Optimal Allocation And State Estimation Algorithm For Power System Measurement

State estimation is the basis for power flow calculation,short-circuit current calculation and stability analysis.It is the first treatment of the data obtained from the field.With the development of modern technology,more and more kinds of equipment appear,and the precision is more and more high,meanwhile,the distribution network is more complex than the transmission network in network structure,and distributed power,loop network and even micro network appeared in distribution network.Therefore,in this paper,a new type of measurement equipment PMU and the traditional measurement equipment in the loop distribution network in the hybrid configuration scheme was carried out in-depth study.The optimal configuration of the measuring device in the loop distribution network is actually to use the less cost to obtain higher accuracy state estimation result,at the same time,it is necessary to ensure the observability of the system.In this paper,the harmony search algorithm is used as the framework,and lower cost of the PMU and the traditional measurement device and the measurement precision is higher is used to the goal.The multi-objective pareto solution set is obtained,and the state estimation is obtained by the measurement.The accuracy of the estimation of the state estimation is used to measure the measurement accuracy in the objective function.In the simulation process,the IEEE33 closed contact switch is used to realize the loop distribution network.The simulation experiment is carried out on the system of closing the connection switch between 9 nodes and 15 nodes,8 nodes and 21 nodes.And the ideal configuration scheme is obtained.The state estimation has a mature algorithm framework.In static state estimation,least squares estimation is used.However,due to the non-linear relationship between the measurement vector and the state vector,the least squares estimation needs Gauss-Newton iteration,the cost is higher.It is possible to use neural networks to train a network for state estimation to solve nonlinear problems.In this paper,we use the sparse self-encoder(SAE)and the feedforward(BP)neural network to combine the parameters of the particle swarm optimization(PSO)algorithm.Finally,we get a network of state estimation based on IEEE 14.State estimation,we can not only shorter calculation time,but also get a higher accuracy of state estimation.Finally,we propose a new method for estimating the state of the neural network.