The present work deals with the experimental investigation of additively manufactured models with the objective to determine the influences of different surface roughness and cooling channel configurations on the associated heat transfer behaviour and the pressure loss. Since the manufacturing-related roughness cannot be easily scaled to larger dimensions, multiple real-scale models of a cooling channel structure were produced and investigated. The model size on the scale of millimetres and the requirement to obtain flow conditions similar to those in the application pose particular challenges to the test rig, the measurement data acquisition and the experimental evaluation. A new evaluation methodology is needed to obtain locally resolved information about the heat transfer on the inaccessible inner surfaces of the cooling channels. In transient experiments, infrared measurements are conducted on the outer, optically accessible side, from which the heat transfer on the inner side is determined. By a numerical simulation, the inverse method is validated and the expected differences between direct and inverse observation are shown. The applicability of the inverse method is demonstrated on a small-scale model with negligible surface roughness. To compare the determined local heat transfer distributions, heat transfer measurements are performed on an up-scaled model using the well-established transient liquid crystal method. The potential measurement errors occurring in the heat transfer experiments with both measurement methods are determined by uncertainty analyses.weiterlesen