Distributed Predictive Control Rules for Memory-Limited Energy Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Model predictive control in energy systems, such as wind turbines and grids, is challenging due to complexity and limited computing resources, especially when systems require local optimization and have inadequate memory for explicit MPC approaches.
Innovation Solution
A method utilizing a main computing system to maintain and optimize a predictive control model, generating control rules that are then stored and updated in local computing systems, allowing for effective model predictive control without being constrained by computational or storage limitations, with the main computing system able to scale resources and update rules as needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If explicit model predictive control with precomputed lookup tables is used to reduce online computational demands, then control speed improves, but memory requirements become inadequate for wind turbine environments
Solution Approach 1:
The patent segments the control system into a central computing system that performs complex offline optimizations and generates control rules, and local wind turbine controllers that execute these rules with minimal computation. This segmentation allows the memory-intensive lookup tables to be created centrally and transferred to controllers, reducing their memory requirements while maintaining fast local control execution.
Solution Approach 2:
The central computing system performs preliminary offline optimizations to generate control rules and lookup tables before deployment to local controllers. By precomputing the complex model predictive control solutions offline and storing them as executable rules, the system eliminates the need for heavy online computation and reduces memory requirements at the local level.
2Reliability
If complex model predictive control optimization is performed locally to achieve effective control, then control effectiveness improves, but computational resource requirements exceed available local capacity
Solution Approach 1:
The patent introduces a central computing system as an intermediary between the complex optimization algorithms and the local control systems. This intermediary performs the computationally intensive model predictive control optimizations centrally, then communicates the resulting control rules to local controllers, enabling effective control without requiring high computational power at remote wind turbine locations.
Solution Approach 2:
The patent replaces the mechanical/computational system of local optimization with a centralized computational approach. Instead of each wind turbine performing complex local optimizations, the system substitutes this with a central computing system that performs all heavy computational work and distributes simplified control rules, reducing the computational burden on individual turbines.
3Device complexity
If local optimization is performed at each system to simplify control architecture, then system complexity reduces, but overall system optimization performance deteriorates
Solution Approach 1:
The patent merges the optimization function from individual local controllers into a single central computing system. By combining all wind turbine data and performing centralized model predictive control optimization, the system achieves overall system optimization while maintaining relatively simple local controller architectures that only need to execute received control rules.
Data Source
AI summary
A main computing system maintains and optimizes a predictive control model for an energy system, wherein the main computing system receives state information for the energy system, optimizes the predictive control model, and generates control rules for control of the energy system. The one or more local computing systems, each have a local memory for storing control rules for controlling the associated local state. The main computing system receives local state information and updates control rules, wherein the updated control rules comprise a subset of the control rules generated by the main computing system selected to be appropriate to the local state information received at the main computing system.