基于ARIMA模型和BP人工神经网络的产品质量预测  

Product Quality Prediction Based on ARIMA Model and BP Artificial Neural Network

作　　者：曹学晨 张顺堂[1] CAO Xue-chen;ZHANG Shun-tang(College of Management Science and Engineering,Shandong Institute of Business and Technology,Yantai 264000,China)

机构地区：[1]山东工商学院管理科学与工程学院,烟台264000

出　　处：《价值工程》2018年第35期190-193,共4页Value Engineering

基　　金："煤矿安全生产过程仿真培训系统研究与开发";安全生产科技发展指导性计划项目(06519)

摘　　要：在汽车零部件制造行业中,部分企业对产品的质量检验方式采用的是产出后由人工进行,目前针对残次品的处理方法仍然是由人工即时发现即时处理,因此在机床加工产品的同时对产品进行质量预测并由维护人员对机床采取提前预防措施有了很重要的意义。本文采用基于ARIMA模型与BP人工神经网络算法作为产品质量预测方法。从机床原始数据中提取可用信息,对提取后的所有相关参数用SPSS分析模块中的ARIMA模型进行建模、分析与预测,形成质量预测数据,结合应用由Python的Keras库与TensorFlow框架构建的BP人工神经网络在非线性拟合上的优势,构建组合预测模型,实际验证了该模型在短期质量预测方面的可行性。In the auto parts manufacturing industry,the quality inspection of some enterprises is manually performed after output.At present,the processing method for the defective products is still artificial instant processing.Therefore,it is very important to predict the quality of the product while the machine tool is processing the product and take precautionary measures against the machine tool by the maintenance personnel.This paper adopts ARIMA model and BP artificial neural network algorithm as product quality prediction methods.The available information is extracted from the machine raw data,and all relevant parameters are extracted,modeled,analyzed and predicted by the ARIMA model in the SPSS analysis module to form quality prediction data.Combining the advantages of BP artificial neural network constructed by Python's Keras library and TensorFlow framework on nonlinear fitting,a combined forecasting model is constructed and the feasibility of the model in short-term quality prediction is verified.

关 键 词：BP人工神经网络 ARIMA模型 TensorFlow 预测 质量控制 

分 类 号：U468.4[机械工程—车辆工程]