IEEE Systems, Man and Cybernetics Magazine - January 2023 - 5

A
wind turbine rotor system is a typical networked
industrial control system. The security
of its operation is very important to energy
systems and users. In this article, the artificial
intelligence algorithm is used to predict the
security operation of a wind turbine rotor system, and a prediction
method of system security monitoring based on a convolutional
neural network (CNN) is proposed. First, the
dynamic analysis of the operation principle of the wind turbine
rotor system is carried out, and the industrial control
model of the rotor system is established by using the relevant
data of the wind turbine. The relevant data required for
the security prediction of the wind turbine rotor system are
obtained, and its dataset is established. Then, the CNN is
trained with limited datasets, and the trained CNN is used
to accurately predict the pitch angle. The residual information
is obtained by comparing the predicted pitch angle with
the real output pitch angle of the wind turbine rotor changing
system. Finally, the security prediction results are
obtained according to the residual and the decision index.
The proposed security trend prediction method for wind
turbine rotor systems can accurately and effectively predict
the change of the fault amplitude, provide detection and
estimate decision results, and improve the system security.
Introduction
With the development of today's society, energy issues have
attracted more and more attention. Ro and Choi [1] point out
that the current energy issues facing humankind are mainly
how to effectively use fossil fuels, how to safely use nuclear
energy, and how to develop and utilize renewable energy.
Digital Object Identifier 10.1109/MSMC.2022.3211690
Date of current version: 10 January 2023
Wind power is a typical renewable energy. In recent years,
wind power generation technology, mainly based on wind turbine
system has attracted more and more attention. With the
development of technology, wind power has become more
cost competitive with traditional power generation methods.
Wind power generation can be classified according to the
type of generator used, power control method, constant- or
variable-speed operation, and interconnection with the grid.
Many wind turbines are connected by the network to form a
large networked industrial control system. The system structure
is far more complex than an ordinary power generation
system. At the same time, considering the utilization of wind
energy, wind turbines are mostly installed in places with
perennial high temperatures, frequent storms, and other
extreme weather. The harsh external environment reduces the
operation reliability of wind turbines, and the system is prone
to failure, bringing potential security hazards and economic
losses to wind farm owners. The operation of the security
trend prediction system can trigger an alarm when the industrial
control system has security risks and accurately reflect
the location of the potential risks and possible fault types.
Therefore, to improve the operation stability of wind turbines
and reduce maintenance costs, it is necessary to study
a variety of wind turbine security prediction methods [2]-[9].
In the past 10 years, scholars, combined with different engineering
applications in the field of industrial control system
security prediction, have done a lot of research work, developing
a variety of diagnosis and prediction methods.
On the theoretical level, Zhang et al. [10] pointed out that
closed-loop system faults and microfaults are the difficulties
in current industrial control system security research.
For closed-loop system faults, Zhang and Jiang [11], [12]
pointed out that not all existing security prediction methods
can be used in the design of fault-tolerant closed-loop
January 2023 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE
5
IEEE Systems, Man and Cybernetics Magazine - January 2023

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - January 2023
