基于组合赋权-云模型的LSTM氨合成工艺控制风险预警研究  

Risk warning for ammonia synthesis process control using LSTM and combination weighting cloud model

作　　者：陈樑 石舒文[1] 朱君烨 雷坚 郭冰 马小义 CHEN Liang;SHI Shuwen;ZHU Junye;LEI Jian;GUO Bing;MA Xiaoyi(School of Public Safety and Emergency Management,Kunming University of Technology,Kunming Yunnan 650500,China;School of Environmental Science and Engineering,Kunming University of Technology,Kunming Yunnan 650500,China;Yunnan Shuifu Yuntianhua Co.,Ltd.,Zhaotong Yunnan 657000,China;Yunnan Tian'an Chemical Co.,Ltd.,Kunming Yunnan 650300,China)

机构地区：[1]昆明理工大学公共安全与应急管理学院,云南昆明650500 [2]昆明理工大学环境科学与工程学院,云南昆明650500 [3]云南水富云天化有限公司,云南昭通657000 [4]云南天安化工有限公司,云南昆明650300

出　　处：《工业安全与环保》2025年第4期70-78,共9页Industrial Safety and Environmental Protection

基　　金：云南省科技计划项目(202003AC100002)。

摘　　要：提出了一种基于组合赋权-云模型的长短期神经网络(LSTM)氨合成工艺控制风险预警模型,该模型通过LSTM采集氨合成工艺7个主要控制参数的运行数据,预测未来240 min的数据变化,运用以期望、熵和超熵3个数字特征为主的云模型评估方法来表达控制参数风险隶属度变化关系,最终通过组合赋权确定工艺的综合风险等级。结果表明,该模型能精准预测未来240 min内氨合成工艺控制参数风险动态变化规律,及时提醒操作人员辅助自动控制系统干预调整以实现降低装置运行的安全风险,为类似危险化工工艺装置过程风险控制和工艺参数优化管理提供了实际应用价值。A risk warning model for ammonia synthesis process control based on a combined weighting-cloud model and Long Short-Term Memory(LSTM)neural network is proposed.The model collects operational data for seven key control parameters of the ammonia synthesis process using LSTM,and predicts data changes for the next 240 minutes.The cloud model evaluation method,based on three digital features of expectation,entropy,and hyper-entropy,is applied to express the changes in the risk membership degree of control parameters.Finally,a combined weighting method is used to determine the comprehensive risk level of the process.The results show that the model can accurately predict the dynamic changes in the risk of ammonia synthesis process control parameters over the next 240 minutes,and provide timely warnings to operators,enabling intervention and adjustment by the automatic control system to reduce the operational safety risks of the facility.This model offers practical value for process risk control and parameter optimization management in similar hazardous chemical processes.

关 键 词：神经网络 云模型 层次分析法 熵权法 组合赋权 氨合成工艺 风险预警 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]