These days many research efforts on internal combustion engines are centred on optimising turbocharger matching and performance on the engine. In the last years a number of studies have pointed out the strong effect on turbocharger behaviour of heat transfer phenomena. The main difficulty for taking into account these phenomena comes from the little information provided by turbocharger manufacturers. In this background, Original Engine Manufacturers (OEM) need general engineering tools able to provide reasonably precise results in predicting the mentioned heat transfer phenomena.Therefore, the purpose of this work is to provide a procedure, applicable to small automotive turbochargers, able to predict the heat transfer characteristics that can be used in a lumped 1D turbocharger heat transfer model. This model must be suitable to work coupled to whole-engine simulation codes (such as GT-Power or Ricardo WAVE) for being used in global engine models by the OEM. Moreover, the procedure must be capable to predict heat transfer effects using available data as external geometrical parameters of the turbocharger.To reach these several purposes, a description of the procedure to obtain correlations for heat transfer behaviour to be used in a lumped model is given. The procedure is based on several generalised correlations obtained from the evaluation of heat transfer properties of different turbochargers. The validity of the procedure is confirmed by simulations performed in GT-Power environment compared to experimental results both in a gas stand and in an engine test bench. The results of the validation show an acceptable level of accuracy using the proposed procedure. Furthermore, the advantages of the procedure are shown by comparing its results with the results of standard GT-Power simulations. In these last simulations only turbocharger maps are used without a turbocharger heat transfer model.Further analysis of these results evidences that the simple procedure of using general correlations for heat transfer properties in a lumped model is accurate enough to predict turbine outlet temperature much better than standard GT-Power model. This parameter is crucial for after treatment modelling and design as well as for two stage turbocharging.