The aim of this paper is to present and to compare some linear and non-linear control laws for an electropneumatic positioning system both for point-to-point control and for tracking control. The experimental results are presented in terms of repeatability for each control law implemented on the same device: an in-line electropneumatic servo-drive. Different kinds of model, namely a non-linear affine model, linearized tangent model and a reduced linearized tangent model, are presented to synthesize the different control laws. For this a new mathematical modelling of the servo-distributor flow stage is described.

As an important driving element, the pneumatic cylinder is widely used in industrial applications because of its reliable, cheap and excellent performance in an industrial servo system. However, along with the development of control technology, the requirement for control precision gets higher and higher. In many cases, in order to achieve the satisfied control performance, the effects of nonlinear factors contained is considered in the plant. Backstepping is one of these nonlinear control techniques which has attracted a great deal of research interest in recent years. It is mainly applicable to strict-feedback system. This paper presents a new form of backstepping controller for an electropneumatic system. Notice that the model of this system is not under strict-feedback form. Then, a modified methodology is proposed, synthesized and implemented on an experimental test bench. Experimental results are presented and discussed.

Mass flow rates in the half bridges " " of the servovalve, with or. Interchamber leakage mass flow rate. Variable restriction between and .

This paper focuses on a comparison between two control laws of an half meter stroke electropneumatic dissymetrical cylinder controlled by two three ways servo-distributors: a classical pole placement method and a with scheduling gains control law. Using physical laws, a nonlinear model of this process is described in the first part of this paper and the difficulty for the obtention of the servodistributors mass flow rates is carried out. Nevertheless, a particular choice of the control input leads to a single input one and around an equilibrium point, a tangent linearized model is obtained. Two control laws are described and implemented using a well kwown dSPACE interface card. Experimental results obtained for point to point control are presented and discussed.

System inversion is a concept that often appears either implicitly or explicitly in the design of control systems (e.g. decoupling, feedback linearization or feedforward control ). This paper presents an algorithm for the representation of an inverse systems model using bond graph techniques. Structural invertibility conditions are investigated graphically and the structure of the inverse model is discussed. This graphical inversion method is carried out in parallel with classical inversion algorithms using causality manipulations in bond graphs. The physical approach adopted here is enhanced with applications presented in Part 2 which follows.