Edge Computing Multi-Agent Load Regulation for Micro-Grids
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current load management systems in intelligent micro-grid environments lack effective methods for balancing and regulating distributed power supply and load, particularly in scenarios involving multiple energy sources like wind, photovoltaic, and flexible loads, and fail to consider network losses and coupling relationships between regional power grids.
Innovation Solution
A multi-agent load regulating and controlling method based on edge computing, which divides the power grid into regional grids connected through transformer substations, builds data models for thermal, wind, and photovoltaic power generation, and flexible loads, and uses edge computing to perceive voltage and phase angles across tie-lines, optimizing economic benefits and power balance through an augmented Lagrange method.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a completely-distributed computing framework is used for power supply and load regulation, then autonomy and reactivity of individual agents are improved, but implementation feasibility based on actual edge computing technological means deteriorates
Solution Approach 1:
The patent segments the power grid into multiple regional grids, each with its own edge computing node and multi-agent system. This segmentation allows each region to operate autonomously while maintaining overall system coordination, making the complex distributed control framework implementable through modular edge computing devices at each substation.
Solution Approach 2:
The patent introduces edge computing nodes as intermediaries between the centralized control system and distributed agents. These nodes run local multi-agent systems that mediate between global optimization goals and local implementation constraints, bridging the gap between theoretical distributed control and practical edge computing capabilities.
2Device complexity
If distributed control is applied to single category of distributed power supply or load, then control simplicity is improved, but synergistic regulation and control of controllable load and distributed power supply deteriorates
Solution Approach 1:
The patent designs a universal multi-agent framework that can handle multiple types of distributed resources (wind, photovoltaic, thermal power) and controllable loads simultaneously. The same agent architecture and control methodology apply to different resource types, enabling synergistic regulation while maintaining relatively simple implementation through standardized protocols and models.
3Ease of operation
If regional power grids are divided and connected through transformer substations, then local regulation capability is improved, but system complexity and coupling relationships between regions worsen
Solution Approach 1:
The patent divides the power grid into regional grids centered around transformer substations, with each region having independent edge computing nodes and multi-agent systems. This segmentation enables local autonomous regulation while the overall system complexity is managed through modular architecture and standardized interconnection protocols.
Solution Approach 2:
The patent uses parameter optimization techniques to manage the coupling between regional grids. By adjusting control parameters and operating points of distributed resources and loads, the system achieves coordinated regulation across regions without requiring complex real-time optimization of all interconnections simultaneously.
4Productivity
If new energy generation resources and flexible loads are accessed in large scale, then power grid regulating and controlling capacity is improved, but system stability and reliability worsen
Solution Approach 1:
The patent implements multi-layer feedback mechanisms in the multi-agent system, including local feedback at each agent level, regional feedback through edge computing nodes, and global feedback through inter-region coordination. These feedback loops enable the system to maintain stability and reliability by continuously adjusting to changes in new energy generation and flexible load conditions.
Solution Approach 2:
The patent employs predictive control and constraint management in the multi-agent optimization framework to cushion against potential instability. By anticipating the impact of large-scale new energy and flexible load integration, the system pre-adjusts operating parameters and maintains reserve capacities to prevent stability issues before they occur.
Data Source
Figure 1~2
AI summary
The invention provides a multi-agent load regulating and controlling method based on edge computing, and relates to the technical field of load management under intelligent micro-grid environment. The method comprises the steps of: firstly, building electric power node data models within regional power grids; considering coupling relationship between the regional power grids, building perception model based on edge computing, to perceive the voltage and the phase angle of two ends of the tie-line; computing line losses between the regional power grids, and superimposing generation power and consumption power in the regional power grids to obtain net active power of the regional power grids; then setting the object function of the regional power grids; deducing the optimal object function of economic benefits of the interval power grids according to the object function of the regional power grids; converting the economy optimization and load regulation problem of the interval power grids into the problem of balancing and coupling of whole economic benefits of the distributed regional power grids and power between regions; and converting the object function into an augmented Lagrange function, solving the function by using the ADMM algorithm to achieve regulation and control of load of the power grid system.