Search Results

Neural networks have been used for engine computations in the recent past. One reason for using neural networks is to capture the accuracy of multi-dimensional CFD calculations or experimental data while saving computational time, so that system simulations can be performed within a reasonable time frame. This paper describes three methods to improve upon neural network predictions. Improvement is demonstrated for in-cylinder pressure predictions in particular. The first method incorporates a physical combustion model within the transfer function of the neural network, so that the network predictions incorporate physical relationships as well as mathematical models to fit the data. The second method shows how partitioning the data into different regimes based on different physical processes, and training different networks for different regimes, improves the accuracy of predictions.

This paper will focus on the spray characteristics of a high pressure (up to 155 MPa) accumulator type injector in a high pressure (chosen density) quiescent spray chamber. The injector uses a standard single orifice nozzle which produces a full cone spray. Using this apparatus, we are examining the fundamental aspects of high pressure spray formation under controlled conditions. Experimental data was collected using high speed photography (10,000 frames per second) which used a pulsed copper-vapor laser as a light source. Two photographic techniques are being utilized. Direct attenuation allows measurement of tip penetration, spray cone angle, and injection duration. Scattering from a sheet of laser light perpendicular to the camera field of view is being developed in an attempt to resolve inner spray cone structure. In addition to the quantitative data from the high speed photography, injector accumulator pressure, supply pressure and injection rate histories were recorded.