The Research Of Harmonic Sources Location Method Based On Compressed Sensing

With the increase of nonlinear elements and nonlinear loads in power system.The damage caused by the power harmonic injection into the system becomes more and more serious,which not only endanger the safe use of electricity users,but also causes a great impact on the economic operation of the power grid.Therefore,there is an urgent need for a precise harmonic source localization method to find out the distribution and the state of the harmonics,to provide the basis for the management of harmonic sources.Before the location of harmonic source,we need to obtain the power system state estimation,However,considering economic factors,it is not possible to install the harmonic measurement device on each bus and line in the system,Therefore,the need for a reasonable allocation of the system to achieve considerable of harmonic measurement device.This paper mainly analyzes the measurement configuration based on 0-1 configuration method and the binary particle swarm optimization algorithm,and the results are analyzed in the IEEE123 system.The observability measure configuration from the 36 node to the 15 node is successful implementated.The general measurement schemes can only generate a single solution,but this method can also generate a variety of the equivalent observability measurement scheme,which has very strong adaptability in practical engineering application.In the premise of achieving considerable power system,the harmonic source location problem is transformed to the problem of estimating injected harmonic current into the system after obtaining the small phasor measurement and measurement matrix.Since the traditional least square method can't solve this kind of problem in the measurement equation of harmonic source location,and the space of the distribution of harmonic sources in power system is sparse,this paper uses compressed sensing domain algorithm to achieve precise harmonic sources location.The branch harmonic current which is not liable to be affected by the line harmonic impedance parameters is taken as the measurement variable,and the injected harmonic current as the state variable(the size of the load harmonic current injected to the system).The gradient projection method of compressed sensing is used to convert the l0 norm in harmonic source localization in the mathematical model into the l1 norm to locate the harmonic source under the condition of underdetemined measurement.The effectiveness of-the method is tested in the IEEE 123 bus system,and the results are compared with the result calculated by least square method.The simulation results show that the proposed method can accurately locate the harmonic source,which has the advantages of good adaptability and strong anti-interference ability.The node harmonic voltage which is not liable to be influenced is taken as the measurement variable,and the injected harmonic current as the state variable.The smoothing l0 norm(the approximate mathematical model of harmonic source location l0 norm smoothing)of the compressed sensing algorithm is applied,and the concept of support set is introduced to locate the harmonic source under the condition of underdetermined measurement.The effectiveness of the algorithm is verified in the IEEE 123 node system,and the results are compared with that calculated by least square method.From the analysis of simulation results,we can see that this algorithm can accurate locate the harmonic sources in different orders of harmonic source and under strong noise condition.Thus effectiveness and accuracy of the algorithm is verified.