Study on thermal placement optimization of 3D high-power microwave module

This paper takes a 3D high-power microwave module as research object and analyses the law that temperature distribution, thermal-via and design variables influence on internal maximum junction temperature of the module base on the model established with finite element simulation software. In view of limitation of the traditional single varying parameters appraisal method for thermal analysis, the analysis method which combines orthogonal experiment and response surface method is applied to analyze thermal characteristics of 3D high-power microwave module under the influence of mutual parameters. Hybrid optimization algorithm which combines BP neural network with the genetic algorithm is applied to the optimization for chip thermal placement of embedded power chip microwave module in connection with influence of chip placement effect on the temperature distribution. Internal chip coordinates are selected as design variables and internal chip temperature are the response variables. The optimal placement scheme which makes internal junction temperature lowest with the BP-GA hybrid algorithm proposed in this paper can be obtained. The results show the structure and material parameters of the cooling plate and heat metal sheet have greater impact on module temperature, and the structure and material parameters of the substrate and encapsulation layer have less impact temperature. The quadratic regression equation in the paper can accurately represent the relationship between thermal parameters and internal module temperature, and the BP network prediction model for chip thermal placement can accurately express the functional relationship between the chip placement and the maximum junction temperature of the internal module. The BP-GA hybrid algorithm proposed in this paper can get an optimal solution of chip thermal placement, and it lay a good foundation for the thermal analysis and thermal placement of 3D high-power microwave module.