Experimental investigation of the effect of intermittent operation on membranes in wind-powered SWRO plants, focusing on frequent start-stop scenarios

• Simulated frequent start-stops of a wind-powered SWRO desalination plant. • Analyzed the evolution of four normalized performance parameters. • Conducted membrane autopsies on the first and last system membranes. • Performed characterization, integrity tests, and fouling analysis. Given the temporal variability of wind, the intermittent operation of off-grid wind-powered seawater reverse osmosis (SWRO) desalination plants has been proposed. However, it has also been suggested that intermittency may damage plant membranes. This paper presents the results and analysis of testing a 100 m 3 /day rated capacity SWRO plant (two pressure vessels, 1:1 arrangement with 5 and 2 membrane elements [Hydranautics 6-SWC4-LD] per pressure vessel). For over 4 months (totaling 3,136 h), the SWRO plant was operated intermittently—1 h on and 1 h off—24 h each day, resulting in 1,568 start-stop cycles. The evolution of various parameters was analyzed using machine learning techniques, hypothesis testing, and effect size analysis. At the end of the test period, membrane autopsies were conducted on the first and last membranes of the system. The results show that cumulative intermittent operating hours significantly influenced the normalized permeate flow (NPF increased slightly by 0.0065 m 3 /h), normalized salt rejection (NSR increased by 0.12 %), normalized salt permeate concentration (NSPC decreased by 4.5 mg/L), and normalized pressure differential (NPD showed no significant change). Statistical analysis indicated that the statistically significant differences between the means of these normalized parameters at the start and end of the test are primarily the result of changes in operating parameters—such as feed water temperature (decreased by 1.2 °C) and conductivity (decreased by 527 mg/L)—rather than membrane degradation. According to the autopsies, only slight fouling was detected and no telescoping effect was observed. The integrity tests were negative, except for the methylene blue test, which detected some abrasion marks on the polyamide layer due to displacement of the feed channel spacers. These results provide useful information about the behavior of wind-powered SWRO systems with a high number of starts and stops, suggesting that frequent intermittent operation does not significantly damage SWRO membranes over the tested period.