Rockwell Automation (Canada)

Abstract The problem of reinforcement learning (RL) in an unknown nonlinear dynamical system is equivalent to the search for an optimal feedback law utilizing only data from the simulations/rollouts of the dynamical system. Most RL techniques search over a complex global nonlinear feedback parametrization making them suffer from high training times as well as variance. Instead, we advocate searching over a local feedback representation consisting of an open-loop sequence, and an associated optimal linear feedback law completely determined by the open-loop. We show that this alternate approach results in highly efficient training, the answers obtained are globally optimum, repeatable with negligible variance, and hence reliable, and the resulting closed performance is superior to global state-of-the-art RL techniques. Finally, if we replan, whenever required, which is feasible due to the fast and reliable local solution, it allows us to recover the optimal global feedback law.

Multimotor drives have become increasingly important in modern industrial applications due to their ability to provide superior performance, efficiency, and flexibility compared to single-motor systems. Hence, this article presents an overview of recent advancements in multimotor drives, focusing on three main areas: structural diversity, advanced control, and emerging challenges and solutions. First, the various structural configurations of multimotor drives are summarized, which include parallel, cascaded, and hybrid configurations. The features as well as component motors and converters of each configuration are discussed, along with the selection rules of a particular configuration for a given application. Second, from the perspective of different performance requirements, the advanced control technologies used for multimotor drives are discussed. Then, this article highlights the technical challenges associated with multimotor drives, including coordination control, mutual interference, communication, interdependent fault diagnosis, and power quality. Meanwhile, viable solutions to these challenges are summarized. Finally, a discussion of the future directions and opportunities for further research and development in the field of multimotor drives is presented. Through this article, scholars and engineers can gain a comprehensive understanding of current and future developments in multimotor drives, contributing to continued research in this field and facilitating successful integration into various applications.

A novel method is proposed in this paper for inter-turn short circuit detection in three-phase power transformers. This detection method only requires primary side currents and can determine the fault severity and locate the faulty phase(s). Phase shift information is extracted out of primary side currents and can act as an effective fault indicator. Machine learning techniques are adopted for feature generation and decision region separation. Continuous degradation monitoring is feasible with the proposed approach. To improve the robustness and accuracy of the detection algorithm, Gaussian mixture model is introduced from the perspective of probabilistic machine learning. Both simulation and experiment results are provided to validate the algorithm.

Dry ice blast cleaning originated in the aircraft industry when they were looking for alternative ways to strip paint off older aircraft, at that time. The technology did not become commercially available until around 1987. The dry ice cleaning process begins with the creation and use of pellet or granular shapes made from liquid Carbon Dioxide (CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>).This type of cleaning method is non-abrasive and does no damage to the base substrate materials. Therefore, it can even be used on sensitive mechanical and electrical equipment. The process is not electrically conductive and can be safely used on electric motors and other electrical equipment.Dry ice exists as a liquid only when under very high pressure. When the pressure drops to near normal atmospheric pressure, approximately half of the dry ice turns back into a gaseous form and half turns to a solid. These solids, usually in the form of fluffy snow-like material, are then compressed to form dry ice blocks, pellets, or nuggets for use in the cleaning process.

Cascaded H-bridge (CHB) motor drive is an attractive solution for high-power industrial motor drives due to its modularity and scalability. The existing phase-shifting transformer in the CHB drive can eliminate the low-order harmonics at the grid, which makes CHB drives outperform other MV topologies due to their “perfect” harmonic profile. However, in the applications where frequent braking or regenerative capability are required, the conventional CHB drives cannot be employed due to the diode-front-end (DFE) structure. To obtain the regeneration capability, several solutions based on IGBT active front-end rectifier (AFE) have been proposed to convert the regenerative power to grid in each CHB power cell. However, these IGBT-based methods in regenerative CHB drives introduce the extra switching sideband harmonics that cannot be effectively eliminated by the existing phase-shifting transformers. The advantage of the CHB drives is thus sacrificed. In this paper, a new solution is discussed by replacing the DFE rectifier with an SCR (Silicon-Controlled Rectifier)-Front-End rectifier/inverter in each power cell. Unlike the existing IGBT-based solutions, the harmonics generated by the SCR rectifier/inverter can be eliminated by the existing phase-shifting transformer. Thus, to comply with the grid harmonic standard IEEE std 519–2014, a small <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$L$</tex> <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> filter between the grid and SCR rectifier/inverter is required in each power cell. To further remove the additional filters while complying with IEEE standards, in this paper, a new 54 pulse phase-shifting transformer is introduced for a seven-level CHB drive and studied with an improved harmonic profile. Finally, the proposed SCR-Front-End solution is validated on a seven-level CHB inverter in simulation. This method can be generalized for any n-level CHB motor drive.

Medium voltage motor control centers, designed around the requirements of UL 347 [1], have been successfully utilized in the petrochemical industries for many decades. As they continue to provide consistent control of large medium voltage assets, they have also become key components within the intelligent control requirements of today’s complex petrochemical facilities. Through the incorporation of proven technologies employed in other safety-based products used globally, these control products can be moved beyond their traditional safety requirements specified by the current product design and testing standards. This paper will cover the implementation and utilization of these proven technologies that can provide additional levels of personnel safety within these traditional designs.

Switchgear is an integral part of power systems network. Any fault occurring in switchgear may leads to significantly long outage duration and the affected load, or area, can be enormously large. Partial discharge (PD) is one of such faults which are very common in metal-clad switchgear due to insulation defects in bushings, CTs, PTs, bus bar insulations and connections. Online and offline PD measurements have been recently employed to investigate the health condition of such insulations. It is more likely in such equipment that multiple PD faults exist concurrently and hence the identification and separation of multiple faults sometimes gets difficult. This paper presents separation of multiple PD faults in metal-clad switchgear based on the significant features extracted from PD pulse shape and PD current waveform using K-mean clustering technique. For this purpose, a conventional PD measurement technique given in IEC 60270 standard is adopted to record the PD signals under several insulation faults in the switchgear. Also, a PD current probe is used to detect their corresponding PD current signals. The characteristics of the detected PD signals were studied to estimate the severity of insulation faults. The significant features of time domain PD pulse shape have been compared for the separation of multiple insulation faults in switchgear. Therefore, eight-dimensional feature space has been developed. Finally, K-mean clustering algorithm has been implemented based on the extracted features of PD pulse shape and PD current waveform to define their clusters in the feature space. It has been revealed that the proposed framework is effective in separating the mixed data of PD faults in metal- clad switchgear.

Conversion Congress and Exposition (ECCE 2022) will be held in

The diode-front-end cascaded H-bridge (CHB) inverters have prevailed in the nonregenerative industry drive domain for high-power medium-voltage applications. The regenerative version of the CHB drives is made possible by adding the extra active-front-end rectifier in each power cell, such as a three-phase PWM rectifier. However, due to the instantaneous power unbalance, the dc-link capacitors of the regenerative power cell need to be overdesigned to maintain a stable low ripple dc-link voltage. To reduce the dc-link capacitance, this article proposes a novel closed-loop voltage ripple controller for the regenerative CHB drive without adding extra sensors. In the proposed method, dc-link voltage ripple amplitude and phase angle are accurately detected with a high-performance adaptive filter. Moreover, a latent instability issue is discussed and is avoided in the proposed controller. The performance of the proposed control strategy is validated experimentally on a seven-level regenerative CHB drive.

The subject of safety continues to be at the forefront of all forest products industry segments. However, a high frequency of serious electrical injuries and fatalities are still occurring around electrical equipment. Some of these events are attributed to the lack of proper regular routine maintenance practices. These events also result in high monetary losses associated with damaged property and production losses. This paper will outline some of the areas of concern and offer some perspective on how to plan better and schedule appropriate preventive maintenance.