Search

Shannon's channel coding theorem states the existence of long random codes that can make the error probability arbitrarily small. Recently, advanced error-correcting codes such as turbo and low-density parity-check codes have almost reached the theoretical Shannon limit for binary additive white Gaussian noise channels. However, designing optimal high-rate short-block codes with automatic bit-labeling for various wireless networks is still an unsolved problem. Deep-learning-based autoencoders (AE) have appeared as a potential near-optimal solution for designing wireless communications systems. We take a holistic approach that jointly optimizes all the components of the communication networks by performing data-driven end-to-end learning of the neural network-based transmitter and receiver together. Specifically, to tackle the fading channels, we show that AE frameworks can perform near-optimal block coded-modulation (BCM) and differential BCM (d-BCM) designs in the presence and absence of the channel state information knowledge. Moreover, we focus on AE-based designing of high-rate short block codes with automatic bit-labeling that are capable of outperforming conventional networks with larger margins as the rate R increases. We also investigate the BCM and d-BCM from an information-theoretic perspective. With the advent of internet-of-things (IoT) networks and the widespread use of small devices, we face the challenge of limited available bandwidth. Therefore, novel techniques need to be utilized, such as full-duplex (FD) mode transmission reception at the base station for the full utilization of the spectrum, and non-orthogonal multiple access (NOMA) at the user-end for serving multiple IoT devices while fulfilling their quality-of-service requirement. Furthermore, the deployment of relay nodes will play a pivotal role in improving network coverage, reliability, and spectral efficiency for the future 5G networks. Thus, we design and develop novel end-to-end-learning-based AE frameworks for BCM and d-BCM in various scenarios such as amplify-and-forward and decode-and-forward relaying networks, FD relaying networks, and multi-user downlink networks. We focus on interpretability and understand the AE-based BCM and d-BCM from an information-theoretic perspective, such as the AE's estimated mutual information, convergence, loss optimization, and training principles. We also determine the distinct properties of AE-based (differential) coded-modulation designs in higher-dimensional space. Moreover, we also studied the reproducibility of the trained AE framework. In contrast, large bandwidth and worldwide spectrum availability at mm-wave bands have also shown a great potential for 5G and beyond, but the high path loss (PL) and significant scattering/absorption loss make the signal propagation challenging. Highly accurate PL prediction is fundamental for mm-wave network planning and optimization, whereas existing methods such as slope-intercept models and ray tracing fall short in capturing the large street-by-street variation seen in urban cities. We also exploited the potential benefits of AE framework-based compression capabilities in mm-wave PL prediction. Specifically, we employ extensive 28 GHz measurements from Manhattan Street canyons and model the street clutters via a LiDAR point cloud dataset and 3D-buildings by a mesh-grid building dataset. We aggressively compress 3D-building shape information using convolutional-AE frameworks to reduce overfitting and propose a machine learning (ML)-based PL prediction model for mm-wave propagation.

The aim of this study is to exploit the gap in entrepreneurship literature by using mixed methods to advance our theoretical and empirical understanding of the role played by the family in developing the entrepreneurial intentions (EI) of Indian youth. This is of specific relevance from an Indian point of view since family in India is considered the cornerstone of its culture, and is considered to be responsible for passing on values and beliefs that are socially acceptable. The study combines an enthnographic study of six families in Delhi with a survey of 1,123 complete responses from final year undergraduates. The ethnographic study advances our knowledge of Ajzen's Theory of Planned Behaviour by identifying four elements of family which influence EI: (i) support and encouragement from grandparents, (ii) support and encouragement from parents, (iii) experience and experimentation with entrepreneurship, and (iv) image of entrepreneurship. Informed by the results of the ethnographic study, structural equation modelling (SEM) was used to formally test the direct relationships between the four elements of family and EI. The study finds strong relationships between Support and Encouragement – Parents to EI and also Experience and Experimentation to EI. The SEM also investigates the mediating role of: (i) Perceived Behavioural Control (PBC), (ii) Personal Attitudes (PA) and (iii) Subjective Norms (SN) between EI and the four elements of family influence. The study has been successful in establishing the crucial role played by an individual's family in the development of their intentions towards entrepreneurship. The researcher hopes that this understanding along with the help of appropriate structural interventions can help to motivate the present young generation who have entrepreneurial aspirations but are constantly challenged and influenced by societal and familial pressures.

Corporate Social Responsibility (CSR) has been conceptualised in a variety of ways and interest groups have approached the subject differently. CSR theorists and practitioners have attributed the lack of a consistent approach to CSR to the absence of uniformity in the manner in which the meaning of CSR has been articulated. This has meant that organisations have operationalized the concept differently. Evidence of this can be seen in the way construction firms in the UK have reported on CSR. Construction firms operating in the UK have been found to vary in terms their scope, content and extent of their CSR reporting. CSR reporting and practices of UK construction firms have come under increased scrutiny after the sector has been plagued by ethical issues in the form of bribery, employment of illegal migrants and health & safety negligence to name a few. The effects of such unethical practices can be seen to haVing a detrimental impact on construction firms. As a result of this construction firms have been increasingly keen to demonstrate their commitment to CSR. The prevailing literature privileges establishing a positive link between corporate performance and CSR engagement. However, commenters are beginning to focus on how organisational actors interpret the concept of CSR to suit their organisational requirements. Collective sensemaking was adopted as the theoretical lens for the multiple case study design to interrogate the empirical data from the document review and semi-structured interviews. The unit of analysis for the case studies is the construction firm. The findings show how the process CSR strategy development is initiated, structured and managed by key actors within the firms. The case studies show the variations in the use of management structures for CSR strategy development and the different contextual features that shape and characterise each firms' CSR strategy.

This thesis examines the role of low carbon energy projects in widening energy access, progressing energy transitions and furthering development goals in rural India. Currently in development contexts, energy access and transitions are mobilised through micro energy projects like solar lanterns and micro-grids. The successes and failures of these projects are primarily assessed quantitatively – number of villages covered, number of households connected etc. However, this approach fails to understand how energy transitions projects perform in people’s everyday lives. It does not capture the reasons why they work for particular groups of people and not for others. To go beyond the quantitative understanding, this thesisfocuses on the micro-politics of everyday life that shape the effects energy transitions projects have on different groups of people. Itconsiders how power, politics and culture are vital for understanding the successes and failures of these projects. Theoretically the thesis conceives low carbon projects as low carbon assemblages to understand their fluid and contingent nature and the ways in which they are configured and reconfigured through relationships of power and everyday politics. Engaging with governmentality studies, it further considers how different, pre-existing and newly configured relationships of power conduct people’s conducts and sometimes lead to resistances for low carbon projects. The thesis critically examines three crucial aspects of low carbon energy projects by engaging with three key ideas – trusteeship, significances and resistances. Firstly, it explores how, by positioning themselves as trustees, particular actors seek to assemble and govern low carbon projects in order to achieve specific outcomes. Secondly, it investigates how, by focusing on particular significances of electricity, these interventions work to achieve particular development goals in different spaces of everyday life. Finally, it asks how and why different pressures and contestations emerge as everyday resistances in low carbon transitions. The thesis takes an ethnographic route of enquiry in order to examine energy in everyday life, using participant observations, interviews and photography. It explores two different low carbon projects – Lighting a Billion Lives (LaBL) solar lanterns project and Husk Power Systems (HPS) biomass micro-grids project – and contrasts them against the central grid and kerosene oil networks, in five villages in Bihar. Three key arguments emerge from the thesis. Firstly, electricity access should be understood as a spatially heterogeneous and temporally fluid idea. Its firm quantification and standardisation are problematic because electricity access is geographically and socially differentiated. It needs to be explored in an ethnographic manner, in which not onlyquestions of ‘how much’ but also of who, how and where are critical. Secondly, in energy and development projects, context matters. The society, culture, politics and economy of spaces in which projects are implemented mediate their impacts. Finally, the upkeep and maintenance of low carbon energy projects is not just about economies and supply chains of spare parts but also about cooperation and coordination between the project designers and users. Being able to fulfil people’s changing electricity requirements by building flexibility in the projects is critical to respond to these three issues. This will make projects more sustainable and increase people’s trust on low carbon projects leading to a convergence between energy access and energy transitions.

This thesis is a study of the international dimensions of the Kashmir dispute between India and Pakistan from before its outbreak in October 1947 till the Tashkent Summit in January 1966. By focusing on Kashmir’s under-researched transnational dimensions, it represents a different approach to this intractable territorial conflict. Concentrating on the global context(s) in which the dispute unfolded, it argues that the dispute’s evolution was determined by international concerns that existed from before and went beyond the Indian subcontinent. Based on new and diverse official and personal papers across four countries, it foregrounds the Kashmir dispute in a twin setting of Decolonisation and the Cold War and investigates the international understanding around it within the imperatives of these two processes. In doing so, it traces Kashmir’s journey from being a residual irritant of the British Indian Empire, to becoming a Commonwealth embarrassment and its eventual metamorphosis into a security concern in the Cold War climate(s). A princely state of exceptional geo-strategic location, complex religious composition and unique significance in the context of Indian and Pakistani notions of nation and statehood, Kashmir also complicated their relations with Britain, the United States, Soviet Union, China, the Commonwealth countries and the Afro-Arab-Asian world. The thesis begins with British anxieties regarding independent India’s international identity that arose in 1945-47 and covers the international involvement in the first Kashmir conflict (1947-49). Next, it undertakes a survey of the initial American attitude to India (1945-47) and situates the early American approach to Kashmir (1947-49) in that light. The thesis then shows the transformation of Kashmir from being a Commonwealth concern to becoming an American affair (1949-53). Further, it traces the dispute’s transition from the prism of Western pact-politics to that of Subcontinental package proposal (1953-61). The thesis ends with comparing the last Anglo-American intervention in Kashmir (1962-63) with its Soviet counterpart (1965-66).

Dendritic cells are efficient antigen-presenting-cells that can be used in tumour-antigen specific vaccination for malignant disease. Melanoma patients were recently treated with a dendritic cell vaccine expressing gp100 and Melan-A antigens after non-viral (CL22 peptide) transfection. Although clinical and immunological responses were noted, there was no correlation between responses and whole antigen expression levels in the vaccine cells that varied widely. Here, it is established that patient cells expressed detectable levels of Class I restricted epitopes from both antigens, although there was no correlation with whole antigen detection. CL22 transfected dendritic cells could simultaneously present a viral antigen (EBNA1) to CD8 and CD4 T-cells, which had not previously been demonstrated. Using RNA transfection, it was demonstrated that early after transfection cells are whole Melan-A positive yet negative for Class I epitopes and with time Melan-A antigen levels fall whilst Class I epitopes are generated. Loss of whole antigen expression seemed related to lysosomal function and, unlike the viral antigen EBNA1, Class I presentation from Melan-A was lysosome-dependent. For Class II presentation of EBNA1, cellular localisation seems to determine access to the Class II pathway although this depends on the time-scale over which epitope presentation is assessed.

Short-term stocks trading can be based on quantitative or qualitative factors. Company earnings reports seem to provide quantitative signals at regular intervals which is used by short-term traders and long-term investors alike. In this study, we tackle the question of whether we can create stock portfolios that outperform the broader US market based on stock price movement around earnings release dates. Thus, in essence, we test whether the semi-strong form of the Efficient Market Hypothesis (EMH) is applicable to the US stock market. We select a large set of stocks from the overall market and use historical movements of each stock price around the earnings date to calculate a profit-to-loss ratio. Next, we construct a portfolio based on the risk taking capacity of the trader and the stock profit-to-loss ratios. The performance of this portfolio is compared with a long-term buy and hold strategy and the S&P 500 index.

India's higher education system aims to develop well-rounded individuals with intellectual curiosity, a spirit of service, and a strong ethical foundation. However, the current state of technical higher education in India is criticized for mono-disciplinarity and rigid boundaries that do not provide an appropriate context for developing socially conscious individuals who make productive contributions to society. To address this growing concern, the Department of Higher Education in India introduced the Student Induction Program (SIP). The primary objective of the SIP is to foster purpose, which is not only meaningful to the students themselves but also to their nation; as such, the focus of this research is twofold: to assess the primary objective of the SIP and its implications on students’ well-being in a methodologically robust manner and to perform a detailed psychometric analysis of the instruments administered in an Indian context. This convergent parallel mixed methods study, grounded in the pragmatism paradigm, used closed and open-ended survey questions from 975 individuals recruited from two prestigious national institutes of India. Preintervention and postintervention questionnaires were used to assess the program’s impact and demographic measures to identify any distinguishing outcomes. Quantitative analysis showed that the SIP boosted other-oriented purposes but did not find any statistically significant differences in purpose among the demographic groups. Qualitative analysis revealed that living in peaceful coexistence with everyone and familial purpose exhibited positive change, possibly due to sessions conducted on Universal Human Values and likely owing to family obligation, respectively. Female participants preferred familial, harmonious living and ecologically sensitive purposes, while male goals pertained to careers. Overall, the SIP boosted other-oriented purpose among many participants but still needs further development to foster purposes that are consistent with institutional ethos. Cluster analysis was used to identify purpose profiles and found that other-oriented purposeful youth demonstrated greater well-being than disengaged youth. Psychometric analysis of the revised Stanford Purpose Instrument was satisfactory, and the Flourishing Scale was aligned with the original study. Multiple imputations were used to impute missing data. The study contributes purpose, well-being, and program assessment literature. The findings may inform similar educational interventions in India and abroad.

To correlate and be able to tune the photophysical properties of molecular scale chromophores with the mechanical properties of the macroscale supramolecular polymer such as viscosity, modulus etc. is a complex phenomenon. To tune these properties, we utilize dynamical coordination bonds such as Metal-Ligand interactions. In this work, we present how we can use the photophysical properties specifically probing excited state arising from these tunable dynamical interactions to probe the mechanical properties of the macroenvironment real time in-situ. Utilizing this approach, we were able to detect the viscosity of supramolecular polymeric assembly by probing emission lifetime from triplet electronic state of [Cu(diptmp)2]2+. However, this correlation is not quantitatively ubiquitous in every polymeric host. We probed the reversibility of the polymeric assembly by swelling and deswelling the polymer assembly which is attributed to micro-viscosity. The photophysical changes were of significant magnitude to be able to detect even small reversible changes during swelling and deswelling the polymer.Furthermore, worked on synthesizing a bio-based polymer and made a composite material with multiple ligand moieties. We present how these moieties interact with the polymeric environment and quantify them through their affinity to different metals. The photophysical changes in emission of a bio-based Cu-curcumin polymeric composite material upon applying stress to the composite material has been discussed. We also present to show whether we can induce the photophysical changes in NIR emission of a water soluble [Cr(ddpd)3]3+ complex termed as molecular ruby exhibiting upon subjecting it to thin supramolecular films with increasing modulus and looked at its stress-response.

Recent advances in computer vision tasks are mainly due to the success of large deep neural networks. The current state-of-the-art models have high computational costs during inference and suffer from a high memory footprint. Therefore, deploying these large networks on edge devices remains a serious concern. Furthermore, training these over-parameterized networks is computationally expensive and requires a longer training time. Thus, there is a demand to develop techniques that can efficiently reduce training costs and also be able to deploy neural networks on mobile and embedded devices. This dissertation presents practices like designing a lightweight network architecture and increasing network resource utilization. These solutions improve the efficiency of large networks during training and inference. We first propose an efficient micro-architecture (slim modules) to construct a light-weight Slim-CNN to predicting face attributes. Slim modules uses depthwise separable convolutions with pointwise convolutions, making them computationally efficient for embedded applications. Next, we investigate the problem of obtaining a compact pruned model from an untrained original network in a single-stage process. We introduce our RAPID framework that distills knowledge to a pruned student model from a teacher model under online settings. Next, we analyze the phenomena of inactive channels in a trained neural network. We take a deep dive into the gradient updates of these channels and discover that these channels have no weight update after a few early epochs. Thus, we present our channel regeneration technique that reinitializes batch normalization gamma values of all inactive channels. The gradient updates of these channels improve after the regeneration step, resulting in an increase in the contribution of these channels to the network performance. Finally, we introduce a method to improve computational efficiency in pre-trained vision transformers by reducing redundancy in visual data. Our method selects image windows or regions with high objectness measures, as these regions may contain an object of any class. Across all works in this dissertation, we extensively evaluate our proposed methods and demonstrate that our techniques improve the computational efficiency of deep neural networks during training and inference.

III-Vs like GaN and InP are the most promising substrates for high performance RF, photonics, and power electronics. However, III-V processing has been a gating factor in achieving smaller features and better devices. A primary cause of this setback has been large feature sizes resulting from limitations of older equipment. Older equipment must be used because it is compatible with III-V wafer sizes, which are smaller, costlier, and of lower quality compared to silicon wafers. To use the same cutting-edge tools and systems that silicon has immensely benefited from, III-V wafers must be made in a 300 mm format so that they can be processed in the same facilities as Si wafers. This thesis proposes and demonstrates one method of reconstituting gallium nitride (GaN) to create a large wafer that can be processed using large wafer tooling. Investigating these techniques also develops thermocompression bonding process capability that can be transferred to eventual III-V to Si vertical monolithic integration. For example, a GaN power amplifier can be vertically bonded (3D stacked) to a silicon dielet, decreasing the scale of power delivery networks by almost 1000x (from mm scale to �m scale). However, integration of gold into active silicon devices would require significant reliability study that is outside the scope of this work.In this thesis, GaN on (111) silicon was chosen as a starting III-V material. After dicing, these dielets were bonded face-to-face to a (100) silicon handler with a gold-to-gold thermocompression bonded interface. The bulk of the dielet thickness is silicon, and it is removed with a hydrofluoric acid, nitric acid, and acetic acid etch (HNA). This leaves behind a thin (< 1 �m) film of III-V substrate, which is planarized and passivated with plasma enhanced chemical vapor deposition (PECVD) oxide and chemical mechanical planarization (CMP). The techniques developed in this thesis can be adapted to reconstitute other III-V materials, like GaN on sapphire, indium phosphide (InP), and gallium arsenide (GaAs).

The impact-induced fracture analysis has a wide range of engineering and defence applications, including aerospace, manufacturing and construction. An accurate simulation of impact events often requires modelling large-scale complex geometries along with dynamic stress waves and damage propagation. To perform such simulations in a timely manner, a highly efficient and scalable computational framework is necessary. This thesis aims to develop a high-performance computational framework for analysing large-scale structural problems pertaining to impact-induced fracture events. A hierarchical grid-based mesh containing octree cells is utilised for discretising the problem domain. The scaled boundary finite element method (SBFEM) is employed, which can efficiently handle the octree cells by eliminating the hanging node issues. The octree-mesh is used in balanced form with a limited number of octree cell patterns. The master element matrices of each pattern are pre-computed while the storage of the individual element matrices is avoided leading to a significant reduction in memory requirements, especially for large-scale models. Further, the advantages of octree cells are leveraged by automatic mesh generation and local refinement process, which enables efficient pre-processing of models with complex geometries. To handle the matrix operations associated with large-scale simulation, a pattern-by-pattern (PBP) approach is proposed. In this technique, the octree-patterns are exploited to recast a majority of the computational work into pattern-level dense matrix operations. This avoids global matrix assembly, allows better cache utilisation, and aids the associated memory-bandwidth limited computations, resulting in significant performance gains in matrix operations. The PBP approach also supports large-scale parallelism. In this work, the parallel computation is carried out using the mesh-partitioning strategy and implemented using the message passing technique. It is shown that the developed solvers can simulate large-scale and complex structural problems, e.g. delamination/fracture in sandwich panels with approximately a billion unknowns (or DOFs). A massive scaling can be achieved with more than ten thousand cores in a distributed computing environment, which reduces the computation time from months (on a single core) to a few minutes.

Friction element welding (FEW) has been advocated as a solution to weld different materials together, with the ability to join high-strength materials for a range of thicknesses with low input energy and a short processing time. This work develops a coupled thermal-mechanical finite element model to better understand the physical mechanisms involved in the process and to predict temperature and material flow during the process. Furthermore, microstructural analysis is performed for the steel layer using a scanning electron microscope and Vickers microhardness tester to understand the variation in its grain structure and hardness. Results from the finite element model and microstructural analysis are correlated to help understand influence of process parameters on the microstructure and weld strength. The simulation results for sets with varying process parameters show that the friction element's rotational speed has the greatest influence on the amount of friction heat generated, followed by the time for different steps. The experimental results show that grain refinement and improvement in hardness value is observed in the steel layer due to high temperature levels and/or high stress. In addition to investigating the effects of process parameters on the heat generation and resultant weld quality, chipping phenomena during the FEW process is investigated using finite element modeling. The numerical analysis indicates that chipping is severe at the beginning when the friction element contacts the workpiece and tends to reduce as the former advances into the latter. Chipping can be minimized by using high rotational speeds coupled with low feed rates.

Antibiotic resistance is one of the major global issues in the field of public health and medicine. Good antibiotic candidates need to be selectively toxic, inhibit cellular target, and effectively penetrate and accumulate in bacterial cells. The last factor is a formidable barrier in the development of antimicrobials effective in Gram-negative bacteria, due to the presence of two layers of cell envelope. The first half of my thesis focuses on understanding the permeation of small molecules through this formidable cell envelope, distribution inside the cell of Gram-negative bacteria, and design of novel methods to make small molecules effectively cross the cell envelope. The second half of my thesis focuses more on the crosstalk between Gram-negative bacteria and host immune system during systemic infection and sepsis. More specifically we studied the contribution of inflammasome activation and pyroptosis during pathogenesis of Salmonella systemic infection. In the first project, I studied the accumulation and distribution behavior of fluoroquinolone class of antibiotics inside Gram-negative bacteria using E. coli as a model. Although several studies have been focused regarding the correlation between compound’s cellular accumulation and their effectiveness against Gram-negative bacteria but no correlation between accumulation of antibiotics and their efficacy has been observed. In this study, we measured the concentration of nine fluoroquinolones accumulated in the subcellular compartments of E. coli. Good correlation between the MIC and the cytoplasmic accumulation, but not whole cell accumulation, was observed using a pair of isogenic wild type and drug-efflux deficient strains. Our results supported the explanation that the efficacy cannot be determined by the whole cell accumulation alone. Accumulation in the target region as well as the intrinsic potency determines the effectiveness of an antimicrobial compound. In the second project, I explored whether conjugation of biotin to small molecules can increase the permeation of small molecules through the Gram-negative cell envelope. We used a florescent molecule pair, Atto565 and Atto565-biotin as model compounds and studied their permeation behavior in E. coli. Our results indicated that biotinylation helped the molecule Atto565 to cross the outer membrane of E. coli through OmpC porin. Moreover, in the third project, I studied how the inflammasome activation and pyroptosis play a role in pathogenesis of Salmonella systemic infection. We found that Salmonella systemic infection causes severe inflammation as indicated by very high plasma concentration of pro-inflammatory cytokines, IL-1β, IL-6 and TNF-α. Furthermore, it also caused disseminated intravascular coagulation (DIC) as indicated by increased prothrombin (PT) time and plasma thrombin-antithrombin (TAT) levels. Deficiency of caspase 1 protected the mice from Salmonella induced inflammation, coagulation and death during acute systemic infection. Similarly, deficiency of NAIP and GSDMD significantly reduced the Salmonella induced inflammation in vivo. Use of flagellin and Salmonella pathogenicity island 1 (SPI1) region knockout strains of Salmonella induced significantly less cytokine release in the plasma, however, could not protect from the coagulation and lethality. In vitro studies showed that deficiency of Caspase 1, NAIP and GSDMD also protects the bone marrow derived macrophage’s (BMDM’s) death upon Salmonella infection during early phase of infection. In vitro, inflammasome activation and BMDM death was also completely abolished when flagellin or SPI1 deficient strains of Salmonella were used during early but not late phase of infection. During late phase of infection, caspase 11 and NLRP3 dependent pyroptosis had major contribution. These results indicate that during acute Salmonella systemic infection under flagellin and SPI1 expressing conditions, severe inflammation occurs mainly through NAIP/Caspase 1/GSDMD axis. However, coagulation could also be induced also by factors other than flagellin and SPI1 that contributes to lethality. Flagellin and SPI1 independent coagulation was protected by NLRP3 and caspase 11 deficiency.

Rhythmic motion in organisms is driven by central pattern generators (CPG). A CPG converts high-level input from the brain to low-level oscillatory input to the muscles via neuron spikes. Here a particular CPG, a reciprocal inhibition oscillator(RIO), is connected to an underdamped pendulum with position feedback to form a closed-loop system. Two types of neuron models are considered, both based on Hodgkin-Huxley dynamics, self-bursting and adapting. The closed-loop system is simulated and the pendulum oscillation is observed to indicate under which conditions the oscillation entrains to the pendulum's mechanical resonance frequency. We find that entrainment depends on the configuration (negative/positive feedback and excitatory/inhibitory connections) and the ratio of pendulum resonant frequency to RIO intrinsic frequency. This is true of both neuron models, although the exact quantities and regions of entrainment vary between the two.

Abstract: The outer shell of firefighter protective clothing is prepared from high performance polymer fibers such as Kevlar®, Nomex®, PBI, PBO and LCP. They are designed to withstand extreme temperatures and provide protection against service hazards to firefighters. However, these fabrics show continuous deterioration of mechanical properties when exposed to heat, UV, and moisture. This thesis work focusses on studying eight different outer shell fabrics with varying fiber contents and fabric structures in the unaged condition and after immersion of the fabric specimens in water at varying temperatures between 60°C and 90°C for up to 1200 hours. The effect of hydrothermal ageing was studied by analysing the residual tensile strength of these fabric specimens. This study also gives us quantitative information about the effect of fiber content and the influence of fabric structure on the fabric resistance to hydrothermal ageing and creates a foundation for the development of end-of-life sensors for fire protective clothing. A second part of the work relates to the development of a fabrication process for the production of robust end-of-life sensors for the outer shell fabrics that can sustain the normal conditions experienced by the firefighter protective clothing. The overarching theme of the work is to develop an end-of-life sensor for firefighter garment, whereas the scope of my thesis study was to develop sensors that are free from premature failure. Specifically, I worked on reinforcing encapsulation by careful selection of electrode materials, developing processing methods, and establishing the testing protocols to monitor the deterioration of the sensor properties. The sensor development work was still in progress at the time of my thesis defense. This ageing study provides a basis for understanding the behaviour of these outer shell fabrics when exposed to water to help in estimating the service life of these fabrics and protect firefighters from safety hazards. The sensor study can impact the fields of wearable sensor and e-textiles by providing insights in developing effective encapsulation technologies and in testing the reliability of these devices.

The explanatory power of earnings per share (eps) is on the decline as firms are focusing more on intangible assets and are disclosing more marketing metrics when they announce their earnings (e.g., subscribers for the telecom & media industry and monthly active users for social media industry). However, the performance of these marketing output metrics beyond market/analysts’ expectations (i.e., surprises) requires marketing resources, which may reduce current profitability but may also signal a higher future cash flow. Therefore, building on information economics, we assess if there is information content in marketing metric surprises, and how the stock market reacts to such surprises. Further, we argue that the information content of marketing metric surprises varies under different information signals by firms (strategic emphasis) and screening cues by investors (marketing expenditure). We also investigate the temporal variations in the effect of marketing metric surprises and also examine the relative importance of marketing metric surprises as compared to earnings surprises across multiple industries. We test the claims using an event study methodology around earnings announcement on S&P 1500 firms consisting of firms disclosing industry-specific marketing metrics and non-disclosing firms. We account for sample selection bias and correcting for potential endogeneity concerns of surprises marketing metrics. Our findings suggest that (1) although an increase in marketing metric surprise affects the stock market returns positively, (2) this effect is strengthened when firms signal strategic emphasis on value appropriation relative to value creation whereas (3) it is attenuated when investors screen for firms with higher unanticipated marketing expenditure, (4) the effect of marketing metric surprises increases over time whereas it decreases for earnings surprise, and (5) the effect of marketing metric surprise is higher in the telecom and media industry as compared to earnings surprises. The study helps to improve marketing accountability at the time of earnings announcement by improving the overall earnings quality of firms.

Abstract: There is an increasing need of clean renewable energy sources to address the growing problems of greenhouse gas emission, global warming and climate change. The rapid consumption of fossil fuel resources has also contributed to toxic emissions and raised a question on sustainability of present fuels. Hence, alternative renewable fuels such as biofuels have been considered as a promising low pollution source. Biomass is composed of the green carbon i.e. carbon that belongs to the present biological cycle, making it a cleaner fuel. Biomass has an inherent energy that comes from the sun and has possibility to regrow in a short period of time. Lignocellulosic biomass mainly derived from agricultural and forest can be processed through pyrolysis, gasification, steam reforming and hydrothermal liquefaction to obtain high energy density biofuels. Predominantly, all biomass materials contain high moisture, which makes the hydrothermal liquefaction as the most effective technique for conversion of biomass to biofuels such as bio-oil, hydrochar and gases. In this work, hydrothermal liquefaction process was used for conversion of lignocellulosic biomass (corn stover) into liquid, solid and gaseous fuels at moderate temperature in the range of 250 – 375 °C, final pressure (Pf) in range of 1100 – 3400 psi and retention time (tr) in range 0 – 60 min. The key focus of this research work is to understand feedstock properties, develop experimental methodology for hydrothermal liquefaction and apply it for conversion of biomass to biofuels. The study also attempts to optimize the process parameters with an aim to obtain quantitatively and qualitatively better biofuels. The experiments were performed in a 250 mL, high pressure autoclave batch reactor having temperature and pressure upper limit of 500 °C and 5000 psi (34.47 MPa) respectively. An inert nitrogen environment was maintained. The final recovered slurry was processed through a detailed separation procedure in order to obtain the bio-oil and hydrochar separately for analysis. In order to understand the properties of products, characterization tests such as elemental analysis and GC-MS was performed on bio-oil and elemental analysis, FTIR and SEM were carried out on hydrochar. The quantitative results obtained at different operation conditions gave a highest yield of heavy oil (29.25 wt.%) obtained at 300 °C, final pressure, Pf, of 2200 psi and 0 min retention time. The highest yield of hydrochar (30.21 wt.%) was obtained at 350 °C, Pf of 3150 psi and soak period, tr, of 15 min. Based on elemental analysis and energy calculation, highest carbon content and higher heating value for heavy oil was 76.32 wt.% and 35.13 MJ/kg at 375 °C, Pf of 600 psi and tr of 15 min; and for hydrochar it was 68.23 wt.% and 24.7 MJ/kg at 350 °C, Pf of 3150 psi and tr of 15 min. The GC-MS results for heavy oil indicated that majority of the compounds were phenolic in nature. FTIR results confirmed the decomposition of protein and carbohydrate and formation of new aromatic bonds in hydrochar during HTL. The morphology results for different hydrochar indicates the breaking of fibrous structure and formation of a more porous material. Even though, quality of oil obtained in this study was good, more studies need to be carried out in a direction to upgrade oil by reducing the oxygen content and increasing the higher heating value; moreover, developing better understanding by analyzing viscosity and total acid number value.

Abstract: The energy sector is responsible for a significant portion of global greenhouse gas emissions, water withdrawal, and water consumption. There are strong dependences between energy production and water use, and the adoption of more clean energy production technologies will affect water use. Such technological changes are not well understood. There is very little research assessing the water use associated with clean energy pathways in the energy sector. The objective of this research is to understand and evaluate the water-use impacts of Canada’s renewable energy transition. This research developed an integrated energy-water model to assess clean energy scenarios and the resulting technology penetration, water use, and cumulative and marginal greenhouse gas (GHG) emissions abatement costs. The energy sectors considered in this work are the oil and gas sector and the electricity generation sector. The Canadian Water Evaluation and Planning model (WEAP-Canada) was developed and integrated with the Long-range Energy Alternative Planning (LEAP-Canada) system model to determine integrated water-greenhouse gas footprints for the electricity generation sector and the oil and gas sector for the years 2005-2050. This research develops integrated water-greenhouse gas footprints for future electricity generation mix pathways in Canada with a focus on deep decarbonization. The LEAP model of Canada’s electricity system was developed by using technology system capacity requirements, technology capacity addition, and technology and economic inputs to provide electricity generation technology capacities and generation, system costs, and GHG emissions. A Water Evaluation and Planning model for Canada’s electricity generation sector was also developed by considering one-hundred-fifty-six electricity generation water demand sites and seventy-four major rivers. The two models were integrated to analyze electricity production and its associated greenhouse gas emissions and water use. Two scenarios were developed and evaluated for a planning horizon of 2019 to 2050. First scenario was based on current policy and second scenario was developed with a deep decarbonization target of 100%. In the current policy scenario, water consumption is projected to decrease by 5% and GHG emissions to decrease by 81% by 2050 compared to 2019. In the 100% decarbonization scenario, the water consumption is projected to increase by 3% by 2050 compared to 2019. Setting decarbonization targets of 100% resulted in a marginal water consumption of 3.6% and marginal GHG abatement costs of $23 per tonne of CO2 equivalent. There are water consumption tradeoffs in the 100% decarbonization scenario, but they are iii relatively small compared to those in other water-consuming sectors in Canada. Assessing the water-use impacts and costs of GHG emission reduction pathways helps to identify the co-benefits or tradeoffs associated with decarbonizing this sector and may be useful in policy development. This research also develops water use footprints for future oil and gas production in Canada by considering different energy pathways. The oil and gas section of the thesis focuses on developing a baseline for future water-GHG analysis in the oil and gas sector rather than understanding the water-GHG trade-off for clean energy scenarios. The WEAP model of Canada’s oil and gas sector (WEAP-COG) was developed using production, water-use intensities, and production shares based on watersheds to provide water withdrawal and consumption at the watershed level. Using the six water withdrawal sectors of the oil and gas sector (oil sands mining, oil sands in situ production, bitumen upgrading, conventional crude oil production, natural gas production, and crude oil refining), we modelled twenty rivers and forty-five demand sites. Energy scenarios were adopted from literature for high-price and low-price cases and then run in the WEAP-COG model. The results were compared to the reference scenario results to obtain marginal water use and GHG footprints for each scenario. The GHG emissions, water withdrawal, and water consumption for the oil and gas sector will increase by 80%, 21%, and 39% by 2050 from 2005, respectively, because of the increase in oil production. GHG emissions from the oil and gas sector will exceed Canada’s nationally determined contributions (NDC) target. The adoption of clean energy, less water-consuming technologies is recommended in order to achieve the current NDC target or to mitigate future water stress. The results developed in this research can be used by the decision makers in the government and industry for investment decision and policy formulation.

Cold atmospheric pressure plasmas are being widely studied for applications in biology and medicine. Non-thermal plasma-based decontamination has been one of the important areas of investigation due to the capability of cold plasmas to generate reactive species with antimicrobial properties close to room temperatures. Foodborne viruses and bacteria have been major causes of disease outbreaks, especially through the consumption of minimally processed foods and fresh produce. Conventional non-thermal food processing technologies face limitations and affect the properties of food. This work on plasma-based decontamination was motivated by the necessity to develop an effective and efficient non-thermal food decontamination technology. We used different plasma sources operating in air at atmospheric pressure for surface decontamination experiments. These experiments were performed with feline calicivirus (FCV), a surrogate of human norovirus, and Salmonella Heidelberg on stainless steel discs as model for food contact surfaces. The substrates were treated in dry and wet surface conditions. The plasma sources were designed such that the substrates are treated either directly such that there is direct contact between the plasma and the substrate or remotely such that the reactive species in the afterglow are transported to the substrate. It was shown that humidity on the surface of the substrate strongly enhances the decontamination effectiveness. Only direct treatment was seen to be significantly effective against dry samples. All the plasma sources achieve complete inactivation of wet samples. Synergistic effects between O3 and NO2 in the afterglow of the DBDs was recognized through positive control tests. With the use of a kinetic model to study the interaction among the chemical species in the afterglow of the plasma, the virucidal effects induced by the DBDs were attributed to the generation of gas-phase N2O5. The influence of surface humidity was studied through the quantification of reactive species in the water layer covering the substrate. It was found that the reactive species concentrations increased as the volume of the water decreased and this effect was consistent with the observed decontamination effects. The importance of enhancing the transport of reactive species from the gas to liquid phase for optimal decontamination efficacy of remote plasma treatment was identified. Finally, plasma sources were compared based on the energy consumption requirements for achieving complete decontamination. To gain an understanding on the possibilities of adoption of plasma technologies by the food industry, the plasma sources were compared with UV-C radiation in terms of energy costs and practical utility. For surface decontamination, direct plasma treatment was found to be four times more energy-intensive than UV-C treatment. A significant enhancement of energy efficiency was achieved using remote plasma treatment in a lab-scale batch reactor prototype employing a surface discharge generating predominantly O3, leading to energy per unit area requirements similar to that for UV-C. Preliminary estimates suggest that the energy efficiency might be further increased for a source enabling disinfection by reactive nitrogen species.

Irrigation is a policy focus in Sub-Saharan Africa and is viewed as an important mechanism to improve farmers’ income and livelihoods while reducing the impacts of climate change. Water, energy, and food are linked in intricate ways in irrigated agriculture, and understanding the interplay of these components is crucial for sustainable and profitable crop production. Although studies have been conducted in different parts of the world to understand water and energy use at a field scale under large irrigation systems, little is known about linkages under farmer-managed mechanized irrigated schemes in Sub-Saharan Africa. This study evaluates water-energy-food linkages, engineering and economic performance, current irrigation decision making, and challenges faced around water management in a community-based mechanized irrigation scheme. The research synthesizes intraseasonal water and energy use data for selected crops in a shared center-pivot irrigation scheme in Rwanda. The major cultivated crops are maize and beans (French beans, dry beans, common beans). A daily soil-water balance is central to estimate actual irrigation water requirement (IWR) and is simulated in FAO-CROPWAT 8.0. The study further investigates the variation in water requirements, and the relationship and impacts of this variability on crop yield. Assessment of irrigation performance is done by estimating and comparing crop water productivity (CWP) and crop water use efficiency with global and local averages. Observed irrigation decision-making analyses demonstrate a lack of irrigation planning during growth stages and significant field-to-field variation in irrigation; this is linked to yield reduction in major crops. An econometric model assessment is used to understand the relationship between yield and energy inputs. The energy use assessment includes both direct (electricity) and indirect energy inputs (fertilizers, pesticides, machinery, labor, etc.). This study has implications for understanding irrigation policies in the context of the water-energy-food nexus and decision-making in Sub-Saharan Africa. Advisors: Nicholas Brozović and Derek M. Heeren

Data structured in the form of overlapping or non-overlapping sets are found in a variety of domains, sometimes explicitly but often subtly. For example, teams, which are of prime importance in industry and social science studies are “sets of individuals”; “item sets” in pattern mining of customer transactions are sets, and for various types of analysis in language studies a sentence can be considered as a “set or bag of words”. Although building models and inference algorithms for structured data has been an essential task in the fields of machine learning and statistics, research on “set-like” data remains less explored. Relationships between pairs of elements can be modeled as edges in a graph. However, for modeling relationships that involve all members of a set, hyperedges in a Hypergraph are more natural representations. Hypergraphs are less known graph-theoretic structure as compared to graphs. Because of this popularity graphs have been employed prolifically to model data of all kinds. Little attention is given to the fact that whether the data is naturally being generated as dyadic interactions or not. We think that much data is even deliberately converted to a graph for the sake of fitting it into a graph-based model and destroying the precious information present when it was originally generated. This thesis describes analyzing complex group structured data from domains like social networks, customer transaction data, and general categorical data, via the lens of Hypergraphs. To do so, we propose the Hypergraph Analytics Framework, under which we shall be interested in three higher-level questions pertaining to the hypergraph modeling. Firstly, how to model higher-order hypergraph information and what kind of lower-order approximations are available or sufficient depending upon the problem at hand. This question is addressed across the thesis as we employ different hypergraph models contingent upon the problem at hand. Secondly, we shall be interested in understanding what kind of inferences are possible over the hypergraph structure and what kind of probabilities can be learned. For this, we shall be dissecting the problem of hypergraph inference into various hyperedge prediction sub-problems and developing inference methods for each of them. We develop inference methods for both cross-sectional analysis: when we ignore the time information about group interactions into account, as well as longitudinal analysis: where we leverage temporal data. We also develop separate methods for conducting inference over observed and unobserved regions of the hypergraph structure. This variety of inference mechanisms on hypergraph structure together constitute the first part of the thesis, which we refer to as the \textit{Spatial Analysis} within our Hypergraph Analytics framework. Lastly, we are interested in learning what kinds of compression algorithms are possible for hypergraphs and how effective these techniques are. Here we develop techniques to compress the hypergraph topology to lower-dimensional latent space. We shall be chiefly considering hyperedge compression or hyperedge embeddings. We examine two different embedding approaches. First, is an algebraic approach which leverages leverage the relationship between hypergraphs and symmetric higher-order tensors. Symmetric tensor decomposition techniques are then developed to learn embeddings. Second, is a neural networks based solution which employs auto-encoders regularized by hypergraph structure. Together, both these approaches constitute the second part of the thesis, which we refer to as \textit{Spectral Analysis} within the proposed Hypergraph Analytics framework.

There has been a continuous evolution in deep neural network architectures since Alex Krizhevsky proposed AlexNet in 2012. Part of this has been due to increased complexity of the data and easier availability of datasets and part of it has been due to increased complexity of applications. These two factors form a self sustaining cycle and thereby have pushed the boundaries of deep learning to new domains in recent years. Many datasets have been proposed for different tasks. In computer vision, notable datasets like ImageNet, CIFAR-10, 100, MS-COCO provide large training data, with different tasks like classification, segmentation and object localization. Interdisciplinary datasets like the Visual Genome Dataset connect computer vision to tasks like natural language processing. All of these have fuelled the advent of architectures like AlexNet, VGG-Net, ResNet to achieve better predictive performance on these datasets. In object detection, networks like YOLO, SSD, Faster-RCNN have made great strides in achieving state of the art performance. However, amidst the growth of the neural networks one aspect that has been neglected is the problem of deploying them on devices which can support the computational and memory requirements of Deep Neural Networks (DNNs). Modern technology is only as good as the number of platforms it can support. Many applications like face detection, person classification and pedestrian detection require real time execution, with devices mounted on cameras. These devices are low powered and do not have the computational resources to run the data through a DNN and get instantaneous results. A natural solution to this problem is to make the DNN size smaller through compression. However, unlike file compression, DNN compression has a goal of not significantly impacting the overall accuracy of the network. In this thesis we consider the problem of model compression and present our end-to-end training algorithm for training a smaller model under the influence of a collection of "expert" models. The smaller model can be then deployed on resource constrained hardware independently from the expert models. We call this approach a form of compression since by deploying a smaller model we save the memory which would have been consumed by one or more expert models. We additionally introduce memory efficient architectures by building off from key ideas in literature that occupy very small memory and show the results of training them using our approach.

There is a high demand for low step-down dc voltage conversions. Many conventional power converters that are currently being used have a moderate conversion ratio and this may not be sufficient to meet the demand. This can be achieved by either cascading power converters or using converters with a low step-down conversion ratio. Cascading the converters increases the power conversion stage complexity and increase the order of the system, while also affecting the stability. Using converters with a high conversion ratio seems to be a more intelligent option to root. Thisdissertation tackles to analyze one such converter, called tapped-inductor buck dc-dc converter.A tapped-inductor buck dc-dc converter, capable to produce higher conversion ratios compared to the conventional converters is used. An analysis describing a detailed steady-state operation of the converter is provided. The expected voltage and current waveforms across different components at different points of time during the entire operation of the converter are analytically derived. Design equations for the converter have also been provided. Power lost across various converter components are predicted. The overall converter efficiency is calculated. The dynamics of the system are predicted. For this a model of the tapped-inductor buck dc-dc converter is derived using circuit averaging technique. Transfer functions relating the output to the input and the control voltages are derived. Various poles and zeros affecting the system magnitude and phase plots were analytically defined. A controller closed-loop system is implemented. Various time and frequency domain parameters affecting the system response are measured and compared to the open-loop system.All the theoretically obtained responses are implemented using MATLAB and verified using saber circuit simulator. The verified model responses from the simulations are also validated through hardware implementation.

In many applications, models of physical systems have known structure but unknown parameters. By viewing the unknown parameters as constant states, nonlinear estimation methods such as the extended Kalman filter, unscented Kalman filter, and ensemble Kalman filter can be used to estimate the states of the augmented system, thereby providing estimates of the parameters along with the dynamic states. These methods tend to be computationally expensive due to the need for Jacobians, ensembles, or adjoints, especially when the models are high-dimensional. This dissertation presents retrospective cost parameter estimation (RCPE), which does not require gradients, ensembles, or adjoints. Rather, RCPE estimates unknown parameters from a single trajectory, and requires updating an adaptive integrator gain for each unknown parameter. RCPE is applicable to parameter estimation in linear and nonlinear models, where the parameterization may be either affine or nonaffine. The main contribution of this work is to show that the parameter estimates may be permuted in an arbitrary way, and thus a permutation is needed to correctly associate each parameter estimate with the corresponding unknown parameter. RCPE is illustrated through several numerical examples including the Burgers equation and the Global Ionosphere Thermosphere Model (GITM), where the goal is to estimate representational parameters such as eddy diffusion coefficient and thermal conductivity coefficients using measurements of atmospheric variables such as total electron content, density, temperatures etc. The next part of the dissertation focuses on forgetting in the context of recursive least squares (RLS) algorithm. It is a well-known fact that classical RLS with forgetting diverges in the cases where the excitation is not persistent. In this work, an information-driven directional forgetting technique is proposed, which constrains the forgetting to directions in which new information is available, thereby allowing RLS to operate without divergence during periods of loss of persistency. In the last part of this dissertation, retrospective cost adaptive control (RCAC) is extended to the problem of control allocation in overactuated systems. In particular, it is shown that the applied control input lies in the range of the target model used in RCAC, thereby providing a simple technique to constrain the control input to a desired subspace. Finally, RCAC is extended to asymptotically enforce output constraint by formulating the problem as a problem of following conflicting commands, and is used to prevent a scramjet combustor from unstarting using pressure measurements.

The purpose of this research is to implement and analyze the true lean transformation in a manufacturing start-up organization. Often, lean transformations are observed in developed manufacturing organizations having sophisticated production lines and numerous employees, where lean tools are utilized to reduce waste while increasing profit. However, this type of transformation is narrowly focused on quantifiable process results and falls short of the ultimate goal, establishing a true lean culture within the organization. As a result, it is recommended that true lean principles, philosophies, culture, operations environment and tools be applied at the very beginning stages of an organization, or during the start-up phase to embed true lean thinking and application in the entirety of the organization. In this analysis, a case study is performed on a team embarking on a start-up manufacturing enterprise. The team is trained on true lean and all the four aspects of true lean are applied for successfully implementing true lean in the start-up. In addition, a true lean implementation approach was generated and applied using PDCA (plan, do, check and act) model. The results of this case study are presented in this work.

With the objective of minimizing carbon footprint of vehicles, different organizations across the world are increasingly enforcing higher fuel efficiency targets for the automobile manufacturers. To improve the fuel economy while retaining or further improving the structural integrity, the automobile industry is vigorously shifting towards substituting conventional heavy materials like cast iron with new age materials such as aluminum alloys, steel alloys, etc. which are not only much lighter but also offer superior strength-to-weight ratio. Engineers use a mix of these new age materials with the aim of maximizing the benefits from each material. However, the utilization of such materials is currently limited in the industry as welding them using conventional methods such as resistance spot welding or fusion welding process, is plagued with inherent difficulties such as formation of brittle inter-metallic compounds, irreversible and adverse changes in the thermal and mechanical properties of the materials. Dissimilar material joining is of critical importance in aiding the manufacturers realize the crucial objective of a safer and more fuel efficient vehicle. Friction element welding (FEW), a friction based joining process, has been proposed for joining highly dissimilar materials in minimal time and with low input energy. FEW process can join a variety of materials which differ significantly in their mechanical, thermal, and metallurgical properties without inducing any of the defects associated with conventional welding methods. The fundamental governing mechanisms that characterize the FEW process needs to be investigated to help optimize the process for specific applications. Conducting experimental investigation is undesirable and infeasible due to the highly complex thermal-mechanical procedures occurring simultaneously in a very short period of time of about one second. As such, the utilization of a finite element model to simulate and analyze the FEW process is warranted which would help understand the underlying mechanisms of the process in detail and provide an efficient yet effective tool to observe the effect of different process parameters on the weld quality. A coupled thermal-mechanical finite element model (FEM) is developed in this work to simulate the FEW process and gain an understanding of the physical mechanisms involved in the process and help predict the influence of variation of process parameters on the evolution of temperature, material flow, and their effect on weld quality. The primary difficulty in simulating a highly transient process like FEW, wherein not only the workpiece is subjected to deformation but also the auxiliary joining element i.e. friction element undergoes extensive deformation, is that the mesh elements are prone to distortion failure while trying to capture such high amount of deformation. The presence and importance of temperature effect on material properties further complicate the FEM. To help eliminate the distortion issue while simultaneously achieving an accurate simulation of the FEW process, the coupled Eulerian-Lagrangian (CEL) approach is adopted. The novelty of the current approach employed lies in using a Eulerian definition for the tool as against the more traditional convention of adopting a purely Lagrangian definition. The Eulerian definition enables to simulate the extreme deformation of friction element and capture the material flow without any computational issues. To inspect for the accuracy of the FEM results, mechanical deformation for different parts observed in the FEM is compared against the experimental results. To further validate the FEM, experimental measurements of temperature at different locations at the interface of two layers of workpiece are compared against the FEM results at same locations in the model. With respect to, both, thermal and mechanical measurements comparisons good agreement is shown between the simulation results and the experimental data. The simulation results for sets with varying process parameters show that the rotational speed of the friction element has the highest influence on the amount of frictional heat generated followed by the time period for different steps. Higher amount of heat is generated and conducted into the top aluminum layer for longer Penetration time, whereas for more heat concentration into the friction element to achieve the required deformation, longer Welding step with higher rotational speed is desired.