Operation-State Abnormality Diagnosis With Reused Model Parameters

Solution Overview

Problem

Existing abnormality diagnosis methods struggle with erroneous determinations when facilities experience normal changes due to maintenance or operation mode shifts, and require extensive data collection and calculation, especially when operation states are unpredictable.

Innovation Solution

An abnormality diagnosis method that classifies data by operation state, assesses data sufficiency, calculates and associates parameter values without overlaps, and creates a trained model to determine abnormality based on these parameters, reducing the need for extensive data collection and calculation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If training data is collected for each new operation state before diagnosis, then diagnosis accuracy is improved, but diagnosis time is increased

Engineering Contradiction:

Improvediagnosis accuracyVSAvoiddiagnosis time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary classification of operation states and pre-calculates parameter values for each state category during system operation. When diagnosis is needed, the already-classified operation state and pre-calculated parameters enable immediate accurate diagnosis without requiring new training data collection

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically adapts to new operation states by classifying them into existing state categories based on similarity. Instead of requiring static pre-collected training data for every possible state, the system dynamically matches new states to closest known states and uses their parameter values for diagnosis

Inventive Principle:

Principle #15Dynamics

2Adaptability or versatility

If training data is collected for unpredictable operation states, then diagnosis coverage is improved, but data collection period is increased

Engineering Contradiction:

Improvediagnosis coverageVSAvoiddata collection period

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system performs preliminary classification of operation states and pre-calculates parameter values for each state category during system operation. When diagnosis is needed, the already-classified operation state and pre-calculated parameters enable immediate accurate diagnosis without requiring new training data collection

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system creates a universal diagnosis framework that handles both predictable and unpredictable operation states through a unified operation state classification mechanism. This single system serves multiple functions: classifying known states, matching new states to existing categories, and providing diagnosis coverage across all operational conditions

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Measurement precision

If training data is selected and models are created for each diagnosis target, then diagnosis accuracy is improved, but computational load is increased

Engineering Contradiction:

Improvediagnosis accuracyVSAvoidcomputational load

Core Design Contradiction:

Measurement precisionVSPower

Solution Approach 1:

The system segments the complex diagnosis task into distinct phases: operation state classification, parameter value calculation and storage, and diagnosis execution. By pre-calculating and storing parameter values for each operation state category, the system eliminates redundant computations during actual diagnosis, significantly reducing computational load while maintaining accuracy

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary classification of operation states and pre-calculates parameter values for each state category during system operation. When diagnosis is needed, the already-classified operation state and pre-calculated parameters enable immediate accurate diagnosis without requiring new training data collection

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250224718A1Abnormality diagnosis method, abnormality diagnosis apparatus, non-transitory computer-readable medium storing abnormality diagnosis program, and abnormality diagnosis system

Publication Date: 2025.07.10 MITSUBISHI ELECTRIC CORP

AI summary

A method of diagnosing whether there is abnormality includes obtaining data having a quantity of state of one or more assessment items from the facility, classifying, for each operation state of the facility, data obtained from the facility, assessing sufficiency of the number of pieces of data for each classified data group, calculating multiple parameter values configuring a trained model in accordance with the sufficiency and holding the parameter values in association with the operation state, obtaining the parameter values configuring the trained model without overlaps for each parameter, in accordance with the operation state of data to be diagnosed and a status of holding of the parameter values associated with each operation state, creating the trained model to diagnose data, with the parameter values, calculating a degree of abnormality for determination as to abnormality based on the trained model, and determining whether or not there is the abnormality.