非线性电路故障诊断模型的建立及神经网络求解的方法研究

Research on Approach of Deriving Nonlinear Network Fault Model and Solving It by ANN′S

分析并推导了含有非线性流控型电阻元件和非线性压控型电阻元件组成的非线性电阻网络的故障诊断方程,这是以支路故障标志量为未知的一组线性代数方程.在推导非线性元件的"故障标志量"时,消除由工作点变化引起的电参量的变化,保留由伏 安特性变化引起的电参量变化,它标志非线性元件本身发生了故障;文章还提出一种神经网络,可在线快速求解非线性电阻网络的故障方程的解,从而可在线对故障进行定位.

After analysis of the fault characters of nonli near resistance network consisting of current-controlled resistors and voltage- controlled resistor s, the fault diagnosis equations of nonlinear resistance network was deriv ed. The unknown quantities of these equations, which are algebraic linear equati ons, are quantities indicating fault of each branch of network . When derivin g the quantities, the variation of electrical parameter caused by v ariation of operating point is eliminated, and the variation caused by variation of v-i cha racteristic is kept, because the later represents the fault occurring in the nonlinear element. In order to solve these fault diagnosis equations with hi gh speed computation on line, we adopted a method based on ANN′s.