Parametric Failure Models for Equipment Reliability Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current solutions for predicting equipment failures rely heavily on sensor data, which is not available for many equipment assets, leading to inaccurate estimates of failure probability and remaining useful life, and lack interactive visualizations and granular analysis capabilities.
Innovation Solution
The implementation of a system using parametric probability models, such as the Weibull distribution, to predict failure mode-specific reliability characteristics of equipment assets, enabling the generation of accurate and granular analytical data without sensor instrumentation, and providing interactive visualizations for users to adjust analysis granularity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data is used for predicting equipment failures, then prediction accuracy is improved, but equipment complexity and cost increase due to required sensor instrumentation
Solution Approach 1:
The patent uses parametric probability models as mathematical copies that replicate failure patterns observed from historical data. These models enable prediction of equipment failures without requiring physical sensor copies on each equipment unit, thus maintaining prediction accuracy while avoiding the complexity of extensive sensor instrumentation.
Solution Approach 2:
The patent replaces the mechanical sensor-based monitoring system with a computational model-based system. Instead of using physical sensors to detect equipment state, the system uses parametric probability models that compute failure likelihood based on equipment characteristics and historical failure data, eliminating the need for complex sensor hardware.
2Device complexity
If general failure prediction is used, then system simplicity is maintained, but granularity and actionable insights of analytical data are reduced
Solution Approach 1:
The patent segments failure predictions by specific failure modes rather than providing only general failure probability. The parametric models are trained on and can predict different failure modes separately, enabling users to obtain granular analytical data about specific types of failures while maintaining system simplicity through a unified modeling framework.
Solution Approach 2:
The patent enables users to obtain locally optimized analytical data by allowing selection of specific equipment models and failure modes of interest. The system provides detailed, granular predictions tailored to specific equipment and failure type combinations rather than generic predictions, enhancing information quality without requiring separate complex systems for each case.
3Device complexity
If equipment is not instrumented with sensors, then device complexity is reduced, but ability to accurately estimate failure probability and remaining useful life is worsened
Solution Approach 1:
The patent performs preliminary action by training parametric probability models on historical failure data before deployment. These pre-trained models capture failure patterns and characteristics, enabling accurate failure probability estimation and remaining useful life prediction when applied to new equipment without requiring sensor instrumentation on those specific units. The modeling work is done in advance, transferring knowledge to unsensorized equipment.
Solution Approach 2:
The patent introduces parametric probability models as intermediary elements between historical failure data and current equipment prediction needs. These models act as mediators that translate historical patterns into predictions for equipment without direct sensor data, bridging the gap between lack of instrumentation and need for reliable failure probability estimation.
Data Source
AI summary
Techniques for predicting failure mode specific reliability characteristics of tangible equipment using parametric probability models are disclosed. In some example embodiments, a computer system receives a model training configuration entered via a user interface, trains a failure curve model for a selected failure mode of a selected equipment model based on the model training configuration at a time indicated by training schedule data, and generates analytical data for the selected failure mode of the selected equipment model using the trained failure curve model. The failure mode corresponds to a specific way in which the equipment model is capable of failing. In some example embodiments, the training of the failure curve model comprises determining a shape parameter and a scale parameter for the failure curve model based on a fitting of failure event data to a continuous probability distribution, and storing the parameters for use in generating the analytical data.