机械加工尺寸在线建模与预测

ON LINE MODELLING AND FORECASTING OF MACHINEDDIMENSION OF MECHANICAL PARTS

在对实际检测数据进行深入分析的基础上，提出了对机械加工尺寸数据序列进行在线建模和预测的时间序列分析原理和方法，通过加工实验对所提出的原理和方法进行了验证。

In machining process of mechanical parts, machining precision is affected by many factors including assignable random and time varying ones, the dimensional measurement data of a batch of machined parts are usually a random data series, i.e., a time series. Based on thorough investigation of the data series, a new method is proposed for the on line modelling and forecasting of machined part dimension by applying time series analysis principle, and the method is discussed in detail in the paper. Through machining experiment the method is proved effective, and Auto Regressive Moving Average (ARMA) model well suited for the dimensional data series. The work of this paper provides a basis for the predictive control of machining process.