LAUSR

Background Low-frequency (20–200 Hz) noise (LFN) from wind turbines has received much public attention due to potential health concerns. This work tries to estimate the sound power level of wind turbines ( L W ,A (dB)) at 20–200 Hz, which are not provided by manufacturers but essential for estimating LFN exposure ( L Aeq ) of nearby residents. Methods L W ,A at 20–200 Hz at three wind farms, each with a different brand of wind turbine (Brands A, B and C, respectively) were estimated using propagation model ISO 9613-2 under different weather conditions (rain, wind speed and wind direction) and validated with LFN monitoring data. The feasibility of using validated L W ,A as inputs for ISO 9613-2 to simulate residents’ exposure ( L Aeq ) to LFN were assessed considering interferences from rain, wind speed and wind directions. Results The average L W ,A at 20–200 Hz ranged between 93.2 and 100.4 dB, 97.8 and 107.2 dB, and 96.5 and 110.4 dB for turbines of Brands A, B, and C, respectively, operating under wind speeds from 2 to 12 m/s. The L W ,A at wind speed of 2–8 m/s increased on average by 1.4, 1.9 and 1.7 dB per 1 m/s increase for Brands A, B and C, respectively. The differences in modeled L eq obtained through the input of L W ,A into the ISO 9613-2 model and the measured L Aeq for the three studied wind farms all fall within 1.5 dB. Conclusion This study successfully determined and validated the L W ,A of wind turbines of three brands, and subsequent residents’ LFN exposure (with 1.5 dB difference) at three wind farms. Accurately obtaining LFN exposure will serve as the basis for assessing LFN exposure-health relationship. As wind power widely use worldwide, health impact should be assessed based on validated LFN exposure assessment.