针对模糊推理建模得到的HX方程(组),提出了一种线性化方法。对局部HX方程(组)建立精确T-S模型,再将“三角波”隶属函数换成“矩形波”,从而实现线性化,使变系数非线性模型化为变系数线性模型。该方法只针对HX方程(组)中的的非线性项进行局部线性化,避免了对非线性项局部线性化时连带对一次项的零次化。为了避免只针对非线性项中的一个变量进行局部线性化而出现的偏颇,给出了联合方程的方法。仿真实验表明,这种新的线性化方法具有较高的逼近精度。
A linearization method of the HX equations is proposed, where the HX equations are obtained by fuzzy inference modeling. The T-S fuzzy models are exactly built to represent the local HX equations. Then we substitute the "triangular wave" by using "rectangular wave". The linearization are established. The HX equations are turned into linear differential equations with variable coefficients. The method presented in this paper only linearizes the nonlinear terms of the HX equations and avoids zero order processing of some linear term. The "joint equation" method is proposed to avoid the partiality by only linearizing one variable. The simulation results show that the linearization methods are of higher approximation precision.