Abstract
A unique feature of bond-graph techniques is that they provide the modeler with a graphical representa tion of the causality (i.e., the input-output rela tionships) in the underlying system. An understand ing of causality makes the modeler aware of algebraic loops and implicit differential equations at an early stage in the solution process. It also clarifies the effects of adding or deleting features of the model. The modeler can use causal information to create simulation programs for nonlinear systems even when some variables cannot readily be expressed in equation form. The resulting program is equivalent to a set of state equations but contains subroutines instead of algebraic expressions. For many complex systems, using vector bond graphs to represent causal interactions among large subsystems can help the modeler write programs using sets of variables rather than individual variables.

This publication has 0 references indexed in Scilit: