Structural Optimization of the Drift Chamber at FermiLAB

The drift chamber model design involved several analyses that made use of a number of multi-disciplinary commercial software packages whose calculations were coupled using a multiobjective optimization software framework.

First the chamber’s working conditions including the operating temperature, the in-service loads and the boundary conditions were determined.

Next sets of parametric simulations for each particular loading condition were set up in Ansys Workbench:

• a parametric CAD model representing the geometry of the chamber was developed in Design Modeler
• the mesh generation and the Finite Element Analysis (FEA) were set up in the Ansys simulation environment. The FEA analyses involved one Static Structural Analysis, one Linear Buckling Analysis and one Modal Analysis to determine the Natural Frequencies and Natural Modes of the Drift Chamber in various configurations

The composite material configuration covering material sets, reinforcement types and lay-up configuration was generated parametrically using ESAComp, and was coupled with the Ansys tools to feed them with the updated materials definition for each particular design configuration that was evaluated. The parametric definition and its assignment to the geometrical model of the material properties automatically accounted for the manufacturability of each evaluated material system by the coupling in Ansys Composite PrePost (ACP) of the structure with the mesh, geometry sources, ply draping over the mold, ply orientation, fiber-angle correction and ESAComp for the material properties and definitions. Conventional tape laminates and sandwich laminates were analyzed as possible material system solution; the sandwich laminates proved to be the best material choice with respect to the structural objectives that had been set for the project. The material selected consisted of unidirectional (UD) carbon-epoxy prepegs produced by SAATI for the skin reinforcement and carbon foams produced by Graftech for the core portion of the sandwich structure.

ACP was further used to post-process the results of the FEA analyses since local stress components, Saftey Factors (SF) or Reversed Savety Factors (RF) need to be accessed at a single lamina level to evaluate the compliance of each design with the established structural requirements. These requirements expressed in terms of Strength, Stiffness and Weight of the structure were directly calculated within ACP and then used as Objective Functions to be either maximized or minimized during the solution of the overall Design Optimization Problem which was set up in a multi-optimization software framework.

The multioptimization software framework used to couple different software simulation tools and to provide an Automatic Adaptive Design Improvement Strategy by means of Optimization Algorithms,  known as Genetic Algorithms (GA),  was modeFRONTIER.  A value was assigned to each of the deisgn variables/parameters within the multi-optimization software framework to generate a possible candidate design. The required sequence of simulations was run and the results of each simulation were collected in order to rank each design. This was done multiple times for different sets of parameters (designs). The initial population of designs was generated using Design of Experiment Techniques (DOE) and was used to train the optimization GA which was then used to generate designs that produced results that met the prescribed objective functions as close as possible.