This paper introduces an approach with the Transformer Neural Networks model for partial discharge patterns classification, that consists of corona discharge, internal discharge and surface discharge. The PD measuring...
详细信息
ISBN:
(数字)9798350374605
ISBN:
(纸本)9798350386165
This paper introduces an approach with the Transformer Neural Networks model for partial discharge patterns classification, that consists of corona discharge, internal discharge and surface discharge. The PD measuring circuit suggested in IEC 60270:2000 is used to record Partial discharge signals. Independent parameters such as phase and charge of PD patterns were recorded. The phase value will be encoded into the charge array and Transformer Neural Network is constructed using Positional Embedding and Transformer Encoder Layer. 80% of the recorded data will be used as a training data and 20% recorded data was used for testing of the classification models. Impacts of neuron numbers and network architecture on the PD classification performance will be observed
This paper presents an FPGA-based low-power acceleration of sound source localization in HARK, open-source software for robot audition. Due to the massive matrix operations, sound source localization in HARK takes sub...
详细信息
ISBN:
(数字)9798350384147
ISBN:
(纸本)9798350384154
This paper presents an FPGA-based low-power acceleration of sound source localization in HARK, open-source software for robot audition. Due to the massive matrix operations, sound source localization in HARK takes substantial processing time in edge computing devices. To balance processing time and low power consumption, two functions in sound source localization that include many matrix operations are targeted and migrated on an FPGA SoC board called M-KUBOS. Compared to CPU-based computing on ARM Cortex A53, our implementation achieved a 2.0× speedup and 1.7× lower energy consumption.
Having access to a reliable and accurate prediction of the short-term power demand is a fundamental step for the widespread adoption of Electric Vehicles (EVs), as their charges may have a significant impact on the po...
Having access to a reliable and accurate prediction of the short-term power demand is a fundamental step for the widespread adoption of Electric Vehicles (EVs), as their charges may have a significant impact on the power system balancing. In this direction, we propose a short-term load demand predictor, based on distributed Long Short-Term Memory Networks, that employs consensus and fully-decentralized Federated Learning (FL) algorithms to seek cooperation among multiple points of charge without the requirement of sharing any user-related data.
This paper proposes a linear quadratic regulator (LQR)-based controller for a grid-connected photovoltaic (PV) with a supercapacitor (SC) system to ensure a smooth transition of control mode when the SC is unavailable...
详细信息
ISBN:
(数字)9798350376067
ISBN:
(纸本)9798350376074
This paper proposes a linear quadratic regulator (LQR)-based controller for a grid-connected photovoltaic (PV) with a supercapacitor (SC) system to ensure a smooth transition of control mode when the SC is unavailable (completely charged) during fault operations. The SC improves the low-voltage ride-through (LVRT) capability of the PV system by quickly balancing power between the PV system and the grid. When the SC reaches its full charge capacity, the DC-link voltage will be controlled by the inverter’s controller, and the PV system may need to curtail its power generation. During the transition, fluctuations in the DClink voltage can occur which can compromise the stability of the PV system, especially when the control mode switching happens during severe fault operations. The proposed LQR-based controller is validated on a hardware testbed platform to demonstrate its effectiveness in reducing DC-link overvoltage during the control mode transitions. As a result, the PV system can remain compliant with LVRT requirements while maintaining stability during fault operations.
With the rapid development of recycling and remanufacturing technologies, disassembly line balancing problems (DLBP) have drawn great attention. Considering the limitation of disassembly by humans or robots alone, thi...
详细信息
In the last years' digitalization process, the creation and management of documents in various domains, particularly in Public Administration (PA), have become increasingly complex and diverse. This complexity ari...
详细信息
This paper addresses the problem of adaptive output regulation of a minimum-phase MIMO LTI system in the presence of un-modeled (fast) input dynamics. The adoption of a post-processing tunable internal model makes it ...
This paper addresses the problem of adaptive output regulation of a minimum-phase MIMO LTI system in the presence of un-modeled (fast) input dynamics. The adoption of a post-processing tunable internal model makes it possible to implement standard methods for the analysis of two-time-scale systems. The proposed adaptation law guarantees, under suitable hypotheses, convergence to zero of the regulation error as well as of the parameter estimation error.
Artificial neural networks (ANN) have been shown to be flexible and effective function estimators for the identification of nonlinear state-space models. However, if the resulting models are used directly for nonlinea...
Artificial neural networks (ANN) have been shown to be flexible and effective function estimators for the identification of nonlinear state-space models. However, if the resulting models are used directly for nonlinear model predictive control (NMPC), the resulting nonlinear optimization problem is often overly complex due to the size of the network, requires the use of high-order observers to track the states of the ANN model, and the overall control scheme does not exploit the available autograd tools for these models. In this paper, we propose an efficient approach to auto-convert ANN statespace models to linear parameter-varying (LPV) form and solve predictive control problems by successive solutions of linear model predictive problems, corresponding to quadratic programs (QPs). Furthermore, we show how existing deep-learning methods, such as SUBNET that uses a state encoder, enable efficient implementation of MPCs on identified ANN models. Performance of the proposed approach is demonstrated by a simulation study on an unbalanced disc system.
The paper examines a new approach to controlling deterministic dynamical chaotic mode of power system with two electric generate-sources. A control law is synthesized in the class of one-parameter structurally stable ...
The paper examines a new approach to controlling deterministic dynamical chaotic mode of power system with two electric generate-sources. A control law is synthesized in the class of one-parameter structurally stable mappings from catastrophe theory. Analytical calculations and a numerical experiment in MatLab showed that the proposed system excludes the generation of a deterministic chaotic mode for any changes in the uncertain system parameters. The study of the control system is carried out by the gradient-velocity method Lyapunov vector functions
暂无评论