Engine Vibration Diagnosis Using Deep Learning Classifiers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional engine diagnosis during production is limited to simple state checks, making it difficult to accurately identify and classify abnormal states, leading to increased costs and man-hours, as well as potential delivery of vehicles with defective engines to customers.
Innovation Solution
A deep learning-based method using primary and secondary diagnostic models to classify engine vibrations through big data analysis, employing algorithms like recursive neural networks, convolutional neural networks, and ensemble deep neural networks, to accurately diagnose normal or abnormal states during the end-of-line production process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If simple state check is used for engine diagnosis during production, then the diagnosis process is simple and quick, but the classification accuracy of abnormal states is low
Solution Approach 1:
The diagnosis system is segmented into multiple specialized classifiers, each designed to detect specific types of abnormal states (e.g., knocking, pre-ignition, misfire). This segmentation allows each classifier to focus on particular patterns, improving overall classification accuracy while maintaining manageable system complexity through modular design
Solution Approach 2:
The system transitions from simple frequency level comparison to multi-dimensional analysis by examining vibration signals across multiple parameters including but not limited to frequency spectrum, time-domain characteristics, and pattern recognition features. This dimensional expansion enables more accurate differentiation between various abnormal states
2Measurement precision
If deep learning-based composite diagnosis is implemented, then the classification accuracy of abnormal states is improved, but the diagnosis process becomes more complex and time-consuming
Solution Approach 1:
Vibration signals are continuously monitored and analyzed in real-time during the engine production process. The deep learning classifiers are pre-trained with extensive data, enabling them to rapidly classify abnormal states without adding significant processing time to the production line
Solution Approach 2:
The system replaces manual or simple automated diagnosis methods with deep learning-based artificial intelligence classifiers. These AI classifiers automatically process and interpret complex vibration patterns, reducing both the time and expertise required for accurate diagnosis while improving precision
3Ease of manufacture
If simple frequency level comparison is used, then the diagnosis method is easy to implement, but it cannot accurately diagnose composite abnormal states
Solution Approach 1:
The deep learning-based diagnosis system is designed to be universal, capable of detecting and classifying multiple types of abnormal states (knocking, pre-ignition, misfire, and other vibrations) using a single integrated platform. This multi-functional approach maintains ease of implementation while significantly improving reliability for composite abnormalities
Data Source
AI summary
A method of diagnosing an engine condition may include measuring a vibration of an engine, wherein assembly of the engine is completed in an automated end of line (EOL) process of an engine production, and primarily diagnosing the vibration of the engine during the EOL process using a primary deep learning classification model in which vibration signals of the engine are classified according to a feature through learning using a plurality of algorithms on the basis of big data with respect to multiple vibration signals measured at three or more positions on the engine, wherein the assembled state of the engine in the EOL process is classified into a normal state or an abnormal state by performing the primary diagnosing of the vibration of the engine.