This paper presents an algorithm for the re-adhesion control of electric rolling stocks. Generally, an engineer manually adjusts parameters used in any re-adhesion control algorithm before a commercial operation. However, it depends heavily on the experience of the engineer. Therefore, developing an automatic adjustment system for re-adhesion control is imperative. To this end, we apply the particle swarm optimization (PSO) algorithm to re-adhesion control. In our approach, a cost function based on the averaged tangential force of an electric bogie is optimized by the PSO algorithm, and the parameters for re-adhesion control are automatically adjusted. In addition, experiments using a scale model railway vehicle are carried out to validate the proposed method.

With the rapid spread of the cloud server market, hard disk drives (HDDs) are required to achieve higher truck following speeds and storage capacities. To meet these demands, it is necessary to improve the accuracy of positioning control of the magnetic head that writes data to the disk. In this paper, we propose a controller designed by recurrent neural network (RNN-) based reinforcement learning for the “benchmark problem of designing a control system for a hard disk drive with two-stage actuators” and evaluate the performance of the RNN-based controller. We also propose a method to convert an RNN-based controller to a state-space linear controller, which guarantees stability and enables performance analysis. This contributes to eliminating the black box drawback of reinforcement learning. Performance verification on the benchmark problem shows that the 3σ value of the head position of the track pitch of the RNN-based controller shows an accuracy improvement of approximately 1.8-fold compared to the reference (example) controller. The sensitivity function of the RNN-based controller shows superior disturbance attenuation especially for the disturbance due to external vibration caused by air-cooling fans and other HDDs in a storage box. The article describes the structure of the proposed RNN-based controller and presents an evaluation of the benchmark results⁽¹⁾.

This paper presents a novel method for compensating PWM ripple currents without delay, enhancing the responsiveness of oversampling current control. Deadbeat control is known for its high responsiveness in PMSM control but lacks robustness against modeling errors, such as variations in resistance and inductance. Previous research has demonstrated that employing current oversampling can mitigate this issue. However, this method introduces voltage distortion due to discrepancies between real and reference voltages. While prior studies have proposed direct compensation methods, these often rely on approximations for current. This paper addresses the voltage distortion caused by PWM ripple currents. The proposed method eliminates PWM ripple from the detected current and the compensated signal is fed back to the controller to reduce voltage distortion. The effectiveness of this approach is validated through both simulation and experimental verification. Consequently, the proposed method enhances the responsiveness of current control without relying on any approximations.

The force-current factor is a control parameters in bearingless motors. It is defined as the ratio of the magnetic suspension force and suspension current. We have proposed a method to estimate the force-current factor of the radial suspension force from the back-EMF of the suspension winding. In this study, the force-current factor of a prototype machine was estimated experimentally and compared to that calculated using 3D-FEM analysis.

This paper discusses asymmetric slit structures in a rotor that can reduce torque ripple as well as increase torque and output for concentrated winding interior permanent magnet synchronous motors. A motor with rotor asymmetric slits in forward rotation can reduce the torque ripple as opposed to a symmetric motor without asymmetric slits because of the effect of the asymmetric slits in increasing the instantaneous minimum torque. Furthermore, the more the outer diameter tip of the asymmetric slits is moved in the direction opposite to rotation, the smaller the torque ripple can be. In the reverse direction, torque ripple increases from forward rotation by the effect of asymmetric slits. However, by adding notches on the rotor core surface, the instantaneous maximum torque can be lowered and the torque ripple can be reduced.

This paper proposes a hierarchical optimization framework for automatic generation of maintenance worker schedules at rolling stock depots. Maintenance worker schedules at rolling stock depots are daily schedules assigned to maintenance worker teams such as cleaning rolling stock and performing inspection during turnaround operations of superior trains. To evaluate the effectiveness of the proposed method, the schedules generated by the proposed method are compared with actual schedules, those generated by a mathematical programming method, and local search. The results indicate that the proposed method can generate maintenance worker schedules satisfying non-relaxable constraints and high priority relaxable constraints within a practical computation time of three minutes with levelized and reduced worker loads.

This study examines the gate magnetic coupling technique, a gate drive technology for series-connected power semiconductor devices, and discusses the generation principle of gate voltage oscillation specific to the technique and its countermeasures. Additionally, this study proposes two gate-drive circuits designed to suppress gate voltage oscillation. The experimental results obtained using 3.3kV/750A SiC-MOSFET/SiC-SBD power modules demonstrate that the proposed methods can suppress gate voltage oscillation and reduce the voltage imbalance between the series-connected elements.

The large-scale aggregation of fast demand response (DR) from hundreds or thousands of building multi-type air-conditioners, i.e., variable refrigerant flow (VRF) air-conditioners, is anticipated to serve as a flexibility resource for balancing supply and demand in future smart grids. This paper proposes a dynamic equilibrium DR aggregation method, inspired by birth-death regulating balance of billions of cells in biological systems. We developed a virtual model of an office building block comprising 50 buildings and 660 multi-type air-conditioners. Using the model, we conducted digital simulations and found that the proposed dynamic equilibrium method exhibits better power reduction precision than the conventional set-temperature shift and group-rotation methods.

This letter proposes the concept of short-time operation when a ripple current exceeds the allowable value. The operable time is determined using the temperature of the electrolytic capacitor. The temperature characteristics of electrolytic capacitors are measured by applying a pulsating voltage. The calculated value and value measured by experiment are compared for transient temperature changes and operable time. The possibility of estimating the operable time is confirmed through a comparison of transient temperature changes.