基于SVM的纱线生产工艺智能优化研究

Intelligent optimization study of yarn production processes based on SVM

针对纺织企业在纱线生产工艺控制方面存在的问题,以及现有纱线质量控制方法存在的不足,将质量相似度模型和支持向量机应用到纱线生产工艺调整问题中。利用遗传算法对支持向量机的2个关键参数C、σ进行寻优,得到支持向量机纱线质量预测模型,在原有工艺方案的基础上,根据生产质量要求,调整局部工艺参数,然后通过质量预测模型进行质量预测,实验显示,该方法能满足纱线生产的质量控制和决策要求,为纱线多品种、小批量生产下的工艺设计和优化提供了新的思路和方法。

Aiming at some existing problems on the yarn production technology control and defects on the yarn quality control method,quality similarity model and support vector machine is applied to yarn production process adjustment. The two key parameters C and σ of support vector machine can be optimized by using genetic algorithm and yarn quality prediction model of support vector machine can be obtained. Based on the original process scheme,the local process parameters can be adjusted according to the production quality requirements and then quality prediction can be made through the quality prediction model. Experiments show that this method can meet the quality control and decision requirements of yarn production and provide a new idea and method for process design and optimization of multi varieties and small batch production of yarn.