Efficient loss modeling for automated computational design synthesis using a next-generation hybrid dual-clutch transmission

Estimating transmission power losses or energy efficiency with sufficient accuracy early in the design phase is essential to obtain an optimal transmission design. High-accuracy simulation models usually require large computation times, hence we focus on the development of fast yet accurately designed parametric loss models for the essential gearbox components. To enhance the models’ accuracy, experiment-based research into the distinct analytical loss models for the key transmission components (gears, bearings, synchronizers, clutches, and seals) is required. The main contribution of this work lies in the thorough experimental validation of the existing load-independent and load-dependent loss models with particular attention to gear mesh, clutch drag and gear wheel churning losses; verification and validation of the transmission modeling method, and providing insights into the breakdown of these losses at transmission system level using a next-generation hybrid dual-clutch transmission. The results showcase that the provided loss models with the essential component input parameters from the literature and the provided geometrical parameters of the chosen gearbox can predict the losses dynamically and over the full range of operation. Hereby, an accuracy of less than 15% error by comparing it with a large number of 75 gearboxes, measured on more than 15 different test rigs, over a certified reference drive cycle (WLTC) has been achieved. Furthermore, in most of the operating points, the predicted losses are within the experimental measurement spread, originating from hardware piece-to-piece variation and test bench-to-test bench variation.