donderdag 28 mei 2015

Faster Production with Evolutionary Optimisation

Robot manipulators, and the automation systems using them, face increasingly demanding and complex requirements. The pre-programmed control of such automation systems specifies how the machines operate and how the material is moved from machine to machine. To obtain the maximal production rate, it is essential that the control is optimised for each product. Simple trial and error by manually making ad hoc changing is not very reliable to do this. It requires a lot of expertise from operators and is very time-consuming and costly. Researchers from the Research Team Production Technology West at University West in Trollhättan, Western Sweden, investigate methods and algorithms to support reprogramming and optimisation of the automated control of industrial production systems. The goal is to enable automatic offline reprogramming and optimisation of complex industrial production systems. This would be very time- and cost-saving for the industry.

In this research project, there is a close collaboration between University West and the Volvo Cars Corporation. This allows the researchers to test their newly developed methods, algorithms and tools on real-world problems. Thereby gaining knowledge about the performance, applicability and robustness of those methods on industrial problems in a real-world context. From the results of these real-world tests, the Volvo Cars Corporation gets a very reliable indication of how much it can benefit from this automatic reprogramming and optimisation.

At the current stage of this research project, a novel evolutionary algorithm is developed and tested to increase the production rate by optimise the control of material handling. This work is presented in Emile Glorieux’s licentiate thesis titled Constructive Cooperative Coevolution for Optimising Interacting Production Stations. The control of the material handling determines when and how fast the robots move that pick up products from one machine and place these into the next machine. Evolutionary algorithms for optimisation have been around for several decades. These algorithms are inspired by biological evolution and include mechanisms such as reproduction, recombination, mutation and competitive selection. They have been successfully applied on many different types of problems. The evolutionary optimisation algorithm developed in this research project at University West is specifically designed for optimising the control of production systems. This optimisation algorithm works in concert with a computer simulation model of the production system. This is convenient because simulation models are already largely used in the industry. Using a simulation models enables that the optimisation is done offline, without interrupting production.   

The developed evolutionary optimisation algorithm was tested in a case study. Several sheet metal press lines from Volvo Cars Corporation that produce body panels were optimised. These body panels are stamped in the multiple presses in a press line, as illustrated in Figure 1. Later in the production process, these panels are then welded together to constitute the car body. Due to the high investment cost, it is crucial to get the maximal production rate and consequently maximal return on investment. To get the maximal production rate, the transfer of plates between the different presses requires that the coordination between robots and presses (i.e. the material handling) is optimised. The size and the power of the presses makes it difficult and dangerous to online manually optimise it. Hence, the need for offline optimisation tools and support.

Figure 1: 2D drawing of sheet metal press line
In the case study, the performance of the developed evolutionary optimisation algorithm is compared with online manual optimisation by experienced operators and also with other optimisation algorithms. The results show that the developed evolutionary algorithm achieves a production rate that is up to 18% better than manual optimisation. Whereas, the performance of the other optimisation algorithms is not better than manual optimisation.

In the next stage of this research work the goal is to not only maximise the production rate but also minimise the energy consumption and the equipment wear at the same while. The relevance of the energy consumption is to further reduce the operating cost and increase the return on investment of a sheet metal press line. On the other hand, the equipment wear is of importance to avoid excessive wear and consequently more frequent production interruption for maintenance interventions.

For more info, contact Emile Glorieux or contact Research Team Production Technology West

Geen opmerkingen:

Een reactie posten

Klik rechts onder het commentaar-kader op "Aanmelden via e-mail" indien u via mail op de hoogte wilt blijven van de nieuwe reacties op deze post.