Self-optimizing production processes

Flexible process control for robust systems

Self-optimization is an important control principle in adaptive systems. The extension of classical control principles to include autonomous goal redefinition makes it possible to establish artificial intelligence in technical systems. In conjunction with the availability of real time information, this paves the way for the creation of robust and at the same time flexible production systems even in highly dynamic Industry 4.0 environments.

In the “Integrative Production Technology for High-Wage Countries”, the Fraunhofer IPT along with institutes and facilities affiliated with the RWTH Aachen University is investigating new means of applying the principles of artificial intelligence to expand significantly the range of applications and services in which closed control loops can be used. The overriding goal is to make machines and equipment more autonomous and more intelligent in order to enhance their flexibility as well as their robustness when exposed to interference factors. Self-
optimization as an approach to flexible and reactive automation makes an important contribution in this context.

Assembly is a classical field of application. Principles of self-optimization for the assembly of high-performance laser systems are developed and applied at the Fraunhofer IPT. High-precision orientation of optical components is one of the crucial quality criteria for the assembly of such laser systems. It is vital to meet the most exacting demands in terms of manufacturing and process tolerances. Self-optimizing assembly systems can rise to the challenge by using models to interpret the sensor data. This is achieved by drawing on the optical characteristics such as beam profile, optical power or imaging quality, evaluating them and using the results in a closed control loop in order to correct the position of individual optics. In a self-optimizing system, tolerance minimization is replaced by function-oriented assembly process management. As a result, the planning effort required for complex assembly processes can be greatly simplified whilst maintaining a robust production system.  

In collaboration with the research partners in the excellence cluster, the scientists will transfer the principle of self-optimizing control from production processes to a number of other applications – from assembly through welding to optimization of weaving looms. The goal of the research work is always to achieve a significant reduction in the effort and cost involved in the initial process set-up. In this case, a previously unattainable level of flexibility was achieved – and with it, an important step toward the automation of low volume production series and the manufacture of individual products. The intensification in control of these processes in comparison with that achieved in classical applications increases their robustness and therefore their reliability very considerably, even given the rapidly changing parameters in highly flexible factory operation.