Search

Accurately estimating evaporation is essential for water managers to formulate effective rules and policies. The complexity arising from intricate interactions within the soil-atmosphere system makes evaporation a challenging parameter to predict. In the past decade, Machine Learning techniques rooted in soft computing have emerged as potent tools for addressing the intricacies and non-linearities in hydrology. Surprisingly, there has been no prior research exploring the impact of reservoir water temperature on evaporation modeling. Consequently, this study aimed to employ a hybridized Adaptive Neuro-Fuzzy Inference System (ANFIS) integrated with four optimization algorithms: Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Harris Hawks Optimization (HHO), and Salp Swarm Algorithm (SSA). The focus was on modeling the monthly evaporation of the Boukourdane Dam in Algeria for the period of September 1996 to August 2016 and examining how the reservoir's water temperature influences model performance based on four performance indicators, Root Mean Square Error (RMSE), Mean Absolute Error (MAE), scatter index (SI) and Correlation Coefficient (R). The findings underscored the significance of incorporating the water temperature parameter. Among the models, ANFIS-HHO demonstrated the highest accuracy with MAE, RMSE and R values of 1.28, 1.21 mm and 0.92, respectively during test period. Moreover, results revealing a notable impact of reservoir water temperature on evaporation forecasting and the addition of this parameter provide an increase in R and decrease in RMSE about 4.54% and 17.98% respectively. [ABSTRACT FROM AUTHOR]

• An autonomous water pumping system (WPS) is suggested. • A hybrid energy system is optimized through seagull optimization algorithm (SOA). • The impacts of load management (LM) strategy in system sizing is analyzed. • Techno-economic evaluation in the presence and absence of load management is made. This study introduces an innovative irrigation system tailored for a tomato field in Algeria, optimizing its elements using Seagull Optimization Algorithm (SOA) over a 20-year lifespan. The system integrates photovoltaic (PV), wind turbine (WT), diesel generator (DG), battery, and tank storage for water pumping. The novelty lies in the incorporation of Load Management (LM), creating a dynamic water demand profile across four tomato development stages. Without LM, the optimal configuration yields a Net Total Present Cost (NTPC) of $14,140, corresponding to a Cost of Energy (COE) of 0.292 $/kWh. This configuration ensures a 24% Loss of Power Supply Probability (LPSP), a 41.2 % Loss of Water Supply Probability (LWSP), and an emission factor of 0.169 kg/kWh during 279 h of DG operation annually. With LM, the COE drops to $0.298/kWh, reducing the NTPC to $12,080. This LM strategy gives an 18 % LPSP, 20 % LWSP, and an Energy Fraction (EF) of 0.140 kg/kWh during 168 h of DG operation annually. The present application is a promising and cost-effective solution for optimizing Water Pumping Systems (WPS) in agricultural settings. [ABSTRACT FROM AUTHOR]