Hierarchical AI Control for DER and EV Charging Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional models for integrating distributed energy resources (DERs) and electric vehicles (EVs) in power system operation become computationally expensive and intractable as the number of DERs and EVs increases, failing to effectively manage real-time control and operational constraints, leading to complex and obsolete energy management optimization models.
Innovation Solution
A hierarchical combination of artificial intelligence and optimization using a deep deterministic policy gradient (DDPG) algorithm, which identifies integrated hybrid resources (IHRs) comprising energy storage, solar generating units, and EVs, and executes a DDPG algorithm with critic and actor deep neural networks to generate a charging schedule, ensuring real-time coordinated operation and optimizing power distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional optimization models are used to integrate DERs and EVs, then operational constraints and quality of service can be maintained, but computational complexity becomes intractable as the number of resources increases
Solution Approach 1:
The patent segments the power system into multiple hierarchical levels: transmission system level, distribution system level, and resource level. Each level handles specific optimization tasks independently, with the transmission level managing overall coordination, distribution level handling local constraints, and resource level executing specific control actions. This segmentation divides the intractable global optimization problem into manageable sub-problems that can be solved efficiently at each level.
Solution Approach 2:
The patent transforms the conventional optimization approach by changing the parameter representation from detailed individual resource parameters to aggregated group parameters. Resources are grouped into categories (e.g., EV charging groups, DER clusters) and represented by aggregate variables, significantly reducing the dimensionality of the optimization problem while preserving essential operational characteristics and constraints.
2Adaptability or versatility
If the number of DERs and EVs increases to achieve higher renewable penetration, then green energy integration improves, but the optimization model becomes obsolete and computationally intractable
Solution Approach 1:
The patent introduces a temporal dimension to the optimization by implementing real-time control capabilities. The system continuously updates optimization decisions based on current system states, renewable generation forecasts, and load conditions. This dynamic approach allows the system to adapt to increasing numbers of DERs and EVs by solving simplified real-time problems rather than attempting to solve complex static problems that become intractable with scale.
Solution Approach 2:
The patent enables distributed resources and local distribution systems to make autonomous decisions through local controllers that operate independently with limited communication requirements. Each distribution system operator can manage local DERs and EVs using local information and simple control algorithms, reducing the computational burden on centralized systems and enabling scalable integration of large numbers of resources.
Data Source
AI summary
A computer system for real-time coordinated operation of power distribution systems and electric vehicles identifies a set of integrated hybrid resources (IHRs), wherein each IHR within the set of IHRs comprises one or more of: energy storage (ES) systems, solar generating units, electric vehicles (EVs), and/or inflexible loads. The computer system executes, at an IHR selected from the set of IHRs, a deep deterministic policy gradient (DDPG) algorithm, the DDPG algorithm utilizing a critic deep neural network and an actor deep neural network. The critic deep neural network estimates a Q-value of an action for a given state, and the actor deep neural network estimates a best action for the given state. Based upon an output of the DDPG algorithm, the computer system generates a charging schedule for the ES systems and the EVs within the IHR.