Diesel engine pollutant emissions legislation is becoming more and more stringent. New driving cycles, including increasingly severe transient engine operating conditions and low ambient-temperature conditions, extend considerably the engine operating domain to be optimized to attain the expected engine performance. Technological innovations, such as high pressure injection systems, Exhaust Gas Recirculation (EGR) loops and intake pressure boosting systems allow significant improvement of engine performance. Nevertheless, because of the high number of calibration parameters, combustion optimization becomes expensive in terms of resources. System simulation is a promising tool to perform virtual experiments and consequently to reduce costs, however models must account for relevant in-cylinder physics to be sensitive to the impact of technology on combustion and pollutant formation. In particular, soot is one of the major pollutants of Diesel engines and its kinetic is highly dependent on local mixture properties into the cylinder. This is a challenge for 0-Dimensional (0D) combustion approaches, as it implies 3-Dimensional (3D) phenomena. In this work, to tackle this aspect, the 0D Dual Flame Model (DFM) combustion model was enriched with quasi dimensional features based on the conceptual spray combustion model proposed by Sandia National Laboratories (SNL). This allows to identify local key phenomena depending on mixture thermochemical properties driving soot kinetics. The model was tested on a comprehensive experimental database generated at IFP Energies nouvelles (IFPEN), to investigate the capability of the new approach to predict the impact of engine operating conditions, injection strategy and dilution rate on soot emissions. The quality of the results and the reduced computational time make this approach suitable for engine design and control activities.