Hierarchical Asset Models for Interpretable MTBF Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating Mean Time Between Failures (MTBF) in equipment, such as those based on the Telcordia TR-332/SR-332 standard, face challenges in model interpretability and accuracy when dealing with multiple causes of failures and failure states, leading to convoluted transition probabilities.
Innovation Solution
A multi-layer predictive model is developed that collects historical data from different hierarchical levels of equipment (component, circuit, and logical path) to generate operational state models, which are then used to create a top-level operational model for determining maintenance and replacement timing, utilizing both synchronous and asynchronous data collection methods and machine learning models to improve MTBF predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If state-based models with multiple failure states are used to estimate MTBF, then model accuracy improves, but model interpretability deteriorates due to convoluted transition probabilities
Solution Approach 1:
The patent segments the complex failure analysis into multiple hierarchical levels (component level, circuit level, logical path level, and equipment level). Each level has its own operational state model with transition probabilities, allowing the system to maintain accuracy while improving interpretability by analyzing failures at appropriate granularities rather than using a single convoluted model.
Solution Approach 2:
The patent introduces a hierarchical dimension to the failure analysis model. Instead of using a flat, single-level state model with many convoluted transitions, the system creates multiple levels of abstraction where transition probabilities are computed and aggregated hierarchically from component to equipment level, making the model both accurate and interpretable.
2Measurement precision
If hierarchical multi-layer models are used to capture detailed operational transitions, then MTBF prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The computational workload is segmented across multiple hierarchical levels. Each level processes transition probabilities for its specific scope (components, circuits, logical paths, or equipment), allowing parallel computation and avoiding the need to compute a single massive transition matrix. This segmentation reduces overall computational complexity while maintaining prediction accuracy.
Solution Approach 2:
The patent merges results from multiple hierarchical levels through aggregation of transition probabilities. The system combines component-level, circuit-level, and logical path-level models to produce equipment-level MTBF predictions, achieving accurate results through systematic combination rather than requiring a single complex monolithic model.
Data Source
AI summary
A method for generating a multi-layer predictive model includes collecting historical observable data from one or more pieces of equipment of a same type, wherein the historical observable data is collected at different hierarchical levels of the one or more pieces of equipment; collecting operational state indications of the pieces of equipment corresponding to the collected historical observable data; generating, from the collected historical observable data, a set of operational state models, wherein each operational state model corresponds to one of the different hierarchical levels; and generating, from outputs of the set of operational state models, a top-level operational model for the piece of equipment. The top-level operational model is operable to determine maintenance and replacement timing for the piece of equipment.