Wind Turbine RUL Prediction Using Multi-Error Breakdown Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting the remaining useful life of wind turbine components, particularly pitch systems, face challenges due to the lack of vibrational data in all turbines, the reactive nature of sensor data, and the impact of harsh weather conditions, leading to inaccurate linear or exponential calculations that do not account for the unique operational conditions of each turbine.
Innovation Solution
An error-based method that calculates the remaining useful life by mapping cumulative error counts in an n-dimensional error space, using relative error weights based on temporal distances and frequencies, and integrating data from similar apparatuses to provide a more precise prediction, which is then used to train machine learning algorithms for improved accuracy.
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 training can be derived, but the calculation does not accurately represent the useful life at each time due to ignoring extreme operational conditions and turbine-specific variability
Solution Approach 1:
The patent transforms the remaining useful life calculation from a time-based parameter (linear/exponential interpolation over time) to an error-space parameter (distance in n-dimensional error space). This parameter change allows the model to account for turbine-specific operational conditions and extreme events, improving accuracy without significantly increasing computational complexity.
Solution Approach 2:
The patent introduces an n-dimensional error space where each dimension represents a different type of error or deviation from normal operation. By mapping breakdown points and test points in this multi-dimensional space, the method captures operational variability and extreme conditions that single-dimensional time-based methods miss, thereby improving prediction accuracy.
2Reliability
If vibrational data is used for predicting remaining useful life, then mechanical failure can be detected, but not all wind turbines are equipped with such sensors and electro-mechanical failures cannot be detected
Solution Approach 1:
The patent creates a universal error-based framework that can process various sensor data types (temperature, pressure, flow rates, etc.) from different turbine configurations. The n-dimensional error space accommodates multiple sensor inputs, making the method adaptable to both turbines with vibrational sensors and those without, thereby improving versatility while maintaining reliability.
Solution Approach 2:
The patent introduces error metrics as an intermediary between raw sensor data and remaining useful life prediction. Instead of directly using raw sensor values or requiring specific sensor types, the method transforms all sensor inputs into error deviations from expected behavior, which then feed into the prediction model. This intermediary approach enables broad applicability across different turbine configurations.
3Loss of time
If sensor data is used for prediction, then key performance indicators can be monitored, but the data is reactive in nature and provides less advance indication of failure
Solution Approach 1:
The patent performs preliminary mapping of historical breakdown points in the error space before making predictions. By pre-processing and storing breakdown patterns in the n-dimensional error space, the system can quickly compare current operational states against known failure modes, providing advance warning while maintaining prediction accuracy through pattern recognition.
Solution Approach 2:
The patent implements a feedback mechanism where the error-based prediction model continuously compares current operational parameters against historical breakdown patterns. The deviation metrics provide feedback on how close the system is to failure conditions, enabling early warning while maintaining accuracy through continuous comparison with actual breakdown data.
Data Source
Figure 1
Figure 2
Figure 3
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.