Inverse modeling and industrial processing

The Next-Generation Manufacturing Project (NGM) state Simulation and Modeling as one of the key issues to meet the future challenges in manufacturing enterprises; "one which can drive the rate of implementation of new products and technology". They regard Simulation and Modeling in a broad perspective, involving all company decisions from management to manufacturing processes. This thesis will address the problem of virtual prototyping, the introduction of new products typically initiates the creation of a mock-up and trial and error methods. By virtual prototyping the trial and error methods can be moved from the shop to a virtual environment. The virtual environment enables fast evaluations of different designs, and the designer gains further knowledge and understanding of the simulated process through visualization. With better insight, adjustments can be made much more efficient. This is presently the typical utilization of virtual prototyping. The process simulation and modeling will be addressed within the metal forming domain. The domain has received increasing academic and industrial interest during the last decade due to the development of both stable, robust and reliable simulation software. Metal forming is, therefore, regarded as a fairly matured domain where highly advanced simulation software and constitutive models are available. Moreover, the area is subject to continuing development and improvement. An example of design optimization combined with numerical simulation is given in the paper: Least Square Formulation of the Object Function, applied on Hydro Mechanical Tube Forming, the paper was submitted to the She-Met 2001 conference. The objective is to generate an optimal set of process parameters, i.e. internal pressure and material feeding. The problem is solved through optimization where the objective is defined as: * Minimize the geometric derivation between the part/tube and an ideal geometry * The algorithm should intent to utilize a uniform thickness distribution. The utilization of simulation tools initiates calibration of the applied model, i.e. coherence between the simulated output and physical output through calibration of constitutive parameters. NGM lists this as one of the barriers for further implementation of simulation tools in manufacturing enterprises. Due to the lack of verified usable models and unsystematic calibration and maintenance of parameters. Inverse modeling is regarded as a solution to the problem of identification of constitutive parameters; the method supplies a systematic approach both for identification and calibration of constitutive parameters. Furthermore, the method could provide a systematic strategy for updating constitutive parameters by establishing a feedback loop from the production to the virtual product. The problem is to identify data from the production which is easy to sample and which contains a sufficient amount of information to provide significant gradients. Supervisor: Karl Brian Nielsen. Project duration: Fall 1999-Fall 2002. (Benny Endelt Kristensen)