Data Center Microgrid Cluster Dispatch Using Virtual Battery Aggregation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies lack a method for characterizing the flexibility of data center microgrid clusters and coordinating the scheduling of multiple flexibility resources, which is crucial for efficient and low-carbon operation of large-scale data centers.
Innovation Solution
A robust dispatch method is developed for large-scale data center microgrid clusters, utilizing a virtual battery model to represent flexibility regulation capacity, aggregating flexibility regions through Minkowski summation, and implementing a day-ahead-intraday two-stage robust dispatch model to optimize resource-load scheduling and adjust for uncertainties.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing optimal scheduling methods are applied to individual data center microgrids, then scheduling efficiency is improved, but the ability to characterize and coordinate flexibility resources across multiple microgrids is lost
Solution Approach 1:
The patent merges multiple individual data center microgrid scheduling problems into a unified cluster-level scheduling framework. By introducing a virtual battery model that aggregates flexibility resources across multiple microgrids and using Minkowski summation to combine flexibility regions, the system achieves coordinated optimization that maintains individual scheduling efficiency while gaining cluster-level flexibility characterization capabilities.
2Adaptability or versatility
If flexibility resources are aggregated across multiple data center microgrids, then overall flexibility regulation capability is improved, but system complexity increases
Solution Approach 1:
The patent introduces a virtual battery model as an intermediary that simplifies the aggregation of flexibility resources across multiple microgrids. This virtual battery abstracts the complex interactions between distributed generator sets, energy storage equipment, and demand response characteristics into a unified flexibility representation, making the scheduling system more manageable despite the increased scale.
Solution Approach 2:
The patent transforms the multi-dimensional flexibility characteristics of distributed resources into a unified flexibility region representation using Minkowski summation. By projecting complex resource constraints into a standardized flexibility region space, the system manages aggregation complexity while preserving the essential flexibility characteristics needed for coordinated scheduling.
3Loss of energy
If day-ahead scheduling is performed with forecast information, then operational cost is reduced, but uncertainty in renewable energy output and workload causes scheduling deviations
Solution Approach 1:
The patent performs day-ahead scheduling using forecast information to establish an initial optimal resource-load scheduling plan, capturing the economic benefits of advance planning. The system then uses intraday rolling optimization to adjust this plan in real-time based on actual conditions, ensuring both cost efficiency and scheduling accuracy by combining preliminary planning with adaptive correction.
Solution Approach 2:
The patent implements a two-stage robust dispatch model where day-ahead scheduling provides the initial plan and intraday rolling optimization provides real-time feedback adjustments. This feedback mechanism allows the system to correct deviations caused by forecasting uncertainties while maintaining the economic advantages of advance scheduling, achieving both cost reduction and improved reliability.
Data Source
AI summary
The robust dispatch method for flexibility resources of large-scale data center microgrid clusters characterizes overall scheduling capacity of multiple flexibility resources in each data center microgrid by establishing a virtual battery model, constructs an equivalent model of scheduling capacity of data center microgrid cluster based on constraint space superposition methods of Minkowski summation, and fully exploits flexibility of output power regulation of distributed generator sets and energy storage of the data center microgrid cluster as well as demand response characteristics of loads, so as to obtain regulation domain of aggregate power of the data center microgrid cluster.