最优加权组合模型在管道腐蚀预测中的应用
|
摘要
为构造海底混输管道腐蚀最优加权组合预测模型,针对传统非等间距GM(1,1)管道腐蚀预测模型中初始条件的选取问题,提出了新信息优先原理下的NEGM(1,1,τ)海底管道腐蚀速率预测模型,以充分发挥建模序列中各分量对预测系统的修正作用;引入ARIMA预测模型,在3个不同定权准则下与NEGM(1,1,τ)模型形成管道腐蚀加权组合预测模型,并通过评价指标函数实现组合模型的性能评价。研究结果表明:组合模型2的海底混输管道腐蚀速率预测值与实际值的平均相对误差为0. 495 4%,评价指标函数RMSE,AARD和MAPE的值分别为0. 1936%,0. 1275%和0. 595 3%,均优于其他2个准则下的组合模型。建立NEGM(1,1,τ)-ARIMA海底管道腐蚀速率最优加权组合预测模型,从多角度挖掘了管道腐蚀速率序列中的可靠信息,预测结果的可信度更高。
In order to construct the optimal weighted combination prediction model for the corrosion of submarine multiphase pipeline,aiming at the problem of selecting the initial values in the traditional non-equal spacing GM( 1,1) pipeline corrosion prediction model,a NEGM( 1,1,τ) prediction model on the corrosion rate of submarine pipeline based on the principle of new information prioritization was put forward,so as to fully exert the correction effect of each component in the modeling sequence on the prediction system. The ARIMA prediction model was introduced and combined with the NEGM( 1,1,τ)model to form the weighted combination prediction model of pipeline corrosion under three different weighting criteria,and the performance evaluation of the combination model was realized by the evaluation index functions. The results showed that the average relative error of the prediction values for the corrosion rate of submarine multiphase pipeline by combination model2 compared with the practical values was 0. 4954%,and the values of evaluation index functions RMSE,AARD and MAPE was 0. 193 6%,0. 127 5% and 0. 595 3% respectively,which were all better than those of combination models under other two criteria. The NEGM( 1,1,τ)-ARIMA optimal weighted combination prediction model for the corrosion rate of submarine pipeline was established,which revealed the reliable information in the corrosion rate sequence of pipeline from multiple perspectives,with a higher reliability of prediction results.
In order to construct the optimal weighted combination prediction model for the corrosion of submarine multiphase pipeline,aiming at the problem of selecting the initial values in the traditional non-equal spacing GM( 1,1) pipeline corrosion prediction model,a NEGM( 1,1,τ) prediction model on the corrosion rate of submarine pipeline based on the principle of new information prioritization was put forward,so as to fully exert the correction effect of each component in the modeling sequence on the prediction system. The ARIMA prediction model was introduced and combined with the NEGM( 1,1,τ)model to form the weighted combination prediction model of pipeline corrosion under three different weighting criteria,and the performance evaluation of the combination model was realized by the evaluation index functions. The results showed that the average relative error of the prediction values for the corrosion rate of submarine multiphase pipeline by combination model2 compared with the practical values was 0. 4954%,and the values of evaluation index functions RMSE,AARD and MAPE was 0. 193 6%,0. 127 5% and 0. 595 3% respectively,which were all better than those of combination models under other two criteria. The NEGM( 1,1,τ)-ARIMA optimal weighted combination prediction model for the corrosion rate of submarine pipeline was established,which revealed the reliable information in the corrosion rate sequence of pipeline from multiple perspectives,with a higher reliability of prediction results.
引文
[1]穆龙新,潘校华,田作基,等.中国石油公司海外油气资源战略[J].石油学报,2013,34(5):1023-1030.MU Longxin,PAN Xiaohua,TIAN Zuoji,et al.The overseas hydrocarbon resources strategy of Chinese oil-gas companies[J].Acta Petrolei Sinica,2013,34(5):1023-1030.
[2]骆正山,车朝阳.基于TVR的腐蚀油气管道失效概率及安全寿命研究[J].中国安全生产科学技术,2018,14(9):95-99.LUO Zhengshan,CHE Chaoyang.Research on failure probability and safe life of corroded oil and gas pipelines based on TVR[J].Journal of Safety Science and Technology,2018,14(9):95-99.
[3]王晓光,周慧.输气管道腐蚀的最优灰色组合预测模型[J].腐蚀科学与防护技术,2009,21(5):496-498.WANG Xiaoguang,ZHOU Hui.An optimal grey combinatorial forecasting model for corrosion of gas pipilines[J].Corrosion Science and Protection Technology,2009,21(5):496-498.
[4]LUO Zhengshan,WANG Wenhui,WANG Xiaowan,et al.Soil corrosion prediction of buried pipeline based on the model of RS-PSO-GRNN[J].Materials Protection,2018,51(8):47-52,79.
[5]LI G D,WANG C H,YAMAGUCHI D,et al.A study on the corrosion process of gas pipeline applying grey dynamic model[J].International Journal of Reliability and Safety,2010,4(1):1-5.
[6]ZHANG X,ZHAO M,MANAGEMENT S O.Application of an improved grey-markov model in corrosion depth prediction of oil and gas pipeline[J].Geological Science&Technology Information,2017,36(6):286-291.
[7]谭开忍,肖熙,等.基于灰色理论的海底管道腐蚀剩余寿命预测方法[J].上海交通大学学报,2007,41(2):186-188.TAN Kairen,XIAO Xi.The forecast of remaining life of corrosive submarine pipelines based on grey theory[J].Journal of Shanghai Jiaotong University,2007,41(2):186-188.
[8]姜峰,郑运虎.预测腐蚀管道剩余寿命的新方法研究[J].机械强度,2015,37(3):539-545.JIANG Feng,ZHENG Yunhu.Study of new method for the remaining life of prediction of corrosion pipeline[J].Journal of Mechanical Strength,2015,37(3):539-545.
[9]WANG Y,DANG Y,LI Y,et al.An approach to increase prediction precision of GM(1,1)model based on optimization of the initial condition[J].Expert Systems with Applications,2010,37(8):5640-5644.
[10]MADHI M,MOHAMED N.An initial condition optimization approach for improving the prediction precision of a GM(1,1)Model[J].2017,22(1):21.
[11]DAHONG H U,AMP C S,UNIVERSITY H.Unequal Interval GM(1,1)Model Based on Optimization of Background Value and Original Condition[J].Journal of Hubei University of Arts and Science,2016,37(11):20-22.
[12]陈耀辉,李楚霖.分数阶ARIMA模型的参数估计与预测[J].系统工程,2004,22(6):87-90.CHEN Yaohui,LI Chulin.Parameter estimatiom and forecast in fractional order ARIMA models[J].Systems Engineering,2004,22(6):87-90.
[13]KURAKIM J R.KIM J.Combining forecasts using optimal combination weight and generalized autoregression[J].Journal of Forecasting,2010,27(5):419-432.
[14]乔梁,张露,许懿,等.基于最大-最小贴近度和诱导有序加权算子的风电功率短期预测模型[J].电力系统保护与控制,2014(19):114-121.QIAO Liang,ZHANG Lu,XU Yi,et al.Wind power short-term forecast model based on maximum-minimum approach degree and induced ordered weighted operator[J].Power System Protection and Control,2014(19):114-121.
[15]丁松,党耀国,徐宁,等.非等间距GM(1,1)模型性质及优化研究[J].系统工程理论与实践,2018,38(6):1575-1585.DING Song,DANG Yaoguo,XU Ning,et al.Research on properties and optimization of unequal interval GM(1,1)model[J].Systems Engineering-Theory and Practice,2018,38(6):1575-159.
[2]骆正山,车朝阳.基于TVR的腐蚀油气管道失效概率及安全寿命研究[J].中国安全生产科学技术,2018,14(9):95-99.LUO Zhengshan,CHE Chaoyang.Research on failure probability and safe life of corroded oil and gas pipelines based on TVR[J].Journal of Safety Science and Technology,2018,14(9):95-99.
[3]王晓光,周慧.输气管道腐蚀的最优灰色组合预测模型[J].腐蚀科学与防护技术,2009,21(5):496-498.WANG Xiaoguang,ZHOU Hui.An optimal grey combinatorial forecasting model for corrosion of gas pipilines[J].Corrosion Science and Protection Technology,2009,21(5):496-498.
[4]LUO Zhengshan,WANG Wenhui,WANG Xiaowan,et al.Soil corrosion prediction of buried pipeline based on the model of RS-PSO-GRNN[J].Materials Protection,2018,51(8):47-52,79.
[5]LI G D,WANG C H,YAMAGUCHI D,et al.A study on the corrosion process of gas pipeline applying grey dynamic model[J].International Journal of Reliability and Safety,2010,4(1):1-5.
[6]ZHANG X,ZHAO M,MANAGEMENT S O.Application of an improved grey-markov model in corrosion depth prediction of oil and gas pipeline[J].Geological Science&Technology Information,2017,36(6):286-291.
[7]谭开忍,肖熙,等.基于灰色理论的海底管道腐蚀剩余寿命预测方法[J].上海交通大学学报,2007,41(2):186-188.TAN Kairen,XIAO Xi.The forecast of remaining life of corrosive submarine pipelines based on grey theory[J].Journal of Shanghai Jiaotong University,2007,41(2):186-188.
[8]姜峰,郑运虎.预测腐蚀管道剩余寿命的新方法研究[J].机械强度,2015,37(3):539-545.JIANG Feng,ZHENG Yunhu.Study of new method for the remaining life of prediction of corrosion pipeline[J].Journal of Mechanical Strength,2015,37(3):539-545.
[9]WANG Y,DANG Y,LI Y,et al.An approach to increase prediction precision of GM(1,1)model based on optimization of the initial condition[J].Expert Systems with Applications,2010,37(8):5640-5644.
[10]MADHI M,MOHAMED N.An initial condition optimization approach for improving the prediction precision of a GM(1,1)Model[J].2017,22(1):21.
[11]DAHONG H U,AMP C S,UNIVERSITY H.Unequal Interval GM(1,1)Model Based on Optimization of Background Value and Original Condition[J].Journal of Hubei University of Arts and Science,2016,37(11):20-22.
[12]陈耀辉,李楚霖.分数阶ARIMA模型的参数估计与预测[J].系统工程,2004,22(6):87-90.CHEN Yaohui,LI Chulin.Parameter estimatiom and forecast in fractional order ARIMA models[J].Systems Engineering,2004,22(6):87-90.
[13]KURAKIM J R.KIM J.Combining forecasts using optimal combination weight and generalized autoregression[J].Journal of Forecasting,2010,27(5):419-432.
[14]乔梁,张露,许懿,等.基于最大-最小贴近度和诱导有序加权算子的风电功率短期预测模型[J].电力系统保护与控制,2014(19):114-121.QIAO Liang,ZHANG Lu,XU Yi,et al.Wind power short-term forecast model based on maximum-minimum approach degree and induced ordered weighted operator[J].Power System Protection and Control,2014(19):114-121.
[15]丁松,党耀国,徐宁,等.非等间距GM(1,1)模型性质及优化研究[J].系统工程理论与实践,2018,38(6):1575-1585.DING Song,DANG Yaoguo,XU Ning,et al.Research on properties and optimization of unequal interval GM(1,1)model[J].Systems Engineering-Theory and Practice,2018,38(6):1575-159.