Cellular networks in Africa have grown exponentially over the past 10 years and their data centres (DCs) on average consume 3 MW of electrical power. They require a reliable electrical power supply and can have a downtime loss of over a million dollars per hour. Power quality, reliability and availability have emerged as key issues for the successful operation of a data centre. Investigations are carried out into emerging technologies and their application in data centre power distribution systems for cellular networks in Africa. Best practices are applied to develop a power distribution system (PDS) with the objective of achieving optimal reliability and availability. Analytical techniques are applied to determine and compare the reliability and availability of various power systems. Minimal cut set simulations identify system weak points and confirm component selection. Components’ inherent characteristics (CIC) and system connectivity topology (SCT) are key factors in the improvement of data centre availability. The analysis practices can be used by engineers and managers as a basis for informed decision making in determining power system reliability and the availability of an existing or a new data centre design. Weak points in the PDS of a data centre causing downtime are identified through analysis, and accurate solutions can be determined to prevent or minimise downtime. System connectivity topology (SCT) techniques were identified that could increase the reliability and availability of data centres for cellular networks in Africa. These techniques include multiple incomers from the utility company, redundancy levels of critical equipment and parallel distribution paths. Two case studies were carried out on data centres for a cellular network, one in Nigeria and one in Cameroon. The reliability and availability of both data centres was improved, with substantial reduction in downtime per year. The outcome of the case studies shows the importance of designing and implementing the power distribution system with sufficient levels of redundancy for critical equipment, and parallel distribution paths. / MSc (Engineering Sciences in Nuclear Engineering), North-West University, Potchefstroom Campus, 2014

Commercial and private cloud providers offer virtualized resources via a set of co-located and dedicated hosts that are exclusively reserved for the purpose of offering a cloud service. While both cloud models appeal to the mass market, there are many cases where outsourcing to a remote platform or procuring an in-house infrastructure may not be ideal or even possible. To offer an attractive alternative, we introduce and develop an ad hoc cloud computing platform to transform spare resource capacity from an infrastructure owner’s locally available, but non-exclusive and unreliable infrastructure, into an overlay cloud platform. The foundation of the ad hoc cloud relies on transferring and instantiating lightweight virtual machines on-demand upon near-optimal hosts while virtual machine checkpoints are distributed in a P2P fashion to other members of the ad hoc cloud. Virtual machines found to be non-operational are restored elsewhere ensuring the continuity of cloud jobs. In this thesis we investigate the feasibility, reliability and performance of ad hoc cloud computing infrastructures. We firstly show that the combination of both volunteer computing and virtualization is the backbone of the ad hoc cloud. We outline the process of virtualizing the volunteer system BOINC to create V-BOINC. V-BOINC distributes virtual machines to volunteer hosts allowing volunteer applications to be executed in the sandbox environment to solve many of the downfalls of BOINC; this however also provides the basis for an ad hoc cloud computing platform to be developed. We detail the challenges of transforming V-BOINC into an ad hoc cloud and outline the transformational process and integrated extensions. These include a BOINC job submission system, cloud job and virtual machine restoration schedulers and a periodic P2P checkpoint distribution component. Furthermore, as current monitoring tools are unable to cope with the dynamic nature of ad hoc clouds, a dynamic infrastructure monitoring and management tool called the Cloudlet Control Monitoring System is developed and presented. We evaluate each of our individual contributions as well as the reliability, performance and overheads associated with an ad hoc cloud deployed on a realistically simulated unreliable infrastructure. We conclude that the ad hoc cloud is not only a feasible concept but also a viable computational alternative that offers high levels of reliability and can at least offer reasonable performance, which at times may exceed the performance of a commercial cloud infrastructure.

Introduction: The incidence of anterior cruciate ligament (ACL) injuries in female floorball is relativley high, and the risk for sustaining a second ACL injury is greater compared to previously uninjured. Existing evaluation tests in rehabilition may not be discriminative enough to guide decisions on return to sport Aim: To evaluate the withinAsession and testAretest reliability of knee kinematics in floorball athletes and controls during a hop encompassing a sudden unanticipated change of direction. A second aim was to investigate the discriminative validity by comparing the test outcomes between the athletes and a control group of nonAathletes. Method: 11 elite floorball athletes and 8 controls were tested on two occassions separated by one to three weeks. Knee kinematics, ground contact time and number of succesfull hops were analyzed. Relative reliability was quantifyed by Intraclass correlation coefficient (ICC) and absolute reliability by standard error of measurement (SEM). Results: ICCs for knee kinematics withinAsession reliability were excellent (0.83A0.96) for athletes and poor to excellent (0.40A0.94) for the controls. For the testAretest reliability, the athletes had good to excellent (0.56A0.96) reliability and the controls had poor to excellent (0.26A0.93) reliability. Only two measured kinematic variables were significantly different between the groups. Conclusion: This pilot study indictate good to excellent reliability for the majority of the kinematic variables tested and, therefore, it could be assumed to be adequatley reliable. Discriminative validity needs to be further evaluated in a larger material.

This thesis applies structural reliability measures for the performance based design of residential PV system structures. These measures are intended to support designers in delivering systems with quantified and consistent reliability. Existing codified practices prescribe global factors (allowable stress design) and partial factors (load and resistance factor design) intended to provide an acceptable level of reliability as defined by historical practice. When applied to residential PV systems this prescriptive approach has two flaws, (1) calibration efforts needed to ensure consistency across structural system types have not kept up with the commercially available system types and (2) the actual expected reliability is not quantified and available to support decisions. The proposed reliability measures include probability of failure conditioned to wind speed in a fragility curve and the reliability index β, both of which are commonly used in performance based design. The approach is demonstrated through the application of the reliability measures to code compliant designs. Diverse system types are utilized to demonstrate how the existing code prescribed approach may lead to non-uniform structural performance. For each of the system types on which the reliability measures are demonstrated, a code compliant design is developed for three roof slopes, wind tunnel testing is conducted to provide an experimental measure of wind pressure coefficients, system specific fragility curves are generated to quantify the probability of failure conditioned to a set of wind speeds, and then, a site specific wind model is applied to produce a probability of failure and reliability index β. Through the performance based approach proposed in this thesis, two key outputs show non-uniform and unanticipated structural performance of PV systems designed according to the prescriptive code method. The two key outputs which illustrate this finding are fragility curves which illustrate the probability of failure over a range of wind speeds and reliability index, β values which couple the structural and wind distributions for a single measure of reliability.

No description available.

The objective of this work is to develop an effective TCAD based hot-carrier degradation model in predicting the damage that a SiGe HBT undergoes as it is stressed across bias, time and temperature.

The objective of this research is to characterize and manage lifetime reliability, microarchitectural performance, and power tradeoffs in multicore processors. This dissertation is comprised of three research themes; 1) modeling and simulation method of interacting multicore processor physics, 2) characterization and management of performance and lifetime reliability tradeoff, and 3) extending Amdahl’s Law for understanding lifetime reliability, performance, and energy efficiency of heterogeneous processors. With continued technology scaling, processor operations are increasingly dominated by multiple distinct physical phenomena and their coupled interactions. Understanding these behaviors requires the modeling of complex physical interactions. This dissertation first presents a novel simulation framework that orchestrates interactions between multiple physical models and microarchitecture simulators to enable research explorations at the intersection of application, microarchitecture, energy, power, thermal, and reliability. Using this framework, workload-induced variation of device degradation is characterized, and its impacts on processor lifetime and performance are analyzed. This research introduces a new metric to quantify performance-reliability tradeoff. Lastly, the theoretical models of heterogeneous multicore processors are proposed for understanding performance, energy efficiency, and lifetime reliability consequences. It is shown that these system metrics are governed by Amdahl’s Law and correlated as a function of processor composition, scheduling method, and Amdahl’s scaling factor. This dissertation highlights the importance of multidimensional analysis and extends the scope of microarchitectural studies by incorporating the physical aspects of processor operations and designs.

Techniques of performance-based design in fire safety have developed notably in the past two decades. One of the reasons for departing from the prescriptive methods is the ability of performance-based methods to form a scientific basis for the cost-risk-benefit analysis of different fire safety alternatives. Apart from few exceptions, observation of past fires has shown that the structure’s contribution to the overall fire resistance was considerably underestimated. The purpose of this research is to outline a risk-based design approach for structures in fire. Probabilistic methods are employed to ascertain uniform reliability indices in line with the classical trend in code development. Modern design codes for complex phenomena such as fire have been structured to facilitate design computations. Prescriptive design methods specify fire protection methods for structural systems based on laboratory controlled and highly restrictive testing regimes. Those methods inherently assume that the tested elements behave similarly in real structures irrespective of their loading, location or boundary conditions. This approach is contested by many researchers, and analyses following fire incidents indicated alarming discrepancy between anticipated and actual structural behaviour during real fires. In formulating design and construction codes, code writers deal with the inherent uncertainties by setting a ceiling to the potential risk of failure. The latter process is implemented by specifying safety parameters, that are derived via probabilistic techniques aimed at harmonising the risks ensuing different load scenarios. The code structure addresses the probability of failure with adequate detail and accuracy. The other component of the risk metric, namely the consequence of failure, is a subjective field that assumes a multitude of variables depending on the context of the problem. In codified structural design, the severity of failure is implicitly embodied in the different magnitudes of safety indices applied to different modes of structural response. This project introduces a risk-based method for the design of structures in fire. It provides a coherent approach to a quantified treatment of risk elements that meets the demands of performance-based fire safety methods. A number of proposals are made for rational acceptable risk and reliability parameters in addition to a damage index with applications in structural fire safety design. Although the example application of the proposed damage index is a structure subjected to fire effects, the same rationale can be easily applied to the assessment of structural damage due to other effects.

Thesis (MScEng)--University of Stellenbosch, 2001. / ENGLISH ABSTRACT: First-order second-moment structural reliability methods are used to assess the reliability of a prestressed concrete beam. This beam was designed for imposed office floor loads and partitions following the limit states design method as provided for by the applicable South African structural codes, viz SABS 0100-1:1992 and SABS 0160:1989. The reliability is examined at two limit states. At the ultimate limit state of flexure the ultimate moment of resistance must exceed the applied external moment at the critical section, while at the serviceability limit state of deflection the deflection must satisfy the codespecified deflection criteria. Realistic theoretical models are selected to express the flexural strength and deflection of the prestressed concrete member, while appropriate probabilistic models are gathered from the literature for loading, resistance and modelling uncertainties. The calculated reliability index at the ultimate limit state of flexure (3.10) is lower than expected in view of the fact that this represents a non-critical limit state in the case of a Class 2 prestressed concrete member. This condition can be explained with reference to the relatively high uncertainty associated with the modelling error for flexural strength. The calculated reliability index at the serviceability limit state of deflection (l.67) compares well with acceptable practice. The study further focuses on the sensitivity of the reliability at the two limit states of interest to uncertainty in the various design parameters. The ultimate limit state of flexure is dominated by the uncertainty associated with the modelling error for flexural strength, while the contribution to the overall uncertainty of the ultimate strength and area of the prestressing steel and the effective depth is less significant. In comparison the reliability at the serviceability limit state of deflection is not dominated by the uncertainty associated with a single basic variable. Instead, the uncertainty associated with the modelling error, creep factor and prestress loss factor are all significant. It was also demonstrated that the variability in beam stiffness is not a major source of uncertainty in the case of a Class 2 prestressed concrete member. It is recommended that the present code provisions for ultimate strength and deflection should be reviewed to formulate theoretical models with reduced systematic and random errors. The effect of the uncertainty associated with the creep and prestressed loss factors should also be adressed by adjustment of the partial material factor for concrete at the serviceability limit state of deflection. Furthermore, research must be directed towards formulating an objective failure criterion for deflection. The uncertainty in the deflection limit must therefore be quantified with a probability distribution. / AFRIKAANSE OPSOMMING: Eerste-orde tweede-moment struktuur betroubaarheid metodes word ingespan om die betroubaarheid van 'n voorspanbeton balk te bereken. Hierdie balk is ontwerp vir opgelegte kantoor vloerbelasting en partisies volgens die grenstoestand ontwerp metode soos beskryf in die toepaslike Suid-Afrikaanse boukodes, naamlik SABS 0100-1: 1992 en SABS 0160: 1989. Die betroubaarheid word ondersoek by twee grenstoestande. By die swiglimiet van buiging moet die weerstandsmoment die eksterne aangewende moment oorskrei by die kritieke balksnit, terwyl die defleksie die kriteria soos voorgeskryf deur die kode moet bevredig by die dienslimiet van defleksie. Realistiese teoretiese modelle word gebruik om die buigsterkte en defleksie van die voorspanbeton balk te bereken. Verder is geskikte waarskynlikheid modelle uit die literatuur versamelom die belasting, weerstand en modelonsekerhede te karakteriseer. Die betroubaarheid indeks soos bereken vir die swiglimiet van buiging (3.10) is laer as wat verwag sou word in die lig van die feit dat hierdie nie 'n kritieke grenstoestand verteenwoordig in die geval van 'n Klas 2 voorspan element nie. Dit kan verklaar word met verwysing na die relatiewe groot onsekerheid wat geassosieer word met die modellering fout vir buigsterkte. Die berekende betroubaarheid indeks vir die dienslimiet van defleksie (1.67) vergelyk goed met aanvaarde praktyk. Die studie fokus verder op die sensitiwiteit van die betroubaarheid by die twee grenstoestande onder beskouing ten opsigte van die onsekerheid in die verskillende ontwerp parameters. By die swiglimiet van buiging word die onsekerheid oorheers deur die bydrae van die modelering fout vir buigsterkte. Die bydraes tot die totale onsekerheid deur die swigsterkte en area van die voorspanstaal sowel as die effektiewe diepte is minder belangrik. By die dienslimiet van defleksie word die betroubaarheid nie oorheers deur die onsekerheid van 'n enkele basiese veranderlike nie. In stede hiervan is die onsekerheid van die modellerings fout, kruipfaktor en voorspan verliesfaktor almal noemenswaardig. Daar word verder aangetoon dat die veranderlikheid in balkstyfheid nie 'n belangrike bron van onsekerheid in die geval van 'n Klas 2 voorspan element is nie. Daar word aanbeveel dat die bestaande voorskrifte in die kode vir buigsterkte en defleksie aangespreek moet word deur teoretiese modelle met klein modelonsekerhede te formuleer. Die uitwerking van die onsekerheid van die kruip- en voorspan verliesfaktore kan aangespreek word deur 'n aanpassing te maak in die parsiële materiaalfaktor vir beton in die geval van die dienslimiet van defleksie. Navorsing moet verder daarop gemik wees om 'n objektiewe falingskriterium vir defleksie te formuleer. Die onsekerheid van die toelaatbare defleksie moet dus gekwatifiseer word deur 'n waarskynlikheidsverdeling.

published_or_final_version / Statistics / Doctoral / Doctor of Philosophy