中国安全科学学报 ›› 2023, Vol. 33 ›› Issue (8): 149-155.doi: 10.16265/j.cnki.issn1003-3033.2023.08.1389

• 安全工程技术 • 上一篇    下一篇

模糊推理-组合赋权的化工工艺本质安全度研究

张任栋(), 张礼敬**(), 陶刚   

  1. 南京工业大学 安全与科学学院,江苏 南京 211816
  • 收稿日期:2023-03-11 修回日期:2023-06-13 出版日期:2023-10-08
  • 通讯作者:
    **张礼敬(1962—),男,江苏高邮人,博士,教授,主要从事化工过程及装置安全、失效分析、安全技术及工程、金属材料及热处理、金属焊接等方面的研究。E-mail:
  • 作者简介:

    张任栋 (1997—),男,山东青岛人,硕士研究生,研究方向为化工过程风险评价控制理论与技术。E-mail:

    张礼敬 教授

    陶 刚 研究员

Research on inherent safety of chemical process based on fuzzy reasoning-combination weighting

ZHANG Rendong(), ZHANG Lijing**(), TAO Gang   

  1. College of Safety Science and Engineering,Nanjing Tech University,Nanjing Jiangsu 211816,China
  • Received:2023-03-11 Revised:2023-06-13 Published:2023-10-08

摘要:

为提高化工企业安全生产水平,基于本质安全指数法(ISI)选取10个指标,建立化工工艺过程本质安全评价指标体系,划分各评价指标等级,结合Matlab环境下的Mamdani模糊推理方法进行模糊推理;依据四级指标参数下的安全度等级评分数值,建立各三级指标下的规则输出矩阵,最大程度上保留模糊理论型本质安全评价结果的灵敏性,得到可视化三维模糊推理曲面图;为精确本质安全评价结果,建立兼顾主客观权重且距离最小的目标函数,优化利用归一法和熵权法计算出的各评价指标的主客观组合权重,减少组合赋权结果偏差;应用该方法评估 4 条工艺路线的本质安全度。结果表明:基于模糊推理-组合赋权的本质安全评价法,在兼顾主客观权重的同时,最小化主客观权重距离偏差,线性加权综合计算各项安全指标评分,精确输出组合赋权后的本质安全评估结果数值。该方法的评估结果与前人文献中评价结果基本吻合,证明该方法具有一定的科学性和有效性。

关键词: 模糊推理, 组合赋权, 化工工艺, 本质安全度, 评价指标

Abstract:

In order to improve the production safety level of chemical enterprises, 10 indexes were selected based on the intrinsic safety index method (ISI), and an intrinsic safety evaluation index system for chemical processes was established. Each evaluation index level was divided, and fuzzy reasoning was carried out with Mamdani fuzzy reasoning method in the Matlab environment. The rule output matrix under each three-level index was established according to the safety grade score value under the four-level index parameters. The sensitivity of fuzzy theoretical intrinsic safety evaluation results was preserved to the maximum extent, and the visual 3D fuzzy inference surface diagram was obtained. In order to accurately evaluate the intrinsic safety evaluation results, an objective function with the minimum distance and both subjective and objective weights was established, and the combined subjective and objective weights of each evaluation index calculated by the normalization method and entropy weight method were optimized to reduce the bias of the combined weight results. This method was applied to evaluate the intrinsic safety degree of four process routes, and the evaluation results were basically consistent with the evaluation methods proposed by predecessors. The research shows that: Based on the fuzzy inference-combination weighting intrinsic safety evaluation method, the distance deviation of subjective and objective weights is minimized while taking into account the subjective and objective weights. The evaluation results of this method are basically consistent with those of previous literature, which proves that this method has certain scientificity and effectiveness.

Key words: fuzzy logic, combination weighting, chemical process, inherent safety, evaluation indexes