Robot ‘modelling’ assuages cautious aerospace manufacturers

01 April 2010

The aerospace industry, which regards skilled manual assembly as the norm, is tentatively embracing robotic automation in some of its manufacturing processes. While robotics are treated with a degree of caution by this industry, recent research using modern software technologies and tools to simulate and subsequently compensate for the shortcomings have helped to allay its fears

The Aerospace Manufacturing Technology Centre (AMTC) is a new initiative born out of a partnership between the National Research Council and Canada Economic Development for Quebec regions. Located in Montréal, Canada, the aim of AMTC is to develop modern aerospace manufacturing technologies that have the potential for significant cost savings, while maintaining high levels of quality, reliability and performance.

Michel Lambert, a researcher at the AMTC, is investigating the use of robotic technologies for automating several aerospace manufacturing tasks that involve contact with the environment. These tasks include drilling of fuselage and wing parts for assembly, surface finishing tasks like polishing, grinding and machining, as well as metallic panel forming using shot peening. All of these tasks require precise and repeatable motion from the robotic systems but a particular aspect of the robot arm design makes this project very difficult, as Mr Lambert explains.

“The motor drives in the joints of a robot arm have an inherent elasticity. This results in several problems such as unwanted vibrations from the desired path and chatter, thus making it almost impossible to carry out tasks to the precision we need in many cases.” Mr Lambert sought to study the effect of dynamic parameters, such as elasticity on robotised contact tasks, and to develop suitable control algorithms that would compensate for the elasticity in the joints. To do this, he first developed a high-fidelity model of a six-degrees-of-freedom robot arm that included the joint elasticity. He used Maplesoft’s MapleSim physical modelling tool, which allowed him to define the layout of a multi-body mechanism – rigid beams, flexible beams and joints – in two and three dimensions, from which the kinematic and dynamic equations of motion can be extracted.

In a very short time, he was able to demonstrate an exact correlation between predicted behaviour described by the MapleSim model and a classical model derived from first principles. This validated the MapleSim model, which meant it could serve as a reference for both the actual robot and other analytical models developed in-house.

Now that he has developed the model, he can use the power of the Maple application to perform analyses, such as stability and sensitivity analysis, find eigenvalues and eigenvectors, and visualise the dynamic behaviour of the robot. He will then be able to validate the results of the model with the measured response of the actual robot. Mr Lambert has no doubt that the main benefit of using MapleSim and Maple is speed.

“I can achieve in a day what would normally take a week to do manually. Maple is extremely intuitive and has a great interface. The block-diagram approach is very easy to understand and the tasks document themselves very well. This is particularly useful if I have not been working on the project for some time. I don’t have to spend time trying to remind myself of what I have done. I can simply pick up from where I left off which improves my own efficiency enormously.”

The work done by Mr Lambert and the AMTC team is making significant contributions to increase the confidence of the aerospace industry in automated processes, and tools from Maplesoft are playing no small part in this endeavour.