板带轧机随机干扰模型与ARMA谱分析

Stochastic excitation model for a plate ant strip rolling mill and its ARMA spectral analysis

板带轧机随机干扰模型不易建模,是制约轧机随机动力特性研究的主要难点之一。基于某铝热连轧机轧制1235铝合金的轧制力随机数据,得到平稳、正态、零均值的时间序列,采用时序分析法建立轧制力的ARMA模型及轧机动态干扰功率谱密度模型,分析了轧制力的ARMA谱特性,采用麦夸特法+通用全局优化算法,得出便于工程应用的轧制力功率谱表达式,构建轧制力随机干扰模型。对轧制力的预报、轧机动刚度的确定及对轧机与轧制过程的随机动态特性研究与分析具有一定的意义。

One of the main difficulties for study on stochastic dynamic characteristics of a strip rolling mill is to establish its stochastic excitation model. Based on stochastic rolling force data acquired from an aluminum hot strip tandem mill, stationary, normal and zero mean time series were obtained. The ARMA model for the rolling force and the stochastic excitation power spectral density （PSD） model were established with the time series analysis method. The ARMA spectral characteristics of the rolling force were analyzed by means of random process theory. The rolling forces PSD expressions for engineering application were deduced by utilizing Levenberg-Marquardt method combined with the generalized global optimization algorithm, and the stochastic excitation model was built. The results provided a basis for prediction of rolling force, identification and parameter estimation of dynamic rigidity of a mill, and analyzing and studying stochastic dynamic characteristics of rolling process and a plate and strip rolling mill.