Modelisation causale et acausale d'un systeme electro-hydraulique

• Main Author: Gagnon, Tommy
• Format: Dissertation
• Language: French
• Subjects: engineering, mechanical; engineering, system science
• ISBN: 9780494763711; 049476371X

In a context were time allotted to design need to be as short as possible to stay competitive, the systems to design are still gaining in complexity. In that case, advanced design tools become essential. These tools help the designer prototyping virtual models to find optimal solution faster than using conventional approaches. Computer models of a system can be done in different ways. First, a mathematical model can be developed from physics laws that describe system evolution. This model is then linearized to form a transfert function, or directly used in bloc diagrams. Recently, a new multi-physics simulation approach, called acausal modeling, increases in popularity. This approach uses a mathematical model structured like a physical model.
An electro-hydraulic actuated pendulum has been modeled using three different approaches to evaluate if multi-physics simulation is an improvement compared to classical simulation, called causal, and if its response represents the one from the real system. The first model uses linear system theories. The second approach is a non-linear model using block diagrams in Simulink. Finally, a physical model is built using SimScape. Those three models are used to design a PID controller to adjust the pendulum position. Validation and code qualification are then compared using the real system response as reference.
After adding some delays to the three different models, their responses match the one from the real system. The error between the multi-physics and the real system is less than 15% for the raise time and 6.7% for the response time. In conclusion, all the steps required to build models show that it is more advantageous to use an acausal modeling approach due to its reusability, simplicity of usage and readability.
Key words : Acausal modeling, causal modeling, PID, electro-hydraulic system, servovalve, hydraulic flow, SimScape, Simulink.