Distributed Wind Farm Wake Control Using Turbine-Level PSO
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wind farm control methods fail to optimize power extraction efficiently due to wake effects, leading to suboptimal operation, long convergence times, and high computational and communication loads, especially in large wind farms with variable turbine configurations.
Innovation Solution
A decentralized control method that allows each wind turbine to perform optimization calculations independently using standard commercial equipment, with a distributed particle swarm optimization algorithm and adjustable interaction parameters to adapt to changing wind conditions and turbine configurations, reducing computational and communication loads while maintaining system resilience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If centralized control methods are used to optimize wind farm power extraction, then power extraction efficiency is improved, but computational load and communication requirements increase significantly
Solution Approach 1:
The centralized control system is segmented into decentralized control units, with each wind turbine equipped with an independent optimization module. This distributes the computational load across multiple nodes rather than concentrating it in a central system, reducing the complexity and resource requirements of each individual unit while maintaining collective optimization capability.
Solution Approach 2:
Each wind turbine performs its own optimization calculations independently using local measurements and pre-stored interaction parameters. The turbines self-configure their control strategies without requiring continuous centralized computation or communication, enabling autonomous operation that reduces both computational burden and communication infrastructure requirements.
2Ease of operation
If centralized communication networks are deployed for real-time control, then coordination between turbines is improved, but system reliability decreases due to single point of failure
Solution Approach 1:
The communication architecture is segmented from a centralized network into peer-to-peer connections between turbines. Each turbine communicates only with its immediate neighbors, creating a distributed network with no single point of failure. This segmentation maintains coordination capability while significantly improving system robustness.
Solution Approach 2:
The system changes from requiring continuous real-time communication to using pre-configured interaction parameters that are adjusted based on wind direction and turbine configuration. This parameter adaptation allows turbines to coordinate effectively with reduced communication frequency and complexity.
3Productivity
If complex optimization algorithms are used to account for wake effects, then power extraction is maximized, but convergence time increases making real-time control difficult
Solution Approach 1:
Interaction parameters characterizing wake effects between turbine pairs are pre-calculated and stored in lookup tables before operation. During real-time control, turbines simply retrieve and apply these pre-computed parameters based on current wind direction and configuration, avoiding the need for complex real-time calculations while maintaining optimization accuracy.
Solution Approach 2:
The optimization approach changes from solving complex differential equations in real-time to selecting from pre-determined parameter sets based on measured wind conditions. This parameter selection method dramatically reduces computational time while preserving the ability to account for wake effects and maximize power extraction.
4Ease of manufacture
If standard commercial equipment is used for control calculations, then cost is reduced, but computational precision and capability are limited
Solution Approach 1:
The control methodology changes from requiring high-precision real-time simulation to using pre-computed interaction parameters with sufficient accuracy for operational control. This approach allows standard commercial equipment to achieve the necessary precision by working with simplified parameter representations rather than full physical models.
Solution Approach 2:
The system uses inexpensive pre-computed parameter tables instead of expensive high-performance computing resources. These parameter sets are calculated once offline and reused repeatedly in real-time operation, providing cost-effective precision that matches the actual needs of wind farm control without requiring expensive hardware.
Data Source
Figure 1~3
Figure 4~6
Figure 5
AI summary
A wind turbine control method comprising: - determining (105) for each wind turbine a downstream group of neighboring wind turbines influenced by wake effect, an upstream group of neighboring wind turbines influential by wake effect, and a lead wind turbine from among the upstream group, - implementing a particle swarm optimization (PSO) algorithm (107) for each of the wind turbines (11i) in a distributed manner so as to obtain a control variable value (αi), - transmitting (1075, 1079) intermediate results of the algorithm to at least some of the neighboring wind turbines according to the determined groups, - receiving (1075, 1079) intermediate results from at least some of the neighboring wind turbines according to the determined groups, so that the intermediate results feed the algorithm, - generating (109) intended control signals according to the obtained control variable value (αi).