To see the other types of publications on this topic, follow the link: Window detection.
Dissertations / Theses on the topic 'Window detection'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Window detection.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Sipahigil, Oktay. "Multiple Window Detectors." Master's thesis, METU, 2010. http://etd.lib.metu.edu.tr/upload/12612600/index.pdf.
Full textAbstract:
Energy or DFT detector using a fixed window size is very efficient when signal start time and duration is matched with that of the window'<br>s. However, in the case of unknown signal duration, the performance of this detector decreases. For this scenario, a detector system composed of multiple windows may be preferred. Window sizes of such a system will also be fixed beforehand but they will be different from each other. Therefore, one of the windows will better match the signal duration, giving better detection results. In this study, multiple window detectors are analyzed. Their false alarm and detection probability relations are investigated. Some exact and approximate values are derived for these probabilities. A rule of thumb for the choice of window lengths is suggested for the case of fixed number of windows. Detectors with overlapping window structure are considered for the signals with unknown delay. Simulation results are added for these types of detectors.
2
Gonzalez-Garcia, Abel. "Image context for object detection, object context for part detection." Thesis, University of Edinburgh, 2018. http://hdl.handle.net/1842/28842.
Full textAbstract:
Objects and parts are crucial elements for achieving automatic image understanding. The goal of the object detection task is to recognize and localize all the objects in an image. Similarly, semantic part detection attempts to recognize and localize the object parts. This thesis proposes four contributions. The first two make object detection more efficient by using active search strategies guided by image context. The last two involve parts. One of them explores the emergence of parts in neural networks trained for object detection, whereas the other improves on part detection by adding object context. First, we present an active search strategy for efficient object class detection. Modern object detectors evaluate a large set of windows using a window classifier. Instead, our search sequentially chooses what window to evaluate next based on all the information gathered before. This results in a significant reduction on the number of necessary window evaluations to detect the objects in the image. We guide our search strategy using image context and the score of the classifier. In our second contribution, we extend this active search to jointly detect pairs of object classes that appear close in the image, exploiting the valuable information that one class can provide about the location of the other. This leads to an even further reduction on the number of necessary evaluations for the smaller, more challenging classes. In the third contribution of this thesis, we study whether semantic parts emerge in Convolutional Neural Networks trained for different visual recognition tasks, especially object detection. We perform two quantitative analyses that provide a deeper understanding of their internal representation by investigating the responses of the network filters. Moreover, we explore several connections between discriminative power and semantics, which provides further insights on the role of semantic parts in the network. Finally, the last contribution is a part detection approach that exploits object context. We complement part appearance with the object appearance, its class, and the expected relative location of the parts inside it. We significantly outperform approaches that use part appearance alone in this challenging task.
3
Vajhala, Rohith, Rohith Maddineni, and Preethi Raj Yeruva. "Weapon Detection In Surveillance Camera Images." Thesis, Blekinge Tekniska Högskola, Institutionen för tillämpad signalbehandling, 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-13565.
Full textAbstract:
Now a days, Closed Circuit Television (CCTV) cameras are installedeverywhere in public places to monitor illegal activities like armedrobberies. Mostly CCTV footages are used as post evidence after theoccurrence of crime. In many cases a person might be monitoringthe scene from CCTV but the attention can easily drift on prolongedobservation. Eciency of CCTV surveillance can be improved by in-corporation of image processing and object detection algorithms intomonitoring process.The object detection algorithms, previously implemented in CCTVvideo analysis detect pedestrians, animals and vehicles. These algo-rithms can be extended further to detect a person holding weaponslike rearms or sharp objects like knives in public or restricted places.In this work the detection of weapon from CCTV frame is acquiredby using Histogram of Oriented Gradients (HOG) as feature vector andarticial neural networks performing back-propagation algorithm forclassication.As a weapon in the hands of a human is considered to be greaterthreat as compared to a weapon alone, in this work the detection ofhuman in an image prior to a weapon detection has been found advan-tageous. Weapon detection has been performed using three methods.In the rst method, the weapon in the image is detected directly with-out human detection. Second and third methods use HOG and back-ground subtraction methods for detection of human prior to detectionof a weapon. A knife and a gun are considered as weapons of inter-est in this work. The performance of the proposed detection methodswas analysed on test image dataset containing knives, guns and im-ages without weapon. The accuracy rate 84:6% has been achievedby a single-class classier for knife detection. A gun and a knife havebeen detected by the three-class classier with an accuracy rate 83:0%.
4
Zhao, Meng John. "Analysis and Evaluation of Social Network Anomaly Detection." Diss., Virginia Tech, 2017. http://hdl.handle.net/10919/79849.
Full textAbstract:
As social networks become more prevalent, there is significant interest in studying these network data, the focus often being on detecting anomalous events. This area of research is referred to as social network surveillance or social network change detection. While there are a variety of proposed methods suitable for different monitoring situations, two important issues have yet to be completely addressed in network surveillance literature. First, performance assessments using simulated data to evaluate the statistical performance of a particular method. Second, the study of aggregated data in social network surveillance. The research presented tackle these issues in two parts, evaluation of a popular anomaly detection method and investigation of the effects of different aggregation levels on network anomaly detection.<br>Ph. D.
5
Shakil, Sadia. "Windowing effects and adaptive change point detection of dynamic functional connectivity in the brain." Diss., Georgia Institute of Technology, 2016. http://hdl.handle.net/1853/55006.
Full textAbstract:
Evidence of networks in the resting-brain reflecting the spontaneous brain activity is perhaps the most significant discovery to understand intrinsic brain functionality. Moreover, subsequent detection of dynamics in these networks can be milestone in differentiating the normal and disordered brain functions. However, capturing the correct dynamics is a challenging task since no ground truths' are present for comparison of the results. The change points of these networks can be different for different subjects even during normal brain functions. Even for the same subject and session, dynamics can be different at the start and end of the session based on the fatigue level of the subject scanned. Despite the absence of ground truths, studies have analyzed these dynamics using the existing methods and some of them have developed new algorithms too. One of the most commonly used method for this purpose is sliding window correlation. However, the result of the sliding window correlation is dependent on many parameters and without the ground truth there is no way of validating the results. In addition, most of the new algorithms are complicated, computationally expensive, and/or focus on just one aspect on these dynamics. This study applies the algorithms and concepts from signal processing, image processing, video processing, information theory, and machine learning to analyze the results of the sliding window correlation and develops a novel algorithm to detect change points of these networks adaptively. The findings in this study are divided into three parts: 1) Analyzing the extent of variability in well-defined networks of rodents and humans with sliding window correlation applying concepts from information theory and machine learning domains. 2) Analyzing the performance of sliding window correlation using simulated networks as ground truths for best parameters’ selection, and exploring its dependence on multiple frequency components of the correlating signals by processing the signals in time and Fourier domains. 3) Development of a novel algorithm based on image similarity measures from image and video processing that maybe employed to identify change points of these networks adaptively.
6
Sean, Viseth. "Exploration Framework For Detecting Outliers In Data Streams." Digital WPI, 2016. https://digitalcommons.wpi.edu/etd-theses/395.
Full textAbstract:
Current real-world applications are generating a large volume of datasets that are often continuously updated over time. Detecting outliers on such evolving datasets requires us to continuously update the result. Furthermore, the response time is very important for these time critical applications. This is challenging. First, the algorithm is complex; even mining outliers from a static dataset once is already very expensive. Second, users need to specify input parameters to approach the true outliers. While the number of parameters is large, using a trial and error approach online would be not only impractical and expensive but also tedious for the analysts. Worst yet, since the dataset is changing, the best parameter will need to be updated to respond to user exploration requests. Overall, the large number of parameter settings and evolving datasets make the problem of efficiently mining outliers from dynamic datasets very challenging. Thus, in this thesis, we design an exploration framework for detecting outliers in data streams, called EFO, which enables analysts to continuously explore anomalies in dynamic datasets. EFO is a continuous lightweight preprocessing framework. EFO embraces two optimization principles namely "best life expectancy" and "minimal trial," to compress evolving datasets into a knowledge-rich abstraction of important interrelationships among data. An incremental sorting technique is also used to leverage the almost ordered lists in this framework. Thereafter, the knowledge abstraction generated by EFO not only supports traditional outlier detection requests but also novel outlier exploration operations on evolving datasets. Our experimental study conducted on two real datasets demonstrates that EFO outperforms state-of-the-art technique in terms of CPU processing costs when varying stream volume, velocity and outlier rate.
7
Ahmed, Ejaz. "Monitoring and analysis of internet traffic targeting unused address spaces." Thesis, Queensland University of Technology, 2010. https://eprints.qut.edu.au/34075/1/Ejaz_Ahmed_Thesis.pdf.
Full textAbstract:
Today’s evolving networks are experiencing a large number of different attacks ranging from system break-ins, infection from automatic attack tools such as worms, viruses, trojan horses and denial of service (DoS). One important aspect of such attacks is that they are often indiscriminate and target Internet addresses without regard to whether they are bona fide allocated or not. Due to the absence of any advertised host services the traffic observed on unused IP addresses is by definition unsolicited and likely to be either opportunistic or malicious. The analysis of large repositories of such traffic can be used to extract useful information about both ongoing and new attack patterns and unearth unusual attack behaviors. However, such an analysis is difficult due to the size and nature of the collected traffic on unused address spaces. In this dissertation, we present a network traffic analysis technique which uses traffic collected from unused address spaces and relies on the statistical properties of the collected traffic, in order to accurately and quickly detect new and ongoing network anomalies. Detection of network anomalies is based on the concept that an anomalous activity usually transforms the network parameters in such a way that their statistical properties no longer remain constant, resulting in abrupt changes. In this dissertation, we use sequential analysis techniques to identify changes in the behavior of network traffic targeting unused address spaces to unveil both ongoing and new attack patterns. Specifically, we have developed a dynamic sliding window based non-parametric cumulative sum change detection techniques for identification of changes in network traffic. Furthermore we have introduced dynamic thresholds to detect changes in network traffic behavior and also detect when a particular change has ended. Experimental results are presented that demonstrate the operational effectiveness and efficiency of the proposed approach, using both synthetically generated datasets and real network traces collected from a dedicated block of unused IP addresses.
8
Hradiš, Michal. "Sdílení lokální informace pro rychlejší detekci objektů." Doctoral thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2014. http://www.nusl.cz/ntk/nusl-261243.
Full textAbstract:
Cílem této dizertační prace je vylepšit existující detektory objektů pomocí sdílení informace a výpočtů mezi blízkými pozicemi v obraze. Navrhuje dvě metody, které jsou založené na Waldově sekvenčním testu poměrem pravděpodobností a algoritmu WaldBoost. První z nich, Early non-Maxima Suppression , přesunuje rozhodování o potlačení nemaximálních pozic ze závěrečné fáze do fáze vyhodnocování detektoru, čímž zamezuje zbytečným výpočtům detektoru v nemaximálních pozicích. Metoda neighborhood suppression doplňuje existující detektory o schopnost zavrhnout okolní pozice v obraze. Navržené metody je možné aplikovat na širokou škálu detektorů. Vyhodnocení obou metod dokazují jejich výrazně vyšší efektivitu v porovnání s detektory, které vyhodnocují jednotlivé pozice obrazu zvlášť. Dizertace navíc prezentuje výsledky rozsáhlých experimentů, jejichž cílem bylo vyhodnotit vlastnosti běžných obrazových příznaků v několika detekčních úlohách a situacích.
9
Dybäck, Matilda, and Johanna Wallgren. "Pupil dilation as an indicator for auditory signal detection : Towards an objective hearing test based on eye tracking." Thesis, KTH, Skolan för teknik och hälsa (STH), 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-192703.
Full textAbstract:
An early detection of hearing loss in children is important for the child's speech and language development. For children between 3-6 months, a reliable method to measure hearing and determine hearing thresholds is missing. A hearing test based on the pupillary response to auditory signal detection as measured by eye tracking is based on an automatic physiological response. This hearing test could be used instead of the objective hearing tests used today. The presence of pupillary response has been shown in response to speech, but it is unstudied in response to sinus tones. The objective of this thesis was to study whether there is a consistent pupillary response to different sinus tone frequencies commonly used in hearing tests and if yes, to determine reliably the time window of this response. Four different tests were done. The adult pupillary response in regard to sinus tone stimuli with four frequency levels (500 Hz, 1000 Hz, 2000 Hz and 4000 Hz), and four loudness levels (silence, 30 dB, 50 dB and 70 dB) was tested (N=20, 15 females, 5 males). Different brightness levels and distractions on the eye tracking screen were investigated in three substudies (N=5, 4 females, 1 male). Differences between silence and loudness levels within frequency levels were tested for statistical significance. A pupillary response in regard to sinus tones occurred consistently between 300 ms and 2000 ms with individual variation, i.e. earlier than for speech sounds. Differences between silence and loudness levels were only statistically significant for 4000 Hz. No statistical difference was shown between different brightness levels or if there were distractions present on the eye tracker screen. The conclusion is that pupillary response to pure sinus tones in adults is a possible measure of hearing threshold for at least 4000 Hz. Larger studies are needed to confirm this, and also to more thoroughly investigate the other frequencies.<br>En tidig upptäckt av hörselnedsättning hos barn är viktig för barnets tal- och språkutveckling. För barn mellan 3-6 månader saknas det en tillförlitlig metod för att mäta hörsel och bestämma hörtrösklar. Ett hörseltest baserad på pupillreaktion på ljud som mäts med en eye tracker bygger på en automatisk fysiologisk reaktion och skulle kunna användas istället för de objektiva test som används idag. Hitintills har pupillreaktion på tal påvisats, men det saknas studier som studerat eventuella reaktioner på sinustoner. Syftet med denna uppsats var att undersöka om det finns en enhetlig pupillreaktion på de olika frekvenserna av sinustoner som vanligen används i hörseltest. Vidare var studiens syfte att fastställa ett tillförlitligt tidsfönster för pupillreaktion. Fyra olika typer av tester utfördes. Pupillreaktionen mot sinustoner med fyra olika frekvensnivåer (500 Hz, 1000 Hz, 2000 Hz och 4000 Hz), och fyra olika ljudnivåer (tystnad, 30 dB, 50 dB och 70 dB) undersöktes i ett test på vuxna deltagare (N=20, 15 kvinnor, 5 män). Olika ljusnivåer och distraktioner på eye tracker-skärmen undersöktes i tre test (N=5, 4 kvinnor, 1 man). Skillnaderna mellan ljudnivåer och frekvensnivåer testades med statistiska tester. Resultaten visade att pupillreaktion på sinustoner inträffade konsekvent mellan 300 ms och 2000 ms med individuella variationer. Denna reaktionstid inträffar tidigare än för taljud. En statistisk signifikant skillnad mellan tystnad och olika ljudnivåer kunde endast ses för frekvensnivån 4000 Hz. Ingen statistisk skillnad uppmättes mellan olika ljudnivåer eller om det fanns distraktioner på eye tracker-skärmen. De i studien framkomna resultaten tyder på att pupillreaktioner mot rena sinustoner hos vuxna är en möjlig metod för att identifiera hörseltrösklar för åtminstone 4000 Hz. Större studier behöver göras för att fastställa detta och en noggrannare undersökning behöver genomföras för de andra frekvenserna.
10
Rankine, Luke. "Newborn EEG seizure detection using adaptive time-frequency signal processing." Thesis, Queensland University of Technology, 2006. https://eprints.qut.edu.au/16200/1/Luke_Rankine_Thesis.pdf.
Full textAbstract:
Dysfunction in the central nervous system of the neonate is often first identified through seizures. The diffculty in detecting clinical seizures, which involves the observation of physical manifestations characteristic to newborn seizure, has placed greater emphasis on the detection of newborn electroencephalographic (EEG) seizure. The high incidence of newborn seizure has resulted in considerable mortality and morbidity rates in the neonate. Accurate and rapid diagnosis of neonatal seizure is essential for proper treatment and therapy. This has impelled researchers to investigate possible methods for the automatic detection of newborn EEG seizure. This thesis is focused on the development of algorithms for the automatic detection of newborn EEG seizure using adaptive time-frequency signal processing.
The assessment of newborn EEG seizure detection algorithms requires large datasets of nonseizure and seizure EEG which are not always readily available and often hard to acquire. This has led to the proposition of realistic models of newborn EEG which can be used to create large datasets for the evaluation and comparison of newborn EEG seizure detection algorithms. In this thesis, we develop two simulation methods which produce synthetic newborn EEG background and seizure. The simulation methods use nonlinear and time-frequency signal processing techniques to allow for the demonstrated nonlinear and nonstationary characteristics of the newborn EEG.
Atomic decomposition techniques incorporating redundant time-frequency dictionaries are exciting new signal processing methods which deliver adaptive signal representations or approximations. In this thesis we have investigated two prominent atomic decomposition techniques, matching pursuit and basis pursuit, for their possible use in an automatic seizure detection algorithm. In our investigation, it was shown that matching pursuit generally provided the sparsest (i.e. most compact) approximation for various real and synthetic signals over a wide range of signal approximation levels. For this reason, we chose MP as our preferred atomic decomposition technique for this thesis.
A new measure, referred to as structural complexity, which quantifes the level or degree of correlation between signal structures and the decomposition dictionary was proposed. Using the change in structural complexity, a generic method of detecting changes in signal structure was proposed. This detection methodology was then applied to the newborn EEG for the detection of state transition (i.e. nonseizure to seizure state) in the EEG signal. To optimize the seizure detection process, we developed a time-frequency dictionary that is coherent with the newborn EEG seizure state based on the time-frequency analysis of the newborn EEG seizure. It was shown that using the new coherent time-frequency dictionary and the change in structural complexity, we can detect the transition from nonseizure to seizure states in synthetic and real newborn EEG.
Repetitive spiking in the EEG is a classic feature of newborn EEG seizure. Therefore, the automatic detection of spikes can be fundamental in the detection of newborn EEG seizure. The capacity of two adaptive time-frequency signal processing techniques to detect spikes was investigated. It was shown that a relationship between the EEG epoch length and the number of repetitive spikes governs the ability of both matching pursuit and adaptive spectrogram in detecting repetitive spikes. However, it was demonstrated that the law was less restrictive forth eadaptive spectrogram and it was shown to outperform matching pursuit in detecting repetitive spikes.
The method of adapting the window length associated with the adaptive spectrogram used in this thesis was the maximum correlation criterion. It was observed that for the time instants where signal spikes occurred, the optimal window lengths selected by the maximum correlation criterion were small. Therefore, spike detection directly from the adaptive window optimization method was demonstrated and also shown to outperform matching pursuit.
An automatic newborn EEG seizure detection algorithm was proposed based on the detection of repetitive spikes using the adaptive window optimization method. The algorithm shows excellent performance with real EEG data. A comparison of the proposed algorithm with four well documented newborn EEG seizure detection algorithms is provided. The results of the comparison show that the proposed algorithm has significantly better performance than the existing algorithms (i.e. Our proposed algorithm achieved a good detection rate (GDR) of 94% and false detection rate (FDR) of 2.3% compared with the leading algorithm which only produced a GDR of 62% and FDR of 16%).
In summary, the novel contribution of this thesis to the fields of time-frequency signal processing and biomedical engineering is the successful development and application of sophisticated algorithms based on adaptive time-frequency signal processing techniques to the solution of automatic newborn EEG seizure detection.
11
Rankine, Luke. "Newborn EEG seizure detection using adaptive time-frequency signal processing." Queensland University of Technology, 2006. http://eprints.qut.edu.au/16200/.
Full textAbstract:
Dysfunction in the central nervous system of the neonate is often first identified through seizures. The diffculty in detecting clinical seizures, which involves the observation of physical manifestations characteristic to newborn seizure, has placed greater emphasis on the detection of newborn electroencephalographic (EEG) seizure. The high incidence of newborn seizure has resulted in considerable mortality and morbidity rates in the neonate. Accurate and rapid diagnosis of neonatal seizure is essential for proper treatment and therapy. This has impelled researchers to investigate possible methods for the automatic detection of newborn EEG seizure. This thesis is focused on the development of algorithms for the automatic detection of newborn EEG seizure using adaptive time-frequency signal processing. The assessment of newborn EEG seizure detection algorithms requires large datasets of nonseizure and seizure EEG which are not always readily available and often hard to acquire. This has led to the proposition of realistic models of newborn EEG which can be used to create large datasets for the evaluation and comparison of newborn EEG seizure detection algorithms. In this thesis, we develop two simulation methods which produce synthetic newborn EEG background and seizure. The simulation methods use nonlinear and time-frequency signal processing techniques to allow for the demonstrated nonlinear and nonstationary characteristics of the newborn EEG. Atomic decomposition techniques incorporating redundant time-frequency dictionaries are exciting new signal processing methods which deliver adaptive signal representations or approximations. In this thesis we have investigated two prominent atomic decomposition techniques, matching pursuit and basis pursuit, for their possible use in an automatic seizure detection algorithm. In our investigation, it was shown that matching pursuit generally provided the sparsest (i.e. most compact) approximation for various real and synthetic signals over a wide range of signal approximation levels. For this reason, we chose MP as our preferred atomic decomposition technique for this thesis. A new measure, referred to as structural complexity, which quantifes the level or degree of correlation between signal structures and the decomposition dictionary was proposed. Using the change in structural complexity, a generic method of detecting changes in signal structure was proposed. This detection methodology was then applied to the newborn EEG for the detection of state transition (i.e. nonseizure to seizure state) in the EEG signal. To optimize the seizure detection process, we developed a time-frequency dictionary that is coherent with the newborn EEG seizure state based on the time-frequency analysis of the newborn EEG seizure. It was shown that using the new coherent time-frequency dictionary and the change in structural complexity, we can detect the transition from nonseizure to seizure states in synthetic and real newborn EEG. Repetitive spiking in the EEG is a classic feature of newborn EEG seizure. Therefore, the automatic detection of spikes can be fundamental in the detection of newborn EEG seizure. The capacity of two adaptive time-frequency signal processing techniques to detect spikes was investigated. It was shown that a relationship between the EEG epoch length and the number of repetitive spikes governs the ability of both matching pursuit and adaptive spectrogram in detecting repetitive spikes. However, it was demonstrated that the law was less restrictive forth eadaptive spectrogram and it was shown to outperform matching pursuit in detecting repetitive spikes. The method of adapting the window length associated with the adaptive spectrogram used in this thesis was the maximum correlation criterion. It was observed that for the time instants where signal spikes occurred, the optimal window lengths selected by the maximum correlation criterion were small. Therefore, spike detection directly from the adaptive window optimization method was demonstrated and also shown to outperform matching pursuit. An automatic newborn EEG seizure detection algorithm was proposed based on the detection of repetitive spikes using the adaptive window optimization method. The algorithm shows excellent performance with real EEG data. A comparison of the proposed algorithm with four well documented newborn EEG seizure detection algorithms is provided. The results of the comparison show that the proposed algorithm has significantly better performance than the existing algorithms (i.e. Our proposed algorithm achieved a good detection rate (GDR) of 94% and false detection rate (FDR) of 2.3% compared with the leading algorithm which only produced a GDR of 62% and FDR of 16%). In summary, the novel contribution of this thesis to the fields of time-frequency signal processing and biomedical engineering is the successful development and application of sophisticated algorithms based on adaptive time-frequency signal processing techniques to the solution of automatic newborn EEG seizure detection.
12
Pesaranghader, Ali. "A Reservoir of Adaptive Algorithms for Online Learning from Evolving Data Streams." Thesis, Université d'Ottawa / University of Ottawa, 2018. http://hdl.handle.net/10393/38190.
Full textAbstract:
Continuous change and development are essential aspects of evolving environments and applications, including, but not limited to, smart cities, military, medicine, nuclear reactors, self-driving cars, aviation, and aerospace. That is, the fundamental characteristics of such environments may evolve, and so cause dangerous consequences, e.g., putting people lives at stake, if no reaction is adopted. Therefore, learning systems need to apply intelligent algorithms to monitor evolvement in their environments and update themselves effectively. Further, we may experience fluctuations regarding the performance of learning algorithms due to the nature of incoming data as it continuously evolves. That is, the current efficient learning approach may become deprecated after a change in data or environment. Hence, the question 'how to have an efficient learning algorithm over time against evolving data?' has to be addressed. In this thesis, we have made two contributions to settle the challenges described above.
In the machine learning literature, the phenomenon of (distributional) change in data is known as concept drift. Concept drift may shift decision boundaries, and cause a decline in accuracy. Learning algorithms, indeed, have to detect concept drift in evolving data streams and replace their predictive models accordingly. To address this challenge, adaptive learners have been devised which may utilize drift detection methods to locate the drift points in dynamic and changing data streams. A drift detection method able to discover the drift points quickly, with the lowest false positive and false negative rates, is preferred. False positive refers to incorrectly alarming for concept drift, and false negative refers to not alarming for concept drift. In this thesis, we introduce three algorithms, called as the Fast Hoeffding Drift Detection Method (FHDDM), the Stacking Fast Hoeffding Drift Detection Method (FHDDMS), and the McDiarmid Drift Detection Methods (MDDMs), for detecting drift points with the minimum delay, false positive, and false negative rates. FHDDM is a sliding window-based algorithm and applies Hoeffding’s inequality (Hoeffding, 1963) to detect concept drift. FHDDM slides its window over the prediction results, which are either 1 (for a correct prediction) or 0 (for a wrong prediction). Meanwhile, it compares the mean of elements inside the window with the maximum mean observed so far; subsequently, a significant difference between the two means, upper-bounded by the Hoeffding inequality, indicates the occurrence of concept drift. The FHDDMS extends the FHDDM algorithm by sliding multiple windows over its entries for a better drift detection regarding the detection delay and false negative rate. In contrast to FHDDM/S, the MDDM variants assign weights to their entries, i.e., higher weights are associated with the most recent entries in the sliding window, for faster detection of concept drift. The rationale is that recent examples reflect the ongoing situation adequately. Then, by putting higher weights on the latest entries, we may detect concept drift quickly. An MDDM algorithm bounds the difference between the weighted mean of elements in the sliding window and the maximum weighted mean seen so far, using McDiarmid’s inequality (McDiarmid, 1989). Eventually, it alarms for concept drift once a significant difference is experienced. We experimentally show that FHDDM/S and MDDMs outperform the state-of-the-art by representing promising results in terms of the adaptation and classification measures.
Due to the evolving nature of data streams, the performance of an adaptive learner, which is defined by the classification, adaptation, and resource consumption measures, may fluctuate over time. In fact, a learning algorithm, in the form of a (classifier, detector) pair, may present a significant performance before a concept drift point, but not after. We define this problem by the question 'how can we ensure that an efficient classifier-detector pair is present at any time in an evolving environment?' To answer this, we have developed the Tornado framework which runs various kinds of learning algorithms simultaneously against evolving data streams. Each algorithm incrementally and independently trains a predictive model and updates the statistics of its drift detector. Meanwhile, our framework monitors the (classifier, detector) pairs, and recommends the efficient one, concerning the classification, adaptation, and resource consumption performance, to the user. We further define the holistic CAR measure that integrates the classification, adaptation, and resource consumption measures for evaluating the performance of adaptive learning algorithms. Our experiments confirm that the most efficient algorithm may differ over time because of the developing and evolving nature of data streams.
13
Pavlíček, Tomáš. "Segmentace pro časově-variantní systémy a jejich implementace." Master's thesis, Vysoké učení technické v Brně. Fakulta elektrotechniky a komunikačních technologií, 2014. http://www.nusl.cz/ntk/nusl-220605.
Full textAbstract:
This thesis is interested in describing stationary random discrete signals, especially) in music discrete signals. Here is described when is signal stationary and when is not stationary. It contains tip for preprocessing of signal for accurate recognition of local stationarity. Thesis contain mathematical definition of parameters of random digital signals, which are used for stationarity recognition. It is followed by description of basic windows, their categories, describing of their parameters and comparing of each. In next part of thesis are described mothods of segmentations with constant window constant overlap save, constant window constant ovelap add, variable window constant overlap save, variable window constant ovelap add and variable window variable overlap add. It is followed by analyzing of windows used in segmentations with variable lengths of segments. As next point of thesis are transients made by step changes of coefficients of filter in filtering of segments with variable lengths. At the end is investigated the best accurate method of signal stationarity detection. Segments made by accurate method of detection are analyzed. thesis contains exapmle of music signal segmentation.
14
Draisbach, Uwe, Felix Naumann, Sascha Szott, and Oliver Wonneberg. "Adaptive windows for duplicate detection." Universität Potsdam, 2012. http://opus.kobv.de/ubp/volltexte/2012/5300/.
Full textAbstract:
Duplicate detection is the task of identifying all groups of records within a data set that represent the same real-world entity, respectively. This task is difficult, because (i) representations might differ slightly, so some similarity measure must be defined to compare pairs of records and (ii) data sets might have a high volume making a pair-wise comparison of all records infeasible. To tackle the second problem, many algorithms have been suggested that partition the data set and compare all record pairs only within each partition. One well-known such approach is the Sorted Neighborhood Method (SNM), which sorts the data according to some key and then advances a window over the data comparing only records that appear within the same window.
We propose several variations of SNM that have in common a varying window size and advancement. The general intuition of such adaptive windows is that there might be regions of high similarity suggesting a larger window size and regions of lower similarity suggesting a smaller window size. We propose and thoroughly evaluate several adaption strategies, some of which are provably better than the original SNM in terms of efficiency (same results with fewer comparisons).<br>Duplikaterkennung beschreibt das Auffinden von mehreren Datensätzen, die das gleiche Realwelt-Objekt repräsentieren. Diese Aufgabe ist nicht trivial, da sich (i) die Datensätze geringfügig unterscheiden können, so dass Ähnlichkeitsmaße für einen paarweisen Vergleich benötigt werden, und (ii) aufgrund der Datenmenge ein vollständiger, paarweiser Vergleich nicht möglich ist. Zur Lösung des zweiten Problems existieren verschiedene Algorithmen, die die Datenmenge partitionieren und nur noch innerhalb der Partitionen Vergleiche durchführen. Einer dieser Algorithmen ist die Sorted-Neighborhood-Methode (SNM), welche Daten anhand eines Schlüssels sortiert und dann ein Fenster über die sortierten Daten schiebt. Vergleiche werden nur innerhalb dieses Fensters durchgeführt.
Wir beschreiben verschiedene Variationen der Sorted-Neighborhood-Methode, die auf variierenden Fenstergrößen basieren. Diese Ansätze basieren auf der Intuition, dass Bereiche mit größerer und geringerer Ähnlichkeiten innerhalb der sortierten Datensätze existieren, für die entsprechend größere bzw. kleinere Fenstergrößen sinnvoll sind. Wir beschreiben und evaluieren verschiedene Adaptierungs-Strategien, von denen nachweislich einige bezüglich Effizienz besser sind als die originale Sorted-Neighborhood-Methode (gleiches Ergebnis bei weniger Vergleichen).
15
Дрозд, В. П. "Застосування гістограми орієнтованих градієнтів (HOG) для виявлення пішохода на зображенні". Thesis, Сумський державний університет, 2014. http://essuir.sumdu.edu.ua/handle/123456789/39124.
Full textAbstract:
Проблема виявлення пішохода полягає в тому, що люди дуже різноманітні за статурою та можуть приймати різні пози, у зображення можуть бути різні спотворення. Існує ряд методів для виявлення
пішохода: методи основані на Haar wavelet признаках, нейронні мережі, гістограми направлених градієнтів та інші. В даній роботі пропонується розгляд варіанту застосування HOG.
16
Špaňhel, Jakub. "Re-identifikace vozidla pomocí rozpoznání jeho registrační značky." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2015. http://www.nusl.cz/ntk/nusl-264932.
Full textAbstract:
This thesis aims at proposing vehicle license plate detection and recognition algorithms, suitable for vehicle re-identification. Simple urban traffic analysis system is also proposed. Multiple stages of this system was developed and tested. Specifically - vehicle detection, license plate detection and recognition. Vehicle detection is based on background substraction method, which results in an average hit rate of ~92%. License plate detection is done by cascade classifiers and achieves an average hit rate of 81.92% and precision rate of 94.42%. License plate recognition based on Template matching results in an average precission rate of 60.55%. Therefore the new license plate recognition method based on license plate scanning using the sliding window principle and neural network recognition was introduced. Neural network achieves a precision rate of 64.47% for five input features. Low precision rate of neural network is caused by small amount of training sample for some specific license plate characters.
17
Korba, Jonathan (Jonathan James) 1977. "Windows NT attacks for the evaluation of intrusion detection systems." Thesis, Massachusetts Institute of Technology, 2000. http://hdl.handle.net/1721.1/86454.
Full textAbstract:
Thesis (S.B. and M.Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2000.<br>Includes bibliographical references (leaves 99-101).<br>by Jonathan Korba.<br>S.B.and M.Eng.
18
Ito, Stephen Hiroyuki. "The effects of gaze-contingent windows on detection of peripheral targets." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 1997. http://www.collectionscanada.ca/obj/s4/f2/dsk2/ftp03/MQ28730.pdf.
Full text
19
Ganesan, Balaji. "TCP/IP stack fingerprinting for patch detection in a distributed Windows environment." Morgantown, W. Va. : [West Virginia University Libraries], 2004. https://etd.wvu.edu/etd/controller.jsp?moduleName=documentdata&jsp%5FetdId=3510.
Full textAbstract:
Thesis (M.S.)--West Virginia University, 2004.<br>Title from document title page. Document formatted into pages; contains ix, 109 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 56-58).
20
Řehánek, Martin. "Detekce objektů pomocí Kinectu." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2012. http://www.nusl.cz/ntk/nusl-236602.
Full textAbstract:
With the release of the Kinect device new possibilities appeared, allowing a simple use of image depth in image processing. The aim of this thesis is to propose a method for object detection and recognition in a depth map. Well known method Bag of Words and a descriptor based on Spin Image method are used for the object recognition. The Spin Image method is one of several existing approaches to depth map which are described in this thesis. Detection of object in picture is ensured by the sliding window technique. That is improved and speeded up by utilization of the depth information.
21
Ayichiluhm, Theodros, and Vivek Mohan. "IPv6 Monitoring and Flow Detection." Thesis, Blekinge Tekniska Högskola, Sektionen för datavetenskap och kommunikation, 2013. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-4165.
Full textAbstract:
IPv6 Privacy extensions, implemented in major operating systems, hide user’s identity by using a temporary and a randomly generated IPv6 addresses rather than using the former, EUI-64 format where the MAC address is part of the IPv6 address. This solution for privacy has created a problem for network administrators to back-trace an IPv6 address to a specific MAC address, since the temporary IP address used once by the node is removed from the interface after a period of time. An IPv6 Ethernet test bed is setup to investigate IPv6 implementation dynamics in Windows 7 and Ubuntu10.04 operating systems. The testbed is extended to investigate the effects of temporary IPv6 addresses due to IPv6 privacy extensions on the on-going sessions of different applications including ping, File Transfer Protocol (FTP) and video streaming (HTTP and RTP). On the basis of the knowledge obtained from investigations about dynamics of IPv6 privacy extensions, this work proposes Internet Protocol version 6 Host Tracking (IPv6HoT), a web based IPv6 to MAC mapping solution. IPv6HoT uses Simple Network Management Protocol (SNMP) to forward IPv6 Neighbor table from routers to Network Management Stations (NMS). This thesis work provides guidelines for configuring IPv6 privacy extensions in Ubuntu10.04 and Windows 7; the difference of implementation between these two operating systems is also presented in this work. The results show that temporary IPv6 addressing has a definite effect on the on-going sessions of video streaming and FTP applications. Applications running as server on Temporary IPv6 address encountered more frequent on-going session interruptions than applications running as a server over public IPv6 address. When temporary IPv6 addresses were configured to host FTP and video streaming applications, their on-going sessions were permanently interrupted. It is also observed that LFTP, a client FTP application, resumes an interrupted session.<br>theodrosmek11@gmail.com, iamvivek86@gmail.com
22
Kremer, H. Steven. "Real-time intrusion detection for Windows NT based on Navy IT-21 audit policy." Thesis, Monterey, Calif. : Springfield, Va. : Naval Postgraduate School ; Available from National Technical Information Service, 1999. http://handle.dtic.mil/100.2/ADA378151.
Full textAbstract:
Thesis (M.S. in Software Engineering) Naval Postgraduate School, September 1999.<br>"September 1999". Thesis advisors(s): Neil C. Rowe, Ronald Broersma. Includes bibliographical references (p. 49). Also available online.
23
Ronquist, Anton, and Birger Winroth. "Estimation and Compensation of Load-Dependent Position Error in a Hybrid Stepper Motor." Thesis, Linköpings universitet, Reglerteknik, 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-129554.
Full textAbstract:
Hybrid stepper motors are a common type of electric motor used throughout industry thanks to its low-cost, high torque at low speed and open loop positioning capabilities. However, a closed loop control is often required for industrial applications with high precision requirements. The closed loop control can also be used to lower the power consumption of the motor and ensure that stalls are avoided. It is quite common to utilise a large and costly position encoder or resolver to feedback the position signal to the control logic. This thesis has explored the possibility of using a low-cost position sensor based on Hall elements. Additionally, a sensorless estimation algorithm, using only stator winding measurements, has been investigated both as a competitive alternative and as a possible complement to the position sensor. The thesis work summarises and discusses previous research attempts to adequately measure or estimate and control the hybrid stepper motors position and load angle without using a typical encoder or resolver. Qualitative results have been produced through simulations prior to implementation and experimental testing. The readings from the position sensor is subject to noise, owing to its resolution and construction. The position signal has been successfully filtered, improving its accuracy from 0.56° to 0.25°. The output from the sensorless estimation algorithm is subject to non-linear errors caused by errors in phase voltage measurements and processing of velocity changes. However, the dynamics are reliable at constant speeds and could be used for position control.
24
Li, Jie, and Yuting Lu. "Rootkits." Thesis, Linnaeus University, School of Computer Science, Physics and Mathematics, 2010. http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-8378.
Full textAbstract:
<p>Abstract：The kernel system of Windows is more thoroughly exposed to people. So, thekernel-level Rootkits techniques are now laid on greater emphasis. It is very importantto maintain the security of computers and to conduct an in-depth research on theoperational mechanism by using kernel-level Rootkits in hiding its traces. Since theinvolved core techniques are beginning to catch on nowadays, we should analyzesome new key techniques employed for application of Rootkits, discuss the specificmethods and propose a set of defense strategy for computer security.</p>
25
White, Andrew J. "Identifying the unknown in user space memory." Thesis, Queensland University of Technology, 2013. https://eprints.qut.edu.au/64181/1/Andrew_White_Thesis.pdf.
Full textAbstract:
This thesis is a study of how the contents of volatile memory on the Windows operating system can be better understood and utilised for the purposes of digital forensic investigations. It proposes several techniques to improve the analysis of memory, with a focus on improving the detection of unknown code such as malware. These contributions allow the creation of a more complete reconstruction of the state of a computer at acquisition time, including whether or not the computer has been infected by malicious code.
26
Amzajerdian, Farzin. "Remote sensing of atmospheric winds by utilizing speckle-turbulence interaction and optical heterodyne detection /." Full text open access at:, 1988. http://content.ohsu.edu/u?/etd,170.
Full text
27
Kuz, Kadir. "Design And Implementation Of A Monitoring Framework." Master's thesis, METU, 2009. http://etd.lib.metu.edu.tr/upload/12610557/index.pdf.
Full textAbstract:
In this thesis work, the symptoms in Windows XP operating system for fault monitoring are investigated and a fault monitoring library is developed. A test GUI is implemented to examine this library. Performance tests including memory and CPU usage are done to see its overhead to the system and platform tests on the current version of Windows operating system series (Windows Vista) are done to
see for compatibility.
In this thesis, fault monitor-fault detector interface is also defined and implemented. To monitor a symptom that is not implemented in the monitoring library, projects can implement their own monitors. A monitoring framework is designed to control and coordinate these monitors with the main one. To create monitors for Java projects easily, a monitor creator library is developed.
28
Joseph, Liselle AnnMarie. "Transition Detection for Low Speed Wind Tunnel Testing Using Infrared Thermography." Thesis, Virginia Tech, 2014. http://hdl.handle.net/10919/78145.
Full textAbstract:
Transition is an important phenomenon in large scale, commercial, wind tunnel testing at low speeds because it is an excellent indicator of an airfoil performance. It is difficult to estimate transition through numerical techniques because of the complex nature of viscous flow. Therefore experimental techniques can be essential. Over the transition region the rate of heat transfer shows significant increases which can be detected using infrared thermography. This technique has been used predominantly at high speeds, on small models made of insulated materials, and for short test runs. Large scale testing has not been widely undertaken because the high sensitivity of transition to external factors makes it difficult to detect.
The present study records the process undertaken to develop, implement and validate a transition detection system for continual use in the Virginia Tech Stability Wind Tunnel: a low speed, commercial wind tunnel where large, aluminium models are tested. The final system developed comprises of two high resolution FLIR A655sc infrared cameras; four 63.5-mm diameter circular windows; aluminium models covered in 0.8-mm silicone rubber insulation and a top layer of ConTact© paper; and a series of 25.4-mm wide rubber silicone fiberglass insulated heaters mounted inside the model and controlled externally by experimenters. This system produces images or videos of the model and the associated transition location, which is later extracted through image processing methods to give a final transition location in percentage chord.
The system was validated using two DU96-W-180 airfoils of different chord lengths in the Virginia Tech Stability Wind Tunnel, each tested two months apart. The system proved to be robust and efficient, while not affecting the airfoil performance or any other system in use in the wind tunnel. Transition results produced by the system were compared to measurements obtained from pressure data and stethoscope tests as well as the numerical predictions of XFOIL. The transition results from all four methods showed excellent agreement with each other for the two models, for at least two Reynolds numbers and for several angles of attack on both suction and pressure side of the model. The agreement of data obtained under such different conditions and at different times suggests that the infrared thermography system efficiently and accurately detects transition for large aluminium models at low speeds.<br>Master of Science
29
Kuhner, Joseph T. "Automating the Detection of Precipitation and Wind Characteristics in Navy Ocean Acoustic Data." ScholarWorks@UNO, 2018. https://scholarworks.uno.edu/td/2567.
Full textAbstract:
A challenge in Underwater Acoustics is identifying the independent variables associated with an environment’s ambient noise. A strict definition of ambient noise would focus on non-transient signatures and exclude transient impacts from marine mammals, pelagic fish species, man-made sources, or weather events such as precipitation or wind speeds. Recognizing transient signatures in acoustic spectra is an essential element for providing environmental intelligence to the U.S. Navy, specifically the acoustic signatures from meteorological events. While weather event detection in acoustic spectra has been shown in previous studies, leveraging these concepts via U.S. Navy assets is largely an unknown. Environmental intelligence collection can be improved by detecting precipitation events and establishing wind velocities with acoustic signatures. This will further improve meteorological models by enabling validation from both manned and unmanned sub-surface assets.
30
Huška, Michal. "Automatické značkování prezentací." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2007. http://www.nusl.cz/ntk/nusl-412801.
Full textAbstract:
This thesis deals with a development of application running on mobile devices with Windows Mobile operating system. The main task of this application is observing canvas with running presentation and saving time marks, that inform about slide change or animation. Description of system requirements, system analysis using UML language, solutions on image processing level, decription of implementation in C++ language and application tests results are described in the text. Problems of mobile device software development are also outlined in the document. A great part is dedicated to work with multimedia on Windows Mobile 5.0 system, especially to problems linked with DirectShow technology.
31
Jain, Sharad. "Skidding and fault detection in the bearings of wind-turbine gearboxes." Thesis, University of Cambridge, 2013. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.608104.
Full text
32
Bhamidipati, Kanthi Latha. "Detection and elimination of defects during manufacture of high-temperature polymer electrolyte membranes." Diss., Georgia Institute of Technology, 2011. http://hdl.handle.net/1853/43616.
Full textAbstract:
Defect generation and propagation in thin films, such as separation membranes, can lead to premature or catastrophic failure of devices such as polymer electrolyte membrane fuel cells (PEMFC). It is hypothesized that defects (e.g., air bubbles, pin-holes, and holes) originate during the manufacturing stage, if precise control is not maintained over the coating process, and they propagate during system operation. Experimental and numerical studies were performed to detect and eliminate defects that were induced during slot die coating of high-viscosity (1 to 40 Pa-s), shear-thinning solutions. The effects of fluid properties, geometric parameters and processing conditions on air entrainment and coating windows (limited set of processing conditions for which defect-free coating exists) were studied. When smaller slot gaps and coating gaps were used, relatively small bubbles were entrained in the coated film. The air bubble sizes increased as the viscosity of the coating solution decreased. A semi-empirical model correlating the maximum coating speed to a solution's material properties, geometric parameters and processing conditions was developed. Such a predictive model will enable engineers to determine the maximum coating boundary for shear-thinning and Newtonian solutions within certain constraints. Smaller coating gaps and low-viscosity solutions produced higher coating speeds. The surface tension property of the coating solution provided stability to the coating bead. Therefore, solutions with higher surface tension could be processed at higher coating speeds.
33
Schillinger, Douglas J. "Wind speed estimates and precipitation detection using ambient sound in the ocean." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 2000. http://www.collectionscanada.ca/obj/s4/f2/dsk2/ftp01/MQ55540.pdf.
Full text
34
Ten, Chui Fen. "Loss of mains detection and amelioration on electrical distribution networks." Thesis, University of Manchester, 2011. https://www.research.manchester.ac.uk/portal/en/theses/loss-of-mains-detection-and-amelioration-on-electrical-distribution-networks(b7680a62-7caa-4fd3-89d4-d45e649f8ef9).html.
Full textAbstract:
Power system islanding is gaining increasing interest as a way to maintain power supply continuity. However, before this operation become viable, the technical challenges associated with its operation must first be addressed. A possible solution to one of these challenges, out-of synchronism reclosure, is by running the islanded system in synchronism with the mains whilst not being electrically connected. This concept, known as 'synchronous islanded operation' avoids the danger of out-of-synchronism reclosure of the islanded system onto the mains. The research in this thesis was based on the concepts presented in [1-3] and specifically applied to multiple-DG island scenarios. The additional control challenges associated with this scenario are identified and an appropriate control scheme, more suited for the operation of multiple-DG synchronous islands, is proposed. The results suggest that multiple-DG synchronous islanded operation is feasible, but a supervisory controller is necessary to facilitate the information exchange within the islanded system and enable stable operation.For maximum flexibility, the synchronous island must be capable of operating with a diversity of generation. The difficulties become further complicated when some or all of the generation consists of intermittent sources. The performance of the proposed control scheme in the presence of a significant contribution of renewable sources within the island is investigated. Two types of wind technologies were developed in PSCAD/EMTDC for this purpose, they are a fixed speed induction generator (FSIG) based wind farm and a doubly-fed induction generator (DFIG) based wind farm. The results show that although synchronous islanded operation is still achievable, the intermittent output has an adverse effect on the control performance, and in particular limits the magnitude of disturbances that can happen in the island without going beyond the relaxed synchronisation limits of ±60o.Energy storage is proposed as a way to reduce the wind farm power variation and improve phase controller response. A supplementary control is also proposed such that DFIG contributes to the inertial response. The potential of the proposed scheme (energy storage + supplementary control) is evaluated using case studies. The results show massive improvement to the load acceptance limits, even beyond the case where no wind farm is connected. The benefit of the proposed scheme is even more apparent as the share of wind generated energy in the island grows.
35
Lapira, Edzel R. "Fault Detection in a Network of Similar Machines using Clustering Approach." University of Cincinnati / OhioLINK, 2012. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1339250832.
Full text
36
Tutivén, Gálvez Christian. "Fault detection and fault tolerant control in wind turbines." Doctoral thesis, Universitat Politècnica de Catalunya, 2018. http://hdl.handle.net/10803/663289.
Full textAbstract:
Renewable energy is an important sustainable energy in the world. Up to now, as an essential part of low emissions energy in a lot of countries, renewable energy has been important to the national energy security, and played a significant role in reducing carbon emissions. It comes from natural resources, such as wind, solar, rain, tides, biomass, and geothermal heat. Among them, wind energy is rapidly emerging as a low carbon, resource efficient, cost effective sustainable technology in the world. Due to the demand of higher power production installations with less environmental impacts, the continuous increase in size of wind turbines and the recently developed offshore (floating) technologies have led to new challenges in the wind turbine systems.Wind turbines (WTs) are complex systems with large flexible structures that work under very turbulent and unpredictable environmental conditions for a variable electrical grid. The maximization of wind energy conversion systems, load reduction strategies, mechanical fatigue minimization problems, costs per kilowatt hour reduction strategies, reliability matters, stability problems, and availability (sustainability) aspects demand the use of advanced (multivariable and multiobjective) cooperative control systems to regulate variables such as pitch, torque, power, rotor speed, power factors of every wind turbine, etc. Meanwhile, with increasing demands for efficiency and product quality and progressing integration of automatic control systems in high-cost and safety-critical processes, the fields of fault detection and isolation (FDI) and fault tolerant control (FTC) play an important role. This thesis covers the theoretical development and also the implementation of different FDI and FTC techniques in WTs. The purpose of wind turbine FDI systems is to detect and locate degradations and failures in the operation of WT components as early as possible, so that maintenance operations can be performed in due time (e.g., during time periods with low wind speed). Therefore, the number of costly corrective maintenance actions can be reduced and consequently the loss of wind power production due to maintenance operations is minimized. The objective of FTC is to design appropriate controllers such that the resulting closed-loop system can tolerate abnormal operations of specific control components and retain overall system stability with acceptable system performance. Different FDI and FTC contributions are presented in this thesis and published in different JCR-indexed journals and international conference proceedings. These contributions embrace a wide range of realistic WTs faults as well as different WTs types (onshore, fixed offshore, and floating). In the first main contribution, the normalized gradient method is used to estimate the pitch actuator parameters to be able to detect faults in it. In this case, an onshore WT is used for the simulations. Second contribution involves not only to detect faults but also to isolate them in the pitch actuator system. To achieve this, a discrete-time domain disturbance compensator with a controller to detect and isolate pitch actuator faults is designed. Third main contribution designs a super-twisting controller by using feedback of the fore-aft and side-to-side acceleration signals of the WT tower to provide fault tolerance capabilities to the WT and improve the overall performance of the system. In this instance, a fixed-jacket offshore WT is used. Throughout the aforementioned research, it was observed that some faults induce to saturation of the control signal leading to system instability. To preclude that problem, the fourth contribution of this thesis designs a dynamic reference trajectory based on hysteresis. Finally, the fifth and last contribution is related to floating-barge WTs and the challenges that this WTs face. The performance of the proposed contributions are tested in simulations with the aero-elastic code FAST.<br>La energía renovable es una energía sustentable importante en el mundo. Hasta ahora, como parte esencial de la energía de bajas emisiones en muchos países, la energía renovable ha sido importante para la seguridad energética nacional, y jugó un papel importante en la reducción de las emisiones de carbono. Proviene de recursos naturales, como el viento, la energía solar, la lluvia, las mareas, la biomasa y el calor geotérmico. Entre ellos, la energía eólica está emergiendo rápidamente como una tecnología sostenible de bajo carbono, eficiente en el uso de los recursos y rentable en el mundo. Debido a la demanda de instalaciones de producción de mayor potencia con menos impactos ambientales, el aumento continuo en el tamaño de las turbinas eólicas y las tecnologías offshore (flotantes) recientemente desarrolladas han llevado a nuevos desafíos en los sistemas de turbinas eólicas. Las turbinas eólicas son sistemas complejos con grandes estructuras flexibles que funcionan en condiciones ambientales muy turbulentas e impredecibles para una red eléctrica variable. La maximización de los sistemas de conversión de energía eólica, los problemas de minimización de la fatiga mecánica, los costos por kilovatios-hora de estrategias de reducción, cuestiones de confiabilidad, problemas de estabilidad y disponibilidad (sostenibilidad) exigen el uso de sistemas avanzados de control cooperativo (multivariable y multiobjetivo) para regular variables tales como paso, par, potencia, velocidad del rotor, factores de potencia de cada aerogenerador, etc. Mientras tanto, con las crecientes demandas de eficiencia y calidad del producto y la progresiva integración de los sistemas de control automático en los procesos de alto costo y de seguridad crítica, los campos de detección y aislamiento de fallos (FDI) y control tolerante a fallos (FTC) juegan un papel importante. Esta tesis cubre el desarrollo teórico y también la implementación de diferentes técnicas de FDI y FTC en turbinas eólicas. El propósito de los sistemas FDI es detectar y ubicar las degradaciones y fallos en la operación de los componentes tan pronto como sea posible, de modo que las operaciones de mantenimiento puedan realizarse a su debido tiempo (por ejemplo, durante periodos con baja velocidad del viento). Por lo tanto, se puede reducir el número de costosas acciones de mantenimiento correctivo y, en consecuencia, se reduce al mínimo la pérdida de producción de energía eólica debido a las operaciones de mantenimiento. El objetivo de la FTC es diseñar controladores apropiados de modo que el sistema de bucle cerrado resultante pueda tolerar operaciones anormales de componentes de control específicos y retener la estabilidad general del sistema con un rendimiento aceptable del sistema. Diferentes contribuciones de FDI y FTC se presentan en esta tesis y se publican en diferentes revistas indexadas a JCR y en congresos internacionales. Estas contribuciones abarcan una amplia gama de fallos WTs realistas, así como diferentes tipos de turbinas (en tierra, en alta mar ancladas al fondo del mar y flotantes). El rendimiento de las contribuciones propuestas se prueba en simulaciones con el código aeroelástico FAST.
37
Purarjomandlangrudi, Afrooz. "Application of machine learning technique in wind turbine fault diagnosis." Thesis, Queensland University of Technology, 2014. https://eprints.qut.edu.au/70624/2/Afrooz_Purarjomandlangrudi_Thesis.pdf.
Full textAbstract:
In this study, a machine learning technique called anomaly detection is employed for wind turbine bearing fault detection. Basically, the anomaly detection algorithm is used to recognize the presence of unusual and potentially faulty data in a dataset, which contains two phases: a training phase and a testing phase. Two bearing datasets were used to validate the proposed technique, fault-seeded bearing from a test rig located at Case Western Reserve University to validate the accuracy of the anomaly detection method, and a test to failure data of bearings from the NSF I/UCR Center for Intelligent Maintenance Systems (IMS). The latter data set was used to compare anomaly detection with SVM, a previously well-known applied method, in rapidly finding the incipient faults.
38
Gulán, Filip. "Detekce, lokalizace a určení plochy chronických ran." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2018. http://www.nusl.cz/ntk/nusl-385966.
Full textAbstract:
The aim of this diploma thesis is to design and implement a multiplatform application for detection, localization and determination of the extent of chronic wounds. The application is intended to assist nurses, doctors and healthcare assistants to monitor and evaluate chronic wounds in the course of treatment. The application is based on the Typescript programming language, on the Ionic hybrid application framework and on the Electron desktop application framework. Chronic wound assessment runs on the server-side where the Python programming language is used. The Flask application framework is used for the RESTful application interface and the OpenCV library is used for image processing.
39
Zafar, Jawwad. "Winding short-circuit fault modelling and detection in doubly-fed induction generator based wind turbine systems." Doctoral thesis, Universite Libre de Bruxelles, 2011. http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/209854.
Full textAbstract:
Abstract<p><p>This thesis deals with the operation of and winding short-circuit fault detection in a Doubly-Fed Induction Generator (DFIG) based Wind Turbine Generator System (WTGS). Both the faulted and faultless condition of operation has been studied, where the focus is on the electrical part of the system. The modelled electrical system is first simulated and the developed control system is then validated on a test bench. The test-bench component dimensioning is also discussed.<p><p>The faultless condition deals with the start-up and power production mode of operation. Control design based on the Proportional Integral (PI) control technique has been compared for power and torque control strategies against the Linear Quadratic Gaussian (LQG) control technique, at different operating points through the variable-speed region of WTGS operation following the maximum power curve of the system. It was found that the torque control strategy offered less degradation in performance for both the control techniques at operating points different for the one for which the control system was tuned. The start-up procedure of the DFIG based WTGS has been clarified and simplified. The phase difference between the stator and the grid voltage, which occurs due to the arbitrary rotor position when the rotor current control is activated, is minimized by using a sample-and-hold technique which eliminates the requirement of designing an additional controller. This method has been validated both in simulation and experiments.<p><p>The faulted condition of operation deals with the turn-turn short-circuit fault in the phase winding of the generator. The model of the generator, implemented using the winding-function approach, allows the fault to be created online both in a stator and a rotor phase. It has been demonstrated that the magnitude of the current harmonics, used extensively in literature for the Machine Current Signature Analysis (MCSA) technique for winding short-circuit fault detection, is very different when the location of the fault is changed to another coil within the phase winding. This makes the decision on the threshold selection for alarm generation difficult. Furthermore, the control system attenuates the current harmonics by an order of magnitude. This attenuation property is also demonstrated through experiments. The attention is then shifted to the negative-sequence current component, resulting from the winding unbalance, as a possible fault residual. Its suitability is tested in the presence of noise for scenarios with different fault locations, fault severity in terms of the number of shorted-turns and grid voltage unbalance. It is found that due to the presence of a control system the magnitude of the negative-sequence current, resulting from the fault, remains almost the same for all fault locations and fault severity. Thus, it was deemed more suitable as a fault residual. In order to obtain a fast detection method, the Cumulative Sum (CUSUM) algorithm was used. The test function is compared against a threshold, determined on the basis of expected residual magnitude and the time selected for detection, to generate an alarm. The validation is carried out with noise characteristics different from the ones used during the design and it is shown that the voltage unbalance alone is not able to trigger a false alarm. In all the scenarios considered, the detection was achieved within 40 ms despite the presence of measurement filters.<br>Doctorat en Sciences de l'ingénieur<br>info:eu-repo/semantics/nonPublished
40
Costello, Luke H. "Condition Monitoring and Fault Detection of Blade Damage in Small Wind Turbines Using Time-series and Frequency Analyses." DigitalCommons@CalPoly, 2021. https://digitalcommons.calpoly.edu/theses/2280.
Full textAbstract:
Condition monitoring systems are critical for autonomous detection of damage when operating remote wind turbines. These systems continually monitor the turbine’s operating parameters and detect damage before the turbine fails. Although common in utility-scale turbines, these systems are mostly undeveloped in distributed, small-scale turbines due to their high cost and need for specialized equipment. The Cal Poly Wind Power Research Center is developing a low-cost, modular solution known as the LifeLine system. The previous version contained monitoring equipment, but lacked decision-making capabilities.
The present work builds on the LifeLine by developing software-based detection of blade damage. Detection is done by monitoring of tower vibrations, rotor speed, and generator power output. First, testing is completed to inform algorithm design: the tower vibrational response is recorded, and blade damage is simulated by adding a mass imbalance to one blade. From these results, several algorithms are developed, and their performance is analyzed in a cross-validation study. The time-series method known as the Nonlinear State Estimation Technique and Sequential Probability Ratio Test (NSET+SPRT) is implemented first. This algorithm is highly successful, with a 93.3% rate of correct damage detection; however, it occasionally raises false alarms during normal operation. A custom-built algorithm known as the Adaptive Fast Fourier Transform (AFFT) is also built; its strength lies in its elimination of false alarms. The final system utilizes a joint monitoring approach, combining the benefits of the NSET+SPRT and AFFT. The final algorithm is successful, correctly categorizing 95.5% of data when operating above 120RPM, and raising no false alarms in normal operation. This version is then implemented for live monitoring on the Cal Poly Wind Turbine, allowing for robust and autonomous detection of blade damage.
41
Allen, Jeffrey R. "An Analysis of SeaWinds Simultaneous Wind/Rain Retrieval in Severe Weather Events." Diss., CLICK HERE for online access, 2005. http://contentdm.lib.byu.edu/ETD/image/etd704.pdf.
Full text
42
Bravo, Jimenez Ismael. "Detection and removal of wind turbine ice : Method review and a CFD simulation test." Thesis, Högskolan i Gävle, Energisystem, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:hig:diva-27798.
Full textAbstract:
Nowadays, the energy sector is facing a huge demand that needs to be covered. Wind energy is one of the most promising energy resources as it is free from pollution, clean and probably will arise as one of the main energy sources to prevent global warming from happening. Almost 10% of the global energy demand is coming from renewable resources. By 2050 this percentage is expected to grow to 60%. Therefore, efforts on wind turbine technology (i.e. reliability, design…) need to be coped with this growth. Currently, large wind energy projects are usually carried out in higher altitudes and cold climates. This is because almost all of the cold climates worldwide offer profitable wind power resources and great wind energy potential. Operating with wind turbines in cold climates bring interesting advantages as a result of higher air density and consequently stronger winds (wind power is around 10% higher in the Nordic regions). Not only benefits can be obtained but extreme conditions force to follow harsh conditions. Low temperatures and ice accretion present an important issue to solve as can cause several problems in fatigue loads, the balance of the rotor and aerodynamics, safety risks, turbine performance, among others. As wind energy is growing steadily on icy climates is crucial that wind turbines can be managed efficiently and harmlessly during the time they operate. The collected data for the ice detection, de-icing and anti-icing systems parts was obtained through the company Arvato Bertelsmann and is also based on scientific papers. In addition, computer simulations were performed, involving the creation of a wind tunnel under certain conditions in order to be able to carry out the simulations (1st at 0ºC, 2nd at -10ºC) with the turbine blades rotating in cold regions as a standard operation. In this project, Computational Fluids Dynamics (CFD) simulation on a 5MW wind turbine prototype with ice accretion on the blades to study how CL and CD can change, also different measures of ice detection, deicing and anti-icing systems for avoiding ice accumulation will be discussed. Simulation results showed a logical correlation as expected, increasing the drag force about 5.7% and lowering the lift force 17,5% thus worsening the turbine's efficiency.
43
Curt, Jordan. "Damage detection for wind turbine towers with Digital Image Correlation." Thesis, université Paris-Saclay, 2021. http://www.theses.fr/2021UPAST008.
Full textAbstract:
La technologie des éoliennes est mature et le marché est avec le temps devenu compétitif. Un levier pour diminuer les coûts de l’exploitation des éoliennes réside dans l’optimisation de la maintenance des turbines. Dans ce contexte, cette étude se focalise sur l’état de la tour de l’éolienne. La durée de vie de la tour est influencée par des paramètres incertains liés à l’environnement et les aléas matériaux. Ceci rend très difficile la prédiction de la durée de vie d’un mât et des marges importantes sont prises lors de son dimensionnement. Cependant des éoliennes se sont déjà effondrées dans le monde, à chaque fois lors d'épisodes climatiques violents. Les deux phénomènes de ruine identifiés sont le flambement local de la tour et l'arrachement de sa partie supérieure. Des sources académiques ou industrielles ont également rapporté la découverte de fissures sur quelques turbines aux interfaces soudure-matériau. Une réflexion concernant la criticité d'une fissure vis à vis du flambement et de sa propagation brutale a été menée dans le cas où la tour est soumise à des chargements violents. Des simulations éléments finis ont permis de montrer que la propagation brutale de fissure était le phénomène dimensionnant pour la prolongation de durée de vie d'une tour d'éolienne fissurée.Pour s'assurer de l'intégrité des structures des techniques de suivi de santé ont été développées. Celles-ci sont en général réalisées à partir de capteurs extrapolant un état de santé global à partir de données locales (accéléromètres, jauges de déformations). Le risque est qu'un défaut reste invisible mais soit tout de même critique pour la structure. Il est donc essentiel de revisiter les outils de prédiction à partir de données de sites relatives à l’endommagement de la tour et des structures les supportant. Dans le cadre de ce travail de thèse des méthodes basées sur l'imagerie et plus particulièrement la Corrélation d'Images Numériques (CIN) ont été développées. Pour aborder le problème, deux approches à différentes échelles ont été considérées.La première, à l’échelle structurale, consiste à reconstruire le champ de déplacement de l'éolienne vue comme une structure unidimensionnelle. La présence d'un défaut induira une perte de raideur latérale, et donc une (quasi-)discontinuité dans le champ de rotation. Les défis relatifs à cette partie sont doubles : l'absence de contraste sur la tour et la prise de photos en extérieur sur une grande structure. Une technique de CIN intégrée a été utilisée afin de diminuer le nombre d'inconnues du problème et réduire les incertitudes de mesures. Il a été montré que l'influence d'un défaut, plus particulièrement d'une fissure, serait trop faible par rapport aux incertitudes de mesures. Cependant, suivant cette approche, une méthode innovante d'analyse modale de tour d'éolienne a été développée et les deux premières fréquences propres ont pu être déterminées avec précision.La seconde est une stratégie de contrôle de la structure à l'échelle mésoscopique. Celle-ci se base sur un dispositif de caméras bon marché à l'intérieur de l'éolienne couvrant les soudures circonférentielles au niveau des zones de plus fortes contraintes. L'idée est d'établir pour chaque caméra lorsque la structure est considérée comme saine une base modale de déplacement caractéristique de cet état grâce à des techniques de réduction de modèle. Au cours du temps, si un défaut sous-jacent ou traversant apparait, celui-ci induira une perturbation du champ de déplacement qui pourra être détectée à l'aide d'indicateurs globaux tels que l'écart en déplacement ou les résidus de corrélation.Afin de déterminer si la détection d'un défaut aux deux échelles est faisable ou non, la prise en compte de l'incertitude de mesure en regard de l'influence d'un défaut est primordiale. Alors, un cadre mathématique de la CIN optimale à N champs a été proposé et validé<br>Wind turbine technology is mature and the market has become very competitive over time. A lever to reduce the costs of wind turbine operation lies in optimizing turbine maintenance. In this context, this study focuses on the wind turbine tower. The lifetime of the tower is influenced by uncertain parameters related to the environment and material hazards. This makes it very difficult to predict the tower lifetime, and large safety factors are used when dimensioning it. However, around the world, a few wind turbines have already collapsed, each time during extreme climatic episodes. The two identified ruin causes are the local buckling of the tower and the tearing off of its upper part. Academic and industrial sources have also reported the discovery of cracks on some turbines at the weld-material interfaces. An investigation of the criticality of a crack with respect to buckling and its fatal propagation has been carried out in the case where the tower is subjected to violent loads. Finite element simulations showed that the fatal crack propagation was the limiting phenomenon for the life extension of a cracked wind turbine tower.To ensure structural integrity, health monitoring techniques were developed over time. These are generally carried out using sensors extrapolating a global state of health from local data (accelerometers, strain gauges). The risk is that a defect remains invisible but is nevertheless critical for the structure. It is therefore essential to revisit the prediction tools based on site data. Within the framework of this work, methods based on imagery and more particularly Digital Image Correlation (DIC) have been developed. To address the problem, two approaches at different scales have been considered.The first, at structural scale, consists in reconstructing the wind turbine displacement field considered as a one-dimensional structure. The presence of damage will induce a loss of lateral stiffness, and thus a (quasi-)discontinuity in the field of rotation. The challenges for this part are twofold: the lack of contrast on the tower and the capture of outdoor photographs on a large structure. An integrated DIC technique was used in order to reduce the number of unknowns in the problem and to lower the measurement uncertainties. It was shown that the influence of a damage, especially a crack, would be too small regarding the measurement uncertainties. However, an innovative method of wind turbine tower modal analysis was developed and the first two natural frequencies could be determined accurately.The second strategy is focused on a mesoscopic scale. It is based on a low-cost camera system, inside the wind turbine, covering the circumferential welds in the regions of highest stress. The idea is to establish for each camera, when the structure is considered sound, a displacement modal basis using model reduction techniques. Over time, if an underlying or through defect appears, it will induce a disturbance in the displacement field which can be detected using global indicators such as displacement deviation or DIC residuals.In order to determine whether the detection of damage at both scales is feasible or not, it is essential to take into account the measurement uncertainty. Therefore, a mathematical framework of the optimal CIN with N fields has been proposed and validated
44
Karuturi, Hemanth Surya, and Megha Sanjeev Reddy Karri. "Raspberry Pi Based IoT System for Bats Detection at Wind Farms." Thesis, Blekinge Tekniska Högskola, Institutionen för matematik och naturvetenskap, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-20899.
Full textAbstract:
Context: Large numbers of bats are killed by collisions with wind turbines and there is at present no accepted method of reducing or preventing this mortality. We designed a system, which detects and records any bats’ activity in and around the surroundings of wind turbines. The system can help to study bats by identifying the species that are present in that particular locality. Objectives: The main objective of this thesis is to design an ultrasound-based IoT system, which detects the bats to prevent them from clashing with wind turbines. The design is based on a study of bats’ behaviors. Methods: The system has been developed using User-Driven Design, UDD, approach. The required functionalities have been embedded into IoT based system. An ultrasonic technology along with other sensors are used. The sensors are intended to activate monitoring during favorable conditions for bat activity. Results: A model of a system has been developed. The model was implemented into a prototype. Recorded bats’ activities are uploaded to a server by employing a suitable app, which informs the user about the activities of bats' various sub-species. Conclusions: A surveillance for bats approaching the wind farms within 80 m has been developed. The monitoring system is activated when the weather conditions are favorable for bat activities.
45
Venkatramanan, Adithya. "Design of control electronics for the Ram Energy Distribution Detector." Thesis, Virginia Tech, 2015. http://hdl.handle.net/10919/56594.
Full textAbstract:
The bulk motion of the neutral gas at altitudes between about 200 and 600 km is an important factor in predicting the onset of plasma instabilities that are known to distort and/or disrupt high frequency radio communications. A ram wind sensor is a space science instrument that, when mounted on a satellite in low-Earth orbit, makes in-situ measurements of the component of the neutral gas velocity that lies along the orbit track of the satellite. The instrument works by changing the voltage on one of a set of grids and measuring a corresponding electron current generated by ions flowing through the grid stack and detected by the microchannel plate, generating a function of current vs. voltage called an I-V curve. Traditionally, the size and power requirements of ram wind sensors has limited their use to larger satellites. In this thesis, the electrical design and basic testing of a cubesat compatible RWS known as the ram energy distribution detector (REDD) are described.
The mechanical design of the REDD sensor is first described, and the requirements of the electrical design are presented. The electrical requirements are based on both the characteristics of the ionosphereic flight environment, and on the size and power requirements typical of the small cubesat platforms for which the instrument is intended.
The electrical hardware is then described in detail. The microcontroller design is reviewed as well, including the instrument's operating mode, and timing scheme.
Test data showing the basic functionality of the instrument are then presented. Bench tests validate the design by confirming its ability to control voltages and measure small electron currents. End-to-end tests were also performed in a vacuum chamber to mimic the ionospheric environment. These data are presented to show the ability of the REDD sensor to meet or exceed its design specifications.<br>Master of Science
46
Åkerberg, Ludvig. "Using Unsupervised Machine Learning for Outlier Detection in Data to Improve Wind Power Production Prediction." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-200336.
Full textAbstract:
The expansion of wind power for electrical energy production has increased in recent years and shows no signs of slowing down. This unpredictable source of energy has contributed to destabilization of the electrical grid causing the energy market prices to vary significantly on a daily basis. For energy producers and consumers to make good investments, methods have been developed to make predictions of wind power production. These methods are often based on machine learning were historical weather prognosis and wind power production data is used. However, the data often contain outliers, causing the machine learning methods to create inaccurate predictions. The goal of this Master’s Thesis was to identify and remove these outliers from the data so that the accuracy of machine learning predictions can improve. To do this an outlier detection method using unsupervised clustering has been developed and research has been made on the subject of using machine learning for outlier detection and wind power production prediction.<br>Vindkraftsproduktion som källa för hållbar elektrisk energi har på senare år ökat och visar inga tecken på att sakta in. Den här oförutsägbara källan till energi har bidragit till att destabilisera elnätet vilket orsakat dagliga kraftiga svängningar i priser på elmarknaden. För att elproducenter och konsumenter ska kunna göra bra investeringar har metoder för att prediktera vindkraftsproduktionen utvecklats. Dessa metoder är ofta baserade på maskininlärning där historiska data från väderleksprognoser och vindkraftsproduktion använts. Denna data kan innehålla så kallade outliers, vilket resulterar i försämrade prediktioner från maskininlärningsmetoderna. Målet med det här examensarbetet var att identifiera och ta bort outliers från data så att prediktionerna från dessa metoder kan förbättras. För att göra det har en metod för outlier-identifikation utveklats baserad på oövervakad maskininlärning och forskning har genomförts på områdena inom maskininlärning för att identifiera outliers samt prediktion för vindkraftsproduktion.
47
Verma, Anoop Prakash. "Performance monitoring of wind turbines : a data-mining approach." Diss., University of Iowa, 2012. https://ir.uiowa.edu/etd/3398.
Full textAbstract:
The rapid growth of wind turbines in terms of turbine size, number of installations and rated capacity has a huge impact on its operations and maintenance costs. Monitoring the performance of wind turbines and early fault prediction is highly desirable. To date, traditional maintenance strategies such as reactive maintenance, periodic maintenance etc. are more prevalent in wind industry. However, over the last couple of years, the research pertaining to wind turbine has been shifted towards the condition monitoring and maintenance.
Condition monitoring approaches have shown their potential in wind industry by providing continuous monitoring of the wind turbines, and identifying fault signatures in the event of faults. However, most of the studies reported in literature are based on the simulated dataset, or in constrained experiments. In reality, the external environment plays an important role in governing the turbine operations. Moreover, the cost associated with condition monitoring cannot be justified as it often requires installations of specific sensors, equipment. Another stream of research focuses on utilizing historical turbine data for turbine performance assessment in real time. The cost associated with such approaches is almost negligible as most of the wind farms are equipped with SCADA systems which records turbine performance data in regular time-interval. Such approaches are called as performance monitoring.
In this dissertation, the performance monitoring of wind turbines is accomplished using the historical wind turbine data. The information from SCADA operational data, and fault logs is used to construct accurate models predicting the critical wind turbine faults. Depending upon the nature of turbine faults, monitoring wind turbines with different objectives is studied to accomplish different research goals.
Two research directions of wind turbines performance are pursued, (1) identification and prediction of critical turbine faults, and (2) monitoring the performance of overall wind farm. The goal of predicting critical faults is to facilitate planned maintenance, whereas, monitoring the performance of overall wind farm provides the status-quo of all wind turbines installed in a wind farm. Depending on the requirement, the performance of overall wind farm can be assessed on a daily, weekly, or monthly basis. Solution methodologies presented in the dissertation are generic enough to be applicable to other industries such as wastewater treatment facilities, flood prediction, etc.
48
Liu, Zongchang. "A Systematic Framework for Unsupervised Feature Mining and Fault Detection for Wind Turbine Drivetrain Systems." University of Cincinnati / OhioLINK, 2016. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1471348052.
Full text
49
Mcclelland, Hunter Grant. "Towards Detecting Atmospheric Coherent Structures using Small Fixed-Wing Unmanned Aircraft." Diss., Virginia Tech, 2019. http://hdl.handle.net/10919/90667.
Full textAbstract:
The theory of Lagrangian Coherent Structures (LCS) enables prediction of material transport by turbulent winds, such as those observed in the Earth's Atmospheric Boundary Layer. In this dissertation, both theory and experimental methods are developed for utilizing small fixed-wing unmanned aircraft systems (UAS) in detecting these atmospheric coherent structures. The dissertation begins by presenting relevant literature on both LCS and airborne wind estimation. Because model-based wind estimation inherently depends on high quality models, a Flight Dynamic Model (FDM) suitable for a small fixed-wing aircraft in turbulent wind is derived in detail. In this presentation, some new theoretical concepts are introduced concerning the proper treatment of spatial wind gradients, and a critical review of existing theories is presented. To enable model-based wind estimation experiments, an experimental approach is detailed for identifying a FDM for a small UAS by combining existing computational aerodynamic and data-driven approaches. Additionally, a methodology for determining wind estimation error directly resulting from dynamic modeling choices is presented and demonstrated. Next, some model-based wind estimation results are presented utilizing the experimentally identified FDM, accompanied by a discussion of model fidelity concerns and other experimental issues. Finally, an algorithm for detecting LCS from a single circling fixed-wing UAS is developed and demonstrated in an Observing System Simulation Experiment. The dissertation concludes by summarizing these contributions and recommending future paths for continuing research.<br>Doctor of Philosophy<br>In a natural or man-made disaster, first responders depend on accurate predictions of where the wind might carry hazardous material. A mathematical theory of Lagrangian Coherent Structures (LCS) has shown promise in ocean environments to improve these predictions, and the theory is also applicable to atmospheric flows near the Earth’s surface. This dissertation presents both theoretical and experimental research efforts towards employing small fixed-wing unmanned aircraft systems (UAS) to detect coherent structures in the Atmospheric Boundary Layer (ABL). These UAS fit several “gaps” in available sensing technology: a small aircraft responds significantly to wind gusts, can be steered to regions of interest, and can be flown in dangerous environments without risking the pilot’s safety. A key focus of this dissertation is to improve the quality of airborne wind measurements provided by inexpensive UAS, specifically by leveraging mathematical models of the aircraft. The dissertation opens by presenting the motivation for this research and existing literature on the topics. Next, a detailed derivation of a suitable Flight Dynamic Model (FDM) for a fixed-wing aircraft in a turbulent wind field is presented. Special attention is paid to the theories for including aerodynamic effects of flying in non-uniform winds. In preparation for wind measurement experiments, a practical method for obtaining better quality FDMs is presented which combines theoretically based and data-driven approaches. A study into the wind-measurement error incurred solely by mathematical modeling is presented, focusing on simplified forms of the FDM which are common in aerospace engineering. Wind estimates which utilize our best available model are presented, accompanied by discussions of the model accuracy and additional wind measurement concerns. A method is developed to detect coherent structures from a circling UAS which is providing wind information, presumably via accurate model based estimation. The dissertation concludes by discussing these conclusions and directions for future research which have been identified during these pursuits.
50
Tureček, Martin. "Detekce obličejů v obraze z kamery na mobilním telefonu s WM." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2010. http://www.nusl.cz/ntk/nusl-235553.
Full textAbstract:
This thesis deals with a face detection on mobile phones. It especially focuses on Windows Mobile platform. The introduction is therefore devoted to this operating system and alternatives of working with the camera. The next part of the text refers to general problems of the face detection in the image considering the weak performance of the target device. Another part of this thesis is a description of the acquisition of images from the camera using DirectShow multimedia framework and creation of a custom transformation filter for the face detection. Achieved results are summarized in the conclusion. It takes a form of tests examining different mobile devices. All difficulties arising during Windows Mobile developing are also mentioned.