Inference Model Configuration for Faster Asset-Specific Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing remote condition monitoring systems require significant time and effort to develop accurate statistical models for different assets within the same category due to differences in construction, necessitating substantial rewriting of source code, which is typically done by skilled computer scientists.
Innovation Solution
A system and method for building and deploying inference models using a configuration file or compiled configuration database with customizable model configuration data, allowing a model builder application to operate in learning and runtime modes to construct and apply inference models adaptable to specific assets, reducing the need for extensive rewriting of source code.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If statistical models are customized for different asset models, then detection accuracy is improved, but development time and effort increase significantly
Solution Approach 1:
The system segments the model development process into distinct components: asset model definitions, parameter configurations, and inference model construction. Each asset model within a family is defined with specific parameters that can be independently configured, allowing accurate customization without complete redevelopment. The configuration files separate asset-specific parameters from the core inference logic, enabling efficient modification.
Solution Approach 2:
The system performs preliminary action by providing pre-built inference model templates and configuration frameworks for different asset families. These templates include common parameters and structures that can be directly applied to multiple asset models. The configuration files are prepared in advance with default settings that can be quickly adjusted, eliminating the need to build models from scratch for each asset variant.
2Measurement precision
If statistical models are customized for different asset models, then detection accuracy is improved, but complexity of the system increases
Solution Approach 1:
The system implements universality through a family-based asset modeling approach where a single inference model framework serves multiple asset models. The configuration files use standardized structures that can accommodate different asset types within a family, allowing the same core system to handle diverse assets with minimal customization. Common parameters and data structures are reused across multiple asset models, reducing overall system complexity.
Solution Approach 2:
The system manages complexity by allowing customization through parameter changes rather than structural modifications. Each asset model is distinguished by specific parameter values in configuration files (e.g., operational envelopes, threshold values, sensor configurations) rather than requiring different code structures. This parameter-based approach enables accurate customization while maintaining a unified, simple system architecture.
3Measurement precision
If source code rewriting is performed for different assets, then model accuracy is improved, but ease of operation deteriorates
Solution Approach 1:
The system introduces configuration files as an intermediary layer between the user and the inference model system. These files provide a user-friendly interface for specifying asset-specific parameters without requiring direct manipulation of source code. The configuration files translate high-level asset characteristics into the detailed parameters needed by the inference models, making the system accessible to domain experts rather than requiring programming expertise.
Solution Approach 2:
The system uses copying by providing template configuration files that can be replicated and adapted for different asset models. Instead of creating new models from scratch, users can copy existing configuration templates and modify specific parameters to match their asset characteristics. This copying approach maintains consistency across asset models while enabling quick customization, improving ease of operation without sacrificing accuracy.
Data Source
AI summary
A system and related method for building and deploying one or more inference models for use in remote condition monitoring of a first fleet of a first asset. The system includes model configuration data for subsequent use by a model builder application to construct one or more desired inference models for the first asset. The model configuration data is customized to the first asset and the desired one or more inference models, and is provided in a format which is easily readable and editable by a user of the system. The model configuration data is separate from the underlying processing algorithms which are employed by the model builder application in the constructing of the one or more desired inference models during a learning mode of operation of the system.