Profitable and reliable EV charging infrastructure: A time-series power flow model for improved voltage and power stability

Efficient and sustainable planning of Electric Vehicle (EV) charging infrastructure requires balancing technical, economic, and environmental factors. Existing EV Charging Station (EVCS) designs often overlook user convenience and grid reliability, while failing to account for uncertainties, which can lead to inefficiencies and suboptimal system performance. This study presents an intelligent approach for profitable and reliable EV charging infrastructure using a Time-Series Power Flow (TSPF) model to enhance voltage and power stability. The proposed method integrates the Opposition-Based Botox Optimization Algorithm (OBOA) and Spatio-Temporal Field Neural Network (STFNN), referred to as the OBOA-STFNN technique. The BOA optimizes the siting and sizing of EVCSs to balance operator profit, grid stability, and user convenience, while STFNN predicts individual EV charging behavior and station demand. The effectiveness of the technique is evaluated in MATLAB and compared with Particle Swarm Optimization (PSO), Modified Snake Optimization (MSO), and Convolutional Neural Network (CNN) approaches. Simulation results demonstrate that the OBOA-STFNN method significantly reduces energy consumption to 38.74 MWh and energy loss to 418 MWh, while achieving a lower optimal cost, mean, and standard deviation, along with reduced total computation time. These results highlight the superior efficiency, reliability, and practicality of the proposed approach for EVCS planning and operation. • OBOA-STFNN optimizes EVCS siting for profit, grid stability, and convenience. • STFNN predicts EV charging patterns to improve station demand forecasting. • Time-Series Power Flow model enhances voltage and power stability in grids. • Method outperforms MSO, PSO, and CNN in cost, energy loss, and computation.