Two adjoint formulations are developed for the design sensitivity analysis of linear and nonlinear dynamic thermoelastic systems. The variation of a general design functional is expressed in explicit form with respect to variations of the design quantities, i.e. material properties, applied loads, prescribed boundary conditions, initial conditions, and shape. The design functional is expressed in terms of these explicit quantities and the implicit response variables: displacement, temperature, stress, strain, heat flux vector, temperature gradient, reaction forces, and reaction surface flux. The convolution is employed, in lieu of the time mappings used in other approaches to account for transient response. The design sensitivities for dynamic thermoelastic, nonlinear uncoupled dynamic thermoelastic problems are presented here for the first time. The use of the convolution for sensitivity analysis is also presented here for the first time. By formulating these sensitivities, numerical optimization algorithms currently used for elastostatic structure redesign, can be applied to nonlinear dynamic thermoelastic systems. Sensitivities for linear elastodynamic and dynamic thermoelastic problems are derived by incorporating the reciprocal theorem between a variation of the real design and an adjoint system. The Lagrange multiplier method is utilized to formulate sensitivities for nonlinear uncoupled dynamic thermoelastic problems. These formulations may also be specialized for linear and nonlinear transient conduction problems. The finite element method is used to demonstrate the numerical implementation of the formulations. Shape sensitivities are evaluated and compared with finite difference calculations to validate the results. A numerical optimization algorithm is combined with the sensitivity analyses to redesign the geometry of a mold in a casting problem and the shape of structure whose response is governed by a nonlinear quasi-static thermoelastic system.

Extensive numerical methods for computing design sensitivity are included in the text for practical application and software development. The numerical method allows integration of CAD-FEA-DSA software tools, so that design optimization can be carried out using CAD geometric models instead of FEA models. This capability allows integration of CAD-CAE-CAM so that optimized designs can be manufactured effectively.

Extensive numerical methods for computing design sensitivity are included in the text for practical application and software development. The numerical method allows integration of CAD-FEA-DSA software tools, so that design optimization can be carried out using CAD geometric models instead of FEA models. This capability allows integration of CAD-CAE-CAM so that optimized designs can be manufactured effectively.

Multidisciplinary design optimization (MDO) has recently emerged as a field of research and practice that brings together many previously disjointed disciplines and tools of engineering and mathematics. MDO can be described as a technology, environment, or methodology for the design of complex, coupled engineering systems, such as aircraft, automobiles, and other mechanisms, the behavior of which is determined by interacting subsystems.

Extensive numerical methods for computing design sensitivity are included in the text for practical application and software development. The numerical method allows integration of CAD-FEA-DSA software tools, so that design optimization can be carried out using CAD geometric models instead of FEA models. This capability allows integration of CAD-CAE-CAM so that optimized designs can be manufactured effectively.

Extensive numerical methods for computing design sensitivity are included in the text for practical application and software development. The numerical method allows integration of CAD-FEA-DSA software tools, so that design optimization can be carried out using CAD geometric models instead of FEA models. This capability allows integration of CAD-CAE-CAM so that optimized designs can be manufactured effectively.

Extensive numerical methods for computing design sensitivity are included in the text for practical application and software development. The numerical method allows integration of CAD-FEA-DSA software tools, so that design optimization can be carried out using CAD geometric models instead of FEA models. This capability allows integration of CAD-CAE-CAM so that optimized designs can be manufactured effectively.