Improving the efficiency of EVAP system development by utilizing CAE simulations and surrogate models

Evaporative gases from fuel tanks in cars and motorcycles are harmful to human health, which prompting the implementation of strict regulations. To meet these regulations, many gasoline vehicles use evaporative emission control systems (EVAP systems) that utilize carbon canisters. Developing EVAP systems is not just about meeting these rules but also about making sure the engine drivability is not affected. Moreover, as future regulations become increasingly strict, there will be a growing demand for higher-performance systems that reduces evaporative gas into the atmosphere. This will make the control systems more complicated and increase production costs. To overcome these challenges, more efficient ways of development are necessary. One effective method is model-based development (MBD) using CAE technology. In recent years, CAE simulations have become more advanced, and combining different types of CAE simulations allows for faster and more accurate results. Another approach that has been gaining attention is the use of surrogate models. These models use machine learning to predict the results of CAE simulations, helping to save a lot of time while keeping results accurate. This paper discusses a case study on how MBD with CAE simulations and surrogate models that can make EVAP system development more efficient. One major challenge in MBD is the balance between calculation time and accuracy. By combining CAE simulations and surrogate models in the right way during the development process, this study shows that EVAP system development can be improved.