Chiller Fault Prediction Using Adaptive Single and Cluster Models
Solution Overview
Problem
Chillers in HVAC systems often experience unexpected faults leading to unplanned shutdowns, causing efficiency losses and damage to other equipment, due to various influencing factors, making it challenging to predict chiller failures effectively.
Innovation Solution
A chiller fault prediction system using machine learning and deep learning models, which generates single and cluster prediction models to accurately predict chiller faults by analyzing performance data, labeling models based on accuracy, and re-evaluating models to retain or switch states, allowing for proactive maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If single chiller prediction models are generated for each chiller, then prediction accuracy for individual chillers is improved, but system complexity and computational resources increase
Solution Approach 1:
The system divides chiller prediction into two segments: individual chiller models for personalized prediction and cluster models for grouped prediction. This segmentation allows the system to maintain high accuracy for individual chillers while managing complexity through modular model structures that can be independently trained and evaluated.
Solution Approach 2:
Cluster chiller prediction models serve multiple functions by handling groups of chillers simultaneously, reducing the overall number of models needed. These universal cluster models can predict faults for multiple chillers with similar characteristics, thereby reducing system complexity while maintaining adequate prediction accuracy through shared patterns.
2Adaptability or versatility
If cluster chiller prediction models are used for chillers without historical data, then prediction capability is provided for new chillers, but model accuracy may be reduced compared to single chiller models
Solution Approach 1:
Cluster models are pre-trained on historical data from multiple chillers before being applied to new chillers without historical data. This preliminary training establishes robust prediction patterns that can be immediately applied to new chillers, providing prediction capability from day one while the system accumulates individual chiller data for future model refinement.
Solution Approach 2:
The system continuously monitors the performance of cluster models applied to new chillers and uses this feedback to refine and update the cluster models over time. As more data becomes available from newly commissioned chillers, the cluster models are retrained to improve accuracy while maintaining the ability to handle new chillers with limited historical data.
3Reliability
If models are continuously re-evaluated and retrained, then prediction accuracy is maintained over time, but computational resources and processing time increase
Solution Approach 1:
The system implements periodic re-evaluation and retraining of prediction models at scheduled intervals rather than continuously. This periodic approach maintains prediction reliability by updating models with new data while significantly reducing computational resource consumption compared to continuous retraining, as models are updated at manageable intervals based on operational cycles or time periods.
Data Source
AI summary
A fault prediction system for building equipment includes one or more memory devices configured to store instructions that, when executed on one or more processors, cause the one or more processors to receive device data for a plurality of devices of the building equipment, the device data indicating performance of the plurality of devices; generate, based on the received device data, a plurality of prediction models comprising at least one of single device prediction models generated for each of the plurality of devices or cluster prediction models generated for device clusters of the plurality of devices; label each of the plurality of prediction models as an accurately predicting model or an inaccurately predicting model based on a performance of each of the plurality of prediction models; and predict a device fault with each of the plurality of prediction models labeled as an accurately predicting model.