基于贝叶斯网络的机床加工质量异常预警方法
首发时间:2020-01-02
摘要:针对智能制造车间机床加工过程中,导致加工质量异常的各类影响因素来源广、频率高、随机性强且难以管控的难题,提出一种基于贝叶斯网络的机床加工质量异常预警方法。首先,以机床加工质量异常影响因素为节点,构建了一种基于贝叶斯网络的机床加工质量异常模型;其次,采用ChiMerge聚类算法将连续状态数据离散化,运用EM算法对离散化后的不完整数据集进行参数学习;最后,结合贝叶斯公式以最大后验假设完成机床加工质量异常预警贝叶斯网络的概率推理,并通过实例验证了该方法的可行性和有效性。
关键词: 机床加工质量 贝叶斯网络 异常预警 离散化 参数学习
For information in English, please click here
Machining Process Pre-warning Method for Abnormal Quality Based on Bayesian Network
Abstract:During machining process in intelligent manufacturing workshop, abnormal machining quality occurs frequently and is difficult to control due to various factors such as wide source, high frequency and randomness. To address this, an early warning model for machine tools abnormal quality based on Bayesian network is proposed. Firstly, building a model for abnormal machining quality based on Bayesian network by taking the factors affecting the abnormal quality of machine tools as nodes. Secondly, the continuous state data is discretized by using Chimerge clustering algorithm, and the incomplete data sets are learned by using EM algorithm. Finally, the probability reasoning of Bayesian network for early warning of abnormal quality during machining process is completed by combining Bayesian formula with the maximum a posterior hypothesis, and the feasibility and validity of this method are verified by an example.
Keywords: machining quality Bayesian Network abnormal pre-warning discretization parameter learning
引用
No.****
同行评议
共计0人参与
勘误表
基于贝叶斯网络的机床加工质量异常预警方法
评论
全部评论0/1000