State Diagnosis Apparatus Model Validity Monotonic Check
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing state diagnosis systems for industrial devices rely on machine learning models to detect deterioration, but there is a need to validate the effectiveness of these models in sensing changes in device signals over time, ensuring they accurately predict deterioration before it leads to device malfunction.
Innovation Solution
A state diagnosis apparatus that includes a processing circuit to determine the validity of a model by analyzing whether the numerical values output from the model monotonically change with the progression of device deterioration, using multistage deterioration data and a model determination unit to assess the model's ability to correctly sense and quantify deterioration states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a machine learning model is used to diagnose device state, then automation and productivity are improved, but the reliability of the diagnosis is uncertain without validation
Solution Approach 1:
The patent implements a feedback mechanism where the model's output values are evaluated against expected monotonic change patterns. The validation unit provides feedback on whether the model's predictions are reliable, enabling automatic identification of valid models without manual intervention while maintaining diagnostic reliability.
Solution Approach 2:
The system performs self-validation where the model evaluates its own outputs against predetermined criteria (monotonicity of state numerical values). This self-service approach allows the system to automatically determine model validity without external intervention, resolving the contradiction between automation and reliability.
2Reliability
If manual inspection is performed to ensure diagnosis accuracy, then reliability is improved, but productivity and efficiency deteriorate
Solution Approach 1:
The validation unit enables the system to automatically validate model predictions by checking if output values change monotonically with device deterioration. This self-validation mechanism replaces manual inspection while maintaining diagnostic accuracy, thus improving productivity without sacrificing reliability.
Solution Approach 2:
The patent replaces manual mechanical inspection with an automated computational validation system. The validation unit uses algorithmic checks (monotonicity assessment) to substitute human judgment, thereby maintaining diagnostic accuracy while dramatically improving inspection speed and productivity.
3Reliability
If model validation is performed to ensure effectiveness, then reliability is improved, but device complexity and processing requirements increase
Solution Approach 1:
The validation mechanism focuses on a single key parameter - the monotonicity of state numerical values. By validating only this critical parameter rather than all model outputs, the system ensures prediction validity while keeping the validation process simple and computationally efficient.
Solution Approach 2:
The validation unit applies localized quality control by checking only the essential monotonicity property of model outputs rather than进行全面 validation. This targeted approach ensures reliability where it matters most while minimizing overall system complexity and processing requirements.
Data Source
AI summary
According to one embodiment, a state diagnosis apparatus includes a processing circuit. The processing circuit executes a model receiving, as an input, first data relating to a state of a device at each of a plurality of stages along a time series and outputting a first numerical value quantitatively indicating the first data for each of the stages. The processing circuit determines whether or not first numerical values output from the model monotonously change along the time series.