An optimization based technique for power system harmonic filter design

This paper presents a novel optimization design technique for power system shunt filters to ensure harmonic reduction and noise mitigation on the electrical utility grid. The proposed filter design is based on the minimization of a composite objective function depicting the three main filter objectives of minimum harmonic current penetration into the electric grid system, maximum current absorption by the harmonic filter and minimum harmonic voltage distortion at points of common coupling. In addition to these objectives, near-parallel resonance conditions on the combined system/filter configuration are to be avoided. Good knowledge of the driving-point system impedance (measured or estimated) range, or the actual fitted frequency domain transfer function, is essential for designing the most cost-effective filters to reduce dominant harmonics in the specified frequency range. Both continuous and discrete type constraints can be included in the filter parameter estimation. The min-max optimization is performed over a specified set of discrete dominant offending harmonics.