Font Size: a A A

Research For Power System Extended Short-term Load Forecasting In Electrical Market

Posted on:2010-10-21Degree:MasterType:Thesis
Country:ChinaCandidate:F M ZhangFull Text:PDF
GTID:2132360275482506Subject:Power system and its automation
Abstract/Summary:PDF Full Text Request
In the recent 20 years, there has been a significant reform taken place in the power industry. Although the goals and the modes of their reforms are different, their directions are almost the same: introducing competitive mechanism into the conventional monopolistic system of power industry—electricity market. The purpose of enhancing economic performance of generation, transmission and distribution of electricity can be achieved by means of deregulation and introduction of competition. Along with the construction of our nation's socialist market economic system, the market-oriented reform has been moving forward steadily.Short-term load forecasting is an important component of technical support system of electricity market, which is of vital importance to the reliable, safe and economical operation of power system. High-speed and accurate load forecasting is required in the operation, dispatch and economical load dispatching of power system. In the operation of electricity market, high-speed and accurate load forecasting is helpful to arrange the operation and the repair schedule of generating units economically and properly and is also of great help to the safety and stability of electrical network. Meanwhile, electricity power enterprises can adjust their offers properly and reduce generating cost on the base of prediction data; electricity power operation enterprises can buy electricity with a lower prize in electricity market. Therefore, rapid and accurate load forecasting can maximize the electrical power resources.In the electricity market environment, electricity department accomplishes short-term load forecasting one day ahead and confirms the generating schedule of the following day, subsequently monitors the implementation of the schedule. When there is a relatively large deviation between the original plan and actual load (larger than 3%), timely re-forecasting and adjustment of the plan of the remaining period of time during that day is required, which is called rolling generating schedule. Nowadays, short-term load forecasting focuses on the load curve forecasting one day or several days ahead. Super short-term load forecasting focuses on the prediction one hour ahead. Apparently, super short-term load forecasting and short-term load forecasting can not meet the requirement of modification of rolling generating schedule.In view of the establishment of rolling generating schedule, extended short-term load forecasting is presented in this paper. The main idea of this method is: predict the load of the moment and a few hours later using the available latest information (including load information, meteorological information and etc.). Researches have indicated that: As a result of accomplishing the load forecasting of the remaining period of time using the known information, the predictable accuracy of the proposed method is apparently higher than conventional ones. Therefore, accurate extended short-term load forecasting is fundamental to plan the rolling generating schedule. Three methods are proposed in this paper, they are as follows: 1. extended short-term load forecasting based on dynamical supported vector machine. 2. extended short-term load forecasting based on similar days weight. 3. combined extended short-term load forecasting. The simulation result is satisfactory, and can meet the demand of actual production.
Keywords/Search Tags:electricity market, extended short-term load forecasting, supported vector machine, similar days, weight, combined forecasting
PDF Full Text Request
Related items