Title: Industry 4.0 - Machine Learning for Industrial Assembly
Subject: Computer science, Robotics
Level: Advanced, Basic
Description: Background:

At our assembly lines, we have multiple measurement points from which data is collected to ensure that our customers receive products that meet their expectations. These collection points provide multitude of opportunities to find relationships and analyse to extract value from this data.

One such point of interest in the assembly is the calculation of shims. Shim thickness is calculated using a mathematical formula which involves tolerance and measurement. It is important to measure the thickness of a shim before using it - one reason being the position of the shim on the gear determines the backlash on the unit. Angular displacement is then calculated and if it is not within the specification, a rework is needed and the cycle is repeated.

Goal:

The aim of this project work, is to create a machine learning model to predict or estimate the right shim that goes into the final unit.

A model is required which should train on the data we are collecting and be capable of automating the prediction of the thickness of the shim to be used. This model should be efficient to reduce the work that is redone if the right shim is not chosen.

Requirements:

• Verifying data quality, and/or ensuring it via data cleaning
• Defining datasets to be used for training
• Supervisory control and data acquisition
• Understanding the business objective and developing an efficient model that will help to make customer oriented decisions, along with metrics to track their progress
• Exploring and visualizing data to gain an understanding of it, then identifying differences in data distribution that could affect performance when deploying the model in the real world

Qualification:
Proficiency in Machine learning with an understanding of mechanical engineering concepts.

Time Duration of the project:
20 weeks


Start date:
End date:
Prerequisites: Proficiency in Machine learning with an understanding of mechanical engineering concepts.
IDT supervisors: Shaibal Barua (shaibal.barua@mdh.se), Shahina Begum (shahina.begum@mdh.se))
Examiner: Mobyen Uddin Ahmed (mobyen.ahmed@mdh.se)
Comments:
Company contact: GKN Manasi Jayapal, IT Business Analyst | GKN ePowertrain Box 961 I 73129 Köping I Sweden Phone: +46 221 76 20 95 I Mobile +46 72 53 93 541 manasi.jayapal@gkn.com I www.gkn.com