Font Size: a A A

Study And System Realization Of Load Forecasting Considering Weather Factors

Posted on:2010-02-28Degree:MasterType:Thesis
Country:ChinaCandidate:H X ZhangFull Text:PDF
GTID:2132360275482393Subject:Power system and its automation
Abstract/Summary:PDF Full Text Request
Load forecasting is the subject deserved to be researched deeply. Many algorithms and models have appeared after a long-term study by experts. The keystone of these theories is how to get the law of object and the relation between the object and its influencing factors. Therefore we need to analyze and summarize the law of electrical load and the relation between the electrical load and its influencing factors, and carefully design the platform of electrical load forecasting in close combination with the actual local situations.This paper carefully analyzes the weather and other factors that influence electrical load change. And the paper takes the electrical load of Hunan province as an example to interpret the four step influence mode of weather factors. The paper analyzes the excellent selecting similar-days methods that consider weather influence and their difficulty, and a new algorithm of selecting similar days is proposed, the new algorithm is based on decision tree and particle swarm optimization. First the data of historical days are automatically clustered by means of decision tree, then the characteristic vector related to the electrical load change is confirmed, and the best connection weight of every factor is trained by particle swarm optimization, lastly, the best similar day can be evaluated from the outcome of decision tree by weighted Euclidean distance. After valid similar-days are selected, a novel ultra-Short term load forecasting method based on gray model is put forward. The gray model vertically forecasts an approximate electrical load sequence, and then the approximate line is improved based on the latest electrical load news. Vertical forecasting can reflects the general law of the electrical load, horizontal adjusting based on the latest electrical load trend can express the real-time load change. The method combining vertical forecasting with horizontal adjusting can make reparation for each other.This paper introduces the realization of Load Forecasting and Management system of Province and Region Integration. The paper designs the system processes, network, functional modules and database of the system, and introduces the application development framework. The system implements variety of functions with the core of load forecasting and management, therefore it can meet the need of load forecasting precision under the condition of quick evolution of power market.
Keywords/Search Tags:Power System, Load Forecasting, Similar-days, Decision tree, Particle Swarm Optimization, Gray Model, Browser/Server
PDF Full Text Request
Related items