摘要
提出了一种注塑工艺参数自适应控制系统的设计方法,通过构建注塑工艺和质量指标之间的多维非线性预测模型,在过程工艺参数的扰动下,基于满意优化理论,系统能根据满意度要求实时自动调整多个工艺参数,优化综合质量指标。多目标综合满意度函数的权系数可以由实验数据或实时数据库通过熵值赋权法确定,减少了人为因素对于优化设计的影响。实验结果验证了该方法的有效性和实用性。
A new approach was presented for the adaptive-control system design of injection molding process parameters. A neural network model was established to predict the multi-dimensioned non-linear relationship between injection process and quality index. When injection process parameters changed, the adaptive-control system based on satisfactory optimization theory could rectify several parameters to meet the comprehensive quality requirement. The weight of each objective in comprehensive satisfactory function could be determined by the data from experiments or real-time conditions with entropy method, which could avoid influence of artificial factors. The experimental results demonstrated the effectiveness and practicability of the approach.
出处
《塑料工业》
CAS
CSCD
北大核心
2009年第9期39-42,49,共5页
China Plastics Industry
基金
国家自然科学基金(50765001)
广西教育厅科研基金(200708MS028)
关键词
满意优化
注塑
自适应控制
BP神经网络
熵值赋权法
Satisfactory Optimization
Injection Molding
Adaptive-control
BP Neural Network
Entropy Method
