Papers by Keyword: Wavelet Transform (WT)

Abstract: In order to improve the accuracy of diagnosis pumping, and accelerate the speed of diagnosis, a fault diagnosis model based on improved extreme learning machine (RWELM) was proposed. Firstly, it extracted the energy characteristic eigenvector of dynamometer cards of an oilfield in northern Shanxi by using wavelet packet decomposition method. Then through simulation of fault diagnosis, and compare with the extreme learning machine (ELM), RBF neural networks and support vector machine (SVM). The experimental results show that the accuracy and the speed of fault diagnosis based on the RWELM are better than the ELM, RBF neural network and SVM.

Abstract: For the pulse wave signal acquired by the pulse analyzer, it’s necessary to remove the large number of noise signal by noise reduction processing. For signal noising processing using wavelet transform method in the time domain and frequency domain and the signals can characterize the ability of local information, and it is self-adaptive. Determine the main wave peak and trough points based on extreme value method and pulse wave periodicity, using the threshold method to determine the dicrotic wave starting point and dicrotic wave peaks, we also use the MATLAB programming to implement the algorithm.

Abstract: A new scheme to enhance the solution of the problems associated with Transmission line protection with Statcom connected is presentedin this paper.Static Synchronous Compensator (STATCOM) is a shunt type FACTS device connected at the midpoint of the transmission line to maintain the voltage atdesired level by injecting/absorbing the reactive power. This connection affects the performance of distance protection relay during line faults. Thefault detectionis carried out byusingenergy of the detail coefficients of the phase signals and artificial neutral network algorithm used for fault distance location for all thetypes of faults for transmission line. For each type of fault separate neural network is prepared for finding out the fault location.

Abstract: Using the principle of wavelet transform in the aspect of signal singularity detection analyzes and detects the electric power system fault signal. Then we extract signal feature near the fault moment and sent the feature vectors into the neural network. The simulation results fully prove the effectiveness and superiority of combining wavelet transform and neural network in electric power system fault recognition.

Abstract: An identification method for cracked eggs by means of the digital image technology was proposed in this paper. Firstly, an ideal machine vision system was built and the images of good eggs and cracked eggs were obtained by CCD camera. Secondly, each image was decomposed on two layers of wavelet, so 6 high-frequency sub-images and 2 low-frequency sub-images were extracted. Then joint probability matrix after wavelet transform had been calculated and 5 parameters for each high-frequency sub-images were extracted, so the total of the joint probability matrix parameters was 30 for 6 high-frequency sub-images. At the same time, 10 wavelet energy parameters were obtained. Thirdly, four main factor component scores were selected from above 40 feature parameters after principal component analysis, which were input to support vector machine. Finally, classification model was built by support vector machine. Experiments show that the proposed method was effective to identify the cracked eggs from good eggs and the identification rate was 93.75%.

Abstract: A partial discharge (PD) signal processing method based on dynamic measurement theory and wavelet transform is proposed in this paper. The deterministic component was separated by polynomial fitting, and the random component of the remaining residual after the separation was estimated using autoregressive (AR) model; The true value estimate and dynamic measurement uncertainty of noisy signal were obtained by the deterministic component and random component;Db8 wavelet and the soft threshold based on Stein’s Unbiased Estimate of Risk were used to smoothly denoise for better PD signal processing. Finally, the effectiveness of the method was verified by MATLAB simulation and experimental noisy PD signal extraction.

Abstract: A power network disturbance identification system (PNDIS) was developed as a new function module of wide-area security defense system (WASDS), which has been commissioned in the dispatching center of a real power system. This paper describes the implementation of its software platform which consists of three parts: data sharing service, calculation engine and visualization module. The paper focuses on the design of the calculation engine, which is the key part of the platform, and the visualization module. Wavelet transform (WT) is used to identify the time and location of disturbance in the power system. Tests on measured data recorded by WAMS are presented in order to illustrate the benefits of the software platform and show its excellent performance.

Abstract: Since the strong coupling property between partial discharge and external interference, how to obtain an acceptable Resolution is a huge challenge during partial discharge measure. In terms of interference suppression, many methods associated with analog and digital technique are devised. Whatever the methods are, the way to eliminate pulse interference is the toughest task due to the closest feature in the respect of time and frequency field compared with partial discharge. In this paper, we propose a wavelet based interference separation method to separate partial discharge signal compounding with random pulse signal, mobile communication signal and white noise, providing valid data for the following partial discharge measurement. Both simulation and practical results verify the effectiveness of our proposed method.

Abstract: A new recognition method of gait is presented in this paper. In this method, the feature vector of gait is found by multidistinguish analysis of wavelet transform, and gait is recognized by genetic algorithm (GA). This method is different from the traditional method of correlation matching recognition gait. First, the stored space reduces greatly because recognition model is used to replace the store of gait profile image template. Thus this method reduces stored memory greatly. Second real-time is ensured in the process of gait recognition by using GA. The experiments of recognition using the three kinds of gait databases are performed. The experiment results show the feasibility and effectiveness of the proposed method.

Abstract: Since the infrared image has the problem of poor contrast, edge blur, low noise ratio, noisy and other shortcomings, this paper presents an improved threshold denoising algorithm based on curvelet transform, using a combination of hard and soft threshold methods to form a new threshold based functions. Compare to the conventional hard threshold, soft threshold, de-noising based on wavelet transform method, the proposed denoising method can obtain better PSNR and visual characteristics, retain more details