基于遗传算法优化的花式捻线机转速值预测  

Fancy Twister Rotate Speed Prediction Based on Optimized Genetic Algorithm

作　　者：王延年[1] 翟伟勋 宋功庆 WANG Yannian;ZHAI Weixun;SONG Gongqing(Xi'an Polytechnic University,Shaanxi Xi'an,710048)

机构地区：[1]西安工程大学

出　　处：《棉纺织技术》2020年第2期25-29,共5页Cotton Textile Technology

基　　金：陕西省科技厅工业领域一般项目(2019GY-109);西安市科技局科技计划项目[201805030YD8CG14(1)];西安工程大学柯桥纺织产业创新研究院项目(19KQYB02)

摘　　要：探讨基于遗传算法优化的花式捻线机转速值预测模型。针对纺织厂花式捻线机生产中工艺参数转换至罗拉和锭子转速值的预测问题,采用了遗传算法来优化传统以BP神经网络为基础的预测模型,利用遗传算法的全局寻优特点对BP神经网络的权值和偏置进行优化,再通过BP神经网络算法进行罗拉和锭子转速值的预测,改进了BP神经网络容易陷入局部极小值和收敛速度慢的问题。试验数据表明:基于遗传算法优化的BP神经网络的预测数据精确、误差小。认为:该预测模型可以满足花式捻线机转速值预测的需要。The rotate speed prediction model for fancy twister based on optimized genetic algorithm was discussed.Aimed at the prediction problems of the roller and spindle rotate value transferred from the production processing parameters of fancy twister in textile mills,genetic algorithm was used to optimize the traditional prediction model based on BP neural network.The weight and polarization by BP neural network were optimized by using the global optimization property of genetic algorithm.Then,the prediction of the rotate speeds for roller and spindle were predicted with BP neural network algorithm.The problems of BP neural network easily falling to local minimum and slow convergence rate were improved.The test results showed that the prediction data of optimized BP neural network based on genetic algorithm is more accurate and with smaller deviation.It is considered that the prediction model can meet the requirement of fancy twister rotate speed prediction.

关 键 词：遗传算法 BP神经网络 花式捻线机 花式纱线 超喂比 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] TS103.23[自动化与计算机技术—计算机科学与技术]