RESEARCH ARTICLE Experimental validation and comparison between different active control methods applied to a journal bearing supported rotor Matheus Freire Wu | Katia Lucchesi Cavalca Laboratory of Rotating Machinery, School of Mechanical Engineering, UNICAMP, Campinas, Brazil Correspondence Matheus Freire Wu, Laboratory of Rotating Machinery, School of Mechanical Engineering, UNICAMP, 200, Rua Mendeleyev, Campinas, CEP: 13083‐860, SP, Brazil. Email: matheusfw@fem.unicamp.br Funding information Fundação de Amparo à Pesquisa do Estado de São Paulo, Grant/Award Num- bers: 2015/20363‐6, 2016/13059‐1 and 2017/15494‐0; São Paulo Research Foun- dation (FAPESP), Grant/Award Numbers: 2017/15494‐0, 2016/13059‐1 and 2015/ 20363‐6 Summary Rotors supported by cylindrical hydrodynamic journal bearings usually face rotational speed limitations due to the unstable motion caused by oil whip effect. An active magnetic bearing is utilized as auxiliary component to sup- press oil whip and reduce unbalance lateral vibration amplitude at critical speed. Three distinct control methods are considered for the actuation force (dynamic H ∞ , μ synthesis, and polynomially parameter‐dependent static H ∞ ), which must be designed to perform in a situation with varying parame- ters and uncertainties. These techniques are tested in scenarios with collocated and noncollocated feedback (emulating a magnetic actuator with magnetic flux sensors). The different resultant performances are successfully compared with each other numerically and experimentally. KEYWORDS active control, oil whip, H ∞ , parameter dependence, rotating machinery, μ synthesis 1 | INTRODUCTION Cylindrical hydrodynamic journal bearings are widely employed in contexts of industries and power plants because of their interesting set of advantages, such as good load/size ratio, high damping capability, relatively easy and low‐cost manufacture, and long life span. On the downside, continuous oil input is needed, and the system is susceptible to an autoexcitation provoked by the pressure distribution in the supporting oil film. This autoexcitation, known as oil whirl, appears as a response at approximately half the rotor operating frequency. At about twice the first critical speed, the oil whirl coincides with the resonance and the autoexcitation usually turns into an unstable motion, 1 known as oil whip or fluid‐induced instability, which imposes an operational threshold. The oil whip effect is better described by nonlinear dynamics. However, for an easier control approach, it can be approximated by speed dependent linear coefficients 2 with sufficient accuracy. In this case, the oil whirl subsynchronous (half harmonic) component does not appear in simulation. Still, the energy transfer from rotational speed to lateral vibration is represented by decreasing equivalent bearing damping coefficients dependent on the rota- tional speed. Damping loss eventually leads to instability, which, in linear systems perspective, amplifies the rotor vibra- tion indefinitely. In practice and in nonlinear models, the oil whip will be limited by the journal bearing gap limit circle. Nonetheless, it is an undesirable condition that can break the oil film, causing direct contact between solid parts and resulting in severe damage to the rotor components. There are several passive methods to overcome main issues as oil whip and unbalance: project improvement, balancing, installation of additional dampers, bearing geometry modification, application of tilting pads, and so forth. Received: 28 March 2019 Revised: 2 July 2019 Accepted: 6 August 2019 DOI: 10.1002/stc.2446 Struct Control Health Monit. 2019;e2446. https://doi.org/10.1002/stc.2446 © 2019 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/stc 1 of 16