Generalized Thermal Agents with Multiple Boundary Conditions and Three-Dimensional Thermal Agents

The temperature profile across an object can be computed by iterative methods. However, the time required for iterative solutions to converge for multiple objects in a complicated configuration impedes the exploratory analysis of engineering systems. A rapidly computed initial guess could greatly increase the speed of convergence for an iterative thermal solver. This study continues research testing various systems for creating thermal agents that provide such initial guesses. During an off-line training process, genetic programming is used to locate a thermal agent by training in several sets of boundary conditions. It is demonstrated here that thermal agents with two boundary temperatures require only one training case. This study examines thermal agents with three-boundary temperatures conditions using different physical geometries. In addition the first appearance of three-dimensional thermal agents appears here.