Energy Resilience in Water Treatment Infrastructure under Seasonal Hydrological Stress: Neural Network-controlled Solar Generation Integration in Southern Nigeria
Nwandu Charles Arnold
Nigerian Building and Road Research Institute, Km 10, Idiroko Road, Ota, Ogun State, Nigeria.
ThankGod Sylvanus Ntem
*
Apple Networks Nigeria Limited, Nigeria.
O. Nwafor Christiana
Nigerian Building and Road Research Institute, Km 10, Idiroko Road, Ota, Ogun State, Nigeria.
Olumide Omojola
Nigerian Building and Road Research Institute, Km 10, Idiroko Road, Ota, Ogun State, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Aims: This study quantifies the impact of seasonal hydrological variability on energy demand at the Cross River State Water Board (CRSWB) treatment facility in Calabar, Nigeria, and evaluates the effectiveness of neural network-controlled distributed solar generation (DSG) in enhancing grid resilience during turbidity-driven peak demand periods.
Study Design: A quantitative, model-based case study using historical water quality and load profile data with MATLAB/Simulink and ETAP simulation over a three-year horizon (2013–2015).
Place and Duration of Study: Cross River State Waterboard water treatment facility, Calabar, Nigeria; data collection 2013–2015, simulation conducted in 2024.
Methodology: Turbidity records (NTU), hourly load profiles (kW), and plant operational data were obtained from CRSWB. Seasonal partitioning followed Nigerian hydrological cycles: dry season (November–March) and rainy season (April–October). Linear regression modelled the turbidity–energy relationship. A 33 kV network diagram in ETAP was reconfigured in MATLAB/Simulink to integrate a 217.12 kW solar PV unit. A Bayesian-regularized neural network (R = 0.991) automated DSG dispatch during peak hours (09:00–18:00). Validation employed two-sample and paired t-tests.
Results: Dry-season turbidity peaked at 194 NTU versus 17 NTU in the rainy season. This 4.2-fold increase was associated with 23.5% higher daily energy demand (24,542 kW vs. 19,876 kW) and 45% greater load variability (σ = 142 kW vs. 98 kW). DSG integration reduced daily energy consumption by 9.1%, lowered peak demand by 30% (680 kW to 476 kW), and decreased load variability by 37% (σ = 89 kW). All differences were statistically significant (P < .001).
Conclusion: Seasonal turbidity is a primary driver of energy demand instability in tropical water treatment infrastructure. Neural network-controlled solar DSG substantially mitigates peak demand and load variability, offering a replicable, evidence-based blueprint for energy-resilient water utilities across sub-Saharan Africa.
Keywords: Energy resilience, seasonal turbidity, distributed solar generation, neural network control, Cross River State