纺织生产在线测试分析的集成模型及误差分析

摘要

针对纺织生产非电量检测与转换中,采用回归分析模型进行电测值与非电量值转换时存在较大误差的情况,给出了建立基于回归模型和灰色残差模型集成的思路.通过对实例建模过程的分析,发现集成模型的拟合精度远远高于单纯的回归分析模型.集成模型的拟合误差呈现递减的变化趋势,这符合灰色系统理论的“新息优先”和“新陈代谢”原理.由于利用集成模型可以替代人工测试分析和计算,将其应用于纺织生产过程控制,可以通过实时检测转换实现纺织生产自动化的目标.%Due to statistically significant errors by using regression model in the electric measuring and the transforming for nonelectric parameters, an integrated application of the residual grey model and the regression model was presented. By simulation modeling and analysis, it showed a significant higher fitting precision in prediction versus the results of only the regression model was used. Further analysis showed a declining trend of the fitting error in the integration model, which was in accordance with the basic principle of Priority Innovation and Metabolism in grey system theory. Consequently, the integration model as an alternative to manual calculation can be applied in the production process control of textile industry. This technology adopted in real-time test offers new options in automated textile production.