A mountainous wind power prediction method and device based on terrain physical constraints

By constructing a terrain physical feature field and a target physical association structure, and combining a multi-head attention mechanism and a temporal convolutional network, the problems of accuracy and efficiency in wind speed and power prediction in mountainous wind farms are solved, achieving high-precision and low-cost wind power prediction.

CN122390166APending Publication Date: 2026-07-14BEIJING EAST ENVIRONMENT ENERGY TECH +6
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING EAST ENVIRONMENT ENERGY TECH
Filing Date
2026-06-17
Publication Date
2026-07-14

AI Technical Summary

Technical Problem

Existing technologies struggle to achieve high-precision and efficient calculations for wind speed and power prediction in mountainous wind farms with complex terrain. Traditional methods either fail to incorporate terrain physical constraints or have high computational complexity, making it difficult to balance prediction accuracy with engineering practicality.

Method used

A method for predicting wind power in mountainous areas based on topographic physical constraints is constructed. By integrating topographic data, meteorological data, and wind turbine operation data, a topographic physical feature field is generated. The spatial influence relationship between wind turbines is determined using slope aspect parameters, topographic flow tube functions, and pressure gradient coefficients. The prediction is performed by combining a multi-head attention mechanism and a temporal convolutional network, and physical constraint correction is applied.

Benefits of technology

It achieves high prediction accuracy and low computational cost in complex mountainous scenarios, balancing prediction accuracy and engineering practicality, and improving the accuracy and efficiency of wind power prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122390166A_ABST
    Figure CN122390166A_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of power prediction, and discloses a mountainous wind power prediction method and device based on terrain physical constraints. The method comprises the following steps: firstly, a terrain physical characteristic field is constructed by fusing terrain data, meteorological data and wind turbine operation data; secondly, complex physical mechanisms such as mountain obstruction, canyon acceleration, flow around and terrain shear are quantified into the model, so as to make up for the defects of the traditional statistical method, such as lack of terrain physical constraints and poor generalization; thirdly, the wind turbine is taken as a node, and the target physical correlation structure is generated by combining the dominant wind direction, the slope direction parameter of the position of the wind turbine, the terrain flow pipe function and the pressure gradient coefficient, so as to depict the spatial influence relationship of the wind turbine under the irregular distribution of the mountainous wind farm and to fuse the space-time correlation information of the wind turbine; finally, the wind power prediction is carried out based on the physically reasonable topological structure, so that the model has the rigor of the physical mechanism and the high efficiency of the data driving at the same time, and the problem that the existing technology cannot balance the prediction accuracy, the calculation efficiency and the engineering practicability is effectively solved.
Need to check novelty before this filing date? Find Prior Art