Browsing by Author "Barendse, Paul"
Now showing 1 - 20 of 20
Results Per Page
Sort Options
- ItemOpen AccessAI-based hybrid optimisation of multi-megawatt scale permanent magnet synchronous generators for offshore wind energy capture(2019) Lilla, Abdurahman Daleel; Khan, Mohamed; Barendse, PaulThe finite nature of earth’s natural resources has become a post-industrial reality. Despite their alarming depletion, fossil fuels still dominated the global final energy landscape. Technological advances and rapid deployment of various renewable energy technologies have demonstrated their potential at reducing the worlds dependency on fossil fuels and their negative impacts. Presently, wind energy is the most cost-effective means of renewable energy conversion in the developed world and has currently has a price point that is in direct competition with fossil fuel. Coupled with the low price, the adoption of wind power has seen capacity increases in excess of 650% over the last ten years. Permanent Magnet Synchronous Generators (PMSGs) have become prominent in large wind energy system applications. The Radial Flux machine topology has become particularly attractive. In order to improve the competitiveness of large wind energy systems, the main focal point of current research is toward reducing the Levelised Cost of Energy (LCOE) of the systems. A proven method of reducing the LCOE of wind power generation is by upscaling RF-PMSGs to the multi mega-watt (MW) range. For the much wider adoption of wind power generation, the cost of energy (price/MWh) needs to be driven down further, by the development of more efficient and cost-effective ways to harvest the vast amounts of energy. The main objective of this dissertation is the drive-train selection, detailed design, sizing and optimisation of a 10.8 MW permanent magnet radial flux synchronous generator (RF-PMSG) to be used in the next generation of offshore wind farms. From an analytical viewpoint, the results suggested the use of a medium speed RF-PMSG utilizing a single-stage geared drivetrain, together with a MV voltage rating (3.3kV) for the 10.8 MW RF-PMSG designed in the thesis. Finally, this dissertation proposes a promising hybrid, analytical-numerical optimisation of a 10.8 MW RF-PMSG to be used for offshore Wind Energy Conversion. The hybrid optimisation utilises a two-stage optimisation strategy that incorporates both an analytical and a numerical (FEA) optimisation; using the DE algorithm and the Taguchi method respectively. Although the permanent magnet losses are neglected in the analytical loss calculations, they are included in the numerical FE portion of the hybrid optimisation. The initial stage (STAGE I) of the hybrid optimisation utilised the DE algorithm. The objective function was set to reduce the initial cost (!"#"$%&) of the RF-PMSG, by reducing the active material mass ('()$"*+) in the generator, i.e. NdFeB PM mass (',-), copper mass (').), and active steel in the stator lamination and rotor core ('/0$%&1$++&), while maintaining a pmsg efficiency (23456 ≥ 97%). The initial stage saw a reduction in initial cost by 25.5%, while maintaining an efficiency of 23456 = 97.8%. The final stage (STAGE II) of the hybrid optimisation utilising the Taguchi method, to make improvements on the performance of the machine, by optimising the Torque and back EMF characteristics while further reducing the NdFeB PM mass. The Magnet Fill Factor (APM), the Slot opening (bs0), the thickness of the permanent magnet poles (ℎ34) and the equivalent length of the air gap (?6) were used as optimisation variables. The final stage saw a decrease in cogging torque (@)06) by 53.4%, an increase in average torque (@%*) by 1.1%, a reduction in the total harmonic distortion of the back EMF (@AB) by 8.0%, a reduction in the required mass of the NdFeB permanent magnet material by 12.43%, while maintaining a torque ripple (@C"3) < 10%. The RF-PMSG characteristics optimised using the hybrid analytical-numerical optimisation were hypothesised to contribute in a reduction of the LCOE of offshore wind energy both in terms of Operational expenditure (OPEX) and Capital expenditure (CAPEX).
- ItemOpen AccessApplication of Sweep Frequency Response Analysis and Locked-Rotor Test for Stator Winding Inter-Turn and Broken Rotor Bar Fault Detection of Squirrel-Cage Induction Motors(2021) Boniface, Chigozie Jacob; Barendse, PaulThe field of electromagnetic energy conversion has been dominated by induction motors, where more than 80% are three-phase squirrel-cage induction motors (SCIM) [1], [2]. Despite their flexibility, robustness, and high efficiency, they are susceptible to various types of failures due to external or internal factors. Several studies have shown that 30%-40% [3], [4] and 5- 10% [5], [6] of the total failures of induction motor are due to the stator winding breakdown and rotor failures, respectively. Most induction motors have a long service life regardless of types and ratings and warrant minimum maintenance to ensure they work properly. Regular maintenance is recommended to check for signs of insulation damage or other parts that may derate the motor. This reduces motor repair costs, minimizes unscheduled downtime, and enhances industrial processes' reliability [7]. Stator winding faults include inter-turn fault, phase to phase fault, coil to coil fault, and phase to ground fault. The stator winding insulation usually begins with inter-turn faults comprising of a small number of winding turns which makes it more probable than other faults. Manufacturing defects and excessive start-stop cycles or frequent speed changes are mainly the causes of broken rotor bar and end-ring among different rotor fault types [8], but our focus in this study will be on broken rotor bar. Hence, early detection of these faults, both during the manufacturing and operation stages, is essential. Different techniques have been used for assessing the operating conditions of a rotating machine. One of the more popular techniques is motor current signature analysis (MCSA) [9]- [10], which analyses the spectral content of the stator current in determining the state of health. A popular technique for assessing the stator winding insulation is the DC Hipot test, which involves applying a dc voltage higher than that of the peak ac voltage [11]. Unfortunately, this procedure may lead to winding failure and replacement of the affected coil or winding. The sweep frequency response analysis (SFRA) approach, developed by Dick and Erven [12], allows for the detection of small changes in the windings due to the effect of short-circuits by analysing the winding impedance over its bandwidth. This technique is mostly used for transformers and very sensitive to inductance and capacitance between winding turns; therefore, any change in the winding geometrical structure can affect frequency response. Therefore, this project aims to model two devices (inductor to understand better the impedance behaviour of a coil and stator winding) analytically using the impedance transfer function approach. SFRA and locked-rotor test are carried out experimentally on the device under test (DUT) using the impedance transfer function measurement to detect the inter-turn and broken rotor bar fault, respectively. The SFRA technique relies on comparing two measurement results under normal and faults conditions over a wide frequency bandwidth while locked-rotor test relies on a higher excitation current for stator winding impedance extraction as a function of rotor angle to diagnose the SCIM rotor state properly. The results of this study show that SFRA and locked-rotor test are reliable techniques that enables the detection of damages to SCIM's winding and rotor at their incipient stage before a catastrophic failure occurs. It is observed that inter-turn fault mostly affects the inductive and capacitive region of the impedance response plot which in turns reduces the impedance peak and increases the resonant frequency. The appearance of the sinusoidal profile for the substituted impedance from the locked-rotor test indicates that SCIM rotor bar is faulty. Therefore, applying these two techniques for this study to diagnose SCIM will minimise unplanned outages, thereby reducing repair costs and improving the productivity of the system that uses it.
- ItemOpen AccessBroadband excitation of fuel cells for online condition monitoring using different switch-mode DC-DC mode topologies(2023) Mahlangu, Surprise; Barendse, PaulThere is a great demand for renewable energy sources, and these include solar and wind energy sources. However, a renewable energy source with a continuous energy supply is desired, but energy sources from wind and solar energy sources are intermittent, i.e., solar energy is only available during the day, and wind energy's availability is dependent on the season of the year and climate in the area. This, in essence, makes fuel cell systems desirable as a renewable energy source and storage in the form of green gases obtained from electrolysis or other processes using these intermittent green energy sources. Online condition monitoring of fuel cell systems without the need for additional hardware is desired in both stationery and transportation applications. Recent work has explored the use of online impedance spectroscopy for batteries using multi-sine signals through a single dc-dc converter. However, for fuel cells, the non-linearity of both the converter and the fuel cell poses a challenge to the online monitoring process. There is also a need for faster diagnostics due to the internal operating conditions of the fuel cell needing immediate control and regulation. This study demonstrates the use of the pseudo-random binary sequence (PRBS) to perform impedance spectroscopy in comparison to a single sinusoid injection. This is done to show the comparison between the two signals and to show the decrease in impedance estimation time brought about by PRBS in comparison to multi-sine signals. PRBS is a faster and easier technique to implement compared to the single-sine signals used in the condition monitoring of batteries. The tolerance of error brought by the implementation of impedance spectroscopy through PRBS and single-sine signals through a dc-dc converter is compared to the benchmarked theoretical results. This is demonstrated in simulation and experimentally. Results obtained from the Frequency Response Analyzer (FRA) are compared to the results obtained through a dc-dc converter using PRBS and Electrochemical Impedance Spectroscopy (EIS) as testing signals. Thereafter, this study demonstrates the feasibility and compares the use of the buck-boost and boost converters to perform impedance spectroscopy, and this is done by modelling and designing the converters for a linearized region of operation to accommodate different operating conditions of the fuel cell. This is achieved mainly using the double loop control strategy, which is rarely used in literature yet brings the benefits of controlling the power transfer and IS implementation. A practical buck-boost converter has discontinuous input current, more switching harmonics, and noise compared to the boost converter. As such, the contribution and influence of these factors are explored with regard to impedance estimation. The comparison of tolerance of error brought about by the implementation of impedance spectroscopy through PRBS and multi-sine signals using the boost and buck-boost converters are compared to the benchmarked theoretical results and Frequency Response Analyzer results.
- ItemOpen AccessComparative analysis of Polymer Electrolyte Membrane (PEM) fuel cells(2018) Balogun, Emmanuel O; Barendse, Paul; Chamier, JessicaPer-Fluoro-Sulphonic-Acid (PFSA) ionomers have been singled out as the preferable ionomers for making the Polymer Electrolyte Membrane Fuel Cells (PEMFC) membranes owing to their extensive intrinsic chemical stability and super sulfonic acid strength which is core to the PEMFC proton conductivity. This thesis presents a deeper analysis into these PFSA ionomer membrane electrode assemblies (MEA), presenting an electrochemical-analytical comparative analysis of the two basic types, which are the Long-Side-Chain (LSC) Nafion® and the ShortSide-Chain (SSC) Aquivion® ionomer MEA with emphasis on performance and durability which are currently not well understood. In particular, electrochemical circuit models and semiempirical models were employed to enable distinguishable comparative analysis. Also, in this thesis, we present a further probe into the effect of ionomer ink making processes, critically investigating the effect of the High Share Dispersion (HSD) process on both the Nafion® and Aquivion® ionomer membrane electrode assembly (MEA). The findings in this research provides a valuable insight into the performance and durability of PFSA ionomer membrane under various application criteria. The effect of operating parameters and accelerated stress testing (AST) on the PFSA ionomers was determined using electrochemical impedance spectroscopy (EIS) and electronic circuit model (ECM) analysis. The result of this study, shows that the ionomer ink making process for Nafion® and Aquivion® MEAs are not transferrable. Analysis of the PEMFC performance upon application of the high shear dispersion (HSD) process showed that Nafion® MEA had a 10.47% increase in voltage while the Aquivion® MEA had a 2.53% decrease in voltage at current density of 1.14A/cm2 . Also, upon accelerated stress testing, the Nafion® showed a 10.49% increase in its voltage while the Aquivion® on the other hand had a 7.16% decrease in voltage at 0.66A/cm2 . Thus indicating the HSD process enhances the performance of the Nafion® MEA and inhibits the performance of the Aquivion® MEA.
- ItemOpen AccessCondition monitoring and fault detection of inverter-fed rotating machinery(2017) Ipurale, Andrew; Barendse, PaulCondition monitoring of rotating machinery is crucial in industry. It can prevent long term outages that can prove costly, prevent injury to machine operators, and lower product quality. Induction motors, often described as the workhorse of industry, are popular in industry because of their robustness, efficiency and the need for low maintenance. They are, however, prone to faults when used improperly or under strenuous conditions. Gearboxes are also an important component in industry, used to transmit motion and force by means of successively engaging teeth. They too are prone to damage and can disrupt industrial processes if failure is unplanned for. Reciprocating compressors are widely used in the petroleum and the petrochemical industry. Their complex structure, and operation under poor conditions makes them prone to faults, making condition monitoring necessary to prevent accidents, and for maintenance decision-making and cost minimization. Various techniques have been extensively investigated and found to be reliable tools for the identification of faults in these machines. This thesis, however, sets out to establish a single non-invasive tool that can be used to identify the faults on all these machines. Literature on condition monitoring of induction motors, gearboxes, and reciprocating compressors is extensively reviewed. The time, frequency, and time-frequency domain techniques that are used in this thesis are also discussed. Statistical indicators were used in the time domain, the Fourier Transform in the frequency domain, and Wavelet Transforms in the time-frequency domain. Vibration and current, which are two of the most popular parameters for fault detection, were considered. The test rig equipment that is used to carry to the experiments, which comprised a modified Machine Fault Simulator -Magnum (MFS-MG), is presented and discussed. The fault detection strategies rely on the presence of a fault signature. The test rig that was used allows for the simulation of individual or multiple concurrent faults to the test machinery. The experiments were carried out under steady-state and transient conditions with the faults in the machines isolated, and then with multiple faults implemented concurrently. The results of the fault detection strategies are analysed, and conclusions are drawn based on the performances of these tools in the detection of the faults in the machinery.
- ItemOpen AccessCondition Monitoring of VSD-fed Induction Motors(2023) Ralikalakala, Lebohang; Barendse, PaulInduction motor drives are presently the most fully fledged technology amidst the various motor drives not utilising commutators [1]. They consist of an induction motor (IM) fed from a variable frequency AC inverter. They are mostly used in applications which need good dynamic and steady state performance over a large range of speeds [2]. The induction motor is often referred to as the workhorse of the industry due to its predominance in the industry – constituting around 90% of all the motors in the industry [3], [4]. Its admired features include relatively low cost, high efficiency, robustness, and ease of maintenance [5], [6]. These motors are of two types, the wound-rotor and the squirrel cage [1], [5], with the three-phase squirrel cage induction motor being the most common, constituting over 80% [7], [8]. On the other hand, variable frequency AC inverters are also of two types, the voltage source inverter (VSI) and the current source inverter (CSI). VSI is used for most induction motor drive applications due to its simpler control and efficient operation in addition to its low weight, cost, and volume. Despite their high reliability and robustness, induction motors are still prone to faults [9] due several factors such as being subjected to diverse, continuous, and harsh conditions [10], being subjected to frequent start-stop cycles leading to wear and cracking of machine elements, or being subjected to continued motor overload for long periods required by some industrial processes, resulting in increased thermal stress [11]. Studies have shown that one category of faults that an induction motor may suffer from is the rotor faults [12]. Of these, the broken rotor bar fault in squirrel cage induction motors is the most common [13], constituting 5-10% of all induction motor faults [14]. Besides the induction motor, the state of health of the driving inverter is also important to the overall reliability, integrity, and availability of the induction motor drive system. The power electronic converter (inverter) constitutes about 82.5% of all the faults that occur in the inverter-fed motor drives [15] and of this, 38% occur due to failures of the power electronic switching transistors [16], [17]. These faults include short circuit faults, open circuit faults and intermittent gate-misfiring faults [18]. Open switch faults and short circuit faults are very common in inverters. Consequently, the continuous monitoring of the induction motor drive components (e.g., the induction motor or inverter) is crucial for an early and timely detection of faults to avoid the propagation of the faults and the total breakage of the drive or part of the drive. This will in turn reduce the risk of reduced output, increased emergency maintenance costs and out-of-service problems [10]. The process of continuously monitoring the condition or state of health of a system is called condition monitoring. It is very indispensable in electric drives. One condition monitoring technique commonly used for the detection of broken rotor bar fault is the motor current signature analysis (MCSA), which is focused on the application of the Fourier transform to machine's stator current under steady state operating conditions [7]. The MCSA assumes that the motor is sufficiently loaded and operating under steady state conditions [19]. Despite its robustness, this fault diagnostic tool suffers nonnegligible drawbacks, especially when applied in the industry, where there are many real-time factors such as speed variation or load changes which affect the operation of the induction motor. Similarly, diagnostic variables technique (DVT) is a condition monitoring technique commonly used in the detection and diagnosis of inverter open switch faults. Contrary to MCSA, it uses diagnostic variables and current average values to detect and localise open switch faults [17], [20], [21]. This method can detect both single open switch and double open switch faults [21]. The objective of this study is to design and develop an open-loop volts-per-hertz controlled induction motor drive for study of faults across the various parts of the drive. In particular, the study focuses on the broken rotor bar fault on the motor-end of the drive and the inverter open switch faults on the converter side of the drive. It commences with investigation of the impacts of the various inverter open switch faults (single switch open circuit, single phasing, and double switch open circuit faults) on the performance of the induction motor, before performing fault detection and diagnosis of these faults to validate the diagnostic variables technique. Then it experimentally establishes some of the shortcomings of MCSA in the detection of the broken rotor bar fault of a Class B inverter-fed 250 W, 190 V, 50 Hz and 2-pole, star-connected threephase squirrel cage induction motor. Finally, it proposes a complementary broken rotor bar fault detection technique, for inverter-fed induction motor, that is premised on the variation of impedance with rotor position at standstill to supplement MCSA. The results for investigation of the impact of open switch faults show that both simulation and experimentation produced results with high concordance. They show that double switch open switch fault results in the largest increase in distortion and in DC offset voltage and current than the other two faults, while relatively less changes on these parameters are observed for single phasing fault. A similar trend is observed in speed and torque pulsations of the motor under the respective faults. Similarly, the method of diagnostic variables has proven valid for both simulation and experimental results and is able to detect and localise inverter open switch fault for single and double switch fault. Furthermore, the MCSA is shown to be incapable of detecting broken rotor bar fault under no-load conditions because the sideband harmonics overlap the fundamental component. Likewise, inverter switching harmonics mask and interfere with higher ordered sideband harmonics in the case of inverter-fed motors, thus decreasing the confidence level in this technique. Finally, the proposed standstill impedance variation test shows large change, to the value of 9.59% of average impedance, in impedance as function of rotor position, and this is used to characterize the broken rotor bar fault. The inverter harmonics and switching noise do not impact on this variation in impedance. Further exploration of the standstill impedance variation approach in inverter-fed induction motors showed that it is possible to detect anomalies within the motor due to broken rotor bar fault by utilizing the geometrical orientation of the three-phase stator windings instead of changing the rotor position. That is, by manipulating the inverter switching such that the pulsating flux is aligned with each of the three magnetic axis of the three-phase stator windings, three impedances (one along each magnetic axis of the windings) are obtainable. The average of the absolute differences of these three impedances is large enough to characterize the broken rotor bar fault except when the rotor is positioned at 30°, 70° and 150° relative to the chosen reference position. These positions are referred to as critical rotor positions since the difference in impedance at these positions is very small and can lead to the possible misdiagnosis of the fault (false negative indication of the fault). It is noteworthy that this extension of the standstill impedance variation test is only applicable to 2-pole machines only since their mechanical degrees is equal to their electrical degrees.
- ItemOpen AccessDesign and Implementation of A Three-Level Boost converter for Battery Impedance Spectroscopy(2020) Mosunmola, Faloye Omolola; Barendse, PaulLithium-ion batteries are the most are widely used as electrical storage device in various applications such as portable electronics, electric vehicles, Photovoltaic application, telecommunication etc due to the characteristics of the batterie such as high-power density, long cycling and high-power efficiency. Extensive condition monitoring of the battery should be implemented due to the usage of the battery so that there will be an increase in all the overall performance and expectancy. This research is focused on implementing an online condition monitoring on the Li-ion battery using a signal injection through a power converter. The implemented technique in this research is known as the Electrochemical Impedance Spectroscopy (EIS). The EIS is a widely known technique used in determining the internal impedance of a battery cell. The estimated impedance can be used to determine the state of charge (Soc) and State of health (SoH) of a battery. The EIS is used to characterize the electrochemical behaviour thereby monitoring the change in the impedance of the cell of the battery. The EIS technique is accomplished by sinusoidally injecting current at different frequencies and measuring the voltage response. A standard Frequency Response Analyser (FRA) is used as an offline test while the battery is disconnected from the Load. The limitation of this standard FRA analyser is that it is bulky and Expensive. Attempts have been made to migrate the techniques to online operations, each having their own challenges. For an online Implementation, the interfacing power converter is used for Signal injection to measure the impedance of the battery. This work explores the low current ripple advantage of a threelevel boost converter to implement EIS on lithium ion battery.
- ItemOpen AccessDetection of faults in a scaled down doubly-fed induction generator using advanced signal processing techniques.(2023) Hamatwi, Ester; Barendse, PaulThe study ventures into the development of a micro-based doubly fed induction generator (DFIG) test rig for fault studies. The 5kW wound rotor induction machine (WRIM) that was used in the test rig was based on a scaled-down version of a 2.5MW doubly fed induction generator (DFIG). The micromachine has been customized to make provision for implementing stator inter-turn short-circuit faults (ITSCF), rotor ITSCF and static eccentricity (SE) faults in the laboratory environment. The micromachine has been assessed under the healthy and faulty states, both before and after incorporating a converter into the rotor circuit of the machine. In each scenario, the fault signatures have been characterised by analyzing the stator current, rotor current, and the DFIG controller signals using the motor current signature analysis (MCSA) and discrete wavelet transform (DWT) analysis techniques to detect the dominant frequency components which are indicative of these faults. The purpose of the study is to evaluate and identify the most suitable combination of signals and techniques for the detection of each fault under steady-state and transient operating conditions. The analyses of the results presented in this study have indicated that characterizing the fault indicators independent of the converter system ensured clarity in the fault diagnosis process and enabled the development of a systematic fault diagnosis approach that can be applied to a controlled DFIG. It has been demonstrated that the occurrence of the ITSCFs and the SE fault in the micro-WRIM intensifies specific frequency components in the spectral plots of the stator current, rotor current, and the DFIG controller signals, which may then serve as the dominant fault indicators. These dominant components may be used as fault markers for classification and have been used for pattern recognition under the transient condition. In this case, the DWT and spectrogram plots effectively illustrated characteristic patterns of the dominant fault indicators, which were observed to evolve uniquely and more distinguishable in the rotor current signal compared to the stator current signal, before incorporating the converter in the rotor circuit. Therefore, by observing the trends portrayed in the decomposition bands and the spectrogram plots, it is deemed a reliable method of diagnosing and possibly quantifying the intensity of the faults in the machine. Once the power electronic converter was incorporated into the rotor circuit, the DFIG controller signals have been observed to be best suited for diagnosing faults in the micro-DFIG under the steady-state operating condition, as opposed to using the terminal stator or rotor current signals. The study also assessed the impact of undervoltage conditions at the point of common coupling (PCC) on the behaviour of the micro-DFIG. In this investigation, a significant rise in the faulted currents was observed for the undervoltage condition in comparison to the faulty cases under the rated grid voltage conditions. In this regard, it could be detrimental to the operation of the micro-DFIG, particularly the faulted phase windings, and the power electronic converter, should the currents exceed the rated values for extended periods.
- ItemOpen AccessDevelopment of a converter for grid-tied and isolated operation of an interior permanent magnet synchronous generator, coupled to a twin-shaft gas turbine(2018) Molaoa, Molaoa; Khan, M A; Barendse, PaulSouth Africa’s overreliance on coal fired power generation has led to the government’s commitment to diversifying the country’s energy mix. Gas turbine generators are poised to play a larger role in South Africa’s energy mix, due to the country’s abundance in natural gas reserves. Therefore, there is a need to developed gas turbine emulation systems to investigate how this transition is to be implemented and to discover new efficient ways to generate power through gas turbines. This thesis presents the development of a twin-shaft gas turbine emulator. A DC-machine that accepts both torque and speed references is used to emulate the behaviour of the gas turbine according to a modified Rowen gas turbine model. The emulator is coupled to a 1.5kW interior permanent magnet synchronous generator (IPM). The power density of a DC-machine is significantly lower than that of a gas turbine of the same rating. Thus, the DC-machine is rated at double the rating of the IPM to overcome the high inertia it has when compared to a gas turbine of the same rating. This means that the DC-machine can produce large toques to successfully emulated the dynamic behaviour of the gas turbine. A maximum error 2.5% in the emulation of the gas turbine’s speed is reported. A two-level active converter is used to compare control strategies for an IPM. Ninety-degree torque angle (NTA) control, maximum torque per ampere (MTPA) control and unity power factor (UPF) control are compared for performance. The UPF and MTPA control result in the lowest and second lowest DC-link utilisation respectively when compared to NTA control. This is due to a negative d-axis current component as opposed to a zero d-axis current component in the case of NTA control. It is also concluded that to achieve a high power factor and torque development, a negative d-axis current component is required. UPF and MTPA control perform well in both categories, with UPF control and MTPA control resulting in the highest power factor and developed torque respectively. A fourth control strategy that maximises the efficiency of the IPM is developed experimentally. The maximum efficiency (ME) control strategy minimises mechanical, core, windage and conduction losses. It also results in near unity power factor and near maximum developed torque. A nonconventional control structure that involves control of the DC-link from the generatorside converter is presented. This frees the outer-loop control of load-side converter to regulate voltage across the load when the system is supplying power to an isolated load. This control structure also allows the grid-side converter to employ reactive power compensation, without having to regulate the DC-link voltage at the same time. In doing so, large grid currents are avoided. A recursive least squares (RLS) algorithm is used to separate negative and positive sequence current components during grid voltage unbalance. A method to minimise the presence of negative sequence components in the load current is presented and implemented successfully in an experiment.
- ItemOpen AccessDevelopment of a DFIG wind energy system for grid integration studies(2013) D'Oliveira Pio, Manuela Maria Amilha; Khan, Azeem; Barendse, PaulDoubly Fed Induction Generators (DFIGs) are common for use in wind power generation. This can be attributed to factors such as the variable speed operation of the rotor, which can be adjusted through control systems with varying wind speed, ensuring more efficient power conversion.The stator windings in a DFIG are directly connected to the grid while the rotor windings are connected to the grid via back to back converters. This allows for independent active and reactive power control through the rotor side convertor and lower rating of converter which leads to lower cost of the system. The system below is a schematic of experimental set-up of the doubly fed induction generator connected to the grid.
- ItemOpen AccessDevelopment of an online broadband impedance spectroscopy extraction system of PV modules through switch-mode converters(2023) Shelembe, Lindani; Barendse, PaulSolar PV systems have traditionally relied on terminal DC current and voltage characteristics for online diagnostics. This approach monitors the current and voltage levels of a module, and a change in these levels indicates a system anomaly. However, this information only alerts to a decrease in the current or voltage levels, but not the possible cause(s) for the anomaly. To enable more detailed online diagnostics, linear impedance spectroscopy, a frequency-based characterization technique implemented through a DC-DC switch-mode converter, has been shown to provide fast and reliable impedance information that can be utilized to diagnose the condition of batteries and fuel cells in real-time. However, the implementation of this technique in a PV system presents challenges that have not been identified or addressed in current research. To develop such a system for a solar PV module, an excitation source with wide loop bandwidth capabilities is required to account for the broader impedance bandwidth of a PV module compared to batteries or fuel cells. The excitation signals must be designed such that they can minimize the nonlinearities observed at lower and higher frequencies, which result from induced excitation produced by the nonlinear DC-DC switch-mode converter.
- ItemOpen AccessDevelopment of new methods for nonintrusive induction motor energy efficiency estimation(2020) Aminu, Muhammad; Barendse, Paul; Khan, AzeemInduction motors (IMs) are the most widely used motors in industries. They constitute about 70% of the total motors used in industries and are the largest energy consumers in industrial applications. As a result of the increasing need for energy savings and demand-side management, the development of methods for accurate energy efficiency estimation has become a crucial area of research. While several methods have been proposed for induction motor efficiency determination, majority of the methods cannot be easily applied in the field owing to the intrusive nature of the test procedures involved. This PhD work presents some novel methods for nonintrusive efficiency estimation of induction motors operating on-site using limited motor terminal measurements and nameplate data. The first method is developed for induction motors operating on sinusoidal supply source (line-fed). The method uses a modified inverse Г-model equivalent circuit with series core loss arrangement to mitigate the inherent problems of higher computational burden and parameter redundancy associated with the conventional equivalent circuit method. Furthermore, a new method is presented for estimating the friction and windage loss using the airgap torque and motor nameplate data. The proposed Nonintrusive Field Efficiency Estimation (NFEE) technique was validated experimentally on four different induction motors for both balanced and unbalanced voltage supply conditions. The results demonstrate the accuracy of the proposed NFEE method and confirm its advantage over the conventional equivalent circuit method. In addition to the problem of unbalanced voltage supply, the presence of harmonics significantly affects the operation of induction motors. The second novel approach for estimating efficiency proposed in this PhD work extends the NFEE method to cover for non-sinusoidal supply condition. The method considers the variation of core loss, rotor bar resistance and leakage inductance due to time harmonics and skin effects. Finally, the efficiency estimations are compared to the IEC/TS 60034-2-3 in the case of a balanced non-sinusoidal supply condition. This allows not only the efficiency comparison but also the loss segregation analysis on the various components of the motor losses. In the case of an unbalanced supply, the efficiency results are compared to measured values obtained based on the direct input-output method. In both the first and second methods, a robust Chicken Swarm Optimization (CSO) algorithm has been used for the first time in conjunction with a simplified inverse Г-model EC to correctly determine the induction motor parameters and hence its losses and efficiency while inservice. As Variable Frequency Drives (VFDs) continue to dominate industrial process control, there is a need for stakeholders to quantify the converter-fed motor losses over a wide range of operating frequency and loading conditions. Although there is an increase in legislative activities, particularly in Europe, towards the classification and improvement of energy efficiency in electric drive systems, the handful of available standards for quantifying the harmonic losses are still undergoing validation. One of such standards is the IEC/TS 60034-2-3, which has been lauded as a step in the right direction. However, its limitation to rated motor frequency has been identified as one of its main weaknesses. Therefore, the third method proposed in this research demonstrates how the IEC/TS 60034-2-3 loss segregation methodology at nominal frequency can be extended over the constant-torque region of an induction motor (IM). The methodology has been validated by testing two motors using a 2-level voltage source inverter (VSI) in an open-loop V/F control mode. The results provide good feedback to the relevant IEC standards committee as well as guidance to stakeholders.
- ItemOpen AccessImpedance spectroscopy techniques for condition monitoring of polymer electrolyte membrane fuel cells(2018) Aroge, Fabusuyi Akindele; Barendse, PaulEnergy continues to remain the spine of all human development. As we continue to make advances in various levels, the need for energy in quantity, and even more recently, quality, continues to increase. The fuel cell presents itself as a promising prospect to solve one of mankind’s current challenge - clean energy. The fuel cell is essentially an electrochemical conversion system which takes in fuel supply to produce electricity. Some key features make the fuel cell attractive as a power source. Firstly, its efficiency in practical applications is approximately 50% compared to the typical efficiency of 40% for a typical internal combustion engine [1]. Secondly, unlike the systems such as the internal combustion engine that typically releases carbon-monoxide which is a major greenhouse gas, the typical fuel cell system, produces just water and heat, alongside the useful electrical energy. These characteristics make it attractive as a clean energy supply capable of replacing the fossil-based supplies that are currently the mainstay. Unfortunately, the fuel cell is far cry from an ideal system. Despite significant advantages of the fuel cell as a power supply, various challenges still exist which have hindered its widespread acceptance and deployment. The fuel cell at its core is a highly multi-physics system and its operational intricacies makes it highly prone to a series of fault conditions. This begs the question of durability - an important requirement of a viable power source. Another challenge is the fact that humanity currently struggles with an efficient method of producing hydrogen which is the fuel of choice for the fuel cell. Given the promises of the fuel cell however, research efforts continue to increase to further improve its viability as an energy source competitive enough to meet mankind’s need of clean energy. This work presents results bordering on efficient diagnostic approaches for the fuel cell, aimed at improving the durability of the fuel cell. Particularly, two techniques targeted at improving the popular Electrochemical Impedance Spectroscopy (EIS) are presented. Conventional EIS takes significant amount of time, rendering it unsuitable for real-time diagnostics. Multi-frequency perturbation signals have been proposed to address this challenge. These however introduces concerns surrounding the accuracy of the resulting impedance measurement. Part of this work addresses some of the challenges with the fuel cell multi-sine impedance spectroscopy, such as measurement accuracy, by defining an optimized signal synthesis formulation. The proposed approach is validated in simulation and compared to the popular exponential frequency distribution approach using the appropriately defined error metric. Secondly, the chirp – as a frequency rich signal, is investigated as an alternative perturbation signal. Consequently, the use of the wavelet transform as an analysis tool of choice is presented. The characteristic nature of the chirp signal makes a broadband frequency sweep over time possible, hence enabling a faster impedance estimation. The resulting decomposition is harnessed for impedance calculation. The approach is tested in simulation and results for equivalent circuits are presented. It is shown that the resulting impedance spectrum well approximates the theoretical values. To further validate both techniques in practice, a low-cost active load is designed and built. The active load enables the injection of an arbitrary signal using the load modulation technique. The device is tested and benchmarked against commercial frequency response analyzer (FRA) using the conventional single sine EIS technique. Both approaches developed – the improved multi-sine scheme and the chirp signal perturbation are demonstrated with the aid of the active load on a single cell fuel cell station. Outcomes of the experiment show significant accuracy from the two techniques in comparison with results obtained from the FRA equipment which implements the single sine technique. In addition, the two schemes enabled impedance results to be taken in a few seconds, compared to conventional single sine EIS which takes several minutes. Impedance measurements are also carried out in the presence of two prominent faulty conditions – flooding and drying, using the developed techniques. This demonstrates the capability of the proposed system to perform real-time diagnostics of the PEMFC using impedance information.
- ItemOpen AccessMultilevel inverters for renewable energy systems(2018) Chiwaridzo, Pride; Barendse, PaulVoltage source inverters have become widely used in the last decade primarily due to the fact that the dangers and limitations of relying on fossil fuel based power generation have been seen and the long term effects felt especially with regards to climate change. Policies and targets have been implemented such as from the United Nations climate change conference (COPxx) concerning human activities that contribute to global warming from individual countries. The most effective way of reducing these greenhouse gases is to turn to renewable energy sources such as the solar, wind etc instead of coal. Converters play the crucial role of converting the renewable source dc power to ac single phase or multiphase. The advancement in research in renewable energy sources and energy storage has made it possible to do things more efficiently than ever before. Regular or 2 level inverters are adequate for low power low voltage applications but have drawbacks when being used in high power high voltage applications as switching components have to be rated upwards and also switch between very high potential differences. To lessen the constraints on the switching components and to reduce the filtering requirements, multilevel inverters (MLI's) are preferred over two level voltage source inverters (VSI's). This thesis discusses the implementation of various types of MLI's and compares four different pulse width modulation (pwm) techniques that are often used in MLI under consideration: three, five, seven and nine level inverters. Harmonic content of the output voltage is recorded across a range of modulation indices for each of the three popular topologies in literature. Output from the inverter is filtered using an L only and an LC filter whose design techniques are presented. A generalized prediction algorithm using machine learning techniques to give the value of the expected THD as the modulation index is varied for a specific topology and PWM switching method is proposed in this study. Simulation and experimental results are produced in five level form to verify and validate the proposed algorithm.
- ItemOpen AccessNon-intrusive efficiency estimation of inverter-fed induction machines(2020) Chirindo, Mathews; Khan, Azeem; Barendse, PaulMotorised loads using induction machines use approximately 60% of the electricity globally. Most of these systems use three-phase induction motors due to their robustness and lower cost. They are often installed in continuously operating industrial plants/applications that require no operational interruptions. Whilst most of these induction machines are supplied from ideally sinusoidal supplies, applications are emerging where induction machines are fed from non-sinusoidal supplies. In particular, pulse width modulated inverters realize efficient control of induction machines in many automated industrial applications. From an energy management perspective, it is vital to continually assess the efficiency of induction machines in order to initiate replacement or economic repair. It is therefore of paramount importance that reliable and non-intrusive techniques for efficiency estimation of induction machines be investigated, that consider sinusoidal and non-sinusoidal supplies. This work proposes a non-intrusive efficiency estimation technique for inverter–fed induction motors that is based on harmonic regression analysis, harmonic equivalent circuit parameter estimation and harmonic loss analysis using limited measured data. Firstly, considerations for inverter-fed induction motor equivalent circuit modelling and parameter estimation techniques suitable for non-intrusive efficiency estimation are presented and the selection of one equivalent circuit for analysis is justified. Measured data is obtained from two different induction motors on a flexible 110kW test rig that utilises an HBM Gen 7i data acquisition system. By measuring voltage, current and input power at the supply terminals of the inverter-fed motor, the fundamental equivalent circuit parameters are estimated using population based incremental learning algorithm and compared with those obtained from the IEC 60034-2-1 Standard. The harmonic parameters are estimated using the bacterial foraging algorithm basing on the input impedance of the motor at each harmonic order. A finite harmonic loss analysis is carried out on the tested induction motors. The proposed techniques and harmonic loss analysis provide accurate efficiency estimates of within 1.5% error when compared to the direct method. Lastly, a related non-intrusive efficiency estimation technique is proposed that caters for a holistic loss contribution by all harmonics. The efficiency results from the proposed techniques are compared to those obtained from the IEC-TS 60034-2-3 Technical Specification and a direct method. The estimated efficiencies are comparable to those measured by the Technical Specification and a direct method within 2% error when tested on 37kW and 45kW PWM inverter-fed motors across the loading range. Furthermore, this work conducts a comprehensive non-intrusive rotor speed estimation comparative analysis in order to recommend the best technique(s), in terms of intrusiveness, accuracy and computational overhead. Errors of less than 1% have been reported in literature and experimental verification when using vibration analysis, Motor Current Signature Analysis (MCSA), Rotor Slot Harmonic (RSH) and Rotor Eccentricity Harmonic (REH) analysis techniques in inverter-fed IMs.
- ItemOpen AccessOnline condition monitoring of lithium ion batteries by performing impedance spectroscopy using a DC-DC converter(2018) Moore, Sean; Barendse, PaulLi-Ion batteries are currently being used extensively in a variety of applications such as portable electronics, electric vehicles and grid storage applications, due to the high demand for high power and high energy density storage batteries. However, the usage of Li-Ion batteries requires extensive condition monitoring to increase overall performance and life expectancy. This research is focused on implementing rapid online condition monitoring techniques, using signal injection via the associated power converter in the battery management system. The technique implemented in this work is known as Electrochemical Impedance Spectroscopy (EIS). EIS is a well-known technique, that has been used to characterize an electrochemical cell’s behaviour and state by monitoring changes in the cell's impedance. This is accomplished by injecting currents at different frequencies into the battery and measuring the voltage response. This technique is conventionally implemented, using standard Frequency Response Analysers (FRA), while the battery is disconnected from the load (offline) due to the long procedural times involved. This has limited the use of EIS to laboratory testing. In recent years, there has been literature regarding incorporating EIS testing into the battery system, which is done in mainly 2 ways, by including a linear amplifier into the system to inject the current perturbations directly, or by using the existing circuitry in the BMS (typically the DC-DC converter) to inject the current perturbations using various control techniques. Although, these strategies have been applied in literature for online systems, they are still riddled with a lengthy EIS measurement time issue. This work seeks to significantly reduce the associated testing time with the use broadband signals to implement Impedance Spectroscopy for online systems via the associated BMS converter. Broandband Impedance Spectroscopy (BIS) is implemented by injecting a signal with multiple frequencies concurrently as this allows for quicker measurement. The main contribution of this work is the real-time implementation of a multi-sine broadband excitation via a bi-directional converter that can be used in a varying system. The results obtained were compared to results from an industry standard FRA and showed to produce Nyquist plots with a reasonable error.
- ItemOpen AccessOnline condition monitoring of lithium-ion and lead acid batteries for renewable energy applications(2018) Alao, Olakunle Oluwatosin; Barendse, PaulElectrochemical Impedance Spectroscopy (EIS) has been largely employed for the study of reaction kinetics and condition monitoring of batteries during different operational conditions, such as: Temperature, State of Charge (SoC) and State of Health (SoH) etc. The EIS plot translates to the impedance profile of a battery and is fitted to an Equivalent Electric Circuit (EEC) that model the physicochemical processes occurring in the batteries. To precisely monitor the condition of the batteries, Kramers-Kronig relation: linearity, stability and causality as well as the appropriate perturbation amplitude applied during EIS should be adhered to. Regardless of the accuracy of EIS, its lengthy acquisition time makes it impracticable for online measurement. Different broadband signals have been proposed in literature to shorten EIS measurement time, with different researchers favouring one technique over the other. Nonetheless, broadband signals applied to characterize a battery must be reasonably accurate, with little effect on the systems instrumentation. The major objective of this study is to explore the differences in the internal chemistries of the lithium-ion and lead acid batteries and to reduce the time associated with their condition monitoring using EIS. In this regard, this study firstly queries the methodology for EIS experiments, by investigating the optimum perturbation amplitude for EIS measurement on both the lead acid and lithium-ion batteries. Secondly, this study utilizes electrochemical equations to predict the dynamics and operational conditions associated with batteries. It also investigates the effect of different operational conditions on the lead acid and lithium-ion batteries after EEC parameters have been extracted from EIS measurements. Furthermore, different broadband excitation techniques for rapid diagnostics are explored. An online condition monitoring system is implemented through the utilization of a DC-DC converter that is used to interface the battery with the load. The online system is applied alongside the different broadband signals. The deviation in the broadband impedance spectroscopy result is compared against the Frequency Response Analyzer (FRA) to determine the most suitable technique for battery state estimation. Based on the comparisons, the adoption of a novel technique – Chirp Broadband Signal Excitation (CBSE) is proposed for online condition monitoring of batteries, as it has the advantage of being faster and precise at the most important frequency decade of the impedance spectrum of batteries.
- ItemOpen AccessReclosing transients in standard and premium efficiency induction machines in the presence of voltage unbalance(2018) Gabaraane, Tumelo; Barendse, PaulThe impact of restarting an induction machine, while coasting, was investigated in this study. When restarting an induction machine, high peaking current and torque transients appear. As a result, possible damage to the rotor shaft could occur together with a destabilising of the power system which could lead to fluctuating supply voltage levels. This dissertation investigated the restarting of two different efficiency class induction machines and analyses how a shift in the phase between the supply voltage and the residual voltage affects the restarting transients. Previous studies have been carried out on standard efficiency induction machines, however this study aimed to compare the impact of restarting on the standard and premium plus efficiency induction machines. The design differences between standard and energy efficient induction machines play a significant role in how these transients behave. The energy efficient machines were found to be made of higher grade lamination steel and larger sized conductors in order to reduce the stator and rotor copper losses. Using material with a higher permeability, as is the case with energy efficient machines, ensures that the machines are able to retain less residual flux at the point of disconnect than standard motors. It is for this reason that the energy efficient machines have a larger residual voltage at the point of disconnect. However, the speed of the decay of the residual flux of the standard efficiency machine is higher than the energy efficient machine and, as a result, the energy efficient machine resulted with a higher residual voltage at the 180° phase shift. This leads to a larger potential difference for the energy efficient machines than the standard efficiency machines. However, the stator impedance of the energy efficient machine was established as a dominating factor in the outcome of the tests results. Energy efficient machines have a significantly lower impedance than standard efficiency machines and due to this, they are more susceptible to the detrimental effects of out-of-phase restarting. It was important to establish a critical point when the current and torque transients are at a maximum as the motors speed is winding down. It is necessary to minimise the effects of restarting at this critical point as the resulting developed torque can be mechanically detrimental to the shaft. The residual voltage needs to be minimised dramatically to a point where the out-of-phase reclosing will not give rise to significantly high torque transients. Additional to this study was the observation of the current and torque transients in the presence of voltage unbalance. Voltage unbalance presents a power quality issue which can potentially have a negative effect on the efficiency of an induction machine. Both machines were tested for 0%, 1% and 2% voltage unbalance with a focus on start-up and 180° restart. In addition, voltage unbalance was introduced at the rated voltage of 400 V with 10% variations above and below this rating for over and undervoltage respectively. It was seen that the energy efficient machines were more vulnerable to unbalance effects due to a larger negative sequence air gap torque being developed under transient conditions.
- ItemOpen AccessSingle-Phase Bi-directional Ćuk Inverter for Battery Applications(2019) Shelembe, Lindani; Barendse, PaulBidirectional inverters are widely applied in photovoltaic and wind systems that require battery power backup. They are advantageous over unidirectional inverters because of their ability to convert DC power into AC power and then AC power back into DC power to recharge for storage purposes. Generally, bidirectional inverters are designed to have multiple power stages and/or make use of transformers for isolation and voltage/current gain. This usually increases the cost of production and oftentimes reduces the efficiency of the system. At the same time, attempts at eliminating usage of transformers and reduction in the number of power stages limits the range of bidirectional inverters’ capabilities. This is because battery applications today require low voltage DC-AC inverters with AC-DC power flow capability to store energy for later use. As such, only buck-boost based topologies are majorly being proposed and used for this functionality. The buck boost converter is the most widely used in such applications because of its higher efficiency, low component count and simple structure. It has drawbacks, however, such as: pulsating input and output currents - this leads to lower high electromagnetic interference; lower power factor during AC-DC power flow rectification when the batteries are being recharged; and external filter is also required during this power flow to keep the charging voltage constant. This research proposes a bidirectional inverter that attempts to overcome the drawbacks of the widely used buck-boost converter-based topology. The bidirectional inverter proposed in this work is based on a bidirectional Ćuk converter. The Ćuk converter has both continuous input and output currents. A galvanic isolation option on a Ćuk converter is simpler than a buck boost converter - this is important for grid tied systems. The inverter is based on a pseudo DC-link architecture - it uses a front end Ćuk converter cascaded with an unfolding bridge to convert DC power into AC power. The switches in the converter stage are switched at high frequency, while the switches in the unfolding stage are switched slower at the grid frequency. This configuration is desirable over the two-stage topologies because the switching losses in the unfolding bridge are lower because of this low switching frequency used. This configuration also ensures good switch utilization at the unfolding stage by lowering the parasitic effects on the power transfer. The proposed inverter has 4 modes of operation: during modes I and II the power is positive, and it converts DC power into AC power; during modes III and IV the power is negative, and it converts AC power back into DC power. The inverter is designed such that during DC-AC power flow, the input and output inductor currents and coupling capacitor voltage are continuous for improved efficiency. During the AC-DC power flow, the coupling capacitor voltage is discontinuous to achieve a higher input power factor by improving the AC line current, thereby simultaneously increasing the efficiency. The inverter was analysed in terms of: the dead time inserted into the switches to avoid shoot through and shortcircuiting switches; the parasitic effects on the power transfer ratio. Because the Cúk inverter is a high order system, several robust control strategies, such as sliding mode and current control have been proposed. These control methods require complex theory and present practical challenges to be reviewed. As such a new nested loop control strategy was proposed based on the dynamics of the coupling capacitor as the primary energy storage in the Cúk inverter. The control strategy uses 2 loops: an inner current loop and an outer voltage loop. Lead compensators were designed for both the current and voltage loops to achieve good dynamic response at a high bandwidth. Both simulated and experimental results showed that the bidirectional inverter was able to meet the design specifications. The control strategy showed good dynamic response and disturbance rejection under several inverter variations. Although the efficiency during simulations was above 96%, the experimental efficiency dropped significantly because the inverter was built on a Vero board for easy manipulation. The AC input power factor was > 0.95 for both simulated and experimental results.
- ItemOpen AccessThe impact of voltage unbalance and regulation on the life expectancy of LV induction machines(2019) Mponwana, Samuel; Barendse, PaulThe induction machine is the most widely used electrical machine in the world, they are used for industrial, commercial and industrial applications. When manufactured they have nameplate ratings that stipulate the voltages and currents at which they may be operated. The quality of supply from power utilities can lead to them being operated under unbalanced voltage conditions. Power utilities experience voltage unbalances and voltage dips when they provide services to end users. In South Africa ESKOM is the primary power utility, it has a license agreement with NERSA stipulating the allowable voltage levels. In most cases the operating levels are within the agreed limits however many consumers are exposed to voltage levels at the fringe of these specified limits. This can be detrimental to electrical equipment if operated under these conditions for considerable lengths of time. NRS048 part 2 provides limits for voltage regulation and voltage unbalance which can exist on the power network at various voltage levels. It is incumbent on each utility to ensure that the quality of power supplied to end users complies with the minimum standards specified in the NRS048. Most customers connected to rural 11/22kV networks are farmers, where a large portion of the load is pumps driven by low voltage AC induction motors. NRS048 dictates the voltage limits as ±10% of the nominal supply voltage of 400V. The voltage unbalance on three phase systems is limited to 2%, and 3% for predominantly single-phase systems. Utilities such as Eskom have standards and operating procedures in place to run the networks optimally within these limits. Variations in the voltage levels on rural networks can be significant during load changes. Voltage unbalance levels can also be high due to the use of single-phase loads on these networks. Normally, utilities have operating procedures in place to allow operation of the supply network within the limits specified in the NRS048. Although these operating levels are within specified limits, a large portion of the customer’s plant can however be exposed to voltage levels well above or below the rated value of the equipment. Operating under these conditions can eventually lead to failure of the customer equipment. The impact of operating AC motors with voltages levels above or below the rated levels has not been fully determined on rural networks in South Africa. Research is therefore required to assess the impact of these operating conditions on customer equipment in rural areas and whether utility operating procedures need to be revised to take into consideration customer equipment. This has consequences for Eskom Distribution that relates to: a) The way in which the network voltages are managed, and b) Claims lodged against the Distribution business resulting from damage to customer’s motors The purpose of this research project is to investigate and quantify the impact of voltage regulation (over and under) and unbalance conditions on a typical rural feeder on the lifespan of induction machines. There are various definitions of voltage unbalance by NEMA, the IEC and the IEEE. The IEC definition is used in this research report, it is known as the true definition and incorporates both magnitude and phase information. To estimate the loss of life in induction machines operating under unbalanced conditions, the positive and negative sequence per phase equivalent circuits must be determined and the thermal model needs to be obtained and quantified as well. The losses obtained from the per phase sequence circuits are inputs to thermal model which in turn is used to predict the induction machine stator windings temperatures. Factors that are considered when analysing the impact of voltage unbalance and regulation on the life expectancy of machines include, the induction machine manufacturer, the size of the induction machine, the voltage rating of the induction machine and the efficiency class of the induction machine. The research presented in this report is primarily focused on the impact of voltage unbalance and regulation on the life expectancy of low voltage induction machines. The operating conditions considered in this report are prevalent on a typical rural feeder. Most customers connected to rural 11/22kV networks are farmers and a large portion of load pumps are driven by low voltage induction machines. This report presents the effect of these operating conditions on the life expectancy of the machines. The thermal model presented is suitable for continuously operated machines (S1). Since the operating conditions considered are primarily prevalent on typical rural feeders, the machines considered were also the machines primarily used in those regions. This has consequence that relates to the way in which the network voltages are managed, and claims lodged against power utilities resulting from damage to end user’s induction machines. The thermal model presented in this report can be incorporated as an algorithm and be implemented in microprocessor devices which enhance the level of accuracy and flexibility. As a practical application of the thermal model real time data can be processed according to the firmware thermal algorithm program and results are compared with the expected values and stored in memory. If a machine protection device is used, it computes an analog value which is then compared with the output of the thermal model algorithm. In practice the MPD usually triggers the digital outputs if the compared analog values exceed the set thermal threshold.