Stephanie Ness

The incorporation and integration of advanced metering infrastructure (AMI) and real-time monitoring have improved energy distribution through smart grid, being introduced at fast speeds. Electricity theft is, however, a key issue as it causes the world to lose more than 96 billion dollars annually, with an estimated 40 percent recorded in developing countries occasioned by the illegal meter messing and open connections. The continuously changing fraud tactics have increasingly rendered traditional rule-based and manually driven detection methods ineffective hence the need to implement the intelligent learning, evolving systems. The proposed framework is a hybrid deep learning solution in which several machine learning algorithms can be combined with Long Short-Term Memory (LSTM) networks to identify unusual patterns of energy consumption. The models were trained on time-series manifestations of consumer actions by using data about electricity consumption in January, 2014 through October, 2016. The best accuracy was produced by KNN+LSTM architecture 81.32, whereas the RF+LSTM one had a rather low accuracy level of 50.34. Both XGB+LSTM and LR+LSTM gave similar outcomes with the accuracy of 73.16% and 73.78%, respectively. These results prove that optimized feature extraction along with temporal modelling is able to improve fraud detection. The suggested hybrid system is a flexible and auto mechanism of resolving issues on grid protection and minimizing energy losses in superior power distribution systems.

Accurate assessment of the Remaining Useful Life (RUL) is critical for preventing catastrophic equipment failures, but the machine learning models used for these predictions are highly vulnerable to adversarial attacks that can corrupt sensor data and produce dangerously incorrect outputs. To address this critical security gap, this paper introduces a novel active learning framework designed to bolster the adversarial robustness of RUL models. By strategically selecting the most informative and uncertain data points for labeling, our method enhances defensive capabilities while significantly reducing the reliance on large, pre-labeled datasets. Experiments conducted on the NASA C-MAPSS FD001 dataset against CW, FGSM, and JSMA attacks demonstrate that this approach yields significant gains. Specifically, the Random Forest model paired with active learning achieves a superior accuracy of 89%. This robust performance is achieved while requiring 42% fewer labeled instances compared to a traditional supervised approach, at the cost of a 28% increase in iterative training time. This work provides a practical pathway toward developing more secure, efficient, and resilient predictive maintenance systems for safety-critical applications.

Stable operations of modern power networks depend on the reliability of smart grids. The expansion of modern power systems combined with renewable energy applications and decentralized networks creates difficulties for traditional fault detection methods because of high-dimensional data combined with measurement noise and delayed fault identification processes. A deep learning solution based on Phasor Measurement Unit (PMU) image data detection for smart grid faults is presented in this work. The proposed method achieves accurate and efficient fault classification through the conversion of PMU data to images and then utilization of advanced deep learning models VGG16 combined with Convolutional Neural Networks (CNNs). The experimental findings show that VGG16 reaches 98.75% accuracy in fault classification while surpassing CNNs that delivered 94.44% accuracy. In addition to reporting high classification accuracy, the study proposes a new hybrid deep learning framework that combines transfer learning and adaptive thresholding techniques. The image-based deep learning system extracts features automatically, reducing the need for human preprocessing to effectively detect fault patterns. The methodology enables immediate fault detection, allowing operators to make critical decisions to avoid major power outages. According to the research, artificial intelligence has the potential to significantly improve smart grid resilience, reduce economic losses, and improve power system operational efficiency. The research advances AI developments in energy management by developing advanced, powerful fault detection methods that will guide future changes to power system networks.

Continuous operations in extreme environments require a reliable energy supply for autonomous systems. The proposed research presents a novel AI-driven system that averts power system blackouts by utilizing autonomous robots trained in artificial intelligence. The robots monitor power system signals in real-time to detect anomalies while employing their intelligence to select optimal distribution methods. This research combined multiple machine learning models utilizing power system attack data from Mississippi State University and Oak Ridge National Laboratory to ascertain their efficacy in anomaly detection and enhancing energy resilience. In power system threat detection, the ExtraTree Classifier and Random Forest Classifier demonstrate optimal accuracy at 95%, confirming their ability to ensure energy stability. In addition to XGBoost (88%), Bagging (91%), and Decision Tree (86%), the ExtraTree Classifier demonstrates superior performance; however, Gradient Boosting (77%) and Logistic Regression (63%) need improvement to ensure reliable performance in critical applications. The proposed method implements XAI techniques to make the system explainable while improving decision-making reliability. Future research will build on this work by augmenting data availability, integrating renewable energy systems into the research framework, and enhancing self-healing capabilities and robotic mobility to further improve power grid resilience. The developed findings on AI-driven energy automation of this mathematical applicability exploration created scalable intelligent solutions to tackle power stability challenges in unpredictable environmental conditions.

Anomaly detection in network traffic is a critical aspect of network security, particularly in defending against the increasing sophistication of cyber threats. This study investigates the application of various machine learning models for detecting anomalies in network traffic, specifically focusing on their effectiveness in addressing challenges such as class imbalance and feature complexity. The models assessed include Isolation Forest, Naive Bayes, XGBoost, LightGBM, and SVM classification. Through comprehensive evaluation, this research explores both supervised and unsupervised approaches, comparing their performance across key metrics like accuracy, F1-score, and recall. The results reveal that while models like XGBoost and LightGBM exhibit impressive performance, with LightGBM achieving near-perfect training accuracy (1.0) and solid test accuracy (0.85), others like Isolation Forest show limitations with low accuracy. The study highlights the strengths and weaknesses of each model, providing valuable insights into their practical application for network anomaly detection. By comparing different algorithms, this research contributes to advancing the application of machine learning in network security, offering guidance on model selection and optimization for improved detection of cyber threats.

Smart grids themselves have emerged as vital structures of the up-to-date practical power systems or electricity networks that incorporate high technologies and information handling. Yet, they are more susceptible to an adversarial attack that can interfere with the critical functions like energy distribution and faults detection. This paper therefore proposes a new alternative to developing a DL and ML framework for identifying adversarial attacks on smart grids. After analyses of the performances of Logistic Regression, Perceptron, Gaussian Naive Bayes and Multi-Layer Perceptron, LSTM network has better results with an accuracy of 99.81%. The suggested framework strengthens smart grid immunity to cyber threats such as DoS attacks, back door injections, and adversarial perturbations while increasing energy distribution stability and security. For enhancing smart grid security, our results emphasize the importance of integration of ML and DL techniques and provide such an understanding of threat environment for future research and development on threat identification.