Refrigerant Leak Detection Using Expected Level Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Refrigeration systems often lack accurate monitoring and analysis capabilities, making it difficult for users to detect refrigerant leaks and maintain optimal performance, which can affect efficiency, safety, and profitability.
Innovation Solution
A system comprising a refrigerant level sensor, system sensors, a model database, and a notification module that uses linear regression and hour-compensated models to detect refrigerant leaks by comparing actual and expected refrigerant levels, generating notifications for deviations beyond control limits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If refrigeration systems use basic monitoring without advanced analysis capabilities, then the system simplicity is maintained, but the ability to detect refrigerant leaks and maintain optimal performance deteriorates
Solution Approach 1:
The patent introduces an intermediary processing layer that includes a model database storing multiple refrigeration models, a model selecting module that chooses appropriate models based on system conditions, and a refrigerant level prediction module that generates predictions. This intermediary layer processes sensor data through selected models to detect refrigerant leaks, thereby improving reliability without requiring the entire system to be fundamentally complex.
Solution Approach 2:
The system performs self-service by automatically selecting appropriate refrigeration models based on current system conditions and autonomously predicting refrigerant levels without requiring external expert intervention. The model selecting module automatically chooses from stored models based on sensed conditions, and the prediction module generates leak detection alerts autonomously, improving reliability while keeping operation simple.
2Measurement precision
If the system stores multiple refrigeration models in a model database, then the prediction accuracy is improved, but the memory requirements and data processing load increase
Solution Approach 1:
The system dynamically selects from multiple stored refrigeration models based on current system conditions rather than using a single static model. The model selecting module chooses the most appropriate model for the current operating state, allowing the system to maintain high prediction accuracy across varying conditions without needing to process all models simultaneously, thus managing data volume efficiently.
Solution Approach 2:
Different refrigeration models are stored for different operating conditions or system states. Each model is optimized for specific local conditions, and the model selecting module chooses the appropriate model based on current sensor readings. This allows the system to have high prediction accuracy for each specific condition while managing overall data volume by only loading and processing one model at a time.
3Reliability
If the system performs continuous monitoring with multiple sensors and model comparisons, then the refrigerant leak detection reliability is improved, but the computational energy consumption increases
Solution Approach 1:
Multiple refrigeration models are pre-stored in the model database during system setup or initialization, rather than being generated in real-time during operation. The model selecting module and prediction module then efficiently retrieve and apply these pre-computed models to current sensor data, significantly reducing computational energy consumption during continuous monitoring while maintaining high detection reliability.
4Loss of time
If the notification module generates alerts for deviations beyond control limits, then the timeliness of leak detection is improved, but the number of false alarms may increase
Solution Approach 1:
The system uses different control limit parameters (upper control limit and lower control limit) to define the threshold for generating notifications. By establishing specific parameter thresholds based on normal system variation, the system can timely detect actual refrigerant leaks while filtering out normal fluctuations, thus improving notification accuracy and reducing false alarms.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A refrigerant leak detection system is herein disclosed. The system can detect the existence of a slow leak based on the refrigerant level and the system data. They system uses data models, either stored or dynamically created, to calculate an expected refrigerant level. Based on the expected refrigerant level and the actual refrigerant level and statistical process control data, a leak can be identified.