In this work, three-dimensional computational fluid dynamics (CFD) studies of sulphur oxides (SOX) and sulphuric acid (H2SO4) formation processes in a large, low speed two-stroke marine diesel engine are carried out. The current numerical study aims to investigate the conversion of sulphuric dioxide (SO2) to sulphuric trioxide (SO3) and the possibility of H2SO4 condensation which are the prerequisites to better understand the corrosion-induced wear phenomenon. This is achieved with the aid of the implementation of a multicomponent surrogate model, which comprises a skeletal n-heptane mechanism and a reduced sulphur subset mechanism. In the present work, performance of the coupled CFD-chemical kinetic model is evaluated using both qualitative and quantitative methods. The modelling results show that the temporal and spatial evolutions of SOX predicted by the skeletal model are similar to those by the base mechanism. Predictions of the variations of SOX and the associated SO2 to SO3 conversion in response to the change of fuel sulphur content, swirl velocity, start of injection, scavenge pressure and humidity qualitatively agree with numerical and experimental results from the literature. The model is further evaluated using the measured SO2 to SO3 conversion levels in a low load, low scavenge pressure case and a low load, high scavenge pressure case. The absolute values of simulated and measured conversion levels are close, although the former appear to be higher. The current results show that the flame impinges at the cylinder liner near top dead centre. The gas is cooled rapidly by the wall temperature and H2SO4 is produced in the region where the local temperature is less than 600 K. Based on the flue gas correlation, the acid dew point temperature is higher than the wall temperature, suggesting that acid condensation may begin early at the top part of the cylinder liner. The predicted distribution corresponds well with the distribution of corroded parts observed in service engines. The model is expected to serve as an important tool to simulate the rates of SO2 absorption into lubricating oil film and H2SO4 condensation in this combustion system. (C) 2017 Elsevier Ltd. All rights reserved.