面向变压器油色谱趋势预测的深度递归信念网络
|
摘要
油色谱数据及其变化趋势是评估变压器健康状态的重要依据。现有研究表明,深度信念网络(deep belief network,DBN)在油色谱数据预测领域已取得一定成果,为变压器的运行维护提供了参考。但在实际应用过程中,仍存在因网络结构限制导致油色谱时域相关性表述不充分的情况,其预测结果呈现显著的"时移"误差,从而使得基于该方法的设备状态预测结果与实际不符。针对此问题,提出了一种面向油色谱预测的深度递归信念网络算法(deepre current belief network,DRBN),该算法构建了具有时序关联特征的深度网络结构,使预测结果呈现的"时移"误差得以消除,更新了误差的迭代修正过程,使误差在网络层间和层内得以同时流动,从而提升了预测准确率。测试结果表明,文中所提出的方法可以有效克服"时移"误差,其预测准确率可达95.16%以上,为变压器的状态预测和故障预判提供了依据。
Oil chromatography data and their variation trend provide key basis for evaluation of transformer health state. Existing studies show that deep belief network(DBN) has achieved a few results in the field of oil chromatography data prediction, providing a reference for operation and maintenance of transformers. However, in practical application, there is still insufficient expression of time-domain correlation of oil chromatography due to the limitation of model structure. Obvious "time-shift" error could be observed in chromatography prediction results, making the equipment state prediction results based on this model inconsistent with actual situation. Aiming at this problem, a deep recurrent belief network(DRBN) model for transformer state prediction is proposed based on time series theory and oil chromatography data characteristics. A self-adaptive delay network with timing correlation features is constructed, able to eliminate "time-shift" error in the prediction results. The iterative correction process of the error is updated so that the error flows simultaneously between and within network layers, thereby improving prediction accuracy. Field case studies are performed, verifying that the model proposed in this paper can availably overcome the "time-shift" error, and its prediction accuracy can reach more than 95.16%.
Oil chromatography data and their variation trend provide key basis for evaluation of transformer health state. Existing studies show that deep belief network(DBN) has achieved a few results in the field of oil chromatography data prediction, providing a reference for operation and maintenance of transformers. However, in practical application, there is still insufficient expression of time-domain correlation of oil chromatography due to the limitation of model structure. Obvious "time-shift" error could be observed in chromatography prediction results, making the equipment state prediction results based on this model inconsistent with actual situation. Aiming at this problem, a deep recurrent belief network(DRBN) model for transformer state prediction is proposed based on time series theory and oil chromatography data characteristics. A self-adaptive delay network with timing correlation features is constructed, able to eliminate "time-shift" error in the prediction results. The iterative correction process of the error is updated so that the error flows simultaneously between and within network layers, thereby improving prediction accuracy. Field case studies are performed, verifying that the model proposed in this paper can availably overcome the "time-shift" error, and its prediction accuracy can reach more than 95.16%.
引文
[1]Liao R J,Zheng H B,Grzybowski S,et al.Fuzzy information granulated particle swarm optimization-support vector machine regression for the trend forecasting of dissolved gases in oil-filled transformers[J].IET Electric Power Applications,2011,5(2):230-237.
[2]王慧芳,赵婉芳,杜振东.基于寿命数据的电力变压器经济寿命预测[J].电网技术,2015,39(3):810-816.Wang Huifang,Zhao Wanfang,Du Zhendong.Economic life?}???X÷D4???x?D€?oà???o`”ˉ?@?o°oprediction of power transformers based on the lifetime data[J].Power System Technology,2015,39(3):810-816(in Chinese).
[3]赵文清,祝玲玉,高树国,等.基于多源信息融合的电力变压器故障诊断方法研究[J].电力信息与通信技术,2018,16(10):25-30.Zhao Wenqing,Zhu Lingyu,Gao Shuguo,et al.Research on fault diagnosis method of power transformer based on multi-source information fusion[J].Electric Power Information and Communication Technology,2018,16(10):25-30(in Chinese).
[4]常方圆,李玉凌,李二霞,等.基于SoC芯片的配电变压器监测终端的研究与设计[J].电力信息与通信技术,2018,16(10):60-64.Chang Fangyuan,Li Yuling,Li Erxia,et al.Research and design of distribution transformer monitoring terminal based on SoCchip[J].Electric Power Information and Communication Technology,2018,16(10):60-64(in Chinese).
[5]Duval M,dePabla A.Interpretation of gas-in-oil analysis using new IEC publication 60599 and IEC TC 10 databases[J].IEEE Electrical Insulation Magazine,2002,17(2):31-41.
[6]Hinton G E,Salakhutdinov R R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5786):504-507.
[7]Hinton G E,Osindero S,Teh Y.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554.
[8]梁智,孙国强,李虎成,等.基于VMD与PSO优化深度信念网络的短期负荷预测[J].电网技术,2018,42(2):598-606.Liang Zhi,Sun Guoqiang,Li Hucheng,et al.Short-term load forecasting based on VMD and PSO optimized deep belief network[J].Power System Technology,2018,42(2):598-606(in Chinese).
[9]黄丽.基于改进PSO优化RBF神经网络的变压器故障诊断[J].电力信息与通信技术,2018,16(9):66-72.Huang Li.Transformer fault diagnosis based on an improved PSOoptimized RBF neural network[J].Electric Power Information and Communication Technology,2018,16(9):66-72(in Chinese).
[10]张婧昕.考虑分布式发电与需求响应的电网优化调度与配网规划研究[D].杭州:浙江大学,2018.
[11]代杰杰,宋辉,杨祎,等.基于深度信念网络的变压器油中溶解气体浓度预测方法[J].电网技术,2017,41(8):2737-2742.Dai Jiejie,Song Hui,Yang Yi,et al.Concentration prediction of dissolved gases in transformer oil[J].Power System Technology,2017,41(8):2737-2742(in Chinese).
[12]张小博,王婷,秦浩,等.基于循环神经网络的95598小尺度网络流量预测[J].电力信息与通信技术,2019,17(2):9-14.Zhang Xiaobo,Wang Ting,Qin Hao,et al.A prediction method for95598 small scale network traffic based on RNN[J].Electric Power Information and Communication Technology,2019,17(2):9-14(in Chinese).
[13]Affonso E T,Rosa R L,Rodriguez D Z.Speech quality assessment over lossy transmission channels using deep belief networks[J].IEEESignal Processing Letters,2018,25(1):70-74.
[14]卞建鹏,廖瑞金,杨丽君.应用弱化缓冲算子与最小二乘支持向量机的变压器油中溶解气体浓度预测[J].电网技术,2012,36(2):195-199.Bian Jianpeng,Liao Ruijin,Yang Lijun.Concentration prediction of gases dissolved in transformer oil based on weakening buffer operator and least square support vector machine[J].Power System Technology,2012,36(2):195-199(in Chinese).
[15]肖燕彩,陈秀海,朱衡君.用改进的灰色多变量模型预测变压器油中溶解气体的浓度[J].电网技术,2006,30(10):86-89.Xiao Yancai,Chen Xiuhai,Zhu Hengjun.Prediction of gas dissolved in power transformer oil by improved multivariable grey model[J].Power System Technology,2006,30(10):86-89(in Chinese).
[16]Graves A.Supervised sequence labelling with recurrent neural networks[M].Studies in Computational Intelligence,2012:5-12.
[17]张智晟,孙雅明,王兆峰,等.优化相空间近邻点与递归神经网络融合的短期负荷预测[J].中国电机工程学报,2003,23(8):45-50.Zhang Zhisheng,Sun yaming,Wang zhaofeng,et al.A new STLFapproach based on the fusion of optimal neighbor points in phase space and the recursive neural network[J].Proceedings of the CSEE,2003,23(8):45-50(in Chinese).
[18]陈涵,邓长虹,李大路.基于循环神经网络的动态等值模型辨识[J].高电压技术,2008,34(5):1001-1004.Chen Han,Deng Changhong,Li Dalu.Recurrent neural network-based dynamic equivalencing and identification[J].High Voltage Engineering,2008,34(5):1001-1004(in Chinese).
[19]荣智海,齐波,张鹏,等.基于油中溶解气体Canopy高维模型的变压器异常状态快速识别方法[J].中国电机工程学报,2018,38(13):3987-3996.Rong Zhihai,Qi Bo,Zhang Peng,et al.Anomalous state detection of dissolved gases in transformer oil based on the canopy hyper sphere model[J].Proceedings of the CSEE,2018,38(13):3987-3996(in Chinese).
[20]范刘洋,汪可友,吴巍.多时间尺度的电力系统鲁棒调度研究[J].电网技术,2017,41(5):238-245.Fan Liuyang,Wang Keyou,Wu Wei.A study of multi-time scale robust schedule and dispatch methodology[J].Power System Technology,2017,41(5):238-245(in Chinese).
[21]杨琦,张建华,王向峰.基于小波-神经网络的风速及风力发电量预测[J].电网技术,2009,33(17):44-48.Yang Qi,Zhang Jianhua,Wang Xiangfeng.Wind speed and generated wind power forecast based on wavelet-neural network[J].Power System Technology,2009,33(17):44-48(in Chinese).
[22]于小青,马素霞,郑智聪.基于堆栈稀疏降噪自编码的电压暂降源识别方法[J].电力信息与通信技术,2018,16(11):47-52.Yu Xiaoqing,Ma Suxia,Zheng Zhicong.A voltage sag sources identification method based on stacked spars de-noising auto-encoder[J].Electric Power Information and Communication Technology,2018,16(11):47-52(in Chinese).
[23]Ho S L,Yang S,Yao Y,et al.Robust optimization using a methodology based on cross entropy methods[J].IEEE Transactions on Magnetics,2011,47(5):1286-1289.
[24]栗然,刘会兰,卢云,等.基于交叉熵理论的配电变压器寿命组合预测方法[J].电力系统保护与控制,2014,42(2):97-103.Li Ran,Liu Huilan,Lu Yun,et al.A combination method for distribution transformer life prediction based on cross entropy theory[J].Power System Protection and Control,2014,42(2):97-103(in Chinese).
[25]朱龙珠,徐宏,刘莉莉.基于深度学习的95598重大服务事件识别研究[J].电力信息与通信技术,2018,16(11):19-23.Zhu Longzhu,Xu Hong,Liu Lili,Research on recognition of 95598significant service events based on deep learning[J].Electric Power Information and Communication Technology,2018,16(11):19-23(in Chinese).
[26]齐波,张鹏,徐茹枝,等.基于分布模型的变压器差异化预警值计算方法[J].高电压技术,2016,42(7):2290-2298.Qi Bo,Zhang Peng,Xu Ruzhi,et al.Calculation method on differentiated warning value of power transformer based on distribution model[J].High Voltage Engineering,2016,42(7):2290-2298(in Chinese).
[27]Silver D,Schrittwieser J,Simonyan K,et al.Mastering the game of Go without human knowledge[J].Nature,2017,550(7676):354.
[2]王慧芳,赵婉芳,杜振东.基于寿命数据的电力变压器经济寿命预测[J].电网技术,2015,39(3):810-816.Wang Huifang,Zhao Wanfang,Du Zhendong.Economic life?}???X÷D4???x?D€?oà???o`”ˉ?@?o°oprediction of power transformers based on the lifetime data[J].Power System Technology,2015,39(3):810-816(in Chinese).
[3]赵文清,祝玲玉,高树国,等.基于多源信息融合的电力变压器故障诊断方法研究[J].电力信息与通信技术,2018,16(10):25-30.Zhao Wenqing,Zhu Lingyu,Gao Shuguo,et al.Research on fault diagnosis method of power transformer based on multi-source information fusion[J].Electric Power Information and Communication Technology,2018,16(10):25-30(in Chinese).
[4]常方圆,李玉凌,李二霞,等.基于SoC芯片的配电变压器监测终端的研究与设计[J].电力信息与通信技术,2018,16(10):60-64.Chang Fangyuan,Li Yuling,Li Erxia,et al.Research and design of distribution transformer monitoring terminal based on SoCchip[J].Electric Power Information and Communication Technology,2018,16(10):60-64(in Chinese).
[5]Duval M,dePabla A.Interpretation of gas-in-oil analysis using new IEC publication 60599 and IEC TC 10 databases[J].IEEE Electrical Insulation Magazine,2002,17(2):31-41.
[6]Hinton G E,Salakhutdinov R R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5786):504-507.
[7]Hinton G E,Osindero S,Teh Y.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554.
[8]梁智,孙国强,李虎成,等.基于VMD与PSO优化深度信念网络的短期负荷预测[J].电网技术,2018,42(2):598-606.Liang Zhi,Sun Guoqiang,Li Hucheng,et al.Short-term load forecasting based on VMD and PSO optimized deep belief network[J].Power System Technology,2018,42(2):598-606(in Chinese).
[9]黄丽.基于改进PSO优化RBF神经网络的变压器故障诊断[J].电力信息与通信技术,2018,16(9):66-72.Huang Li.Transformer fault diagnosis based on an improved PSOoptimized RBF neural network[J].Electric Power Information and Communication Technology,2018,16(9):66-72(in Chinese).
[10]张婧昕.考虑分布式发电与需求响应的电网优化调度与配网规划研究[D].杭州:浙江大学,2018.
[11]代杰杰,宋辉,杨祎,等.基于深度信念网络的变压器油中溶解气体浓度预测方法[J].电网技术,2017,41(8):2737-2742.Dai Jiejie,Song Hui,Yang Yi,et al.Concentration prediction of dissolved gases in transformer oil[J].Power System Technology,2017,41(8):2737-2742(in Chinese).
[12]张小博,王婷,秦浩,等.基于循环神经网络的95598小尺度网络流量预测[J].电力信息与通信技术,2019,17(2):9-14.Zhang Xiaobo,Wang Ting,Qin Hao,et al.A prediction method for95598 small scale network traffic based on RNN[J].Electric Power Information and Communication Technology,2019,17(2):9-14(in Chinese).
[13]Affonso E T,Rosa R L,Rodriguez D Z.Speech quality assessment over lossy transmission channels using deep belief networks[J].IEEESignal Processing Letters,2018,25(1):70-74.
[14]卞建鹏,廖瑞金,杨丽君.应用弱化缓冲算子与最小二乘支持向量机的变压器油中溶解气体浓度预测[J].电网技术,2012,36(2):195-199.Bian Jianpeng,Liao Ruijin,Yang Lijun.Concentration prediction of gases dissolved in transformer oil based on weakening buffer operator and least square support vector machine[J].Power System Technology,2012,36(2):195-199(in Chinese).
[15]肖燕彩,陈秀海,朱衡君.用改进的灰色多变量模型预测变压器油中溶解气体的浓度[J].电网技术,2006,30(10):86-89.Xiao Yancai,Chen Xiuhai,Zhu Hengjun.Prediction of gas dissolved in power transformer oil by improved multivariable grey model[J].Power System Technology,2006,30(10):86-89(in Chinese).
[16]Graves A.Supervised sequence labelling with recurrent neural networks[M].Studies in Computational Intelligence,2012:5-12.
[17]张智晟,孙雅明,王兆峰,等.优化相空间近邻点与递归神经网络融合的短期负荷预测[J].中国电机工程学报,2003,23(8):45-50.Zhang Zhisheng,Sun yaming,Wang zhaofeng,et al.A new STLFapproach based on the fusion of optimal neighbor points in phase space and the recursive neural network[J].Proceedings of the CSEE,2003,23(8):45-50(in Chinese).
[18]陈涵,邓长虹,李大路.基于循环神经网络的动态等值模型辨识[J].高电压技术,2008,34(5):1001-1004.Chen Han,Deng Changhong,Li Dalu.Recurrent neural network-based dynamic equivalencing and identification[J].High Voltage Engineering,2008,34(5):1001-1004(in Chinese).
[19]荣智海,齐波,张鹏,等.基于油中溶解气体Canopy高维模型的变压器异常状态快速识别方法[J].中国电机工程学报,2018,38(13):3987-3996.Rong Zhihai,Qi Bo,Zhang Peng,et al.Anomalous state detection of dissolved gases in transformer oil based on the canopy hyper sphere model[J].Proceedings of the CSEE,2018,38(13):3987-3996(in Chinese).
[20]范刘洋,汪可友,吴巍.多时间尺度的电力系统鲁棒调度研究[J].电网技术,2017,41(5):238-245.Fan Liuyang,Wang Keyou,Wu Wei.A study of multi-time scale robust schedule and dispatch methodology[J].Power System Technology,2017,41(5):238-245(in Chinese).
[21]杨琦,张建华,王向峰.基于小波-神经网络的风速及风力发电量预测[J].电网技术,2009,33(17):44-48.Yang Qi,Zhang Jianhua,Wang Xiangfeng.Wind speed and generated wind power forecast based on wavelet-neural network[J].Power System Technology,2009,33(17):44-48(in Chinese).
[22]于小青,马素霞,郑智聪.基于堆栈稀疏降噪自编码的电压暂降源识别方法[J].电力信息与通信技术,2018,16(11):47-52.Yu Xiaoqing,Ma Suxia,Zheng Zhicong.A voltage sag sources identification method based on stacked spars de-noising auto-encoder[J].Electric Power Information and Communication Technology,2018,16(11):47-52(in Chinese).
[23]Ho S L,Yang S,Yao Y,et al.Robust optimization using a methodology based on cross entropy methods[J].IEEE Transactions on Magnetics,2011,47(5):1286-1289.
[24]栗然,刘会兰,卢云,等.基于交叉熵理论的配电变压器寿命组合预测方法[J].电力系统保护与控制,2014,42(2):97-103.Li Ran,Liu Huilan,Lu Yun,et al.A combination method for distribution transformer life prediction based on cross entropy theory[J].Power System Protection and Control,2014,42(2):97-103(in Chinese).
[25]朱龙珠,徐宏,刘莉莉.基于深度学习的95598重大服务事件识别研究[J].电力信息与通信技术,2018,16(11):19-23.Zhu Longzhu,Xu Hong,Liu Lili,Research on recognition of 95598significant service events based on deep learning[J].Electric Power Information and Communication Technology,2018,16(11):19-23(in Chinese).
[26]齐波,张鹏,徐茹枝,等.基于分布模型的变压器差异化预警值计算方法[J].高电压技术,2016,42(7):2290-2298.Qi Bo,Zhang Peng,Xu Ruzhi,et al.Calculation method on differentiated warning value of power transformer based on distribution model[J].High Voltage Engineering,2016,42(7):2290-2298(in Chinese).
[27]Silver D,Schrittwieser J,Simonyan K,et al.Mastering the game of Go without human knowledge[J].Nature,2017,550(7676):354.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |