基于双阶跃信号输入的Volterra模型辨识

Volterra model identification based on two-step signal input

Volterra模型作为非线性领域的一种非线性模型,由于其对工业过程可以以任意精度逼近,使得该模型有很广泛的应用研究意义。在将该模型运用到实际控制系统中之前,模型的高精度辨识显得尤为重要。在以往针对Volterra模型的辨识算法中,基本上主要是采用通用辨识算法识别模型参数,比如最小二乘法及各种改进的最小二乘法。这些通用的辨识算法在辨识Volterra模型时,不能充分考虑其非线性特点,同样不能在辨识过程中充分利用该特点。本文在充分考虑Volterra模型非线性的前提下,提出了一种基于双阶跃信号输入的Volterra模型辨识算法,该算法辨识原理简单,计算量较小,论文最后将该辨识算法应用到典型非线性CSTR系统的的辨识中,辨识结果证明了算法的有效性。

Volterra model as a nonlinear model nonlinear field, because it can be approached with arbitrary precision industrial process, making the model has a very wide range of applied research significance. Before applying the model to the actual control system, high-precision identification model is particularly important. In the past, Volterra model for the identification algorithm, basically universal identification algorithm is mainly used to identify the model parameters, such as least-squares method of least squares and various improvements. These generic identification algorithm in the identification of Volterra model, cannot fully consider the nonlinear characteristics, the same cannot take full advantage of the features in the identification process. In this paper, taking full account of the premise nonlinear Volterra model, we propose a two- step signal input based on Volterra model identification algorithm that identification is simple in principle, the smaller amount of calculation Finally, the identification algorithm is applied to a typical nonlinear CSTR system identification, the identification results show the effectiveness of the algorithm.