电路瞬态过程中响应模型的参数估计

摘要

当电路瞬态过程的电压或电流响应是以离散数据点给出时,为了测量响应模型的各个参数,提出了一种基于曲线拟合的利用非线性最小二乘优化进行参数估计的算法.首先构造二乘残差目标函数,给出各个参数初始值的计算方法,然后采用拟牛顿法搜索目标函数的极小值.该方法可以一次性求出模型的全部参数.模拟算例和仿真结果表明:算法的稳定性较强,收敛速度和计算精度都较高.%In order to measure each parameter on the response model, an algorithm on parameters estimation is put forward. It is based on curve-fitting's nonlinear least squares optimization, when the response of electric voltage or current is supplied by discrete data in transient process of circuit. Firstly, the algotithm structure squared residual objective function and offered calculating method on each initial values of parameters, then it searched the minimum value of objective function using Quasi Newton Method. This method can calculate all the parameters in model at a time. So the emulation examples and the simulation results show that it has better stability and higher rapidity of convergence and precision of calculation.