Machine Learning Based Approach On Full Stack Development Utilizing Mongodb To Diagnose Faults In Complex Mechatronics System

Abstract

In the context of computerization and industrial plants, the concept addresses the memory safeguarding overhead and the significance of creating a diagnosis component for complex mechatronic systems. In order to minimize needless repairs and cut expenses, the module seeks to detect errors (Fault De- tection) and isolate their underlying causes (Fault Isolation). FastAPI Python, Docker, which is MongoDB, as well as machine learning algorithms make up the tech stack used, and AWS S3, AWS EC2, along with AWS ECR are needed for the infrastructure. MongoDB is shown in the project architecture, and Com- pass needs to be set up for data storage. For the purpose of improving system performance, productivity, and safety, the evaluation section is essential. Time series readings from sensors make up the dataset utilized for fault detection, where SensorID stands in for a temperature detector in a manufacturing envi- ronment. The goal of the issue of binary classification is to determine if a par- ticular part of the heavy-duty vehicle's Air Stress System (ASS), which uses compressed air stress for braking, was the cause of the failure.