In recent years, not only CO/HC/NOx regulations have been restricted but Particulate Matter (PM) regulations have also become stricter for gasoline engines. This is especially true in China and Europe, where not only PM weight but also Particulate Number (PN) is included in the regulations. Reduction of PM and PN is an urgent issue for gasoline engines. On the other hand, the development of Gasoline Particulate Filter (GPF) technology that reduces PM and PN has been promoted around the world. Honda has been selling GPF-equipped vehicles since 2018 and the current layout was basically an adopted Honda system which has two pieces of Three-Way Catalysts (TWC). The high performance GPF is required not only for filtering particulate matter and decreasing particulate number but also for maintaining a low pressure drop and a high purification performance. Therefore, it is important for the high performance GPF to maximize those three performances with the best balance. However, it is difficult to find quantitatively prime factors affecting GPF performances because there are a lot of features, especially concerning the complicated GPF microstructures. Therefore, in our previous research, we conducted a parameter study of microstructure factors for the three GPF performances. As a result, it was found that surface pore diameter and the average of large neck diameter were one of the prime factors, but only those two factors were not sufficient to explain the three GPF performances. Hence, the objective of this study was set to find the other prime factors. For that, a machine learning model which correlates the GPF performances and descriptors such as the GPF structures, the structures of substrate, and the amount of catalyst, was made using previous research data. In addition, to enhance the accuracy of the machine learning model, descriptors were selected based on the test result. As a result, it was found in a detail that the prime factors were several kinds of features concerning about the way of coating catalyst.

Mr. Atsushi Furukawa, Assistant Chief Engineer, Honda R&D Co., Ltd.