A Novel Hybrid Approach for Hydrothermal Scheduling Using Mathematical and Metaheuristic Search Methods

Hydrothermal scheduling is a significant concern in the field of power system economics that seeks to reduce the overall cost of generation by optimizing the hourly output of generators. However, this scheduling process suffers from a non-linear and complex problem due to set of uncertainty constraints from hydro and thermal units. Hence
$\vartheta $
-constraint method has been proposed in which ramp constraints are considered that supply a feasible power region of the thermal units and minimize the uncertainty constraints. In addition, the existing solution techniques of hydro and thermal unit, commitment problems have not considered thermal constraints and their losses. This results in the frequent swapping of hydro units which decreases the energy conversion efficiency by a percentage at each repetition and increases the cost of the system. Therefore, a novel Chaotic geometric programming and chaotic approximation approach has been proposed that balance water discharge level without frequency swapping based on Armijo’s rule and Hazen Williams’s rule. Further, dynamic indexing is performed to increase the energy conversion efficiency. These balanced values are given to multiple-wave neural networks, to solve the hydro and thermal unit commitment problems. Furthermore, the cost of the system is minimized using Enhanced Ebola Optimization Search (EEOS) Algorithm in which the Srate parameter has been modified, thereby normalizing the hydrothermal scheduling problem for matured convergence. The proposed mathematical model has been implemented in the MATLAB R2022b on a 2.10 GHz, AMD Ryzen 5 3500U with Raden Vega Mobile Gfx processor, 8 GB RAM, 64 bit personal computer, and the results obtained show better performance than the previous approaches in terms of mean, median and standard deviation.