Embedded Sensor Analytics for In-Circuit Motor Fault Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motor diagnostic methods, such as Motor Current Signature Analysis (MCSA), are limited to controlled environments and cannot analyze motors during in-circuit functioning, failing to account for operational and environmental hazards that cause thermal stress and degrade motor insulation, leading to complex and costly failure prevention.
Innovation Solution
Applying machine learning to sensor data from real-time sensors on Hardware Under Test (HUT) to predict future failures by analyzing parameters like vibration, voltage, and temperature, using smaller data sets and embedded ML models for predictive maintenance, enabling continuous assessment of HUT health.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Motor Current Signature Analysis (MCSA) is used to diagnose motors, then motor faults can be detected, but the analysis cannot be performed during in-circuit functioning and only in controlled environments
Solution Approach 1:
The patent replaces traditional MCSA methodology with a machine learning-based diagnostic system that uses multiple sensor inputs (vibration, temperature, current) to predict motor failures. This substitution enables the system to function in real-world operational environments rather than controlled lab settings, as the ML model can process noisy, real-time data from actual motor installations.
Solution Approach 2:
The patent introduces machine learning models as an intermediary between raw sensor data and fault diagnosis. The ML model acts as a mediator that processes complex multi-sensor data streams and translates them into actionable failure predictions, enabling reliable operation monitoring in diverse field conditions without requiring controlled environments.
2Reliability
If traditional motor monitoring methods are used, then failure prevention can be achieved, but the process becomes complex and costly
Solution Approach 1:
The patent merges multiple monitoring functions (vibration analysis, temperature monitoring, current measurement) into a single integrated machine learning model. Instead of running separate analysis systems for each parameter, the ML model processes all sensor inputs simultaneously to predict failures, thereby reducing overall system complexity while maintaining comprehensive monitoring capability.
Solution Approach 2:
The patent transforms the monitoring approach by changing from traditional threshold-based parameter analysis to machine learning-based predictive modeling. This parameter transformation allows the system to handle complex multi-variable interactions automatically, reducing the need for complex rule-based systems while improving failure prediction accuracy.
3Reliability
If comprehensive sensor data collection is implemented for predictive maintenance, then future failures can be predicted, but larger data sets require more processing resources
Solution Approach 1:
The patent extracts and utilizes only the most critical features from comprehensive sensor data sets for failure prediction. Rather than processing entire raw data sets, the ML model identifies and processes key predictive features (specific vibration frequencies, temperature trends, current harmonics), thereby maintaining high prediction accuracy while significantly reducing computational energy requirements.
Solution Approach 2:
The patent applies partial action by focusing computational resources on the most predictive subset of sensor data rather than processing all available data equally. The ML model selectively processes only those data elements that contribute most to failure prediction, achieving effective predictive maintenance with reduced energy consumption compared to comprehensive full-data processing.
Data Source
AI summary
A test and measurement instrument includes one or more ports to connect to hardware under test (HUT), a set of sensors connected to the HUT and to the instrument, a display to display one or more signal representations from at least one of the HUT and one or more sensors from the set of sensors, and one or more processors configured to execute code to cause the one or more processors to: acquire data from the set of sensors; form one or more data sets from the data acquired from the set of sensors; apply one or more machine learning models to the one or more data sets; and receive a predictive analysis from the one or more machine learning model about the HUT.