Condition Monitoring

Cyber-Physical Production Networks (CPPNs) represent the backbone of modern smart manufacturing systems, integrating physical industrial processes with computational control and communication infrastructures. Fault diagnosis in such systems is increasingly being addressed using data-driven approaches that leverage the rich sensory data streams and machine learning models to predict and identify faults proactively. This paper explores recent advances in data-driven fault diagnosis techniques tailored for CPPNs, with a focus on their implementation, evaluation, and integration within intelligent production environments. We classify existing approaches, evaluate their performance and challenges, and propose a unified framework integrating multi-source data fusion and adaptive learning models. Furthermore, the paper highlights opportunities for enhancing fault resilience in CPPNs and outlines emerging directions in fault prediction leveraging edge-AI, federated learning, and self-healing control architectures.

Induction motors are mainly used for variable load applications and it is vital to have a condition monitoring system with capabilities to diagnose motor faults in variable load conditions. The environment noise varies has non-linear relation with motor load and it challenges the decision making capability of the condition monitoring system. This paper addresses the issue of reliable decision making on the existence of bearing faults in variable load conditions. Two type of threshold schemes have been proposed to reliably diagnose bearing faults in Park vector modulus spectrum. The performance of the developed threshold based condition monitoring system has been analyzed theoretically and experimentally.

which funded my graduate education and provided the test equipment and resources necessary for laboratory set-up and data collection. I would like to personally thank Larry Hall for his guidance and support of this research and his contribution to my personal and professional development. I am fortunate to have the support of several others at Moog Inc. who allowed me to dedicate time to this research in concert with my daily work initiatives. Special thanks go to Rick Fosdick, Paul Stoelting, and Duff Bushway. I would also like to thank my thesis advisor, Dr. Jason Kolodziej, who provided constant guidance and support. vi

Purpose The purpose of this paper is mainly to highlight how a simplified and streamlined approach to the condition monitoring (CM) of industrial rotating machines through the application of frequency domain data combination can effectively enhance the eMaintenance framework. Design/methodology/approach The paper commences by providing an overview to the relevance of maintenance excellence within manufacturing industries, with particular emphasis on the roles that rotating machines CM of rotating machines plays. It then proceeds to provide details of the eMaintenance as well as its possible alignment with the introduced concept of effective vibration-based condition monitoring (eVCM) of rotating machines. The subsequent sections of the paper respectively deal with explanations of data combination approaches, experimental setups used to generate vibration data and the theory of eVCM. Findings This paper investigates how a simplified vibration-based rotating machinery faults classific...

This article studies the girth and cogirth problems for a connected matroid. The problem of finding the cogirth of a graphic matroid has been intensively studied, but studies on the equivalent problem for a vector matroid or a general matroid have been rarely reported. Based on the duality and connectivity of a matroid, we prove properties associated with the girth and cogirth of a matroid whose contraction or restriction is disconnected. Then, we devise algorithms that find the cogirth of a matroid M from the matroids associated with the direct sum components of the restriction of M. As a result, the problem of finding the (co)girth of a matroid can be decomposed into a set of smaller sub-problems, which helps alleviate the computation. Finally, we implement and demonstrate the application of our algorithms to vector matroids.

This study proposes a novel approach to automate the interpretation of transformer frequency response analysis (FRA) data, addressing the inherent subjectivity of traditional methods. It introduces a framework that converts conventional FRA Bode plots into polar plots, which intuitively combine magnitude and phase information into a single, visually rich image. The study then evaluates the use of image processing techniques to analyze these polar plots for fault detection and classification. By simulating common transformer faults, such as winding deformation and short circuits, we demonstrate that each fault type produces a unique and identifiable pattern on the polar plot. The proposed methodology, which includes image preprocessing, feature extraction, and classification, achieves a remarkable 97.5% overall accuracy in diagnosing the transformer's condition. A comparative analysis with traditional statistical methods, like the correlation coefficient, confirms the superior performance and objectivity of our image-based framework. The results indicate that this approach can significantly reduce reliance on expert judgment, improve diagnostic reliability, and enable the development of standardized automated systems for transformer condition monitoring.

Induction motors are mainly used for variable load applications and it is vital to have a condition monitoring system with capabilities to diagnose motor faults in variable load conditions. The environment noise varies has non-linear relation with motor load and it challenges the decision making capability of the condition monitoring system. This paper addresses the issue of reliable decision making on the existence of bearing faults in variable load conditions. Two type of threshold schemes have been proposed to reliably diagnose bearing faults in Park vector modulus spectrum. The performance of the developed threshold based condition monitoring system has been analyzed theoretically and experimentally.

Dissolved Gas Analysis (DGA) is universally recognized as the most effective non-intrusive method for monitoring the condition of oil-immersed power transformers, providing critical insights into incipient thermal and electrical stresses. However, the diagnostic reliability of DGA is often challenged by the limitations of individual interpretation techniques, which can yield ambiguous or conflicting results for the same dataset. This article presents a comprehensive, client-side software analyzer designed to mitigate these uncertainties by integrating a suite of established and novel graphical DGA methods into a single, interactive platform. The tool automates the complex plotting and interpretation of the complete Duval Triangle suite (DT1, DT4, DT5), the advanced Duval Pentagons (DP1, DP2), and the prognostic Low Energy Degradation Triangle (LEDT). By synthesizing results from these complementary methodologies and incorporating advanced trend analysis, the platform enhances diagnostic accuracy and provides asset managers with a holistic, data-rich view of transformer health, facilitating more reliable and informed maintenance decisions.

This research work explores the feasibility of in- creasing system reliability by applying condition monitoring systems to selected circuit breakers and transformers. Smart grid technology decreases maintenance costs by deploying condition monitoring systems that informs the operator of impending failures or ameliorates those conditions. Since condition monitoring systems can themselves increase operating costs, only selected critical transformers and circuit breakers are chosen to be equipped with these systems. This paper describes a method to identify the most critical transformers and circuit breakers with the aid of contingency ranking methods. Different summer load cases are studied and the results are compared to find the most severe contingencies that affect the study area. Once the critical components of the system are identified, corrective actions are found for each severe contingency to be applied.

Im Rahmen dieses Beitrags werden Ergebnisse einer Untersuchung an feststoffgeschmierten Wälzlagern vorgestellt. Betrachtet werden dabei Lager, welche einen speziellen, modifizierten Käfig verwenden. Die Käfigtaschen des Käfigs dienen dabei, zusätzlich zu ihrer ursprünglichen Funktion, der Führung der Wälzkörper, als Schmierstoffdepot. Es werden zunächst Prüfaufbau und die Versuchsbedingungen erläutert und in diesem Zusammenhang wird gezeigt, dass der in diesem Beitrag verwendete Aufbau, verglichen mit dem Aufbau vorangegangener Arbeiten, eine deutlich reduzierte Streuung aufweist. Als eine nicht zu vernachlässigende Fehlerquelle bei der gravimetrischen Bestimmung des Käfigtaschenverschleißes wurde das hygroskopische Verhalten des Polymercompounds identifiziert. Einer Verfälschung dieser Messergebnisse durch die unkontrollierte Feuchtigkeitsaufnahme aus der Umgebung, muss durch einen zeitlich vorgelagerten Trocknungsprozess unter definierten Bedingungen vorgebeugt werden. Zudem wird ...

The global manufacturing sector is rapidly transitioning towards smarter, datadriven operations to enhance productivity, reduce operational costs, and improve overall equipment effectiveness. Predictive Maintenance (PdM) stands out as one of the most transformative approaches within this paradigm. By anticipating equipment failures before they occur, PdM ensures that maintenance is carried out only when required, leading to optimal utilization of resources. Machine learning (ML) techniques form the core of PdM, leveraging data to detect patterns and anomalies that precede system failures. This paper presents a comprehensive comparative study of leading ML algorithms-both traditional and deep learning-based-deployed for PdM in manufacturing systems. Algorithms such as Decision Trees, Random Forests, Support Vector Machines, K-Nearest Neighbors, Multilayer Perceptrons, Long Short-Term Memory (LSTM) networks, and Convolutional Neural Networks (CNNs) are rigorously evaluated on standardized datasets. Each model is analyzed in terms of prediction accuracy, interpretability, computational efficiency, and suitability for deployment in industrial settings. The results reveal key trade-offs and offer recommendations tailored to different types of manufacturing environments. The paper also discusses the current challenges in PdM adoption and offers insights into future research directions, including the integration of explainable AI and real-time edge computing solutions.

In this paper, novel Wavelet Spectral Kurtosis (WSK) technique is applied for early diagnosis of gear tooth faults. Two variants of Wavelet Spectral Kurtosis technique called Variable Resolution WSK and Constant Resolution WSK are considered for diagnosis of the pitting gear fault. The gear residual signal obtained by filtering the gear mesh frequencies is used as the input to the SK algorithm. The gain obtained by using the wavelet SK techniques when compared to classical Fourier Transform (FT) based SK, is confirmed by estimating tooth wise Fisher's Criterion of diagnostic features. The final diagnosis decision is made by a three stage decision making technique based on weighted majority rule. The probability of the correct diagnosis is estimated for each SK technique for comparison. An experimental study is presented in detail to test the performance of the wavelet spectral kurtosis techniques and the decision making technique.

The early detection of rotor imbalance is crucial for maintaining the reliability and operational efficiency of wind turbines. In this study, we present a comprehensive methodology that leverages vibrationresponse-only accelerometer measurements and utilizes advanced machine learning techniques, specifically the extended isolation forest (EIF), for the early identification of rotor imbalance.

A new methodology for wind turbine condition monitoring that utilizes vibration data and Mel-frequency cepstral coefficients (MFCCs) is being developed to accurately assess the condition of the wind turbine gearbox. Traditional condition monitoring techniques rely on physical inspections, which can be time-consuming and labor-intensive . This new approach offers a more efficient and cost-effective solution.

The reliability requirements of wind turbines have increased significantly in recent years in the search for a lower impact on the cost of energy. In addition, the trend towards larger wind turbines installed in remote locations has significantly increased the cost of repair or replacement of the component. In the wind industry, therefore, condition monitoring is crucial for maximum availability . This contribution makes a review of supervised machine learning classification techniques for wind turbine condition monitoring using only SCADA data already available. That is, without installing extra sensors or costly purpose-built data sensing equipment. Although there has been extensive research into the use of machine learning techniques for wind turbine monitoring, the more recent trend in this type of literature is to focus on a specific WT subassembly: the bearings and planetary gearbox [2], the generator and power converter [3], the blades [4], etc. Oil debris systems can detect pitting failures but cannot detect cracking faults. Vibration based systems can detect both pitting and cracking, but most cannot determine the health of components in the planetary section. This work approaches condition monitoring of various wind turbine components (torque actuator, pitch actuator, pitch sensor, and generator speed sensor) with a unique strategy. In particular, for this purpose, a review of supervised machine learning classification techniques is performed and analyzed.

The reliability requirements of wind turbine (WT) components have increased significantly in recent years in the search for a lower impact on the cost of energy. In addition, the trend towards larger WTs installed in offshore locations has significantly increased the cost of repair of the components. In the wind industry, therefore, condition monitoring is crucial for maximum availability. In this work, without using specific tailored devices for condition monitoring but only increasing the sampling frequency in the already available sensors of the SCADA system, a data-driven multi-fault diagnosis strategy is contributed. The advanced WT benchmark proposed by [1] is used. That is a 5 MW modern WT simulated with the FAST [2] software and subject to various actuators and sensors faults of different type. The measurement noise at each sensor is modeled as a Gaussian white noise. First, the SCADA measurements are pre-processed and feature transformation based on multiway principal component analysis (MPCA) is realized. Then, 10-fold cross validation support vector machines (SVM) based classification is applied. In this work, SVMs were used as a first choice for fault detection as they have proven their robustness for some particular faults [3-5] but never accomplished, to the authors' knowledge, at the same time the detection and classification of all the proposed faults taken into account in this work. To this end, the choice of the features as well as the selection of data are of primary importance. Simulation results show that all studied faults are detected and classified with an overall accuracy of 98%. Finally, it is noteworthy that the prediction speed allows this strategy to be deployed for real-time condition monitoring in WTs.

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson

This paper reviews and demonstrates how artificial intelligence (AI) methods can enhance the operational performance, reliability, and economics of a typical 5 kW rooftop photovoltaic (PV) system. We synthesize state-of-the-art AI approaches across four key functions: (1) maximum power point tracking (MPPT) and adaptive control, (2) short-term power forecasting for energy management, (3) automated fault and soiling detection, and (4) predictive maintenance and lifetime estimation. We propose an integrated AI framework combining reinforcement-learning (RL) MPPT, Spatio-temporal LSTM forecasting, convolutional neural network (CNN)-based visual inspection, and unsupervised anomaly detection for maintenance scheduling. Equations describing PV behavior and the AI objective functions are provided. A simulation/case study for a canonical 5 kW rooftop system illustrates potential improvements: energy yield increases (1 6% depending on conditions) and reduced downtime through earlier fault detection.

Effective sensorless dynamic speed estimation is desirable for both speed sensorless motor-drive applications and for on-line induction motor condition monitoring and assessment. In this paper, a sensorless neural adaptive speed filter is developed for induction motors running off a voltage source inverter. The filter is demonstrated by comparisons with experimental speed measurements and spectral speed estimates. In addition to nameplate information required for the initial set-up, the proposed neural speed filter uses only measured motor currents and voltages. Initial training of the speed filter is accomplished off-line, using rotor slot harmonic-based speed estimates. The developed speed filter is tested on data from a healthy motor and motor with 4 broken rotor bars. The resulting speed filtering accuracy compares favorably with speed estimation results reported in the literature. The study demonstrates the feasibility of neural adaptive speed filters for effective inverter-fed induction motor speed estimation, without explicit use of motor model parameters and speed measurements.

Partial discharge (PD) measurements have proved their reliability for health monitoring of insulation systems in power system components including synchronous generators, power transformers, switchgear and cables etc. Online condition monitoring and pro-active detection of PD faults have been highly demanded over the last two decades. This paper provides results from a research project to develop advanced non-intrusive sensing technologies that are cost effective, reliable and efficient for early detection of PD faults in medium voltage (MV) and high voltage (HV) air-insulated switchgear. Three sensors (high frequency E-field (D-dot) sensor, Rogowski coil and loop antenna) have been developed and tested under various PD faults and their performance were evaluated in comparison with high frequency current transformer (HFCT) which is being used commercially for PD testing and measurement. Among these three sensors, it is shown that D-dot sensor and Rogowski coil are more dependable wh...

In this study, the relationship between vibration and tool wear is investigated during high-speed dry turning by using statistical parameters. It is aimed to show how tool wear and the work piece surface roughness changes with tool vibration signals. For this purpose, a series of experiments were conducted in a CNC lathe. An indexable CBN tool and a 16MnCr5 tool steel that was hardened to 63 HRC were both used as material twins in the experiments. The vibration was measured only in the machining direction using an acceleration sensor assembled on a machinery analyzer since this direction has more dominant signals than the other two directions. In addition, tool wear and work piece surface roughness are measured at different cutting time intervals where the cutting speed, radial depth of cut, and feed rate are kept constant. The vibration signals are evaluated using statistical analysis. The statistical parameters in this study are the Root Mean Square (RMS), Crest Factor, and Kurtosis values. When the flank wear increases, the Kurtosis value and RMS also increase, but the Crest factor exhibited irregular variations. It is concluded that these statistical parameters can be used in order to obtain information about tool wear and work piece surface roughness.

Monitoring systems are prone to record huge amount of data, only a minor part of which could be of interest. Maintaining a giant data base from the monitoring system of a large structure equipped with dozens of sensors is a costly challenge. One way for solving this question lies in discarding everything that does not enter in the scope of what was considered as "interesting event " at the design stage, with a big risk of missing an unattended but essential event. Another way could be to store less data, making a kind of "survey " instead of recording continuously, with a possibility to rebuild events from the discontinuous records. This way of doing is suitable in the case of minutes lasting phenomena only, it is clear that a sudden and brief phenomenon could be missed in such a configuration. This method was used in the case of vortex shedding excitation of the deck of a large cable stayed bridge, in order to rebuild the history of passed excitations and evalua...

The bearing is very essential component in any mechanism. It is always running under high speed and high load conations. Failures in bearings often results in lengthy industrial downtime that has economic consequences. Prompt diagnostics of bearings failure is essential not only for the safe operation of machines, but also for the reduction of maintenance cost of any machine. Different methods are used for detection and diagnosis of defects of bearings; they may be broadly classified as vibration and acoustic measurements, temperature measurements and wear debris analysis. The most fundamental tool for bearing diagnosis is vibration analysis. The strength and weakness of vibration represents the location of defects. This paper discuss about the bearing failure diagnosis using vibration analysis. Crest factor is defined as the ratio of the peak value of a waveform to its RMS value. Comparative analysis is made between damaged and healthy Rolling and Ball bearings.FFT analyzer and accelerometer is use to perform experimental measurements in axial and radial directions.

The IEC 1000-4-7 standard is currently mandated to fix both procedures and instruments for evaluating the level of harmonics and inter-harmonics in electromagnetic interference measurements for power supply systems and connected equipment. Prescriptions and recommendations related to inter-harmonics seem to be not well detailed; this is particularly true for those showing a nonstationary behaviour. As a matter of fact, only a digital signal-processing approach, based on the Discrete Fourier Transform, is strongly recommended; little or nothing is said both about the sampling rate to be adopted and the time window to be observed. The paper aims at designing suitable measurement procedures based on advanced digital signal-processing algorithms in order to overcome the aforementioned limits. In particular, to allow non-stationary inter-harmonics to be monitored and properly analysed, a new approach, based on the use of Time-Frequency Representations, improved by a recently introduced r...

Substantial efforts have been devoted to developing Condition Monitoring techniques to provide timely preventative measures for ensuring a safe and cost-effective operation of electromechanical systems. High investment of installation and implementation in using conventional condition monitoring techniques such as vibration based monitoring makes it difficult to be used in most industries such as petrochemical processing, food and drinking processes, paper mills and so on where large number of motor drives are used but operational profits are very limited. To overcome the shortages of vibration based monitoring, this project focuses on developing condition monitoring techniques based on electrical signal analysis which can offer great savings as electric signatures that can monitor a large system are generally available in most motor drives. However, fault signatures in electrical signatures such as instantaneous current and voltage signals are very weak and contaminated by noise. T...

There are considerable benefits to applying Artificial Intelligence (AI) to automation in modern industries, including oil and gas. Oil and Gas Industries have many complex systems in their production and flow line. Condition monitoring of oil and gas equipment provides the fault diagnosis of components, machines, systems, equipment, data hardware, and even software. The predictive maintenance and industry 4.0 applications have more advantages in petrochemical industries to make a safer economic environment. To understand the effectiveness of AI in condition monitoring for oil and gas industries, a case study related to the condition monitoring of drilling machines is conducted, and an ANN and XG Boost ensemble algorithm is used to analyze and predict potential failures. In the signal collection, acoustic emission sensors, signal conditioning devices, NI-data acquisition cards, and LabVIEW software were used. The dataset was collected using a healthy tool, a worn tool, and a tool with a broken tip. The novelty of this work is the proposal of an approach for tool wear monitoring in drilling using acoustic emission sensors and considering Wavelet Packet Decomposition (WPD) for further analysis for feature extraction. The extracted features from WPD are given as input for the ANN and XG Boost Algorithm for classification to identify the health of the drill bit and machine. The result shows that the prediction accuracy is higher in the XG Boost algorithm than ANN. Average accuracy of 94% is achieved in XG Boost whereas a 90% average accuracy is achieved in ANN. So, a detailed analysis has been made to compare the ANN and XG Boost performance, finding the effectiveness of AI-based condition monitoring in enhancing the monitoring effectiveness and equipment safety in Oil and Gas Industries.

A fault recognition technique for the internal combustion engines using time-frequency representations of vibration signal has been presented in this paper. Engine block vibration results as a sum of many excitations mainly connected with engine speed and their intensity increases with the appearance of a fault or in case of higher engine elements wearing. In this paper an application of acceleration signals for the estimation of the influence of piston skirt clearance on diesel engine block vibrations has been described. Engine body accelerations registered for three simulated cases representing piston skirt clearance variations were an object of preliminary analysis. The presented procedures were applied to vibration and pressure signals acquired for a 0.5 dm3 Ruggerini, air cooled diesel engine. There are important features of the vibration signal that are sensitive to the change of IC engine condition. For that reason the DWT transform was applied. Based on the results, authors propose detection and piston skirt clearance monitoring algorithm.

Accurate monitoring of tool conditions is critical for optimizing productivity in modern manufacturing. This research presents a novel dual-method approach for tool state identification in milling operations, integrating transfer learning models with vision-based image processing to enhance tool conditions detection and classification. The study uses scalogram images derived from acoustic emission signals and input into several pretrained deep learning models, including a modified ResNet50 enhanced with dropout layers. The modified ResNet50 achieved the highest validation accuracy of 93%, while testing accuracy ranged from 51 to 72% across all models due to variations in cutting parameters. Notably, the modified ResNet50 demonstrated superior performance with a testing accuracy of 75%, showcasing its ability to handle noisy data through improved generalization. Despite these advances, sensor noise and cutting parameter variations significantly impacted model performance, particularly when distinguishing between worn and broken tools, with accuracy dropping below 50% for noisy datasets. To address these limitations, the study integrates a vision system that directly analyzes the tool's physical condition by comparing individual cutting edges using pixel differences. This dual-method approach enhances the accuracy of tool condition identification, providing more reliable assessments that minimize production interruptions, reduce misclassifications, and optimize tool replacement decisions. This integration of sensor-based and vision-based tool condition monitoring offers a robust solution for modern manufacturing challenges.

An analytical approach is presented in this paper to minimize the influence of disruptive current spikes occurring at low frequency operation in a Variable Frequency Drive (VFD) induction motor. The recorded test data of load current at low frequency of a high power induction motor coupled with a reciprocating pump in a process industry, is analyzed by Prony Analysis (PA) and Hilbert transform (HT). Optimal values of motor design parameters that deliver sufficient mechanical power at low frequency and cause the least disruptive current spikes is arrived at, by using intelligent optimization techniques like Artificial Bee Colony (ABC), Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Simulated Annealing (SA). Using the most optimal design values obtained, Taguchi method is applied on the VFD control parameters for getting minimal current spikes. The output of the Taguchi method is compared with the experimentally set value at site. The comparison proved that with the opt...

Prognostics and Health Management (commonly called PHM) is a field that focuses on the degradation mechanisms of systems in order to estimate their health status, anticipate their failure and optimize their maintenance. PHM uses methods, tools and algorithms for monitoring, anomaly detection, cause diagnosis, prognosis of the remaining useful life (RUL) and maintenance optimization. It allows for permanently monitoring the health of the system and provides operators and managers with relevant information to decide on actions to be taken to maintain the system in optimal operational conditions. This paper aims to present the emergence of the PHM thematically to describe the subjacent processes, particularly prognosis, how it supplies the different maintenance strategies and to explain the benefits that can be anticipated. More specifically, this paper establishes a state of the art in prognostic methods used today in the PHM strategy. In addition, this paper shows the multitude of po...

Diagnosis of stator current has been presented by several authors for identifying various bearing and rotor faults in an induction motor. However, there is very limited literature covering the effect of interturn faults on the stator current pattern. This is due to the reason that the effect of motor interturn insulation faults in the initial stages being not discernable and any changes in the capacitive current components are negligibly very small and of very high frequency content. This paper deals with a simple technique, utilizing the effect of the interturn fault in modifying the high frequency components of the applied pulse width modulated (PWM) voltage. Insulation Resistance (IR) and Polarization Index (PI) are two universally accepted diagnostic tests for insulation tests. The change in insulation strength between the turns affects the capacitive component of the stator line current. Resulting changes in wave shapes of applied voltage have been studied with respect to both the distance of interturn faults from line end and reduction in the insulation strength; and hence in the insulation resistance value. The studies have been made by computer simulation and validated by experiments. There is ample evidence that an impending and progressing interturn fault can be identified in Adjustable Speed Drives driven by frequency converters by studying line end coil voltage waveforms.

This article investigates the effect of various lightning impulses on the aged epoxy insulations. To secure the reliability of insulation materials, impulse testing is very crucial at design stage. For this reason, electrical equipment is subjected to a lightning impulse voltage test with a standard waveform of [(1.2 × 50 that natural lightning conforms not only the standard impulse voltage waveform but also non-standard lightning voltage waveforms. Considering this fact, a standard waveform of [(1.2 on-standard lightning impulse (NSLI) platform. Each waveform has a peak voltage of 100 kV. Besides, pure epoxy, epoxy-alumina and epoxy-titania nanocomposite insulation samples are prepared in laboratory. These prepared insulation samples are subjected to electro-thermal aging with an interval of 100 hours for a total of 600 hours. After each aging step, polarization and depolarization currents (PDC) are measured. These PDC measurements are then used to develop an Extended Debye Model (EDM) of the insulation within a simulation platform. Therefore, the simulated impulse voltages are applied to the insulation models and corresponding impulse responses are analysed. The results indicate that impulse currents increase with aging, and epoxy-alumina nanocomposites exhibit superior dielectric properties compared to other samples.

Ich m öchte mich bei allen bedanken, die mich bei der Erstellung dieser Arbeit unterstützt haben. Zuerst gilt mein Dank dem Betreuer dieser Arbeit, Herrn Dr. Robert Glawar, der durch Anregungen und konstruktive Kritik bei der Erstellung maßgeblich mitgewirkt hat. Weiters möchte ich mich bei DI Florian Öhlinger und DI Jakob Giner für die vornehmliche Übernahme der Projektbetreuung bedanken. Herrn DI Sebastian Meixner und Christoph H albwidl M Sc möchte i ch f ür d ie Z usammenarbeit und technische Unterstützung danken. Zuletzt gilt mein Dank meiner Familie und meiner Freundin, deren Unterstützung den Abschluss dieses Studiums ermöglicht.

An analytical investigation has been conducted to examine the dynamic characteristics of a dual rotor system of an electromagnetic retarder. It focuses on the effects of mass imbalance in rotors and the system's inclination, which are essential for ensuring stability and optimal performance. The governing equations that describe the nonlinear dynamic response of the system under varying levels of unbalance were derived using the Lagrange method along with finite element modeling, which incorporates lumped mass and Euler beam models. The system's critical speeds and mode shapes are determined by solving an eigenvalue problem. The forced response analysis indicated that the rotational amplitude surpasses the transverse amplitude due to the overhung rotor configuration, which amplifies the centrifugal force as speed increases, influenced by different degrees of unbalanced mass. The amplitude of vibration is lower within the operational range; however, the system's tilt is critical and should be considered during the retarder design.

Anomaly detection is the characterization of a normal behavior of a system or process and identification of any deviation from such normal behavior. Anomaly detection of critical systems provides an important financial and client competitive advantage since it gives the decision-maker lead-time and flexibility to manage the health of the system. Fuel systems are complex and mission critical systems that require high operational availability because of the high costs associated with the services they provide. In complex systems, it is not uncommon to monitor a quality-related response which relies on the functional form between several variables using a non-linear relationship. We present in this paper a new monitoring framework for smart fuel systems utilizing outlying observations detection and monitoring using ccharts. The traditional control charts based on the Hotelling's T 2 statistic were deficient in detecting SFS anomalies and a new approach was necessary to isolate faulty profiles. The proposed methodology requires a simple quality performance test that can be performed once assembly is completed to assure readiness for client use or completion of a job. The results were tested and validated using scaled data that mimic an actual system. The methodology presented in this paper is scalable and can be applied to a wide range of systems to assess their health from an inspection check to anticipate and avoid failures.

Anomaly detection is the characterization of a normal behavior of a system or process and identification of any deviation from such normal behavior. Anomaly detection of critical systems provides an important financial and client competitive advantage since it gives the decision-maker lead-time and flexibility to manage the health of the system. Fuel systems are complex and mission critical systems that require high operational availability because of the high costs associated with the services they provide. In complex systems, it is not uncommon to monitor a quality-related response which relies on the functional form between several variables using a non-linear relationship. We present in this paper a new monitoring framework for smart fuel systems utilizing outlying observations detection and monitoring using ccharts. The traditional control charts based on the Hotelling's T 2 statistic were deficient in detecting SFS anomalies and a new approach was necessary to isolate faulty profiles. The proposed methodology requires a simple quality performance test that can be performed once assembly is completed to assure readiness for client use or completion of a job. The results were tested and validated using scaled data that mimic an actual system. The methodology presented in this paper is scalable and can be applied to a wide range of systems to assess their health from an inspection check to anticipate and avoid failures.

The fully automated manufacturing system suffers from breakage and deterioration of tools. This paper proposes a novel setup and methodology to monitor the tool flank wear from emitted sound signals while turning operation is on. The proposed method is efficient then the other similar methods were a rough estimation of the tool condition into slight, medium and severe wear categories only can identified. But the proposed technique is free from using complex mathematical techniques and delivers the state of the tool wear in terms of voltage using the attached hardware circuits. By the conversion of the emitted sound pressure into corresponding voltage the correlation between the voltage and tool flank wear helped us to perfectly do the classification of carbide inserts which are created with Aluminium Titanium Nitride. The classification is done with the help of standard deviations from the class average of different categories. The classification accuracy is confirmed with correlati...

This paper proposes a new method for extraction of low frequency oscillation from the envelope analysis of Motor Startup Current for diagnosis of Broken Rotor Bar Fault of Induction motor through Wavelet and Hilbert Transforms. The Envelope is the modulus of ...

This paper establishes a mathematical model of induction machine connected to a frequency inverter necessary to adjust the electric motor drive. The mathematical model based on the Park's theory allows the analysis of the whole spectrum (electric car – frequency inverter) to drive the electric trolley bus made on ASTRA Bus Arad (Romania). To remove higher order harmonics, the PWM waveform of supply voltage is used, set in the general case. Operating characteristics of electric motor drive are set to sub-nominal frequency (f < 50 Hz) and for different angles of the control switching elements in the construction of the inverter. Tensions in the stator of the equivalent machine are determined by three components: the fundamental component; the second component corresponding to the higher order harmonics, which rotate in direct connection; and the appropriate third component of higher order harmonics, which are rotated in reverse. The mathematical model developed will help to det...

Vibration sensors have gained increasing popularity as valuable tools for Prognostics and Health Management (PHM) applications, enabling early detection of mechanical failures in industrial machines. Vibration signals comprise two main sources of information: periodic vibrations from components, phase-locked to the rotating speed (e.g., gears), and non-deterministic broadband vibrations associated with bearings, structure, and background noise. In PHM applications, it is important to decompose vibrations to optimize the use of different diagnostic methods for each signal component. In practice, the decomposition should be cost-effective by working without supplementary information about system operating conditions and kinematics. A major application of this method is to separate the bearing and gear signals in a gearbox. This is because the gear signals are typically much stronger than the bearing signals, and can mask the presence of bearing faults. Existing methods of vibration source separation commonly rely on an auto-regression (AR) model of vibrations and employ adaptive filtering techniques to estimate its parameters. However, these methods suffer from degraded accuracy in complex geared vibrations containing numerous periodic components and requiring large filter length to promise high frequency resolution in component separation. To address these challenges, we propose a new method that utilizes dilated Convolutional Neural Networks (CNNs) instead of adaptive filtering to improve the accuracy of decomposing complex vibration signals, all without the need for any supplementary information. To evaluate the performance of the new method, we conducted experiments using both simulated signals and realworld vibrations. The simulation results demonstrate improved accuracy in signal decomposition when our method is used instead of adaptive filtering. Additionally, the new method applied to real vibrations, showcases significant enhancement in bearing failure detection through accurate isolation of bearing-related vibrations. This study reveals the potential of our new method in various PHM applications requiring highly accurate diagnostics and prognostics in complex geared vibrations, particularly when supplementary information about operating conditions and system kinematics is unavailable.

The use of artificial intelligence (AI) is increasing in various sectors of photovoltaic (PV) systems, due to the increasing computational power, tools and data generation. The currently employed methods for various functions of the solar PV industry related to design, forecasting, control, and maintenance have been found to deliver relatively inaccurate results. Further, the use of AI to perform these tasks achieved a higher degree of accuracy and precision and is now a highly interesting topic. In this context, this paper aims to investigate how AI techniques impact the PV value chain. The investigation consists of mapping the currently available AI technologies, identifying possible future uses of AI, and also quantifying their advantages and disadvantages in regard to the conventional mechanisms.

Insulating paper aging of power transformers, a reliability factor for this equipment, can be evaluated by non-destructive methods such as measuring chemical markers' content in transformer oil. Since the sampling temperature of oil can affect the accuracy of laboratory measurements, the correlation between novel aging marker methanol and the degree of polymerization (DP) of insulating paper has been studied in this study. Temperature correction factors have been obtained for the markers (2-Fal, water, and methanol), and the under-load transformers' data has improved the correlation formula. Sampling from a real transformer was carried out to optimize the calculated estimation formula. Various experiments have been performed on them, and their results have been analyzed. The real DP was also measured for the insulating papers sampled at the overhaul stage. The optimized model for insulating paper DP calculation using a temperature-corrected concentration of methanol in insulating oil has been provided in this work.

In several cases the vibration signals generated by rotating machines can be modeled as cyclostationary processes. A cyclostationary process is defined as a non-stationary process which has a periodic time variation in some of its statistics, and which can be characterized in terms of its order of periodicity. This study is focused on the use of cyclic spectral analysis, as a tool to analyze second-order periodicity signals (SOP), such as, those who are generated by either localized or distributed defects in bearings. Cyclic spectral analysis mainly consists of the estimation of the random aspects as well as the periodic behavior of a vibration signal, based on estimation of the spectral correlation density. The usefulness of cyclic spectral analysis for the condition monitoring of bearings, is demonstrated in this paper, through the analysis of several sections of vibration data collected during an endurance test of one of the two main gearbox transmissions of a helicopter.

The railway substructure is a significant infrastructure, requiring significant investment and serving as a vital national blood vessel. The surface layer of a railway track bed is a multilayer structure with three sub layers: Ballast, Sub-Ballast, and Subgrade. Ground Penetrating Radar (GPR) is a crucial tool for assessing track bed conditions, but its accuracy is challenged by resolution limitations and similar permittivity of track material sub layers. This study aims to verify and optimize GPR simulations using gprMax to improve track bed thickness and material classification in railway infrastructure. The research methodology involves simulating various parametric conditions, such as ballast fouling, track bed layer thicknesses, and moisture content scenarios. A structured approach is used, starting with establishing study area characteristics and configuring geometry and materials in gprMax. The appropriate GPR antenna and frequency settings are defined, and simulation settings and boundary conditions are established for numerical stability and accuracy. Results are analyzed using post-processing techniques, and visualization capabilities are used to compare simulated GPR scans under different conditions. The study concludes that GPR simulations in gprMax can effectively model the impact of ballast fouling, layer thickness variations, and moisture content on GPR signals. These simulations provide valuable insights into improving GPR data interpretation, promoting costeffective maintenance strategies by reducing the need for extensive physical testing. This research contributes to enhancing the reliability and efficiency of GPR in railway infrastructure maintenance.

As a significant and growing source of the world’s energy, wind turbine reliability is becoming a major concern. At least two fault detection techniques for condition monitoring of wind turbine blades have been reported in early literature, i.e. acoustic emissions and optical strain sensors. These require off-site measurement. The work presented here offers an alternative non-contact fault detection method based on the noise emission from the turbine during operation. An investigation has been carried out on a micro wind turbine under laboratory conditions. 4 severity levels for a fault have been planted in the form of added weight at the tip of one blade to simulate inhomogeneous debris or ice build up. Acoustic data is obtained at a single microphone placed in front of the rotor. Two prediction methods have been developed and tested on real data: one based on a single feature - rotational frequency spectral magnitude; and another based on a fuzzy logic interference using two input...