Development of a Genetic Algorithm-Based Framework for Automated Load Demand Response and Smart Grid Optimization in 33kV Distribution Network

Abstract

imbalances that necessitate manual load shedding practices often
characterized by inefficiency and inequitable allocation. This study develops
a Genetic Algorithm (GA)-based Automated Load Demand Response
(ALDR) framework for optimizing power allocation on a 33 kV distribution
Optimization Toolbox for optimization, MATLAB App Designer for real-
time operator interaction, and an ETAP-in-the-loop feasibility module to
ensure technical operability. The GA model incorporates loss-aware power
balancing, socio-economic priority indices, critical-load floors, and a
fairness-enhancing grid-Gini metric. These enable the system to guarantee
minimum supply for essential services such as hospitals, waterworks, higher
institutions, and security installations, while equitably distributing residual
capacity among non-critical loads. An adaptive repair operator maintains
allocation feasibility under ETAP-estimated losses, voltage constraints,
thermal loading, and radial topology requirements. The framework was
evaluated across multiple shortage scenarios (20, 15, and 10 Mega Watts,
Particle Swarm Optimization (PSO), Simulated Annealing (SA), ANN-
assisted GA, and heuristic allocation approaches. Results show that the
proposed GA framework improves the Satisfaction Index by 15 25%,
reduces Energy Not Supplied (ENS) by 30 40%, and enhances reliability
indices (SAIFI, SAIDI) by 10 20% compared to existing methods. ETAP
simulations further validated allocation feasibility and maintained technical
-driven prioritization, fairness shaping, multi-algorithm benchmarking, physical-
feasibility validation, and an operator-oriented Graphic User Interface, GUI.
The resulting ALDR architecture improves fairness, transparency, and
resilience in 33kV distribution networks. Future work will explore hybrid
GA PSO/ANN models, renewable energy coupling, and real-time
deployment for feeder automation.