Research Publications (Engineering and Built Environment)

Permanent URI for this collectionhttp://ir-dev.dut.ac.za/handle/10321/215

Browse

Search Results

Now showing 1 - 10 of 31

Towards a hierarchical and distributed power management framework for SGs
(IEEE, 2023-11-16) Nleya, Bakhe; Shezi, Nokwanda
A resources allocation scheme for joint optical wireless transport networks
(IEEE, 2023-08-03) Molefe, Mlungisi; Sibiya, Khulekani; Nleya, Bakhe
As the future of networking dives into a new era of connecting every single physical device into the internet termed Internet of Things (loT), this significantly means a rapid increase in the number of online connected devices, which leads to more bandwidth hungry and data consuming devices. The fifth generation (5G) of mobile communication has been deployed already in multiple countries, therefore researchers have migrated their focus to the sixth generation (6G) of mobile communication to cater for extensive coverage and massive number of loT devices. A promising architecture and technology to cope with massive number of online devices and extensive coverage is a joint optical wireless transport network which offers comparably ultra-high systems capacity and extremely low latency while maintaining an improved quality of service. Furthermore, an optical wireless transport network can accommodate high speed mobility for frequently moving end user devices which is essential for 6G. In this paper our focus is to explore and propose an ultimate optical wireless transport network architecture scheme that will cater for loT as well as networks beyond 5G. We thus propose an innovative Optical-Backhaul and Wireless Access (OBWA) network architecture as a favorable solution for future networks. We further present a joint channel and route allocation (JCRA) scheme for achieving optimal quality of experience. Performance evaluation of the proposed JCRA scheme for OBW A network architecture show a significant improvement in the network throughput as well as the network end-to-end delay despite varying load traffic or varying flow channels.
Efficient index modulation techniques for 5G and beyond
(IEEE, 2023-11-16) Mgobhozi, Bhekinkosi; Nleya, Bakhe
Index modulation (IM) techniques are emerging as promising approaches to improve spectral efficiency and reliability for 5G/future wireless networks. This paper pro vides an overview of key IM schemes including spatial modula tion, orthogonal frequency division multiplexing with index modulation (OFDM-IM), and enhanced variants like SIM OFDM. The spectral efficiency gains of these techniques are quantified mathematically. Spatial modulation uses antenna indices to convey additional information bits, while OFDM-IM utilizes subcarrier indices. SIM-OFDM further activates sub carriers selectively based on incoming bits. Experiments demonstrate spectral efficiency improvements over classical OFDM.
Resources allocation for hybrid cloud-edge computing in 5G network slicing
(TELKOM, 2023-09-01) Molefe, Mlungisi; Sibiya, Khulekani; Nleya, Bakhe; Smuts, Martin; Taute, Anton
In typical heterogeneous networks such as 5G and beyond, innovative technologies such as edge computing and network slicing can enhance overall network performance in terms of handling critical mission services as they often require extremely low latencies. Notably, network slicing facilitates the provisioning of virtual slices with different characteristics to serve different end-user requirements. The Network operator achieves this goal by utilizing the already existing physical wireless network resource. Current resource provisioning schemes suffer inadequacies in scalability and flexibility Thus to support both Cloud and Edge Computing in 5G and beyond networking, the work herein proposes a novel low latency scheme that affords dynamic and intelligent allocation of multi-dimensional resources. It bases on a Hybrid Cloud-edge Network Slicing (HCENS) architecture on leveraging both Cloud and Edge Computing The proposed scheme creates a flexible, scalable as well as energy efficient resource provisioning. Its architecture comprises both centralized units (CUs) and distributed units (DUs). These provide storage, that in turn enhances function partitioning for various network slices. Several agent-based simulations scenarios are carried out in evaluating the efficacy of the proposed scheme. Obtained analytical and simulation results indicate drastic reductions in network latencies for critical mission end user services. This couples with reductions in storage requirements.
Utilizing lean techniques through a P-D-C-A approach to drive built-in-quality in a thermoforming line for an automotive component manufacturer
(Ponte Academic Journal, 2023-08) Makua, Edward Sontaga; Nleya, Bakhe; Dewa, Mendon
Automotive component manufacturers are faced with competitive challenges globally. At the heart of these challenges is the need to produce parts that are of high-quality standards. Now more than ever before, organizations need to work as a unit to drive the fundamentals of built in-quality if they are to maintain good quality standards and overall cost competitive leadership. An automotive component manufacturer was struggling with managing the thermoforming line outflow of defects to their customers. Substantial amounts of resources in the form of inspectors were being added on their manufacturing lines to form quality gates and inspect for quality. This has resulted in cost implications and major profitability losses and further expanded their cost of quality. The severity of the problem was further exacerbated by the acceptance of the quality gates as a status quo with little or no initiatives to turn around the situation. The aim of this study was to subdue the traditional quality management approach of inspecting for quality by introducing a series of lean manufacturing techniques that process capability and built-in quality. The study streamlined a sequence of lean manufacturing techniques that supported each other in their findings and results to develop a Plan-Do-Check-Act (PDCA) based strategic approach to drive in-process quality capability in the manufacturer’s Thermoforming line. The results of study were significant manpower reduction on the lines, improved quality capability results, reduced expected scrap rates and reworks, and an overall strategic guideline for the implementation of built-in-quality in automotive component manufacturer’s thermoforming line
A D2D communication based lightweight customer side data securing scheme in smart grids
(2022-08-20) Nleya, Bakhe; Khumalo, Phlani
With the emergence of modernized power grids into smart equivalents referred to as smart grids (SGs) the bulk generation, transmission, distribution, and end-user infrastructures must be appropriately long-term planned concurrently with the required privacy and security. Notably, the objectives of modern SGs are to minimize power energy losses through theft or physical dissipation. The embedded device-to-device (D2D) communication technology in 5G networks will enable an affordable fail-safe ICT subsystem platform for the SGs. However, Privacy preservation is necessary for D2D services in SGs. In this paper, we propose an anonymity privacy-preserving, and data aggregation scheme. We carry out both security and performance and obtained theoretical analysis and simulation results the privacy algorithm is effective and at the same time, fewer communication overheads are exchanged.
A lightweight based data aggregation scheme for smart grid power systems
(2022-08-20) Khumalo, Philani; Nleya, Bakhe
To accomplish data aggregation securely and efficiently, it is necessary to design a scheme that is low in both computational as well as communication overheads. Thus in this paper, we propose and analyze a novel secured data aggregation scheme that ensures both privacy preservation as well as data integrity. The scheme is centered on forecasting power consumption demands for a particular neighborhood, and overall, because most attacks occur, during the transmission of data across the ICT subsystem, it thus focuses on limiting that. It does so by first forecasting its demands, and only links with the utility operator when adjustments have become necessary. The scheme utilizes a lightweight efficient noninteractive authentication mechanism in the generation and sharing of session keys. Overall, both the security analysis and performance evaluation demonstrate its efficacy in guaranteeing both privacy and security in addition to minimizing computational and communication overheads.
Power demand and supply optimization in islanded microgrids with distributed generation
(IEEE, 2022-01-25) Chidzonga, Richard; Nleya, Bakhe; Khumalo, Philani
In the power sector, a shift from the present fossildominated generation to renewable as well as energy-efficient generation and distribution is firmly underway. The transition is mostly driven by the digitalization of the energy systems to what has been coined ENERNET meaning energy network. Numerous benefits for both utility and consumers accrue. Digitalization enables more activity in the power trading market and a large amount of consumer data becomes available in the sector. Overall, strides are being made in the integration of Demand-Side Management (DSM) in the planning of Isolated/Islanded Microgrids (IMGs) as these will potentially reduce total OPEX costs at both customer and utility levels as well as increase renewable energy utilization. However, there is paucity in literature regarding distributed generators (DGs) non-convex cost function. Notably, not much has been covered regarding microgrid optimal load-dispatching especially with regards to optimizing algorithms. In this paper, we focus on formulating the day-ahead dispatch problem of microgrids with DGs subject to non-convex cost function and load dynamics. We first propose an operational framework that addresses the DG's 'valve point' loading effect as well as optimizing its performance. The impact of DSM on convex and non-convex EMS problems with different load participation levels is investigated. Further, the day-ahead scheduling horizon of fifteen-minute resolution time is considered to examine the effect of load dynamics in the microgrid. A Quantum Particle Swarm based approach is employed to solve non-convex DGs cost optimization. It is demonstrated that the proposed algorithm efficiently solves the non-convex EMS problem. Simulation results yield a 5% reduction in OPEX costs without compromising customer satisfaction.
An energy efficient resource provisioning scheme for joint all photonic and wireless networks
(IEEE, 2022-08-04) Molefe, Mlungisi; Nleya, Bakhe
Flexible joint all photonic and wireless transport networks are a promising backbone network technological solution to accommodate the various dynamic bandwidth natured applications. In this paper, we mitigate methods aimed at maximizing available resources in a joint-photonic and wireless transport network in an energy-efficient manner. In so doing, we take into account challenges posed by transmission impairments as they tend to degrade signals and reduce their optical reach. We thus propose a state-of-the-art and innovative network architecture that can efficiently process a large amount of data. This architecture is designed to cater for bandwidth-hungry and bandwidth-dynamic applications and services. We further propose a load aware energy efficient resource optimization scheme using LERA algorithm, which couples with related service reconfiguration functions to maximize higher spectral efficiencies and minimal blocking in both optical and wireless sections of the composite network. Performance analysis shows that the proposed architectural scheme based on the LERA algorithm outperforms traditional benchmark techniques in achieving minimal blocking while efficiently maximizing spectral utilization compared to legacy benchmark approaches. Hence this scheme drastically enhances the utilization of the available resources and the overall network throughput.
An energy efficient resources allocation scheme for flexible translucent optical transport networks
(Psychology and Education, 2021-03-02) Nleya, Bakhe; Molefe, Mlungisi; Chidzonga, Richard
The present study attempts to explore how academic streams and learning styles play role in the preferences of coping strategies among prospective teachers. A quantitative approach was selected to explore the relationship. A survey was conducted with 300 prospective teachers (150 of science stream and 150 of humanities stream). A multi-stage random sampling technique was used to collect relevant information. Research instrument to measure coping strategies was developed by the researcher himself and Learning Style Inventory (LSI) by Ritu Dangwal & Sugata Mitra, 1997 was used to measure learning styles of prospective teachers. Statistical techniques i.e. mean, S.D., multivariate ANOVA were applied. Results revealed an essential significant effect of academic streams and learning styles on preference of coping strategies among prospective teachers. It is recommended that teacher training institutions should establish guidance or counseling centers to provide counseling to prospective teachers regarding coping skills and learning styles

Research Publications (Engineering and Built Environment)

Browse

Filters

Settings

Sort By

Results per page

Search Results