Control and diagnosis of servo motion systems

Servo motion systems play a huge role in industrial applications, such as robotics, automotive, aerospace, and production lines. Controlling these complex systems can be arduous when they exhibit resonance or time delay in their model. Furthermore, the maintenance of these expensive systems is crucial to ensure their proper functioning and avoid costly downtimes. Therefore, developing fault diagnosis methods for servo motion systems is of paramount importance in industrial settings. This thesis presents two main theoretical contributions to address the challenges associated with resonance and time delay in servo motion systems. Control literature presents several strategies to manage these problems. A notch filter can be employed to suppress unwanted resonance frequencies. However, it can degrade the performance of the closed-loop system if it is not correctly designed. To control time-delayed systems, a well-known strategy is the Smith Predictor approach. Yet, it exhibits poor performance when applied to servomechanisms, which are often integrative systems. This thesis proposes an automatic design method to tune a notch filter inserted in the open-loop transfer function of a feedback control system, in order to suppress resonance frequencies while ensuring closed-loop stability. A second theoretical contribution is related to the analysis of the inability of Smith Predictor strategy to control time-delayed systems with an integral action in terms of tracking and disturbance rejection performance. Subsequently, a Modified Smith Predictor approach is proposed to control time-delayed systems with an integral action, with the aim of ensuring the closed-loop stability while optimizing the closed-loop tracking and disturbance rejection performance. Furthermore, a robust extension of the Modified Smith Predictor approach is developed to consider uncertainty in the model of the system. Both strategies have been verified on a real experimental servomechanism during a period abroad at the Eindhoven University of Technology (TU/e). A significant application contribution of the thesis concerns the exploration of fault diagnosis methods for servo motion systems. In particular, the focus is on a real servo motion system called Liftronic, manufactured by Scaglia Indeva S.p.A. company. The Liftronic system is a self-balancing manual manipulator that assists operators in handling heavy loads with ease and precision. This thesis details the various models identified for the Liftronic, which were derived using several methodologies and under different operating conditions. Furthermore, specialized fault diagnosis methodologies were developed focusing on the Liftronic system's rope, which is the most critical component of the servomechanism.