Refrigeration Controller Using Load Prediction to Limit Compressor Cycling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional refrigeration system control methods, such as PID and fuzzy logic algorithms, require extensive tuning by experts and struggle to accurately forecast system behavior, leading to inefficient operation and excessive cycling, resulting in wasted energy and increased maintenance costs due to overshot suction pressure and changing load conditions.
Innovation Solution
A method that determines a desired rate of change in refrigeration system parameters using historical data to predict component behavior, calculates appropriateness factors, and selectively cycles components based on ranking, utilizing a neural network controller to modulate compressor capacity and adjust for changing loads without extensive tuning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional PID or fuzzy logic control algorithms are used for suction pressure control, then the system can maintain basic control functionality, but the system requires extensive expert tuning and cannot accurately forecast changing load conditions
Solution Approach 1:
The neural network controller automatically learns and adapts to system behavior patterns without requiring expert tuning. The controller self-adjusts by processing historical data and making predictions about load changes, eliminating the need for manual PID or fuzzy logic tuning while maintaining adaptability to changing conditions.
Solution Approach 2:
The neural network forecasts future system behavior and load conditions before they occur, allowing the controller to proactively adjust compressor capacity. This predictive capability enables the system to prepare for upcoming changes rather than merely reacting to them, improving adaptability without increasing tuning complexity.
2Productivity
If traditional control methods are used to adjust compressor capacity, then the system can respond to load changes, but the system overshoots target suction pressure resulting in inefficient operation
Solution Approach 1:
The neural network controller continuously monitors actual system performance and compares it with predicted values, using this feedback to refine future predictions and adjustments. This closed-loop learning process enables accurate control that prevents overshooting while maintaining rapid response to load changes, reducing energy waste from excessive cycling.
Solution Approach 2:
The controller dynamically adjusts compressor capacity in real-time based on predicted load changes and actual system response. This dynamic control approach allows the system to optimize capacity adjustments continuously, preventing overshoot conditions while maintaining high productivity through rapid adaptation to changing demands.
3Stability of the object's composition
If fixed capacity compressors are used in parallel, then the system can provide stable operation, but the system cannot efficiently handle varying load conditions without excessive cycling
Solution Approach 1:
The neural network predicts upcoming load changes and proactively adjusts which compressors should be cycled on or off before the load changes occur. This predictive approach allows fixed capacity compressors to handle varying loads efficiently by preparing capacity adjustments in advance, maintaining stability while improving adaptability to changing conditions.
Solution Approach 2:
The controller implements strategic cycling of fixed capacity compressors based on neural network predictions, optimizing the timing and sequence of compressor on/off transitions. This periodic control strategy reduces excessive cycling by coordinating compressor operations to match predicted load patterns, maintaining operational stability while adapting to varying demands.
Data Source
AI summary
A system and method of controlling components of a refrigeration system includes determining a desired rate of change of a refrigeration system operating parameter, using historical data to predict an expected rate of change of the operating parameter for cycling each component, calculating an appropriateness factor for each component corresponding to a difference between the expected rate of change for the component and the desired rate of change, ranking the components based on the appropriateness factor, and selectively cycling at least one component based on the ranking.