Over the past few years, Artificial Intelligence (AI) techniques have become one of the most exciting technologies of our age. This fascinating field paves the way for new possibilities and extends to almost all areas of industry and research. This paper illustrates a Short-Circuit (SC) diagnosis strategy for a multi-three-phase brushless AC motor drive based on Machine Learning (ML). A thorough model of the electrical machine is obtained through specific finite-element simulations and used to replicate various fault scenarios. This model quickly yields a large amount of data, which is then employed for training the ML algorithms. Once trained, the ML models are tested, with experimental data directly obtained from a prototype of multiphase drive, to verify the effectiveness of the diagnostic algorithms. The ML algorithms allow monitoring the health condition of the machine, distinguishing and localizing two types of faults, i.e., inter-turn SCs and Extra Turns (ETs), placed in different coils of the machine.

Recently, multiphase drives have emerged as a notable alternative for three-phase drives in several domains, including electric marine propulsion, locomotive traction, and high-power industrial applications where fault tolerance and high reliability are required. This study uses a graphical approach to develop a new carrier-based Pulse-Width Modulation (PWM) that reduces the root-mean-square value of the output current ripple in two-level multiphase inverters with an odd number of phases. The optimization process calculates the optimal common-mode voltage for each switching period. Numerical models and experiments with five- and seven-phase inverters demonstrate the effectiveness of the developed PWM technique, which can reduce the number of commutations while obtaining the lowest output current ripple.

This study presents a comprehensive analytical model for the real-time detection of intern-turn short circuits (ITSCs) in asymmetrical dual-three-phase permament magnet synchronous motors (DTPPMSMs). The electrical signals are decoupled into independent complex planes via vector space decomposition (VSD) and controlled using a specifically designed control system. The proposed method enables the real-time identification and diagnosis of an ITSC affecting the stator winding without further measurements, hence decreasing maintenance and monitoring costs. Moreover, it is demonstrated that the ITSC slightly alters the magnetic axis of the faulty phase, and according to this effect, the relative position of the coil within the faulty pole pair can be determined. Experiments with a prototype of a DTPPMSM are carried out to assess the effectiveness of the proposed method.

The Modular Multilevel Converter (MMC) has gained significant popularity over the past decade due to its versatility. The MMC features have been leveraged in numerous fields, including high-voltage DC transmission, electric vehicle power trains, motor drives, and wind energy conversion. In controlling the MMC, the circulating current (i.e., the current flowing through both the upper and lower converter arms without delivering power to the load) has consistently been the most critical variable. In early applications, it was perceived as a source of losses, but more recently, it has become evident that injecting a specific current could reduce voltage and energy ripples. This paper presents a theoretical framework, based on time-scale analysis, useful for modeling and controlling MMCs. The new approach is adopted for generating the circulating current reference, which is expressed as a linear combination of orthogonal functions. The goals are to decouple the control of the voltages of the upper and lower converter arms and manage additional harmonic components of the circulating current for voltage ripple reduction on module capacitors. The simulations and experimental results demonstrate the effectiveness of the proposed control strategy.

Dual Three-Phase Permanent-Magnet Synchronous Machines (DTPPMSMs) are an appealing alternative for a variety of applications that require high power levels and reliability. Regardless of the machine architecture, the winding insulation of an electrical machine is frequently subjected to thermal and mechanical stress as a result of high-performance requirements. This may cause Inter-Turn Short Circuit (ITSC) failures, compromising the machine's health and reducing the overall reliability of the drive system. Diagnosing the presence of an ITSC may be challenging, especially when only a few turns are shorted and the machine operates at low speed. In this circumstance, the effects of a short circuit are externally undetectable until the speed reaches a critical value. This research introduces a method utilizing a Beta-Variational Autoencoder ($\beta$-VAE) neural network to assess the presence of an ITSC by using appropriate features for training the Machine Learning (ML) model. Moreover, a statistical approach is introduced to diagnose the fault for different confidence intervals. The proposed method is assessed under various operating conditions using a prototype of DTPPMSMs.