Predictive Multi-Objective Optimization of Water–Energy–Carbon Trade-Offs in a Renewable-Powered Smart Irrigation System Under Meteorological Variability

The increasing demand for food production under water scarcity and climate variability requires intelligent irrigation systems capable of jointly managing water and energy resources. This challenge is particularly relevant in Morocco, where irrigation relies heavily on energy-intensive pumping and where the increasing penetration of photovoltaic generation introduces additional intermittency and operational uncertainty. This study proposes a predictive multi-objective irrigation management framework integrating soil–water dynamics, FAO-56 Penman–Monteith evapotranspiration, precipitation, and renewable-energy availability. A multi-objective particle swarm optimization (MOPSO) strategy is formulated to simultaneously minimize soil-moisture deviation and the combined burden of irrigation-water consumption and CO2 emissions, while satisfying power balance, battery state of charge (SOC) and state of health (SOH), and soil-moisture constraints. The proposed approach is benchmarked against a conventional rule-based controller (RBC) over a four-day horizon under two scenarios: without precipitation and with precipitation. The MOPSO generates a diverse Pareto front, which is subsequently processed through crowding-distance-based diversity preservation, similarity filtering to eliminate redundant solutions, and the extraction of three representative operating profiles—Optimal Irrigation, Environmental, and Balanced—thereby providing decision alternatives without requiring subjective objective weighting. Compared with the RBC, the proposed strategies achieve substantial improvements across both scenarios. Under dry conditions, irrigation-water consumption is reduced by up to 75.52%, total operating cost by 82.45%, CO2 emissions by 67.73%, and mean soil moisture increases by up to 24.72%. When precipitation is available, the benefits are further enhanced for the Environmental and Balanced strategies, with reductions of up to 80.27% in irrigation water, 98.28% in total operating cost, and 75.63% in CO2 emissions, accompanied by an improvement of up to 13.99% in mean soil moisture; the Optimal Irrigation profile, which prioritizes tight soil-moisture tracking, still reduces water use in this scenario but shows a modest 9.7% increase in CO2 emissions relative to the RBC, reflecting its higher irrigation frequency. These results demonstrate that the proposed framework can automatically exploit rainfall availability while coordinating irrigation and renewable-energy resources, without relying on scenario-specific control rules. The stable and diversified Pareto solutions observed across both scenarios further indicate the consistency of the proposed optimization and decision-making strategy under the investigated meteorological variability and renewable-energy intermittency.