparative studies among multilayer neural network (MNN), extension theory, and K-means are conducted on 200 sets of field-test PD patterns with rather encouraging results.

Using thermal image matter-element to design a circuit board fault diagnosis system

This paper presents a thermal image matter-element used to design a circuit board signal fault diagnosis system. When a circuit element presents faults the temperature distribution will skew. Therefore, extension theory is used to build several kinds of thermal image matter-element models with fault circuits. According to the matter-element and correlation function, the fault type in the testing circuit is detected by analyzing the correlation degree between the typical fault models and test circuit boards. This new method can attain fast fault determination and reduced manpower.

Grey-Extension Method for Incipient Fault Forecasting of Oil-Immersed Power Transformer

In this article, a grey-extension fault forecasting method (GEFFM) based on the grey and extension theories is proposed for oil-immersed power transformers. First, a modified grey model (MGM) is applied to predict the oil-dissolved gas trend of power transformers. Then, a novel extension diagnosis method is presented for fault forecasting according to the gas predicted values of power transformers. The novel fault diagnosis method is based on the matter-element model and extended relation functions. Thus, future faults in power transformers can directly be identified by the degree of correlation. Therefore, the maintenance engineers can switch them safely and improve the reliability of power systems. The application of the proposed method to some power transformers has given promising results.

Extension neural network

A novel extension neural network (ENN) is proposed in this paper. This new neural network is a combination of extension theory and neural network. It uses an extension distance (ED) to measure the similarity between data and cluster center. The learning speed of the proposed ENN is shown to be faster than the traditional neural networks and other fuzzy classification methods. Moreover, the new scheme has been proved to have high accuracy and less memory consumption. Experimental results from the iris data classification problem verify the effectiveness.

A novel fault diagnosis method based-on modified neural networks for photovoltaic systems

The main purpose of this paper is to propose an intelligent fault diagnostic method for photovoltaic (PV) systems. First, Solar Pro software package was used to simulate a photovoltaic system for gathering power generation data of photovoltaic modules during normal operations and malfunctions. Then, the collected power generation data was used to construct matter-element models based on extension theory for PV systems. The matter-element model combines with the neural networks to form an intelligent fault diagnosis system for PV systems. The proposed fault diagnosis method was adopted to identify the faulty types of a 3.15kW PV system. The simulation results indicate that the proposed fault diagnosis method can detect the malfunction types of PV system rapidly and accurately with less time and memory consumption.

An Intelligent Traffic Light Control Based on Extension Neural Network

This paper presents an intelligent traffic light control method based on extension neural network (ENN) theory for crossroads. First, the number of passing vehicles and passing time of one vehicle within green light time period are measured in the main-line and sub-line of a selected crossroad. Then, the measured data are adopted to construct an estimation method based on ENN for recognizing the traffic flow of a standard crossroad. Some experimental results are made to verify the effectiveness of the proposed intelligent traffic flow control method. The diagnostic results indicate that the proposed estimated method can discriminate the traffic flow of a standard crossroad rapidly and accurately.