State Estimation And Analysis Of The Local Power Management Information System Of Provincial Scale

Local Plant Management Information System (LPMIS) of provincial scale is a platform that supervises and publishes information such as determined power by heat and cogeneration of heat and power (CHP). For the reason that the data transmitted from thermal power terminals in plants were instantly collected in the course of the communication of the data in DCS/SIS/MIS system and power production and different automation technical levels of fundamental measurement, automatization and informationization, the collected data by LPMIS or historic data in LPMIS probably includes a number of gross errors and fault information. The occurrence of errors of the uncertainly precise data not only obstructed the application and analysis of following advanced software, but also severely weakened the support for the implementation of national policies in the plants running LPMIS, which impede the establishment of credibility. Therefore, prior to the second phase of project, it is necessary to analyze historic data that accomplishes the detection and correction and accordingly carry out the improvement of data quality in plants. For increasing the quality of the data with unknown errors, there are two plans. One is to improve the original collection sources of power plant that it may increases the investment and workload. The other one is to analyze the data errors by heightening the accuracy via modern state estimations which probably reduces the investment and workload.Optimization of the historic data in LPMIS utilizing modern state estimation techniques was studied and discussed. In the beginning, the author briefed the history, prospects and applications of state estimation techniques, described the theory of state estimation techniques applied in LPMIS, basic devices in substation and systematic model of its running state, analyzed the collected data and introduced the widely used WLS (weighted least square) method in state estimation techniques.Focusing on the mathematical model of LPMIS, via analyzing the thermal power model and taking a specific plant for an instance, the author put forward the network equations and measurement equations. The equations were solved and simulated by making use of WLS. The result of long time operation of state estimation was analyzed. Furthermore, for testing bad data, the author deduced three kinds of state estimation method and proved their probability and practicability by gradual simulations.It was confirmed that from this case the precision of data was enhanced by state estimation techniques, the quantity of estimation errors was less while comparing with measurement errors and the uncertainty and dispersion were lowered. The existence of redundancy could reduce the amount of collection points from thermal power terminals, lessen the cost and heighten the utilization of hardware, in the meantime, in the collecting course, missing points were able to be obtained, bad values were probably being detected and distinguished and more advanced system applications were guaranteed.Although the state estimation has not been widely applied in thermal power systems, its application of data collection in thermal power system represents the prospects with respect to the steady theoretical basis and convenient calculations. State estimation renders possibilities to enhance the utilization and veracity and improve the supervisory quality of LPMIS and assures the stability and reliability for following advanced software. Having not tested the state estimation online, all of the present estimation was done offline. However, in the near future, the state estimation will play an important role in second phase of project LPMIS.