Development of a CFD method for vehicle's thermodynamic transient cycle

The present Master Thesis represents the result of the work done at the Heat Management and Cooling performances department of Volvo Cars, settled in Göteborg, Sweden. The purpose of this work is to develop a strategy to couple CFD and thermal transient simulations, in order to decrease the computational time needed to achieve the results of the parts temperatures, while a vehicle is performing a transient drive cycle. In fact full-vehicle simulations of city cycles are really long and heavy since millions of cells are needed to accurately solve the RANS equations and since the physical time of the simulation is quite long. As a result, current fully-transient CFD simulations, which take few months to provide results, are not suitable for vehicle pre-development purposes. Moreover these cycles are also heavily transient due to the rapid and continuous changes between accelerating and decelerating conditions, so the developed method should be also able of capturing the unsteady behaviour of the parts temperatures with the right accuracy degree. The softwares involved in this coupled procedure are ANSA for the geometry clean-up and meshing of the thermal model, Taitherm for the thermal simulations and STAR-CCM+ for the CFD simulations. In addition, Cotherm is used in order to automatize the entire procedure. Experimental measurements, on different parts of the reference Volvo car S90 T6, were performed in order to validate the coupled methodology. Results demontrated that the implemented strategy is able to meet the industrial target related to time constrains since the entire procedure needs less than one week to generate the results against the time needed by fully-transient CFD simulations which is estimated based on previous works around to six months. On the other hand as regars the accuracy of the results, besides the fact that the results are physically consistent with the experimental measurements and also in line with what found in previous studies, is not enough. Margin for improvement in the methodology exists and future work to be done has been suggested in the final chapter in order to optimize few critical aspects.

Questo lavoro è il risultato del tirocinio svolto in Volvo Cars presso il dipartimento di "Heat management and cooling performances" a Göteborg, in Svezia. Lo scopo di questa tesi è lo sviluppo di una metodologia per fare simulazioni CFD dell'intero veicolo durante cicli di guida in città. Questi cicli sono di natura altamente transitoria e comportano tempi computazionali enormi. Di conseguenza, allo stato attuale, non è possibile includere questo tipo di simulazioni nelle fasi di design del veicolo, poichè al contrario i vincoli in termini di tempistiche sono molto stretti. Per quanto riguarda le analisi CFD è stato usato il software commerciale STAR-CCM+, per le analisi termiche il software TAItherm e infine è stato usato Cotherm per automatizzare l'intera procedura di accoppiamento tra software. I risultati del lavoro hanno dimostrato che questa metodologia è in grado di essere in linea con i vincoli temporali imposti all'industria dell'automotive. L'accuratezza dei risultati, pur coerenti con i dati sperimentali e in linea con i lavori precedenti, non è abbastanza. Tuttavia, il problema analizzato è molto esteso e complesso e un ampio margine di miglioramento è stato individuato per incrementare l'accuratezza di alcuni risultati.