Predictive Diesel Combustion Using DI-Pulse in GT-Power

Abstract

Rising fuel prices and stricter emission regulations have put a great demand on developing new engines with high fuel efficiency and low emissions. This has led to the development of several new concepts such as different types of EGR, variable valve timing, different injection strategies, turbulence enhancement techniques, etc. Virtual simulations, particularly 1D simulation tools for gas exchange have played a critical role in the last decade to reduce the lead time for the development of these new concepts. These simulation tools employ a 0D combustion model. However, one of their major limitation is the use of a non-predictive or fixed burn rate combustion model. With this kind of model it is not possible to evaluate the above mentioned concepts with reliability. In order to overcome this problem major tool developers such as Gamma Technologies have developed a predictive combustion model, which can predict the combustion rate based on the in cylinder conditions. But these models can provide reliable results only if they are well calibrated against test data. The aim of this thesis was to calibrate a predictive combustion model for a Diesel engine in GT-Power using the DI-Pulse combustion object. The performance of the fully calibrated model was evaluated by analysing its capability to predict key operating parameters such as IMEP, CA at 50% burn, peak pressure and NOx . In order to calibrate the model, data was collected by conducting tests in a single cylinder test cell and was subsequently validated thoroughly before using it for calibration. It was concluded that the model was able to predict the key operating parameters mentioned previously within the suggested thresholds except for NOx at low loads and low speeds and peak pressure at high loads and high speeds.