Bond graph modeling - Energy basics

Energy processes have a highly non-linear behavior, mainly due to the mutual interaction of several phenomena of different natures, combining technological components that use different types of energy (mechanical, chemical, thermodynamic, etc.). The dynamic behavior of this type of system is generally described by non-linear differential equations. The classical approach based on the equations of the first principle (consisting in studying exchanges in stationary or transient conditions) becomes complicated due to the multi-energy and non-stationary nature of this type of process.

The multi-disciplinary bond graph tool uses a unified language to explicitly display the nature of power exchanges in the system, such as energy storage, transformation and dissipation phenomena, and to highlight the physical nature and location of state variables. Furthermore, the bond graph model is scalable, making it easy to refine the model (depending on the modeling assumptions adopted) simply by adding new elements (heat loss, inertia effect, etc.), without having to start from scratch. The method is suitable for obtaining integrated models, which is why the benefits of graph bond modeling of phenomena encountered in energy and process engineering are obvious.

The bond graph tool was first defined in 1959 at MIT (Boston, USA) by Paynter , and only introduced in Europe in the 1970s. This modeling approach is highly instructive for understanding and analyzing system dynamics. The methodology was further developed from 1996 onwards, mainly by Karnopp, Rosenberg and Thoma .

Bond graph modeling of structurally rigid systems (mechanical, electrical) has undergone enormous development; however, modeling of thermal, thermodynamic or simply energy and process engineering systems is still an open field, due to the complexity of these phenomena. Yet it is these types of processes, which are present in high-risk industries, that require increasingly accurate and usable models for their understanding and control.

This language is particularly well-suited to these needs thanks to the following features:

...