545 / 2017-05-13 19:11:34

Study on Fast Early Detecting and Rapid Accurate Fault Parameters Estimation Method for Short-circuit Fault

会场 > Fault Current Limiting Technologies

Abstract- Nowadays, because of increasing complexity and capacity of power network, it will be a severe test for the system when short-circuit faults happen. Traditional power supply protection equipment, for instance circuit breaker, have a lag time of tens of milliseconds, from faults happening to automatically switching off fault circuit. Moreover, the ability of those equipment on cutting fault current is limited. Therefore, it’s necessary to detect and identify short-circuit fault rapidly and effectively, especially for large electric power system. Fast early detecting method for short-circuit fault has been a hot research topic nowadays. In this paper, an early detecting method based on Wavelet Packet Decomposition is proposed and implemented. Due to complicated control process of varies of nonlinear load, the original current signal is interfused by several kinds of interference noises, including white noise and impulse noise. It is one of difficult points to filter out the interference signal. The generalized morphological filter is used to remove the positive and negative pulse components in this paper. The morphological filter originates from the different combinations of the two fundamental operators, erosion and dilation, and a kernel called structuring element. The size of the structuring element determines the effect of the filter on current signal. The open operator with different sizes of structuring element, which is defined as the sequence erosion-dilation, is used in this paper. The filtering effect is shown in Fig.1. By taking IEEE 14-node system as an example, the short-circuit fault occurred load branch of 6th-node (Fig.2). Then, detail component of the fourth scale d4 and its higher order detail component d5, can be got by applying wavelet packet decomposition algorithm to filtered current signal Is. By studying the value of d4 and d5 in different fault phase angle and fault time, it is proven that d5 is more proper as fault eigenvector. The optimum value of d5 is determined，comparing the range of d5 under normal operation and fault condition. Finally the proposed algorithm is implemented based on the DSP, and the short-circuit fault signal is generated by arbitrary waveform generator (AWG). The experiment of fault detection time have been carried out, and experimental results show that the method can quickly detect the short-circuit fault, less than 1ms.