Abstract:Diagnosis of ice accretion on wind turbine blades is all the time a hard nut to crack in condition monitoring of wind farms. Existing methods focus on mechanism analysis of icing process, deviation degree analysis of feature engineering. However, there have not been deep researches of neural networks applied in this field at present. Supervisory control and data acquisition (SCADA) makes it possible to train networks through continuously providing not only operation parameters and performance parameters of wind turbines but also environmental parameters and operation modes. This paper explores the possibility that using convolutional neural networks (CNNs), generative adversarial networks (GANs) and domain adaption learning to establish intelligent diagnosis frameworks under different training scenarios. Specifically, PGANC and PGANT are proposed for sufficient and non-sufficient target wind turbine labeled data, respectively. The basic idea is that we consider a two-stage training with parallel GANs, which are aimed at capturing intrinsic features for normal and icing samples, followed by classification CNN or domain adaption module in various training cases. Model validation on three wind turbine SCADA data shows that two-stage training can effectively improve the model performance. Besides, if there is no sufficient labeled data for a target turbine, which is an extremely common phenomenon in real industrial practices, the addition of domain adaption learning makes the trained model show better performance. Overall, our proposed intelligent diagnosis frameworks can achieve more accurate detection on the same wind turbine and more generalized capability on a new wind turbine, compared with other machine learning models and conventional CNNs.

Detection of Aircraft In-flight Icing in Non-steady Atmosphere Using Artificial Neural Network

Research on inflight parameter identification and icing location detection of the aircraft

Inflight parameter identification and icing location detection of the aircraft: the time-varying case

In-flight detection and identification and accommodation of aircraft icing

Model Identification of Aircraft Vertical Tail Using Neural Network

Development of a Real-Time In-Flight Ice Detection System via Computational Aeroacoustics and Bayesian Neural Networks

Aircraft Inflight Icing Detection Based on Statistical Decision Theory

Time Invariant Parameter Identification of Inflight Aircraft Icing

Prediction of Temperature Distribution on an Aircraft Hot-Air Anti-Icing Surface by ROM and Neural Networks

Adaptive Neural Networks-Based Dynamic Inversion Applied to Reconfigurable Flight Control and Envelope Protection Under Icing Conditions

H ∞ Parameter Identification and H 2 Feedback Control Synthesizing for Inflight Aircraft Icing

Aircraft icing detection via flow-induced ambient random noise: from theory to wind tunnel experiments

Machine Learning and IoT Trends for Intelligent Prediction of Aircraft Wing Anti-Icing System Temperature

Intelligent Icing Detection Model of Wind Turbine Blades Based on SCADA data

Development and Implementation of Physics-Informed Neural ODE for Dynamics Modeling of a Fixed-Wing Aircraft under Icing/Fault

Modeling and Simulation of Aircraft Ice Shape Prediction

Aircraft icing mechanism analysis methods

Research on Comprehensive Parameter Identification of Inflight Aircraft Icing

Innovative AI-Enhanced Ice Detection System Using Graphene-Based Sensors for Enhanced Aviation Safety and Efficiency

Neural Network and Dynamic Inversion Based Adaptive Control for a HALE-UAV Against Icing Effects

Aircraft Icing Detection Technology