Establishment of predictive models for head loss and filtration efficiency of pontoon mesh rotary filter based on dimension analysis.

Filters serve as core equipment to ensure the normal operation of micro-irrigation systems, with head loss and filtration efficiency serving as the two key indicators for evaluating performance. In this study, we used a pre-pump filter-pontoon mesh rotary filter as the research object and conducted physical model tests under the flow rate (798-1050 L h), sand content (0.5-2.5 g L), and aperture of the filter screen (0.125-0.180 mm). We then adopted range analysis, variance analysis (ANOVA), dimensional analysis, and the multiple linear regression (MLR) method to analyze the results. The results showed that the order of factors affecting the head loss of the assessment indices, from large to small, was as follows: flow rate, sand content, and aperture of the filter screen. The order of factors affecting the filtration efficiency of the assessment indices from large to small was as follows: sand content, flow rate, and aperture of the filter screen. Predictive models for head loss and filtration efficiency were developed, with coefficients of determination R of 0.969 and 0.954, and root mean square error (RMSE) values of 0.1041 and 0.0183. The model exhibited high accuracy and could be used to predict the head loss and filtration efficiency of the pontoon mesh rotary filter. In the test range, the optimal working condition of this filter was a flow rate of 930 L h, sand content of 2.0 g L, and 0.150 mm aperture of the filter screen. In addition, the head loss under this condition was 0.281 m, and the filtration efficiency was 84.01%. These results could serve as a reference for the further optimization and application of the pontoon mesh rotary filter, while also enriching the hydraulic performance and filtration performance of the pre-pump filter.