Distribution Network Three Phase State Estimation Base On The Innovation Graph

The distribution network which is facing to the terminal user is the last linkof the power transmission, the strength of whose the monitoring of infectioncontrol affects the level of electricity reliability and power quality assessment.Among different kind of monitoring strategies, three phase state estimation in thedistribution network, which has remarkable charateristics in economy, reliabilityand practicality, is talent showing itself. However, distribution network threephase state estimation one of the bottlenecks is the ability of identifying theanomalies was not strong, that is one of the bottlenecks in the development ofdistribution network three phase state estimation. The theory of innovation graphhas been used in transmission network and single phase distribution networksuccessfully, based on above, this paper put forward that innovation graph appliedto unbalanced three phase distribution network.First of all, the modeling of distribution network based on innovation graph isrelying on the characteristics of innovation graph only meeting kirchhoff’s currentlaw. If the potential of the neutral point in the equivalent circuit of the star is notzero, then voltage source replaced the neutral point. Put all the neutral points andthe branches whit voltage source into generalized node to solve the problem ofphase coupling in three phase state estimation which is decoupled into threeindependent single phase innovation graph system. Experiments show that reallycan take advantage of the data in one of the phase to identify the anomalies.According to the need of forming the innovation network in distribution network,different potential node and the ground node is connected to support as thegrounding tree branch.Thus, the formation of the model of distribution networkbased on innovation graph is completed.After fomating the model of distribution network based on innovation graph,according to each of calculation formula for innovation in innovation graph,calculate the innovation vector, the link reckoning innovation vector andinnovation difference. On this basis, even support the link reckoning innovationvector new calculation method, according to the terminal node load even supportthe link reckoning innovation vector to tree branch innovation vector, making themodel set of graphics and data as a whole, and the analysis whit data and graphicsbetween each of innovation.Secondly, under the dual support of the model and the innovation data, withthe identification of bad data and sudden load in distribution network the firstplace, respectively, summarize and generalize the principle in identifying single anomaly. Due to the presence of bad data, the innovation vector is far away fromzero in the innovation network, that will affect the identification of sudden laodchange. Therefore, this paper uses the analysis of examples on bad data on the linkbranch and sudden load change to explain the difference, which greatly improvesthe ability to identify bad data and sudden load change.Finally, excluding with the impact of bad data and sudden load change indistribution network, tap the characteristics of topology errors and propose theprinciple. At the same time, the method can identify topology errors with bad data.