Internet of Things-Based Energy Consumption-Aware Framework Design for Smart Grid Environment

Area of Science:

• Electrical Engineering
• Computer Science
• Artificial Intelligence

Background:

• Smart grids leverage Internet of Things (IoT) for enhanced monitoring and control.
• Maintaining grid stability is difficult due to unexpected drops in electricity production from fault-prone units.
• Reactive demand-side control strategies can cause temporary imbalances and operational stress.

Purpose of the Study:

• To propose an Artificial Intelligence of Things (AIoT)-based adaptive energy management framework.
• To enable online adaptive demand-side control through anomaly detection and priority-aware load management.
• To achieve preventive, data-driven intervention for enhanced grid stability.

Main Methods:

• Offline learning of normal production behavior and deviation identification using machine learning.
• Fault modeling for scenario-based training data generation.
• Employing Logistic Regression (LR), Random Forest (RF), and Extreme Gradient Boosting (XGBoost) for anomaly detection and control magnitude estimation.
• Regulating IoT-enabled devices based on priority levels via an MQTT communication infrastructure.

Main Results:

• Early anomaly detection is insufficient without accurate estimation of required demand reduction.
• The proposed AIoT framework effectively manages demand-side control.
• Critical loads are preserved, supporting resilient smart grid operations.

Conclusions:

• The AIoT framework provides a preventive and adaptive approach to smart grid energy management.
• Accurate estimation of demand reduction is crucial for effective intervention.
• The system enhances smart grid resilience by selectively controlling loads and protecting critical services.