反窃电系统的研究与应用
|
摘要
作为基础产业之一的供电企业,在计算机技术、通信技术和网络技术方面有了较快发展,信息技术、自动化技术得到广泛应用,但是我国目前供电企业在电量自动化管理方面,技术手段相对落后,线路损失率居高不下,窃电现象严重,经济损失巨大,供电可靠性、供电质量都有待进一步提高。以现代化信息处理系统为基础的电量管理及反窃电系统的实施,适应了供电企业提高管理水平和经济效益的迫切要求。电力企业积累了大量的历史用电数据,而绝大部分供电企业面对大量用户历史电量信息,并不能快速、有效、全面的对用电数据和窃电行为进行分析,人工神经网络即时就是满足企业这一迫切需求的强有力的工具之一。
本文在对当下反窃电现状分析的基础上,指出了窃电行为的本质所在和传统反窃电技术的不足,分析了人工神经网络的内涵和部分算法原理,提出了应用人工神经网络的反窃电模型。介绍了反窃电评价指标模型构建的原则和流程,并根据窃电的特点以及参考供电行业同业经验,选取了对窃电嫌疑系数产生影响的八个指标,完整的构建了反窃电评价指标模型体系,为应用人工神经网络对窃电情况进行分析奠定了基础。接着,分析了人工神经网络的原理和部分算法,并将其与反窃电模型相结合,分析用户正常用电和非正常用电的特征区别,得出影响用户用电区别的主要特征。最后结合实例探讨了在反窃电系统中运用人工神经网络的过程以及应该注意的问题,通过建立神经网络,可以说明利用神经网络这一工具对窃电嫌疑分析具有可行性。
The power corporations as one of the basic industries have developed very much in application of computer, communication and Internet technology. But the level of the power-automation management in our country remains low and the methods of technology management relatively drop behind. The ratio of power loss in power line has been high for a long time and the situation of power theft is serious which take great economic losing to power corporations. The reliability of power supply and the quality of power supply need to be enhanced. The application of the system of Power management and guarding against power theft based on modern information management technology satisfied the impendence need of enhancing the efficiency and quality of management and economic benefit. A large amount of basic power data and history Power data have been accumulated while many power corporations didn’t make rapid effective and general analysis on these data and the doing of power theft.
On the basis of introducing the status of the guarding against power theft, this paper analyses the shortages of traditional technology on guarding against power theft and the essence of power theft,analyses the connotation and algorithm principles of artificial neural network, bring forward a kind of model guarding against power theft bases on artificial neural network. This paper elaborates the whole process of applying artificial neural network into the system of guarding against power theft. Then it introduced the model of guarding against power theft evaluation principles and processes, and select eight impact factors of the suspect of stealing indicators according to the characteristics of tampering with the electricity supply sector and the reference of experience. It complete the construction of the guarding against power theft index model system for the application of artificial neural network analysis of the situation. Then we analyzed the principles of artificial neural networks and some algorithms, and combined with the guarding against power theft model to analyze the electrical characteristics of normal and abnormal users.
Finally, The paper discuss the processing and problems that should pay attention combined with the examples of using artificial neural networks in guarding against power theft system. It can explain the artificial neural network is feasible as a tool in guarding against power theft system through the establishment of neural networks.
本文在对当下反窃电现状分析的基础上,指出了窃电行为的本质所在和传统反窃电技术的不足,分析了人工神经网络的内涵和部分算法原理,提出了应用人工神经网络的反窃电模型。介绍了反窃电评价指标模型构建的原则和流程,并根据窃电的特点以及参考供电行业同业经验,选取了对窃电嫌疑系数产生影响的八个指标,完整的构建了反窃电评价指标模型体系,为应用人工神经网络对窃电情况进行分析奠定了基础。接着,分析了人工神经网络的原理和部分算法,并将其与反窃电模型相结合,分析用户正常用电和非正常用电的特征区别,得出影响用户用电区别的主要特征。最后结合实例探讨了在反窃电系统中运用人工神经网络的过程以及应该注意的问题,通过建立神经网络,可以说明利用神经网络这一工具对窃电嫌疑分析具有可行性。
The power corporations as one of the basic industries have developed very much in application of computer, communication and Internet technology. But the level of the power-automation management in our country remains low and the methods of technology management relatively drop behind. The ratio of power loss in power line has been high for a long time and the situation of power theft is serious which take great economic losing to power corporations. The reliability of power supply and the quality of power supply need to be enhanced. The application of the system of Power management and guarding against power theft based on modern information management technology satisfied the impendence need of enhancing the efficiency and quality of management and economic benefit. A large amount of basic power data and history Power data have been accumulated while many power corporations didn’t make rapid effective and general analysis on these data and the doing of power theft.
On the basis of introducing the status of the guarding against power theft, this paper analyses the shortages of traditional technology on guarding against power theft and the essence of power theft,analyses the connotation and algorithm principles of artificial neural network, bring forward a kind of model guarding against power theft bases on artificial neural network. This paper elaborates the whole process of applying artificial neural network into the system of guarding against power theft. Then it introduced the model of guarding against power theft evaluation principles and processes, and select eight impact factors of the suspect of stealing indicators according to the characteristics of tampering with the electricity supply sector and the reference of experience. It complete the construction of the guarding against power theft index model system for the application of artificial neural network analysis of the situation. Then we analyzed the principles of artificial neural networks and some algorithms, and combined with the guarding against power theft model to analyze the electrical characteristics of normal and abnormal users.
Finally, The paper discuss the processing and problems that should pay attention combined with the examples of using artificial neural networks in guarding against power theft system. It can explain the artificial neural network is feasible as a tool in guarding against power theft system through the establishment of neural networks.
引文
[1]王春生,彭建春,卜水红,配电网线损分析与管理系统的研制[J],中国电力,1999,32(9),48—32
[2]吴安官,倪保珊,电力系统线损[M],北京:中国电力出版社,1999
[3]迟玉军,适应电力供需形势,提升负荷管理系统实用价值[J].电力需求侧管理,2003.3:24-25.
[4]吴小美.防丢窃电的技术措施与管理[J].电测与仪表,2003.40(446):8-10.
[5]冯璐.王成文.申晓留.谭忠富.基于数据挖掘的供电企业客户关系管理系统研究与设计[J].电力信息化,2007,5(7)
[6]刘军,高亮,蔡振江等.单相全电子式防窃电电能表的研制.河北农业大学学报,2003,26(2):93一96
[7]孙洪波.2003三相电能计量系统防窃电技术研究:「硕士学位论文」.重庆:重庆大学.
[8]周爽.防窃电实用技术的研究和应用[D].辽宁:东北大学,2005.
[9]吴小美.防丢窃电的技术措施与管理[J].电测与仪表,2003.40(446):8一10.
[10]邹玉明,冯庆东,张海忱.防窃电装置在无线电负荷管理系统中的应用「J].电力自动化设备,2002.22(6):68一70.
[11]李海林.大学生综合素质评价体系与评价方法研究[D].中国石油大学(华东), 2008.
[12]刘伟,郭志忠.配电网安全性指标的研究[J].中国电机工程学报,2003,23(8):85-90.
[13]肖峻,王成山,周敏.基于区间层次分析法的城市电网规划综合评判决策[J].中国电机工程学报,2004,24(4):50-57.
[14]吴开贵,王韶,张安邦,等.基于RBF神经网络的电网可靠性评估模型研究[J].中国电机工程学报,2000,20(6):9-12.
[15]阎平凡,张长水.人工神经网络与模拟进化计算[M].第2版.北京:清华大学出版社,2005:1-2304-310.
[16]赵林明,胡浩云,魏德华,王树谦.多层前向人工神经网络.郑州:黄河水利出版社1999,69-90
[17] Ghaboussi.J,Joghataie.A.Aetiveeontrolofstrueturesusingneuralnctworks[J].Eng·Mec,1995,4(55):554-567
[18] Hoffmann A.Paradigms of Artificial Intelligence:A Methodological and Computational Analysis[M].Singapore:Springer,1998.
[19] Kung S.-Y.,Hwang J.-N.Neural networks for intelligent multimedia processing.Proceedings of the IEEE,1998[C]:1244-1272.
[20]李祚泳,彭荔红.BP网络学习能力与泛化能力满足的不确定关系式[J].中国科学,2003,33 (10):887-895.
[21]魏东.非线性系统神经网络参数预测及控制[M].北京:机械工业出版社,2008,34-38.
[22]金广厚,宋建永,程晓荣.基于模糊神经网络的电能质量指标的经济性评估.电力系统及其自动化学报,2004,16(6):18～23
[23]李洪刚,吕辉,李刚.一种BP神经网络的改进方法及应用.中国工程科学,2005,17(15):98～103
[24] Saad E.W.,Prokhorov D.V.,Wunsch D.C.Comparative Study of Stock Trend Prediction UsingTime Delay,Recurrent and Probabilistic NN[J].IEEE Transactions on Neural Networks,1998,9:1456-1469.
[25]陈刚,刘发升(2006).基于BP神经网络的数据挖掘方法.计算机与现代化134(10):20-22.
[26]韩力群.人工神经网络理论、设计及应用[M].第2版.北京:化学工业出版社,2007:47-51 65-66.
[27] Fahlman S E,Lebiere C.The cascade correlation learning rchitecture.Advances in Neural Information Processing Systems.Morgan Kaufmann,LOS Altos CA,1990,524～532.
[28] Frean M.The upstart algorithm:Amethod for constructing and trainingfeed forward neural networks.Neural Computation,1990,2(2):l98～209
[29] LiMin F.An Expert Network for DNA Sequence Analysis[J].IEEE Intelligent Systems,1999,14(1):65-71.
[30] Hornik K.Approximation capabilities of multilayer feedforward networks[J].Neural Networks,1991,4(2):251-257.
[31] Hornik K.,Stinchcombe M.,White H.Multilayer feedforward networks are universalapproximators[J].Neural Networks,1989,2(5):359-366.
[32]陈天平.神经网络及其在系统识别应用中的逼近问题[J].中国科学A辑,1994,(1):1-7.
[33]董聪.前向网络全局最优化问题研究[J].中国科学基金,1997,(01):23-29.
[34] Roberto B.First-and second-order methods for learning:between steepest descent and Newton'smethod[J].Neural Computation,1992,4(2):141-166.
[35] Ackley D.H.,Hinton G.E.,Sejnowski T.J.A learning algorithm for boltzmann machines[J]. Cognitive Science,1985,9(1):147-169.
[36] Hagan M.T.,Demuth H.B.,Beale M.H.Neural Network Design[M].Boston:MA:PWS Publishing,1996.
[37]韩哲,陈治平,熊斯毅,黎湖广,人工神经网络及其在电力短期负荷预测中的应用研究[J].科学技术与工程,2009,27(05),34-36
[38] Marin, F.J.;Garcia-Lagos,F.;Joya,G;Sandoval,F.Global model for short-term load forecasting.Electrical and Computer Engineering,2003.IEEE CCECE 2003.Canadian Conference on , Volume 3,4-7 May 2003
[39] Amjady N.Short-term hourly load forecasting using time series modeling with peak loadestimation capability[J].IEEE Trans on Power Systems,2001,16(4):795-805.
[40] Doulamis A D, Doulamis N D,Kollias S D.An Adaptable Neural-Networks Model for Recursive Nonlinear Traffic Prediction and Modeling of MPEG Video Source.IEEE Trans on Neural Networks,2003,14(1):150-166.
[41]杨光亮,乐全民,郁惟铺,王忠民.基于小波神经网络和故障录波数据的电网故障类型识别[J].中国电机工程学报,2006,26(10):99一103·
[42]王忠民,乐全明,杨光亮,郁惟铺.基于提升小波和神经网络的超高压电网故障类型识别[J].华东电力,2006,34(2):30一34·
[43]刘志刚,张友刚,钱清泉.小波网络在电力系统故障信号处理中的应用研究[J].电网技术,2003,27(4):7一10.
[44]郭付军,林军.用多个对应的后向神经网络进行同杆双回线故障识别及测距的模式[J].电网技术,2002,26(10):14一15.
[45]张举,王兴国,李志雷.小波包能量嫡神经网络在电力系统故障诊断中的应用[J].电网技术,2006,30(5):72一76.
[46]房鑫炎,钟聪,沈芳良.基于人工神经网络的自适应电流保护[J].继电器,2000,28(5):16一20.
[47]张兆宁,范春菊,郁惟铺.模糊神经网络在超高压输电线路速动保护中的应用[J].继电器,2002,30(3):39一42.
[48]章剑光,周浩,项灿芳.基于SuperSAB神经网络算法的主变压器故障诊断模型[J].电工技术学报,2004,19(7):49一53·
[49]吕干云,董立新,程浩忠.基于最小二乘加权融合集成神经网络的电力变压器故障识别[J].电网技术,2004,28(16):52一55.
[50]董立新,肖登明,吕干云,刘奕路.一种基于油中溶解气体分析的变压器绝缘故障诊断新方法[J].上海交通大学学报,2005,39(3):82一86.
[51]熊卫华,赵光宙.基于自适应高斯基表示的神经网络在电力系统故障和振荡识别中的应用[J].电网技术,2006,30(3):22一26.
[52]毛鹏,张兆宁,林湘宁,张雅明.基于小波神经网络的电力系统振荡和故障识别[J].电力系统自动化,2002,26(11):9一14.
[53]郁惟铺,李航,康建洲.基于人工神经网络的电力系统中振荡与短路模式识别的研究[J].继电器,2000,25(6):7一10.
[54]欧阳森,王建英,耿英三,宋政湘.基于小波和神经网络的电能质量辨识方法[J].电工电能新技术,2003,22(3):32一36.
[2]吴安官,倪保珊,电力系统线损[M],北京:中国电力出版社,1999
[3]迟玉军,适应电力供需形势,提升负荷管理系统实用价值[J].电力需求侧管理,2003.3:24-25.
[4]吴小美.防丢窃电的技术措施与管理[J].电测与仪表,2003.40(446):8-10.
[5]冯璐.王成文.申晓留.谭忠富.基于数据挖掘的供电企业客户关系管理系统研究与设计[J].电力信息化,2007,5(7)
[6]刘军,高亮,蔡振江等.单相全电子式防窃电电能表的研制.河北农业大学学报,2003,26(2):93一96
[7]孙洪波.2003三相电能计量系统防窃电技术研究:「硕士学位论文」.重庆:重庆大学.
[8]周爽.防窃电实用技术的研究和应用[D].辽宁:东北大学,2005.
[9]吴小美.防丢窃电的技术措施与管理[J].电测与仪表,2003.40(446):8一10.
[10]邹玉明,冯庆东,张海忱.防窃电装置在无线电负荷管理系统中的应用「J].电力自动化设备,2002.22(6):68一70.
[11]李海林.大学生综合素质评价体系与评价方法研究[D].中国石油大学(华东), 2008.
[12]刘伟,郭志忠.配电网安全性指标的研究[J].中国电机工程学报,2003,23(8):85-90.
[13]肖峻,王成山,周敏.基于区间层次分析法的城市电网规划综合评判决策[J].中国电机工程学报,2004,24(4):50-57.
[14]吴开贵,王韶,张安邦,等.基于RBF神经网络的电网可靠性评估模型研究[J].中国电机工程学报,2000,20(6):9-12.
[15]阎平凡,张长水.人工神经网络与模拟进化计算[M].第2版.北京:清华大学出版社,2005:1-2304-310.
[16]赵林明,胡浩云,魏德华,王树谦.多层前向人工神经网络.郑州:黄河水利出版社1999,69-90
[17] Ghaboussi.J,Joghataie.A.Aetiveeontrolofstrueturesusingneuralnctworks[J].Eng·Mec,1995,4(55):554-567
[18] Hoffmann A.Paradigms of Artificial Intelligence:A Methodological and Computational Analysis[M].Singapore:Springer,1998.
[19] Kung S.-Y.,Hwang J.-N.Neural networks for intelligent multimedia processing.Proceedings of the IEEE,1998[C]:1244-1272.
[20]李祚泳,彭荔红.BP网络学习能力与泛化能力满足的不确定关系式[J].中国科学,2003,33 (10):887-895.
[21]魏东.非线性系统神经网络参数预测及控制[M].北京:机械工业出版社,2008,34-38.
[22]金广厚,宋建永,程晓荣.基于模糊神经网络的电能质量指标的经济性评估.电力系统及其自动化学报,2004,16(6):18～23
[23]李洪刚,吕辉,李刚.一种BP神经网络的改进方法及应用.中国工程科学,2005,17(15):98～103
[24] Saad E.W.,Prokhorov D.V.,Wunsch D.C.Comparative Study of Stock Trend Prediction UsingTime Delay,Recurrent and Probabilistic NN[J].IEEE Transactions on Neural Networks,1998,9:1456-1469.
[25]陈刚,刘发升(2006).基于BP神经网络的数据挖掘方法.计算机与现代化134(10):20-22.
[26]韩力群.人工神经网络理论、设计及应用[M].第2版.北京:化学工业出版社,2007:47-51 65-66.
[27] Fahlman S E,Lebiere C.The cascade correlation learning rchitecture.Advances in Neural Information Processing Systems.Morgan Kaufmann,LOS Altos CA,1990,524～532.
[28] Frean M.The upstart algorithm:Amethod for constructing and trainingfeed forward neural networks.Neural Computation,1990,2(2):l98～209
[29] LiMin F.An Expert Network for DNA Sequence Analysis[J].IEEE Intelligent Systems,1999,14(1):65-71.
[30] Hornik K.Approximation capabilities of multilayer feedforward networks[J].Neural Networks,1991,4(2):251-257.
[31] Hornik K.,Stinchcombe M.,White H.Multilayer feedforward networks are universalapproximators[J].Neural Networks,1989,2(5):359-366.
[32]陈天平.神经网络及其在系统识别应用中的逼近问题[J].中国科学A辑,1994,(1):1-7.
[33]董聪.前向网络全局最优化问题研究[J].中国科学基金,1997,(01):23-29.
[34] Roberto B.First-and second-order methods for learning:between steepest descent and Newton'smethod[J].Neural Computation,1992,4(2):141-166.
[35] Ackley D.H.,Hinton G.E.,Sejnowski T.J.A learning algorithm for boltzmann machines[J]. Cognitive Science,1985,9(1):147-169.
[36] Hagan M.T.,Demuth H.B.,Beale M.H.Neural Network Design[M].Boston:MA:PWS Publishing,1996.
[37]韩哲,陈治平,熊斯毅,黎湖广,人工神经网络及其在电力短期负荷预测中的应用研究[J].科学技术与工程,2009,27(05),34-36
[38] Marin, F.J.;Garcia-Lagos,F.;Joya,G;Sandoval,F.Global model for short-term load forecasting.Electrical and Computer Engineering,2003.IEEE CCECE 2003.Canadian Conference on , Volume 3,4-7 May 2003
[39] Amjady N.Short-term hourly load forecasting using time series modeling with peak loadestimation capability[J].IEEE Trans on Power Systems,2001,16(4):795-805.
[40] Doulamis A D, Doulamis N D,Kollias S D.An Adaptable Neural-Networks Model for Recursive Nonlinear Traffic Prediction and Modeling of MPEG Video Source.IEEE Trans on Neural Networks,2003,14(1):150-166.
[41]杨光亮,乐全民,郁惟铺,王忠民.基于小波神经网络和故障录波数据的电网故障类型识别[J].中国电机工程学报,2006,26(10):99一103·
[42]王忠民,乐全明,杨光亮,郁惟铺.基于提升小波和神经网络的超高压电网故障类型识别[J].华东电力,2006,34(2):30一34·
[43]刘志刚,张友刚,钱清泉.小波网络在电力系统故障信号处理中的应用研究[J].电网技术,2003,27(4):7一10.
[44]郭付军,林军.用多个对应的后向神经网络进行同杆双回线故障识别及测距的模式[J].电网技术,2002,26(10):14一15.
[45]张举,王兴国,李志雷.小波包能量嫡神经网络在电力系统故障诊断中的应用[J].电网技术,2006,30(5):72一76.
[46]房鑫炎,钟聪,沈芳良.基于人工神经网络的自适应电流保护[J].继电器,2000,28(5):16一20.
[47]张兆宁,范春菊,郁惟铺.模糊神经网络在超高压输电线路速动保护中的应用[J].继电器,2002,30(3):39一42.
[48]章剑光,周浩,项灿芳.基于SuperSAB神经网络算法的主变压器故障诊断模型[J].电工技术学报,2004,19(7):49一53·
[49]吕干云,董立新,程浩忠.基于最小二乘加权融合集成神经网络的电力变压器故障识别[J].电网技术,2004,28(16):52一55.
[50]董立新,肖登明,吕干云,刘奕路.一种基于油中溶解气体分析的变压器绝缘故障诊断新方法[J].上海交通大学学报,2005,39(3):82一86.
[51]熊卫华,赵光宙.基于自适应高斯基表示的神经网络在电力系统故障和振荡识别中的应用[J].电网技术,2006,30(3):22一26.
[52]毛鹏,张兆宁,林湘宁,张雅明.基于小波神经网络的电力系统振荡和故障识别[J].电力系统自动化,2002,26(11):9一14.
[53]郁惟铺,李航,康建洲.基于人工神经网络的电力系统中振荡与短路模式识别的研究[J].继电器,2000,25(6):7一10.
[54]欧阳森,王建英,耿英三,宋政湘.基于小波和神经网络的电能质量辨识方法[J].电工电能新技术,2003,22(3):32一36.