Modélisation du procédé de meulage sous-marin pour la réfection robotisée de surfaces d'acier

Abstract

Hydroelectric dam gâtes are constituted of metallic structure embed in concrète. Over time, corrosion and water pressure damage the built-in structures. To ensure their long-term sustainability, they must be periodically repaired. A submersible mechanical and control System has been developed by the Institut de recherche d'Hydro-Québec (IREQ) to diagnose and conduct the maintenance procédure underwater. A 3D scan of the surface to be repaired is firstly obtained using a laser caméra vision System. Then, it is planned to integrate a grinding robot which will be controlled from the surface to re-grind the damaged portion of the underwater structure (this part of the project is still in development). This Master degree's thesis présents an approach for modeling the material removal rate (MRR) of an underwater grinding process using straight snagging wheel. The objective of this research work is to détermine the MRR in terms of the process parameters such as wheel speed and grinding power over a wide range of grinding diameters. Grinding is a complex process influenced by many parameters. The literature contains a high number of papers on the topic, but none concems directly underwater précision grinding. The understanding of the process is the key for accurate repair of the built-in structures. In order to better understand the underwater grinding process and to develop a modeling strategy, a bench test has been created to conduct underwater experiments. Two instruments hâve been assembled to measure important aspect of the process. In a first step, a rétractable linear probe is designed and used to measure the removed volume of steel between two consécutive sets of grinding passes. Being submersible, this tool reduces expérimental measurements by up to 10 times compared to out-of-water measurements. Secondly, a load cell table has been constructed. The workpiece can be installed on it and the normal grinding force can be measured during underwater grinding experiments. A software, based on motor identification and empirical relations among System and process parameters is developed with graphical user-friendly interface. Expérimental results showed that it leads to good prédictions of the depth of eut for underwater grinding using this software. A comparative study for dry and underwater grindings is also conducted by experiments and presented in this thesis. Finally, an optimization algorithm is proposed to facilitate the sélection of process parameters. A simulation shows that important gains can be made using this optimization tool. In this particular case, it is predicted that the grinding maintenance for a single structure can be reduced by 4 hours, which is équivalent to 15%) saving of time and cost.

Date	9 Apr 2010
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Zhaoheng Liu (Supervisor) & Henri Champliaud (Co-supervisor)