Fast and Accurate Deep Learning Framework for Secure Fault Diagnosis in the Industrial Internet of Things

Abstract

This paper introduced a new deep learning frame-1 work for fault diagnosis in electrical power systems. The frame-2 work integrates the convolution neural network and different re-3 gression models to visually identify which faults have occurred in 4 electric power systems. The approach includes three main steps, 5 data preparation, object detection, and hyper-parameter opti-6 mization. Inspired by deep learning, evolutionary computation 7 techniques, different strategies have been proposed in each step 8 of the process. In addition, we propose a new hyper-parameters 9 optimization model based on evolutionary computation that can 10 be used to tune parameters of our deep learning framework. 11In the validation of the framework's usefulness, experimental 12 evaluation is executed using the well known and challenging VOC 13 2012, the COCO datasets, and the large NESTA 162-bus system. 14The results are very promising against many current state-of-the-15 art solutions in terms of runtime and accuracy performances.

RETRACTED: Fault diagnosis of control system based on neural network data fusion

“…The procedure consists of three stages: gathering relevant data, locating the target item, and ne-tuning the method's hyper-parameters. Different methodologies have been suggested at each phase of the process, inspired by deep learning and evolutionary computing approaches [12]. An IoT technique was developed and applied to diagnose bearing faults at changing speeds using an Internet of Things (IoT) node.…”

A systematic literature review of machine learning applications in IoT

“…By utilizing past data gatheblack for a particular activity to automatically optimize performance without iteration, ML may be utilized to enhance the performance of various IoT devices. [15][16][17][18][19][20][21] Specifically, the following points are the primary reasons why ML is significant in IoT applications: • The configurations that IoT systems frequently screen have a dynamic structure. As a result, it is critical to develop IoT systems that can effectively respond to these changes as a whole.…”

“…The IoT enables the connection of millions of objects, including sensors and mobile phones/devices, to carry out numerous functions. By utilizing past data gatheblack for a particular activity to automatically optimize performance without iteration, ML may be utilized to enhance the performance of various IoT devices 15–21 . Specifically, the following points are the primary reasons why ML is significant in IoT applications: The configurations that IoT systems frequently screen have a dynamic structure.…”

A self-attention integrated spatiotemporal LSTM approach to edge-radar echo extrapolation in the Internet of Radars

“…Recently, artificial neural networks have been widely used and achieved outstanding performances in fields such as computer vision [12], edge computing [13,14,15,16,17,18], anomaly detection [19,20,21], data mining [22,23,24,25], algorithm optimization [26,27,28,29], medical diagnosis [30,31], and climate prediction [32], etc. This has attracted widespread attention from researchers in the field of weather forecasting, and they began to apply these networks to radar echo extrapolation.…”