Optimization of sample size and location for evaluating the hydraulic performance of microirrigation subunits using simulation and genetic algorithms

Abstract

Rapid and accurate field evaluation of hydraulic performance is critical for the operation of a microirrigation system. However, the optimal sample size and the specific locations of the emitters selected in one subunit for field tests have not been determined. A model (Hydraulic Analysis of Pressurized Irrigation System,HAPIS) was constructed for hydraulic analysis of a pressurized irrigation system by coupling MATLAB and EPANET. The random sampling method (RSM) and uniform sampling method (USM) were optimized for emitters selected through simulation, aiming to achieve higher estimation accuracies of the mean emitter discharge rate of the subunit ( $\overline{q})$ and the Christiansen uniformity coefficient (CU) while minimizing the number of emitters tested. In addition, a linear sampling method at predetermined emitter locations (LSMPE) was developed to simplify the evaluation process using a genetic algorithm (GA). The results indicate that the appropriate sample size range for RSM was 20–40, in which the maximum percentage difference between $\overline{q}$ and CU was maintained at ±10%. For the USM, a sample size of approximately 18 can provide relatively accurate estimations of $\overline{q}$ and CU, while it is recommended that the sampled emitters be distributed over three to five laterals. The optimal sample size of LSMPE could be decreased to approximately 10, and the selected emitters were arranged along the sampling line. The absolute relative estimation error of $\overline{q}$ and CU could be maintained at <1%.