مقاله انگلیسی رایگان در مورد بررسی شبکه های عصبی مصنوعی در سیستم های انرژی باد – الزویر 2018

Wind energy has become one of the most important forms of renewable energy. Wind energy conversion systems are more sophisticated and new approaches are required based on advance analytics. This paper presents an exhaustive review of artificial neural networks used in wind energy systems, identifying the methods most employed for different applications and demonstrating that Artificial Neural Networks can be an alternative to conventional methods in many cases. More than 85% of the 190 references employed in this paper have been published in the last 5 years. The methods are classified and analysed into four groups according to the application: forecasting and predictions; design optimization; fault detection and diagnosis; and optimal control. A statistical analysis of the current state and future trends in this field is carried out. An analysis of each application group about the strengths and weaknesses of each ANN structure is carried out. A quantitative analysis of the main references is carried out showing new statistical results of the current state and future trends of the topic. The paper describes the main challenges and technological gaps concerning the application of ANN to wind turbines, according to the literature review. An overall table is provided to summarize the most important references according to the application groups and case studies.

Introduction and motivation

Nowadays, wind energy is one of the most important renewable energy sources. In 2016, wind energy systems (WES) provided more than 420 GW, and this is expected to rise to more than 1000 GW in the 2030s [1]. WES are undergoing a modernization process where the number of requirements has increased to ensure efficient energy production [2,3]. There has been an increase in WES and their complexity, as well as the demand for new techniques and methods to improve reliability [4], maintenance [5] and investments [6], leading to greater competitiveness in the energy market. The wind energy market, as one of the most exploitable and growing markets, requires both technical and economic advances. Regarding the technical aspects, efforts are oriented towards harnessing the wind to a maximum level. There are many issues addressed in the literature, such as the aerodynamic optimization of wind turbines (WT) [7], the optimization of blade shapes [8], the study of power curve under different circumstances [9], the optimization of WT position in a wind farm [10], etc. With respect to the economic issues, the main objective is to maximize profits obtained from the available resources. For this purpose, the literature covers topics such as wind speed modelling [11], strategies based on energy price forecasting [12], the study of the interactions between wind energy and the power market [13], wind turbine life cycle analysis [14], etc. This paper shows an exhaustive review of the current techniques and methods concerning these issues employing artificial neural networks (ANN) WTs are equipped with a large number of devices to evaluate the humidity, temperature, vibration, etc. [15]. Data acquisition systems measure all the variables in order to determine the system condition [16]. Data processing requires robust algorithms [17] that enable as much information as possible to be gathered from the available data [18]. Machine learning algorithms are widely employed due to their ability to process a large amount of data, ANNs being one of the most employed methods [19]. ANNs are complex structures based on biological neurons. These structures provide a good solution to problems that cannot be analytically defined. An ANN consists of neurons which are simple processing units, and weighted connections between those neurons. A typical structure corresponds to the multilayer perceptron (MLP), shown in Fig. 1 [20]. The ANN receives a dataset and starts a training process to adjust the weights of the interconnections between neurons. The training will be supervised if the output is known, otherwise it will be named unsupervised training [21].