Universität Rostock, 2021
Abstract: During energy conversion in marine diesel engines, especially in transient operation, particles are released. The overall aim of the work is to reduce these emissions with the help of predictions. This study highlights the complexity of particle formation and the need for data-based modelling. Based on several hundred load changes, four models are developed, all of which make use of artificial neural networks. After model training, validation and testing, the models are compared and evaluated according to previously defined criteria.
doctoral thesis free access