Reaearch On State Estimaiton For Distribution System Considering AMI Data And Optimal Location For AMI

In the twenty-first century, many countries engage in the construction boomof smart grid due to advances in science and human’s depending on electricpower industry. Distribution network is the bond which is connected withelectricity production and electricity consumption. So the smart distributionsystem is the key point to achieve the intelligence of electricity. AdvancedMetering Infrastructure is the important hardware foundation of smartdistribution system and it could improve the level of automation of distributionsystem comprehensively. State estimation is an effective tool for control centerto monitor the operation state of power system. However, state estimation israrely used in current distribution system because of its low level of automation.Thus, under the new situation of smart distribution system, it is meaningful toimprove the performance of distribution state estimation by considering AMI,which is a new style of measurement system. In this paper, we consider problemthat using AMI data to improve the distribution system state estimationperformance, as well as the AMI location problem. This paper includes thefollowing three aspects:First, for the feature that AMI could provide the real-time load measurementdata, it proposes a layered method for distribution system state estimation basedon AMI. This method decouples the distribution system into several subsystemsby AMI node and use the weighted least squares method to achieve thedistributed parallel computing for the state estimation. It uses an example tocompare the traditional method and layered method on the aspects of calculatingspeed and accuracy.Second, for the actual situation that distribution system has the lowmeasurement redundancy and it needs the pseudo measurements to participate inthe state estimation, it proposes a method for calculating load pseudomeasurements based on AMI. This method uses the idea of load distributioncoefficient predicting and combines the predicted coefficient and AMI real-timedata to produce load pseudo measurements. The example shows the results oftwo different methods which are traditional one and the one based on AMI.Finally, it considers the proposal for the problem of AMI location toimprove the computing speed of layered method for state estimation. Itintroduces the node computational burden to characterize the state estimation computing speed and establish an optimization problem model with an objectivefunction which is equivalent computational burden of layered method for stateestimation. The genetic algorithm is used to solve the optimization problem.Then it analyses several cases of the AMI location.