Building Equipment Vibration Anomaly Detection With ML
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analyzing large data sets from building equipment for operational issues is time-consuming and costly due to the need for human analysts, and manually parsing these data sets can be difficult with limited personnel.
Innovation Solution
A method utilizing machine learning models to analyze vibration data sets from building equipment, incorporating operator comments and FFT spectra, to identify normal or abnormal operations and initiate corrective actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human analysts manually analyze vibration data sets to detect equipment abnormalities, then detection accuracy is improved, but analysis time and cost increase significantly
Solution Approach 1:
The patent introduces machine learning models as an intermediary between raw vibration data and human analysts. The ML models automatically process vibration data sets, extract features, and generate preliminary anomaly assessments, thereby reducing the time and effort required for manual analysis while maintaining detection accuracy through the models' ability to identify patterns in the data
Solution Approach 2:
The patent replaces the mechanical process of manual data analysis with an automated computational system. Machine learning models substitute for human analysts in the initial screening and pattern recognition tasks, automatically processing vibration data sets to identify abnormalities without requiring manual parsing of each data point
2Productivity
If more human analysts are hired to parse large data sets, then analysis coverage is improved, but training cost and time increase
Solution Approach 1:
The machine learning models perform self-learning by training on historical vibration data sets with known outcomes. Once trained, the models autonomously analyze new data sets without requiring human intervention or training, thereby increasing analysis capacity without the need to hire and train additional analysts
Solution Approach 2:
The patent transforms the analysis capacity parameter from being dependent on human resources to being dependent on computational resources. By changing the system from manual analysis to automated ML-based analysis, the capacity to handle large data sets increases significantly without the linear scaling costs associated with hiring and training more analysts
3Measurement precision
If manual parsing of vibration data sets is performed, then detailed inspection is improved, but difficulty increases with limited personnel
Solution Approach 1:
The patent extracts the most critical features and patterns from vibration data sets using machine learning models. Instead of requiring manual parsing of entire data sets, the ML models extract relevant anomalies and present condensed findings to analysts, thereby maintaining inspection quality while reducing the operational difficulty of processing large volumes of data
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Automatically identifies equipment abnormalities, reducing the workload on human analysts and ensuring timely corrective actions, thereby improving operational efficiency and reducing costs.
Implementation Method 1
analyzing the vibration data set and the operator comments using one or more machine learning models to identify the operation of the equipment as normal or abnormal
Implementation Method 2
analyzing the vibration data set includes performing one or more fast Fourier transforms on the vibration data set to generate one or more fast Fourier transform (FFT) spectra
Data Source
AI summary
A method for correcting abnormal operation of equipment includes obtaining a vibration data set including vibration measurements recorded by one or more vibration sensors while operating the equipment during a time period, obtaining operator comments including observations from an operator characterizing operation of the equipment during the time period, analyzing the vibration data set and the operator comments using one or more machine learning models to identify the operation of the equipment as normal or abnormal, and initiating a corrective action responsive to identifying the operation of the equipment as abnormal. In some embodiments, the method includes generating model reasoning indicating a reason why the operation of the equipment is identified as normal or abnormal by the one or more machine learning models.