| With China’s efforts to develop renewable energy,accelerate the construction of electrified railway and urban rail transit,and large-scale construction of electric vehicle charging facilities,the problem of harmonics in the power system has become increasingly serious.At the same time,with the rapid development of China’s economy,the significant improvement of residents’ living standards,the wide use of various sophisticated equipment and the increasing attention to product quality in the production field,the requirements of power users for power quality have been further improved.Harmonics is a very important aspect of power quality problem.In order to improve power quality,control harmonics quickly and effectively,reduce harmonic pollution,improve power efficiency of power grid,and improve power environment,it is necessary to accurately estimate the harmonic current injected into harmonic source and determine the location of harmonic source.Harmonic State Estimation method is to use the data provided by the synchronous measuring equipment installed on the selected bus and line to estimate the harmonic state of the whole power network without knowing the specific information of the harmonic source,and find out the harmonic source in the system.At present,when the grid topology and system parameters are unknown,the fixed point algorithm is often used to estimate the harmonic state in the system.However,the accuracy of harmonic state estimation is not high because of the poor stability of the fixed point algorithm and the disadvantages of easily falling into local optimal.Moreover,the existing harmonic state estimation mostly ignores the influence of measurement noise or assumes that the information of the noise signal is known.However,in engineering applications,the information of noise signals is often difficult to obtain,which makes the estimation error of harmonic state in noisy environment very large.In order to solve the above problems and improve the precision of harmonic state estimation,the Dreaming Particle Swarm Independent Component Analysis algorithm is applied to the harmonic state estimation,and the precision of harmonic state estimation was improved by optimizing the separation matrix.In order to improve the estimation accuracy when the measurement noise is large,the Variable Bayesian Independent Component Analysis algorithm is applied to the harmonic state estimation,and the noise is regarded as the hidden variable in the Bayesian network,so that the harmonic state can be estimated more accurately without knowing the noise information in advance.In a multi-harmonic source system where the power grid topology and system parameters are unknown,an improved Independent Component Analysis algorithm and Mutual Information(MI)are combined to form a harmonic source location method,and the improved Independent Component Analysis algorithm is used to estimate the injected harmonic current.Then calculate the mutual information(MI)between the estimated harmonic current and the measured harmonic voltage,and find the node with the largest mutual information to determine the location of the harmonic source.The simulation results in IEEE14 node test system show that the DPSO-ICA algorithm has a higher estimation accuracy than the fixed point algorithm when ignoring the influence of noise.The VBICA algorithm can estimate the harmonic current accurately even when the measurement noise information is unknown.Harmonic source localization method combined with improved independent component analysis and mutual information can accurately determine the location of the harmonic source.Finally,the validity and reliability of the harmonic source localization method in practical engineering application are verified by the measured harmonic data of an open and close station on the outgoing line of 10.The paper contains 47 pictures,11 tables and 86 references. |
