Research On Voltage Flicker Detection Method Based On Subspace Decomposition

Accompanied by rapid economic development in our country today, the power system has also been tremendous growth,power electronic devices are widely applied, and the increasing of nonlinear loads in the grid, the resulting power quality issues have been a great deal of attention of the vast number of users, how to meet user demand for high power quality has become a new problem which need to solve in the electricity sector. Voltage fluctuations and flicker as an important aspect of power quality has become one of the focus of attention, so in order to improve power quality, we must depth research and effective inhibition of voltage flicker.In this paper, subspace decomposition algorithm which come from spatial spectrum estimation is applied in the detection of voltage flicker. By decomposing the trigonometric function, flicker signal detection could be transformed into side frequency component detection, did not need to extract the voltage flicker envelope and get high precision flicker signal parameters. The high accuracy in the extraction of flicker signal parameters, as well as the great anti-noise quality of this algorithm was proved by the simulation results. The main research contents:1.Introduced the power quality and the significance of improving the power quality,and introduced the voltage flicker definition, cause and harm, establish a model for the voltage flicker and carried out a detailed analysis of this model, direct detection of flicker by using trigonometric function decomposition, do not need to extract the voltage flicker envelope to simplify the process parameters to strike.2. TAM which come from subspace decomposition algorithm is applied in the detection of voltage flicker, do not need to extract the voltage flicker envelope and get flicker signal parameters, greatly reduces the amount of computation with high accuracy and strong anti-noise. The algorithm based on the rotational invariance of the subspace, using trigonometric function decomposition principle,flicker signal detection turned to side frequency component detection,decompose the covariance matrix of the sampled signal data matrix through singular-value decompostion, according to the size of the singular values to distinguish the signal subspace and noise subspace,the principle of least square method, estimate the frequency and amplitude information signal, calculated by the formula of high precision signal parameters flicker.3. Root-MUSIC is directly applied in the estimation of voltage flicker, this algorithm is one of subspace decomposition algorithm,the core idea is use of the orthogonality of the signal subspace and noise subspace, then with the help of the orthogonality of these two subspaces, a polynomial is constructed. Strike the root of the polynomial, which are located in the unit circle contains information of the signal source,analysis of the roots and get the frequency and amplitude information of the side frequency component. Finally, flicker signal parameters with high precision can be attained through the solution of the formula. The roots obtained by polynomial will not lose signal information and the solution of the signal affected by noise is small, so this method has a strong noise immunity, while the estimation accuracy is also higher.