Model Development and Design Optimization of a Novel Reciprocating Engine Technology

A novel reciprocating steam engine technology that utilizes reed valves has been developed, prototyped and patented by researchers at the Lawrence Livermore National Laboratory (LLNL) in Livermore, CA. To assist in proper sizing of this new technology in follow-on development efforts, and to better understand the interactions between various parameters, a system level computational model of the engine was developed from first principles. This model was developed for the express purpose of performing design optimization studies of the engine technology, and thus various modeling decisions were made in an effort to balance desired model accuracy with necessary computational speed. The developed model takes as inputs various environmental, geometric and kinematic parameters of the engine system and calculates the resulting power, work, torque and thermal efficiency of the proposed engine design. The model consists of numerous sub-models including a flow model for the intake fluid physics as it enters the engine, a dynamic model of the intake valve response, a pressure model of the engine cylinder, a kinematic model of the engine piston movement, and an output model that determines engine performance parameters. In order to capture the performance of the engine over time, a crank angle discretization strategy was employed and each of the engine design sub-models was evaluated for each crank angle position considered producing results based on the data obtained from the sub-model evaluations at the previous crank angle position. This strategy was determined to be necessary for accurately modeling the performance of the engine over time and crank angle position, but obviously created a computational effort challenge in that it required that various flow models and differential equations be solved iteratively within the overall model. To produce a model with sufficient computational speed to be useful within the desired optimization studies various simplifying assumptions and modeling approximations were utilized. The model was tested by performing a set of multi-objective design optimization case studies on the engine model using the geometry and operating conditions of the prototype engine developed by LLNL as a baseline. The results produced were determined to properly capture the fundamental interactions of the engine and demonstrated that the design of engine technology could be improved over the baseline through the use of the developed model.