Graduate School of Engineering and Science

Permanent URI for this collectionhttps://hdl.handle.net/10679/9877

Browse

Now showing 1 - 20 of 58

Metadata only
A software-defined networking approach for wireless systems
(2015-05) Yazıcı, Volkan; Sunay, Mehmet Oğuz; Civanlar, Mehmet Reha; Ercan, Ali Özer; Sunay, Mehmet Oğuz; Civanlar, Mehmet Reha; Ercan, Ali Özer; Tekalp, M.; Sayit, M.; Department of Computer Science; Yazıcı, Volkan
The tremendous growth in wireless Internet use is showing no signs of slowing down. Existing cellular networks are starting to be insufficient in meeting this demand in part due to their inflexible and expensive equipment as well as complex and non-agile control plane. Software-defined networking (SDN) is emerging as a natural solution for the next generation cellular networks as it enables further Network Functions Virtualization (NFV) opportunities and network programmability. In this dissertation, we advocate an all-SDN network architecture with hierarchical network control capabilities to allow for different grades of performance and complexity in offering core network services and provide service differentiation for 5G systems. As a showcase of this architecture, we first introduce a unified approach to mobility, handoff and routing management and offer Connectivity Management as a Service (CMaaS). CMaaS is offered to application developers and over-the-top service providers to provide a range of options in protecting their flows against subscriber mobility at different price levels. Next, we present the implementation details of a distributed SDN controller specifically crafted to realize the proposed all-SDN architecture and investigate the flow-level performance characteristics of the system.
Embargo
Runtime specialization and autotuning of sparse matrix-vector multiplication
(2015-12) Ylmaz, Buse; Aktemur, Tankut Barış; Garzaran, M.; Sözer, Hasan; Kaya, K.; Uğurdağ, Hasan Fatih; Department of Computer Science; Yılmaz, Buse
Runtime specialization is used for optimizing programs based on partial information available only at runtime. In this thesis, we present a purpose-built compiler to quickly specialize Sparse Matrix-Vector Multiplication code for a particular matrix at runtime. There are several specialization methods and the best one depends both on the matrix and the platform. To avoid having to generate all the specialization variations, we use an autotuning approach to predict the best specializer for a given matrix. To this end, we define a set of matrix features for autotuning. Several of these features are unique to our work. We evaluate our system on two machines and show that our approach predicts either the best or the second best method in 91-96\% of the matrices. Predictions achieve average speedups that are very close to the speedups achievable when only the best methods are used. By using an efficient code generator and a carefully designed set of matrix features, we show the total runtime costs of autotuning and specialization can be amortized to bring performance benefits for many real-world cases.
Metadata only
An investigation into fluid flow and heat transfer of high frequency synthetic jets for electronics cooling
(2016-04) Ghaffari, Omidreza; Arık, Mehmet; Arık, Mehmet; Yaralıoğlu, Göksenin; Başol, Altuğ; Solovitz, S. A.; Koşar, A.; Department of Mechanical Engineering; Ghaffari, Omidreza
Modern electronics have been decreasing in size for decades, so their cooling systems must continually improve in efficiency too. In particular, compactness is vital, which is challenging because typical thermal management uses relatively large fans and heat sinks. For more advanced liquid cooling, additional coolant and structure are required, somewhat counteracting the improvement with liquid heat transfer. Ideally, thermal management should be economical, low volume, and localized on the powered devices. Fortunately, some recent advances in synthetic jet devices may provide a potential solution. Synthetic jets use an oscillating structure near an orifice, which produces a periodic jet outflow and sink inflow. When averaging over time, this leads to an axial jet, which can be directed towards a powered device. Unlike traditional impingement cooling, the jet is supplied by ambient fluid, as opposed to an additional coolant. We developed a series of thermal, structural and flow experimental setups in order to test two distinct in house manufactured synthetic jet devices along with one commercial ultrasonic jet. The in house manufactured slot synthetic jets used in this study have a different topology than the previous slot actuators, sandwiching two circular disks together. While a large number of earlier devices placed their orifices normal to the oscillating piezoelectric disk, this new approach placed the orifice along the circumference of the sandwiched pair. Hence, the jet direction was perpendicular to the piezoelectric disk deflection. We later performed a series of CFD simulations to examine the performance of in-house made novel slot synthetic jet and validated it with experimental data. Along with measurements of deflection and thermal performance, we used time-averaged and phase-locked PIV to study the flow physics and their effects on heat transfer. We focused on synthetic jet behavior at several frequencies, both at and away from resonant conditions, which may help in selecting device conditions with lower acoustic noise. For a slot synthetic jet the degradation of heat transfer for small jet-to-surface spacings, like H/Dh = 2, was due to the reduced growth of the vortices. In addition, there was re-entrainment of warm air next to the impinging plate by the vortex back into the jet flow, and some warm fluid was sucked back into the orifice. In a slot jet case the heat transfer is maximum for a jet-to-surface spacing for 5 ≤ H/Dh ≤ 10, which is associated with flow dominated by coherent vortices that grow to full strength before detachment or impact with the wall. At the diaphragm resonant condition and below, the flow structure was similar at all phases, with a single vortex present between the orifice and the wall. This response suggests that there is a critical jet-to-surface spacing for this behavior, Hcrit = Uo/2f, which should relate to the optimal thermal condition. By tuning the actuator frequency to the wall spacing, the vortices can reach the wall in phase at the end of the outstroke. When well-tuned, the thermal response is governed primarily by Reynolds number. There is a superior cooling performance at high Stokes number with the same ReU0 number. The maximum cooling performance of a circular synthetic jet at the close heater to jet distance was not observed at the structural frequency of the jet where the maximum velocity occurs. It occurs at frequencies greater than the structural frequency. Based on efficiency (COP) comparison, a slot synthetic jet has a better cooling performance compared to a circular jet and ultrasonic micro-blower due to the difference in the flow physics of the two jets so the slot synthetic jet is a better candidate for electronics cooling applications. For the micro blower jet the preferred operating frequency of the piezoelectric actuator occurs at an ultrasonic frequency of 25 kHz, meaning that this device can function with low noise. The micro-blower axial velocity profile shows similar behavior to high Reynolds number turbulent free jets in the far field, including a self-similar profile. But in the near field, there has been a significant deviation with turbulent free jets. The average Nusselt number increases sharply up to H/D = 10, and then it shows a gradual increase till H/D=15. There is a fairly flat maximum region for 15 ≤ H/D ≤ 30, followed by a gradual decay for H/D > 30. The heat transfer increases more than three times by moving the jet from H/D = 2 to H/D = 10. This reveals that the jet performance is highly sensitive to the jet-to-surface spacing. The jet cooling performance is sensitive to the frequency, though there is a 1 kHz wide band of similar thermal response about the peak. The coefficient of performance at the best operating heat transfer condition is about 2, which is less than the value of 14 seen for a slot synthetic jet. Thus, while this micro-blower can greatly reduce noise, it has a significant performance penalty.
Metadata only
Near- and far-field thermal radiation in metamaterials and the development of NF-RT-FDTD algorithm
(2016-06) Didari, Azadeh; Mengüç, Mustafa Pınar; Yaralıoğlu, Göksenin; Ertürk, H.; Erkol, Güray; Şendur, K.; Department of Electrical and Electronics Engineering; Didari, Azadeh
In this dissertation, analysis of near-field regime of thermal radiative transfer in metamaterials supporting surface phonon polaritons (SPhPs) is given. Solutions of electromagnetic fields at subwavelength distances are studied where combined effects of surface waves and total internal reflection, result in enhancement of thermal radiation by orders of magnitude when compared against far-field regime of thermal radiation which is obtainable through Planck's blackbody law. We have developed Near Field Radiative Transfer Finite Difference Time Domain (NF-RT-FDTD) algorithm which is developed based on Finite Difference Time Domain (FDTD) method specifically designed to provide full solutions to near-field radiative transfer problems by solving Maxwell's equations combined with fluctuation-dissipation theory. We have extensively investigated the near- and far-field thermal emission and heat flux profiles in different geometries with corrugations and porosities of various size and shapes and report on our findings which reveals a high degree of accuracy is attainable by NF-RT-FDTD method in complex geometries. We have compared our results against solutions of effective medium theory which makes effective medium theory's ability to provide accurate solutions highly questionable. NF-RT-FDTD could be used to provide solutions for complex geometries with different applications, including energy harvesting with near-field thermophotovoltaics, radiative cooling, thermal sensing, nano manufacturing and medical diagnostics.
Metadata only
Application of chemical mechanical polishing process on titanium based medical implants
(2017-05) Özdemir, Zeynep; Başım, Gül Bahar; Başım, Gül Bahar; Yaralıoğlu, Göksenin; Erkol, Güray; Sanyal, R.; Nizamoğlu, S.; Department of Mechanical Engineering; Özdemir, Zeynep
Biomaterials are commonly used as implant materials in the body for dental prostheses, orthopedic applications, heart valves and catheters. Based on the research studies conducted up to date, titanium and its alloys are known to be the most biocompatible materials due to their surface properties as well as extraordinary mechanical properties. Processing methods for the implant materials also affect the surface properties and may lead to contamination that can lessen the biocompatibility and after implantation may cause infection on patients which can be up to 4% in numbers. Changing the surface roughness and forming a surface oxide film have been implemented through various methods in the literature to increase of the biocompatibility and to ensure bio-inertness to the implant material. Sand blasting and chemical etching methods are commonly used for patterning the titanium surfaces to alter the surface roughness which can cause surface contamination. However, the other alternative methods such as high temperature plasma coating and laser patterning are costly. In this dissertation, Chemical Mechanical Polishing (CMP) process is established as an alternative technique to the existing methods in the literature in order to change the implant material surface properties. CMP process is one of the methods used in the semiconductor industry to ensure surface planarization through simultaneous mechanical and chemical actions. The abrasive particles in the polishing slurries provide the mechanical effect during the process enabling nanometer level erosion and cleaning the implant from any potential contamination during its machining. The chemical components of the slurry including the stabilizers, pH adjusters and oxidizers, on the other hand, help form a passive oxide film coating the surface. Generally, CMP is used to form very smooth surfaces but it has been demonstrated that by changing the slurry particle size and the pad material properties, it is possible to generated controlled roughness on the polished surface as well. The protective nature of the generated oxide film enables planarization in semiconductor applications. In implant applications of CMP, it is believed to help reduce the contamination on the surface of the bio-implants in the body environment and reducing the infection risk by stopping the chemical reactions in-vivo. It has been shown in the literature that the application of CMP on Ti films has been successful in terms of creating a smooth surface and a TiO2 oxide film. However, its native oxide film after CMP has not been characterized fully for its protective nature other than the passivating properties of the Ti/TiN films in semiconductor CMP applications. Titanium oxide film is known to promote the biocompatibility, cell adhesion, formation of hydroxyapatite layers. Yet, the oxide films obtained by artificial oxidation methods result in thick films and have porous structures. Therefore, in this study, CMP process has been applied to the Ti plates synergistically to remove the potentially contaminated surface layers and induce controlled roughness on the implant surfaces. In addition, the treated surface oxide layers have been characterized for the nature of the metal oxide layers in terms of their self-protective properties. Furthermore, biocompatibility of the CMP implemented surfaces have been evaluated through cell growth and infection resistance capabilities through biofilm analyses and optimal surface parameters were determined according to the desirability of the surface responses which help promote the cell behavior.In terms of carrying the results of this dissertation to the future studies, development of a 3 dimensional CMP process considering the 3-D nature of the implants is the most important necessity. The application of the 3-D CMP process on the implant surfaces is believed to be both an economical and more effective method on structuring the surface of the titanium based bio-implants. It is aimed to further develop a CMP driven surface nano-structuring methodology to create engineered surfaces on the Ti based bio-implants with self-protective surfaces to minimize chemical and bacterial reactivity, while promoting their biocompatibility through simultaneous surface patterning.
Metadata only
Thermo-mechanical behavior of severely deformed titanium
(2017-06) Sajjadifar, Seyedvahid; Yapıcı, Güney Güven; Yapıcı, Güney Güven; Arık, Mehmet; Bundur, Zeynep Başaran; Yılmazer, H.; Oral, A.; Department of Mechanical Engineering; Sajjadifar, Seyedvahid
Thermo-mechanical processing of metallic materials has attracted noticeable interest due to the fact that these processing methods can be used to improve the mechanical properties. In this investigation, ultrafine grained commercial purity titanium was fabricated utilizing equal channel angular extrusion as a severe plastic deformation technique. Compression tests were performed on severely deformed titanium at various temperatures of 600–900°C and at strain rates of 0.001–0.1/s. It was observed that severe plastic deformation via equal channel angular extrusion can considerably enhance the flow strength of grade 2 titanium deformed at 600 and 700°C. Post-compression microstructures showed that a fine-grained structure can be retained at a deformation temperature of 600°C. The strain rate sensitivity during hot compression of severely deformed Ti was shown to be strongly temperature dependent, where strain rate sensitivity increased with the increase in deformation temperature. High temperature workability of severely deformed grade 2 titanium was analyzed based on the flow localization parameter. According to the flow localization parameter values, deformation at and below 700°C is prone to flow localization. The flow response of the ultrafine grained microstructure is modeled using the Arrhenius type, dislocation density based and modified Johnson-Cook models. The validities of the models were demonstrated with the reasonable agreement in comparison to the experimental stress-strain responses. In order to investigate the influence of purity level on hot characteristics and dynamic softening mechanisms of severely deformed titanium, compression tests were also conducted on severely deformed grade 4 titanium at similar temperatures and strain rates. It was seen that the effects of deformation rate and temperature are significant on obtained flow stress curves of both grades. Higher compressive strength exhibited by grade 2 titanium at relatively lower deformation temperatures was attributed to the grain boundary characteristics in relation with its lower processing temperature. However, severely deformed grade 4 titanium demonstrated higher compressive strength at relatively higher deformation temperatures (above 800°C) due to suppressed grain growth via oxygen segregation limiting grain boundary motion. Constitutive equations were established to model the flow behavior, and the validity of the predictions was demonstrated with decent agreement accompanied by average error levels less than 5% for all the deformation conditions. Severely deformed grade 4 titanium was less stable at the temperature range of 600-800°C. Therefore, warm deformation behavior and microstructure evolution of severely deformed grade 4 titanium were studied at temperatures of 300-600°C and at strain rates of 0.001-0.1/s in order to reveal the pertinent softening mechanisms such as dynamic recovery and dynamic recrystallization. Results suggest that severe plastic deformation is capable of increasing the strength of grade 4 up to 500°C. Above this temperature the severely deformed microstructure was seen to demonstrate complete recovery. The strain rate sensitivity within the warm tension of severely deformed titanium was shown to be strongly temperature dependent, where strain rate sensitivity increased with the increase in deformation temperature. With the rise of temperature, void coalescence and growth of dimples takes place which was attributed to the higher rate of diffusion and growth of recrystallized grains at higher deformation temperatures. Studying monotonic behavior of severely deformed titanium is not the only aim of this research work. Another important aspect is the fatigue behavior. Therefore, cyclic deformation response of coarse-grained and ultrafine-grained grade 4 titanium was probed by low cycle fatigue experiments at elevated temperatures up to 600°C and at strain amplitudes of 0.2-0.6%. It was found that cyclic stability strongly depends on grain size and volume fraction of high angle grain boundaries. Severely deformed titanium showed an improvement in fatigue performance at or below 400°C. Electron backscatter diffraction assisted microstructural findings were seen to stand in reasonable agreement with cyclic mechanical results, where micrographs revealed the occurrence of recrystallization and grain growth at 600°C. Cyclic characteristics of two processing routes, defined as route-8E and 8Bc, were also compared. Accordingly, cyclic deformation response of severely deformed titanium was not sensitive to changing routes in the examined range of temperatures and strain amplitudes. Last but not least, the impact of severe plastic deformation on the tensile and damping properties of titanium grade 4 was also explored to complete this study more effectively. Grain refinement via severe plastic deformation enhanced the strength at room temperature while this effect diminished at a high temperature of 600°C. Ultrafine-grained titanium showed an improvement in damping capacity over its coarse-grained counterpart. High damping capacity of the former was rationalized with the contributions of increased dislocation density and high angle grain boundary fraction. As a summary of this research, thermo-mechanical behavior of severely deformed titanium investigated in detail by demonstrating the thermal stability under monotonic and cyclic loading. It was found that severely deformed grade 2 and grade 4 titanium are monotonically stable up to the homologous temperature of 0.45 and 0.40, respectively. Severely deformed grade 4 titanium also showed cyclic stability up to the homologous temperature of 0.35. These findings form the base for the utilization of ultrafine-grained titanium in high temperature applications.
Metadata only
Automated refinement of models for model-based testing
(2017-07) Gebizli, Ceren Şahin; Sözer, Hasan; Aktemur, Barış; Uğurdağ, Hasan Fatih; Briand, L.; Yılmaz, C.; Department of Computer Science; Gebizli, Ceren Şahin
Model-Based Testing (MBT) enables automatic generation of test cases based on models of a system. It has been successfully applied in various application domains, each of which might introduce specific challenges. In this dissertation, we introduce methods and tools for addressing some of these challenges for the consumer electronics domain. In particular, we focus on the testing of Digital TV systems as our case study. We identified the following 3 problems in this context: i) Models of the system are created based on requirement specifications, which are often incomplete and imprecise. Therefore, these models are subject to accidental omissions of certain system behavior. As a result, critical faults can be left undetected by the generated test cases. ii) Resources are extremely limited in the consumer electronics domain. It is not feasible to attain an extensive coverage of test models. iii) A product family in consumer electronics often includes hundreds of systems. The set of features can highly differ among these systems. Therefore, the MBT process and modeling must be flexible to systematically manage variability and increase the amount of reuse for test models. To tackle the first problem, we introduce an approach and tool for automatically extending test models based on a set of collected execution traces. These traces are collected during Exploratory Testing (ET) activities. Several critical faults were detected in 3 case studies after generating test cases based on extended models. These faults were not detected by the initial set of test cases. They were also missed during the ET activities. As a solution for the second problem, we iteratively update test models in 3 steps to focus the test case generation process only on execution paths that are liable to highly severe failures. We use Markov Chains as test models, in which transitions among states are annotated with probability values. First, we update these values based on usage profile. Second, we perform an update based on fault likelihood that is estimated with static code analysis. Our third update is based on error likelihood that is estimated with dynamic analysis. We generate and execute test cases according the updated values after each iteration of updates. New faults can be detected after each iteration. To address the variability problem, we document variations among tested systems explicitly with a feature model. We map optional and alternative features in the feature model to a set of states in the test model. Transition probabilities in the test model are updated according to the selected features so that the generated test cases focus only on these features. This approach facilitates the reuse of a test model for many systems.
Metadata only
Using eigenvoices and nearest-neighbours in HMM-based cross-lingual speaker adaptation with limited data
(2017-08) Sarfjoo, Seyyed Saeed; Demiroğlu, Cenk; Demiroğlu, Cenk; Şensoy, Murat; Uğurdağ, Hasan Fatih; Saraçlar, M.; Güz, Ü.; Department of Computer Science; Sarfjoo, Seyyed Saeed
Thesis abstract: Cross-lingual speaker adaptation for speech synthesis has many applications, such as use in speech-to-speech translation systems. Here, we focus on cross-lingual adaptation for statistical speech synthesis systems using limited adaptation data. We propose new methods on HMM-based and DNN-based speech synthesis. To that end, for HMM-based speech synthesis we propose two eigenvoice adaptation approaches exploiting a bilingual Turkish-English speech database that we collected. In one approach, eigenvoice weights extracted using Turkish adaptation data and Turkish voice models are transformed into the eigenvoice weights for the English voice models using linear regression. Weighting the samples depending on the distance of reference speakers to target speakers during linear regression was found to improve the performance. Moreover, importance weighting the elements of the eigenvectors during regression further improved the performance. The second approach proposed here is speaker-specific state-mapping which performed signicantly better than the baseline state-mapping algorithm both in objective and subjective tests. Performance of the proposed state mapping algorithm was further improved when it was used with the intra-lingual eigenvoice approach instead of the linear-regression based algorithms used in the baseline system. We propose new unsupervised adaptation method for DNN-based speech synthesis. In this method, using sequence of acoustic features from target speaker, we estimate continuous linguistic features for unlabeled data. Based on objective and subjective experiments, adapted model outperformed the gender-dependent average voice models in terms of quality and similarity.
Metadata only
Immersion cooling of suspended and coated nano-phosphor particles for extending the limits of optical extraction of light emitting diodes
(2017-08) Tamdoğan, Enes; Arık, Mehmet; Arık, Mehmet; Yaralıoğlu, Göksenin; Başol, Altuğ; Budaklı, M.; Graham, S.; Mechanical Engineering; Department of Mechanical Engineering; TAMDOĞAN, Enes; Tamdoğan, Enes
Energy efficiency, long life, exceptional color and performance of solid-state light (SSL) sources have resulted in a rapidly increasing trend in a number of practical applications especially for general lighting after a long history of incandescent lamps. However, LEDs, as a solid state lighting technology, have some limitations as typical electronics so thermal management is vital for LEDs. Moreover, predicting the life and the light quality is essential to assess and enhance the performance of LED performance. Specifically, to improve the heat dissipation one major parameter used to evaluate the LED performance is thermal resistance. The major obstacle in estimating the thermal resistance for LEDs is the accuracy of determining the junction temperature especially for high power LEDs. Ideally, the junction temperature is determined reliably by monitoring the device temperature at a position close to the junction. That could be achieved with small temperature sensors that are placed very close to the junction. But there are still physical limitations to this method due to the sensor itself would be larger than the junction, which would result in an additional error to the measurement and will not be very useful in most applications. On the other hand, to convert blue to white light, GaN LEDs are usually encapsulated with a phosphor-epoxy mixture that cures into a soft material at high temperatures. During LED operation, significant self-heating occurs causing the glass-like epoxy to undergo large displacements due to its high thermal expansion coefficient at a critical temperature. This enhanced displacement inside the LED package may fracture the gold wire bonds and ultimately lead to device failure. Thus, the inclusion of phosphor into a high brightness LED package is another complex task. In a typical 450 nm or 470 nm blue LED, YAG phosphor is introduced and combined with the blue emission to create what appears to the eye is white light. While, the common practice for phosphor carrier medium is typically a silicone with high refraction index, the geometry of the phosphor is a primary design variable and can be classified as dispersed (dispersed inside the liquid coolant as particles), remote (remote coated under the dome), and local (coated over chip). In each case the geometry greatly affects the ultimate optical output of the LED color qualities and conversion efficiency. Since there is limited information about individual losses in an LED system and there is no available correlation to predict the total losses, this study is for filling the gap for both of these fundamental problems; estimating the thermal resistance and increasing the ultimate optical output of an LED with different coating technologies. Thus, the junction temperatures are first measured in the current work with Raman Spectroscopy, Infrared (IR) imaging as well as Forward Voltage method (FVM) for a 455 nm bare blue LED chip (without any phosphor coating). Then, the same samples have been coated with a phosphor-epoxy (13%, 4300 CCT) mixture to convert blue light into white light. After that junction temperatures were measured experimentally with the previously mentioned three methods and compared to each other. While IR imaging shows better capability on capturing the possible hotspots over the surface, Raman and Forward Voltage methods were in reasonably good agreement on measuring the junction temperature for 455 nm blue (uncoated) LED chip. However, the measurements performed after coating have shown slightly different results with Infrared (IR) imaging and Raman method, while Forward Voltage method has still shown meaningful results for coated chips. To be able to have a reasonable comparison, all cases have been measured with FVM simultaneously with one other mentioned method as in couple. As the second phase of the work, after identifying the junction temperatures accurately, LED effective liquid cooling is examined since it is tackling a major challenge "hot phosphor losses" that provides unique information for both fundamental nano-fluid (phosphor based) heat transfer and improved light conversion efficiency. Thus, topside liquid cooling with optically-transparent liquids is utilized to reduce average chip temperatures and to improve the uniformity of chip and phosphor temperature, leading to higher light extraction efficiencies. Furthermore, computational models and experimental studies of heat transfer and optical behavior to validate modeling results were performed for three proposed coating configurations (coated over chip, dispersed inside the liquid coolant as particles and remote coated under the dome) with di-electric liquid cooling. While, the phosphor is in direct contact with the LED chip in the current coating applications, it has shown higher junction temperatures beside of lower conversion efficiency and possible color shifts. The dispersed phosphor idea inside the liquid coolant as particles has resulted with lower conversion efficiency beside of any important thermal enhancements on the LED junction. However, the phosphor in the remote coating system is not affected by the LED temperature and thus maintained a consistent conversion rate and overall color point. Moreover, the remote phosphor with immersion cooling system has extended the Lumen Extraction Limits of White LEDs in excess of 53%, as long as the remote-phosphor system is well designed.
Metadata only
Radiative cooling by spectrally selective materials for building
(2018-01-03) Family, Roxana; Mengüç, Mustafa Pınar; Mengüç, Mustafa Pınar; Mengüç, Mustafa Pınar; Bundur, Zeynep Başaran; Çelik, S.; Şendur, K.; Başol, Altuğ Melik; Department of Mechanical Engineering; Family, Roxana
Buildings utilize more than one-third of the total energy consumed in counties within the Mediterranean climate zones like in Turkey. Particularly during the summer months, the absorption of solar energy by the buildings increases the required cooling load profoundly. In warmer zones, and in Mediterranean countries, air conditioning applications are becoming more common with every passing year, with their sizable negative impact on energy use. A possible solution to this problem is the radiative cooling of the building surfaces and roofs. This requires tailoring of the radiative properties of surfaces to decrease or increase their natural ability to absorb, emit, or reflect radiant energy. It is favorable to have the utmost emission from the surface with the highest reflection of solar energy and that is for situations where a surface is to be kept cool while exposed to the sun. Note that the Earth’s atmosphere is relatively transparent between the wavelength of 8-13 m; therefore, buildings emitting this is called as “transparency window” for electromagnetic waves. This window allows the radiation emitted by the earth to escape to space with no absorption within the atmosphere. This spectral energy loss, versus the radiation absorbed by the Earth is the reason for the atmospheric radiation cooling. If a building surface emits mostly in this window, than the building can be cooled effectively as well. For daytime radiative cooling which was the goal of this study, coating or painting an object with a strong solar reflector can be considered but significantly mutate its color, which may not be desired. By manufacturing a surface that had an absorptivity large in the spectral region of short wavelengths about the peak solar energy, but small in the spectral region of longer wavelengths where the peak surface emission would occur, it might be possible to absorb almost as a blackbody while emitting very little energy that such surfaces are called “Spectrally selective”. Spectrally selective surfaces can also be useful where it is required to cool an object exposed to incident radiation from any high-temperature source. These situations are objects subjected to the sun, such as the roof of a building. In this study, for the first time sustainable and economically viable materials for radiative cooling in the buildings and the roofs were developed. In the first stage a large group of materials mostly sustainable materials were selected and the morphology and optical properties of them (by optical microscope, UV-Visible, and FTIR) were obtained. For the next stage six sustainable materials were chosen. To evaluate the radiative cooling potential of the samples, the power of cooling was calculated and the results were compared with the selective and broadband emitters. Furthermore, the power of the cooling for the summer and winter time, in daytime and night time cases were also calculated and compared with each other. Heat transfer through most materials is not just a surface phenomenon, but it also needs a volumetric analysis. Therefore, a coupled radiation and conduction heat transfer analysis was solved for all six selected samples. Results are discussed for the selection of the best materials, for different applications on building surfaces. Meanwhile, coupled conduction and radiation was solved for two cases of one layer (concrete roof) and three layers (concrete with soil and the moss on the surface of it) and at last the results were compared with each other.
Metadata only
Runtime verification of internet of things using complex-event processing (RECEP)
(2018-06) İnçki, Koray; Arı, İsmail; Arı, İsmail; Sözer, Hasan; Aktemur, Tankut Barış; Baydere, Ş.; Aktaş, M. S.; Department of Computer Science; İnçki, Koray
Increase in the computing power and memory accompanied with decreasing architectural footprints has enabled conquering new frontiers in proliferation of technology in the next industry revolution. More autonomous systems have been deployed thanks to the advancing capabilities provided by embedded systems with such computing power. Internet of Things (IoT) has emerged as an enabler of many achievements in the industry through presenting a seamless integration of computing units, usually in the form of an embedded system, by allowing interconnection of such embedded systems without requiring human interaction. Engineering a system of systems (SoS) constituted by IoT devices has been the new challenge of designing large scale systems, as the scale of such a system could range from tens of devices in an ambient assisted living (AAL) example to thousands of devices in a smart city application. Therefore, the complexity of software engineering and veri fication of those SoS's necessitates new approaches that would facilitate those processes. In this thesis, we tackle the problem of verifying IoT SoS's at runtime. We first propose an event calculus that captures the fundamental behavioral model of IoT messaging primitives. The event calculus allows us to specify interaction of IoT devices in terms of events that represent sending and receiving Constrained-Application Protocol (CoAP) messages. Representing the behavior of CoAP endpoints in EC helps us defi ne complex-event processing (CEP) patterns that will later be used as runtime monitors. Existing research on runtime verifi cation (RV) usually presents a solution with heavy formal methods, which hinders the usefulness of method by intimidating the practitioners. We, therefore, propose a model-driven engineering (MDE) approach for RV of IoT systems, which is expected to promote the utilization of RV in industrial scenarios. We propose an extension to the UML2.5 profi le, which enables us to customize a modeling tool so that we can develop a domain-specifi c model (DSM) for verifying IoT systems. Later, in order to allow automatically generating runtime monitors in the form of CEP statements, we contribute a model-to-text (M2T) transformation utility in the modeling tool. The contributions of the thesis are demonstrated in several case scenarios.
Metadata only
Channel modeling and characterization for visible light communications: indoor, vehicular and underwater channels
(2018-06) Miramirkhani, Farshad; Uysal, Murat; Tekin, Ahmet; Uğurdağ, Hasan Fatih; Başar, Ertuğrul; Baykaş, Tuncer; Department of Electrical and Electronics Engineering; Miramirkhani, Farshad
Despite the increasing attention on visible light communications (VLC) systems, there is a lack of proper visible light (VL) channel models. This is a serious concern since channel modeling is the very first step for efficient, reliable, and robust VLC system design. This dissertation focuses on channel modeling and characterization study for indoor, vehicular and underwater VLC. Our study is based on Zemax®; a commercial optical and illumination design software. Although the main purpose of such software is optical system design, we take advantage of the ray tracing features of this software which allows an accurate description of the interaction of rays emitted from the lighting source within a specified confined space. The simulation environment is created in Zemax® and enables us to specify the geometry of the environment, the objects within as well as the specifications of the sources (i.e., LEDs) and receivers (i.e., photodiodes). For a given number of rays and the number of reflections, the non-sequential ray tracing tool calculates the detected power and path lengths from source to detector for each ray. These are then imported to Matlab® and processed to yield the channel impulse response (CIR). In contrary to existing works which are mainly limited to ideal Lambertian sources and purely diffuse reflections, our approach is capable to obtain CIRs for any non-ideal sources as well as specular and mixed specular-diffuse reflections. Furthermore, we can precisely reflect the presence of objects and wavelength-dependent reflection characteristics of surface materials in channel study. In the first part of this thesis, we propose a realistic indoor channel modeling approach and carry out a detailed channel characterization study. We also investigate the effect of user mobility and receiver orientation on CIRs. In the second part of this thesis, we present VLC channel models for vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) taking into account the asymmetrical pattern of headlamp and street lights, reflections from road surfaces and weather conditions. We further develop a closed-form path loss expression for V2V VLC channel for different weather conditions. In the last part of this thesis, we carry out a detailed underwater optical channel modeling and characterization study taking into account the reflection characteristics of the sea surface and sea bottom as well as the water characteristics, i.e., extinction coefficient, and scattering phase function of particles. We develop a closed-form path loss expression as an explicit function of water type, beam divergence angle and receiver aperture diameter and validate the accuracy of the proposed expression through Monte Carlo simulation results.
Metadata only
Turbulence and flame interaction for control of flame location in diffuser combustor
(2018-08) Nazzal, Ibrahim Thamer Nazzal; Ertunç, Özgür; Ertunç, Özgür; Başol, Altuğ; Mengüç, Mustafa Pınar; ÖZdemir, B.; Koşar, A.; Department of Mechanical Engineering; Nazzal, Ibrahim Thamer Nazzal
Achieving an appropriate flame location is desirable in many combustion system applications especial in the design of the combustion system. Given that the characteristics of turbulent flows influence flame behavior and the design of combustion systems, flame location can be controlled within desirable levels without distorting the design of the combustor geometry by selecting suitable characteristics. A feedback control method utilizes the characteristics of turbulent flows to stabilize the flame location within the desirable level. The primary objective of this study is to introduce a strategy for flame location control using characteristics of turbulence flow. Turbulence intensity and length scale are among the main parameters of turbulent flow. Therefore, the secondary goal of this study is to investigate the influence of turbulence intensity and length scale on flame location. The investigation of the dependency of flame location on turbulence is based on selecting suitable combustor geometry. For this purpose, the axisymmetric diffuser form is used to reveal the response of the flame location of a turbulent premixed flame that has been exposed to various turbulence intensities and length scales. The diffuser is selected because the flow slows down along the direction. Thus, the flame is expected to propagate towards the inlet when the flame speed increases. In this manner, the effect of turbulence can be studied without changing the thermal power. In addition, the diffuser combustor is used to avoid blow-off and flame extinction because the flow slows along the combustor. Two types of diffuser combustors are selected for this study. The first combustor is a cylindrical diffuser, while the second one is a cylindrical diffuser with a conical insert. Numerical simulations are applied to the diffuser combustor for turbulent premixed propane flames by using a coherent flame model integrated to the Reynolds-averaged Navier–Stokes flow model with k-epsilon turbulence model. Firstly, the influence of turbulence on flame location in a diffuser-type combustor is studied under steady-state conditions. Results show that the flame location moves towards the inlet of the diffuser combustor with the increase in turbulence intensity for moderate- and high-turbulence length scales. The behavior of flame location is different in the low-turbulence length scale. The flame location initially decreases with the increase in turbulence intensity and subsequently stabilizes. Furthermore, the flame area density influences the flame location with the increase in turbulence intensity and turbulence length scale. Turbulence intensity and length scale simultaneously influence the flame area density, flame shape, and flame location. Secondly, the results of the unsteady simulations indicate that turbulence intensity, length scale, and flow separation exert a significant effect on the flame location of the premixed turbulent combustion. The flame front moves toward the diffuser inlet as a result of the increase in turbulence intensity and length scale. The flame location drops to the middle of the diffuser for the high turbulence intensity. However, the effect of turbulence intensity is more visible than that of turbulence length scale within the tested range. An increase in turbulent length scale at a constant turbulence intensity causes a decrease in flame location. It is observed that the combustion and inlet turbulence cause a flow separation mainly downstream of the flame front. Consequently, the secondary flow structures influence the flame topology and location. Therefore, the flame location and shape are influenced by the flow separation and the turbulence intensity and length scale. Thirdly, a conical insert is placed in the middle of a diffuser-type combustor to eliminate the flow separation. The influence of turbulence on flame location in two diffuser-type combustors (with and without conical insert) is studied and compared. Results indicate that the flame moves towards the inlet of the diffuser with the increase in turbulence intensity and length scale in the two diffuser-type combustors. At a high-turbulence length scale, the flame rapidly drops at the inlet of the diffuser with a conical insert with the increase in turbulence intensity, whereas the flame drops to an intermediate level when the diffuser is not used with a conical insert. Moreover, a similarity was observed in the trends of the flame location at low turbulence intensities in both cases. Results show that the Taylor-scale Reynolds number is the influential parameter of flame location and not turbulence intensity and length scale. An increase in the Taylor-scale Reynolds number leads the flame location to move towards the combustor inlet. The flame drops to the inlet of the combustor at a high-turbulence Taylor-scale Reynolds number. Flow separation is observed in the diffuser without a conical insert, and flow separation is eliminated by using the conical insert. Finally, the control of the flame location in the diffuser combustor is studied under various turbulent flow characteristics. A control strategy is suggested for this purpose. Control algorithm written as a Java macro is implemented to a commercial CFD software, namely STAR CCM+. This framework is utilized to perform all the simulations for the premixed turbulent flame under unsteady-state controlled conditions. The algorithm is built to adjust the turbulent kinetic energy and turbulent dissipation rate. Feedback control is introduced to stabilize the flame location at the desired level. Results indicate that control of the turbulent kinetic energy at the inlet control the flame location within the targeted level. In addition, it is observed the flame location moved to a low level for high turbulent kinetic energy whilst it moved to the high level for low turbulent kinetic energy.
Metadata only
Effect of pH on particle agglomeration and radiative transfer in nanoparticle suspensions
(2018-08) Al-Gebory, Layth Wadhah Ismael; Mengüç, Mustafa Pınar; Mengüç, Mustafa Pınar; Başol, Altuğ; Ertunç, Özgür; Koşar, A.; Şendur, K.; Department of Mechanical Engineering; Al-Gebory, Layth Wadhah Ismael
Nanoparticle suspensions (NPSs) are solid-fluid mixtures where small dielectric or metallic particles (with sizes <100 nm) used in a base fluid. NPSs have unique and tunable thermo-optical properties, and for that reason they can be used extensively to improve the thermal efficiency of different systems where they show remarkable enhancement in heat transfer compared with those of a base fluid. The effectiveness of solar thermal systems used for photo-thermal energy conversion is measured by their ability of absorb radiative energy by the working medium; for such applications NPSs are much better choice than traditional fluids. NPSs have also been used in coatings as they can be tuned to improve or alter the appearance of an object, as radiative and optical properties play significant roles. Although NPSs are considered very promising for these applications, there is some concern about their stability and their long-term use. Particle agglomeration in NPSs remains one of the most important challenges faced in terms of their usage. In all of these applications, the pH value and its effects on the particle agglomeration may have significant impact on the nanoparticles stability behavior, and consequently on the radiative transfer of energy. Steric and electrostatic stabilization methods are among the two approaches used for particulate suspensions to avoid such problems. In thermal applications, especially in high temperature ones, electrostatic stabilization method is usually preferred. In this dissertation, both experimental and theoretical investigations were carried out to determine the stability and optical properties of individual (water/TiO2 and water/Al2O3) and hybrid (water/TiO2+Al2O3) nanoparticle suspensions. The experimental studies include the preparation, characterization, and optical property measurements of the nanoparticle suspensions. The impact of the electrostatic stabilization (zeta potential and pH values) on the size and structure of particles due to agglomeration behavior are explored. The particle size distribution and the average (effective) particle agglomerate size for the nanoparticle suspensions in different conditions (the pH and particle volume fraction) were measured by using the dynamic light scattering (DLS) technique. The effects of the different particle agglomerates under different pH values on the dependent and independent scattering and their boundaries are investigated and demarcated for different conditions, where the relationship between the distance between particle to particle surface and the incident wavelength for different particle types are explored. The effects of particle agglomeration (similar and dissimilar particle agglomerations), particle size distribution and their contributions to the radiative properties of the nanoparticle suspensions are determined using the UV-Vis spectroscopy technique. The numerical part included the study of the optical and radiative properties and thermal radiation transfer based on the average (effective) particle agglomerate size obtained from the experimental studies. The optical and radiative properties of nanoparticle suspensions are calculated based on the Lorenz-Mie theory applying the single-scattering approximation technique. The influence of the particle size distribution on the scattering coefficient of nanoparticle suspensions is studied theoretically to account for the effect of compact particle agglomerates. The thermal radiation transfer in the nanoparticle suspensions is assessed by solving the radiative transfer equation using the discrete ordinates method, where the volumetric radiative heat flux and the thermal flux efficiency are calculated. The results show the impact of pH value on the stability of individual and vi hybrid nanoparticle suspensions. The different particle agglomerate types, sizes, and shapes yield different behavior of suspensions, including their stability or sedimentation rates, which help formation of optically thicker media. Light scattering in such media is significantly different as a function of the proximity of particles to each other. If they are closer to each other roughly less than dominant wavelength of the radiation, then their behavior is defined as dependent scattering, which is explored in this study. It is shown that a significant enhancement in the radiative properties, specifically in the UV/Vis spectrum, can be observed , which has an important effect on the thermal radiation transfer of the incident solar radiation. The demarcation of dependent and independent scattering regimes is explained for the individual and hybrid nanoparticle suspensions based on their pH value. NPSs with different effective particle agglomerate sizes have a considerable effect on the volumetric radiative heat flux, where the losses in radiative energy were decrease in comparison to those of pure water. The results also show the effects of composite particle agglomerates in the hybrid nanoparticle suspensions on the radiative properties, which are produced from dissimilar suspended particles. The results of this dissertation show that the pH value has a dominant effect on the radiative transfer involving nanoparticle suspensions, compared to other parameters. Adjusting the pH value based on the isoelectric point of the nanoparticle is an efficient method when specific radiative properties are required for specific applications. Such impact of pH value on optical and radiative properties of NPSs is studied for the first time in the literature.
Metadata only
Thermal and radiative energy/exergy analyses of parabolic trough collector systems
(2018-08) Mohammed, Hayder Noori; Mengüç, Mustafa Pınar; Mengüç, Mustafa Pınar; Başol, Altuğ; Ertunç, Özgür; Şendur, K.; Koşar, A.; Department of Mechanical Engineering; Mohammed, Hayder Noori
The concept of exergy is used to determine the maximum energy that can be extracted from a system. It is based on both the first and the second laws of thermodynamics and allows us to determine the irreversibilities throughout a process and the losses from the system. In this dissertation, the fundamentals of spectral radiative exergy are developed and applied to determine the maximum conversion of solar energy in concentrated solar power (CSP) systems. There are five primary objectives of this study. First, a new formulation is developed for the maximum efficiency of the solar radiation conversion by considering the radiative energy transfer between two surfaces at different temperatures for a constant volume system. Exergy of spectral radiative transfer is determined, and the formulation for the exergy efficiency maximization is presented in a direct and practical manner. For the calculation of maximum efficiency, the mean temperature of the environment and the sink temperature are used. Second, a new methodology is presented for spectral radiative energy and radiative exergy calculations to evaluate the performances of CSP systems. Spectral radiative properties and the operating temperature of selective surfaces, along with the temperature of the environment, are considered in these analyses. The fundamental quantities needed for the spectral radiative energy and radiative exergy formulations are introduced, and then the spectral performances of five selective coatings are assessed. The spectral analysis is performed in the wavelength range of 250 nm to 20,000 nm, while thermal analysis is carried out for the temperature range of 325 K to 800 K. The third objective is to introduce a new approach for estimating the exergy value of the monthly average daily horizontal global radiation, including several parameters, as the monthly average daily value of the horizontal extraterrestrial radiation, the number of sunny hours, the day length, the mean temperature and the mean wind velocity. Seven statistical parameters are used to validate the accuracy of all models. The concept is applied to four locations in Iraq and Turkey, to help predicting the maximum available solar radiation based on different weather parameters. The fourth objective is to outline a comprehensive energy analysis for a parabolic trough collector (PTC) system. The analysis considers all heat transfer modes, optical components, and the details of spectral absorption and reflection of solar radiation on the glass envelope. The energy performance of the PTC system is investigated using five gases in an annular space, five selective coatings of the absorber surface, and four common heat transfer fluids following a two-dimensional approach. A model is built using Engineering Equation Solver (EES). The results obtained are compared against the available results from experimental tests and analytical models. This analysis shows the effects of the properties of the absorbing gas, the selective coating and the working fluid on the energy performance of PTC as the key parameters of energy for various operating conditions. The fifth objective of the study is to establish a methodology to analyze PTC systems using the principles of spectral radiative exergy. The fundamental relations for spectral exergy analyses are derived starting from the first and second law of thermodynamics, and the key performance parameters, including exergy losses, destructions, consumption and efficiency are determined using the same parameters mentioned above in the fourth objective. It is noted that the exergy destruction is directly related to irreversibility throughout processes while the exergy losses are due to the thermal and optical losses. Based on these findings, an improvement of PTC design parameters are discussed.
Metadata only
Tools and tecniques for implementation of real-time video processing algorithms
(2018-09) Levent, Vecdi Emre; Uğurdağ, Hasan Fatih; Uğurdağ, Hasan Fatih; Uysal, Murat; Kıraç, Furkan; Demir, O.; Aydın, N.; Department of Computer Science; Levent, Vecdi Emre
Hardware implementation of video processing algorithms, which are usually real-time by nature, need architectural exploration so that we achieve the required performance with minimal cost. In addition, the video algorithm to be implemented may need to be used with di erent frames-per-second and resolution in di erent applications. Hence, we usually need to design a parameterized IP block instead of a xed design. Also, during the hardware design process, the requirements fed from the algorithms team may change as well as the algorithm itself. As a result of these, hardware implementation iterations need to be as fast as the algorithms development iterations. This is only possible with the use of tools and techniques speci cally geared towards hardware design generation for video processing. The tools and techniques discussed in this dissertation include host software, FPGA interface IP, HLS, RTL generation tools, an architectural estimation tool, ow based veri cation approach, and logic synthesis automation as well as architectural concepts (e.g., nested pipelining). The architectural estimation tool estimates many design metrics. These metrics are area, throughput, latency, DRAM usage, interface bandwidth, temperature, and compilation time. While we explain the above tools and techniques within a speci c use case, namely, optical ow, we also present results from another use case, image fusion. Using our methodology and tools, we were able to design and bring up to 11 versions of optical ow and 3 versions of image fusion on 3 di erent FPGAs from 2 di erent vendors. The rst version of these designs (hence the generators) took several months; however, the subsequent design versions each took a few days with a few people. In the case where only architectural trade-o is needed, we were able to generate and synthesize around one thousand designs in a single day on a 48-core server.
Metadata only
Investigation of flame characteristics in a turbulent premixed combustion
(2018-10) Alhumairi, Mohammed Khudhair Abbas; Ertunç, Özgür; Ertunç, Özgür; Başol, Altuğ; Mengüç, Mustafa Pınar; Koşar, A.; Özdemir, B.; Department of Mechanical Engineering; Alhumairi, Mohammed Khudhair Abbas
In this thesis, the turbulent flame closure (TFC) model and coherent flame model (CFM) of turbulent premixed flames as steady state flow are used. In addition, large eddy simulation (LES) model as unsteady state flow is used. The effects of different turbulent parameters such as Reynolds number based in a Taylor micro scale , turbulence length scale turbulence intensity and the constant of the TFC model on the combustion are modelled for steady reacting flow. In addition, the characteristics of sinusoidal wave, such as amplitude of pulsation and the frequency are used for unsteady reacting flow to show the behavior of the flame topology and flame location of jet flow combustor of lean propane-air combustion. The simulations are achieved with 3, 5 and 9 kW thermal loads at constant inlet velocity and equivalence ratio. Transport equations for progress variable (c) are shown in terms of Reynolds and Favre averages, and the reaction rate terms are used to calculate heat release at different turbulent flow conditions. The results are compared with existing experimental data from the combustor performance studies. The lean premixed combustion under the influence of active grid turbulence was computationally investigated, and results were compared with the experiments. In the experiments, the transverse and longitudinal active grids generated turbulence. The experiments were conducted to generate a premixed gas flame at a given inlet power 3, 5 and 9 kW. Turbulent burning modelling such as CFM, TFC and LES models were implemented to conduct simulations under different turbulent flow conditions as steady and unsteady state flows, respectively. The turbulent flow conditions obtained in the simulations were specified by the dissipation rate of turbulence ( ) and a turbulent kinetic energy ( ) at the inlet region for steady reacting flow. All simulations were used to simulate the turbulent reacting flows at the equivalence ratios of 0.606 and 0.588 to estimate the combustion conditions of the propane. The heat release field was used for comparison with experimental cases. Acceptable agreement is found between the simulations and the experimental results. The flame topology is more sensitive to turbulence in CFM model than that simulated by the TFC model, and the flame location moved toward to inlet region by increasing . CFM and TFC models were used as a fundamental parameter. In addition, in the LES model, the turbulence was attained by setting the characteristics of a sinusoidal wave such as and . Three numerical models were used to prediction the flame topology and flame location at different turbulent flow conditions, and three different results were found as compared with experiments. Moreover, the fields of heat release, species mass fraction and temperature distribution in the centerline of the combustor were investigated. After the TFC model was calibrated, the best value of constant that matches the experiment was A = 0.37. All numerical simulations were performed in STAR CCM+ v10.02 and v12.04 software. The turbulent flame speed was derived from the Zimont formula, and the results showed that the flame location and topology were influenced solely by , as suggested by the derived new equation for . The results showed that combustion occurs in the wrinkled and corrugated flamelet regions on the Borghi diagram. At a low value, the flame topology in the TFC model was wrinkled and symmetrical with respect to the vertical axis of the combustor, whereas at medium and large values, the flame topology exhibited cusps. By contrast, the flame topology behaviour in the CFM model was not constant at different and was like a mushroom shape, and the flames moved toward inlet regions by increasing . In addition, in the LES model, the V-shape and the corrugated wings of the flame were formed. The flame changed topology and location at different turbulent flow conditions of amplitude of pulsation and frequency. The flame topology investigation for jet flow combustor can be used to modulate effectively well the gas turbine burner design or other turbulent combustion studies. The investigation of the flame topology in the combustor with various turbulent flow conditions is important in controlling the flame location to reduce emissions and increase power efficiency, or even design pioneering production techniques related to flame.
Metadata only
A semantic policy framework for internet of things
(2018-10-31) Göynügür, Emre; Şensoy, Murat; Şensoy, Murat; Tekin, Ahmet; Sözer, Hasan; Alkaya, A. F.; Özgür, A.; Department of Computer Science; Göynügür, Emre
With the proliferation of technology, connected and interconnected devices (henceforth referred to as IoT) are fast becoming a viable option to automate the day-to-day interactions of users with their environments. However, with the explosion of IoT deployments we have observed in recent years, manually managing the interactions between humans-to-devices, and especially devices-to-devices, is an impractical task, if not an impossible task. This is because devices have their own obligations and prohibitions in context, and humans are not equipped to maintain a bird's-eye-view of the interaction space. Motivated by this observation, in this thesis, we propose a semantic policy framework that (a) supports representation of high-level and expressive user policies to govern the devices and services in the environment; (b) provides e cient procedures to re ne and reason about policies to automate the management of interactions; and (c) delegates similar capable devices to ful ll the interactions, when con icts occur. We then describe how to combine ontology-based policy reasoning mechanisms with in-use IoT applications to customize and automate device behaviors and discuss how the policy framework can be extended with data federation to handle diverse and distributed data sources. We demonstrate that smart devices and sensors can be orchestrated through policies in diverse settings, from smart home environments to hazardous workplaces, such as coal mines. Lastly, we evaluate our approach using real applications with real data and demonstrate that our approach is scalable under high load of data and devices.
Metadata only
Energy and exergy efficiency analyses of high-performance buildings
(2018-11-26) Al-Doury, Raaid Rashad Jassem; Mengüç, Mustafa Pınar; Mengüç, Mustafa Pınar; Kundakçıoğlu, Erhun; Başol, Altuğ Melik; Özyurt, T. O.; Çelik, S.; Department of Mechanical Engineering; Al-Doury, Raaid Rashad Jassem
Reducing building energy density has become one of the global requirements to a decrease fuel consumption and emission production, and consequently making our world sustainable. The high value of energy consumed by buildings highlights the importance of the requirement to decrease building energy demand. The aim of this thesis is to analyze a well-designed existing building exegetically, exergo-economically and environmentally in order to determine the consequent effects of any options for improvement. In the first part of the study, the building was analyzed statically and dynamically (hourly) over a year for the heating season to specifically highlight the differences between them in addition to an accuracy estimation. In the second part, actual experimental processes for improvement that were applied to the buildings at different stages were investigated as well. For this study, the SCOLA Building at Ozyegin University Campus in Istanbul is considered. SCOLA was designed to be one of the least amounts of energy consuming buildings in Turkey. It includes 291 rooms with a floor area of 17,250 m2 and 6 floors. A natural gas boiler that produces hot water is used to heat the building with a distribution network and four fan coils. The hourly operational information about the heating system was recorded and used in tandem with the local weather data. The calculations were applied both in static and dynamic fashion, based on the recommendations from the facility management team. Improved pre-design tools were used to implement these details to the simulations. The simulations reveal that energy demand for the building can be as low as 1.38 MW and exergy is 1.34 MW at peak load, while their annual values are 8.9 GJ and 8.2 GJ, respectively. Based on the calculations, the specific heating demand at the building is found to be 25 W/m2 while the annual specific heating demand is 80 W/m2/yr (if heating is evenly distributed to the year; actual energy density for the building is determined to be 50 W/m2/year, considering summer months). The reduction in energy and exergy demand reaches 6% during working hours of the 21st January 2016, simultaneously reducing the cost and the CO2 emissions. The increase in exergetic cost coefficients reflect the reduction in the exergy efficiency of the heating system components based on the ambient temperature change. It is noted that a dynamic analysis using average monthly temperatures is preferred over a static analysis. However, if a simpler static analysis is to be used, an annual average temperature needs to be identified for a specific climate zone and building type. For Istanbul, an average temperature of 14oC is recommended for a static analysis. We also examine how different engineering implementation strategies, in addition to the original design, can improve the thermal performance of the building. Five different cases (scenarios) are investigated in addition to the original case which is considered to be the base case (1st case). The second is the use of a ground air heat exchanger, whereas the use of better insulation materials and the use of glass and roofs is the third case study representing the traditional approach. The use of solar PV-panels over the entire building constitute the fourth case, and the integration of a campus tri-generation system to the building energy modalities is considered as the fifth case. Applying all these processes to the building simultaneously is considered as the sixth scenario. All these changes have already been implemented in the building where the real time data are being collected. Performance simulations based on exergy, exergoeconomic, and environmental analyses are conducted and presented. Energy and exergy flow diagrams from all sources to the envelope for all cases are also outlined and conclusions are drawn. A Marginal Abatement Cost Curve (MACC) was constructed based on exergy analysis in order to achieve more accurate results. MACC analysis outlines the cost and carbon dioxide emission savings simultaneously. Its results help policymakers to employ the best potential option.
Metadata only
Digital oil refinery: utilizing real-time analytics and cloud computing over industrial sensor data
(2018-12-14) Khodabakhsh, Athar; Arı, İsmail; Arı, İsmail; Şensoy, Murat; Kayış, Enis; Aktaş, M.; Alkaya, A. F.; Department of Computer Science; Khodabakhsh, Athar
This thesis addresses big data challenges seen in large-scale, mission-critical industrial plants such as oil refineries. These plants are equipped with heavy machinery (boilers, engines, turbines, etc.) that are continuously monitored by thousands and various types of sensors for process efficiency, environmental safety, and predictive maintenance purposes. However, sensors themselves are also prone to errors and failure. The quality of data received from them should be verified before being used in system modeling or prediction. There is a need for reliable methods and systems that can provide data validation and reconciliation in real-time with high accuracy. Furthermore, it is necessary to develop accurate, yet simple and efficient analytical models that can be used with high-speed industrial data streams. In this thesis, design and implementation of a novel method called DREDGE, is proposed and presented first by developing methods for real-time data validation, gross error detection (GED), and gross error classification (GEC) over multivariate sensor data streams. The validated and high quality data obtained from these processes is later used for pattern analysis and modeling of industrial plants. We obtained sensor data from the power and petrochemical plants of an oil refinery and analyzed them using various time-series modeling and data mining techniques that are integrated into a complex event processing (CEP) engine. Next, the computational performance implications of the proposed methods are studied and regimes that are sustainable over fast streams of sensor data are uncovered. Distributed Control Systems (DCS) continuously monitor hundreds of sensors in industrial systems, and relationships between variables of the system can change over time. Operational mode (or state) identification methods are developed and presented for these large-scale industrial systems using stream analytics, which are shown to be more effective than batch processing models, especially for time-varying systems. To detect drifts among modes, predictive modeling techniques such as regression analysis, K-means and DBSCAN clustering are used over sensor data streams from an oil refinery and models are updated in real-time using window-based analysis. In addition, the shifts among steady states of data are detected, which represent systems' multiple operating modes. Also, the time when a model reconstruction is required is identified using DBSCAN algorithm. An adaptive window size tuning approach based on the TCP congestion control algorithm is proposed, which reduces model update costs as well as prediction errors. Finally, we proposed a new Lambda architecture for Oil & Gas industry for unified data and analytical processing over DCS. We discussed cloud integration issues and share our experiences with the implementation of sensor fault detection and classification modules inside the proposed architecture.

Browse

Browsing by Issue Date