Data-Driven Mechanical Wear Prediction via Statistical Modeling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for predicting mechanical system wear and anomalies over time are imprecise due to external influences and require manual data manipulation, leading to delayed maintenance and inefficiencies.

Innovation Solution

A data-driven method and system that uses machine learning algorithms on large historical datasets to predict operational states by preprocessing data, selecting a training set, fitting statistical models, and accounting for nuisance variables to provide accurate and timely maintenance planning.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If empirical methods with manual data manipulation are used, then the system can predict wear using simple lookup tables, but the precision is reduced due to outside influences and the process is not automated

Engineering Contradiction:

Improvesimplicity of prediction methodVSAvoidprecision of wear prediction

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent replaces manual empirical methods with an automated data-driven system using machine learning algorithms. The system automatically processes large datasets, fits statistical models, and generates predictions without manual intervention, thereby improving both automation and precision while eliminating the limitations of simple lookup tables.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent transforms the prediction approach by changing from static lookup tables to dynamic statistical models that continuously learn from data. The system fits models to training datasets and updates predictions based on learned patterns, allowing it to account for outside influences and improve precision over time.

Inventive Principle:

Principle #35Parameter changes

2Reliability

If theoretical models based on physics or engineering information are used, then the system can understand system operation and failure progression, but the models use simplifying assumptions and are theoretical rather than data-driven

Engineering Contradiction:

Improveunderstanding of system operationVSAvoidaccuracy of wear prediction

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent replaces theoretical physics-based models with data-driven statistical models. Instead of relying on simplifying assumptions from first principles, the system learns directly from operational data, capturing real-world complexities and outside influences that theoretical models cannot account for, thereby improving prediction accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system continuously collects data during operation and uses this data to automatically update and refine its predictions. The model learns from the system's own operational data without requiring external theoretical input, adapting to actual system behavior and degradation patterns.

Inventive Principle:

Principle #25Self-service

3Quantity of substance

If known methods collect data only when the engine is new, then the initial system state is captured, but continual data collection during flights or tracking of degradation over time is not achieved

Engineering Contradiction:

Improveamount of data collectedVSAvoidprecision of wear tracking over time

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent implements continuous data collection during system operation rather than single-point measurements. The system continuously monitors operational parameters, collects data throughout the system's lifecycle, and uses this continuous stream of data to track degradation over time, thereby improving the precision of wear predictions at any point in the system's life.

Inventive Principle:

Principle #20Continuity of useful action

4Measurement precision

If large rolling averages are required to obtain confident values, then the scatter of individual points is reduced, but time delays occur for corrective action and maintenance prediction

Engineering Contradiction:

Improveconfidence in prediction valueVSAvoidtime delay for maintenance action

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent changes the prediction approach from using large rolling averages to using a trained statistical model that can provide confident predictions from individual or small sets of data points. The model, trained on comprehensive historical data, learns the underlying degradation patterns and can predict wear with confidence without requiring extensive averaging, thereby reducing time delays for maintenance actions.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS8165826B2Data driven method and system for predicting operational states of mechanical systems

Publication Date: 2012.04.24 THE BOEING CO

AI summary

There is provided an automated data driven method for predicting one or more operational states, such as wear or degradation, of a mechanical system over time. The method has the steps of collecting data on the mechanical system from a data recording device, preprocessing the collected data, selecting a training data set that represents a base condition for statistical comparison, fitting a statistical model to the training data set to relate a predicted response to nuisance variables at the base condition, and using an output model to predict what an observed response would have been at the base condition and calculating the difference between the observed response and the predicted response to predict the one or more operational states of the mechanical system.