Adaptive EV Charging Control via Dynamic Policy Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing power and energy systems face challenges in coordinating electric vehicle charging due to uncertainties in behavioral patterns and fleet routing decisions, leading to limited scalability and accuracy in model-based and data-driven control mechanisms.
Innovation Solution
A system and method for dynamically selecting and adapting control policies using a trained agent selection policy that evaluates and deploys the best-performing control agent based on real-time data, incorporating aggregation and clustering to manage electric vehicle charging systems effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If model-based control approaches are used to coordinate electric vehicle charging demand, then control precision can be improved, but device complexity increases due to the need for accurate models of heterogeneous user behavior and vehicle arrival patterns
Solution Approach 1:
The patent segments the control problem by dividing the fleet into multiple clusters based on behavioral similarities, and further segments the solution space by creating multiple control agents (model-based, data-driven, hybrid) that can be selectively applied to different clusters or situations. This segmentation reduces the overall complexity by breaking down the heterogeneous problem into more homogeneous sub-problems.
Solution Approach 2:
The patent implements dynamic selection of control agents based on real-time conditions. The system dynamically determines which control agent to apply to each cluster based on factors like cluster characteristics, available data, and computational resources. This dynamic approach allows the system to adapt its complexity level to match the specific requirements of each situation.
2Adaptability or versatility
If data-driven control approaches are used to coordinate electric vehicle charging demand, then adaptability to real-world conditions can be improved, but device complexity increases due to the scalability challenges of state-action spaces
Solution Approach 1:
The patent applies segmentation by clustering vehicles into groups with similar behavioral patterns, which reduces the effective state-action space that each data-driven control agent needs to handle. Instead of learning from all possible states in the entire fleet, each agent learns from a subset of similar states, making the problem scalable.
Solution Approach 2:
The patent introduces clustering as an intermediary step between raw data and control decision-making. The clustering process transforms the high-dimensional, heterogeneous data into grouped representations that are more manageable for data-driven control agents, reducing computational complexity while preserving essential patterns.
3Adaptability or versatility
If multiple control agents are maintained to handle different scenarios, then adaptability can be improved, but device complexity increases due to the need to manage and select among multiple agents
Solution Approach 1:
The patent segments the control agent selection process by assigning different agents to different clusters based on their characteristics. This segmentation allows the system to maintain multiple specialized agents without requiring complex global selection logic, as each cluster can be independently managed with its appropriate control agent.
Solution Approach 2:
The patent performs preliminary classification of vehicle clusters before selecting control agents. By pre-characterizing clusters based on behavioral patterns and matching them with suitable control agent types, the system avoids complex real-time decision-making about agent selection, reducing operational complexity.
Data Source
AI summary
Systems and methods are provided for dynamically selecting a control policy from among several available control policies for controlling an energy system having multiple controllable assets. The performance of the selected control policy is monitored and a different control policy may be deployed in its place if the different control policy has a higher chance of providing better performance given the current control environment. Thus, as the control environment changes, the control policy that controls the power system may also be changed in an adaptive manner. In this way, the control policies may be changed as the control environment changes to provide an improved real-time performance compared to the use of a single control policy.