Change search
ReferencesLink to record
Permanent link

Direct link
Analyses and optimisation of component selection algorithm for bearing assembly at SKF
2001 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

This thesis present a study of the process to assembling ball bearing components and is intended to increase knowledge about the ball bearing assembly process. The work has been done at SKF Group Manufacturing Development Centre (MDC) in Gothenburg in cooperation with the Department of Materials and Manufacturing Engineering at LuleƄ University of Technology. The pairing algorithm used today is a very static one that makes no consideration of the process state i.e. the target for the pairing algorithm is not responsive to the status in the assembly process. The objectives were to use knowledge about the pairing process to develop a more dynamic pairing algorithm and also to make an feedback loop which gives compensation signals to grinding machines based on recorded measurements of rings received in the pairing buffer. To find the best pairing algorithm, a model of the bearing assembly process was built in the manufacturing simulation program Witness2001. This model is a simplification of the assembly process and does not involve all machines and operations. Simulations with different pairing algorithms showed the possibilities to significantly increase pair-ability by just implementing a dynamic pairing algorithm. The potential problem to implement this is that the assembly process may be slower and increase lead-time for a given batch size. When using a feedback loop to track grinders it is important to keep the number of rings between grinding and measuring as small as possible. Even a small increase in distance will decrease the pair ability and the system will become unstable. The problem with the tested feedback algorithms was that they did not take any consideration of drift over time for grinders. If drift is large, changes in grinding targets must also be made quickly. It should be no problem to make feedback algorithms that compensate the grinders. To make this algorithm successful it is necessary to do a more complete investigation on a specific system. It is important to collect machine capability data so a complete system can be modelled in Witness. A model with all machines from grinding to pairing will make a sensitivity analysis of the new system possible.

Place, publisher, year, edition, pages
Keyword [en]
Technology, SKF, bearing assembly, Witness, simulation, process, control, Ishikawa
Keyword [sv]
URN: urn:nbn:se:ltu:diva-50005ISRN: LTU-EX--01/355--SELocal ID: 74c7053d-d94d-42da-a4ec-daec43d94172OAI: diva2:1023358
Subject / course
Student thesis, at least 30 credits
Educational program
Industrial and Management Engineering, master's level
Validerat; 20101217 (root)Available from: 2016-10-04 Created: 2016-10-04Bibliographically approved

Open Access in DiVA

No full text

Search outside of DiVA

GoogleGoogle Scholar

Total: 7 hits
ReferencesLink to record
Permanent link

Direct link