Control sequence generation system and methods
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing approaches struggle to efficiently manage building states due to complex interactions between buildings, occupants, and external environments, making it difficult to maintain desired conditions like temperature and humidity.
Innovation Solution
A control sequence generation system using a heterogenous physics network with nodes representing equipment behavior, employing thermodynamic models and machine learning to iteratively optimize control sequences until a goal state is reached, minimizing energy costs and improving comfort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional building control methods are used, then equipment operation is simple, but building state control precision deteriorates due to complex interactions between building, occupants, and external environment
Solution Approach 1:
The patent replaces traditional mechanical control systems with a neural network-based computational system. The neural network learns complex nonlinear relationships between building states, equipment operations, and environmental factors, substituting conventional control algorithms with machine learning models that can handle the complexity of building dynamics without requiring explicit mathematical models of all interactions.
Solution Approach 2:
The patent transforms the control problem by changing parameters from direct state control to demand curve optimization. Instead of directly controlling temperature or humidity, the system optimizes energy demand curves over time, allowing the building to naturally respond to these optimized demand patterns while maintaining desired states, thereby simplifying the control architecture.
2Ease of operation
If building state is maintained at desired levels, then occupant comfort improves, but energy consumption increases
Solution Approach 1:
The patent applies preliminary action by using the neural network to predict future building states and optimize equipment operation schedules in advance. The system anticipates when heating or cooling will be needed and pre-conditions spaces during off-peak hours, allowing comfort to be maintained while shifting energy consumption to more efficient times and reducing peak demand.
Solution Approach 2:
The patent implements periodic action through iterative optimization cycles where the neural network continuously refines control sequences. The system periodically updates control strategies based on learned patterns, optimizing equipment operation cycles to balance comfort maintenance with energy efficiency, rather than using continuous or static control approaches.
3Ease of operation
If equipment operates frequently to maintain states, then building comfort is maintained, but equipment lifespan decreases due to short cycling
Solution Approach 1:
The neural network performs preliminary action by predicting equipment needs and scheduling operations to avoid frequent on-off cycling. The system pre-cools or pre-heats spaces before peak demand periods, allowing equipment to run longer, more efficient cycles rather than short, frequent cycles, thereby extending equipment lifespan while maintaining comfort.
Data Source
AI summary
A model receives a target demand curve as an input and outputs an optimized control sequence that allows equipment within a physical space to be run optimally. A thermodynamic model is created that represents equipment within the physical space, with the equipment being laid out as nodes within the model according to the equipment flow in the physical space. The equipment activation functions comprise equations that mimic equipment operation. Values flow between the nodes similarly to how states flow between the actual equipment. The model is run such that a control sequence is used as input into the neural network; the neural network outputs a demand curve which is then checked against the target demand curve. Machine learning methods are then used to determine a new control sequence. The model is run until a goal state is reached.