EPE Association: EPE 2021 - Dialogue session - Data Analysis, Artificial Intelligence and Communication

EPE 2021 - Dialogue session - Data Analysis, Artificial Intelligence and Communication
You are here: EPE Documents > 01 - EPE & EPE ECCE Conference Proceedings > EPE 2021 ECCE Europe - Conference > EPE 2021 - Topic 10: Data Analysis, Artificial Intelligence and Communication Issues > EPE 2021 - Dialogue session - Data Analysis, Artificial Intelligence and Communication

   [return to parent folder]

   Auto Structural Optimization of Toroidal Coils Based on a Neural Network and a Genetic Algorithm
By Junichi KASHIWAGI [View]
[Download]

Abstract: The combination of a neural network (NN) and a genetic algorithm (GA) achieves the autonomousstructural design of a toroidal coil to optimize its resistance (R) and radiated magnetic field noise (B).Instead, of the finite element method, NNs are used to calculate R, inductance (L), and B. The NNs aretrained using the relationship between four structural parameters and the R, L, and B data set. Thetoroidal coil structures are optimized under a constraint condition of L while effectively using trainedNNs and a GA. As a result, R is reduced by 10\%, the power supply efficiency is improved by 1.3\%, andB is reduced by 17.6 dB.

   Enhancing Transient Stability of Power Synchronization Control via Deep Learning
By Amir SEPEHR [View]
[Download]

Abstract: Transient stability of grid-connected converters has become a critical concern to the continuing integration of power electronic converters into the power system. Therefore, this paper aims to study the transient stability of power synchronization control (PSC) and propose a developed control system by employing deep learning methods. A long short-term memory (LSTM) network has been trained and integrated into PSC for adapting the synchronization loop of the converter to the grid transients. The trained LSTM network extracts and predicts the voltage trajectory of the connection point with respect to the converter dynamics and the grid strength. In the developed control system, active power reference and internal voltage of the converter are updated dynamically to both satisfy the low voltage ride through (LVRT) requirements and prevent the loss of synchronization. The developed control system is validated by time-domain simulation results.

   Evaluation of energy flow data in emobility applications to create efficiency maps for simulation setup and system optimisation
By Mathias HERGET [View]
[Download]

Abstract: As part of a research project at Fulda University of Applied Sciences (Germany), the energy efficiencyof battery-electric commercial vehicles in real-world operation is being analysed. For this purpose,measurement data was collected from the driving operation of a battery-electric truck. The paper explains how measured data can be used to optimize the simulation of battery-electric commercial vehicles. The focus is on the generation and application of efficiency maps in order to more realistically represent the efficiencies of electrical components within the vehicle depending on relevant influencing parameters. The generated efficiency maps can also be used to estimate the energy flows of the subcomponents and their effect on the overall energy demand and on the range of the electric vehicle.

   Supervised imitation learning of FS-MPC algorithm for multilevel converters
By Mateja NOVAK [View]
[Download]

Abstract: Model predictive control (MPC) applications for multilevel power electronics converters are often facing problems of a high computation burden. By using supervised imitation learning, it is possible to synthesise computationally light controllers, which can capture the behaviour of computationally heavy MPC. To obtain a high performance controller, which can do the correct control actions, training data generation and pre-processing of the data are of high importance. This paper presents guidelines for training data generation and artificial neural network (ANN) controller design for a multistep-horizon finite set FS-MPC applied to neutral point clamped (NPC) converter. A particular challenge of the selected converter topology is that some control actions are used more often than others, thus the training data will be heavy skewed i.e. it will be difficult for the controller to learn when to apply these actions due to the lack of data. A workaround for solving this challenge is discussed in the paper. The performance and the robustness of the designed controller has been validated in a hardware in the loop (HIL) system, where the limitations of the synthesised ANN controller were explored. It was observed that ANN controller performance can match the performance of the FS-MPC algorithm when operating within the span of training data values and the computational burden was much lower.

   Wireless control of modular multilevel converter autonomous submodules
By Baris CIFTCI [View]
[Download]

Abstract: The wireless control of modular multilevel converter (MMC) submodules might offer advantages for MMCs with a high number of submodules. However, the control system should tolerate the stochastic nature of the wireless communication, continue the operation flawlessly or, at least, avoid overcurrents, overvoltages, and component failures. The previously proposed control methods enabled to control the submodules wirelessly with consecutive communication errors up to hundreds of control cycles. The submodule control method in this paper facilitates the MMC to safely overcome communication errors that last longer and when the MMC experiences significant electrical disturbances during the errors. The submodules are proposed to operate autonomously by implementing a replica of the central controller in the submodules and drive the replicas based on the local variables and the previously received data. The simulation and experimental results verify the proposed control method.

   Wireless control of modular multilevel converter submodules under ac-side faults
By Baris CIFTCI [View]
[Download]

Abstract: Wireless control of modular multilevel converter (MMC) submodules has been offered recently with potentially lower cost and higher availability advantages for the converter station. In this paper, the wireless control of MMC submodules under ac-side faults is investigated. The central controller of the MMC is equipped for the unbalanced grid conditions. Local current controllers in the submodules are operated autonomously in case of loss of wireless communication during the fault. A set of simulations with single line-to-ground, line-to-line, and three-phase-to-ground faults reveal that the MMC rides through the faults in all the cases with the expected communication conditions or when the communication is lost before or after the fault instant.

EPE 2021 - Dialogue session - Data Analysis, Artificial Intelligence and Communication
You are here: EPE Documents > 01 - EPE & EPE ECCE Conference Proceedings > EPE 2021 ECCE Europe - Conference > EPE 2021 - Topic 10: Data Analysis, Artificial Intelligence and Communication Issues > EPE 2021 - Dialogue session - Data Analysis, Artificial Intelligence and Communication
	[return to parent folder]

	Auto Structural Optimization of Toroidal Coils Based on a Neural Network and a Genetic Algorithm By Junichi KASHIWAGI	[View] [Download]
Abstract: The combination of a neural network (NN) and a genetic algorithm (GA) achieves the autonomousstructural design of a toroidal coil to optimize its resistance (R) and radiated magnetic field noise (B).Instead, of the finite element method, NNs are used to calculate R, inductance (L), and B. The NNs aretrained using the relationship between four structural parameters and the R, L, and B data set. Thetoroidal coil structures are optimized under a constraint condition of L while effectively using trainedNNs and a GA. As a result, R is reduced by 10\%, the power supply efficiency is improved by 1.3\%, andB is reduced by 17.6 dB.

	Enhancing Transient Stability of Power Synchronization Control via Deep Learning By Amir SEPEHR	[View] [Download]
Abstract: Transient stability of grid-connected converters has become a critical concern to the continuing integration of power electronic converters into the power system. Therefore, this paper aims to study the transient stability of power synchronization control (PSC) and propose a developed control system by employing deep learning methods. A long short-term memory (LSTM) network has been trained and integrated into PSC for adapting the synchronization loop of the converter to the grid transients. The trained LSTM network extracts and predicts the voltage trajectory of the connection point with respect to the converter dynamics and the grid strength. In the developed control system, active power reference and internal voltage of the converter are updated dynamically to both satisfy the low voltage ride through (LVRT) requirements and prevent the loss of synchronization. The developed control system is validated by time-domain simulation results.

	Evaluation of energy flow data in emobility applications to create efficiency maps for simulation setup and system optimisation By Mathias HERGET	[View] [Download]
Abstract: As part of a research project at Fulda University of Applied Sciences (Germany), the energy efficiencyof battery-electric commercial vehicles in real-world operation is being analysed. For this purpose,measurement data was collected from the driving operation of a battery-electric truck. The paper explains how measured data can be used to optimize the simulation of battery-electric commercial vehicles. The focus is on the generation and application of efficiency maps in order to more realistically represent the efficiencies of electrical components within the vehicle depending on relevant influencing parameters. The generated efficiency maps can also be used to estimate the energy flows of the subcomponents and their effect on the overall energy demand and on the range of the electric vehicle.

	Supervised imitation learning of FS-MPC algorithm for multilevel converters By Mateja NOVAK	[View] [Download]
Abstract: Model predictive control (MPC) applications for multilevel power electronics converters are often facing problems of a high computation burden. By using supervised imitation learning, it is possible to synthesise computationally light controllers, which can capture the behaviour of computationally heavy MPC. To obtain a high performance controller, which can do the correct control actions, training data generation and pre-processing of the data are of high importance. This paper presents guidelines for training data generation and artificial neural network (ANN) controller design for a multistep-horizon finite set FS-MPC applied to neutral point clamped (NPC) converter. A particular challenge of the selected converter topology is that some control actions are used more often than others, thus the training data will be heavy skewed i.e. it will be difficult for the controller to learn when to apply these actions due to the lack of data. A workaround for solving this challenge is discussed in the paper. The performance and the robustness of the designed controller has been validated in a hardware in the loop (HIL) system, where the limitations of the synthesised ANN controller were explored. It was observed that ANN controller performance can match the performance of the FS-MPC algorithm when operating within the span of training data values and the computational burden was much lower.

	Wireless control of modular multilevel converter autonomous submodules By Baris CIFTCI	[View] [Download]
Abstract: The wireless control of modular multilevel converter (MMC) submodules might offer advantages for MMCs with a high number of submodules. However, the control system should tolerate the stochastic nature of the wireless communication, continue the operation flawlessly or, at least, avoid overcurrents, overvoltages, and component failures. The previously proposed control methods enabled to control the submodules wirelessly with consecutive communication errors up to hundreds of control cycles. The submodule control method in this paper facilitates the MMC to safely overcome communication errors that last longer and when the MMC experiences significant electrical disturbances during the errors. The submodules are proposed to operate autonomously by implementing a replica of the central controller in the submodules and drive the replicas based on the local variables and the previously received data. The simulation and experimental results verify the proposed control method.

	Wireless control of modular multilevel converter submodules under ac-side faults By Baris CIFTCI	[View] [Download]
Abstract: Wireless control of modular multilevel converter (MMC) submodules has been offered recently with potentially lower cost and higher availability advantages for the converter station. In this paper, the wireless control of MMC submodules under ac-side faults is investigated. The central controller of the MMC is equipped for the unbalanced grid conditions. Local current controllers in the submodules are operated autonomously in case of loss of wireless communication during the fault. A set of simulations with single line-to-ground, line-to-line, and three-phase-to-ground faults reveal that the MMC rides through the faults in all the cases with the expected communication conditions or when the communication is lost before or after the fault instant.