Abstract This paper reports the quantification of the evaporative emissions from gasoline-poweredmotor vehicles in South Africa under various conditions including high temperatures. Three vehicles with and without evaporative emission control systems have been tested to quantify gasoline evaporative emissions, including both diurnal and running loss emissions. Measurements were made in a Sealed Housing for Evaporative Determination over a range of ambient temperatures. The standard diurnal test shows that evaporative emissions increase progressively with increasing fuel temperature. South African gasoline poweredmotor vehicles exhibitedevaporative emissions 10 times higher than that allowed by the United States Environmental Protection Agency. The extended-time diurnal test shows that the fuel emittedfrom gasoline-poweredmotor vehicles in South Africa is about 97 million litres of gasoline every year.

Thesis (MScEng)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: During the course of the CSIR’s research into the characterisation of explosive sources to devise methods of active intervention against threats, the need has arisen to research a particular means of early identification of the threat, which is the intense light flash during the threat detonation. For this purpose, a low cost rugged fast optical sensor was sought, since the application thereof would imply possible destruction, especially if integrated into an active intervention system later on. Given the average time of about 1ms available for intervention, it is clear that the active intervention system needs to operate within that period, hence the interest in the characteristic light emission of detonations in the pre-millisecond time frame. It was thought that by characterising this emitted light in terms of wavelength (temperature) and amplitude (and maybe other unique phenomena), the size of the threat could be determined and logic decisions derived therefrom. Needless to say, the environment in which the detonation light emission sensor is to operate, is extremely hostile in terms of shock, dust, flying debris, fast rise time of the explosive event, and Electro-magnetic Interference ( EMI) caused by the detonation itself. It must be noted that the light sensor research was driven by the outcome of research tests performed in aid of the development of an active intervention system. During this research the possibility of using commercially available low cost optical detectors at room temperature in combination with cost effective narrow band pass op- tical filters for the relative measurement of the light emission at discrete wavelengths during explosive detonation events were investigated. In 2006, not much applicable lit- erature could be found on this subject, hence the educated “shot-in-the-dark” approach then, which, by a systematic approach of explosive tests and continuous evaluation up to 2011, led to a surprisingly simple and robust low cost optical sensor. The research commenced with a range of optical detector elements selected for their responsivity and bandwidth in the optical spectrum of interest; the optical filtering by means of the recording of the emitted light signal during scaled down explosive tests at the Blast Impact Survivability Research Unit (BISRU) at the University of Cape Town. These tests were followed by full-scale tests at DBEL, and confirmed the findings at BISRU that the light emissions at the longer wavelengths (>2 m) manifest themselves too late for use within the intervention time frame. It was therefore decided to concentrate on the ultra-violet (UV) to near infra-red (NIR) spectrum of the emitted light for further full scale tests, since these discrete spectra showed the most promise for characterisa- tion of the emitted light. During this period a robust sensor housing with detector and filter mounts was designed for protection against blast shock and EMI. During the following years, certain types of optical detectors that were used during previous tests were eliminated according to results obtained, and more discrete narrow band pass filters added in the visible to NIR spectrum. A dedicated fast instrumen- tation amplifier (bandwidth > 1MHz and selectable gain up to 40dB) was developed to amplify weak signals (mainly caused by the heavy load in the detector circuit to improve rise times). However, the emission of light per wavelength in this region was measured to be relatively strong, and actually not as fast as was anticipated. This meant that the load resistor value of the detector element could be increased without affecting the signal negatively (bandwidth sufficient), thus adding to the amplitude of the signal to such a point that amplification in a 10m to 30 meter stand-off scenario was no longer needed. This culminated in an unamplified universal detector element being used with various narrow band pass filters up to 1 m, integrated as a very robust analog sensor at a discrete wavelength, and facilitating the direct comparison of light amplitude/relative intensity of the detonation at discrete spectral points. The sensor was employed in the field at various full scale explosive tests at DBEL, which led to the capture of a vast amount of light emitted data for different types of explosives, at various distances from the detonation, and of varying mass. Analysis of this data showed that the broadband light intensity of the emitted light scales to the explosive mass1/3 (as published by FJ Mostert and M Olivier in the Journal for Applied Physics, October 2011). Further analysis also confirmed the attenuation of the emitted light intensity by the square of the distance. Besides the aforesaid, various other key inputs to a possible active intervention algorithm have been identified. These findings are inputs to the determination of i.a. the detonation threat size, a vital component in the active intervention algorithm. The results of these experiments confirmed that the final low cost analog sensor can measure relative light emission at discrete wavelengths from detonation of explosives in the very early stages of development, and that the sensor has many other applications in the detonics research fields as well. / AFRIKAANSE OPSOMMING: Gedurende die WNNR se navorsing om detonerende bronne te karakteriseer ten einde aktiewe teenmaatreëls daar te stel, het die behoefte na vore gekom om die intense ligflits van ’n detonasie te ondersoek en te karakteriseer. Vir hierdie doel is ’n lae koste ligsensor benodig, synde die uiteindelike aanwending van hierdie ligsensor die vernietiging daarvan sou beteken, aldus die lae koste vereiste. Gegewe die kort tydsduur van die detonasie (’n paar millisekondes), is dit duidelik dat die ligflits karakerisering voor 1ms moet geskied, en daarom moet die ligsensor ook baie vinnig reageer om insette te lewer tot ’n aktiewe teenmaatreëlstelsel. Daar moet op gelet word dat die ligsensor se ontwikkeling uitkomsgedrewe was deur die navorsingstoetse om ’n aktiewe teenmaatreëlstelsel daar te stel. Een van die insette tot so ’n aktiewe teenmaatreëlstelsel is die grootte van die bedreiging: deur die ligflits te karakteriseer met die lae koste ligsensors t.o.v. golflengte, ligamplitude en moontlik ander verskynsels, kan bv. die massa inset verkry word wat nodig is vir die teenmaatreël algoritme. Die omgewing waarin die ligsensor moet funksioneer is baie onvriendelik i.t.v. skok, stof, vlieënde partikels en elektromagnetiese steurings, en sou daarteen beskerm moes word. Gedurende die navoring om so ’n ligsensor te ontwikkel (samelopend met die teen- maatreël navorsing), is kommersiële kamertemperatuur detektors oorweeg en aange- wend, in samewerking met nouband optiese filters. Die doel was om die ligopbrengs per golflengte te karakteriseer m.b.t. die plofstof massa, plofstof tipe en geometrie, en die afstand vanaf die detonasie. Bitter min literatuur oor die ligmeting van detonasies is aanvanklik gevind, aldus is ’n basislyn daargestel en deur sistematiese toetsing, ontleding en verbetering voortgegaan met die navorsing. Dit het gelei tot ’n verrassend eenvoudige en verharde lae koste ligsensor, wat deur meting sleutelinsette kon lewer tot die gesogte aktiewe teenmaatreël algoritme. Kommersiële detektors en nouband optiese filters is uitgesoek na aanleiding van hul prys en prestasie, en waar nodig, is versterking van die seine aangebring. Verskeie toetse met plofstof (op klein en groot skaal) is uitgevoer, waartydens ligmeting by spesifieke golflengtes opgeneem is. Analise van hierdie data het getoon dat die langer golflengtes (>2 m) se verskyning te laat is vir insluiting in die teenmaatreël algoritme, en is dus geleidelik (of sistematies) uitgeskakel. Die klem het geskuif na die detonasie liguitsetting in die UV tot naby infrarooi spektrum, wat nuwe detektors en filters tot gevolg gehad het (uitkoms gebasseerde navorsing). In die proses is ’n instrumentasie versterker ontwerp en gebou, vir buffering en versterking van seine hoër as 1 MHz met ’n selekteerbare aanwins van tot 40dB. Toetse met volskaalse ladings het egter getoon dat die liguitset besonder sterk is in die UV tot naby infrarooi spektrum, en ’n onversterkte ligsensor is aldus op die proef gestel. Hierdie proeflopie het getoon dat die onversterkte ligsensor besonder goed funksioneer op afstande tot en met 30m, en daar is op hierdie model voortgebou. Die verharde onversterkte ligsensor is aangewend in verskeie verdere volskaalse plofstof toetse, en het data gelewer t.o.v. detonasie liguitstraling by spesifieke golflengtes vir tipes plofstof, plofstof massas, plofstof geometrie en afstande vanaf die detonasie. Analise van hierdie data het getoon dat breëband liguitsetting se intensiteit skaal met die plofstof massa1=3 (gepubliseer as ’n artikel deur FJ Mostert en M Olivier in die Journal of Applied Physics’ - Oktober 2011). Verdere analise het verskeie sleutelinsette tot ’n aktiewe teenmaatreël algoritme geïdentifiseer. Die uitkoms van hierdie eksperimentele navorsing het getoon dat die lae koste lig- sensor relatiewe liguitsetting van ’n detonasie by gekose golflengtes vinnig kan meet in die baie vroeë stadia van die detonasie. Buiten dit, het die sensor verskeie ander nuttige aanwending in die detonasie navorsingsveld.

[Truncated abstract] The cochlea presumably possesses a number of regulatory mechanisms to maintain cochlear sensitivity in the face of disturbances to its function. Evidence for such mechanisms can be found in the time-course of the recovery of CAP thresholds during experimental manipulations, and in observations of slow oscillations in cochlear micromechanics following exposure to low-frequency tones (the “bounce phenomenon”) and other perturbations. To increase our understanding of these oscillatory processes within the cochlea, and OHCs in particular, investigations into cochlear regulation were carried out using a combination of mathematical modelling of the ionic and mechanical interactions likely to exist within the OHCs, and electrophysiological experiments conducted in guinea pigs. The electrophysiological experiments consisted of electrocochleographic recordings and, in some cases, measurement of otoacoustic emissions, during a variety of experimental perturbations, including the application of force to the cochlear wall, exposure to very-low-frequency tones, injection of direct current into scala tympani, and intracochlear perfusions of artificial perilymph containing altered concentrations of potassium, sodium, and sucrose. To obtain a panoramic view of cochlear regulation under these conditions, software was written to enable the interleaved and near-simultaneous measurement of multiple indicators of cochlear function, including the compound action potential (CAP) threshold, amplitude and waveshape at multiple frequencies, the OHC transfer curves derived from low-frequency cochlear microphonic (CM) waveforms, distortion-product otoacoustic emissions (DPOAEs), the spectrum of the round-window neural noise (SNN), and the endocochlear potential (EP). ... The mathematical model we have developed provided a physiologically-plausible and internally-consistent explanation for the time-courses of the cochlear changes observed during a number of different perturbations. We show that much of the oscillatory behaviour within the cochlea is consistent with underlying oscillations in cytosolic calcium concentration. We conclude that a number of the discrepancies between the simulation results and the experimental data can be resolved if the cytosolic calcium functions as two distinct pools: one which controls basolateral permeability and one which controls slow motility. This two-calcium-pool model is discussed.