Removing the heat quickly from the power electronics within Navy power systems, such as those planned for DDG-1000, enhances their performance. Reducing the junction temperature of semiconductor power electronic devices enables them to operate at higher electrical current flows. Lowering operating temperatures also reduces thermal stresses on the devices, which leads to improved efficiency and reduced failures. SolidWorks Model of IGBTs, Face plate, and Current Cold Plate from which Algor FEA (Finite Element Analysis) thermal management simulations can be constructed ALGOR based model, using dimensional information from SolidWorks (see Figure 1-1), forms the basis for further advanced analysis. These analytical software tools enable high-level thermal modeling of some of the many possible thermal management solutions. Once the screening is completed, only those solutions with high probability of success can be analyzed using a more elaborate FEA modeling. The prime thermal management candidates identified through the FEA modeling are tested and evaluated in the EMPF laboratory.The existing design uses Thermal Interface Materials (TIM) between the IGBT package and the Face plate and between the Face plate and the Cold Plate. The Face plate serves as a heat spreader, and the Cold Plate removes heat away from the entire assembly. The exploded view of the entire package appears in Figure 1-3 . The model shows the mating of the various heat conducting elements responsible for conducting heat away from the IGBT heat source. To simplify this analysis, heat removal due to convection is neglected since it is small in comparison to heat removal due to conduction.

Since it is cost prohibitive and time consuming to laboratory test all possible combinations and parameters for all configurations of TIMs, coolants and cold plate designs, Algor Thermal Analysis is used as a filtering tool to reduce the number of possible solutions to a selected few by performing “what if” analyses. Once the screening is completed, only those solutions with high probability of success can be analyzed using a more elaborate FEA modeling. The prime thermal management candidates identified through the FEA modeling are tested and evaluated in the EMPF laboratory.

For developing any FEA model, first a 3D solid model of the existing mechanical design needs to be created using any available CAD (Computer Aided Design) software. In this case, SolidWorks 2006 software was used to create the 3D Solid Model as shown in Figure. The Solid Model consists of the four main elements, namely IGBTs, TIMs between the two sets of interfaces, face plate and the cold plate. This solid model is used as an input to the first step of the FEA model in creating a nodal mesh network. Algor FEA software uses the physical dimensions from the 3D SolidWorks Model as input to generate a mesh network using a proprietary algorithm. Once the mesh network is generated, the boundary conditions are applied.

Thermal analysis needs the velocity profile of the coolant running through the cold plate copper tubes. These velocity profiles are computed using the CFD (Computational Fluid Dynamic) simulation part of the Algor software. Once the velocity profiles are generated, the data is used as an input to the thermal analysis model.

The water coolant velocity profile, along with many other properties of the IGBTs, face plate, cold plate, and TIMs, such as specific heat, mass density, thermal conductivity, and others are used by the ALGOR thermal model to calculate the 3D temperature profile. The input to the thermal analysis is simulated as a heat flux imposed on the top surface of the face plate with uniform density.

Figure 4-3 shows the ALGOR output temperature profile of the face plate surface to which the IGBTs are attached. This particular Algor model shown is one of the many possible “what if” scenarios, using water as coolant and aluminum as the TIM. This temperature distribution on the Face plate is a good representation of the critical temperature distribution on the IGBT surface since the two are in intimate contact. These results allow the EMPF to construct the laboratory experiments that will validate cold plate performance.

Figure 4-3. Typical temperature map result of thermal modeling using the Algor and Solid Works software in the case of the current cold plate design using water coolant and aluminum TIM (Thermal Interface Material) to thermally manage the three- IGBT assembly shown in Figure 4-1. Examples of the maximum temperatures in the hottest spots of the plate in multiple “what if” scenarios using this design are shown in Figure 1-4.