Short-time photovoltaic output prediction method based on depthwise separable convolution Visual Geometry group- deep gate recurrent neural network

Frontiers in Energy Research, Journal Year: 2024, Volume and Issue: 12

Published: Aug. 1, 2024

In response to the issue of short-term fluctuations in photovoltaic (PV) output due cloud movement, this paper proposes a method for forecasting PV based on Depthwise Separable Convolution Visual Geometry Group (DSCVGG) and Deep Gate Recurrent Neural Network (DGN). Initially, motion prediction model is constructed using DSCVGG, which achieves edge recognition clouds by replacing previous convolution layer pooling VGG with depthwise separable convolution. Subsequently, results DSCVGG network, along historical data, are introduced into Unit (DGN) establish model, thereby achieving precise output. Through experiments actual Mean Absolute Error (MAE) Squared (MSE) our only 2.18% 5.32 × 10 −5 , respectively, validates effectiveness, accuracy, superiority proposed method. This provides new insights methods improving stability power generation.

Language: Английский

A Secure IIoT Environment That Integrates AI-Driven Real-Time Short-Term Active and Reactive Load Forecasting with Anomaly Detection: A Real-World Application

Sensors, Journal Year: 2024, Volume and Issue: 24(23), P. 7440 - 7440

Published: Nov. 21, 2024

The Industrial Internet of Things (IIoT) revolutionizes both industrial and residential operations by integrating AI (artificial intelligence)-driven analytics with real-time monitoring, optimizing energy usage, significantly enhancing efficiency. This study proposes a secure IIoT framework that simultaneously predicts active reactive loads while also incorporating anomaly detection. system is optimized for deployment on an edge server, such as single-board computer (SBC), well cloud or centralized server. It ensures reliable smart data acquisition systems control, protective measures. We propose Temporal Convolutional Networks-Gated Recurrent Unit-Attention (TCN-GRU-Attention) model to predict loads, which demonstrates superior performance compared other conventional models. metrics load forecasting are 0.0183 Mean Squared Error (MSE), 0.1022 Absolute (MAE), 0.1354 Root (RMSE), forecasting, the 0.0202 0.1077 0.1422 (RMSE). Furthermore, we introduce Isolation Forest detection considers transient conditions appliances when identifying irregular behavior. very promising performance, average all using this being 95% Precision, 98% Recall, 96% F1 Score, nearly 100% Accuracy. To entire system, Transport Layer Security (TLS) Secure Sockets (SSL) security protocols employed, along hash-encoded encrypted credentials enhanced protection.

Language: Английский

Home Assistant platform under DDoS attacks for IPv4 and IPv6 networks

Published: Nov. 13, 2024

Language: Английский