A Stochastic Genetic-Algorithm–Monte Carlo framework for robust Type-II distributed generation planning in radial networks

Seasonal demand variability and the intermittent nature of distributed generation (DG) intensify voltage regulation challenges in long radial distribution networks. Conventional DG planning approaches often rely on deterministic or average operating conditions, leading to DG sizes that may be ineffective under adverse seasonal scenarios. This study addresses this limitation by developing a risk-aware framework for optimal sizing and placement of Type-II DGs, where load demand uncertainty and DG availability are explicitly incorporated. The proposed Stochastic Genetic Algorithm–Monte Carlo Simulation (SGA–MCS) model integrates Monte Carlo-based scenario generation with genetic optimization to evaluate candidate DG configurations across seasonal-stochastic operating conditions. Unlike conventional probabilistic or PPF-based approaches, where uncertainty is mainly used for post-optimization performance evaluation, this study embeds uncertainty directly within the optimization process. Distributional metrics, including expected voltage, lower-tail quantiles, voltage compliance probability, CVaR, and kurtosis, are used to characterize voltage performance and determine a true risk-conditioned DG size. The framework is validated on a 56-bus radial feeder representative of weak distribution networks. Results show that SGA–MCS outperforms benchmark algorithms, achieving substantial voltage improvement, more than 95% voltage compliance across stochastic scenarios, and the lowest tail-risk. Localized Volt–VAR/SVC support further improves voltage robustness at critical buses. The findings confirm that DG sizing should be conditioned on uncertainty rather than nominal operation alone. The proposed framework therefore provides a practical, risk-based planning tool for reliable DG deployment with reduced need for operational intervention. This paper: