Theses and Dissertations (Electrical, Electronic and Computer Engineering)

Permanent URI for this collectionhttp://hdl.handle.net/2263/31930

Browse

Now showing 1 - 20 of 471

Model predictive static programming control applied to mineral processing plants
(University of Pretoria, 2023-05) Le Roux, Johan Derik; z.noome@gmail.com; Noome, Zander Meindert
In a mineral processing plant, the separation of valuable material from ore has multiple stages. Usually, the ore is crushed or ground into smaller parts through multiple crushers or grinding mills. This is called the communition process. This process is typically the first stage for extracting valuable material and is important for further down-stream processes. The output of the communintion stage is usually regulated to achieve a stable throughput and a specific ore particle size. After the ore is crushed and ground to a specified size, the valuable material in the ore needs to be separated from the undesired materials. The properties of the desired material influence the method used for separation. These methods include froth flotation, gravitational separation, magnetic separation and electrostatic separation. The separation process can include multiple process streams to get a high grade of the desired minerals out of the ore. In froth flotation, the main objective is to extract the desired material from the ore to obtain a large mineral recovery. Because the flotation process relies on the flotation of particles, particle size is extremely important. The use of control systems in mineral processing plants has been adopted to improve throughput, optimize power usage, ensure safe process operation and to running at a stable operating condition. The control of these plants makes use of different advanced process control strategies which include but are not limited to cascaded control, where multiple layers of control systems are applied, and model predictive control. These different control strategies can range from regulatory control to supervisory control. Because of the large number of inputs to these plants, efficient controllers are necessary to obtain desired results. The use of Nonlinear Model Predictive Control (NMPC) is an attractive option for most mineral processing plants because of the constraint management capabilities of the controller. Unfortunately, the NMPC method has a large computational load which requires sufficient resources to make it a viable option. Another model predictive control method known as Model Predictive Static Programming (MPSP) has shown promise to improve the computational time of a standard NMPC controller. The MPSP control philosophy generates a static optimization problem which is less computationally difficult to solve compared to the dynamic optimization problem that is generated through NMPC. In this dissertation, the control of a single-stage grinding mill circuit and a four-cell flotation circuit with an MPSP controller to reduce the computational load is proposed. The computational efficiency and the output performance of MPSP controllers are compared to NMPC controllers as a motivation for the use thereof. The comparison is done by simulating two mineral processing stages, namely the communition phase and the separation phase. The simulations considered different configurations for both the MPSP and NMPC controllers. The comparison of the controllers in the simulations shows that the MPSP controller obtained similar or improved plant results while also having a reduced computational time compared to the NMPC controller. The MPSP controller also displays scalability improvements compared to the NMPC controllers which can be beneficial for supervisory control of large-scale processing plants.
Advancing environmental, social, and governance outcomes through process optimisation and control
(University of Pretoria, 2024-02) Le Roux, Johan Derik; Craig, Ian K.; burchell.john@gmail.com; Burchell, John James
Organisations are compelled to integrate Environmental, Social, and Governance (ESG) considerations into their core strategy, with the tightening of regulatory requirements and the mounting pressure from stakeholders for sustainable practices driving a trend toward socially responsible investing. Advanced process optimisation and control provides innovative solutions to support ESG objectives. This thesis explores two case studies aimed at enhancing the consistency of material flow and composition into metallurgical operations to improve overall processing efficiency. The first case study introduces a (μ+λ)-Evolutionary Strategy (ES) to solve the input blending problem for a base metal refinery (BMR), where variability in the feed of contaminants to the operation impact negatively on plant throughput, product quality, and harmful emissions. The algorithm outperforms baseline blending strategies demonstrating a significant improvement in the blended consistency of contaminant feed. In the second case study, a nonlinear Model Predictive Controller (NMPC) is developed and implemented on a surge tank for level averaging control in an industrial tailings reprocessing circuit. A rigorous dynamic model is derived to describe the rate of change of both the volume and density in these surge tanks. By simulation with industrial data it is demonstrated that the significant input disturbances typical to tailings reprocessing circuits drive a gain inversion in the density model of the surge tank. This gain inversion and the multivariable objectives of both density and flow disturbance attenuation motivates for a NMPC solution. Results presented show significant improvements in both the water recovery and the stability of mass flow of tailings in the circuit. These advanced optimisation and control solutions support ESG objectives across multiple dimensions. Improved input stability with the (μ +λ)-ES enhances the efficiency of downstream processes where contaminants are extracted, resulting in lower emissions, especially when hazardous reagents are involved in the extraction process. By improving the efficiency of contaminant extraction the need for rework of product that fail to meet specifications is minimised, which leads to a reduction in waste generation, conservation of resources, and lower energy consumption. Improved water recovery with the NMPC lowers the overall environmental footprint of the tailings reprocessing circuit by reducing water consumption and energy usage, while stability improvements positively impact recoveries, thereby reducing waste and supporting responsible resource management.
Extremum seeking control of grinding mill circuits based on grind curves
(University of Pretoria, 2023-05) Le Roux, Johan Derik; Craig, Ian K.; lukieski@gmail.com; Ziolkowski, Lukasz
Mineral processing plants include several operations to liberate the valuable minerals within raw ore material to produce a concentrate, which is processed into a usable product by a metallurgical refinery. A mineral processing plant consists of a comminution and a separation stage. During the comminution stage, the raw ore material is processed through a grinding mill circuit to liberate the valuable minerals by grinding the ore to fine particles. The product from the comminution stage is then processed at a separation stage, which separates the valuable minerals (concentrate) from the waste material (tailings). The comminution stage plays a crucial role in the mineral processing industry. It significantly impacts the net revenue generated by a mineral processing plant due to the high operating costs associated with liberating the valuable minerals from the ore material. A grinding stage operates efficiently if it is processing the ore material at its maximum capacity, minimizing power consumption while reducing the amount of valuables lost to the tailings stream. Therefore, the ore material should be sufficiently ground for effective separation in subsequent downstream processes. Ideally, the separation stage requires a consistent stream of fine particles for effective separation. It is challenging for plant operators to manually achieve the above-mentioned operational objectives, which motivates the need to adopt a suitable control framework and ensure an efficiently run process. The performance of a grinding mill circuit is measured by its throughput and grind quality. These performance indicators are inversely related to operational objectives. The challenge in controlling the grinding mill circuit arises in determining the optimal operating conditions to maximize the net revenue generated by the plant. The optimal operating conditions vary with different ore types and unknown disturbances, such as varying ore hardness, which can result in the comminution stage operating at sub-optimal operating conditions. Furthermore, grinding mills rely on the cascading motion of the ore material and grinding media to accelerate ore breakage. The cascading motion is a function of the fraction of the mill volume filled with ore and the mill rotating speed, which influences the breakage forces that occur between rocks. Therefore, selecting optimal operating conditions is a difficult task requiring frequent adjustments as the operating conditions vary. Grind curves are a valuable tool that establishes the relationship between the mill load filling and rotational speed to the grinding mill throughput, grind quality and power consumption for a given ore type. Generally, the curves show parabolic features and the peaks vary with changes in the ore characteristics. A model-free adaptive control strategy is proposed for optimizing the performance of a semi-autogenous grinding (SAG) mill based on grind curves to improve throughput or grind quality. The controller explores an unknown map in search of the extremum of the performance indicators along the grind curves. A perturbation-based (PESC), a time-varying parameter estimation-based (TESC), and a Nelder-Mead simplex-based (SESC) extremum seeking control method are considered to optimize the grinding mill performance. Several optimization strategies are investigated for an open grinding mill configuration and a closed grinding mill circuit, where the closed circuit is equipped with a screen or with a hydrocyclone classifier to recirculate oversized ore material for additional grinding. The challenge lies in implementing an efficient optimization model-free control framework that will effectively maximize the performance measures of the complex, non-linear behaviour of the grinding mill circuit.
Model-based estimation and control of wheel slip in locomotives
(University of Pretoria, 2023-10) Le Roux, Johan Derik ; charlvandemerwe@gmail.com; Van de Merwe, Charl Viljoen
This dissertation investigates wheel slip control of locomotive traction systems in the presence of non-linear wheel surface behaviour and varying adhesion conditions. It is difficult to determine when the maximum point of adhesion has been exceeded since the adhesion coefficient cannot be measured directly during the operation of the locomotive. Therefore, classical slip controllers suppress excessive slip by using predetermined thresholds for the slip velocities and accelerations of the axles. The classical methods are convenient but cannot maximise adhesion utilisation. Modern methods continuously modulate the torque and are expected to produce superior performance if implemented effectively. Most continuous controllers calculate the reaction torque using a generated slip ratio (slip velocity divided by the locomotive velocity) reference and a slip ratio estimate feedback. Computing the estimate depends on an accurate locomotive velocity estimate, which is difficult to obtain when all the wheelsets of a locomotive are driven. Slip ratio reference generation generally requires estimates of the slip ratio and adhesion coefficient or adhesion force. This dissertation focuses on producing accurate estimates to enable effective slip control. Adhesion force is the adhesion coefficient multiplied by the normal force. The adhesion coefficient is dependent on the rail conditions. Under constant rail conditions, it varies only with a wheel load and slip ratio change. Therefore, the normal forces, wheel velocities, and locomotive velocity should be modelled accurately to ensure the model produces realistic adhesion coefficients. A linearised railway vehicle model could be well over the 100th order. Such models are helpful for design and validation, but using such complex models in model-based filter or estimator design is impractical. In this dissertation, a new simulation model is developed that includes the longitudinal, pitch, vertical, and wheelset rotational dynamics. In addition, it includes a unique approach to the coupler force by modelling the wagons using a single-axle wheelset model. This model captured the desired dynamics, including wheelset torsional vibrations and oscillations in the pitch dynamics. A linear state-observable estimator is developed to produce estimates of slip ratios and adhesion coefficients. The estimation model is an adaptation of the simulation model, but the adhesion forces and coupler force are modelled as unknown disturbances. This estimator requires measurements of the locomotive longitudinal acceleration and velocity, body pitch angle and rate, and the motor angular velocities. The rail angle and motor torque estimates should be provided to the estimator. The estimates are used in a novel slip ratio reference adaptation method to provide a reference to an adaptive PI controller. The PI controller is used to compute the reaction torque to prevent unstable slip in the rear/reference wheelset, while a speed differential controller is used to prevent slip in the other wheelsets. The simulation results indicate that the estimator and controller configuration can suppress unstable slip under varying adhesion conditions, thereby preventing damage to the wheels and rail while ensuring maximum adhesion utilisation. Maximum adhesion utilisation allows a locomotive to increase its hauling capacity without increasing its mass.
Innovations in advanced regulatory control methods for modern distributed control systems
(University of Pretoria, 2024-09) Craig, Ian Keith; Le Roux, Johan Derik; gustaf.gous@gmail.com; Gous, Gustaf Zacharias
Many modern Distributed Control Systems (DCS) in industry are new replacements of previous versions of the same DCS vendor’s product line. During such upgrades the process is often automated using software to translate existing controller configurations as well as custom software to comply with the new system’s requirements and syntax. Doing this makes the upgrade process much faster and reduces the risk of introducing errors. It does, however, rob the control practitioner from making use of new features and capabilities of the new system. Therefore, there are many DCS in industry where only a small fraction of their newer capabilities are used. Many improvements in advanced regulatory control (ARC) that would improve control performance are available, but are never used. In order to show how modern DCS can enable more complex control solutions, four ARC level controllers and two stiction compensation algorithms, all more complex than current solutions typically found in industry, are introduced as examples of how increased complexity may provide increased control performance.
Current steering with sequential stimulation in cochlear implants
(University of Pretoria, 2017-12) Hanekom, J.J. (Johannes Jurgens); Roux, Johanie
Current steering has been proposed to increase place pitch resolution in cochlear implant (CI) users. Many studies have shown that a current steering effect can be achieved when simultaneous stimulation is used (Koch, Downing, Osberger, Litvak and Greco, 2007, Saoji and Litvak, 2010, Wu and Luo, 2013). Some literature has shown that a current steering effect can also be achieved when sequential stimulation is used (McDermott and McKay, 1994, Kwon and van den Honert, 2006, Swanson, 2008). Literature proposes different features that could underlie place pitch and consequently possibly also current steering effects (McDermott and McKay, 1994, Kwon and van den Honert, 2006, Swanson, 2008, Frijns, Kalkman, Vanpoucke, Bongers and Briaire, 2009, Macherey and Carlyon, 2012, Venter, 2015). The present study confirmed that a current steering effect can be achieved when sequential stimulation is used by using multi-dimensional scaling and statistical analysis in addition to the convention of using cumulative d' values to analyse pitch ranking results of current steering experiments. It was however observed that a current steering effect could only be achieved in listeners who were at least able to pitch rank the pitch of the two individual stimulating electrodes correctly according to expectation. The effect of different stimulation parameters on the pitch ranking ability of CI users during current steering experiments was investigated. Results showed that some parameters only had an effect on the pitch ranking performance of some listeners, while other stimulation parameters affected the results of all the listeners. Wider stimulation pulse widths, for example, led to improved pitch ranking results for some listeners. Most listeners benefited from wider electrode separation distances. Statistical analysis showed that there was a significant improvement in the pitch ranking performance of the listeners during experiments where the stimulation rate was the same as the rate indicated on the clinical MAP of the listener. Person-specific current distribution models were used to predict the cochlear position of different stimuli because of different features that could underlie place pitch, for each of the experiments for four of the listeners who participated in this study. The model predictions were related to the measured pitch ranking results using correlation and mutual information analysis. The results indicated that the current centroid at electrode level, the position of the peak current at the auditory nerve level (because of either an individual stimulating electrode or because of summed currents) and the centroid of neural activation could underlie place pitch. All these features except the position of the peak of the current distribution at the auditory nerve level because of each individual stimulating electrode could underlie current steering effects. Results showed that the centroid of the current distribution at the auditory nerve level probably does not underlie place pitch. Knowledge about the impact that different stimulation parameters have on the ability to achieve a current steering effect could result in more efficient implementation of current steering effects. Proper knowledge of which features underlie place pitch and current steering effects could be used to create models that can be used to predict the results of place pitch experiments.
Viability of cochlear travelling wave signal processing for cochlear implants
(University of Pretoria, 2016-02) Hanekom, J.J. (Johannes Jurgens); Schmidt, Larry
English: The travelling wave encodes acoustic information by stimulating the auditory nerve fibres. Understanding the travelling wave and its process is important for the development of cochlear implants speech processors. The development of a normal hearing auditory model, using a hydrodynamic model of the travelling wave to predict the nerve fibre spiking diagrams, marked the first stage of this study. This study then proceeded to look at the development of a travelling wave speech processing algorithm and model the electrical response due to the stimulation from the vocoder speech processor, and the travelling wave speech processor. The final stage was to predict whether temporal encoding occurred during cochlear implant stimulation for the vocoder speech processor and the travelling wave speech processor. The results showed that the travelling wave normal hearing model was able to predict the nerve fibre characteristics seen in measurements from literature. This showed that the mechanical encoding performed by the travelling wave is vital to the encoding of information in auditory nerve fibres. The travelling wave speech processor was able to encode temporal cues for pitch up to 1060 Hz, where the results for the vocoder speech processor showed the 300 Hz limit seen in other literature of phase locking. Mimicking the travelling wave in cochlear implant speech processors may potentially benefit the delivery of information to the auditory cortex for cochlear implant users. However, these results must be legitimised using animal models and psychoacoustic experiments.
Signal processing by octopus cells for acoustic and electrical hearing : a modelling study
(University of Pretoria, 2016-07) Hanekom, J.J. (Johannes Jurgens); Blignaut, Gertruida Elizabeth
A computational model of a single octopus cell as well as a population of octopus cells was developed. The models were used to investigate the ability of octopus cells to compensate for the travelling wave delay, remove jitter from the neural activity and encode pitch for normal hearing. Furthermore the response of octopus cells to cochlear implant (CI) stimulation with the ACE strategy was investigated to determine whether pitch can be extracted from CI stimulation in the same way as from acoustic stimulation. Their ability to extract the pulse rate from single-electrode stimulation was also investigated. The response of the octopus cells to single-electrode stimulation at different pulse rates was used to predict pulse rate difference limens, which were compared to psychoacoustic measurements found in literature. It was found that octopus cells are sensitive to the delay of synaptic inputs on their dendrites but are broadly tuned to this delay. By evaluating the jitter together with the travelling wave delay present in the activity of auditory nerve fibres (ANFs), it was determined that octopus cells may rather act as coincidence detectors, which extract common interspike intervals (ISIs) from many ANFs. The octopus cell model was found to encode the frequency of pure tones in their ISIs for pure tone acoustic stimulation. They were also found to encode the pitch of a vowel in their ISIs, which was the same as the fundamental frequency extracted from the vowel with a speech processing algorithm. The octopus cell model responded to the pulse rate of the CI stimulation and could therefore not extract the frequency of pure tones from CI stimulation in the same way as from acoustic stimulation. The entrainment of the modelled octopus cell population decreased when the pulse rate of a single electrode increased beyond 300 pps. Pulse rate difference limens were predicted from the standard deviation of the ISIs of the octopus cell population response to single electrode stimulation. The predicted difference limens were in the same range as measured values, which suggests that octopus cells may play a role in the measured perceptual limit at 300 pps. From the findings of this study it is suggested that CI stimulation strategies should be developed to encode pitch in the periodicity of their stimulation to enable octopus cells to extract pitch information from CI stimulation.
Automatic tuning of a MIMO PI controller of a flotation bank
(University of Pretoria, 2024-11) Le Roux, Johan Derik; Craig, Ian K.; albertusr7@gmail.com; Richter, Albertus Viljoen
The literature on the automatic tuning of PID controllers is surveyed. The automatic tuning methods are sorted into model based and model-free methods. The methods are further subdivided into the manner the system is perturbed. The method of Bayesian optimization is presented and discussed within the context of automatic controller tuning. The method used to constrain the Bayesian optimization is presented. The level flotation model is given and linearized. The controllers are given and discussed. The controller tuning strategies for both SISO and MIMO controllers are presented. A Bayesian optimization automatic tuner is implemented on SISO and MIMO PI controllers used to control the pulp levels in a flotation bank. The implemented automatic tuner achieves performance improvement for both SISO and MIMO cases without any noise present. The MIMO controller tuning is also implemented on a system with measurement noise present and the Bayesian optimization automatic tuner settings performed on par with a state of the art forward-feeding controller. The Bayesian optimization automatic tuner is constrained to ensure safety and stability. The constraints are found using a structured singular value analysis.
Model development and validation of an industrial natural gas well production network
(University of Pretoria, 2024-10) Craig, Ian K.; Le Roux, Johan Derik; DzedzemaneR@gmail.com; Dzedzemane, Rudzani
A transient state-space non-linear model is developed for a natural gas production network fed from multiple gas wellheads. The state-space model is developed by making use of the spectral element method for pipeline spatial discretization. Wellhead models are integrated into the pipeline models by making use of suitable boundary conditions based on the characteristic compatibility method. The models are validated against a large scale natural gas well production network. The validation shows that the model has a good prediction performance based on a low normalized root mean square error of at most 5.08% and a high Pearson correlation coefficient with measured plant data of at least 0.94. The good prediction response of the developed transient models make them suitable for use in model-based optimal control of natural gas well production networks. The resulting dynamic model can be easily adapted to a gas network of any configuration due to its modular form.
Deep learning approaches for fingerprint localization in low-power wide area networks
(University of Pretoria, 2024-10) Myburgh, Hermanus Carel; De Freitas, Allan; albert.lutakamale@tuks.co.za; Lutakamale, Albert Selebea
In recent years, low-power wide area networks (LPWANs), particularly long-range wide area networks (LoRaWAN), have been increasingly adopted into large-scale internet of things (IoT) applications due to their ability to offer energy-efficient and cost-effective long-range wireless communication. The need to provide location-stamped communications to IoT applications for meaningful interpretation of physical measurements from IoT devices has increased the demand to incorporate location estimation capabilities into LPWAN networks. Factors such as high-power consumption, high implementation costs, and poor localization performance in urban canyons or environments with many obstructions render outdoor localization solutions based on standalone GPS technology unfit for deployment in large-scale IoT applications, where the emphasis is on energy efficiency and cost-effectiveness. Implementing localization methods in short-range wireless communication networks, such as Bluetooth and ZigBee networks, to estimate locations of target nodes in large outdoor environments is also not economically feasible due to their short-range nature, as there will be a requirement for dense deployment of wireless nodes, leading to high implementation costs. In LoRaWAN (one of the key LPWAN technologies operating in unlicensed frequency bands), fingerprint-based localization methods are known to be robust in challenging environments with multipath and non-line-of-sight phenomena, making them relatively more accurate than range-based methods. However, most currently available fingerprint-based localization methods in LoRaWAN networks rely on conventional ‘shallow’ machine learning models. While such models may yield satisfactory results under specific conditions, their complexity tends to increase as the size of training datasets increases, ultimately resulting in a decline in localization accuracy. In this thesis, driven by the goal of improving the performance and efficiency of fingerprint-based localization methods in LoRaWAN networks, two deep learning-based fingerprint-based methods to estimate the locations of target nodes in LoRaWAN networks are proposed. The first proposed method is a branched convolutional neural network (CNN) localization method enhanced with squeeze and excitation (SE) blocks (referred to as the CNN-SE method). The second proposed method is a hybrid CNN-transformer fingerprint-based localization method (referred to as the CNN-transformer method). The main contribution of the first method is the joint use of CNN (proven to be very efficient in learning useful positional information in structured data) and SE blocks, which improves channel-wise interdependencies. The novel contribution of the second method is the development of a hybrid CNN-transformer fingerprinting-based localization model by leveraging the strengths of both CNNs and transformers. CNNs capture features from the input data at the local level, while the attention mechanism of the transformer captures features from the input data at the global level. Adopting a 0.7/0.15/0.15 data split scheme for the training, validation, and test set, respectively, and using the entire LoRaWAN dataset, the CNN-SE method achieved localization accuracies of 291.51 m and 147.55 m mean and median localization errors, respectively, on the test set, using the powed data representation scheme. With the CNN-transformer method, the localization accuracy of 288.1 m and 143.7 m mean and median localization errors, respectively, were achieved, using the same experimental settings. The localization accuracies achieved by these two methods have outperformed the localization accuracies of the currently available state-of-the-art fingerprint-based localization methods in the literature, evaluated using the same publicly available LoRaWAN dataset. An R2 score of 0.93 obtained by both methods further indicates the high degree to which the proposed methods have been able to fit data in their respective regressors, enabling them to localize target nodes with satisfactory localization accuracies.
Perception of timbre features by cochlear implant users
(University of Pretoria, 2024-08) Hanekom, J.J. (Johannes Jurgens); utiaan@gmail.com; Uys, Rudolph Christiaan
Timbre perception is fundamental to music enjoyment. Various cochlear implant (CI) studies have investigated the identification of musical instruments and the perception of timbre, finding that the ability of CI users to perceive timbre is poor compared to normal hearing (NH) listeners. To better understand timbre perception for CI users, the limitations that cause poor timbre perception in CI users should be investigated when performing timbre discrimination tasks. Therefore, the present study investigated the perception of the timbre by measuring discrimination abilities using just-noticeable differences (jnds) for a set of representative acoustic features that underlie timbre perception. Two spectral features, brightness and irregularity (referred to as brightness and IRR), and a temporal feature, logarithmic rise-time (LRT), were identified in the literature as salient timbre features. The timbre features were used in a synthesis model to create a set of nine synthetic instrument tones. The latter allows for independent variation of the timbre features. Synthetic tones were used in a two-alternative forced-choice (2AFC) experiment to measure jnds for NH individuals and CI users for each of the timbre features. The data showed that the jnds of CI users were larger than those of NH individuals. The findings suggested that CI users had difficulty to attend to the timbre feature when performing the discrimination tasks. To investigate whether or not CI users had access to the timbre features, electrodograms were used to analyse the jnds. Electrical stimulation pulse trains of the original instrument sound were generated and compared with the electrical stimulation pulse trains generated at jnd for each of the CI users. Difference metrics were calculated to determine whether CI users had access to the timbre feature or only to the difference between the reference and probe electrical stimulation signal. Spatial and temporal differences between reference and probe stimulation signals showed that CI users did not have access to the timbre feature, but rather to the differences in electrical pulse trains. The extent to which CI users received the timbre features was investigated using feature information transmission analysis (FITA). This estimated the percentage of available information of the timbre features that CI users received. Confusion matrices were predicted from the jnds of CI users to perform the FITA. The results showed that the information received by CI users is user-dependent and that the information received for each of the features is mostly the same within users. These findings support the notion that CI users probably did not attend to the timbre features and conceivably did not have adequate access to these. The representation of the spectral harmonics of the musical instrument tones by the electrical stimuli was investigated. The spectral representation in the electrical stimulus pattern was found to be a distorted version relative to that of the acoustic sound. The study aimed to answer the question To which extent is musical timbre perceived by CI users and what underlies this? A core objective was to understand what constraints underlie timbre perception. It was concluded that CI users do not attend to the timbre features when performing timbre discrimination tasks, and that the electrical stimuli representing the instrument have a distorted spectrum relative to that of the acoustic sound.
Bandwidth extended designs of double-ridged guide horn antennas
(University of Pretoria, 2024-08) Odendaal, Wimpie; Joubert, Johan; benniejacobs777@gmail.com; Jacobs, Benjamin
The explosive growth in bandwidth requirements on antennas and antenna systems, driven by various industries, also drives the need for the ever‐increasing bandwidth of antennas used for testing. The broadband Double-Ridged Guide Horn (DRGH) antenna finds widespread use in antenna measurement and ElectroMagnetic Compatibility/Interference (EMC/I) testing. An example of a DRGH used extensively for testing is the 1-18 GHz DRGH antenna. A study using ElectroMagnetic (EM) simulation was performed to determine the factors that limit the bandwidth of these antennas. All the parts and sub-assemblies of the DRGH were investigated to determine the impact of each part and or sub-assembly on the electrical performance of the antenna. This was done in simulation using reduced complexity models and parametric studies. The design changes that resulted from this study were implemented in several prototype antennas used for verification. It was found that it is possible to design DRGH antennas with bandwidth ratios of 100:1 and possibly beyond. It is expected that at higher frequencies, the limit will be the manufacturing tolerances and technology, and at lower frequencies, the maximum permissible size of the antenna.
Optimal pathway towards building EPC rating improvements align with building energy performance certificate programme
(University of Pretoria, 2024-10-15) Ye, Xianming; u13166426@tuks.co.za; Mokaile, Sifiso
This dissertation investigates an optimal pathway for commercial buildings to improve their Energy Performance Certificate (EPC) ratings by efficiently adopting Energy Efficiency Measures (EEMs). With buildings contributing approximately 40\% of global energy consumption and 36\% of carbon dioxide emissions, enhancing energy performance is critical. EPCs have been introduced globally to assess and promote energy efficiency in buildings. However, the uptake of EPCs and adherence to the recommended EEMs have been limited, especially in countries like South Africa and Scotland. In South Africa, only 1\% of the commercial buildings requiring EPCs have obtained them, and many building owners are hesitant to implement the recommended measures due to lack of trust, financial, and time constraints. This research addresses these challenges by proposing a model that optimises the selection and implementation of EEMs to achieve higher EPC ratings cost-effectively. The optimisation model, developed using MATLAB’s Genetic Algorithm (GA) solver, minimises the investment required while improving the energy performance of buildings step-by-step. The model considers key EEMs such as lighting system upgrades, HVAC improvements, and the integration of renewable energy sources like solar panels and battery storage. The approach allows building owners to make gradual improvements, balancing cost and time, leading to higher EPC ratings over time. To validate the proposed pathway, the study applied the model to a case study of a commercial building in Pretoria, South Africa. The results revealed that by adopting the optimal sequence of EEMs, the building could achieve significant energy savings while progressing through the EPC rating scales. The model demonstrated that a step-by-step approach can reduce the upfront financial burden compared to an aggressive all-at-once strategy. The study also makes recommendations for policymakers to refine EPC standards and support measures that incentivise building owners to participate in energy saving projects.
Reinforcement learning microservices scheduler in intelligent edge computing
(University of Pretoria, 2024-07-01) Abu-Mahfouz, Adnan Mohammed; Hancke, Gerhard P.; u22851217@tuks.co.za; Afachao, Kevin E.
The proliferation of internet of things (IoT) devices and resource-intensive applications has necessitated the development of intelligent edge computing frameworks. These frameworks aim to address challenges in the resource management, service latency, and data privacy of IoT devices. This research investigates the complex problem of microservice scheduling within intelligent edge computing environments. The focus is on optimising quality of service (QoS) metrics such as the latency, network bandwidth utilisation, and energy consumption during execution of resource-intensive applications. To address this challenge, a novel approach called the Bi-generic A2C Microservice Proxy Policy (BAMPP) is proposed. It leverages reinforcement learning (RL) principles to optimize microservice deployment in dynamic Edge-Cloud ecosystems. BAMPP uniquely considers the intricate inter-dependencies among microservices and adapts to user mobility in real-world scenarios. This research utilises a simulation platform to reproduce the intelligent edge computing environment, integrating real-world datasets to evaluate the performance of BAMPP against comparative algorithms. The research focuses on three key research points: identifying crucial factors influencing microservice scheduler performance, leveraging RL for optimised scheduling, and assessing the impact of random user mobility on service deployment. The results demonstrate BAMPP's superior performance in reducing energy consumption, minimizing network usage, decreasing execution and migration latency, and enhancing reliability in microservice scheduling compared to current systems. This research contributes to the field of intelligent edge computing by introducing a novel modeling approach, developing an advanced algorithm for joint optimization of scheduling and resource management, and providing comprehensive performance evaluations using realistic simulations. The results of this study have important ramifications for raising the effectiveness and performance of microservice applications in intelligent edge environments, potentially leading to cost savings, enhanced sustainability, and widespread implementation across diverse edge computing scenarios.
Optimal integration of solar home systems and appliance scheduling for residential homes under severe national load shedding
(University of Pretoria, 2024-09) Ye, Xianming; u15094822@tuks.co.za; Twala, Sakhile Nqobile
In developing countries such as South Africa, users experienced more than 1 030 hours of load shedding outages in just the first half of 2023 due to inadequate power supply from the national grid. Residential homes that cannot afford to take action to mitigate the challenges of load shedding are severely inconvenienced as they have to reschedule their demand involuntarily. This study presents optimal strategies to guide households in determining suitable scheduling and sizing solutions for solar home systems to mitigate the inconvenience experienced by residents due to load shedding. To start with, this study predicts the load shedding stages that are used as input for optimal strategies using the K-Nearest Neighbour (KNN) algorithm. Based on an accurate forecast of future load shedding outages, this study formulates inconvenience for residents and loss of power supply during load shedding as the objective function. When solving the multi-objective optimisation problem, four different strategies to fight against load shedding are identified, namely (1) optimal home appliance scheduling (HAS) under load shedding; (2) optimal HAS supported by solar panels; (3) optimal HAS supported by batteries, and (4) optimal HAS supported by the solar home system (SHS) with both solar panels and batteries. Among these strategies, appliance scheduling with an optimally sized 9.6 kWh battery and a 2.74 kWp panel array of five 550 Wp panels, eliminates the loss of power supply probability and reduces inconvenience by 92% when tested in the South African load shedding cases in 2023. More than 18.5 million households in South Africa are affected by load shedding. This results in a potential of 18.5 million unique load profiles. Creating unique optimal solar home systems for each household without evaluating similarities in their load profiles risks duplicating solar home system strategies. Clustering is an unsupervised machine learning method that can group households based on their inherent similarities, minimising intracluster similarity and maximising intercluster dissimilarity. K-means clustering is used in a case study of 781 South African households metered for a year, forming representative clusters of energy demand profiles to identify optimal strategies for multiple households that minimise the impact of load shedding. Three load clusters are identified as optimal, using the Davies-Bouldin criterion to minimise the ratio of within- and between- cluster distances. From K-means clustering, 43% of households are clustered in a low-energy demand load profile with an average daily energy consumption of 4.5 kWh, 42% in a medium energy and 15% in a high energy demand profile, with an average daily energy usage of 10.8 kWh and 22.1 kWh, respectively. Additionally, based on a 94.4% accurate, hourly, one-year-ahead prediction of load shedding outages using the KNN algorithm, we formulate each cluster’s inconvenience and loss of power supply due to load shedding as a multiobjective mixed-integer nonlinear optimisation problem. The results show that optimal scheduling of a low, medium and high energy consumption cluster with optimally sized 2.4 kWh and 0.39 kWp, 4.8 kWh and 0.78 kWp and 7.2 kWh and 1.17 kWp, battery and panel arrays, respectively, minimises the loss of power supply and substantially reduces the inconvenience of involuntary rescheduling by 86.8%, 71.1% and 86.4% for clusters 1, 2, and 3, respectively.
Monopulse radar analysis for cross-polarisation jamming
(University of Pretoria, 2023-11-23) Du Plessis, Warren Paul; u14255309@tuks.co.za; Mosoma, Khahliso
Cross-polarisation jamming is an electronic attack (EA) jammer that takes advantage of the design weakness in the radar antenna. The monopulse antenna with symmetric antennas in the four quadrants and feed symmetry has Condon lobes in its cross-polarisation signal component. The peaks of the Condon lobes are in the ±45◦ and ±135◦ diagonal planes. The cross-polarisation jammer receives the tracking signal from the tracking radar, interchanges the polarisation components, and re-transmits it to the tracking radar. If the jammer has a high JSR, the tracking radar will be forced to use one of the Condon lobes as the tracking lobe. Six monopulse antennas are analysed for cross-polarisation jamming. The jammer’s effects on the radar’s angle tracking accuracy are analysed as the JSR increases. How the antenna polarisation purity affects the effectiveness of cross-polarisation jamming is investigated. How the jammer's polarisation inaccuracy affects its ability to induce angular tracking error is investigated. The simulated results are validated using the measurements of the manufactured antenna. The cross-polarisation jammer can induce angular tracking error, but needs high polarisation accuracy. The mathematical models of the antenna cross-polarisation patterns are derived using three different approaches. These models are used to theoretically analyse cross-polarisation jamming and compare the results with the Feko simulations and measurement results. The axial symmetry in antennas causes Condon lobes in their cross-polarisation component. Two antennas with axial symmetry will have two Condon lobes, while four antennas located in four quadrants with axial symmetry will have four Condon lobes in each quadrant. One of the six antennas was used to validate the axial-symmetry effect on the Condon lobes. The analysis shows that the antenna radiating elements must be symmetrical, and the feed network must be symmetrical to result in symmetrical Condon lobes. The size of the Condon lobes is influenced differently in different antennas. The focal-length-to-diameter (F/D) size influences the Condon lobes in the parabolic reflector antennas. To investigate the effects of F/D size on the Condon lobes, a parabolic reflector antenna with different F/D sizes is designed and analysed. The analysis shows that increasing the F/D reduces the Condon lobes and increases the polarisation purity of the antenna.
The analysis of distributed resources on a load sharing reticulation network
(University of Pretoria, 2023-12-01) Naidoo, Raj; vandermerwe.ca@gmail.com; Van der Merwe, Carel Aron
Traditional reticulation network designs are outdated, based on single value static yearly maximum demands, and do not consider the dynamic nature of load-side DR installations. The increasing presence of privately driven downstream renewable and storage system integration (supported by increasing energy costs, maturing of storage, PV, and inverter technology systems, and an unreliable external network supply) requires time-based analysis to advance beneficial, and mitigate detrimental, shared network parameter changes. Fundamental integration network impacts must be re evaluated for grid integration acceptability and a modernised design approach, dependent on the capacity, capability, implementation, load-to-generation balancing, and power management of symbiotic integrated load-side DR (DG and/or ES) systems. These initial performance factors were analysed by conducting time based impact studies. Key concepts and approaches to the integration of PV DG, BESSs, and the combined DR system were identified and modelled at increasing levels of power penetration and energy arbitrage within the main distinctive reticulation network load profile forms in a visualised time based impact analysis. By identifying individual DR operational parameters and limits, an optimal approach to DR utilisation and power control is defined. Variables include load profiles, load diversity, demands, load factors, PV DG and BESS parameters, system power control, voltage profiles, utilisation factors, reactive power requirements, and fault levels. The maximum levels of DR penetration were defined (creating an upper penetration limit) following the evaluation of DR network parameter impacts and forms the foundation of the power flow control algorithm governing PV DG and BESS operation for equipment synergy and the optimisation of integration advantages. The proposed power control enforces permanent load side maximum demand reductions by up to 32%, with additional energy arbitrage operation enabled during peak period demands. This is achieved by limiting bi-directional power flow internally and maximising the combined DR system capability, utilisation, and operational synergy. Intermittent PV DG is selected for generation support, while more controllable BESS operation is chosen for targeted demand reduction applications in a give-and-take interface across all seasonal changes. The time based analysis, integration methodology, DR penetration limits, and the developed power flow control algorithm provide an expectation baseline for future DR network integration studies, guidance for service agreement inclusions, and the modernisation of traditional network designs without the necessity of an external network smart grid system. This will encourage the integration of higher rated privately driven renewable and energy storage systems to enhance grid advancement for both external and load-side DR integrated networks.
Delay- and disruption-tolerant routing algorithms to support human activity on mars
(University of Pretoria, 2024) Palunčić, Filip; Maharaj, Bodhaswar Tikanath Jugpershad; jason.kamps@tuks.co.za; Kamps, Jason Jack
Deep-space activity is expected to increase rapidly in the coming decades. Most notably, crewed missions to Mars will take place. With humans venturing light minutes away from Earth for the first time, communication becomes challenging. Humans have specific communication needs that become difficult to support in deep space where large propagation delays, high error rates, and intermittent connections are prevalent. Delay- and disruption-tolerant networking (DTN) and the Bundle Protocol provide a reliable communication solution in such challenging environments. The overall performance of DTN protocols is highly dependent on their routing algorithms. With Mars being humanity’s next target in our exploration of the Solar System, this study deals with finding and examining the most suitable routing protocols in the context of Earth-Mars communication. Realistic scenarios of space missions are constructed to enable the comparison of various DTN routing algorithms in simulation. Routing algorithm performance is analysed, and an enhancement to Contact Graph Routing (CGR) is proposed to address a deficiency of the algorithm, improving routing performance in networks featuring parallel channels.
Bi-directional DC-DC converter with energy management and protection capabilities For LVDC grids
(University of Pretoria, 2024-06) Gitau, Michael Njoroge; u15209840@tuks.co.za; Doma, Anesu Emmanuel
The work outlines a framework for enhancing the efficacy of current LVDC microgrid protection techniques. Currently, the two most significant challenges are the detection and interruption of fault currents. The primary aim of a protection strategy is to maintain the dependability of a power system by selectively isolating the components that are responsible for the fault occurrence. Consequently, it is imperative to interrupt the fault current before it reaches the components' maximum ratings. A proposal has been put forth for the implementation of a bidirectional converter to verify the functionality of a "converter cascaded with an Impedance Source Circuit Breaker (ISCB)" system. Contemporary investigations on DC microgrids suggest that the converter and impedance source breaker integration is functional; however, these two pivotal components have been analyzed separately, with the presumption of effortless integration. The combination is expected to exhibit fault current interruption capabilities and function as an energy hub. The analysis and design of a converter operating in Average Current Mode control (ACM) and an ISCB are conducted as separate entities. This work presents a proposed methodology for validating protection features. The obtained simulated results provide confirmation of the successful interruption of the circuit and ripple reduction on the DC branch input current.

Browse

Recent Submissions