Degradation Detection Model Transfer Across Similar Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional degradation detection systems face challenges in accurately detecting device degradation due to differences in individual devices, installation conditions, and environmental conditions, requiring similar operation data for accurate model representation, which is often not available, especially when devices operate in a degradation state.
Innovation Solution
A degradation detection system that builds a normal model and degradation determination model for a target device using operation data from another device, allowing for degradation detection without requiring operation data from the target device's degradation state, by processing circuitry that updates models based on normal and degradation data from both devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a diagnostic model is built using operation data of all devices in the same group, then degradation detection can be performed even when the amounts of operation data of individual devices are small, but operation data of each device in the group and operation data of the target device need to be similar, otherwise accurate detection cannot be performed
Solution Approach 1:
The patent segments the data collection and model building process into two distinct phases: (1) collecting operation data from multiple devices in the same group during normal operation, and (2) building a diagnostic model using this aggregated data. This segmentation allows the system to overcome the limitation of insufficient individual device data while maintaining detection accuracy through group-based pattern recognition.
Solution Approach 2:
The patent merges operation data from multiple devices in the same group to build a collective diagnostic model. By combining data from multiple sources, the system creates a more robust model that can handle individual device variations while maintaining overall detection accuracy, resolving the contradiction between data sufficiency and device individuality.
2Measurement precision
If conditions are finely divided taking individual differences, installation conditions, environmental conditions into account, then accurate degradation detection can be performed, but massive amount and types of operation data are required to build each diagnostic model
Solution Approach 1:
The patent creates a universal diagnostic model that serves multiple devices within the same group simultaneously. Instead of building separate models for each device or condition variation, the system develops a single multi-functional model that can accurately detect degradation across diverse devices by leveraging patterns from aggregated group data, thereby reducing the total data requirement.
Solution Approach 2:
The patent changes the approach from device-specific parameter modeling to group-level parameter modeling. By shifting the modeling parameters from individual device characteristics to collective group characteristics, the system achieves accurate detection while requiring less data per device, as the model learns from the aggregated behavior patterns of the entire group.
3Measurement precision
If operation data of each device in the same group as the target device includes data that behaves differently than operation data of the target device, then a built diagnostic model does not accurately represent behavior of the target device, but collecting only data from the target device requires a sufficient amount of operation data in a degradation state which is difficult to obtain
Solution Approach 1:
The patent performs preliminary data collection and model building during the normal operation phase of devices, before degradation occurs. By aggregating data from multiple devices during their normal operation, the system prepares a diagnostic model in advance that can be applied when degradation is detected, eliminating the need to wait for degradation data to accumulate.
Solution Approach 2:
The patent creates a diagnostic model that copies and generalizes the normal operation patterns from multiple devices in the same group. This copied model serves as a reference framework that can detect deviations indicating degradation, allowing the system to leverage data from devices that have not yet degraded while maintaining accuracy for the target device.
Data Source
AI summary
There are included a normal model building unit that builds a normal model of another device on the basis of normal data of the other device; a degradation determination model building unit that builds a degradation determination model of the other device on the basis of the normal data and degradation data of the other device; a normal model rebuilding unit that builds a normal model of a target device on the basis of the normal model of the other device and normal data of the target device; a degradation determination model rebuilding unit that builds a degradation determination model of the target device on the basis of the degradation determination model of the other device and the normal model of the target device; and a degradation determining unit that determines degradation of the target device on the basis of operation data of the target device and the degradation determination model of the target device.