Modeling Water Distribution Uniformity of Medium-Sized Sprinklers Using Artificial Neural Networks

Artificial neural network-(ANN) simulation models have been successfully applied in various fields, including agricultural sciences. This study evaluated the applicability of an artificial neural network for predicting water distribution uniformity in medium-sized sprinklers under varying wind conditions. Using data from 74 field tests, the ANN was developed and trained through supervised learning to optimize its predictions. The water distribution patterns simulated by the ANN closely mirrored observed field data, with 50% of the test set classified as ‘optimal’ in the highest reliability category. Statistical analyses indicated that the ANN performance was not significantly influenced by operating pressure, wind speed, or direction, underscoring its reliability across diverse scenarios. Furthermore, the ANN performed robustly in estimating the Christiansen uniformity coefficient, with statistical indices demonstrating excellent performance (r = 0.9475, d = 0.9689, and c = 0.9181), classifying the results as ‘optimal’. Linear regression analysis further confirmed the model’s robustness, with a slope close to 1 (0.988 ± 0.053) and a mean absolute deviation of 4.33%, indicating high accuracy in the simulations. These findings suggest that the trained ANN provides an accurate and efficient alternative to extensive fields.