Ingeteam (Spain)

Modern railway traction drives require precise traction torque control for regulatory compliance and reliable operation. Inaccurate core loss modeling, flux observer errors, inverter non-linearities and measurement uncertainties are common sources of torque estimation errors. Torque estimation error can be reduced by improving motor models and parameter estimation, inverter modeling, variable measurements, etc. However, this is not trivial, and some level of uncertainty will always remain. In this paper, an Artificial Neural Networks (ANN)-based technique for torque estimation error compensation in railway traction drives is proposed. Data used for ANN training are collected during the commissioning process of the traction drive. The proposed method will be implemented in the 400 kW traction drive of a high-speed train. Torque estimation at high speeds was found to be the most critical. Using the proposed method, the estimation error was reduced from 12% to 2% in this case.

Accurate parameter identification is crucial for the digital twin of power electronic systems. Balancing global exploration and local refinement for high-dimensional, nonlinear systems is not trivial. This paper compares four advanced optimization strategies for parameter identification customized for the digital twin for a bidirectional buck converter: 1) standard PSO, 2) standard IPM, 3) Multi-Start IPM (MS-IPM) and 4) Hybrid PSO-IPM method combining swarm intelligence with gradient-based refinement. It is shown that the sensitivity of gradient-based techniques to the initial guess of temperature-dependent parameters can become critical, making these methods inadvisable unless temperature measurements are available. Hybridization of metaheuristic and gradient-based techniques (Hybrid PSO-IPM) emerges as the optimal choice when thermal measurements are inaccessible. Simulation and experimental results are provided to validate these findings. The research presented in this paper is aimed at onboard railway systems; however, the conclusions have general applicability.

High-power-density converters concentrate heat in smaller areas of printed circuit boards (PCBs), a challenge exacerbated by Gallium Nitride (GaN)-based components in surface-mount device (SMD) packages. While these components reduce parasitic inductance and electromagnetic interference (EMI), they also transfer heat directly to the PCB, increasing local temperatures. Traditionally, PCB thermal design relies on computational fluid dynamics (CFD), which is computationally expensive and requires multiple iterations. This paper presents an analytical thermal model to evaluate and optimise PCB thermal resistance with significantly reduced simulation time. The model examines the influence of design parameters on thermal resistance for through-hole and blind vias, identifying optimal via diameters and spacing to minimise thermal resistance while limiting impacts on copper pouring areas (and parasitic inductance). By addressing both the area beneath the heat source and its surroundings, the model provides a comprehensive understanding of PCB thermal behaviour. Validation against CFD simulations shows a deviation of only 7%, with substantially lower computational time.

The interleaved operation of parallel power converters by applying carrier-based pulse-width modulation (CB-PWM) techniques is a popular method to achieve very high quality output waveforms. However, considering high-power converters that require low switching frequency operation in order keep the power loss below acceptable limits, some of the harmonics may exceed the individual limits defined by grid connection standards. To face this problem, this work proposes to modify the conventional CB-PWM pulse patterns using basic preprogrammed method principles. The proposed method deals not only shaping the required harmonic profile at the grid point-of-common-coupling (PCC), but also reduces the currents total harmonic distortion factors at the load and inverter sides. To consider all these objectives, one high-order polynomial cost function has been defined. The proposed preprogrammed PWM method has been successfully validated in a dual converter system based on neutral point clamped (NPC) converters using separate dc sources.

In rural distribution networks with widely dispersed loads, characterized by the utilization of long overhead lines, voltage profiles with big variations appear along the line due to the significant variations of the load between peak hours and valley hours. This paper describes the problems that the integration of distributed generation based on renewable resources could generate in that scenario, and presents solutions that are commonly used to mitigate these problems and proposes new solutions to solve them. The integration of distributed generation (DG) based on renewable resources in these networks increases the variability of the voltage, and in some cases the power flow of the network also can be bidirectional. Usually this integration of distributed generation is limited by the network operator limiting the maximum voltage variation produced by the distributed generation in the point of common coupling (PCC). In this paper different traditional methods for voltage control are described and analyzed for a widely dispersed rural network scenario. The impact of the DG in these methods of voltage control and the advantages of power electronics in this new scenario are exposed. Power Electronic devices applied to these distribution networks allow more efficient management of the network increasing the integration of distributed generation maintaining the security levels.

On 12 December 2007 a distance relay tripped erroneously by zone 1 for a ground fault in an adjacent line. RTDS simulations and theoretical calculations proved that the relay had operated according to its design parameters and that the settings of the units involved were adequate. Further analysis of the fault recordings showed that the currents followed a single-phase to earth fault pattern, whereas the voltages did not follow any known pattern for the conventional types of faults. In the first part of the paper a theoretical analysis of the problems involved with grounding the VTs in multiple locations is presented. First, the nature of the impedance between different grounding points is studied. Next, the voltage gradient during an earth fault is shown and the sequence voltages affected by a fault involving earth are determined. The deviation from the true value of the faulted phase voltage is also theoretically expounded. The relaying units affected by the multiple groundings are listed and their possible misoperations presented. To end the procedure to determine that such a problem exists and the ways to confirm it are presented along with the calculations that have to be done to determine the performance of the relays involved. In the second part of the paper the analysis of a real case is presented. The theoretical background expounded in the first part of this work is applied to the above mentioned fault and it is shown that, had only been one VT grounding, the relay would have operated correctly.

The railway industry is continuously evolving, necessitating the development of more intricate and dependable software algorithms. To meet these requirements, the creation of new validation tools has become increasingly crucial, with a specific focus on Hardware-In-the-Loop (HIL) simulations. HIL simulations are invaluable during the product development process as they improve system resilience, performance, and overall product value. This study details the design of a HIL platform that can validate the software created for the Traction Control Unit (TCU) of a complex traction system. By producing real interfaces in real-time, this platform offers a practical simulation environment for examining the TCU software. The HIL platform built in this research provides comprehensive test coverage of the TCU software, a feat that is often unattainable through on-train testing. Moreover, the platform provides greater tuning and debugging capabilities, allowing for a virtual certification of the TCU software, which can offer greater security and assurances of success when encountering official tests. Thus, this paper demonstrates how profitable a HIL platform that models the whole traction system can be in order to debug and certificate its TCU software.

Soft Open Point and Shore-to-Ship Power applications permit to deploy smart grid infrastructure into distribution and port networks. Both are based on power electronics converters that require an optimal integration to maximize capacitive reactive power limits at minimum impact over costs and efficiency. To meet this objective, the paper defines a methodology to calculate the optimal connection voltage of power converters used in these applications. The proposed methodology is based on three novelties related to capacitive reactive power limits definition. Firstly, all affecting variables from positive and negative sequences are jointly considered completing partial approaches identified in previous works. Secondly, the impact of negative sequence affecting multiple converter terminals is considered, adding accuracy to reactive power limit calculation of Soft Open Point converters. Thirdly, the influence of two main constraints affecting modulation limits in real converters is described, quantifying its impact over maximum capacitive reactive power limits. These constraints are related to semiconductor characteristics and to digital implementation; usually not considered for dimensioning purposes at application level, but which can have a key impact on capacitive operation. At this point, the authors provide explicit details of industrially available converter designs to support calculations and evaluate its impact over the final solution. Considering the above, the proposed methodology is implemented on a real 20MW 6.6kV medium voltage converter design and results are quantified and compared from a reactive power capacity, efficiency and cost perspective. INDEX TERMS Modulation limits, reactive power capability, shore-to-ship power, soft open point.

This paper presents preliminary experimental results of a high-power resonant converter as a part of the auxiliary systems of a train. A series resonant converter is used developed and tested, since it has interesting characteristics for this type of applications, such as galvanic isolation, good power density and high efficiencies.

Cybersecurity has become a key enabler within today's electricity distribution networks. It is a must-have feature to deliver the functionality expected by end-users, as well as for the development of new services. The increase in the complexity of distribution power grids, mainly due to new market demands, incorporating concepts such as flexibility, distributed generation, or scalability, among others, make it necessary to establish secure and efficient trust mechanisms between all the elements participating in the distribution operation network. Under this scenario, being able to stablish a firm root of trust between every element in the system, and its preservation during their entire life cycle, becomes a critical protection against new generation cybersecurity attacks, such as advanced persistent threads and coordinated cybercriminal actions