Artificial Intelligence (AI) Based Power System State Estimtion

Abstract

For economic dispatch, load frequency control and to control the parameter of power system, data are acquired from the monitoring equipments or different metering de- vice used in power system. Before any control action and security assessment carried out, reliable estimate of the existing state of the system has to be determined. How- ever gross error in any of data causes the result to become useless data, which are analog or digital quantity measured by sensors. Analog measurements passes through analog to digital convertor which contains inaccuracies and free random errors like noise plus some unavoidable errors from instrumentation transformers. These errors ought to be quantified in statical sense. Best estimation are chosen which gives least sum. By using some modern technique like Artificial Intelligence more accurate data can be achievable. In this theses at First place as conventional method Weighted Least Square method is utilised to check and compare the result with standard result and afterwards Particle Swarm Optimisation as an advance technique is applied for power system state estimation in steady state condition. Adopted methods are sub- jected to WSCC 3-Machine 9-Bus standard test system to check effectiveness of each method. From the results of adopted methods it is discernible that Particle Swarm Optimization approach would be better option than conventional method as it has less computational process which leads to lesser consumption of time plus yields ac- curate result. Moreover power system state estimator has four basic operations: (1) Hypothesize structure, (2) Estimate, (3) Detect, (4) Identify. After extrapolating the state it is inevitable to detect and identify the errors in estimated results in order to have reliable data. Subsequently in this theses bad data detection and identification has been carried out by Chi-square method and Largest Normalized Residual(LNR) method. Both the methods are subjected to WSCC 3-Machine 9-Bus standard test system to detect and identify the errors from state estimation results and to check effectiveness of both methods.