石油化工装置长周期运行风险超早期精确预警方法
|
摘要
安全预警技术早已被广泛应用到各石油化工企业中,风险预警结合了异常工况发生的概率与后果,具有较好的预警效果。针对现有的风险预警多使用定性分析方法、准确性低等问题,提出石油化工装置长周期运行风险的超早期精确预警方法。运用损失函数定量评估参数偏离带来的预期损失,并运用剩余时间模型计算异常工况发生概率,从而得到参数偏离所产生的风险值,实现参数的风险预警。案例分析将该方法应用于石油化工企业不同装置不同参数的风险预警。结果表明,该方法可以辨识状态参数的早期偏离,并提前10 min左右检测到参数异常的产生,为识别石油化工装置长周期运行过程中异常工况的发生风险提供了早期预警方法。
Safety early-warning technology has long been widely used in petrochemical industry. Risk-early-warning has been a good method because it combines the probability and consequence of abnormal situation. The current risk early warning methods are mostly based on qualitative methods and the accuracy is not ideal. Therefore, a precise risk-early-warning method for long period running petrochemical devices is proposed. In this method, through applying Loss Function to quantitatively estimate the expected loss caused by parameters deviation, and also using Remaining Time Model to calculate the probability of abnormal situation, the risk value can be generated by analyzing parameters deviation, and thus achieve the parameter risk warning. This method was applied into the risk warning of different parameters on different devices in petrochemical industry. Test results showed that this method can effectively identify the early departure of state parameters, and can detect abnormal parameters ten minutes earlier. The proposed method can provide an accurate early warning method for possible abnormal risk during long period running of petrochemical devices.
Safety early-warning technology has long been widely used in petrochemical industry. Risk-early-warning has been a good method because it combines the probability and consequence of abnormal situation. The current risk early warning methods are mostly based on qualitative methods and the accuracy is not ideal. Therefore, a precise risk-early-warning method for long period running petrochemical devices is proposed. In this method, through applying Loss Function to quantitatively estimate the expected loss caused by parameters deviation, and also using Remaining Time Model to calculate the probability of abnormal situation, the risk value can be generated by analyzing parameters deviation, and thus achieve the parameter risk warning. This method was applied into the risk warning of different parameters on different devices in petrochemical industry. Test results showed that this method can effectively identify the early departure of state parameters, and can detect abnormal parameters ten minutes earlier. The proposed method can provide an accurate early warning method for possible abnormal risk during long period running of petrochemical devices.
引文
[1] 贺丁,舒逸聃,赵劲松.基于改进的时滞分析算法的化工过程故障定位[J].化工学报,2012,63(10):3165-3172.(HE Ding,SHU Yidan,ZHAO Jinsong.An improved time-delay algorithm for chemical processes fault identification[J].CIESC Journal,2012,63(10):3165-3172.)
[2] 张明,冯坤,江志农,等.基于动态自学习阈值和趋势滤波的机械故障智能预警方法[J].振动与冲击,2014,33(24):8-14.(ZHANG Ming,FENG Kun,JIANG Zhinong,et al.A mechanical fault early warning methodology based on dynamic self-learning threshold and trend filtering techniques[J].Journal of Vibration and Shock,2014,33(24):8-14.)
[3] 齐咏生,张海利,高学金,等.基于KECA的化工过程故障监测新方法[J].化工学报,2016,67(3):1063-1069.(QI Yongsheng,ZHANG Haili,GAO Xuejin,et al.Novel fault monitoring strategy for chemical process based on KECA[J].CIESC Journal,2016,67(3):1063-1069.)
[4] 柴令华,陈小虎,孟凯,等.基于自回归预测的故障报警方法研究[J].机械,2010,37(11):6-8.(CHAI Linghua,CHEN Xiaohu,MENG Kai,et al.Research on fault alarming method based on auto-regression forecasting[J].Machinery,2010,37(11):6-8.)
[5] GE Z,SONG Z.Performance-driven ensemble learning ICA model for improved non-Gaussian process monitoring[J].Chemometric & Intelligent Laboratory Systems,2013,123(2):1-8.
[6] WANG S,ZHAO Z,WANG F L,et al.Multi-step ahead fault prediction method based on PCA and EMD[C]//Mianyang:Chinese Control and Decision Conference,2011:2874-2878.
[7] CHANG Y,KHAN F,AHMED S.A risk-based approach to design warning system for processing facilities[J].Process Safety and Environmental Protection,2011,89(5):310-316.
[8] 胡瑾秋,张来斌,王安琪.炼化装置故障链式效应定量安全预警方法[J].化工学报,2016,67(7):3091-3100.(HU Jinqiu,ZHANG Laibin,WANG Anqi.Quantitative safety early warning method of fault propagation for petrochemical plants[J].CIESC Journal,2016,67(7):3091-3100.)
[9] 胡瑾秋,张来斌,伊岩,等.非正常工况下化工过程设备故障实时关联预警研究[J].中国安全科学学报,2016,26(9):140-145.(HU Jinqiu,ZHANG Laibin,YIN Yan,et al.Real-time fault associated early warning for petrochemical process equipment under abnormal condition[J].China Safety Science Journal,2016,26(9):140-145.)
[10] 胡瑾秋,郭放,张来斌.结合改进PSO算法和LSSVM的化工异常工况超早期监测预警研究[J].电子测量与仪器学报,2018,32(2):36-41.(HU Jinqiu,GUO Fang,ZHANG Laibin.Study on ultra-early prediction of chemical abnormal situation based on improved PSO algorithm and LSSVM[J].Journal of Electronic Measurement and Instrument,2018,32(2):36-41.)
[11] HASHEMI S J,AHMED S,KHAN F.Loss functions and their applications in process safety assessment[J].Process Safety Progress,2015,33(3):285-291.
[12] HASHEMI S J,AHMED S,KHAN F.Risk-based operational performance analysis using loss functions[J].Chemical Engineering Science,2014,116:99-108.
[13] WANG H,KHAN F,AHMED S,et al.Dynamic quantitative operational risk assessment of chemical processes[J].Chemical Engineering Science,2016,142:62-78.
[14] 胡瑾秋,郭放,张来斌.基于趋势分析的间歇过程异常工况超早期报警研究[J].石油学报(石油加工),2018,34(1):101-107.(HU Jinqiu,GUO Fang,ZHANG Laibin.Study on abnormal situation ultra-early warning of batch process based on trend analysis[J].Acta Petrolei Sinica (Petroleum Processing Section),2018,34(1):101-107.)
[15] 黄玥诚.高危生产系统动态风险的拓扑模型与定量方法研究[D].北京:中国地质大学(北京),2017.Precise Risk-Ultra-Early-Warning Method for Long Period Running Petrochemical DevicesHU Jinqiu ZHANG Liqiang ZHANG LaibinThe expected loss and occurrence probability of abnormal situation were first calculated by Loss Function and Remaining Time Model,and thus obtained the risk value for risk warning.Test results show that this method can achieve early warning about ten minutes earlier than that from DCS.
[2] 张明,冯坤,江志农,等.基于动态自学习阈值和趋势滤波的机械故障智能预警方法[J].振动与冲击,2014,33(24):8-14.(ZHANG Ming,FENG Kun,JIANG Zhinong,et al.A mechanical fault early warning methodology based on dynamic self-learning threshold and trend filtering techniques[J].Journal of Vibration and Shock,2014,33(24):8-14.)
[3] 齐咏生,张海利,高学金,等.基于KECA的化工过程故障监测新方法[J].化工学报,2016,67(3):1063-1069.(QI Yongsheng,ZHANG Haili,GAO Xuejin,et al.Novel fault monitoring strategy for chemical process based on KECA[J].CIESC Journal,2016,67(3):1063-1069.)
[4] 柴令华,陈小虎,孟凯,等.基于自回归预测的故障报警方法研究[J].机械,2010,37(11):6-8.(CHAI Linghua,CHEN Xiaohu,MENG Kai,et al.Research on fault alarming method based on auto-regression forecasting[J].Machinery,2010,37(11):6-8.)
[5] GE Z,SONG Z.Performance-driven ensemble learning ICA model for improved non-Gaussian process monitoring[J].Chemometric & Intelligent Laboratory Systems,2013,123(2):1-8.
[6] WANG S,ZHAO Z,WANG F L,et al.Multi-step ahead fault prediction method based on PCA and EMD[C]//Mianyang:Chinese Control and Decision Conference,2011:2874-2878.
[7] CHANG Y,KHAN F,AHMED S.A risk-based approach to design warning system for processing facilities[J].Process Safety and Environmental Protection,2011,89(5):310-316.
[8] 胡瑾秋,张来斌,王安琪.炼化装置故障链式效应定量安全预警方法[J].化工学报,2016,67(7):3091-3100.(HU Jinqiu,ZHANG Laibin,WANG Anqi.Quantitative safety early warning method of fault propagation for petrochemical plants[J].CIESC Journal,2016,67(7):3091-3100.)
[9] 胡瑾秋,张来斌,伊岩,等.非正常工况下化工过程设备故障实时关联预警研究[J].中国安全科学学报,2016,26(9):140-145.(HU Jinqiu,ZHANG Laibin,YIN Yan,et al.Real-time fault associated early warning for petrochemical process equipment under abnormal condition[J].China Safety Science Journal,2016,26(9):140-145.)
[10] 胡瑾秋,郭放,张来斌.结合改进PSO算法和LSSVM的化工异常工况超早期监测预警研究[J].电子测量与仪器学报,2018,32(2):36-41.(HU Jinqiu,GUO Fang,ZHANG Laibin.Study on ultra-early prediction of chemical abnormal situation based on improved PSO algorithm and LSSVM[J].Journal of Electronic Measurement and Instrument,2018,32(2):36-41.)
[11] HASHEMI S J,AHMED S,KHAN F.Loss functions and their applications in process safety assessment[J].Process Safety Progress,2015,33(3):285-291.
[12] HASHEMI S J,AHMED S,KHAN F.Risk-based operational performance analysis using loss functions[J].Chemical Engineering Science,2014,116:99-108.
[13] WANG H,KHAN F,AHMED S,et al.Dynamic quantitative operational risk assessment of chemical processes[J].Chemical Engineering Science,2016,142:62-78.
[14] 胡瑾秋,郭放,张来斌.基于趋势分析的间歇过程异常工况超早期报警研究[J].石油学报(石油加工),2018,34(1):101-107.(HU Jinqiu,GUO Fang,ZHANG Laibin.Study on abnormal situation ultra-early warning of batch process based on trend analysis[J].Acta Petrolei Sinica (Petroleum Processing Section),2018,34(1):101-107.)
[15] 黄玥诚.高危生产系统动态风险的拓扑模型与定量方法研究[D].北京:中国地质大学(北京),2017.Precise Risk-Ultra-Early-Warning Method for Long Period Running Petrochemical DevicesHU Jinqiu ZHANG Liqiang ZHANG LaibinThe expected loss and occurrence probability of abnormal situation were first calculated by Loss Function and Remaining Time Model,and thus obtained the risk value for risk warning.Test results show that this method can achieve early warning about ten minutes earlier than that from DCS.