Measurement-Oriented Dynamic Synchronization of Engine and Tailpipe Emission Signals for Comparing Stationary and Dynamic Test Results

Exhaust emission assessment of heavy-duty diesel engines is commonly based on complementary steady-state and transient procedures, represented by the World Harmonized Steady-State Cycle (WHSC) and the World Harmonized Transient Cycle (WHTC). However, under transient operation, tailpipe NOx and CO2 signals cannot be directly assigned to instantaneous engine operating states because the measured response is affected by transport delay, analyser dynamics, and signal dispersion within the measurement chain. This paper proposes a machine-learning-assisted dynamic synchronization framework for aligning engine operating signals with tailpipe emissions under transient conditions. The method uses actual engine torque as the primary dynamic reference and determines local effective alignment between emission readings and the engine operating states that generated them. The synchronized data are then evaluated using an XGBoost-based modelling approach to assess whether emission characteristics obtained from WHSC steady-state operation can be transferred to WHTC transient operation. The results show that the proposed synchronization improves the physical consistency of transient emission data and provides a more reliable basis for comparing stationary and dynamic test outcomes. The transferability analysis indicates good predictive consistency for CO2, whereas NOx shows only partial transferability, reflecting stronger transient sensitivity and more complex formation dynamics. The proposed framework supports intelligent emission-data preprocessing, data-driven interpretation of heavy-duty engine tests, and assessment of the representativeness of steady-state tests under transient operating conditions.