International Journal of Engineering Research and Science & Technology
Power Theft Detection in Smart Grids Using Hybrid Quantum Machine Learning with Variational Quantum Circuits, Data Reuploading, and Entanglement Layers
Chegudi Vijayababu, Dr. Krishna Mohan Ankala, Shaik Lateefa
Thefts of electricity are a plague on smarter grids and dispersed generation of energy from PV systems, because smart meters could be tampered with. The identification of theft at the DG side by utilising typical ML algorithms is hard to achieve and is not appropriate for high-dimensional imbalanced data. To solve these problems, a hybrid Quantum DL model is designed using the QVC and DRC method and applied by a QML technique. The QVC consists of an angle-embedding layer to encode the classical neural outputs into the qubits and a set of trainable parameters of rotational and CNOT gates for the noise reduction on the NISQ devices. If there is still noise, the DRC layer will re-upload the data iteratively to compensate for it. The data set 'Smart Grid Theft Detection' is retrieved from Kaggle and the features in it are divided into 80% training and 20% testing sets. The FH-QVC-DRC model has an accuracy of 88% which is higher than standard algorithms like the XGBOOST and LIGHTGBM with 86% and 73% accuracy, respectively. Furthermore, the Entanglement quantum layer is proposed to boost the performance by using the quantum parallelism, quantum interference and secure state correlations. The developed model is called FH-QVC-DRC-ENT with an accuracy of 91 %. The visualisation, the performance indicators and the prediction output demonstrate that this quantum model, which uses entanglement, is able to better classify events of theft and non-theft, than the other models.
Thefts of electricity are a plague on smarter grids and dispersed generation of energy from PV systems, because smart meters could be tampered with. The identification of theft at the DG side by utilising typical ML algorithms is hard to achieve and is not appropriate for high-dimensional imbalanced data. To solve these problems, a hybrid Quantum DL model is designed using the QVC and DRC method and applied by a QML technique. The QVC consists of an angle-embedding layer to encode the classical neural outputs into the qubits and a set of trainable parameters of rotational and CNOT gates for the noise reduction on the NISQ devices. If there is still noise, the DRC layer will re-upload the data iteratively to compensate for it. The data set 'Smart Grid Theft Detection' is retrieved from Kaggle and the features in it are divided into 80% training and 20% testing sets. The FH-QVC-DRC model has an accuracy of 88% which is higher than standard algorithms like the XGBOOST and LIGHTGBM with 86% and 73% accuracy, respectively. Furthermore, the Entanglement quantum layer is proposed to boost the performance by using the quantum parallelism, quantum interference and secure state correlations. The developed model is called FH-QVC-DRC-ENT with an accuracy of 91 %. The visualisation, the performance indicators and the prediction output demonstrate that this quantum model, which uses entanglement, is able to better classify events of theft and non-theft, than the other models.
Electricity theft could result in massive economic loss and cause potential security problems to the grids. There-fore, it is of great importance to detect electricity theft from both an economic and social perspective. With the help of Advanced Metering Infrastructure (AMI) in smart grids, now it is possible to detect theft activities through analyzing the recorded electricity consuming data using advanced techniques, such as machine learning. However, training a learning model, especially for the ones with a huge amount of parameters, is always resource-consuming, and users could encounter performance bottlenecks due to hardware limitations. To improve the problem, using quantum computing for machine learning tasks has shown to be an effective approach to speed up the learning process. However, the relevant techniques have not been studied in the domain of energy yet. As a pioneering exploration, in this work, we are trying to leverage a quantum classification algorithm to classify the electricity dataset and on that basis to perform a case study on theft activity detection. Specifically, we have given the details of our design in the study, and our experimental results show that the prediction accuracy can be significantly improved with increasing training epochs, demonstrating the feasibility of quantum machine learning for energy data analysis.
Smart grid stability prediction is crucial for providing uninterrupted power and high maintainability while operating smart grids. State-of-the-art research focuses on using Machine Learning (ML) to solve smart grid stability prediction; however, the complex nature of smart grids that serve thousands of users poses challenges for classical ML. Quantum Machine Learning (QML) has emerged as a possible solution to overcome these challenges, making QML a promising approach for energy applications. While QML shows promise in energy research, its application to grid stability problems remains unexplored. In this study, we propose a regularization-enhanced hybrid quantum-classical neural network that combines a classical encoder-decoder architecture with a Variational Quantum Circuit (VQC). Our model achieves 92.25 % accuracy, competitive with parameter-matched classical neural networks while outperforming Gradient Boosting and Random Forest by3.0−4.25%. Noise robustness analysis shows that the proposed hybrid model demonstrates a slight advantage under dropout noise while maintaining comparable performance under Gaussian and uniform noise, outperforming tree-based methods across all noise types.
Non-Intrusive Load Monitoring (NILM) is an essential component of modern smart grid systems, enabling the disaggregation of total power consumption into appliance-level usage. Accurate appliance classification is critical for optimal power flow and uninterrupted grid operation. In state-of-the-art NILM applications, classical machine learning (ML) models require abundant labeled data. However, in real-world scenarios, limited data and measurement noise are common challenges. Quantum machine learning (QML), which leverages quantum mechanical properties, may offer advantages for NILM compared to classical ML. In this study, we propose QLID-Net (Quantum Load IDentification Network), a hybrid quantum-classical neural network for appliance classification in NILM. The proposed architecture consists of a classical encoder, a variational quantum circuit (VQC), and a classical decoder.We evaluated the model on a smart meter dataset (five appliance classes) across four experiments: data efficiency, classical/data noise robustness, baseline comparison, and quantum hardware noise robustness comparing it against XG-Boost, TabPFN, Random Forest, and a parameter-matched classical neural network. In the data efficiency experiment, the proposed model outperformed two classical models at 10 or fewer samples per class and significantly outperformed the parameter-matched classical neural network (p<0.001, Cohen’s d=1.80), with a crossover point at approximately 20 samples per class. Under additive noise (Gaussian, Uniform), the proposed model showed superior robustness, outperforming all classical baselines with up to 17.9% improvement over XGBoost and up to 21.3% improvement over TabPFN. In the baseline comparison, it achieved competitive performance with XGBoost and TabPFN (within 5.1%) and 100% training convergence compared to 70% for the classical neural network. Finally, in the quantum hardware noise tests, results showed that although QLID-Net can compete with the parameter matched classical neural network, it faces a performance decline. Our results identify practical regimes where quantum advantages are beneficial, especially for scenarios with limited data or measurement noise. This study contributes to the growing evidence of practical QML applications and paves the way for quantum-assisted hybrid models in NILM.
Weaknesses arising from heterogeneous device structure, high-dimensional traffic patterns, and more complex nonlinear attack behaviors. The traditional intrusion detection systems (IDS), which rely on the classical machine learning models, often face the challenges of scalability, feature explosion, and low representational ability when handling large-scale IoT traffic data. To meet these challenges, we propose a Hybrid Quantum-Classical Variational Neural Network (HQCVNN) for effective intrusion detection in IoT systems. The model utilizes a Variational Quantum Circuit (VQC) as a nonlinear quantum feature map followed by dense neural layers for overall classification. Angle embedding with entanglement layers enables the encoding of quantum data and leads to a better separability of features in high-dimensional space, whilst the classical optimization guarantees efficient training on NISQ time scales. The model is implemented using Qiskit and evaluated on classical benchmark intrusion detection datasets. Experimental results show the proposed method has better classification accuracy, F1-score, and complexity reduction than conventional ANNs and SVMs. The results demonstrate the promise of hybrid quantum machine learning systems with end-to-end general-purpose capabilities in future IoT security and provide pragmatic impressions on how to deploy them on near-term quantum processors.
Smart grids form the backbone of modern energy infrastructures. They require accurate power prediction to ensure efficient energy management and maintain grid stability. Recent advancements in quantum computing have provided motivation to explore quantum machine learning models for data driven prediction task. In this study, a quantum computing-based power prediction model is investigated using a Variational Quantum Circuit (VQC) to estimate the AC power output in a solar-based smart grid system. The proposed model is studied using a simulation software created using Qiskit framework and real-world solar dataset is used as input. To enable a holistic comparison, the classical regression models, including Linear Regression, Ridge Regression, Random Forest, and XGBoost, are used as baseline models. Experimental findings indicate that classical models have lower prediction errors and much faster execution times. Conversely, the quantum model can only capture the broad patterns of the data because of the constraints that relate to Noisy Intermediate-Scale Quantum (NISQ) devices such as circuit noise, limited qubit count, and optimization limitations. Furthermore, a multi-criteria evaluation is conducted to compare the classical and quantum approaches based on prediction accuracy, execution time, model complexity, and scalability. The findings provide valuable insights into the feasibility of applying quantum computing techniques to smart grid prediction tasks and highlight both the potential and the limitations of quantum machine learning for real-world energy systems.
Hybrid quantum machine learning (QML) algorithms have potentials for current quantum computing technologies since only part of the model is computed by a quantum device. This paper proposes a novel hybrid quantum-classical deep learning model for cybersecurity application: domain generation algorithms (DGA)-based botnet detection. We analyzed our hybrid model’s performance compared with the classical model counterpart to investigate the quantum circuit’s effectivity as a layer in a deep learning model. We used four features of the Botnet DGA dataset: MinREBotnets, CharLength, TreeNewFeature, and nGramReputation_Alexa. The hybrid model’s quantum circuit is a combination of Pennylane’s embedding (Angle Embedding and IQP Embedding) and layers circuit (Basic Entangler Layers, Random Layers, and Strongly Entangling Layers). Also, we implemented noise models for assessing the applicability of the model for current Noisy Intermediate Scale Quantum (NISQ) technology. We found that in some cases, the hybrid model reached high performance (maximum accuracy up to 94,7% using n=100; 93,9% using n=1,000). We discovered that the combination of Angle Embedding and Strongly Entangled delivers high accuracy, superior to the classical deep learning model in n=100 experiments). However, the overall performance is still inferior for the rest cases compared to the classical deep learning model counterpart.