A novel intelligent approach for yaw position forecasting in wind energy systems

https://doi.org/10.1016/j.ijepes.2015.01.030Get rights and content

Highlights

  • •

    The yaw position parameter of a wind turbine is forecasted using multi-tupled inputs.

  • •

    The developed k-NN classifier leads to improve the wind turbine efficiency.

  • •

    Many useful and reasonable outcomes unmentioned in the literature were uncovered.

Abstract

Yaw control systems orientate the rotor of a wind turbine into the wind direction, optimize the wind power generated by wind turbines and alleviate the mechanical stresses on a wind turbine. Regarding the advantages of yaw control systems, a k-nearest neighbor classifier (k-NN) has been developed in order to forecast the yaw position parameter at 10-min intervals in this study. Air temperature, atmosphere pressure, wind direction, wind speed, rotor speed and wind power parameters are used in 2, 3, 4, 5 and 6-dimensional input spaces. The forecasting model using Manhattan distance metric for k = 3 uncovered the most accurate performance for atmosphere pressure, wind direction, wind speed and rotor speed inputs. However, the forecasting model using Euclidean distance metric for k = 1 brought out the most inconsistent results for atmosphere pressure and wind speed inputs. As a result of multi-tupled analyses, many feasible inferences were achieved for yaw position control systems. In addition, the yaw position forecasting model developed was compared with the persistence model and it surpassed the persistence model significantly in terms of the improvement percent.

Introduction

A wind turbine includes many interconnected mechanical components such as blades, rotor, gearbox, bearings, yaw system, pitch system and tower [1], [2]. Yaw and pitch control systems reduce the fatigue loads caused by the aerodynamic forces and increase the production of electrical energy from wind energy [3]. Particularly, yaw control systems track the wind direction and face the wind stream perpendicularly [4]. In addition, yaw control systems also drive the rotor mechanism out of the wind in order to decrease its rotational speed [5]. As a result, yaw position parameter has a critical role in wind energy systems. However, it is difficult to adjust yawing moment in time due to the inertia problem of wind turbine in automatic-oriented yaw control systems [6]. For this reason, yaw position forecasting contributes the efficient and the safe operation of wind turbines.

Farret et al. determined the maximum wind power corresponding to the optimum wind direction and a sensorless yaw control system was realized [7]. Chen et al. designed a fuzzy proportional-integral-derivative system for yaw position control and the wind direction was tracked in a high precision way [8]. Fadaeinedjad et al. simulated the aerodynamic, mechanical and electrical aspects of a fixed-speed wind turbine and yaw errors lead to the voltage and power oscillations [9]. Kusiak et al. optimized the blade yaw angle using an evolutionary computation algorithm and the power output of a wind turbine was upgraded [10]. Lee et al. implemented a maximum power point tracking algorithm and ensured the accurate yawing torque [11]. Rijanto et al. processed wind direction signals in an electronic yaw controller and dissipated the cyclic instabilities of a horizontal-axis wind turbine [12]. Chenghui et al. proposed an intelligent yaw controller based on artificial neuro-endocrine-immunity system and improved the stability and robustness of the yaw control system [13]. Owing to the lack of academic studies in the field of yaw position forecasting, the main objective of this paper is to forecast the yaw position parameter of a wind turbine using air temperature, atmosphere pressure, wind direction, wind speed, rotor speed and wind power parameters in multi-tupled inputs. The developed yaw position forecasting model considers the number of nearest neighbors, the dimension of input parameters, the selected distance metric and minimized the yaw position error remarkably by reducing it to 1.100° of MAE, 0.405% of MAPE and 1.209% of NRMSE in this paper. However, maximum yaw error and standard deviation of 10° in the literature were distributed [14], [15], [16]. On the other hand, the k-NN classifier outperforms with 75.5% improvement in comparison for the persistence model. MAE, MAPE and NRMSE values of the persistence model were obtained as 3.966°, 1.652% and 6.634%, respectively.

This paper is organized as follows. Section ‘Lazy learning model’ focuses on the k-NN classifier as a lazy learning approach and introduces the activity diagram of the yaw position forecasting model developed. Section ‘Yaw position forecasting’ explains the dataset properties, and distance and error metrics used in this study. The yaw position forecasting results based on multi-tupled inputs were compared. Finally, in section ‘Conclusions’, the work was concluded and the future studies were given.

Section snippets

Lazy learning model

Lazy learners store the training instances and do not construct any classification model until receiving a test instance [17]. However, lazy learners enable to model complex decision spaces having hyperpolygonal shapes compared to other learning algorithms [18]. Therefore, lazy learners have a wide range of application in pattern recognition. The k-NN classifier is also based on lazy learning and it initially considers each instance in training and test datasets as a point in an n-dimensional

Yaw position forecasting

Wind speed as a time series shows chaotic characteristics and wind direction often change due to its stochastic nature [21], [22]. Yaw control systems need to be employed to capture the maximum wind power and to improve the wind turbine efficiency [23]. For this reason, the yaw position at 10-min intervals is predicted in this study. Unlike the studies in the literature, in this paper, air temperature (Ta), atmosphere pressure (Pa), wind direction (Wd), wind speed (Ws), rotor speed (Rs) and

Conclusions

In this paper, a novel intelligent approach based on the k-nearest neighbor classification was proposed for yaw position forecasting in wind energy systems. The k-NN classifier proposed in this study employed the air temperature, the atmosphere pressure, the wind direction, the wind speed, the rotor speed and the wind power parameters within 2, 3, 4, 5 and 6-tupled inputs for the nearest neighbor numbers of 1, 2, 3, 4 and 5. From the forecasting analyses, the yaw position forecasting model

References (27)

  • X. Zhao et al.

    Review of evaluation criteria and main methods of wind power forecasting

    Energy Procedia

    (2011)
  • F.P.G. Márquez et al.

    Condition monitoring of wind turbines: techniques and methods

    Renewable Energy

    (2012)
  • L. Wang et al.

    Wind power systems: applications of computational intelligence

    (2010)

    • Cited by (13)

    • Improving wind turbine efficiency through detection and calibration of yaw misalignment

      2020, Renewable Energy
      Citation Excerpt :

      The basic principle of yaw control strategy is when the average yaw angle (or the cumulative yaw angle) for a specific time interval exceeds the threshold, yaw control system works, and then yaw motor drives the nacelle to move toward the wind direction [12–14]. Several studies enhanced the tracking performance by forecasting wind direction or the azimuth angle of nacelle [15–17]. For most wind turbines, wind vane sensor is installed on the nacelle rooftop behind the impeller, so flow distortion caused by blade activities affects the measurement of yaw angle which also causes dynamic misalignment.

    • Investigation of energy output in mountain wind farm using multiple-units SCADA data

      2019, Applied Energy
      Citation Excerpt :

      On the other hand, performance optimization techniques are applied to wind turbines to improve its performance [6], for example, automatic identification [7] and inviscid interaction technique [8]. There are also research works on how to improve energy output by optimizing pitch control based on fuzzy logic [9] and hybrid controller [10], or by optimizing yaw control based on the parametric model for wake effects [11] and k-NN classifier [12]. Recently, the climatic trend of wind energy parameters is calculated based on wind field data for the past 36 years [13].

    • Effect investigation of yaw on wind turbine performance based on SCADA data

      2018, Energy
      Citation Excerpt :

      In the field of yaw control, CFD simulation has shown that yaw setting optimization can increase the energy production in a wind plant [31]; a study on the Princess Amalia Wind Park showed that significant increases in annual energy production by employing wake steering with yaw control [32]. On the basis of the traditional yaw control based on wind direction measurement, yaw control methods based on the output power [33] and rotor speed [34] are also presented; advanced wind LIDAR technology [35,36] and some advanced control theories have been introduced into the application of yaw control, such as FLC [37] and k-NN(k-nearest neighbor classifier) [38]. Besides theoretical and experimental analyzing, SCADA data investigating is a new trend to analyze wind turbine operation mechanism, for example, SCADA data recorded for six months has been used to investigate wake impacts on wind turbine performance and yaw alignment [39].

    • A novel implementation of kNN classifier based on multi-tupled meteorological input data for wind power prediction

      2017, Energy Conversion and Management
      Citation Excerpt :

      In addition, many improved models in the literature have more complicated algorithmic structure and most of academic community ignore the impact of multi-tupled meteorological input data on the wind power prediction. In our previous studies [29,30], we have uncovered that the multi-tupled meteorological inputs adapted with distinct nearest neighbors and distance measures optimize the prediction skills of wind speed and yaw position forecasting models. In this context, the main objective of present paper is to develop an accurate, simple and efficient very short-term wind power prediction model based on the kNN classifier and to provide a strength benchmark opportunity by comparing the prediction results with the persistence reference model.

    • Overall equipment effectiveness as a metric for assessing operational losses in wind farms: a critical review of literature

      2023, International Journal of Sustainable Energy

    • Application of fuzzy logic method in wind turbine yaw control system to obtain maximum energy: a methodological and prototype approach

      2022, Electrical Engineering
    View all citing articles on Scopus
    View full text
    Copyright © 2015 Elsevier Ltd. All rights reserved.