Precise and low-cost computationally efficient method for operational modal analysis in wind turbines

• A new precise and low-cost procedure is presented for the OMA of wind turbines. • The algorithm is computationally efficient and implementable on a microcontroller. • Kalman filters are used to estimate the natural vibration modes in real time. • The damping ratios of the wind turbine vibration modes are precisely calculated. • Using Random Decrement technique the wind turbine modal parameters are estimated. Structural Health Monitoring of wind turbines using Operational Modal Analysis techniques has become increasingly important in the wind industry. This importance is underscored by the fact that many installed wind farms are nearing the end of their operational lifespan and require life extension strategies that ensure safe operation. However, most existing techniques in the state of the art are either imprecise or necessitate complex calculations and high computational costs. These limitations often require extensive data extraction for external processing, the use of complex processors, and the engagement of external services for data analysis, posing significant challenges for wind farm owners. This paper presents an Operational Modal Analysis algorithm designed for Structural Health Monitoring of wind turbines, addressing the aforementioned issues. The proposed algorithm is highly computationally efficient, allowing for implementation on a low-cost electronic node that can autonomously analyze the structural health of the wind turbine with high precision. To achieve this, the algorithm employs a combination of techniques, some of which are novel, such as the modeling of modes and harmonic elimination using linear Kalman filters. Other techniques, such as the Random Decrement Technique and the Ibrahim Time Domain, are well-established in literature. However, the specific combination of these techniques as presented in this paper is also a novelty. All these techniques involve simple calculations, resulting in an efficient algorithm with low computational cost. Moreover, this paper validates the algorithm using both synthetic signals from OpenFAST and real signals from wind turbines. The results are highly satisfactory, outperforming leading techniques in this context and confirming the algorithm's precision. Notably, the algorithm excels in damping estimation, a challenging aspect of Operational Modal Analysis applied to wind turbines, for which no existing Operational Modal Analysis techniques provide precise estimates. In conclusion, the algorithm presented in this paper offers a precise, efficient, and low-cost solution for Structural Health Monitoring of wind turbines, eliminating the need for extensive data processing and external analysis, thereby simplifying and improving the maintenance and operation of wind farms.