Retaining analog intuition in a digital world with bond graphs Original Research Article

Although digital computers have supplanted analog computers for the simulation of physical systems, and digital signal processors have largely replaced analog devices in control systems, most engineers still think about mechatronic systems in analog terms. Skilled circuit designers and many others can use their intuition to solve dynamic problems by modifying components or inserting new elements into a system. This intuition is based not only on experience with physical systems, but also on simulation of continuous signal mathematical models. Digital computers allow the construction of extremely complex mathematical models by the combination of a variety of submodels, but it often happens that an intuitive understanding of the assumptions inherent throughout the system is nearly impossible, so that if the simulation is difficult or the predicted performance is poor, the remedy is far from obvious. Bond graph-based models allow a system analyst to think in physical terms about system components and to appreciate the mathematical and computational aspects of changes in subsystem elements or parameters. Bond graph processors and programs containing libraries of submodels make the process of translation from a physically-based model, to a mathematical model, and then to a computational simulation scheme nearly automatic. In this way, the system designer’s intuition can be brought to bear in solving dynamic problems in mechatronic systems.