How Fast Blades Spin in the Rain, Not How Much It Rains, Drives Wind Turbine Erosion

Rain quietly grinds away the leading edges of wind turbine blades, pitting and roughening the surfaces that slice through the sky at highway-to-racecar speeds. For years, engineers have estimated this damage using rainfall records alone, layering in an assumed rotor speed that the machines are presumed to hold while rain falls. A new study from two wind farms in China shows that this assumption is not a small refinement problem. What the turbines actually do while it rains, the researchers find, can change the estimated erosion exposure by factors of more than two to more than five, and can even reverse which of two sites looks riskier.

The study, published as a preprint under review at the journal Wind Energy Science, was led by Yisheng Liu of Beijing Normal University together with Xinchang Jia, Xiaoya Liu, Meng Jin, Fang Liu, and Yafei Li of Goldwind Science & Technology, one of China’s largest turbine manufacturers. The team did something deceptively simple that previous assessments rarely did: they measured the rain itself, drop by drop, at two wind farms, and then aligned those measurements with the complete operating records of every turbine on site during the same period.

At the wetter site in Guangxi, the instruments reconstructed 513 millimeters of rainfall over the observation period; at the drier site in Jiangsu, 242 millimeters fell. Laser disdrometers recorded the size and fall speed of individual raindrops, capturing the microphysical texture of each storm rather than just a depth of accumulated water. Meanwhile, the supervisory control and data acquisition systems, known as SCADA, logged how fast each rotor actually spun, second by second, through every rainy interval.

The conventional approach to rain-erosion exposure assumes a fixed reference tip speed, typically 90 meters per second, which the researchers note is essentially the rounded value of the fleet’s observed operating-speed ceiling. Erosion models treat impact energy as scaling steeply with speed, so this single number carries enormous weight in any risk calculation. The problem is that turbines do not hold a constant tip speed in rain. They respond to wind gusts, control settings, curtailment orders, and the physics of the storm itself.

When the team replaced the fixed 90-meter-per-second reference with the tip speeds the turbines actually recorded during rain, the results were striking. At the Guangxi farm, the operation-coupled exposure index dropped to 18 percent of its fixed-speed value. At Jiangsu, it fell to 47 percent. In other words, the standard reference speed overstates the true exposure by factors ranging from 2.1 to 5.6, depending on the site. The operating state does not merely refine the estimate; it sets its size.

By contrast, the details of the rain’s microphysics mattered far less than expected. When the researchers swapped a parameterized raindrop fall-velocity relation for the actual measured drop-size and velocity spectra, the same exposure index moved by only 28 to 38 percent. This is a genuinely counterintuitive finding for the field: how hard the drops hit depends more on how fast the blade is moving than on subtle differences in how the drops fall.

The consequences ripple through site selection and fleet management. Under the common reference speed, the wetter Guangxi farm appeared roughly twice as exposed as Jiangsu, with index values of 0.140 versus 0.066. But once real operating states were coupled in, the ranking flipped: Guangxi scored 0.0251 while Jiangsu scored 0.0311. The researchers are careful to note that the bootstrap confidence intervals of these two point estimates overlap, so the reversal is suggestive rather than statistically separated. Still, the message is clear: a site that looks worse on rainfall maps may look better once you watch what its turbines actually do in the rain.

Perhaps the most actionable finding concerns how concentrated the damage is in time. The researchers found that the 10 percent most exposed rain intervals carry 75 to 80 percent of the total exposure, and the five most exposed rain events alone account for 78 to 90 percent. Erosion, in other words, is not a steady drip of damage spread evenly across the year. It is dominated by a handful of intense episodes, which means interventions aimed at those episodes can pay outsized dividends.

That insight points directly to a control strategy. If turbines briefly cap their tip speed at 80 meters per second during rain, the model indicates that 42 to 58 percent of the exposure disappears while touching at most 48 percent of rain time. Because the damage is so concentrated, a modest speed reduction during the worst hours yields a large protective effect. The authors stress this result holds within their model framework and before accounting for power losses or control constraints, but it sketches a concrete lever that operators could pull.

The team is also candid about the limits of their indices. Both are normalized to a literature reference coating and serve as screening quantities rather than lifetime predictions, and the threshold relation they use is extrapolated below its 90-to-150-meter-per-second calibration range. Even so, the study supplies something previous assessments had to assume: the actual operational loading term. As wind turbines grow taller and their blades sweep faster, and as operators weigh expensive leading-edge repairs against lost energy production, knowing what the machines really do when the rain arrives may prove to be the difference between a blade that lasts its design life and one that needs early, costly refurbishment.

Subject of Research: Rain-erosion exposure of wind turbine blades measured with synchronized disdrometer and SCADA data

Article Title: Rain-time operation, not rainfall alone, sets wind turbine blade rain-erosion exposure: synchronized disdrometer and SCADA measurements

Article References: Liu, Y., Jia, X., Liu, X., Jin, M., Liu, F., & Li, Y. (2026). Rain-time operation, not rainfall alone, sets wind turbine blade rain-erosion exposure: synchronized disdrometer and SCADA measurements. https://doi.org/10.5194/wes-2026-187

Image Credits: AI Generated

DOI: 10.5194/wes-2026-187

Keywords: wind energy, rain erosion, blade leading edge, tip speed, disdrometer, SCADA, turbine operation, exposure index, erosion control, site assessment, rainfall microphysics, Wind Energy Science