Open Collections

Abstract

The development of natural gas fueling for diesel engines is a new solution to reducing diesel engine pollutants. To convert existing diesel engines to support the use of natural gas, a new type of injector must be designed. A model to predict the hydraulic and mechanical operation of injectors that introduce the natural gas into the combustion process for diesel engines will significantly lower design times and development costs. The developed injector model couples the physical kinematics of the mechanical components with the fluid mechanics of the diesel fuel and the compressed natural gas. Development of the injector model began with the modeling and validation of the diesel injector, followed by modeling and validation of the more complex gas injector. The final version of the injector simulation uses a finite difference method, where specific reservoirs at important regions in the injector are connected by passages separated into volumes of finite length. The models are validated using experimental means with the use of an Injector Test Rig. Both a diesel injector and gas injector was modified to allow the capture of hydraulic pressure data within the injector. Comparisons of model results and experimental results of this hydraulic pressure show excellent agreement for both injectors. A device was also designed for the gas injector to measure the relative timing of the individual diesel and gas jets. Comparison of the diesel and gas needle lifts from the gas injector model with the experimental data obtained from this device shows good agreement. The final version of the injector model accurately represents both the diesel and dual fuel injectors within an acceptable amount of error associated with experimental conditions. A Lax-Wendroff velocity diffusion algorithm is used to eliminate the instabilities associated with the numerical model. The accuracy of the injector models primarily depends on the precise geometrical representation of the injectors with a secondary effect coming from effect of discharge coefficients or leakage tolerances. A study of the cycle to cycle variability for the experimental data was performed but was identified as being insufficient to establish complete confidence in the consistency of the experiment.