This thesis describes a method of minimising the total losses of an inverter-fed 3-phase squirrel-cage induction motor when the motor ,is subjected to a pulse-width modulated (PWM) voltage waveform. The inverter is supplied from a d.c. link and operates at variable frequency to provide speed control of the motor. Appropriate triggering of the inverter's six main thyristors generates pulse-width modulated voltage waveforms for application to the induction motor. The operation of an induction motor with nonsinusoidal voltage applied results in a reduction of the motor's efficiency due to the harmonics present in the waveform. The aim of the project is to minimise the total losses by obtaining optimum PWM voltage waveforms rather than by improving the design of the motor. This requires a thorough examination of motor losses. The determination of a RiM voltage waveform which may be produced by the described inverter is subjected to ccnstraints which characterise the operation of the drive system. The motor operates with constant airgap flux density throughout its speed range to obtain maximum output power at fixed per unit slip. In addition the switching frequency of the thyristors must not exceed a specified limit to avoid short circuiting of the inverter. The calculation of the motor's steady-state performances for both sine wave and PWM supplies is incorporated in a computer program. The detail of the experimental and theoretical performances are given and comparison is made between sinusoidal and PWM voltage wave systems. Good agreement is obtained between test and calculated results on an inverter-fed 7.5 kw squirrelcage induction motor. It is concluded that the degradation of motor efficiency due to the applied PWM voltage waves is mainly the result of increased copper losses, which are produced by harmonic currents. The minimisation of the losses for continuous, constant-flux, operation of the induction motor is achieved for the given constraints. It is found that the total losses can be further minimised if the d.c. link voltage is variable. This permits improved motor performance but adds complexity and cost to the d.c. link voltage supply.

An opto-electrical modelling platform has been designed to model the effects of illumination spectra on amorphous silicon solar cells of different i-layer thickness and degradation state. The illumination spectra, which were investigated in this work, are solar simulator spectra and solar spectra recorded outdoors at CREST, Loughborough. These spectra are used to probe the effect of spectral variation on a-Si:H solar cell performance and its co-dependence with the state of the device. For the case of indoor evaluation of performance of a-Si:H solar cells, it is shown that the performance of the device remains relative to the illumination source of the solar simulator. Spectra with Average Photon Energy (APE) higher than AM1.5G tend to overestimate the performance parameters (JSC, MPP, VOC) of the device, while spectra with APE lower than AM1.5G tend to underestimate the values of the performance parameters of the device. The maximum power deviation of most class-A solar simulators is less than 1% of the actual STC values, but the performance deviation may arise up to 4% for the case of LED light sources. It is suggested to apply voltage dependant corrections to the J-V characteristics, whenever the spectral mismatch between the illumination spectra and AM1.5G is significant. The effects of outdoor spectral variation on the performance of a-Si:H solar cells has been investigated. The results show that light intensity is primarily responsible for a- Si:H outdoor performance. The APE of the outdoor spectra is also identified a significant factor for the variation of performance. The magnitude of maximum power deviations due to APE changes is in the range of ±3% as compared to power output of the device under the AM1.5G spectrum. The percentage of performance variation to STC differed for a-Si:H solar cells of different i-layer thickness and level of degradation. Specifically devices with thicker i-layer, which have suffered degradation, are prone to performance variations. Finally, the energy yield and the performance ratio of amorphous silicon solar cells were reviewed in respect to outdoor spectral changes. The performance ratio is a useful method for cases where prediction of power output is necessary. However, it is suggested that PV modules should be rated on the basis of their annual energy yield, when possible.

No description available.

This study investigates the feasibility and potential benefits of high frequency alternating current (HFAC) for vehicle auxiliary electrical systems. A 100Vrms, 50kHz sinusoidal AC bus is compared with 14V DC and 42V DC electrical systems in terms of mass and energy efficiency. The investigation is focused on the four main sub-systems of an on-board electrical network, namely: the power generation, power distribution, power conversion and the electrical loads. In addition, a systemlevel inquiry is conducted for the HFAC bus and a comparable 42V DC system. A combination of computer simulation, analytical analysis and experimental work has highlighted benefits for the HFAC power distribution sub-system and for low-torque motor actuators. Specifically, the HFAC conductor mass is potentially 70% and 30% lighter than comparable 14V DC and 42V DC cables, respectively. Also, the proposed cable is expected to be at least 80% more energy efficient than the current DC conductor technology. In addition, it was found that 400Hz AC machines can successfully replace DC motor actuators with a rated torque of up to 2Nm. The former are up to 100% more efficient and approximately 60% lighter and more compact than the existing DC motors in vehicles. However, it is argued that the HFAC supply is not feasible for high-torque motor actuators. This is because of the high energy losses and increased machine torque ripple associated with the use of HFAC power. The HFAC power conversion sub-system offers benefits in terms of simple power converter structure and efficient HFAC/DC converters. However, a significant limitation is the high power loss within HFAC/AC modules, which can be as high as 900W for a 2.4kW load with continuous operation. Similar restrictions are highlighted for the HFAC power generation sub-system, where up to 400W is lost in a 4kW DC/HFAC power module. The conclusion of the present work is that the HFAC system offers mass and energy efficiency benefits for the conventional vehicle by leveraging the use of compact lowtorque motor actuators and lightweight wiring technology.

This work aims at the advancement of state-of-art accelerometer design and optimization methodology by developing an ear-plug accelerometer for race car drivers based on a novel mechanical principle. The accelerometer is used for the measurements of head acceleration when an injurious event occurs. Main requirements for such sensor are miniaturization (2×2 mm), because the device must be placed into the driver earpiece, and its measurement accuracy (i.e. high sensitivity, low crosstalk and low nonlinearity) since the device is used for safety monitoring purpose. A micro-electro-mechanical system (MEMS)-based (bulk micromachined) piezoresistive accelerometer was selected as enabling technology for the development of the sensor. The primary accelerometer elements that can be manipulated during the design stage are: the sensing element (piezoresistors), the micromechanical structure and the measurements circuit. Each of these elements has been specifically designed in order to maximize the sensor performance and to achieve the miniaturization required for the studied application. To achieve accelerometer high sensitivity and miniaturization silicon nanowires (SiNWs) as nanometer scale piezoresistors are adopted as sensing elements. Currently this technology is at an infancy stage, but very promising through the exploitation of the “Giant piezoresistance effect” of SiNWs. This work then measures the potential of the SiNWs as nanoscale piezoresistors by calculating the major performance indexes, both electrical and mechanical, of the novel accelerometer. The results clearly demonstrate that the use of nanoscale piezoresistors boosts the sensitivity by 30 times in comparison to conventional microscale piezoresistors. A feasibility study on nanowires fabrication by both top-down and bottom-up approaches is also carried out. The micromechanical structure used for the design of the accelerometer is an optimized highly symmetric geometry chosen for its self-cancelling property. This work, for the first time, presents an optimization process of the accelerometer micromechanical structure based on a novel mechanical principle, which simultaneously increases the sensitivity and reduces the cross-sensitivity progressively. In the open literature among highly symmetric geometries no other study has to date reported enhancement of the electrical sensitivity and reduction of the cross-talk at the same time. Moreover the novel mechanical principle represents advancement in the accelerometer design and optimization methodology by studying the influence of a uniform mass moment of inertia of the accelerometer proof mass on the sensor performance. Finally, an optimal accelerometer design is proposed and an optimized measurement circuit is also specifically designed to maximize the performance of the accelerometer. The new proposed accelerometer design is capable of increasing the sensor sensitivity of all axes, in particular the Z-axis increases of almost 5 times in respect to the current state-of-art-technology in piezoresistive accelerometer. This occurs thanks to the particular newly developed approach of combination of beams, proof mass geometry and measurement circuit design, together with the use of silicon nanowires as nanoscale piezoresistors. Furthermore the cross-sensitivity is simultaneously minimized for a maximal performance. The sum of the cross-sensitivity of all axes is equal to 0.4%, well below the more than 5% of the state-of-art technology counterpart reported in the literature. Future work is finally outlined and includes the electro-mechanical characterization of the silicon nanowires and the fabrication of the proposed accelerometer prototype that embeds bottom up SiNWs as nanoscale piezoresistors.

Voltage security assessment is becoming a more and more important issue due to the fact that electrical power systems are more prone to voltage instability under increased demand, and it can be time-consuming to determine the actual level of voltage security in large power systems. For this reason, this thesis presents a novel method for calculating the margin of voltage collapse that is based on the Continuation Power Flow (CPF) method. The method offers a flexible and reliable solution procedure without suffering from divergence problems even when near the bifurcation point. In addition, the new method accounts for reactive power limits. The algorithmic continuation steps are guided by the prediction of Q-limit breaking point. A Lagrange polynomial interpolation formula is used in this method in order to find the Q-limit breaking point indices that determine when the reactive power output of a generator has reached its limit. The algorithmic continuation steps will then be guided to the closest Q-limit breaking point, consequently reducing the number of continuation steps and saving computational time. The novel method is compared with alternative conventional and enhanced CPF methods. In order to improve CPF further, studies comparing the performance of using direct and iterative solvers in a power flow calculation have also been performed. I first attempt to employ the column approximate minimum degree (AMD) ordering scheme to reset the permutation of the coefficient matrix, which decreases the number of iterations required by iterative solvers. Finally, the novel method has been applied to a range of power system case studies including a 953 bus national grid transmission case study. The results are discussed in detail and compared against exiting CPF methods.

Traditionally, the electrochemical battery has been the prime medium by which electrical energy is stored for future use. Increasingly, the demands of modern systems such as electric vehicles, renewable energy, distributed generation, smart grid and others has stretched the development of new chemistries, materials and assembly techniques for electrochemical batteries. Additionally, some load profiles in these applications demand extremely high dynamic behaviour which is either undeliverable by conventional electrochemical batteries or is undesirably damaging to these technologies. As such, a family of electrochemical storage, known generally as supercapacitors or ultracapacitors, have been developed and implemented for such applications. In recent years advancements in electrochemical technology has led to hybridisation of high capacitance devices. Lithium-ion capacitors that are used in this work are, with their higher cell voltage and modern packaging, expected to be among the next emerging families of state-of-the-art electrical energy storage devices. The relatively low cell voltage of high capacitance cells requires them to be connected in series to attain a system level voltage. During charging and discharging, manufacturing tolerances between the cells results in voltage mismatch across the stack. Mismatched voltages are an inefficient use of the energy storage medium and can lead to dangerous failures in the cells. Several techniques exist to limit the variance in cell voltages of supercapacitors across a series connected stack. These range from simple systems which discharge the cells at higher voltages through resistors to more complex active converter systems which equalise the cell voltages through charge redistribution via a power electronic converter. Whilst the simpler schemes are effective they are very inefficient and as such are not suitable for use in many applications. A number of active converter voltage equalisation schemes have been proposed in literature, however, each of these equalisation schemes exhibit flaws which either makes them less desirable or less effective for a broad range of applications. Therefore, a new equalisation converter topology is proposed which is designed for greater equalisation effectiveness, modularity and size. The proposed equalisation converter differs from previously published equalisation schemes by allowing energy transfer between any pair of cells without the cumbersome multi-winding transformers employed in existing equalisation converters. The new equalisation scheme uses a bi-directional arrangement of MOSFET switches for galvanostatic isolation allowing the converter to be multiplexed to the stack. This arrangement allows the total size of the equalisation scheme to be reduced whilst maintaining performance.

The aim of this work is a comparative study of the electromechanical transient phenomena and of the stability limits of synchronous power systems, when they are mainly interconnected by a. c. transmission lines, but when one of the a. c. lines is replaced by the d.c. one. A thorough survey of the recent literature on h. v. d. c. technology has been made in order to be conversant with latest developments in the field. Digital computation and numerical analysis are used to study mathematical models of the systems under consideration. For the h. v. d. c. system, three digital programmes have been established to find the characteristics of the converter, to simulate the h.v.d.c. link in the a. c. power system and to study the transient behaviour of the systems, respectively. For the a. c. system the Park's equations describing the behaviour. of the synchronous machines are modified to suit the digital computers Then two digital programmes have been written, for the a. c. system one to draw the swing curves and the other to find the stability boundaries which incorporate the A. V. P. and the speed governor. To make a comparative study of the a.c. - d. c. system and of the equivalent a.c. system, two new comprehensive programmes have been established by incorporating the d.c. programme for the transient studied in the modified a.c. programmes. One of the digital programmes draws the swing curves while the other finds the stability boundaries of both the systems, with or without A. V. R. Thus, by performing different tests, it has been established that from the stability point of view the a. c. -d. c. system can be made superior to the equivalent a. c. one, provided that provision is made for increasing the power through the d. c. line when a fault is sensed.

Bioethanol is an alternative fuel produced mainly by biochemical conversion of biomass. This can be carried out efficiently and economically by simultaneous saccharification and fermentation (SSF) of sugarcane, corn, wheat, cellulose, etc., a process which integrates the enzymatic saccharification of the complex, polymeric sugars to glucose with the fermentative synthesis of ethanol by yeasts (Saccharomyces cerevisiae). However, the SSF unit operation still contributes nearly 50% to the cost of ethanol production. In SSF it is essential that a high sugar yield is obtained in the saccharification of cellulose. This yield is affected by factors such as inhibition of enzyme action by heat and other degradation products, enzyme and substrate concentrations, speed of enzyme action, adsorption of cellulase to cellulose, and degree of agitation. SSF was investigated in an intensified form of plug flow reactor, called the Oscillatory Baffled Reactor (OBR). The effect of agitation on saccharification of microcrystalline cellulose was correlated with the mean strain rates in the reactors. After 168 h of saccharification at 200 Wm-3 (Watts per cubic meters), 91% conversion of the cellulose (~25 g L-1 glucose) was observed in the OBR, whereas in the STR 74% conversion (~21 g L-1 glucose) was observed. At 120 Wm-3, the conversion in the OBR was 69% (~19 g L-1 glucose) within the first 24 h of saccharification and 88% conversion (24 g L-1 glucose) after 168 h. At the same power density the conversions in the STR were 55% (15.3 g L-1 glucose) and 67% (~18.6 g L-1 glucose), differences of 14 and 21% respectively. At 200 Wm-3 the ethanol concentration in a Stirred Tank Reactor (STR) after 72 h was 10.9 g L-1 (80.3% of theoretical yield) equivalent to production yield Yp/s = 0.55 g.g-1 cellulose and a volumetric productivity Qp of 0.15 g L-1 h-1. In the OBR at 200 Wm-3 the final concentration of ethanol after 72 h SSF was 12.5 g L-1 (93.8% of theoretical yield) equivalent to production yield Yp/s = 0.63 g.g-1 cellulose and a volumetric productivity Qp of 0.2 g L-1 h-1. It is hypothesised that the reason for these differences is the differing extents of cellulase deactivation in the two reactors. The OBR has a more uniform shear field than the STR, so the enzyme and yeasts would be exposed to fewer pockets of high shear.

No description available.