AI-FORECASTED TECHNO-ECONOMIC AND ENVIRONMENTAL ASSESSMENT OF BIOGAS, METHANE (CH?), HYDROGEN (H?), AND ELECTRICAL POWER GENERATION AT A DAIRY FARM IN AL-DHLAIL, ZARQA, JORDAN

Jordan?s heavy reliance on imported fossil fuels and its rapidly expanding dairy sector create both an environmentalburden and a largely untapped renewable-energy opportunity. While household-scale and laboratory-scale biogas studies dominate the Middle-Eastern literature, mid-scale fixed-dome plants designed for semi-arid Jordanian conditions have received very littleattention. No published study has yet coupled the techno-economic and environmental assessment of such a plant with an artificial-intelligence (AI) forecasting framework that explicitly resolves biogas into its methane, hydrogen co-production, and electrical-power components. This work addresses that gap. A fixed-dome biogas plant is designed for a 200-cow dairy farm in the Al-Dhlail area of Zarqa Governorate, comprising a total plant volume of 262 m? (170.3 m? digester + 91.7 m? gas holder) and a 32-day hydraulic retention time. Under base-case operation, the plant produces 38.64 m?/day of biogas (60% CH??23.18 m?/day of methane, with a two-stage dark-fermentation H?co-production envelope of about 8.89 m?/day), equivalent to 86,510 kWh/year of thermal energy and, after combined-heat-and-power conversion at 35% electrical efficiency, approximately 30,279 kWh/year of electricity (?3.5 kW continuous). The capital cost of 50,065 JD is offset by annual revenues of 12,365 JD, giving a payback period of ~4 years, a levelized cost of electricity of ~0.066 JD/kWh (? 0.093 USD/kWh) and a negative carbon-abatement cost of approximately 235 JD per ton of CO?avoided, while displacing 28.46 tons of CO?per year (~6 passenger cars or ~470 mature trees). A Long Short-Term Memory (LSTM) recurrent neural network is then used to project plant performance over 2025?2035, forecasting biogas growth to 42.8 m?/day, CH?to 25.7 m?/day, H?potential to 9.8 m?/day, and electrical generation to 33,538 kWh/year by 2035. Scaled to all 250 dairy farms in the Al-Dhlail region, the design would deliver ~21.6 GWh/year (~2.5 MW continuous) more than four times Jordan?s existing UNDP-supported landfill biogas plant and abate ~7,115 t CO?/year, supporting the National Energy Strategy and demonstrating that AI-augmented, multi-product farm-scale biogas is a strategically valuable building block for Jordan?s renewable-energy transition.