Abstract Heat exchangers are used in industrial processes to transfer energy from one source to another. Heat exchangers improve the energy efficiency of processes and therefore increase the profitability and decrease the environmental impact of production. Effectiveness of heat exchangers may be diminished by fouling, in which unwanted material deposits on the heat transfer surface reducing the heat transfer and increasing the pressure drop of the system. Due to fouling, the energy demand, and the operation and maintenance costs of industrial processes increase significantly. In addition, fouling causes considerable environmental effects due to the increased energy demand and the use of additives and cleaning chemicals. In this thesis, the crystallization fouling of calcium carbonate on heat transfer surfaces was studied using experimental and modeling methods with the aim to reduce fouling by increasing the understanding of the crystallization fouling phenomenon and providing a validated model for studying crystallization fouling on heat transfer surfaces. The fouling experiments were conducted in a laboratory scale set-up to study the crystallization fouling mechanism under controlled conditions. Based on the knowledge and data gained from the experiments, a crystallization fouling model was developed, and validated in various conditions. As a result, the most important parameters affecting crystallization fouling were identified, and the governing sub-processes were determined. The developed model was used to optimize the surface temperature of a heat exchanger to provide the maximum heat transfer rate when the heat exchanger is subject to fouling at various cleaning intervals. In addition, the used methods were applied in the study of the abatement of fouling by surface modifications. Based on the results, the studied crystallization process is governed by the surface integration sub-process. In addition, the residence time of the fluid at the wall affects the mass deposition. The results showed that the surface temperature and the shear stress are the most important parameters affecting fouling in the studied conditions, and therefore, they should be defined accurately. For this purpose, CFD was found to provide a useful tool. The developed models were found to predict reliably the experimental conditions. Therefore, the thesis shows that the developed model facilitates the design of heat exchangers, but also assists in minimizing the fouling of heat exchangers. / Tiivistelmä Lämmönvaihtimia käytetään teollisissa prosesseissa siirtämään energiaa kohteesta toiseen. Lämmönvaihtimet parantavat prosessien energiatehokuutta ja siten lisäävät tuotannon kannattavuutta ja vähentävät ympäristövaikutuksia. Lämmönvaihdinten energiatehokkuutta heikentää kuitenkin likaantuminen, jossa lämmönsiirtopinnalle muodostuu kerrostuma, joka heikentää lämmönsiirtoa ja aiheuttaa virtausvastusta. Likaantuminen lisää lämmönvaihdinten energiankulutusta ja käyttökustannuksia sekä aiheuttaa merkittäviä ympäristöpäästöjä kasvaneen energiantarpeen sekä lisä- ja puhdistusaineiden käytön vuoksi. Tässä työssä tutkittiin kalsium karbonaatin aiheuttamaa lämmönsiirtopintojen kiteytyvää likaantumista käyttäen sekä kokeellisia että mallinnusmenetelmiä. Työn tavoitteena oli lisätä tietoa kiteytyvästä likaantumisesta sekä kehittää validoitu laskentamalli, jolla voidaan tutkia lämmönsiirtopintojen kiteytyvää likaantumista. Likaantumiskokeet tehtiin laboratoriomittakaavan koelaitteessa, jolla voidaan tutkia likaantumismekanismia hallituissa olosuhteissa. Kokeista saadun tiedon ja aineiston perusteella kehitettiin kiteytyvää likaantumista kuvaava malli, joka validoitiin eri olosuhteissa. Tulosten perusteella identifioitiin tärkeimmät kiteytyvään likaantumiseen vaikuttavat parametrit sekä määritettiin rajoittavat osa-prosessit. Kehitettyä mallia käytettiin lämmönsiirtoprosessin optimoinnissa. Lisäksi käytettyjä metodeja sovellettiin tutkimukseen, jossa likaantumista pyrittiin vähentämään pintamodifikaatioiden avulla. Tulosten perusteella tutkittu kiteytymisprosessi on pinta-integraation rajoittama. Lisäksi havaittiin, että fluidin viipymäaika lämmönsiirtopinnalla vaikuttaa likaantumisnopeuteen. Tulokset osoittivat, että pintalämpötila ja leikkausjännitys ovat tärkeimmät kiteytyvään likaantumiseen vaikuttavat tekijät tutkituissa olosuhteissa. Näin ollen niiden tarkka määrittely on erittäin tärkeää likaantumisnopeuden mallintamiseksi luotettavasti. Virtauslaskenta (CFD) osoittautui hyödylliseksi työkaluksi näiden parametrien määrittämisessä. Tulosten perusteella kehitetty malli ennustaa luotettavasti kokeellisia olosuhteita. Näin ollen tämä työ osoittaa, että kehitettyä mallia voidaan käyttää apuna paitsi lämmönvaihtimien suunnittelussa myös lämmönsiirtopintojen kiteytyvän likaantumisen vähentämisessä.

This study consists of an experimental investigation into the fluid flow and heat transfer aspects of microchannels. Rectangular copper microchannels of hydraulic diameters 1.05 mm, 0.85 mm and 0.57 mm were considered. Using water as the working fluid, heat transfer and pressure drop characteristics were determined under a constant surface heat flux for different inlet configurations in the laminar and transitional regimes. Three inlet geometries were experimentally investigated: a sudden contraction inlet, a bellmouth inlet and a swirl-generating inlet. The influence of the inlet conditions on the pressure drop, Nusselt number and critical Reynolds number was determined experimentally. Pressure drop results showed good agreement with existing correlations for adiabatic conditions. Diabatic friction factor results for the sudden contraction and bellmouth inlets were overpredicted when using the friction factor results from literature. It is noted that a relationship between the pressure drop and heat flux existed in the laminar regime, where an increase in the heat input resulted in a decrease in the friction factor. The bellmouth inlet condition showed an enhancement of the heat transfer in the transition regime compared with the sudden contraction inlet. The critical Reynolds number for the onset of transition for the sudden contraction inlet was found to be approximately 1 950, with a sharp rise to the turbulent regime thereafter. The bellmouth inlet influenced the originating point of the transition regime, which commenced at a Reynolds number of approximately 1 600. A smoother and more gradual increase to the turbulent regime was observed as an effect of the bellmouth inlet over the sudden contraction inlet. The swirl-generating inlet condition produced higher friction factor results in all three flow regimes. Transition occurred at a Reynolds number of approximately 1 500 and the turbulent regime was quickly ii reached thereafter. The turbulent regime friction factor was found to be significantly higher with the swirl inlet compared with both the sudden contraction and bellmouth inlets. Nusselt numbers continued to increase until the onset of the transition regime, and did not converge to a constant value as stated in theory. Similar enhancement of the transition regime with the bellmouth inlet was observed for the Nusselt numbers as with the friction factors. The initial turbulent regime results followed the trend of the theory for both the sudden contraction and bellmouth inlet conditions for most of the data sets, with deviation occurring in some of the 0.57 mm test cases. The swirl inlet Nusselt number results were significantly underpredicted by the theory in the early turbulent regime. / Dissertation (MEng)--University of Pretoria, 2014. / gm2014 / Mechanical and Aeronautical Engineering / unrestricted

Master of Science / Department of Mechanical and Nuclear Engineering / Steven J. Eckels, Bruce R. Babin / This project studied the pool boiling of R-134a and R-123 on smooth and enhanced tubes. This is the 1st phase of ASHRAE project RP-1316 "Experimental Evaluation of The Heat Transfer Impacts of Tube Pitch in a Highly Enhanced Surface Tube Bundle". A Turbo BII-HP and a Turbo BII-LP enhanced tubes were used in this study. These tubes were manufactured and donated by Wolverine Tube, Inc. Four different boiling cases were tested, R-134a on smooth tube, R-123 on smooth tube, R-134a on Turbo BII-HP tube, and R-123 on Turbo BII-LP tube. The first step in this study was performing a modified Wilson plot analysis, once completed, the average and local refrigerant heat transfer coefficients were determined. This thesis also presents the enthalpy-based heat transfer analysis (EBHT), a new method for determining the heat exchanger's overall heat transfer coefficient as a function of the enthalpy change of incompressible fluids. The test tubes' outer diameter is 19.05 mm and length is 1 m. Tests were conducted in a single tube test section, in which the test tube was water heated. All tests were conducted at a saturation temperature of 4.44 °C. The heat flux range is 9.2-126.6 kW/m[superscript]2 for testing with R-134a on smooth tube, 9.2-58 kW/m[superscript]2 for R-123 on smooth tube, 4.1-135.1 kW/m[superscript]2 for R-134a on Turbo BII-HP tube and 4.7-59.8 kW/m[superscript]2 for R-123 on Turbo BII-LP tube. Results show that the heat transfer coefficient increases with heat flux for all cases except the case of R-134a on Turbo BII-HP tube, where it experiences a trend change. Part of this study was comparing the smooth and enhanced tubes performances. R-134a Turbo BII-HP tube to smooth tube heat transfer coefficient ratio changes from 4 at low heat flux to 1.7 at high heat flux. R-123 Turbo BII-LP tube to smooth tube heat transfer coefficient ratio changes from 24 at low heat flux to 7 at high heat flux. The performance of Turbo BII-HP and Turbo BII-LP was found to be very similar over the tested heat flux range of the Turbo BII-LP tube. Comparison plots with available literature are presented.

Additive Manufacturing (AM) is a group of manufacturing technologies which are capable to produce 3D solid parts by adding successive layers of material. Parts are fabricated in an additive manner, layer by layer; and the geometric data can be taken from a CAD model directly. The main revolutionary aspect of AM is the ability of quickly producing complex geometries without the need of tooling, allowing for greater design freedom. As one of AM methods, Selective Laser Melting (SLM) is a process for producing metal parts with minimal subtractive post-processing required. It relies on the generation and distribution of laser generated heat to raise the temperature of a region of a powder bed to above the melting point. Due to high energy input to enable full melting of the powder bed materials, SLM is able to build fully dense metal parts without post heat treatment and other processing. Successful fabrications of parts by SLM require a comprehensive understanding of the main process controlling parameters such as energy input, powder bed properties and build conditions, as well as the microstructure formation procedure as it can strongly affect the final mechanical properties. It is valuable to control the parts' microstructure through controlling the process parameters to obtain acceptable mechanical properties for end-users. In the SLM process, microstructure characterisation strongly depends on the thermal history of the process. The temperature distribution in the building area can significantly influence the melting pool behaviour, solidification process and thermal mechanical properties of the parts. Therefore, it is important to have an accurate prediction of the temperature distribution history during the process. The aim of this research is to gain a better understanding of process control parameters in SLM process, and to develop a modelling methodology for the prediction of microstructure forming procedure. The research is comprised of an experiment and a finite element modelling part. Experimentation was carried out to understand the effect of each processing control parameters on the final part quality, and characterise the model inputs. Laser energy input, build conditions and powder bed properties were investigated. Samples were built and tested to gain the knowledge of the relationship between samples' density and mechanical properties and each process control factor. Heat transfer model inputs characterisation, such as defining and measuring the material properties, input loads and boundary conditions were also carried out via experiment. For the predictive modelling of microstructure, a methodology for predicting the temperature distribution history and temperature gradient history during the SLM process has been developed. Moving heat source and states variable material properties were studied and applied to the heat transfer model for reliable prediction. Multi-layers model were established to simulate the layer by layer process principles. Microstructure was predicted by simulated melting pool behaviour and the history of three dimensional temperature distribution and temperature gradient distribution. They were validated by relevant experiment examination and measurement.

[pt] O trabalho analisa a transferência de calor no escoamento de fluidos não-newtonianos através de uma contração abrupta circular de razão 4:1, com temperatura prescrita nas paredes sólidas. O escoamento de fluidos elásticos nesta geometria apresenta uma região de recirculação bem maior que no caso de fluidos Newtonianos. Esta alteração no padrão do escoamento altera significativamente o processo de transferência de calor. O escoamento representa uma boa modelagem do processo de extrusão de líquidos poliméricos. Resolvem-se as equações de conservação de momentum e energia desacopladamente, já que foi adotadas a hipótese de não variação das propriedades do fluido com a temperatura. A relação tensão – taxa de deformação foi feita através de dois modelos constitutivos, Newtoniano generalizado e Maxwell convectado. A hipótese de escoamento lento não foi adotada, como é usualmente feito na literatura da área. Deste modo, analisa-se separadamente a influencia dos efeitos elásticos e inerciais. As equações diferenciais foram integradas numericamente pelo método dos volumes finitos e o aclopamento velocidade\ pressão foi feito através do algoritmo SIMPLE. Pelos resultados obtidos, observa-se a importância da modelagem não newtoniana e da inclusão dos termos inerciais no estudo do escoamento e da transferência de calor no processo de extrusão de polímeros. / [en] It is well known that the flow of a non-Newtonian fluid through a sudden contraction exhibits a vortex in the corner region bigger than the one observed in the corresponding flow of a Newtonian Fluid. This change of pattern of the flow affects significantly the heat transfer at the wall. It was investigated the case og a a 4:1 circular contraction, with uniform temperature distrubuition at the solid walls. This problem represents a first approach for the analysis of the polymeric liquids extrusion process. The flow and temperature field have been obtained from the numerical integration of the conservation equations. To account for the flow dependence of the stress tensor, a generalized Newtonian model and a convected Maxwell model have been employed. The creeping flow hypothesis has not been adopted, so it was possible to analyse the elastic effects and the inertial effects separately. The nuemerical solution have been obtained via a finite-volume method. The results show the importance of the non-Newtonian modeling and of the inclusion of inertial terms in the study of the flow and beat transfer in the polymeric liquids extrusion process.

[pt] Na presente tese, a equação de energia é resolvida para uma tira de aço bi-dimensional que se move com velocidade constante através de um forno contínuo e é submetida a várias condições diferentes de radiação de convecção. As equações de discretização foram obtidas utilizando a técnica de diferenças finitas com enfoque em volume de controle. O conjunto destas equações indicou os sete parâmetros adimensionais do problema. Por se tratar de uma equação elíptica, a solução foi inicialmente encaminhada para a determinação dos comprimentos de tira antes e após o forno, além dos quais desapareceria o efeito da condução de calor axial no interior da tira. Os resultados mostraram a possibilidade de substituição da condição de contorno após o forno, passando-se então a considerar uma seção transversal da tira situada próxima à saída do forno. A nova condição de contorno foi estabelecida por um balanço de fluxos de calor por condução, convecção e emissão radiativa na referida seção da tira. A partir da análise de influência de cada parâmetro adimensional sobre a distribuição e temperatura da tira, foram selecionados como mais importantes o número de Peclet relativo à taxa de produção do forno, o parâmetro de radiação incidente relativo ao consumo de combustível e parâmetro geométrico que exprime a razão entre as dimensões da tira. Uma análise da eficiência na transferência de calor foi então realizada, avaliando os citados parâmetros diante dos requisitos de obtenção de uma determinada temperatura média da tira e de uma elevada homogeneidade de temperatura em sua reação transversal localizada próxima à saída do forno. Ao promover-se a divisão do forno em três zonas e a adoção de diferentes fluxos de radiação incidente, foi possível simular situações operacionais típicas que ocorrem nos fornos de reaquecimento de placas de uma usina siderúrgica e efetuar comparações com dados disponíveis em literatura. / [en] In the present thesis the energy equation is solved for the temperature distribution in a two dimensional steel strip, which moves with Constant velocity through a furnace and is submited to several different radiation and convection fluxes. Steady state conditions are supposed to exist, if the reference frame is attached to the furnace walls. Thus, one more term arises in the energy equation to take into account the strip movement. The non-linearity of the problem is restricted to the radiation boundary conditions. The discretized algebraic equations are set up, utilizing a finite difference, scheme with the control volume approach. The complete set of equations has shown that there are seven non-dimensional parameters to be specified. Due to the elliptic nature of the equation, it was stabilished that at some distance before and after the furnace, the axial heat conduction in the strip would vanish. A first attempt was done in order to find these distances. The results have shown that the boundary condition after the furnace, the axial heat conduction in the strip would vanish. A first attempt was done in order to find these distances. The results have shown that the boundary condition after the furnace could be substituted, taking a section of the strip closer to the furnace outlet and simply stating the heat flux balance in terms of conduction, convection and radiant emission, considering that this is an end of the strip. Examining the resultant temperature profiles and the heat exchanged by the strip, the influence of each non-dimensional parameter was investigated. Three of them are believed to be the more important: the péciét number related to the furnace rate of production, the incident radiation parameter related to the rate of fuel consumption and the geometrical aspect ratio of the strip. A thermal abalysis of the furnace was then carried out, in order to achieve prescribed data on averade temperature of the strip at the furnace outlet, degree od temperature homogenelty in the strip and heat transfer efficiency. To make a more realistic simulation of the furnace, it was also subdivided into three cones of different incident radiation fluxes and then typical operational situations could be reproduced, as they usually occur in slab reheating furnaces. These simulations were compared with data available in the literature, showing a reasonable qualitative accordance with them.

[pt] Investigou-se teórica e experimentalmente o comportamento dos equipamentos de transferência de calor e massa conhecidos como Torre de Resfriamento e Lavador de Ar. Para a torre, analisa-se um modelo de simulação, a partir de equações diferenciais ordinárias obtidas através de balanços de energia e massa em um elemento infinitesimal de volume, baseado em propriedades uniformes em uma mesma seção transversal, sendo, portanto unidimensional. O sistema resultante é resolvido numericamente. Já para o lavador de Ar, obtem-se, através dos mesmos balanços juntamente com a hipótese de temperatura da água constante, uma solução analítica. Montou-se, com o intuito de validação do modelo, um sistema experimental onde os parâmetros de interesse pudessem ser medidos e desta forma comparados com os resultados numéricos. A concordância entre valores foi bastante boa, para a faixa investigada. Este mesmo sistema permitiu que se obtivessem dados reais a respeito da operação tanto da torre como do lavador, e como consequência a determinação do produto coeficiente de transferência de massa pela área de troca (Hd Av), para cada caso. Finalmente, apresentam-se correlações para o cálculo deste parâmetro como função das vazões de ar e da área da seção transversal, as quais podem ser utilizadas em projetos futuros destes equipamentos. / [en] In the present work the cooling tower and the Air Washer were investigated, experimentally and theoretically. The one-dimensional model developed to simulate the cooling tower operation was based on mass and energy balances written for a differential colume. The resulting system of differential equations was solved numerically. For the Air Washer, the energy and mass balances, incorporating the assumption of constant water temperature, led to a system of equations wich was solved analytically. The experiments conducted were aimed at validating the models developed. The experimental results obtained were in very good agreement with the predictions, within the range of parameters investigated. Correlations were developed relating average mass transfer coefficient for the equipments with the water and air flow rates.

[pt] O presente trabalho investigou numericamente e experimentalmente as características de transferência de calor e queda de pressão de um escoamento espiralado, turbulento, de caindo através de um duto retangular de elevada razão de aspecto. A condição de escoamento espiralado na estrada do duto foi obtida por meio de tubos paralelos contendo fitas torcidas de modo a produzir vórtices girando em sentidos opostos, dois a dois. Coeficientes de transferência de calor locais e médios foram determinados através da utilização da técnica de sublimação de naftaleno em conjunto com a anologia entre processos de transferência de calor e massa. Os valores locais foram medidos sobre toda a superfície ativa do duto, com o auxílio de uma mesa de coordenadas controlada por microcomputador. Resultados para os coeficientes transferência de calor e da queda de pressão foram obtidas para três valores do número de Reynolds do duto e para três valores da intensidade do escoamento espiralado, dados pela utilização de fitas torcidas com diferentes relações de passo/diâmetro. Para posições axiais próximas da entrada do duto, os resultados revelaram altas taxas de aumento na transferência de calor, relativamente ao caso base, representado por escoamento turbulento sem a imposição do escoamento espiralado. Os resultados para a queda de pressão demonstraram que a presença da componente tangencial da velocidade do escoamento espiralado imposto reduziu os efeitos de recuperação de pressão existentes na região de entrada de dutos com entrada abrupta. Foi também verificado que o comprimento de desenvolvimento hidrodinâmico do escoamento aumenta com a intensidade do escoamento espiralado. As equações de conservação de massa , movimento linear e energia, incorporadndo o modelo de turbulência K – E foram resolvidas numericamente para a configuração em estudo. Os resultados numéricos apresentaram boa concordância com os experimentos, permitindo a / [en] The presente work investigated experimentally and numerically the heat transfer and pressure drop characteristics of turbulent decaying swirl flows through a rectangular duct of high aspect ratio. The swirl flow inlet condition was obtained by a set of parallel tubes with twisted-tape inserts wich produced pairs of counterrotating vortices. Local and average heat transfer coefficients were determined by the utilization of the naphthalene sublimation technique in conjunction with the analogy between heat and mass transfer. Local result were obtained along the whole active surface of the duct, utilizing a computer-assisted coordinate table and depth gage. The study encompassed the investigation of the local and average heat transfer coefficient distribution and pressure drop, for three values of the duct Reynolds number, and for three values of the swirl intensity given by different tape pitch-to-diameter rations. The results showed regions of high heat transfer augmentation situated close to the entrance of the duct, when compared to the base case results characterized by turbulent flow through a duct without the swirl flow inlet condition. The pressure drop results demonstrated that the presence of the tangencial velocity component of the imposed swirl flow was seen to reduce the pressure recover effect present in sharp-edged entrance ducts. It was also verifyied that the hydroninamic developing length increases with the intensity of the swirl flow. The equations governing conservation of mass, linear momentum and energy were solved numerically for the configuration investigated. The K-E turbulence model was employed. The numerical results displayed good agreement with the experiments and provided additional information related to velocity and temperature fields which complemented the experimental program.

Les aliments à base de plantain, connu comme riche en amidon résistant, sont largement consommés dans les pays du sud. Le contrôle de la digestibilité de l'amidon du plantain cuit implique la maîtrise de ses propriétés physicochimiques qui varient en fonction des conditions de cuisson et qui résultent de processus de transport (énergie, eau), de transformations et de réactions. Ces travaux de recherche ont pour but de mieux comprendre ces processus. Pour y parvenir, l'effet combiné de la température (T) et de la teneur en eau (X) sur le degré de gélatinisation (α) a été mesuré par DSC. Les proportions d'amidon rapidement digestible (RDS) et résistant (RS) mesurées par une méthode enzymatique in vitro, ont été étudiées sur un aliment modèle (farine + eau) à partir des variétés Dominico Harton et Harton. Les deux diagrammes d'état de l'amidon ont été représentés par un modèle de Weibull. Les effets combinés de T et X sur RDS ou RS peuvent être regroupés dans α; quel que soit l'historique thermohydrique de l'amidon. Ainsi, deux relations adimensionnées sont proposées pour prévoir le RDS et le RS en fonction uniquement d'α. Les écarts de comportement (gélatinisation et digestibilité) de l'amidon de l'aliment modèle et d'un morceau de plantain ont été mesurés et attribués aux effets de structures. Le modèle de prédiction de la digestibilité a été couplé à un modèle 2D traduisant les transferts d'énergie et d'eau durant la cuisson de cylindres de plantain. Le modèle simule bien les profils de température, de taux de gélatinisation et de gain en eau. Une diffusion de type capillaire a été observée en début de cuisson à 50 °C alors qu'au-delà (75 – 100 °C), le transport d'eau s'opère par diffusion moléculaire. La digestibilité est essentiellement pilotée par les transferts d'énergie. Ainsi pour du plantain prêt à être consommé, les degrés de liberté pour moduler le RDS et le RS sont limités ; en revanche, pour des produits pré-gélatinisés, des cuissons à températures intermédiaires (75 – 80 °C) permettraient d'obtenir des produits à faible impact glycémique. Toutefois, la sélection variétale du type de plantain est certainement le moyen en amont le plus souple pour contrôler la digestibilité des produits.Mots clés : plantain, amidon, gélatinisation, digestibilité, modélisation. / Plantain-based foods are widely consumed in southern countries. Plantain is well known for its high resistant starch content. Controlling the digestibility of cooked plantain starch involves managing its physicochemical properties. These vary depending on the cooking conditions and are a consequence of transport processes (energy, water), transformation and reactions. This research work aims to further our understanding of the different processes. To achieve this, the combined effect of temperature (T) and water content (X) on the degree of starch gelatinization (α) was measured by DSC. The proportions of Rapidly Digestible Starch (RDS) and Resistant Starch (RS) were measured using an in vitro enzymatic method and studied using a model food (flour + water) based on Dominico Harton and Harton varieties. Both starch state diagrams where represented using a Weibull model. The combined effects of T and X upon RDS or RS can be summarized by α, irrespective of the thermohydric history. Thus, two dimensionless relations were proposed to predict RDS and RS as a function of α. The difference in the starch behaviour (gelatinization and digestibility) between the model and a piece of plantain were measured and determined as structural effects. The digestibility prediction model was coupled with a 2D model that predicts the energy and water transfer during the cooking of the plantain cylinders. The model provides satisfactory simulations for temperature, degree of gelatinization and water gain profiles. A capillary-type diffusion was observed at the start of cooking at 50 °C, whereas at higher temperatures (75 – 100 °C), water transport is controlled by molecular diffusion. Digestibility is essentially driven by energy transfers. Thus, for ready-to-eat plantain, the freedom to modulate RDS and RS is limited. However, for pre-gelatinized products, products with a low glycemic impact could be obtained at intermediate cooking temperatures (75 – 80 °C). Prior selection of the plantain variety is certainly the easiest way to control product digestibility.Key words: plantain, starch, gelatinization, digestibility, model.

With the development of electricity transmission systems, more and more insulators have been used in the system to isolate the high potential conductors from the earth. In order to secure the reliability of power system, it is necessary to understand the ageing mechanism of the insulators. A model for energy (heat) transfer between low current discharges and insulation surfaces has been built in this project which contributes to the understanding and prediction of insulator ageing. Within the experimental work, the temperature of low current discharges under different conditions has been measured with the “Best-fit” method. The temperature was found to vary between 1200 K and 2300 K with current level, sub cycle duration, gap length, electrode conductivity and polarity of DC discharges. The discharge temperature increases with the growth of current and sub cycle duration by means of the elevation of energy, while the discharge temperature is hotter between salt water droplets than between tap water droplets even though the former has a lower energy (I^2 Rt). The temperature is insensitive to a change of gap length and a positive discharge has a higher temperature than an equivalent negative one. The specific measurement shows the middle part is hotter than the electrode areas. Within the simulation work, the presence of water droplet(s) on the insulation surface concentrates and enhances the electric field over the surface which increases the risk of partial discharge. For practical insulators, this means the core, rather than the sheds, suffers more from discharges between water droplets as they are aligned to the electric field. This highlights the importance of keeping the core dry. The simulation of heat transfer between low current discharges insulation surface was achieved using COMSOL software. The simulated results show that the surface temperature increases rapidly in the first few seconds and arrives at a thermal equilibrium state after 20 seconds of discharge, which meets the experimental observation. The insulation surface temperature distribution under AC discharge is symmetric and the surface centre has the highest temperature which decreases towards the water droplets. In DC, the surface around the cathode is the hottest and the anode area is the coldest.