基于希尔伯特-小波变换与神经网络的风电接入配电网电能质量检测与辨识方法研究

Research on detection and identification method of power quality for distribution network connected with wind generator based on of hilbert-wavelet transform and neural network

提出了一种希尔伯特变换、小波变换与神经网络相结合的风电接入配电网电能质量检测与辨识方法。该方法利用希尔伯特变换对低频电能质量扰动敏感、小波变换对高频电能质量扰动和短时电能质量扰动敏感及消噪能力强、神经网络模式识别能力强等特性,通过合理的小波变换消噪、希尔伯特变换的时间幅值特性和时间频率特性以及离散小波变换的细节系数、尺度系数和模极大值的计算,可以检测出各个电能质量扰动的实时值、幅值、频率和起止时间;通过合适的神经网络,辨识出各个电能质量扰动的具体类型。该方法具有通用性强、检测信息全面、检测准确度高和辨识准确性高等优点,仿真结果验证了其有效性和可行性。

A new method of power quality detection and identification of the distribution network connected with wind generators based on Hilbert transform,wavelet transform and neural network is proposed.The Hilbert transform whose characteristics is sensitive to low-frequency power quality disturbances and wavelet transform which is sensitive to high-frequency power quality disturbances and short-term power quality disturbances and having strong noise elimination ability and neural network which has strong pattern recognition ability are used in the proposed method.The the reasonable wavelet transform is used to do noise elimination and Hilbert transform is used to product time amplitude characteristics and time frequency characteristics and the calculation of discrete wavelet transform is used to product detail coefficients,scale coefficients and modulus maxima.The real-time value,amplitude,frequency and start stop time of each power quality disturbances can be detected through the proposed method.Through the appropriate neural network,the specific types of power quality disturbances can be identified.The proposed method has the advantages of strong universality,comprehensive detection information,high detection accuracy,high identification accuracy and so on.The simulation results verify the effectiveness and feasibility of the proposed method.