Digital Twins are core enablers of smart and autonomous manufacturing systems. Although they strive to represent their physical counterpart as accurately as possible, slight model or data errors will remain. We present an algorithm to compensate for those residual errors through Reinforcement Learning (RL) and data fed back from the manufacturing system. When learning, the Digital Twin acts as teacher and safety policy to ensure minimal performance. We test the algorithm in a sheet metal assembly context, in which locators of the fixture are optimally adjusted for individual assemblies. Our results show a fast adaption and improved performance of the autonomous system.

Applying the concept of Digital Twin in production processes supports the manufacturing of products of optimal geometry quality. This concept can be further supported by a strategy of finding the optimal combination of individual parts to maximise the geometrical quality of the final product, known as selective assembly technique. However, application of this technique has been limited to assemblies where the final dimensions are just function of the mating parts' dimensions and this is not applicable in sheet metal assemblies. This paper develops a selective assembly technique for sheet metal assemblies and investigates the effect of batch size on the improvements. The presented method utilises a variation simulation tool (Computer-Aided Tolerancing tool) and an optimisation algorithm to find the optimal combination of the mating parts. The approach presented is applied to three industrial cases of sheet metal assemblies. The results show that using this technique leads to a considerable reduction of the final geometrical variation and mean deviation for these kinds of assemblies. Moreover, increasing the batch size reduces the amount of achievable improvement in variation but increases the amount of achievable improvement in the mean deviation.

Selective assembly is a means of obtaining higher quality product assemblies by using relatively low-quality components. Components are selected and classified according to their dimensions and then assembled. Past research has often focused on components that have normal dimensional distributions to try to find assemblies with minimal variation and surplus parts. This paper presents a multistage approach to selective assembly for all distributions of components and with no surplus, thus offering less variation compared to similar approaches. The problem is divided into different stages and a genetic algorithm (GA) is used to find the best combination of groups of parts in each stage. This approach is applied to two available cases from the literature. The results show improvement of up to 20% in variation compared to past approaches.

A preeminent factor in the geometrical quality of a compliant sheet metal assembly is the fixture layout that is utilized to perform the assembly procedure. Despite the presence of a great number of studies about the optimization of assembly fixture layouts, there is not a comprehensive algorithm to optimize all design parameters of fixture layouts for compliant sheet metal assemblies. These parameters are the location and type of hole and slot in each part, the slot orientation, and the number and location of additional clamps. This paper presents a novel optimization method that optimizes all these parameters simultaneously to maximize the geometrical quality of the assemblies. To attain this goal, compliant variation simulations of the assemblies are utilized along with evolutionary optimization algorithms. The assembly springback and contacts between parts are considered in the simulations. After determining the optimal design parameters, the optimal positions of locators are fine-tuned in another stage of optimization. Besides, a top-down design procedure is proposed for applying this method to multi-station compliant assemblies. The presented method is applied to two industrial sample cases from the automotive industry. The results evidence a significant improvement of geometrical quality by utilizing the determined fixture layout from the presented method compared with the original fixture layouts of the sample cases.