Vehicle Wiring Harness Failure Prediction Using Load Cycling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern vehicles face challenges in predicting failures of electrical loads and wiring harnesses, leading to difficult diagnosis and unnecessary repairs.
Innovation Solution
A method is provided that involves selectively enabling and disabling electrical loads, collecting operational data, and training a model to predict failures associated with electrical loads or wiring harnesses. This method includes using sensors to collect current, voltage, and temperature data, and employing machine learning models such as convolutional neural networks or physics-based models for prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of repair
If traditional trial-and-error diagnosis methods are used for electrical failures, then repair personnel can attempt repairs, but the diagnosis is difficult and unnecessary repairs occur
Solution Approach 1:
The system performs preliminary actions by collecting operational data continuously during normal vehicle operation and training prediction models in advance. When a failure occurs, the pre-trained model can immediately predict the failure type and location without requiring trial-and-error diagnosis, thus resolving the contradiction between ease of repair and time loss.
Solution Approach 2:
The system implements feedback by continuously monitoring operational data (current, voltage, temperature) and comparing it against trained prediction models. The model provides feedback about the health status of electrical loads and wiring harnesses, enabling accurate failure prediction without time-consuming manual diagnosis.
2Measurement precision
If comprehensive operational data collection is performed using sensors, then prediction accuracy improves, but system complexity increases
Solution Approach 1:
The system applies universality by using a multi-functional smart energy center that combines sensor data collection, model training, and failure prediction capabilities in a single integrated unit. This approach improves prediction accuracy through comprehensive data collection while minimizing system complexity by avoiding separate dedicated components for each function.
3Reliability
If machine learning models are trained using operational data from multiple vehicles, then prediction reliability improves, but data processing requirements increase
Solution Approach 1:
The system merges data from multiple vehicles into a centralized training dataset, combining operational data across the fleet to train more reliable prediction models. This approach improves prediction reliability by leveraging diverse real-world operating conditions while managing data processing complexity through unified model training procedures.
Data Source
AI summary
Examples described herein provide a method for predicting electrical failures within a vehicle. The method includes selectively enabling and disabling an electrical load for the vehicle. The method further includes collecting operational data about the electrical load during the selectively enabling and disabling of the electrical load. The method further includes training the model based at least in part on the operational data to predict a failure associated with the electrical load or a wiring harness associated with the electrical load.