TS-71
A Bond Graph Model Library for Modelling Diesel Engine
Transient Performance
Eilif PEDERSEN* and Hallvard ENGJA**
ABSTRACT
Using simulation as a tool to evaluate system performance is often argued to be of vital importance for future development of complex machinery systems. In reality efficient and successful use of simulation depend on the availability of mathematical models of correct level of complexity for the problem to be analyzed, in a format that can be efficiently utilized and a flexible modeling and simulation environment.
In this paper a library of submodels for modeling of diesel engine transient performance is developed using the bond graph method as a framework for both model representation and as the unifying modeling methodology. Submodels for the diesel engine cylinder including combustion, heat transfer and air-and exhaust valves, air- and exhaust receivers, exhaust pipes, turbocharger, air-cooler and speed control system are implemented to mention a few. The library offers through a flexible modeling environment models of different formats for use in different simulation software environments.
Key Words: Modelling, Bond Graph, Model Libraries, Thermodynamics, Diesel Engines
1. INTRODUCTION
Models for simulation of the system dynamic behavior of diesel engine systems has been around for some time and has been developed to a fairly sophisticated level. Both the industry and academia have put much effort into developing and optimizing models for their specific engines, but reuse of the models in other contexts or in other applications are often found to be extremely difficult.
With the increase in computing power and advent of general modeling and simulation software of high quality and flexibility, it has become more and more interesting to perform complex system analysis for an increasing number of applications. The applications may range from: evaluation of different systems, components and solutions in the design phase; design and testing of new control strategies; analysis of modifications on existing systems, to increase the human understanding of the performance and response of a complex system; training and testing of operators and for use in maintenance and diagnostic applications.
In an ideal world mathematical models of typical components or physical phenomena necessary to model a complete diesel engine system would have been available in different complexity levels for easy assembly into a complete model for solution of your specific problem. This is unfortunately not the case. Models are often found available in different formats such as computer source code of different languages, block diagrams of mathematical equations or algorithms to name a few. Likewise they are often made dependable of specific solution procedures or methods. In addition yesterdays computer programs are not compatible with tomorrows, making the long term storage of models a premium concern for many companies.
Mainly two directions are selected to make available for the user community tools for analysis of dynamic systems such as the diesel engine system. The first option is the specialized programs made for analysis of specific problems or systems. These are often found to be preferable from a users point of view for solving standard problems. But for more complex problems involving models not available in the standard library, fewer solutions are available. The other option is to use a general purpose digital analysis and simulation program such as ACSL[1] or MatLab[2], and buy or develop necessary component models yourself. This option is more flexible, but offers little or no support for the modeling process itself. Hence, the majority of existing general simulation programs may be considered as "computer aided analysis" and not "computer aided modeling" programs.
The computer aided means of constructing a suitable model are given less attention.
* Norwegian University of Science and Technology
Department of Marine Engineering
Otto Nielsens vei 10. 7491 TRONDHEIM, Norway
FAX: +47 7359 5983 E-mail: Eilif,Pedersen@imm.ntnu.no
**Norwegian University of Science and Technology
