Oil Check Valve Replacement Prediction From Thermal Degradation Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing oil check valves in vehicles suffer from resin material deterioration due to heat, leading to potential oil leaks that are difficult to predict, necessitating early replacement but currently detected only through post-inspection.

Innovation Solution

A prediction device using a classifier model to analyze operation and temperature history data to calculate the replacement timing of oil check valves, incorporating a neural network to estimate thermal damage and vehicle usage patterns.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the resin material in the oil check valve is used, then the valve can operate normally, but the resin material deteriorates due to heat causing oil leakage

Engineering Contradiction:
Improveoil check valve operationVSAvoidresin material service life
Core Design Contradiction:
ReliabilityVSDuration of action of stationary object

Solution Approach 1:

The system performs preliminary detection of oil leakage by monitoring operation history and oil temperature data before actual oil leakage occurs. The classifier model predicts the likelihood of oil leakage in advance, allowing maintenance to be scheduled before the resin material completely deteriorates and causes functional failure.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If early replacement of oil check valve is performed, then sudden failure is prevented, but replacement is performed before actual need based on current inspection-only approach

Engineering Contradiction:
Improveprevention of sudden failureVSAvoidpremature replacement
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system continuously monitors operation history and oil temperature data, feeding this information to the classifier model which updates the likelihood of oil leakage. This feedback mechanism allows the system to dynamically adjust replacement timing predictions based on actual operating conditions, replacing the valve only when the predicted likelihood exceeds the threshold.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If inspection at inspection center is performed, then actual oil leakage is detected, but prediction of leakage timing is not possible

Engineering Contradiction:
Improvedetection of oil leakageVSAvoidprediction capability
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The classifier model performs preliminary prediction of oil leakage timing by analyzing operation history and oil temperature data before actual leakage occurs. This allows the system to forecast when the valve will fail, providing valuable lead time for scheduled maintenance rather than waiting for inspection centers to detect the problem.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20260073736A1Prediction device, prediction method, and prediction program
Publication Date: 2026.03.12 ISUZU MOTORS LTD
  • US20260073736A1 patent drawing
  • US20260073736A1 patent drawing
  • US20260073736A1 patent drawing

AI summary

An object of the present disclosure is to provide a prediction device capable of predicting replacement required timing of an oil check valve mounted on a vehicle. A prediction device according to the present disclosure is for predicting replacement required timing of an oil check valve provided in an oil circulation path of a vehicle. The prediction device includes the following: an acquisition section that acquires, from a storage section, first information related to an operation history of the oil check valve and second information related to an oil temperature history of oil in the oil circulation path; and a calculation section that calculates an indicator related to the replacement required timing of the oil check valve based on the first information and the second information by using a classifier model that has been trained in advance.