A new approach for suppressing harmonic disturbances in distribution system based on regression analysis

Based on state variable techniques, sensitivity and regression analysis were applied to a distribution system model to study harmonic disturbances. In this research work, a framework for reducing harmonic distortion, in particular the ninth harmonic, is presented. Through measurements, a model based on harmonic distortion representation was developed. The model is represented as a load in the form of an inductor, capacitor, and resistor in parallel with a harmonic current source. With state variable representation of the model, the first- and second-order sensitivity eigenvalues along with Taylor series expansion were employed to study the ninth harmonic. Then a comprehensive analysis was done to reduce the ninth harmonic based on parameters perturbation of the system capacitors. To see the effectiveness of the perturbation analysis and compare it to other methods, the results of a conventional simulation using the Electromagnetic Transients Program (EMTP) are presented. To develop an analytical model representation free of the ninth harmonics, quadratic regression models in multivariables were used.