Engine oil stress as a consequence of DPF regeneration on urban service vehicles and maintenance improvement aided by AI

Urban operating conditions are characterized by a high frequency of start-and-stop operation, extended idling periods, low average speed, and high load levels. These operating characteristics are inherently associated with a reduction in engine oil drain intervals. However, a more pronounced reduction has been observed under the current emission standard (EURO VI-D) compared to previous configurations. Modern engines equipped with EURO VI-D–compliant Diesel Particulate Filters (DPF) introduce new challenges, as effective regeneration is difficult to achieve under typical urban service conditions. This leads to increased engine oil stress, characterized by elevated soot content, viscosity increase, and potentially higher wear rates due to the abrasive nature of soot. The original oil drain period suggested by the engine manufacturer should be reduced from 60,000 km to approximately 15,000 km when average soot content in engine oil samples reaches 3.0%. Additionally, limited access to detailed documentation regarding DPF control strategies restricts the development of solutions based on a full physical understanding of the system. On this work, real data are shown related with that problem and some partial solutions for minimising the negative effects are proposed based on a data-driven approach using machine learning and explainable artificial intelligence (XAI) techniques. Using operational data from two representative buses (one assigned to a predominantly high-speed route and another operating mainly under urban conditions) machine learning models combined with XAI techniques such as SHAP and Partial Dependence Plots (PDP) were applied to interpret the DPF regeneration behavior. This analysis made it possible to identify the most desirable regeneration status and to determine the operating ranges of key variables, such as vehicle speed and engine speed, that promote the occurrence of this status. Moreover, these results can be linked to the average speed of each service line to support operational and maintenance decisions, such as identifying routes where favourable regeneration conditions are more likely to occur and prioritizing the assignment of EURO VI-D–compliant buses on those routes.