Smart Event Detection for Physical Asset State Change Surveillance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for operational surveillance of physical assets face challenges in configuring and maintaining rule-based alerts and alarms, and the population of relevant datasets for model training is time-consuming, leading to inefficient detection of operational state changes in industrial equipment and processes.
Innovation Solution
A method and system that involve collecting time-series data, identifying operational state changes, extracting event data, generating label data, and using it to train a machine learning system to detect similar events, allowing for pattern recognition and continuous improvement in identifying future or past events, thereby avoiding the need for threshold-based alerts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If rule-based alerts with thresholds are used to detect operational state changes, then the detection mechanism is simple to implement, but the system generates significant noise and requires frequent threshold adjustments
Solution Approach 1:
The patent replaces the mechanical threshold-based alert system with a machine learning model that learns operational patterns from historical data. Instead of using fixed thresholds that generate noise, the system uses trained models to detect anomalies and operational state changes, significantly reducing false alarms while maintaining ease of deployment through automated learning.
Solution Approach 2:
The system transforms the detection approach by changing from static threshold parameters to dynamic, data-driven parameters. The machine learning models adapt their detection criteria based on learned patterns from historical operational data, allowing the system to maintain high detection accuracy without requiring manual threshold adjustments.
2Measurement precision
If manual population of datasets for model training is performed, then the training data can be curated for quality, but the process is time-consuming
Solution Approach 1:
The system enables automated data collection and preparation for model training. Instead of requiring manual curation of training datasets, the system automatically collects operational data, processes it into appropriate formats, and prepares it for training, significantly reducing the time investment while maintaining data quality through systematic processing pipelines.
Solution Approach 2:
The system performs preliminary data processing and preparation steps automatically before model training begins. By pre-collecting and pre-processing operational data into training-ready formats, the system eliminates the time-consuming manual dataset population process while ensuring high-quality training data is available when needed.
Data Source
AI summary
Methods and systems are provided for monitoring a physical asset, which includes receiving or collecting time-series data related to operation or status of the physical asset; identifying a time period when the physical asset is experiencing a change in operational state; extracting time-series data corresponding to the time period as event data; generating label data that classifies or characterizes the event data as pertaining to a particular type of event; saving the event data and the corresponding label data in a data repository; and using the event data and label data stored in the data repository to train or update a machine learning system to detect the occurrence of events that are similar to the event types of the labeled event data stored in the data repository from time-series data generated by the physical asset or by another physical asset that operates in a similar manner to the physical asset.