The following summaries show the incredibly useful, practical and indeed critical thesis projects that our students have been producing.
The summaries are merely a suggestion of the hard and astute work that is involved; they hypothesize, test and ultimately make recommendations based on their results.
Our students are, by and large, working in industry and their projects certainly reflect this.
They are genuinely contributing to improvements within their industries and more specifically to their workplaces.
They detect problems, inefficiencies, safety concerns, unnecessary cost blow-outs, or indeed a combination of them all. And they then tackle these issues with astounding results.
This list reflects only a small number of the projects that our students have completed. More will be added - watch for a growing number of summaries on this page.
Wastewater Automation – the Development of a Low Cost Distributed Automation System
Authors: Tom Davies
The purpose of this thesis was to address a range of issues at a Wastewater scheme in Western Australia. Originally the alarms at the facility merely indicated the location of a problem. This lack of information made trending, logging, and trouble-shooting very difficult. The implementation of the author’s ‘ZigBee’ technology was deemed successful. It has resulted in a more cost-effective and reliable system and has proved much easier to install.
Process Control Networks Security
Author: Omar Salim Kidege
The purpose of this thesis was to analyse the risks and benefits surrounding the use of Process Control Networks across all critical infrastructure. The author made recommendations to help improve security and mitigate threats to companies utilising PCNs. The recommendations included implementing multiple layers of security, embracing Computer Emergency Response Teams and making use of ethical hacking to identify weak points in security.
Nickel Smelter Converter Slag Skimming Based on Thermal Imaging
Author: Muhammad Naveed-Ul-Hassan
The purpose of this thesis was to help reduce the inaccuracies and dangers of visual slag identification at a Nickel Smelter. Testing involved thermal imaging and remote skimming equipment. They allowed operators to work more accurately and with increased safety. It was determined that the combined use of thermal imaging and IP cameras allow operators to perform the skimming requirements from a remote location, with an improvement in efficiency and accuracy, reducing nickel loss.
Optimal Feed Density and Control for a Reflux Classifier Fluidised Bed Separator
Author: Michael Calvert
The aim of this thesis was to research the optimal parameters in which to conduct reflux classification in coal separation plants. The relevant factors were determined through both research and the analysis of the resulting effects when certain parameters were changed. They were measured in relation to the yield of a coal plant. It was determined that using higher density coal slurry increases yield.
Automatic Machine Vision-Based Quality Inspection of Plastic HDPE Bottle Closures in a Water Bottling Plant
Author: Mhene Givemore
The aim of this thesis was to formulate a system of checking the quality of plastic products as they pass on a conveyor. It was designed to improve on the current quality check systems with the goal of increasing profits for the manufacturing companies. A machine visions system was selected: to analyse each passing product and remove it from the production line if defective. This led to a significant reduction in the rejection of final products because the defective parts were removed before ruining the final product. Defective colour was the only aspect tested during these experiments. In the future, however, shape, dimensions, and deformities are also plausible variables to test for.
Feasibility Study: Installing Online Vibration Monitoring and AN Analysis System in an LNG plant (GE System1)
Author: Kannan Periyaiya
The purpose of the study was to explore the critical aspects of implementing the online vibration monitoring system. The focus: the effectiveness of the GE system 1 continuous condition monitoring software, when used in a gas plant. The main analysis comprised a comparison between the performance of the offline and online condition monitoring systems. It was concluded that the implementation of the new, online system improved productivity, safety, and overall employee satisfaction with the data analysis software, and so was determined to be a feasible replacement within cost requirements.
Optimisation of 2D Matrix Bar-Code Reading and Motion Control
Author: Bryan Crump
The aim of this thesis was to find an optimal solution by comparing methods of automated part positioning in production. The first method involved the existing process which is carried out by many production facilities. The second was developed by the thesis author as a theoretically more optimised solution. Both methods utilise 2D marked barcodes to ensure the accuracy of placements. After experimentation, it was concluded that the 2nd test process was in fact more accurate for a broader range of variation in positioning. Its application in manufacturing systems optimised the speed and accuracy of production.
Developing a Framework to Mitigate the Growing Incidents of Cyber-Security Threats on Process Control Networks(PCN): A Case Study of the Petrochemical Industry
Author: Abimbola Ogunlade
The aim of this research was to analyse the growing trend of cyber-attacks on process control networks, with the goal of constructing a framework for mitigating the threats these attacks pose. The study sought to establish the types of cyber exploits possible in a PCN environment, to detect and prevent vulnerabilities and to determine if these could adapt to the growing sophistication of cyber intrusions.
Using ML analysis, a multiplayer perceptron, and Naïve Bayes, the classification of data as malicious threats or normal traffic ranged from 98%, 97% and 94% respectively. Due to the relatively large gaps in classification success, it was recommended that the systems be implemented in conjunction with human-based classification protocols using dynamic alert systems, allowing ambiguous or suspicious traffic to be either confirmed or denied as malicious traffic by an operator.