Method and controller for controlling a chiller plant for a building and chiller plant
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing chiller plant control systems rely on static assumptions and are not adaptable to varying environmental conditions or different chiller plant implementations, leading to inefficient energy consumption and operational performance.
Innovation Solution
A method using machine learning models to predict cooling load demand and power consumption based on environmental and operational data, allowing for dynamic control signal generation and selection to optimize chiller plant operation, incorporating building and occupancy data for improved accuracy and adaptability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If static rule-based analytics or manual adjustment methods are used for chiller plant control, then implementation simplicity is maintained, but adaptability to varying environmental conditions and different chiller plant configurations is poor
Solution Approach 1:
The patent implements dynamic control by training machine learning models with historical operational data and environmental conditions, enabling the control system to adapt its parameters in real-time based on varying conditions rather than relying on static pre-defined rules. The model continuously learns from data to optimize chiller plant performance across different environmental scenarios.
Solution Approach 2:
The patent changes control parameters dynamically by using machine learning models to predict optimal setpoints for chiller temperature, condenser water flow, and evaporator water flow based on environmental conditions and historical performance. This allows the system to adjust multiple parameters simultaneously to adapt to different operating conditions.
2Use of energy by moving object
If machine learning models are introduced to improve adaptability and optimization of chiller plant operation, then energy consumption and operational performance are optimized, but system complexity and data processing requirements increase
Solution Approach 1:
The patent applies preliminary action by training the machine learning model offline using historical operational data and environmental conditions before deployment. This pre-training phase captures optimal control patterns, allowing the system to make rapid real-time decisions during operation without requiring complex online computation, thus balancing optimization performance with system complexity.
3Measurement precision
If comprehensive training data including environmental data, cooling load demand data, control signals, and cooling power data are collected and used for model training, then prediction accuracy is improved, but data processing time and computational resources increase
Solution Approach 1:
The patent performs comprehensive data processing and model training in advance using historical datasets containing environmental conditions, cooling load demands, control signals, and cooling power outputs. This offline training phase prepares the model to make rapid predictions during real-time operation, achieving high prediction accuracy without incurring processing delays during critical control decisions.
Data Source
AI summary
Environmental data of an environment of the building and cooling load demand data are received as first training data, which are used for training a first machine learning model to predict a cooling load demand from environmental data. Furthermore, control signals for the chiller plant and cooling power data resulting from applying the control signals to the chiller plant are received as second training data which are used for training a second machine learning model to predict a cooling power from control signals. Actual environmental data are received, from which a cooling load demand is predicted by the trained first machine learning model. Furthermore, candidate control signals for the chiller plant are generated, and from which a resulting cooling power is predicted by the trained second machine learning model. From the candidate control signals, applicable control signals are selected for which a predicted cooling power fulfills the predicted cooling load demand.