摘要
针对针阀偶件插配智能化加工需求,设计并建设针阀偶件智能化插配加工单元.对加工尺寸的数据进行分析,改进针阀偶件间隙配对加工方法;构建间隙配对问题数学模型,采用改进的遗传算法对其进行求解,实现针阀偶件间隙配对的最优化,现场使用验证,配对成功率 85%左右;基于迁移学习训练光学字符识别(OCR)模型,识别针阀体生产顺序号,识别准确率达 95%,并将针阀偶件插配过程数据与之关联,实现数据自动记录,便于质量追溯.
Abstract
Design and construct an intelligent machining unit for needle valve fittings in response to the demand for intelligent machi-ning of needle valve fittings.Analyze the machining dimension data and improve the machining method for matching the clearance be-tween needle valve components;Construct a mathematical model for gap pairing problem and solve it using an improved genetic algo-rithm to achieve optimal gap pairing of needle valve components.The model has been validated on site with a success rate of approxi-mately 85%;Based on transfer learning,an OCR model is trained to recognize the production sequence number of needle valve bodies with an accuracy rate of 95% .The data of the needle valve fitting process is associated with it to achieve automatic data recording and facilitate quality traceability.
维普
万方数据