Real-Time Hierarchical Neural Network Based Fault Detection and Isolation for High-Speed Railway System Under Hybrid AC/DC Grid

Abstract

Reliable and comfortable high-speed railway (HSR) has skyrocketed in popularity as a transportation medium for traveling around the world. High-voltage direct current (HVDC) electrification system has been introduced to the HSR gradually. However, the coexistence of AC and DC systems will last for a long time because AC railway systems are still in the dominant position. A detailed HSR traction system transient model operating under the hybrid AC/DC grid was established in PSCAD/EMTDC. We proposed a real-time fault detection and isolation (FDI) method for the simulated model using neural network (NN). Hierarchical structure of the monitoring system has been employed. Low-level sub-monitors supervised the conditions of their local regions and the top-level monitor collected all the feedback from sub-monitors making the final evaluation of the entire HSR system based on a voting strategy. Both off-line and real-time experiments were conducted to validate the effectiveness of the proposed method. In the experiments, the sub-monitors were designed based on Gated Recurrent Unit (GRU) algorithm and implemented on the Xilinx VCU128 FPGA board. For the off-line experiment, the sub-monitors used the training and testing dataset both from PSCAD/EMTDC to construct the architecture of their individual GRU networks and to verify how great the networks can be. For the real-time task, the sub-monitors interfaced with a real-time HSR system emulator running on the Xilinx VCU118 FPGA board to test the performance in the real-time application. The results proved that our proposed FDI method has the capability of real-time detection and can achieve better accuracy within reasonable time and resource consumption than other NN-based methods. Moreover, the method was capable of standing against noises from measured signals to some extent. Index Terms-Data-driven methods, fault detection and isolation (FDI), field-programmable gate array (FPGA), Gated Recurrent Unit (GRU), high-speed railway (HSR), high-voltage direct current (HVDC), hybrid AC/DC grid, long short-term memory (LSTM), machine learning, neural networks, real-time systems.

Real-Time ML-Assisted Hardware-in-the-Loop Electro-Thermal Emulation of LVDC Microgrid on the International Space Station

“…In this scenario, in order to meet the computation ability and memory requirements of real-time systems, the research needs to expand from both algorithm optimization and hardware aspects. For the former, an emerging equivalent prediction fitting approach based on machine learning (ML) and neural networks (NNs) represents a novel way of circuit transient solution, which has been applied in power electronic applications [20], [21]. After learning from given datasets, the NNs produce the Pareto fronts and select the optimal designs, which means the explicit physical significance is only contained inside the input and output variables but not NN itself [22].…”

“…After learning from given datasets, the NNs produce the Pareto fronts and select the optimal designs, which means the explicit physical significance is only contained inside the input and output variables but not NN itself [22]. In the latter case, as opposed to CPUs that adopt fixed computing architecture, field-programmable gate arrays (FPGAs) provide an intrinsic parallelism without predefined hardware architecture causing them to be the ideal real-time emulation hardware acceleration platform for the parallelization of NNs [20]. Furthermore, FPGAs have been extensively used for real-time HIL emulation of AC/DC power systems [23].…”

Machine Learning Based Transient Stability Emulation and Dynamic System Equivalencing of Large-Scale AC-DC Grids for Faster-Than-Real-Time Digital Twin

“…Artificial neural network (ANN) based power system dynamic equivalencing has been previously proposed in [14]; however, traditional ANN-based equivalent models are relatively simple, and cannot meet the requirements for long-term prediction for a large-scale network. Therefore, the Gated Recurrent Unit (GRU) algorithm is utilized, which is a variant of the Long Short-Term Memory (LSTM) network that has a higher accuracy in representing the non-linear parts for dynamic equivalencing [15].…”