Cascade power station mid-downstream power station warehouse-in flow prediction method

Abstract

The invention discloses a cascade power station mid-downstream power station warehouse-in flow prediction method. The method comprises the steps of selecting variable data having relatively strong correlation with the warehousing flow of the downstream power station; standardizing the data; selecting a data subset proportion alpha range, training a locally weighted linear regression model for different alpha values by adopting cross validation to obtain an error sequence, and calculating a root mean square error RMSE; selecting an alpha value enabling the RMSE to be minimum as an optimal subset proportion value; applying the optimal alpha value to local weighted linear regression to obtain a fitting value sequence of training data and a prediction value y1 of a prediction day; establishingan ARIMA (p, d, q) model for the fitting error of the training data; predicting by using a trained ARIMA model to obtain a predicted value y2; and carrying out inverse standardization processing on y1 + y2 to obtain the reservoir inflow estimated value of the downstream power station of the prediction day. The method can improve the prediction precision of the downstream power station warehouse-in flow and the accuracy of daily output plan making, reduces the number of times of power generation plan modification, and plays a certain positive role in guaranteeing the safe and stable operationof the power station.

Description

technical field [0001] The invention discloses a method for predicting the storage flow of cascade power stations in the middle and lower reaches of power stations, belonging to the technical field of flow prediction. Background technique [0002] In cascade hydropower stations, the inflow and outflow and water head of the upstream and downstream hydropower stations with close hydraulic connection affect each other, and they are usually jointly dispatched as a whole. On the premise of ensuring the safety of the cascade hub, how to give full play to the flood control, power generation, and shipping benefits of the cascade reservoir, so as to maximize the comprehensive benefits of the cascade hub, the downstream power station plays an important role. Since the daily power generation plan of the downstream power station is based on the outflow flow of the upstream power station’s daily plan, for the downstream power station with a short distance, the outflow flow of the upstrea...

AI Technical Summary

Problems solved by technology

[0003] At present, in view of the unbalanced inflow and outflow of cascade power stations, Lin Chan et al. proposed a method based on numerical analysis to correct the inbound and outflow flow. By establishing a linear regression model and using the least square method to solve it, the flow correction has been improved to a certain extent. , but the method is not a predictive model

Images