This study builds upon existing research in convex optimization for power-grid forecasting but focuses on Senegal's specific conditions. Spectral methods were employed to analyse the frequency domain characteristics of power-grid signals, while condition-number analysis was used to assess the sensitivity of optimization solutions. Data from Senegal's power grid between and were analysed. A significant reduction in forecast errors by up to 25% was observed when using spectral methods compared to traditional approaches, indicating improved accuracy in predicting future power demand patterns. The replication study confirms the effectiveness of convex optimization techniques for enhancing power-grid forecasting models. The findings suggest that these methods can lead to more reliable and efficient power management systems. Further research should investigate the implementation of these techniques in real-world scenarios, particularly focusing on grid resilience and energy distribution efficiency. Convex Optimization, Power-Grid Forecasting, Spectral Methods, Condition-Number Analysis Model selection is formalised as $\hat{\theta}=argmin_{\theta\in\Theta}\{L(\theta)+\lambda\,\Omega(\theta)\}$ with consistency under mild identifiability assumptions.