用DSP实现智能化自动重合闸

Implementation of intelligent auto-reclosing with DSP

提出使用数字信号处理器(DSP)高速单片机实现智能化自动重合闸方法,该方法结合电力线路工作和跳闸时的参数变化,利用径向基函数(RBF)神经网络所构成的专家系统进行运算,从而确定是否重合闸及重合闸的间隔时间,有效地解决了目前自动重合闸中使电力系统又一次受到故障的冲击和使断路器的工作条件恶化2大难题。由于数学模型分析技术的实现大多依赖于跳闸和合闸时的实时建模,使得这一技术难以在复杂的、时变非线性的、信噪比低的系统中应用。而神经网络具有非线性映射、自学习能力、容错能力、抗干扰能力等特点,能够解决存在的问题。试验证明设计方案可满足实际要求。

A method to implement intelligent auto-reclosing with DSP is proposed,which considers the difference of power system parameters between normal operation and protective trip,and usesexpert system constructed by RBF neural network to decide reclosing and its mterval,ettectively preventing power system from suffering second shock of fault and preventing circuit breaker＇s working conditions from deterioration. Auto-reclosing technique using mathematic model analysis normally deoends on real-time modeling at trip or closing instant,which is difficult to be implemented in complicated,nonlinear time-variant and low SNR systems,while neural network is more suitable for this application with its abilities of non- linear mapping, self-learning ,fault-tolerance and anti-interference. The experimental result shows it meets practical needs.