Power Asset Failure Prediction Using Similarity-Based Remaining Life
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting the time-to-failure and remaining-life of electrical power system assets are inefficient, often requiring time-consuming and costly physical examinations or scheduled replacements, which may occur too early or too late, leading to service disruptions and economic losses.
Innovation Solution
A method that analyzes performance, environmental, and usage characteristics of power system assets over a specified interval, associates them with similar assets in a database, and calculates time-to-failure and remaining-life probability factors to inform utility companies' operational planning, allowing for more accurate and timely repairs and replacements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If physical examinations and testing of individual assets are conducted periodically, then the detection of failing assets is improved, but the time consumption and cost increase significantly
Solution Approach 1:
The patent creates a virtual model (digital twin) of each physical asset that replicates its behavior and performance characteristics. This virtual model allows for continuous monitoring and failure prediction without requiring physical inspection, thereby maintaining high detection reliability while eliminating time consumption associated with periodic examinations.
Solution Approach 2:
The patent replaces mechanical physical inspection methods with computational analysis of operational data. Instead of physically examining assets, the system uses algorithms to analyze performance data from sensors and historical records, substituting mechanical inspection with automated computational prediction.
2Reliability
If assets are replaced according to a predetermined schedule, then the reliability of service is improved, but assets may be replaced when they are functioning normally, well before failure would have occurred
Solution Approach 1:
The patent performs preliminary failure prediction by continuously analyzing asset performance data and comparing it against failure patterns in the database. This allows the system to identify assets that are likely to fail in the near future, enabling proactive replacement only when necessary, rather than following a rigid predetermined schedule.
Solution Approach 2:
The patent transitions from static predetermined replacement schedules to dynamic, adaptive replacement timing. The system continuously updates failure probability assessments based on real-time performance data and historical patterns, allowing replacement schedules to adapt to actual asset conditions rather than following fixed timelines.
3Reliability
If assets are examined and tested periodically, then failing assets can be detected, but the process is expensive and difficult to schedule without interrupting power to end users
Solution Approach 1:
The patent enables assets to self-report their health status through continuous monitoring sensors and data collection systems. The virtual models automatically assess asset condition and predict failures without requiring external inspection, eliminating the need for complex scheduling of manual examinations and allowing continuous operation without interruptions.
4Reliability
If repairs are made before power outage occurs, then service reliability is improved, but the timing and conditions for repairs must be optimized to avoid additional difficulties and expense
Solution Approach 1:
The patent enables preliminary identification of assets at risk of failure by analyzing performance data and comparing it against failure patterns. This allows utility companies to schedule repairs during planned maintenance windows or off-peak periods, avoiding emergency repairs during critical outages and reducing overall repair costs and complexity.
Data Source
AI summary
A method for predicting time-to-failure and remaining-life of an electrical power system asset includes analyzing characteristics of the power system asset over a specified time interval; based on the analysis, associating the power system asset with a pool of similar power system assets having similar historical performance characteristics; and based on the association, calculating time-to-failure and remaining-life probability factors for the power system asset. A computer program product is also provided for carrying out the method, and the method may further include a mechanism whereby the computer program learns ways to modify and enhance the analysis of the performance characteristics to provide for higher confidence levels in time-to-failure and remaining-life probability calculations.