神经网络在电力系统谐波分析中的应用研究
|
摘要
近年来,由于在工业对象中大量地使用电容器组,以及在普通居民用电系统中增加大量非线性负荷和各种灵敏设备,加重了配电网络的谐波畸变程度,谐波畸变问题变得日益严重。谐波已经成为当前电力系统中影响电能质量的重要影响因素。谐波分析已经成为研究和应用电力系统的重要内容,本文就神经网络在电力系统谐波分析中的应用进行了相关研究。
本文首先介绍了谐波的成因以及谐波所产生的危害,目前的谐波治理现状及发展趋势,同时介绍了人工神经网络理论的发展史以及神经网络原理和目前的应用现状。
针对目前谐波分析方法中出现的问题,提出了基于神经网络的电力系统谐波的高精度分析方法。该方法通过对采样信号进行神经网络训练,可获得高精度的电力系统基波的频率、幅值及相位。文中详细介绍了神经网络的分析原理及算法的实现过程。
针对采样频率与实际基波频率异步的情况,提出一种基于神经网络模拟退火的谐波分析算法。该算法通过对基波频率、角频率学习率、谐波幅值及相位等相关参数进行更新,仿真结果表明,该算法对电力系统的频率和幅值及相位角的估算达到了很高精度,大大降低了训练次数,减少估算震荡,优于传统的谐波分析方法与普通的神经网络算法。
针对在谐波分析过程中出现的基波频率突变的情况,提出一种保留部分采样点,同时不断更新样本值的基于BP神经网络的频率跟踪算法。仿真结果表明,该算法在电力系统频率发生跌落、上升等状况时,均可在较短时间内完成高精度的频率跟踪,较好解决了非同步信号分析时由于频率突变所产生的问题。
In recent years, as a result of a wide application of capacitors in the industrial and a large amount of nonlinear loads and a variety of sensitive equipment application in the general population, the harmonic distortion level of distribution network is increasing fast. The harmonic distortion issue become increasingly serious. Harmonic has become an important power quality impact factors of the power system, this paper primarily studies the neural network applications in harmonic analysis area.
Firstly, this paper describes the causes of harmonic and harmonic harm, and the current situation and development trend of harmonic treatment. Also introduces the development history of artificial neural network theory and its current application status.
In order to solve the current problems existed in harmonic analysis method, proposes a kind of high precision analysis harmonics method based on neural networks for power system. By training the sampling signals using neural network method, access to high-precision power system fundamental frequency, amplitude and phase. This paper describes the analysis of neural network theory and algorithm implementation process.
A neural network method based on simulated annealing is proposed for harmonic analysis algorithm, this method is aimed at the system in which the sampling frequency cannot be locked on the actual fundamental frequency. By updating the relevant parameters of angular frequency study rate, phase and amplitude harmonic and etc, the accuracy of the estimates provided by the proposed approach in the asynchronous case is relatively better than that of the estimates obtained with the conventional harmonic analysis methods. Compared to an ordinary neural network algorithm, the method could reduce the training frequency significantly, and reduce the concussion of estimate.
In the process of harmonic analysis for the fundamental frequency of mutations appears in the situation, proposes a frequency tracking algorithm based on BP neural network. Every time, it reserve a part of the sampling points, and update the other part of the value of sampling points constantly. Simulation results shows that at the situation of the power system frequency change, it can be completed tracking the frequency of high-precision in a relatively short period of time, it could solve theproblems which caused of the frequency mutation.
本文首先介绍了谐波的成因以及谐波所产生的危害,目前的谐波治理现状及发展趋势,同时介绍了人工神经网络理论的发展史以及神经网络原理和目前的应用现状。
针对目前谐波分析方法中出现的问题,提出了基于神经网络的电力系统谐波的高精度分析方法。该方法通过对采样信号进行神经网络训练,可获得高精度的电力系统基波的频率、幅值及相位。文中详细介绍了神经网络的分析原理及算法的实现过程。
针对采样频率与实际基波频率异步的情况,提出一种基于神经网络模拟退火的谐波分析算法。该算法通过对基波频率、角频率学习率、谐波幅值及相位等相关参数进行更新,仿真结果表明,该算法对电力系统的频率和幅值及相位角的估算达到了很高精度,大大降低了训练次数,减少估算震荡,优于传统的谐波分析方法与普通的神经网络算法。
针对在谐波分析过程中出现的基波频率突变的情况,提出一种保留部分采样点,同时不断更新样本值的基于BP神经网络的频率跟踪算法。仿真结果表明,该算法在电力系统频率发生跌落、上升等状况时,均可在较短时间内完成高精度的频率跟踪,较好解决了非同步信号分析时由于频率突变所产生的问题。
In recent years, as a result of a wide application of capacitors in the industrial and a large amount of nonlinear loads and a variety of sensitive equipment application in the general population, the harmonic distortion level of distribution network is increasing fast. The harmonic distortion issue become increasingly serious. Harmonic has become an important power quality impact factors of the power system, this paper primarily studies the neural network applications in harmonic analysis area.
Firstly, this paper describes the causes of harmonic and harmonic harm, and the current situation and development trend of harmonic treatment. Also introduces the development history of artificial neural network theory and its current application status.
In order to solve the current problems existed in harmonic analysis method, proposes a kind of high precision analysis harmonics method based on neural networks for power system. By training the sampling signals using neural network method, access to high-precision power system fundamental frequency, amplitude and phase. This paper describes the analysis of neural network theory and algorithm implementation process.
A neural network method based on simulated annealing is proposed for harmonic analysis algorithm, this method is aimed at the system in which the sampling frequency cannot be locked on the actual fundamental frequency. By updating the relevant parameters of angular frequency study rate, phase and amplitude harmonic and etc, the accuracy of the estimates provided by the proposed approach in the asynchronous case is relatively better than that of the estimates obtained with the conventional harmonic analysis methods. Compared to an ordinary neural network algorithm, the method could reduce the training frequency significantly, and reduce the concussion of estimate.
In the process of harmonic analysis for the fundamental frequency of mutations appears in the situation, proposes a frequency tracking algorithm based on BP neural network. Every time, it reserve a part of the sampling points, and update the other part of the value of sampling points constantly. Simulation results shows that at the situation of the power system frequency change, it can be completed tracking the frequency of high-precision in a relatively short period of time, it could solve theproblems which caused of the frequency mutation.
引文
[1] (美)德拉罗萨(Rosa.F.C.).电力系统谐波.第1版.北京:北京机械工业出版社, 2009,21-44
[2]王兆安,杨君,刘进军.谐波抑制和无功功率补偿.北京:北京机械工业出版社, 1998,13-67
[3]毛筱.神经网络理论在谐波测量中的应用与研究, [湖南大学硕士学位论文].长沙:湖南大学,2000, 3-9
[4]裴源.基于神经网络算法的电力系统谐波分析方法,[长沙理工大学硕士学位论文].长沙:长沙理工大学,2007, 24-40
[5] (美)哈根(Hagan,M.T.).神经网络设计.第1版.北京:北京机械工业出版社, 2002,5-29
[6]邢文训,谢金星.现代优化计算方法.第2版.北京:北京清华大学出版社, 2005,77-111
[7]蒋宗礼.人工神经网络导论.第1版.北京:高等教育出版社, 2008,45-63
[8]周开利,康耀红.神经网络模型及其matlab仿真程序设计.第1版.北京:北京清华大学出版社, 2005,2-4
[9] Chang.G.W.,Chen.C.I., Liu.Y.J.et al. Measuring power system harmonics and interharmonics by an improved fast Fourier transform-based algorithm.Generation, Transmission & Distribution, IET Volume 2, Issue 2, UK,March 2008:193– 201
[10]薛惠,杨仁刚.基于FFT的高精度谐波检测算法.中国电机工程学报,2002.22(12):106-110
[11] Kashyap.S., Singh, A.K.Most. Suitable mother wavelet for measurement of power system harmonics using DWT in view of IEEE Standard 1459-2000. Harmonics and Quality of Power, ICHQP 2008. 2008:1-6.
[12]薛惠,杨仁刚.利用Morlet连续小波变换实现非整数次谐波的检测.电网技术,2002.12(26):41-44
[13]周林,夏雪,万蕴杰等.基于小波变换的谐波测量方法综述.电工技术学报,2006.9(21):67-74
[14] Hao Qian, Rongxiang Zhao, Tong Chen. Interharmonics Analysis Based on Interpolating Windowed FFT Algorithm. Power Delivery, IEEE Transactions onVolume 22, Issue 2, 2007.4:1064–1069
[15]黄纯,江亚群.谐波分析的加窗插值改进算法.中国电机工程学报,2005,25(15):26-32
[16] Lin H C. Intelligent neural network based dynamic power system harmonic analysis. IEEE, 2004: 244-248
[17] Manuele B, Giuseppe B, Simone C, et al. Real-time estimation of power system frequency by neural network. IEEE, 2003:87-92
[18]李红,马新瑜.多层前馈神经网络在电力系统谐波测量中的应用.电测与仪表, 2003, 40(2):15-17
[19]王萍,邹宇,郭翠双.基于Hopfield神经网络的谐波电流检测方法.天津大学学报,2007, 12 (40):1431-1435
[20]柴旭峥,文习山,关根志等.一种高精度的电力系统谐波分析算法.中国电机工程学报, 2003,23(9):67-70
[21]肖雁鸿,毛筱,罗瑞琼等.神经网络理论在谐波测量中的应用.电工技术学报,2002,17(2):101-104.
[22]任永峰,李含善,胡洪涛等.基于多层前馈神经网络的并联型电能质量控制器,电工技术学报,2007.8(22):108-113
[23]吕征宇,钱照明, T C Green.并联有源电力滤波器的神经网络预测控制.中国电机工程学报, 1999,19(12):22-26
[24]王群,吴宁,王兆安等.一种基于人工神经网络的电流谐波测量方法.电力系统自动化,1998,22(11):35-39
[25]李红,马新瑜.多层前馈神经网络在电力系统谐波测量中的应用.电测与仪表, 2003.2(446):15-17
[26] Manuele B, Giuseppe B, Simone C,et al. Real-time estimation of power system frequency by neural network. IEEE,2003:87-92
[27] Nguyen T T. Parametric harmonic analysis of power systems. IEEE,1997, 144(1):21-25
[28] EI-Amin I, Arafah I. Artificial neural networks for power systems harmonic estimation. The 8th international conference on harmonics and quality of power, 1998,2: 999-1009
[29] Serno J A. On the use of amplitude shaping pulses as windows for harmonic analysis,IEEE Trans. on Instrumentation and Measurement, 2001,50(6):1556-1562
[30]王群,吴宁,王兆安.一种基于人工神经网络的谐波测量方法.电网技术,1999,23(1):29-32
[31]什么是电力系统频率. www.kingcable.cn/nettest ,2007-11-5
[32]李瑞福,王秀丽.电力谐波的分类、危害及其治理.锦州师范学院学报(自然科学版), 2003,3(24):79-81
[33] S.Kirkpatrick, C.D.gelatt, Jr.,M.P.Vecchi. Optimization by Simulated Annealing. Science,vol.220,1983:671-80
[34] (美)哈姆,科斯塔尼克著.神经计算原理.第1版.北京:北京机械工业出版社, 2007,71-87
[35]谢小荣,韩英铎.电力系统频率测量综述.电力系统自动化, 1999,23(3):54-58
[36] Gallo D, Langella R, Testa A. Desynchronized processing technique for harmonic and interharmonic analysis. IEEE Trans. on Power Delivery,2004,19(3): 993-1001
[37] Testa A, Gallo D,Langella R. Using hanning window in standard framework, IEEE Trans. on Power Delivery, 2004,19(1):28-34
[38]曹晖,袁世英,易珺等.准同步采样法在配电综合测控仪中的应用.继电器, 2005,33(11):45-48
[39]路文喜,梁志瑞,高振弟等.改进双正交滤波器组的电网频率跟踪算法.电网与清洁能源, 2009.6(25):32-35
[40]甘云华,金龙,胡敏强等.基于锁相环的超声波电机频率跟踪控制技术.哈尔滨理工大学学报, 2009.2(14):66-70
[41]徐茂格,宋耀良.一种稳健的混沌调频信号频率跟踪技术,电子与信息学报, 2009.1(31):104-107
[2]王兆安,杨君,刘进军.谐波抑制和无功功率补偿.北京:北京机械工业出版社, 1998,13-67
[3]毛筱.神经网络理论在谐波测量中的应用与研究, [湖南大学硕士学位论文].长沙:湖南大学,2000, 3-9
[4]裴源.基于神经网络算法的电力系统谐波分析方法,[长沙理工大学硕士学位论文].长沙:长沙理工大学,2007, 24-40
[5] (美)哈根(Hagan,M.T.).神经网络设计.第1版.北京:北京机械工业出版社, 2002,5-29
[6]邢文训,谢金星.现代优化计算方法.第2版.北京:北京清华大学出版社, 2005,77-111
[7]蒋宗礼.人工神经网络导论.第1版.北京:高等教育出版社, 2008,45-63
[8]周开利,康耀红.神经网络模型及其matlab仿真程序设计.第1版.北京:北京清华大学出版社, 2005,2-4
[9] Chang.G.W.,Chen.C.I., Liu.Y.J.et al. Measuring power system harmonics and interharmonics by an improved fast Fourier transform-based algorithm.Generation, Transmission & Distribution, IET Volume 2, Issue 2, UK,March 2008:193– 201
[10]薛惠,杨仁刚.基于FFT的高精度谐波检测算法.中国电机工程学报,2002.22(12):106-110
[11] Kashyap.S., Singh, A.K.Most. Suitable mother wavelet for measurement of power system harmonics using DWT in view of IEEE Standard 1459-2000. Harmonics and Quality of Power, ICHQP 2008. 2008:1-6.
[12]薛惠,杨仁刚.利用Morlet连续小波变换实现非整数次谐波的检测.电网技术,2002.12(26):41-44
[13]周林,夏雪,万蕴杰等.基于小波变换的谐波测量方法综述.电工技术学报,2006.9(21):67-74
[14] Hao Qian, Rongxiang Zhao, Tong Chen. Interharmonics Analysis Based on Interpolating Windowed FFT Algorithm. Power Delivery, IEEE Transactions onVolume 22, Issue 2, 2007.4:1064–1069
[15]黄纯,江亚群.谐波分析的加窗插值改进算法.中国电机工程学报,2005,25(15):26-32
[16] Lin H C. Intelligent neural network based dynamic power system harmonic analysis. IEEE, 2004: 244-248
[17] Manuele B, Giuseppe B, Simone C, et al. Real-time estimation of power system frequency by neural network. IEEE, 2003:87-92
[18]李红,马新瑜.多层前馈神经网络在电力系统谐波测量中的应用.电测与仪表, 2003, 40(2):15-17
[19]王萍,邹宇,郭翠双.基于Hopfield神经网络的谐波电流检测方法.天津大学学报,2007, 12 (40):1431-1435
[20]柴旭峥,文习山,关根志等.一种高精度的电力系统谐波分析算法.中国电机工程学报, 2003,23(9):67-70
[21]肖雁鸿,毛筱,罗瑞琼等.神经网络理论在谐波测量中的应用.电工技术学报,2002,17(2):101-104.
[22]任永峰,李含善,胡洪涛等.基于多层前馈神经网络的并联型电能质量控制器,电工技术学报,2007.8(22):108-113
[23]吕征宇,钱照明, T C Green.并联有源电力滤波器的神经网络预测控制.中国电机工程学报, 1999,19(12):22-26
[24]王群,吴宁,王兆安等.一种基于人工神经网络的电流谐波测量方法.电力系统自动化,1998,22(11):35-39
[25]李红,马新瑜.多层前馈神经网络在电力系统谐波测量中的应用.电测与仪表, 2003.2(446):15-17
[26] Manuele B, Giuseppe B, Simone C,et al. Real-time estimation of power system frequency by neural network. IEEE,2003:87-92
[27] Nguyen T T. Parametric harmonic analysis of power systems. IEEE,1997, 144(1):21-25
[28] EI-Amin I, Arafah I. Artificial neural networks for power systems harmonic estimation. The 8th international conference on harmonics and quality of power, 1998,2: 999-1009
[29] Serno J A. On the use of amplitude shaping pulses as windows for harmonic analysis,IEEE Trans. on Instrumentation and Measurement, 2001,50(6):1556-1562
[30]王群,吴宁,王兆安.一种基于人工神经网络的谐波测量方法.电网技术,1999,23(1):29-32
[31]什么是电力系统频率. www.kingcable.cn/nettest ,2007-11-5
[32]李瑞福,王秀丽.电力谐波的分类、危害及其治理.锦州师范学院学报(自然科学版), 2003,3(24):79-81
[33] S.Kirkpatrick, C.D.gelatt, Jr.,M.P.Vecchi. Optimization by Simulated Annealing. Science,vol.220,1983:671-80
[34] (美)哈姆,科斯塔尼克著.神经计算原理.第1版.北京:北京机械工业出版社, 2007,71-87
[35]谢小荣,韩英铎.电力系统频率测量综述.电力系统自动化, 1999,23(3):54-58
[36] Gallo D, Langella R, Testa A. Desynchronized processing technique for harmonic and interharmonic analysis. IEEE Trans. on Power Delivery,2004,19(3): 993-1001
[37] Testa A, Gallo D,Langella R. Using hanning window in standard framework, IEEE Trans. on Power Delivery, 2004,19(1):28-34
[38]曹晖,袁世英,易珺等.准同步采样法在配电综合测控仪中的应用.继电器, 2005,33(11):45-48
[39]路文喜,梁志瑞,高振弟等.改进双正交滤波器组的电网频率跟踪算法.电网与清洁能源, 2009.6(25):32-35
[40]甘云华,金龙,胡敏强等.基于锁相环的超声波电机频率跟踪控制技术.哈尔滨理工大学学报, 2009.2(14):66-70
[41]徐茂格,宋耀良.一种稳健的混沌调频信号频率跟踪技术,电子与信息学报, 2009.1(31):104-107