Thesis (MEng in System Engineering)--University of South Australia, 1995

The research presented in this thesis focuses on the quantification of the effect of limited site investigations on the design and performance of pile foundations. Limited site investigation is one of the main causes of structural foundation failures. Over the last 30 years, most site investigations conducted for infrastructure projects have been dictated by minimum cost and time of completion, rather than meeting the need to appropriately characterise soil properties (Institution of Civil Engineers 1991; Jaksa et al. 2003). As a result, limited site investigations remain common, resulting in a higher risk of structural foundation failure, unforeseen additional construction, and/or repair costs. Also, limited site investigations can result in over-designing foundations, leading to increased and unnecessary cost (ASFE 1996). Based on the reliability examination method for site investigations introduced by Jaksa et al. (2003) and performed by Goldsworthy (2006), this research investigated the effect of limited site investigations on the design of pile foundations. This was achieved by generating three-dimensional random fields to obtain a virtual site consisting of soil properties at certain levels of variability, and by simulating various numbers of cone penetration tests (CPTs) and pile foundations on the generated site. Once the site and the CPTs were simulated, the cone tip resistance (qc) was profiled along the vertical and horizontal axes. The simulated qc profiles yielded by the CPTs were then used to compute axial pile load capacity termed the pile foundation design based on site investigations (SI). In parallel, the axial pile load capacity of the simulated pile foundation utilising the “true” cone tip resistance along the simulated pile was also determined. This is termed “the true” design, or the benchmark pile foundation design, and referred to as pile foundation design based on complete knowledge (CK). At the end of this process, the research compared the pile foundation designs based on SI and those based on CK. The reliability of the foundation design based on SI was analysed with a probabilistic approach, using the Monte Carlo technique. The results indicated that limited site investigations have a significant impact on the design of pile foundations. The results showed that minimum sampling efforts result in a high risk of over- or under-designing piles. More intensive sampling efforts, in contrast, led to a low risk of under- or over-design. The results also indicated that the levels of spatial variability of the soil are notable factors that affect the effectiveness of site investigations. These results will assist geotechnical engineers in planning a site investigation in a more rational manner with knowledge of the associated risks. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1375097 / Thesis (M.Eng.Sc.) - University of Adelaide, School of Civil, Environmental and Mining Engineering, 2009

The research presented in this thesis focuses on the quantification of the effect of limited site investigations on the design and performance of pile foundations. Limited site investigation is one of the main causes of structural foundation failures. Over the last 30 years, most site investigations conducted for infrastructure projects have been dictated by minimum cost and time of completion, rather than meeting the need to appropriately characterise soil properties (Institution of Civil Engineers 1991; Jaksa et al. 2003). As a result, limited site investigations remain common, resulting in a higher risk of structural foundation failure, unforeseen additional construction, and/or repair costs. Also, limited site investigations can result in over-designing foundations, leading to increased and unnecessary cost (ASFE 1996). Based on the reliability examination method for site investigations introduced by Jaksa et al. (2003) and performed by Goldsworthy (2006), this research investigated the effect of limited site investigations on the design of pile foundations. This was achieved by generating three-dimensional random fields to obtain a virtual site consisting of soil properties at certain levels of variability, and by simulating various numbers of cone penetration tests (CPTs) and pile foundations on the generated site. Once the site and the CPTs were simulated, the cone tip resistance (qc) was profiled along the vertical and horizontal axes. The simulated qc profiles yielded by the CPTs were then used to compute axial pile load capacity termed the pile foundation design based on site investigations (SI). In parallel, the axial pile load capacity of the simulated pile foundation utilising the “true” cone tip resistance along the simulated pile was also determined. This is termed “the true” design, or the benchmark pile foundation design, and referred to as pile foundation design based on complete knowledge (CK). At the end of this process, the research compared the pile foundation designs based on SI and those based on CK. The reliability of the foundation design based on SI was analysed with a probabilistic approach, using the Monte Carlo technique. The results indicated that limited site investigations have a significant impact on the design of pile foundations. The results showed that minimum sampling efforts result in a high risk of over- or under-designing piles. More intensive sampling efforts, in contrast, led to a low risk of under- or over-design. The results also indicated that the levels of spatial variability of the soil are notable factors that affect the effectiveness of site investigations. These results will assist geotechnical engineers in planning a site investigation in a more rational manner with knowledge of the associated risks. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1375097 / Thesis (M.Eng.Sc.) - University of Adelaide, School of Civil, Environmental and Mining Engineering, 2009

Security of a Multi-Processor System on Chip (MPSoC) is an emerging area of concern in embedded systems. MPSoC security is jeopardized by Code Injection attacks. Code Injection attacks, which are the most common types of software attacks, have plagued single processor systems. Design of MPSoCs must therefore incorporate security as one of the primary objectives. Code Injection attacks exploit vulnerabilities in \trusted" and legacy code. An architecture with a dedicated monitoring processor (MONITOR) is employed to simultaneously supervise the application processors on an MPSoC. The program code in the application processors is divided into basic blocks. The basic blocks in the application processors are statically instrumented with special instructions that allow communication with the MONITOR at runtime. The MONITOR verifies the execution of all the processors at runtime using control flow checks and either a timing or instruction count check. This thesis proposes a monitoring system called SOFTMON, a design methodology called SHIELD, a design flow called LOCS and an architectural framework called CUFFS for detecting Code Injection attacks. SOFTMON, a software monitoring system, uses a software algorithm in the MONITOR. SOFTMON incurs limited area overheads. However, the runtime performance overhead is quite high. SHIELD, an extension to the work in SOFTMON overcomes the limitation of high runtime overhead using a MONITOR that is predominantly hardware based. LOCS uses only one special instruction per basic block compared to two, as was the case in SOFTMON and SHIELD. Additionally, profile information is generated for all the basic blocks in all the application processors for the MPSoC designer to tune the design by increasing or decreasing the frequency of loop basic blocks. CUFFS detects attacks even without application processors communicating to the MONITOR. The SOFTMON, SHIELD and LOCS approaches can only detect attacks if the application processors communicate to the MONITOR. CUFFS relies on the exact number of instructions in basic blocks to determine an attack, rather than time-frame based measures used in SOFTMON, SHIELD and LOCS. The lowest runtime performance overhead was achieved by LOCS (worst case of 37.5%), while the SOFTMON monitoring system had the least amount of area overheads of about 25%. The CUFFS approach employed an active MONITOR and hence detected a greater range of attacks. The CUFFS framework also detects bit flip errors (reliability errors) in the control flow instructions of the application processors on an MPSoC. CUFFS can detect nearly 70% of all bit flip errors in the control flow instructions. Additionally, a modified CUFFS approach is proposed to ensure reliable inter-processor communication on an MPSoC. The modified CUFFS approach uses a hardware based checksum approach for reliable inter-processor communication and incurred a runtime performance overhead of up to 25% and negligible area overheads compared to CUFFS. Thus, the approaches proposed in this thesis equip an MPSoC designer with tools to embed security features during an MPSoC's design phase. Incorporating security measures at the processor design level provides security against software attacks in MPSoCs and incurs manageable runtime, area and code-size overheads.

Thesis (Ph. D.)--Texas State University-San Marcos, 2007. / Vita. Appendices: leaves 194-207. Includes bibliographical references (leaves 208-209).

Thesis (M.S. in Electrical Engineering)--Naval Postgraduate School, December 2006. / Thesis Advisor(s): Ha, Tri T. ; Su, Weilian. "December 2006." Description based on title screen as viewed on March 12, 2008. Includes bibliographical references (p. 55-56). Also available in print.

Thesis (MSc)--University of Stellenbosch, 2007. / Bibliography. Also available via the Internet.

Thesis (Ph.D.)--Auburn University, 2007. / Abstract. Includes bibliographic references (ℓ. 73-79)

Thesis (Ph.D.) -- University of Newcastle, 2003. / School of Engineering. Includes bibliographical references (leaves 248-256). Also available online.

Thesis (M. Phil.)--University of Hong Kong, 2008. / Also available in print.