Power Plant Fleet Control Using Predictive Economic Dispatch
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Power plant operators face challenges in maximizing economic return due to the complexity of modern power plants with multiple generating units, as conventional control systems struggle to accurately predict and control performance under varying conditions, leading to inefficient operation and suboptimal utilization.
Innovation Solution
A control method and system that combines a power plant model predicting performance under varying conditions with an economic model to optimize profitability, using a plant controller that simulates operation, adjusts setpoints, and incorporates real-time data to minimize fuel costs and maximize efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional control systems are used to manage power plant generating units, then the system structure remains simple and easy to operate, but the ability to accurately predict and control performance under varying conditions deteriorates, leading to inefficient operation
Solution Approach 1:
The control system is segmented into multiple specialized modules: a performance prediction module that uses machine learning models to forecast generating unit performance, an economic dispatch module that optimizes load allocation, and a control module that executes adjustments. This segmentation allows each module to specialize in specific functions, improving overall prediction accuracy while keeping individual modules manageable in complexity.
Solution Approach 2:
An intermediary optimization system is introduced between the conventional control system and the generating units. This intermediary layer processes performance predictions, calculates optimal dispatch strategies, and generates control adjustments, thereby enhancing prediction accuracy and operational efficiency without requiring fundamental changes to the underlying conventional control infrastructure.
2Adaptability or versatility
If power plants operate with static control profiles, then the control system remains simple, but the ability to adapt to changing ambient conditions and degradation deteriorates, resulting in suboptimal utilization
Solution Approach 1:
The control system transitions from static control profiles to dynamic adaptive control. The performance prediction module continuously updates predictions based on real-time ambient conditions and degradation trends, while the optimization module dynamically adjusts dispatch strategies. This dynamic approach enables the system to adapt to changing conditions while maintaining manageable complexity through automated algorithms.
Solution Approach 2:
A feedback mechanism is implemented where actual performance data from generating units is continuously monitored and fed back to the performance prediction module. This feedback loop allows the system to learn from actual operations, refine predictions, and adjust control strategies accordingly, enhancing adaptability while using standard control system components.
3Productivity
If operators manually manage multiple generating units, then the control system remains simple, but the ability to optimize economic return deteriorates due to complexity of coordination
Solution Approach 1:
The optimization system implements self-service capabilities through automated economic dispatch algorithms that independently analyze performance predictions, market conditions, and operational constraints to determine optimal load allocation. This automation eliminates the need for manual coordination while maximizing economic return, and the modular architecture keeps system complexity manageable through standardized interfaces.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A control method (800) for optimizing an operation of a power plant fleet. The power plant fleet may include multiple operating configurations differentiated by a manner in which assets (802) are engaged. The method may include the steps of: sensing and collecting (803) measured values of the operating parameters for the operating of each of the assets; tuning asset models (805) so to configure a tuned asset model for each of the assets (802); simulating proposed operating configurations of the power plant fleet using the tuned asset models; and obtaining simulation results from each of the simulation runs, each of the simulation results including a predicted value for a performance indicator.