For thermal management of electric vehicles, common heat dissipation techniques including forced air cooling and liquid cooling are frequently employed. Such BTMS first seem to be highly effective, but they often increase the overall system's weight and parasitic power requirements by adding extra blowers, fans, pumps, pipes, and other accessories that are large, expensive, and complex [3]. These cooling system limitations, the increased energy consumption of parasitic loads, and the limited battery capacity can all be avoided by using a well-designed and optimized PCM system [4]. Many advantages could result from a novel thermal management approach that uses PCM as a heat dissipation source. A compact, cost-effective, and warp-resistant system without ancillary performance requirements ought to be the primary benefit of such a heat management system [4].

A temperature management system with an implemented cooling circuit requires a manufacturing process that enables the fabrication of complex and fragile structures. Furthermore, a heat resistant material is required to withstand temperature fluctuations within or around the system. Therefore, Direct Metal Laser Sintering (DMLS) is selected to be a suitable manufacturing technology for this application. DMLS is an additive manufacturing technology for fabricating metal components out of powder particles – it is assigned to the Laser Powder Bed Fusion (LPBF) technology.

The DMLS printing system consists of three different chambers as schematically shown in Table 2. These chambers are flooded with inert gas (Nitrogen/Argon – depending on the powder material) to create an oxygen-free atmosphere during the printing process. There is one movable platform in each of the chambers. First, powder material (new/recycled) is filled on the dispenser platform. Here, the powder material has to be manually compacted with the help of spatulas to avoid cavities that could lead to problems during the print job. Then, the recoater arm including the recoater blade moves across all the three platforms, starting on the right (dispenser platform). Here, the recoater blade (steel, ceremic, or soft brush – depending on the powder material and on the geometry) takes away a certain amount of powder material which is then applied on the build platform. The excessive powder material is pushed onto the collector platform. After the first layer of powder material is applied, a high power Yb-fiber laser beam fuses together the single powder particles according to a two-dimensional section of the CAD model. Splashes, which are created by the laser beam, are blown away by a continuous gas flow across the build platform. After the powder layer is exposed, the next powder layer is applied by the recoater arm. This iterative process lasts until the structure to be printed is finished [10–14].

Due to the high process complexity of DMLS, the number of influencing factors (e.g. direction of the gas flow and the direction/movement of the recoater arm) that can have a significant impact on the buildability, the quality, and also the process costs of a part is very high. To keep the risk of a negative impact of the influencing factors low, the overall printing process including its single steps has to be considered, as shown in Figure 4.

The first step of the printing process is the “Part design”. Here, the 3D-model of the part is created by using a CAD software (e.g. Autodesk^® Inventor^®) under consideration of process and material dependent design guidelines. These design guidelines contain relevant information about geometric restrictions (e.g. maximum pin diameter, minimum wall thickness). A thoughtful design can help to save support structure and therefore to reduce the process costs. At the end of the part design, the 3D-model is transferred into a STL file.

The second step is called “Preprocessing”. In this step, the STL file is loaded into a software that represents a virtual build platform (e.g. Materialise Magics^®). Here, the orientation of the part is defined and support structure is generated. The orientation of a part has an enormous influence on the part quality and on the process costs (e.g. build height). After this step, the STL files (part and support structure) are transferred into a slicing software (e.g. EOSPRINT^®). Within the slicing software, the 3D models of the STL files are subdivided into 2D-sections (slices). This step is necessary because the printing system then processes one slice after the other during the process (powder recoating/exposing). Furthermore, the appropriate printing parameter has to be selected within the slicing software. Every powder material requires a certain printing parameter that guarantees a parts’ maximum density and strength. Nevertheless, depending on the application and the part geometry, it is also possible to change some of the process parameters (e.g. laser power, laser speed, hatch distance) to avoid problems (e.g. warpage). Right after the preprocessing step is finished, the virtual print job is sent to the printing system.

Within the third step, which represents the “Print job”, the printing system is prepared for the upcoming process. This step includes the setup of the printing system (filling of powder material, application of the first powder layer, cleaning of the lens, etc.) and the initiation of the process conditions (e.g. flooding of system chamber with inert gas, heating of build platform). Subsequently, the print job can be started. When the print job is finished, the excessive powder material is removed and the build platform including the printed structures is taken out of the printer.

The fourth step is the “Postprocessing”. Here, the parts are removed from the build platform using diverse techniques (e.g. hammer and chisel, band saw, hand saw, wire cutting). Furthermore, mechanical (e.g. turning, milling) and thermal treatments (e.g. hardening) are executed within this step. In general, parts fabricated via DMLS can be treated with all kinds of manufacturing processes, same as conventional parts.

The fifth step of the process represents the “Analysis”. Dimensional inspection, functional testing, and material testing are part of this step.

The process steps of Figure 4 have a mutual influence which is linked to the final result of a print job. Technical aspects (buildability) and the economic aspects (process costs) depend on the optimum correlation between the single process steps. Therefore, it can happen that the entire development process has to undergo several iterations until the result is at its optimum.

In order to accommodate the entire energy of the LIB in the PCM, 0.20 kg of paraffin RT21 per battery cell would be necessary. Since the battery concept provides for additional external heating or cooling and in order to be able to settle the size and weight of the battery pack in a normal range, we decided to reduce the amount of thermal storage mass to 0.05 kg/cell. Furthermore, the aspects of the component design depend on the appropriate printing system. In our case, we decided to use the EOS M400 printing system. The aspects are:

The thickness of the walls in the construction is between 0.5 – 2 mm. The diameter of the pipes is 3.1 mm. The distance between the edge of the honeycomb is 4 mm. Inside the box we have designed a honeycomb structure. The upper and lower part of each honeycomb is a hemisphere so that we do not need any additional support structures inside.

To assess the feasibility and functionality of BTMS, we decided to support the development process with simulations. In today's industry, simulation is an everyday tool, especially in companies with strong engineering ties, such as the automotive industry or mechanical and plant engineering. Therefore, in science as well as in business, there is a strong demand for ways to utilize computers to provide competitive and functional products, and to manufacture and sell them efficiently, with high quality and safety. [27] Computational fluid dynamics (CFD) is a computational mathematical modeling tool that can be viewed as a fusion of theory and experiment in the fields of fluid flow and heat transfer. CFD simulations provide information about flow and fluid properties that can be difficult or expensive to obtain by measurement, providing insight and understanding of product performance or flow behavior in specific situations. CFD is now used in many industries, starting from aerospace and power generation, to consumer goods, biomedical, pharmaceutical, built environment and many more. Where CFD has historically been used to predict or understand the flow performance of existing designs or conditions. The most effective and productive usage is achieved when applied early in the design process to drive the design towards better or more consistent performance. [28] Even the simplest heat exchange system is actually quite complex when viewed through rigorous analysis. Therefore, we decided to perform flow analysis in BTMS using a computational fluid dynamics software tool that offers 3D flow analysis. CFD is a software tool for simulating the behavior of systems involving fluid flow, heat transfer, and other similar physical processes. Solve fluid flow equations over a defined range. I need to know the conditions specified at the boundaries of this region. Creating such CFD models and analyzing the results is quick and easy thanks to high computational power, powerful graphics and interactive 3D manipulation of the models. Advanced solvers contain algorithms that provide robust solutions to flow fields in a reasonable amount of time. For these reasons, CFD is now an established industrial design tool that offers a cost-effective and accurate alternative to scale model testing. There are many different solving methods used in CFD codes, but the most common is known as the finite volume method. In this technique, the region of interest is divided into smaller sub-regions called control volumes. The equations are discretized and solved iteratively for each control volume. As a result, you can get an approximation of the value of each variable at a specific point over the range, giving you a complete picture of the flow behavior. Estimating the heat release of solids due to convection can only be done by abstracting the body under consideration into very simple geometries. However, these empirical approximations lose their meaning when the geometry of the body and the abstraction deviate significantly. Heat sources with very complex geometries cannot be meaningfully represented by abstractions. Calculation of heating power can be mastered only under very simple conditions. To be able to describe the thermal behavior of battery systems with passive PCM cooling, his CFD model of the battery and surrounding thermal mass was created. The CFD tool Ansys/Fluent was used for this. [28] Passive BTMS requires the heat generated in the battery cells of the battery pack to be transferred to the PCM. The resulting heat is stored in the PCM, which should provide the ideal temperature range for the battery cells.

In Chapter 2.2, we explain how we can build a structure so that we don't need a support structure provided by the slicer. The other two critical points are cooling or heating and PCM space. We design pipes inside the structure for cooling and heating. The honeycomb construction also serves as a support framework for the tubes (see Figure 5). To guarantee that the PCM is distributed evenly, we punctured the honeycomb (Figure 5). The middle section of the construction was simulated in Autodesk CFD 2023 to generate preliminary estimations of the temperature distribution. The battery slot's wall temperature was anticipated to be 30°C. The liquid passing through was supposed to be 10°C in temperature. To provide a better idea of the temperature distribution, the temperature difference was considered to be 20°C.

As can be seen in Figure 6 and Figure 7, the temperature distribution between the front view and the side view is very different. In order to achieve an even temperature distribution, an additional cooling loop must be installed in the next step, offset by 90 degrees.