Wind Farm Wind Prediction Using Precomputed Flow Pattern Libraries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting wind conditions across a wind farm require extensive processing power and may not be able to provide accurate, real-time predictions due to computational limitations at the site.
Innovation Solution
A method involving the use of pre-calculated site-specific and non-site-specific wind flow and turbulence libraries, generated offline, allowing real-time prediction of wind conditions by selecting patterns from these libraries based on current weather data and modeling a site-specific wind flow field without requiring extensive on-site processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time calculations based on wind measurements and accurate site specific flow models are performed, then prediction accuracy is improved, but processing power requirements and computational time increase significantly
Solution Approach 1:
The patent pre-calculates and stores wind flow patterns and turbulence patterns in libraries before real-time operation. These pre-computed patterns capture the essential site-specific flow characteristics and turbulence behavior, allowing real-time predictions to be made by selecting and combining appropriate pre-stored patterns rather than performing full computational fluid dynamics calculations, thus dramatically reducing processing power requirements while maintaining prediction accuracy
Solution Approach 2:
The patent divides the wind flow prediction into separate components: mean wind flow patterns and turbulence patterns. Each component is pre-calculated and stored independently in respective libraries. During real-time operation, these segmented components are selected and combined based on current weather conditions, allowing the complex prediction task to be broken down into manageable parts that can be handled with limited computational resources
2Measurement precision
If real-time calculations based on accurate site specific flow models are performed, then prediction accuracy is improved, but prediction time increases making precautionary measures impossible
Solution Approach 1:
The patent performs all computationally intensive flow model calculations in advance and stores the results in pre-computed libraries. During real-time operation, the system only needs to retrieve and combine pre-stored patterns, reducing prediction time from potentially hours or days to seconds or minutes, thereby enabling timely precautionary measures while maintaining accuracy through the use of site-specific pre-calculated patterns
Solution Approach 2:
The patent pre-calculates and stores wind flow patterns and turbulence patterns in libraries before real-time operation. These pre-computed patterns capture the essential site-specific flow characteristics and turbulence behavior, allowing real-time predictions to be made by selecting and combining appropriate pre-stored patterns rather than performing full computational fluid dynamics calculations, thus dramatically reducing processing power requirements while maintaining prediction accuracy
3Productivity
If extensive processing power is available at the wind farm, then real-time accurate predictions can be made, but device complexity and infrastructure requirements increase
Solution Approach 1:
The patent shifts the computational burden from the wind farm site to an offline pre-computation phase. By pre-calculating and storing wind flow and turbulence patterns in libraries, the system enables real-time predictions using simple pattern selection and combination operations that require minimal computational infrastructure at the wind farm, thereby achieving high productivity without increasing device complexity or infrastructure requirements at the site
Data Source
Figure 1
Figure 2
AI summary
A method for real-time prediction of wind conditions across a wind farm (1) comprising a plurality of wind turbines (1), the wind farm (1) being arranged at a wind farm site, is disclosed. A first library (9) of site specific mean wind flow patterns related to the wind farm site, and a second library (10) of non-site specific turbulence patterns, are provided. Weather data is measured at a plurality of positions within the wind farm site, and based on the measured weather data, a mean wind flow pattern is selected based on the first library (9) and a turbulence pattern is selected based on the second library (10). A site specific wind flow field across the wind farm site is modelled, based on the selected mean wind flow pattern and the selected turbulence pattern, and wind conditions across the wind farm (1) are predicted, based on the site specific wind flow field.