Predictive Optimization for Microgrid Asset Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Economic Dispatch methods for microgrids do not effectively integrate a variety of generation, load, and storage assets, including combined heat and power units, renewable generation, and controllable loads, and fail to account for fuel costs, operational costs, unit degradation, and emissions, limiting their ability to achieve global optimization and maintain system normal operation.
Innovation Solution
A predictive optimization control algorithm is used to schedule heat and power generation and consumption across microgrid assets interconnected on electric and thermal grids, optimizing an objective function based on predicted future conditions, incorporating a convex optimization formulation to ensure polynomial-time convergence and handle complex system topologies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional economic dispatch methods are used, then the system can operate with simple control algorithms, but the system fails to integrate diverse assets and achieve global optimization
Solution Approach 1:
The microgrid system is segmented into multiple functional components including controllable assets, uncontrollable assets, and a central controller. Each asset type (renewable generators, storage devices, controllable loads) is modeled separately with its own constraints and characteristics, allowing the complex integration problem to be broken down into manageable segments that can be optimized collectively through the predictive control algorithm
Solution Approach 2:
The system transitions from static conventional dispatch methods to dynamic predictive optimization that adapts to changing microgrid conditions. The controller dynamically adjusts asset scheduling based on predicted future states, incorporating time-varying constraints and objectives to achieve global optimization across diverse asset types while maintaining computational tractability
2Reliability
If conventional dispatch algorithms are used, then the computational process is simple and fast, but the system cannot maintain normal operation under arbitrary topologies and asset configurations
Solution Approach 1:
The system performs preliminary actions by predicting future microgrid states and pre-calculating optimal asset schedules before actual operation occurs. The predictive optimization algorithm computes dispatch strategies in advance based on forecasted conditions, allowing the system to maintain stability and reliability by proactively adjusting to anticipated changes rather than reactively responding to them
Solution Approach 2:
The controller implements feedback mechanisms by continuously monitoring microgrid operation and comparing actual performance against predicted outcomes. The system uses this feedback to refine future predictions and adjust asset scheduling, ensuring reliable operation under varying topologies and configurations while managing computational complexity through iterative optimization
3Adaptability or versatility
If the system integrates more asset types and topologies, then the versatility and applicability increase, but the complexity of control and optimization increases
Solution Approach 1:
The predictive optimization controller is designed as a universal platform that can handle arbitrary numbers and types of microgrid assets including renewable generators, storage devices, and controllable loads. The system uses unified mathematical models and a single optimization framework that adapts to different asset configurations and topologies, providing ease of operation despite the diversity and complexity of integrated assets
Data Source
AI summary
A method of power system dispatch control solves power system dispatch problems by integrating a larger variety of generation, load and storage assets, including without limitation, combined heat and power (CHP) units, renewable generation with forecasting, controllable loads, electric, thermal and water energy storage. The method employs a predictive algorithm to dynamically schedule different assets in order to achieve global optimization and maintain the system normal operation.