This paper reviews bearingless motor technology for medium and high power applications. Historically, bearingless motors have resided in research settings and certain niche commercial applications. Significant literature has been dedicated to developing designs that feature highly-specific geometry, restricted by applications with a unique set of needs, oftentimes with low power requirements. The first part of this paper reviews existing and potentially new use-cases of magnetic levitation in transportation and industrial applications. The second part of the paper reviews bearingless motor technology for these applications, presents a comprehensive literature review of prototype test results, and highlights efforts to generalize and increase the power rating of bearingless motor technology. The final part of this paper derives sizing equations of different bearingless motor topologies, which are summarized as a design tool for developing industrial-scale bearingless motor technology.

This paper uses active flux concept to review fundamental frequency sensorless algorithms for both induction and permanent magnet motors in one framework. Fundamentally, sensorless torque estimation can be directly solved using voltage model (VM) estimator, or indirectly solved using current model (CM) estimator. The latter turns the torque estimation problem into a speed estimation problem. The stator flux in VM and the d-axis angle in CM are deemed as the two sets of original states for sensorless drive. Through change of states, the direct torque estimation can be realized via observer designs; whereas the speed dependency of the unknown state (e.g., active flux and emf) gives rise to a class of speed estimation methods, known as model reference adaptive system (MRAS). The idea of a general speed observer is proposed to summarize various separate speed estimation methods needed for direct torque estimation. It is suggested to adopt inherently sensorless designs such that two-way coupling between torque estimation and speed estimation is avoided. For induction motors, it turns out the unmodelled voltage in the active flux dynamics reveals current flowing in rotor bars and can be further modelled, for which the solutions to regeneration instability problem are discussed, and change of states is recommended to attain global stability. Finally discussed are the results of slow reversal test, where local weak observability of ac motors can be potentially preserved.

As the higher order or generalized extended state observer (ESO), generalized proportional-integral (GPI) observer (GPIO) haves been proposed to enhance the active disturbance rejection (ADR) control (ADRC) system's disturbance rejection ability. However, different ADR controllers can be deduced based on the same order GPIO, resulting in different dynamic performances. In this article, ten different ADR controllers based on the fourth order GPIO are present. In order to reveal the relationship between these ADR controllers and the conventional ADR controllers based on the third order ESO, six different third-order-ESO-based ADR controllers are present and compared with the fourth-order-ESO-based ADR controllers. To ease the comparison between different ADRC systems, a common expression is built for different ADR controllers. A novel frequency-domain analysis method is also introduced to reveal how the observer and feedback control law affects the closed-loop control system's dynamic performance. The effectiveness of the proposed method is verified on the test bench based on dSPACE DS1103.

This paper addresses the problem of stator and rotor resistances identification in speed sensorless induction motor (IM) drives. The IM model that is transformable into the adaptive observer form is sought, for which the speed and rotor resistance adaptive observer can be established with rigorous stability ar- guments under the assumption of constant unknown parameters. However, further inclusion of stator resistance estimation leads to the over-parametrization problem. As a result, the adaptive observer based on the first-order approximation of the error dynamics is proposed. This is feasible because the regressors of the adopted IM model consist of known signals only. Experiment is carried out to verify the effectiveness of a sensorless drive with the proposed adaptive observer. Compared with the existing methods, speed and resistances estimation during a regeneration mode as well as successful slow speed reversal operation is found possible in the experiments.

The electromotive force (EMF) observer based sensorless algorithms are sensitive to temperature rise because it requires the stator resistance (Rs) or the rotor flux-linkage ( Ψr). The temperature measuring of an operating motor can be difficult, therefore adaptive sensorless algorithms with online Rs and Ψr estimation are appreciated. However, existing methods require extra signal injections, transient operations or estimate Rs and Ψr one at a time to overcome the rank-deficient problem. In this paper, an adaptive EMF observer which is able to estimate both Rs and Ψr simultaneously without requiring any signal injections or transient operations is proposed. The method exploits the harmonic back-EMFs (HBEMF) to serve as a novel reference model and regulates the Rs value in the EMF observer accordingly. The rank deficient problem is solved with the extra information provided by the HBEMFs. To make use of the HBEMFs, a parameter insensitive extraction method and an algorithm to calculate the rotor position from them are proposed. With the proposed reference model, an adaptive scheme to regulate Rs is proposed. To obtain the rotor flux-linkage, a Ψr estimator is also developed. The effectiveness of these proposed methods are all verified with experiments and simulations.

Rotor time constant identification is necessary for high-performance indirect field oriented control of induction motor drives. Meanwhile, speed sensorless technique has many advantages such as low cost and reduced hardware complexity. In recent years, the estimation methods for motor speed have respectively grown to maturity, but the research in the identification of rotor time constant remains to be extended. In this paper, an online scheme for parallel motor speed and rotor time constant identification is proposed according to parameter features, with the operating frequencies of all modules carefully considered. First, the speed is estimated via rotor slot harmonic extraction technique. On the premise of estimated speed, a derivative form of induction motor model is proposed for rotor time constant identification, which overcomes the problems of pure integration in rotor flux calculation. Correspondingly, instead of rotor flux, derivative rotor flux is used in the proposed model. Finally, an adjusted particle swarm optimization method is utilized on the proposed model to track the rotor time constant. The robustness and effectiveness of the proposed scheme have been validated experimentally.