基于小波系數(shù)PCA和SaDE-ELM的電能質(zhì)量擾動(dòng)信號(hào)分類
薛正愛(ài)1,黃陳蓉2,張建德2,支昊1,顧飛1
(1 南京工程學(xué)院 電氣工程學(xué)院,江蘇 南京 211167;
2 南京工程學(xué)院 計(jì)算機(jī)工程學(xué)院,江蘇 南京 211167)
摘 要:電能質(zhì)量擾動(dòng)信號(hào)分類是電能質(zhì)量綜合治理的前提,為提高分類精度,提出一種基于主成分分析(PCA) 和自適應(yīng)差分進(jìn)化(SaDE) 優(yōu)化的極限學(xué)習(xí)機(jī)(ELM) 的電能質(zhì)量擾動(dòng)信號(hào)分類方法。對(duì) 8 種擾動(dòng)信號(hào)用 db4 小波進(jìn)行 10 層多分辨分解,與標(biāo)準(zhǔn)能量信號(hào)的能量差系數(shù)作為特征向量,PCA 對(duì)其降維處理,去除冗余特征,得到 4 維數(shù)據(jù)作為分類的樣本數(shù)據(jù)集,利用 SaDE 算法對(duì) ELM 的輸入權(quán)值和隱含層節(jié)點(diǎn)偏置優(yōu)化。通過(guò)仿真實(shí)驗(yàn)表明,提出的 SaDE-ELM 識(shí)別準(zhǔn)確率更高,抗噪性更強(qiáng),更適應(yīng)于電能質(zhì)量擾動(dòng)分類。
關(guān)鍵詞:電能質(zhì)量;多分辨分解;主成分分析;自適應(yīng)差分進(jìn)化;極限學(xué)習(xí)機(jī)
中圖分類號(hào):TM711 文獻(xiàn)標(biāo)識(shí)碼:A 文章編號(hào):1007-3175(2021)04-0006-05
Power Quality Disturbance Signal Classification Based on PCA and SaDE-ELM
XUE Zheng-ai1, HUANG Chen-rong2, ZHANG Jian-de2, ZHI Hao1, GU Fei1
(1 School of Electric Power Engineering, Nanjing Institute of Technology, Nanjing 211167, China;
2 School of Computer Engineering, Nanjing Institute of Technology, Nanjing 211167, China)
Abstract: Power quality disturbance signal classification is the premise of comprehensive power quality control. In order to improve the classification accuracy, this paper proposes a method of power quality disturbance signal classification based on principal component analysis(PCA) and adaptive differential evolution (SaDE) optimization of extreme learning machine (ELM). The 8 kinds of disturbance signals are decomposed by db4 wavelet with 10 layers of multi-resolution, and the energy difference coefficient with the standard energy signal is used as the feature vector, and PCA is used to reduce the dimensionality, redundant features are removed, and 4-dimensional data is obtained as a sample data set for classification. The SaDE algorithm is used to optimize the input weights and hidden layer node bias of ELM. Simulation experiment, the proposed SaDE-ELM has higher recognition accuracy, stronger noise resistance and it is more suitable for power quality disturbance classification.
Key words: power quality; multiresolution decomposition; principal component analysis; adaptive differential evolution; extreme learning machine
參考文獻(xiàn)
[1] 潘從茂,李鳳婷.基于小波變換的暫態(tài)電能質(zhì)量的檢測(cè)與識(shí)別[J]. 電測(cè)與儀表,2013,50(11) :69-72.
[2] 何智龍,蘇娟,覃芳. S 變換在電能質(zhì)量擾動(dòng)中的分析[J]. 電測(cè)與儀表,2015,52(22) :25-30.
[3] 占勇,程浩忠,丁屹峰,等. 基于S 變換的電能質(zhì)量擾動(dòng)支持向量機(jī)分類識(shí)別[J] . 中國(guó)電機(jī)工程學(xué)報(bào),2005,25(4) :51-56.
[4] 陳春玲,許童羽,鄭偉,等. 多類分類 SVM 在電能質(zhì)量擾動(dòng)識(shí)別中的應(yīng)用[J] . 電力系統(tǒng)保護(hù)與控制,2010,38(13) :74-78.
[5] 俞曉冬,周欒愛(ài). 基于改進(jìn) SVM 模型的電能質(zhì)量擾動(dòng)分類[J] . 電力系統(tǒng)保護(hù)與控制,2010,38(3) :15-19.
[6] KUMAR R , SINGH B , SHAHANI D T, et al. Recognition of power-quality disturbances using S-transform-based ANN classifier and rule-based decision tree[J].IEEE Transactions on Industry Applications,2015,51(2) :1249-1258.
[7] HUANG Guangbin, ZHOU Hongming, DING Xiaojian,et al.Extreme learning machine for regression and multiclass classification [J] . IEEE Transactions on Systems,2012,42(2) :513-529.
[8] 苑津莎,張利偉,王瑜,等. 基于極限學(xué)習(xí)機(jī)的變壓器故障診斷方法研究[J] . 電測(cè)與儀表,2013,50(12) :21-26.
[9] 李國(guó)華,李文悍. 基于差分進(jìn)化算法的逆變器 SHEPWM 方法的研究[J] . 電力系統(tǒng)保護(hù)與控制,2019,47(17) :32-38.
[10] 瞿合祚,劉恒,李曉明,等. 一種電能質(zhì)量多擾動(dòng)分類中特征組合優(yōu)化方法[J] . 電力自動(dòng)化設(shè)備,2017,37(3) :146-152.