Harmonic Separation Based on Independent Component Analysis Method

Harmonic separation plays an important role in power quality research due to the harmful impact to power system and electric apparatus. The accurate separation of harmonics is also the demand for filtering. This paper models the problem of harmonic separation as a blind source separation (BSS) task and the paper presents an algorithm called independent component analysis, which is widely used in the field of BSS. Meanwhile, the least square method is also used to solve the problem. During the harmonic separation, ICA is used to determine harmonic frequencies firstly. Then the least square method is used to determine amplitudes and phases of harmonic components. By MATLAB simulation, the result validates that harmonics can be separated effectively with a high precision by using the method proposed under the condition with low real-time demand.