计算机集成制造系统 ›› 2019, Vol. 25 ›› Issue (第1): 71-80.DOI: 10.13196/j.cims.2019.01.007

• 当期目次 • 上一篇    下一篇

基于FFSM的数控机床加工状态建模方法

郭安1,3,于东2,3+,胡毅2,3,4,李浩1,3   

  1. 1.中国科学院大学
    2.高档数控国家工程研究中心
    3.中国科学院沈阳计算技术研究所
    4.沈阳高精数控智能技术股份有限公司
  • 出版日期:2019-01-31 发布日期:2019-01-31
  • 基金资助:
    智能制造综合标准化与新模式应用资助项目(2016ZXFB02002)。

Fuzzy finite state machine based processing state modeling method for computer numerical control machine tools

  • Online:2019-01-31 Published:2019-01-31
  • Supported by:
    Project supported by the Comprehensive Standardization and New Mode Application Program,China(No.2016ZXFB02002).

摘要: 为描述数控机床加工过程中的刀具状态,提出一种基于模糊有限状态机的建模方法。由该方法建立的模型通过“状态叠加”机制定量刻画了刀具的临界属性,从而预测刀具未来状态。围绕建模步骤、模型参数优化和预测方法进行了详细讨论,定性分析了状态转换规则及输入变量,设计了一种基于遗传算法的模型参数定量优化方法,并揭示了当刀具分别处于稳态与临界态时的预测原理。实验表明,在一定预测步长内,其预测误差小于低阶线性自回归模型,同时该模型具有连续预测刀具状态的能力。

关键词: 临界状态, 数控机床, 模糊有限状态机, 遗传算法, 信息物理系统, 数字孪生

Abstract: To synchronize the processing states from a Computer Numerical Control (CNC) machine entity to its counterpart,a modelling method based on Fuzzy Finite State Machine (FFSM) was proposed.The model established by this method was able to quantify the critical attributes of the tool through "state superposition" mechanism for predicting its future states based on prior knowledge.A detailed discussion on modelling steps,parameter optimization and prediction mechanism was expounded,which analyzed the state transition rules and input variables qualitatively,designed a quantitative optimization method based on genetic algorithm for model parameters and revealed the prediction mechanism of tools in steady and critical states respectively.The experiment results showed that the prediction errors of the proposed model was less than the errors of the low-order Linear Autoregressive Model (LAM) within a certain prediction steps,and that the model had the ability to predict machine state continuously.

Key words: critical state, computer numerical control machine tools, fuzzy finite state machine, genetic algorithms, cyber-physical systems, digital twin

中图分类号: