In Production Engineering, there is an area of research on reliability, maintenance, and risks in Production Systems. In this area, the objectives include maintenance policies aimed at predicting component failures and consequently minimizing unexpected downtimes of complex systems. Reliability engineering has successfully utilized machine learning to predict and categorize equipment and machine states. In this thesis, we aim to compare different machine learning algorithms and quantum machine learning in the context of reliability engineering. Specifically, we compare 8 models created with quantum machine learning, using three different bearing datasets: Case Western Reserve University, Machinery Failure Prevention Technology, and Paderborn University. These datasets mainly consist of accelerometer vibration data from the bearings, which must be processed to fit the different quantum machine learning algorithms. We focus on fault detection in rotating equipment, using quantum computing techniques to analyze sensor data installed on the equipment. Our approach considers the healthy state, inner ring faults, and outer ring faults of the bearings, with the aim of achieving fault detection. We compare established circuit designs, such as Real Amplitudes and Quantum Convolutional Circuits, as well as hybrid models that combine quantum circuits and neural networks using the library of TensorFlow with Cirq in python. We obtain results on classical computers using analytical calculations that simulate a quantum computer without experiencing the "quantum noise" delivered by the hardware. The most effective way to input data's key features is through a ZFeatureMap circuit, which assigns a qubit to each extracted feature from the data. Quantum convolutional circuits yield better results than other parameterized circuits, while hybrid models provide higher accuracy than their counterparts that do not utilize neural networks. The achieved accuracy rates on the training dataset are around 96%, suggesting that the parameterized quantum circuits used in this master's thesis yield stable results.
In Production Engineering, there is an area of research on reliability, maintenance, and risks in Production Systems. In this area, the objectives include maintenance policies aimed at predicting component failures and consequently minimizing unexpected downtimes of complex systems. Reliability engineering has successfully utilized machine learning to predict and categorize equipment and machine states. In this thesis, we aim to compare different machine learning algorithms and quantum machine learning in the context of reliability engineering. Specifically, we compare 8 models created with quantum machine learning, using three different bearing datasets: Case Western Reserve University, Machinery Failure Prevention Technology, and Paderborn University. These datasets mainly consist of accelerometer vibration data from the bearings, which must be processed to fit the different quantum machine learning algorithms. We focus on fault detection in rotating equipment, using quantum computing techniques to analyze sensor data installed on the equipment. Our approach considers the healthy state, inner ring faults, and outer ring faults of the bearings, with the aim of achieving fault detection. We compare established circuit designs, such as Real Amplitudes and Quantum Convolutional Circuits, as well as hybrid models that combine quantum circuits and neural networks using the library of TensorFlow with Cirq in python. We obtain results on classical computers using analytical calculations that simulate a quantum computer without experiencing the "quantum noise" delivered by the hardware. The most effective way to input data's key features is through a ZFeatureMap circuit, which assigns a qubit to each extracted feature from the data. Quantum convolutional circuits yield better results than other parameterized circuits, while hybrid models provide higher accuracy than their counterparts that do not utilize neural networks. The achieved accuracy rates on the training dataset are around 96%, suggesting that the parameterized quantum circuits used in this master's thesis yield stable results.
Bearing fault diagnostics is essential for ensuring operational effectiveness and preventing failures across various industries, which often result in substantial financial losses. The development of quantum machine learning has introduced a paradigm shift in defect diagnostic methods, leveraging the computational power of quantum computers. This study investigates the impact of quantum computing and classical computing algorithms on Case Western Reserve University bearing dataset. The proposed methodology encompasses normalization, segmentation, data augmentation, and feature extraction/selection, executed on classical computer. Selected classical features are encoded as quantum state amplitudes, followed by the application of superposition and entanglement to qubits. The classification stage employs a quantum support vector machine, implemented, and simulated using the available IBM quantum resources. The results showcase 99.61% training accuracy and 99.14% testing accuracy with overall F1-score, recall, and precision of 99%. Comparative analysis between classical and quantum machine learning algorithms is conducted perfomace anlysis. These outcomes underscore the potential of quantum machine learning in defect diagnosis, offering heightened precision and reliability in practical applications with fast computing.
The field of Prognostics and Health Management (PHM) aims to predict the behavior of machines to make informed maintenance decisions. In the Oil and Gas industry, fault mode diagnosis, as a PHM activity, has been applied to rotating machinery such as compressors, centrifugal pumps, and submersible motors using traditional Machine Learning (ML) and Deep Learning techniques. With the emergence of a new and rapidly growing research field called Quantum Computing (QC), there is now potential for even more efficient and accurate predictions. The QC has contributed to different purposes and contexts, such as optimization, artificial intelligence, simulation, cybersecurity, pharmaceutics, and the energy sector. Despite the current limitations of hardware, QC has been explored to improve the speed and efficiency of ML models. This master thesis focuses on the application of Quantum Machine Learning (QML) to diagnose rolling bearings which are essential components in rotating machinery, based on vibration signals. We apply hybrid models involving the encoding and construction of parameterized quantum circuits connected to a classical neural network, the Multi-Layer Perceptron (MLP). The study uses the Variational Quantum Eigensolver framework along with rotation gates and different entanglement (two-qubits) gates (CNOT, CZ and iSWAP), and explores the impact of varying the number of layers (1, 5 and 10) in the quantum circuit. We use two databases of different complexity levels not previously explored with QML, namely Case Western Reserve University (CWRU) and Jiangnan University (JNU), with 10 and 12 failure modes, respectively. For CWRU and JNU, all QML models presented higher accuracy than the classical MLP. These results suggest that, despite the current limitations of quantum environments, QML models are promising tools to be further investigated in PHM activities in the Oil and Gas industry.
Academia and enterprises have explored Prognostic and Health Management (PHM) to perform the diagnosis of failure modes via several traditional Machine and Deep Learning methods.However, the computation scenario is heading toward new advances, which include Quantum Processing Units.Due to its promising results in terms of speed and scalability, research centers worldwide began experimenting with models that lay at the intersection of machine learning and quantum computing.In this sense, a new technique that has already been applied in different scenarios is Quantum Machine Learning (QML), which aims to improve conventional methods in terms of performance and results.This work aims to apply QML models for the fault diagnosis of bearings, an important rotating machinery component, by vibration signals.We apply hybrid models involving the encoding and construction of parameterized quantum circuits connected to a classical neural network.The study uses rotation gates and different entanglement gates (CNOT, CZ and iSWAP), and explores the impact of varying the number of the quantum circuits layers.We perform a classical Multilayer Perceptron model for comparisons purposes.We use the database Case Western Reserve University with 10 failure modes.The obtained results suggest that, despite the current limitations of quantum environments, QML models are promising tools to be further investigated in PHM activities.
Prognostics and Health Management (PHM) concerns predicting machines’ behavior to support maintenance decisions through failure modes diagnosis and prognosis. Diagnosis is broadly applied in the context of rotating machines’ state classification using several traditional Machine Learning (ML) and Deep Learning (DL) methods. Recently, Quantum Computing (QC), a new and expanding research field, has contributed to different purposes and contexts, such as optimization, artificial intelligence, simulation, cybersecurity, pharmaceutics, and the energy sector. Despite the current limitations in terms of hardware, QC has been studied as an alternative for improving models’ speed and computational efficiency. Specifically, this paper proposes a Quantum Machine Learning (QML) approach to diagnose rolling bearings, which are essential components in rotating machinery, based on vibration signals. We apply hybrid models involving the encoding and construction of parameterized quantum circuits (PQC) connected to a classical neural network, the Multi-Layer Perceptron (MLP). We consider combinations of the Variational Quantum Eigensolver (VQE) framework with rotation gates and different entanglement (two-qubits) gates (CNOT, CZ and iSWAP). For each PQC configuration, we assess the impact of the number of layers (1, 5 and 10). We use two databases of different complexity levels not previously explored with QML, namely CWRU and JNU, with 10 and 12 failure modes, respectivel. For CWRU, all QML models presented higher accuracy than the classical MLP. For JNU, all QML models were superior to classical MLP as well. These results suggest that, despite the current limitations of quantum environments, QML models are promising tools to be further investigated in PHM.
The study conducted by the prognostics and health management (PHM) entity inquires into the prediction of machine behavior to aid in decision-making through failure modes classification via diagnostics and estimation. This is primarily performed as a classification of the state of rotating machines using diversified The methodologies of Machine Learning (ML) and Deep Learning (DL) have garnered significant attention in contemporary research. In recent times, Quantum Computing (QC) has surfaced as a novel domain of inquiry. With applications in various fields such as optimization, artificial intelligence (AI), gaming, cybersecurity, pharmaceuticals, and the energy sector. Despite hardware limitations, QC is considered an approach to accelerate models and enhance computational power. This paper presents a QML approach for detecting defects in rolling element bearings of rotating machinery using vibration signals. The method employs a hybrid model: a classical neural network to represent a class of parameterized quantum circuits (PQCs), specifically the multi-layer perceptron (MLP). Under different configurations of the Variational Quantum Eigensolver (VQE) system, various types of spin gates and unique two-qubit gates such as CNOT, CZ, and ISWAP are implemented in each PQC configuration. The performance of these PQC configurations is evaluated using two databases of varying complexities: JNU with a failure mode of 10, and CWRRU with a failure mode of 12. Performance results show that all QML models outperform the classical MLP for the two databases considered. This indicates the promising potential of QML models in the field of PHM, despite the current fragile quantum environments.
Quantum Computing (QC) was born as a purely theoretical subject. Thus, in the beginning, the main category of researchers in the area was physicists and mathematicians. Since IBM introduced the first quantum computer, global interest in this emerging technology has been growing more and more, involving computer science and engineering researchers. Machine learning is another rapidly growing technology, used in a wide range of applications where traditional computing is not sufficient. The two technologies are merged in the thesis work. The different ways in which a combination of them can occur are discussed, as well as the explanation why there is not a single definition of Quantum Machine Learning (QML). Firstly, quantum computing is examined in depth. The basic elements of quantum computation theory, based on the fundamental principles of quantum mechanics, are introduced. The attention is focused on the qubit description on the Bloch sphere, the fundamental quantum gates, and examples of quantum algorithms. Therefore, the differences between quantum simulators and quantum processors are analysed, including the frameworks that provide tools to interact with them. Focusing on Quantum Machine Learning, the different definitions of the technology are described. There are four distinct approaches that are differentiated by the type of data used (quantum or classical) and by the device processing them (quantum or classical). In this thesis, QML is defined as the approach that uses classical data to feed a quantum device. The two possible methods for developing QML algorithms are: 1) to translate existing classical models into the terms of quantum computing; 2) to leave the existing models of classical machine learning completely, and explore new possible ways of doing QML to take full advantage of the physical characteristics of a quantum computer. The strategy adopted is the first one, thus it is used a state-of-the-art hybrid model of Quanvolutional Neural Network (QNN), composed of both quanvolutional and classical convolutional layers. The quanvolutional layers reproduce the same concept of the classical ones: they are made of quantum filters that generate feature maps through locally transforming input data. The transformation can be performed by variational quantum circuits or random quantum circuits. The specific field where Quantum Machine Learning is applied is image classification. The considered case study is a labeled dataset with X-ray images of lungs (labels: pneumonia, normal). The images are pre-processed and cropped for being used by quantum circuits, thus some information is lost. The main aim of the thesis is to understand whether and in which ways Quantum Machine Learning could bring advantages in the mentioned classification in terms of timing, accuracy, or dataset reduction, despite the large decrease in images' resolution. Different configurations of the hybrid model (with a different number of quantum filters and different settings of convolutional layers) are analysed and tested. The current quantum devices have many limitations, thus the time is still immature to observe a significant change in perspective. Nevertheless, the obtained results are promising for the future of QML in terms of dataset and hyperparameters reduction.