A Forecast Load Correction Method in Real-time Dispatch Planning

Abstract

The invention provides a prediction load correction method in a real-time scheduling plan. The method comprises the steps that a prediction load time point is randomly extracted during a historical operation time; the actual load value of a sample corresponding time is acquired according to the prediction load time point, and a prediction error, prediction error expectation and standard deviation are calculated; whether the error expectation mu of the sample is zero is checked, and if mu is not zero, the prediction error expectation is used instead; random sample drawing is carried out, and the prediction error, the error expectation and the standard deviation are calculated for a certain time; whether the sample error expectation mu j is equal to mu, and if mu j is not equal to mu, the sample expectation of the error value is used instead; the full-day time error expectation mu j is acquired from the previous two steps; real-time scheduling plan, the corresponding error expectation mu j is selected according to the current prediction load time point, and the prediction load value is corrected. According to the method, the load demand side is considered; the prediction value is corrected; and the real-time scheduling plan can effectively balance power generation and utilization.

Description

technical field [0001] The invention relates to a forecast load correction method, in particular to a forecast load correction method in a real-time dispatch plan. Background technique [0002] When the power dispatching control center formulates a real-time dispatching plan, it calculates the real-time power generation plan based on the predicted load value for the next five minutes, so that some generating units can track the planning mode, and can meet the requirements of "three public" dispatching and intraday dispatching operations. As far as possible to achieve the balance of power generation and power consumption in the power grid, maintain the frequency stability of the power system, ensure the safe and stable operation of the power grid, and provide safe, reliable and high-quality power supply to power users. [0003] Usually, when the real-time power generation plan uses the predicted load, the predicted load error characteristics caused by the forecast method are ...

AI Technical Summary

Problems solved by technology

[0003] Usually, when the real-time power generation plan uses the predicted load, the predicted load error characteristics caused by the forecast method are not considered, and the predicted value is directly used. The calculated total power that should be allocated by the real-time power generation plan is not reasonable enough

[0004] On the one hand, due to the characteristics of the algorithm itself, the prediction algorithm adopted by each dispatch center in the real-time dispatch plan may make the expectation of the prediction error not zero, so that the forecast load is always too large or too small; on the other hand, due to The load curve shows the characteristics of rising, falling, peak, and valley according to different time periods. Even when the same forecasting algorithm predicts these different characteristic periods, its error expectations may be different. More importantly, the load curves of different dispatch centers have different characteristics. Different characteristics, and different forecasting algorithms are used, which makes the error of forecasted load in different periods more complex and changeable

Images