Sensitivity analysis for evaluation of the effect of sensors error on the wind turbine variables using Monte Carlo simulation

Abstract

The dynamics of the wind turbine behaviour and identifying the factors that change turbine performance are very complex and challenging. Quantifying the impact of these factors on improving the wind turbine performance is invaluable. Many attempts have been made to describe the behaviour of the wind turbine variables. Sensitivity analysis and Monte Carlo simulation are methods to identify important parameters affecting model behavior. Here, with these methods, the authors investigate the effect of error in sensors on output power and measured variables of the wind turbine in partial load region and full load region. Then, using the importance factor, the effect of error in the different sensors on the output power of the wind turbine is ranked in the two operating regions, so that, in practice, the sensors that have the greatest impact on the output power are designed as redundant.

Effect of Sensor Faults on the Stresses Caused by Wind Turbine Blades

“…Te controller receives the measured values from sensors such as the PAB, generator speed, and output power and adjusts PABs (β 1,2,3 ref ) and generator torque (τ g ref ). Te nonlinear equation of the WT model is written as follows [21,22]:…”

Quantifying the impact of sensor precision on power output of a wind turbine: A sensitivity analysis via Monte Carlo simulation study

“…While recent research has explored sensorless strategies, exemplified by works such as Aldin et al (2023), Shen et al (2009), Morfin et al (2021), and Benzaouia et al (2021), the WT industry today extensively utilizes sensors for control and monitoring purposes. Researchers like Tian et al (2023), Liu et al (2020), Kini et al (2023), da Silva et al (2022, Biazar et al (2022b), and Maheshwari et al (2023) have utilized sensors in their studies due to Department of Systems and Control Engineering, K. N. Toosi University of Technology, Tehran, Iran the benefits they offer, encompassing superior precision, low algorithmic control complexity, and rapid convergence speed.…”

“…Then, using the conventional regression coefficient, the impacts of the SA were ranked in the partial load (PL) and full load (FL) regions separately. Furthermore, in Biazar et al (2022a), in addition to the faults investigated in Biazar et al (2022b), they studied various faults in the rotor, drive, and power of WT in more detail. It is worth noting, however, that the assumption made in both studies that faults and errors occur independently could limit the practical application of these findings, as real-world scenarios may involve complex interactions between multiple faults and errors.…”

Fault diagnosis-based observers using Kalman filters and Luenberger estimators: Application to the pitch system fault actuators

“…Different researches and studies have been proposed to deal with the diagnosis task of wind turbines using several approaches [3,[5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20]. These proposed methodologies adopt distinctive design of schemes, resulting in different properties according to the used techniques.…”

“…Authors in [17] have proposed a model-based FDI scheme as hybrid model based on adaptive thresholds varying on time in order to guarantee false-alarms for fault detection, approximation, and isolation estimator. Hence, paper [18] considered a Monte-Carlo method for the fault evaluation of sensors in a WT as a sensitivity analysis task. Deep learning algorithms are widely investigated due to their powerful performances for the fault diagnostic and prognosis of WT machine such as in papers [19,20,21,22].…”