There is a large variety of Rotating Machinery in modern industry and everyday life. For the designer of such machines besides the required machine function other aspects like availability, reliability, safety and machinery life time are of importance. Due to this fact a detailed Rotordynamic analysis has to be an essential part in the design process, to avoid irreversible damages, malfunctions and a rapid fatigue of components. Investigations of vibrations in Rotating Machinery are based on the disciplines of mechanics, solid and fluid dynamics, electromechanics and nowadays more and more by mechatronics and digital techniques. With the three tools of Theoretical Modelling, Numerical Analysis and Experimental Analysis (Testing) different tasks in Rotordynamics can be solved. New methods and technologies have been further improved in the past and help design engineers as well as test engineers to realize accepted vibrations of Rotaing Machinery. However, due to the existing world wide transformation processes new challenging tasks in different engineering areas have to be solved. With respect to Rotordynamics different positive effects can be achieved when vibrations are more reduced. Lower vibration levels would for example lead to: - a reduction of mechanical stresses and fatigue,- higher efficiencies due to tighter clearances in bearings and seals, - a reduction of disturbances due to acoustical emissions, - an improvement of work piece qualities in production processes. Some time ago a workgroup of researchers from the IFToMM Technical Committee identified topics of recent advances in Rotordynamics and pointed out the future challenges for science and technology. All of these topics can be grouped in one of the four future challenges: - development of active and passive hardware components, - Theoretical Modeling, including interactions with other disciplines, - improved procedures in Numerical Analysis, - new Experimental Analysis and Testing methodsIn this lecture some of the future challenges in Rotordynamics will be presented. Realized examples in this direction will be discussed development of passive gas foil bearings torsional vibrations due to electromechanical interactions vibration reduction by means of active balancing devices active damping with magnetic bearings and piezo actuators monitoring and failure diagnosis by means of digital twins.

Rotating systems represent a class of machinery with wide application in the power generation industries. The dynamic analysis of these machines is crucial, due to the need to predict operating conditions and ensure their best performance. In this context, predictive and preventive maintenance meets the needs imposed by users of rotating machines. In the context of power generation, the demand becomes even more evident and complex, due to the wide diversity of rotating machines present in this field. Previous projects have focused on the development of rotor analysis software, specifically including geared parallel axis transmission. In view of the above, it becomes feasible to analyse the dynamic conditions of more complex rotating machines, such as those that use hydraulic speed variators associated with planetary gears transmission, allowing the redefinition of safe operating limits, in order to mitigate certain operating conditions critical to integrity, and to investigate new safe set-points to protect the system from adverse operating conditions. In this way, through a more comprehensive analysis of these complex systems, a more robust identification process for operating conditions can be established, favouring actions for fault diagnosis and predictive maintenance planning.

The history of modeling of active magnetic bearings (AMBs) is reviewed from the 1970's up to present. Early models were static and assumed that AMBs could emulate the behavior of stiffness and damping models used for mechanical bearings at the time. These models made no effort to consider the dynamics of the power amplifiers, the many subtleties of the physical implementation of controllers, and the problem of sensor-actuator non-collocation. Over the course of intervening years to the present, industrial and academic application of AMBs has revealed the many shortcomings of these early models and has driven the development of much more detailed models, both linear and nonlinear, capable of very high fidelity predictions of stability and performance. Recognizing that even the most sophisticated AMB models retain sensitivity to parametric uncertainties such as thermal growth of air gaps, the presentation concludes with a few open problems in AMB modeling.

Due to tremendous increase in the computation power, advances in the sensor technology and development in machine learning algorithms, it is now possible to collect enormous valuable information from high-speed mechanical and electrical rotating machinery, process them very fast to extract useful features, and train machine learning algorithms to be useful for the fault detection and its diagnostics. The present keynote presentation briefs research work done at IIT Guwahati in this field for the past more than two decades on variety of rotating machinery, e.g. rolling bearings, gearbox, geared-rotor, centrifugal pumps, induction motor, etc. It covers brief induction to machine learning algorithm, for example, support vector machines, data collections (vibration, pressure, current, etc.), feature extraction from data, feature selection, training to machine leaning algorithms, and finally testing. Latest trends on integrating physical based model with ML based model (hybrid model) will also be briefed. Potential application of the methodology for the practical application will also be addressed.