Remaining Useful Life Estimation Using Multi-Error Breakdown Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting the remaining useful life of wind turbine pitch systems and other electromechanical components face challenges due to the lack of vibrational data in all turbines, reactive nature of sensor data, and harsh weather conditions, leading to inaccurate linear or exponential life calculations.
Innovation Solution
An error-based method that calculates remaining useful life by mapping cumulative error counts in an n-dimensional error space, considering relative error weights based on temporal distances and frequencies, and using these to determine the remaining life through a quotient of distances in the error space, which provides a more precise and realistic estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If linear or exponential interpolation is used to calculate remaining useful life between breakdowns, then a target variable for machine learning can be derived, but the calculation does not represent the useful life accurately at each time due to dependency on extreme conditions and varying run durations
Solution Approach 1:
The patent transforms the remaining useful life calculation from time-based parameters (linear/exponential interpolation over run duration) to error-based parameters (cumulative error counts in n-dimensional error space). This parameter change enables accurate representation of useful life at each time point by considering actual error accumulation patterns rather than assuming uniform degradation over time.
Solution Approach 2:
The patent introduces an n-dimensional error space where each dimension represents a different type of error. This dimensional transformation allows the system to capture multiple error modes simultaneously and calculate remaining useful life based on the device's position in this multi-dimensional error space rather than simple time progression, resolving the inaccuracy caused by varying run conditions.
2Reliability
If sensor data is used for supervised learning, then predictive analysis can be performed, but sensor data often covers key performance indicators that are reactive in nature and thus are less vital in providing an advanced indicator of failure
Solution Approach 1:
The patent introduces cumulative error counts as an intermediary metric that bridges sensor data and failure prediction. Instead of using raw sensor readings directly for prediction, the system first transforms sensor data into error events (when parameters breach thresholds), then accumulates these errors over time. This intermediary transformation converts reactive sensor data into proactive failure indicators, enabling supervised learning to predict remaining useful life based on error accumulation patterns rather than raw sensor values.
3Adaptability or versatility
If not all wind turbines are equipped with sensors capturing vibrational data, then deployment is more widespread, but vibrational data is a good indicator of mechanical failure and its absence limits prediction accuracy for electromechanical failures
Solution Approach 1:
The patent creates a universal error-based methodology that works with any sensor type already deployed on wind turbines (temperature, pressure, flow, etc.), rather than requiring specific vibrational sensors. The system treats errors as universal events that can be detected by various sensor types, making the remaining useful life prediction applicable to all turbines regardless of their specific sensor configuration, while maintaining accuracy for electromechanical failure detection.
Data Source
AI summary
A method for calculating a remaining useful life of an apparatus comprises the following steps. Time-series of previous runs and a current run of the apparatus are provided containing data of sensors configured to monitor parameters of the apparatus. An error occurs when a parameter breaches a threshold. Cumulative counts of errors occurring during a run are calculated. A linearly decreasing remaining useful life is calculated for previous runs. Breakdowns of the apparatus are mapped in an error space. Each dimension of the error space refers to one type of an error. The breakdown points are mapped at coordinates which represent cumulative error counts at the time of the breakdowns. A test point representing cumulative error counts of the current run is mapped. At least two nearest breakdown points to the test point are identified. The remaining useful life of the apparatus is calculated based on the nearest breakdown points.