This thesis is concerned with the design and evaluation of a small scale solarthermoelectric power generation system. The system is intended for electricity generation and thermal energy supply to small scale applications in developing countries of the sunny equatorial regions. Detailed design methodologies and evaluations of both the thermoelectric device and the solar energy collector, which are parts of the combined system, are presented. In addition to experimental evaluations, three theoretical models are presented which allow the design and evaluation of both the thermoelectric module and the solar energy collector. One of the models (a unified thermoelectric device model) concerns the geometrical optimization and performance prediction of a thermoelectric module in power generation mode. The model is unified in the sense that it accounts for the effect of all the parameters that contribute to the performance of the thermoelectric module, a number of which are ignored by the available design models. The unified model is used for a comparative evaluation of five thermoelectric modules. One of these is commercially available and the others are assumed to have optimum geometry but with different design parameters (thermal and electrical contact layer properties). The model has been validated using data from an experimental investigation undertaken to evaluate the commercial thermoelectric module in power generation mode. Results showed that though the commercially available thermoelectric cooling devices can be used for electricity generation, it is appropriate to have modules optimized specifically for power generation, and to improve the contact layers of thermoelement accordingly. Attempts have also been made to produce and evaluate thermoelectric materials using a simple melt-qucnching technique which produces materials with properties similar to those of the more expensive crystalline materials.

The effect of global warming is a growing concern for the global community. This concern is reflected in politics, business, corporate charters, local government charters, electronic and print media and so on. The core to this issue is green house gas (GHG) emissions due to anthropogenic activities. In a developed country such as Australia, residential green house gas emissions are responsible for about 20% of its total GHG emissions. Therefore, sustainability in the housing sector is important towards overall reduction of GHG emissions. Australia's progress in sustainable housing is discussed. There are legislation and financial incentives towards sustainable housing. All the states and territories and the Commonwealth provide financial incentives for PV energy, rain water tank, solar hot water systems to supplement high initial set up cost. A number of high quality rating tools are developed in Australia to facilitate and administer energy efficient design for residential and commercial buildings. There are many good examples of sustainable housing throughout Australia. The Ecohome at Cairnlea, Melbourne is a demonstration home as well as part of this research project. This standard home added with 'off the shelf' sustainable features and having a FirstRate star rating of 6 stars is intended for the volume home market. This thesis presents sustainability outcomes of the Ecohome. Thermal performance of the Ecohome is presented qualitatively and quantitatively. Monitoring data from 14 months' show that approximately two-thirds of the time, indoor temperatures remained in comfort zone (18-26 C) without artificial heating or cooling. Monitoring of indoor air quality included carbon dioxide, humidity and carbon monoxide. Monitoring data from sealed house and while the residents were living in are presented. Humidity level was within 35-45% most of the time while carbon dioxide was under recommended level (1080ppm) in the lower floor (living area and kitchen). Carbon monoxide level within the house was negligible. Significant energy and water savings were realised in the Ecohome. On per capita basis, 45% savings in water usage, 30% savings in electricity usage while similar gas usage was observed compared to average Melbourne home. Residents' feedback on sustainable features was mostly positive (except sisal carpet). Several thermal performance indicators are proposed and presented. These include 'Attenuation Factor', 'Time Lag', 'Heating and Cooling Need' in degree hours, 'Degree C Warmer' and 'Percentage Time in Comfort Zone'. Some of these concepts were used by some authors with different nomenclature; therefore, an attempt was made to unify them. Monitoring data was utilized to measure thermal performance of the Ecohome using these indicators. Prediction of indoor temperature when outside temperature is known is of much interest. Statistical and empirical methods were employed for this purpose. Both statistical method and Givoni method produce reasonably good prediction with deviations from actual observations being in the range of 2 to 3 deg C. This research provides valuable monitoring results in this emerging field. Proposed thermal performance indicators are a significant contribution to the body of knowledge.

This thesis describes the influence of thermal mass on the space conditioning energy consumption and indoor comfort conditions of multi-storey buildings with concrete, steel and timber structural systems. The buildings studied were medium sized educational and commercial buildings. When calculating a building’s life-cycle energy consumption, the construction materials have a direct effect on not only the building’s embodied energy but also on the space conditioning energy. The latter depends, amongst other things, on the thermal characteristics of the building’s materials; thermal mass can also be an influence on comfort conditions in the building. A modelling comparison has been undertaken between three very similar medium-sized buildings, each designed using structural systems made primarily of timber, concrete and steel. The post-tensioned timber version of the building is a modelled representation of a real three-storey educational building that has been constructed recently in Nelson, New Zealand. The concrete- and steel-structured versions have been designed on paper to conform to the required structural codes and meet, as closely as possible, the same performance, internal space layout and external façade features as the real timber-structured building. Each of these three structurally-different buildings has been modelled with two different thermal envelopes (code-compliant and New Zealand best-practice) using a heating, ventilating and air conditioning (HVAC) system with heating only (educational scheme) and heating and cooling (commercial scheme). The commercial system (with cooling) was applied only to the buildings with the best-practice thermal envelope. The analysis of each of these nine different construction and usage categories includes the modelling of operational energy use with an emphasis on HVAC energy consumption, and the assessment of indoor comfort conditions using predicted mean vote (PMV). From an operational energy use perspective, the modelling comparison between the different cases has shown that, within each category (code-compliant, low-energy and low-energy-commercial), the principal structural material has only a small effect on overall performance. The most significant differences are in the building with the best-practice thermal envelope with the commercial HVAC system, were the concrete building has slightly lower HVAC energy consumption, being 3 and 4% lower than in the steel and timber buildings respectively The assessment of indoor comfort conditions during occupied periods through using PMV for each of the three categories shows that the timber structure consistently exhibited longer periods in the over-warm comfort zone, but this was much less pronounced in south-facing spaces. To examine the reasons for the less acceptable PMV in the timber-structure versions, an analysis of indoor timber and concrete surface temperatures was carried out in both buildings. It was found that, particularly in north-facing spaces, there were large diurnal swings in the temperatures of timber surfaces exposed to solar radiation. These swings were much less in the case of concrete surfaces so the environment was perceived to be more comfortable under such conditions because of the reduced influence of higher mean radiant temperatures. To moderate this potential downside of solar-exposed internal timber surfaces, better results are achieved if, when timber is used for thermal mass, the timber is not exposed to direct solar radiation, for example locating it in the ceilings or on the south side of the building. Two other approaches to combating the potential overheating problem in the timber-structured buildings were analysed in an illustrative mode; addition of external louvres to reduce direct solar gains at critical times of day and year; and use of phase change material (PCM) linings to act as light-mass energy buffers. Although external louvres increase comfort conditions significantly by reducing the periods of an overly warm environment, they produce an increase in heating energy consumption through reducing beneficial solar gains. The use of PCM linings shows little benefit to overall indoor comfort conditions for the building of this case-study.

This thesis is concerned with measuring the fabric thermal performance of houses. This is important because the evidence shows that predictions of performance, based upon a summation of expected elemental performance, are prone to significant inaccuracy and in-situ performance is invariably worse than expected the so-called performance gap . Accurate knowledge of the thermal performance of houses could cause a shift in the way that houses are built, retrofitted and managed. It would enable quality-assurance of newly-built and retrofitted houses, driving an improvement in the energy performance of the housing stock. The current barrier to achieving these benefits is that existing measurement methods are impractically invasive for use on a mass-scale. The aim of this research is to address this issue by developing non-invasive fabric thermal performance measurement methods for houses. The co-heating test is currently the most used method for measuring whole-house fabric thermal performance; it is used to measure the Heat Loss Coefficient (HLC) of a house, which is a measure of the rate of heat loss with units of Watts per degree Kelvin. It has been used extensively in a research context, but its more widespread use has been limited. This is due to a lack of confidence in the accuracy of its results and the test s invasiveness (the house must be vacant for two weeks during testing, which has so far been limited to the winter months, and testing cannot be carried out in newly-built houses for a period of approximately one year due to the drying out period). To build confidence in the results of co-heating testing, the precision with which test results can be reported was determined by the combination of a sensitivity analysis to quantify measurement errors, and an analysis of the reproducibility of the test. Reproducibility refers to the precision of a measurement when test results are obtained in different locations, with different operators and equipment. The analysis of the reproducibility of the test was based upon a direct comparison of seven co-heating tests carried out by different teams in a single building. This is the first such analysis and therefore provides a uniquely powerful analysis of the co-heating test. The reproducibility and sensitivity analyses showed that, provided best practise data collection and analysis methods are followed, the HLC measured by a co-heating test can be reported with an uncertainty of ± 10%. The sensitivity analysis identified solar heat gains as the largest source of measurement error in co-heating tests. In response, a new approach for co-heating data collection and analysis, called the facade solar gain estimation method, has been developed and successfully demonstrated. This method offers a clear advancement upon existing analysis methods, which were shown to be prone to inaccuracy due to inappropriate statistical assumptions. The facade method allowed co-heating tests to be carried out with accuracy during the summer months, which has not previously been considered feasible. The demonstration of the facade method included a direct comparison against other reported methods for estimating solar gains. The comparison was carried out for co-heating tests undertaken in three buildings, with testing taking place in different seasons (winter, summer, and spring or autumn) in each case. This comparison provides a unique analysis of the ability of the different solar gain estimation methods to return accurate measurements of a house s HLC in a wide variety of weather conditions. Building on these results, a testing method was developed: the Loughborough In-Use Heat Balance (LIUHB). The LIUHB is a non-invasive measurement method, designed and tested in this study, which can measure the HLC of a house with an accuracy of ± 15% while it is occupied and used as normal. Measurements of energy consumption and internal temperature are discreetly collected over a period of three weeks, and combined with data collected at a local weather station to inform an energy balance, from which the HLC is calculated. This low impact monitoring approach removes the barriers to fabric thermal performance testing on a mass scale. The LIUHB has been successfully demonstrated in several comparative trials versus a baseline measurement provided by the co-heating test. The trials have included the application of extreme examples of synthetic occupancy conditions, testing in an occupied house, and quantification of the effects of a retrofit. Subject to further validation, the LIUHB has the potential to deliver many of the benefits associated with mass-scale measurement and quality assurance of housing performance.

Gas turbine vanes find themselves in very hostile environments – extremely high temperature combustion gases, much exceeding material melting temperatures, flowing over them at enormous pressures. It is necessitated due to the increased efficiency and power output at these conditions. However, this also means that, in spite of the technological advancements made, these parts need frequent repairing compared to parts placed in milder environments. Primarily due to economic reasons, gas turbine parts are repaired by companies other than the original equipment manufacturer (OEM). While multitude of condition monitoring techniques have been developed and are used in the industry for regular maintenance checks, there is no easy way to characterize the impact on thermal performance of the repairing processes involved. This thesis reports the development of a technique to address this issue. It also chronicles the test rig design, experiments conducted, development and significance of the thermal performance metric. Heated air (250 ̊C – 300 ̊C) is flown through the internal cooling passages of 8 samples each of OEM and repaired parts at two different pressure ratios (vane inlet over ambient pressure), 1.1 and 1.3. First, steady state mass flow rates through each airfoil (one part is a cluster of 4 airfoils) is experimentally determined and compared among the OEM and repaired sample sets. Second, a transient experiment is run and the surface temperatures of the airfoils are measured using multiple infrared cameras viewing both the pressure and suction side of the airfoils. A parameter involving localized vane surface temperature, airfoil inlet temperature and ambient temperature is formulated to characterize the vane thermal performance. Using statistical analysis, it is found that there is no significant difference between the OEM and repaired samples tested. The development of the discussed technique, it is expected, will help companies in the gas turbine vane repairing business to qualify their parts in a robust and efficient manner without the need to invest a lot of money in buying precision equipment, or, control chambers. Finally, a couple of further studies are recommended to further improve the qualifying procedure and thereby increase the efficiency of the technique. / Master of Science / Most manufactured parts, during its lifetime, go through wear and tear of some form. Some much more than others – a gas turbine vane is one example, owing to the hostile environments it finds itself in. While repairing turbine vanes make economic sense instead of replacing the worn-out vanes with new ones, due care must be taken to ensure that the repairs pass high quality standards of the original manufactured parts. Most, if not all, companies in the turbine repairing business rely on room-temperature air-flow testing through the internal passages of these vanes to qualify their repaired parts. This is done partly due to the complexity in replicating engine-like conditions in a test environment in addition to being very time-intensive. While room-temperature air flow comparison between repaired and original parts is a necessary test, it does not paint the whole picture. Thermal performance, or, how the vane exchanges heat with the surrounding media, is the other part which completes the puzzle. A plurality of techniques has been developed to ascertain the thermal performance of gas turbine vanes, however, these are limited in the scope of their applicability – the reason why industry is still mostly relying on airflow measurements for their part qualification. In this study, a new technique has been proposed which is agnostic of the unavoidable variations in operating conditions and easy to apply while still upholding high quality standards. This translates to huge savings to organizations which are in the business of repairing original parts, not necessarily restricted to gas turbine industry.

Published Article / The purpose of this paper is to analyse the thermal performance of two buildings. The one has a large thermal mass and the other a highly insulated low thermal mass. A typical 120 m2 suburban building was modelled in Ecotect. As part of the model infiltration rate, wind sensitivity and a central Pretoria weather file were used. New material composites were introduced in the materials database to represent typical building materials used in the construction of heavy and light-weight buildings in South Africa. The thermal characteristics of these new materials were then calculated within Ecotect. Ecomat was used to calculate thermal lag which was used as an additional input into Ecotect. The research indicates that a low thermal mass and highly insulated building have been shown to use 18.3% less annual space heating and cooling energy when compared to the high thermal mass building. The good thermal performance results of the light-weight building will help in clearing scepticism to adopting this construction technology in southern Africa where high thermal mass masonry is still predominant.

At present there are concerns that new housing in the UK under-performs in terms of energy-efficiency. In research studies where the thermal performance of the building fabric has been measured, post-construction, significant gaps between design predictions and the as-built performance of new housing have been found. This thesis is concerned with how thermography may be used to assess the thermal performance of new housing during the construction process. The practical utility of this testing approach, which has not been investigated in-depth by previous research, is to identify performance issues at a stage when undertaking remedial work is less costly and disruptive. Moreover, by identifying issues that could reduce the energy-efficiency of the building fabric, these tests could help address the observed "performance gap". An approach for using thermography at different stages of construction has been developed through practical case studies and experimental work. The complementary use of heat transfer modelling and thermography is also explored. Finally, through interviews with industry professionals, the context of implementing tests within UK housing development practices is examined. The results of applying the testing approach demonstrate that beneficial feedback can be obtained without significant interruption to construction activities. However, test procedures need to be modified according to the stage and method of construction. It is proposed that heat transfer modelling can inform the interpretation of test results and analysis of defect severity. For the testing approach to be implemented more widely, there would need to be further development and trials on different types of construction and the oversight of an independent body to establish its credibility. In conclusion, thermography has the potential alongside other testing and inspection practices to help improve standards of site construction. However, there are limitations to its use which need to be considered if such tests are to be implemented effectively.

Thesis (MEng)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: In this study, one and two-dimensional models are developed for the evaluation of the thermal performance of a delugeable flat tube bundle to be incorporated in the second stage of an induced draft hybrid (dry/wet) dephlegmator (HDWD) of a direct air-cooled steam condenser (ACSC). Both models are presented by a set of differential equations. The one-dimensional model is analysed analytically by using three methods of analysis which are: Poppe, Merkel, and heat and mass transfer analogy. The two-dimensional model is analysed numerically by means of heat and mass transfer analogy method of analysis whereby, the governing differential equations are discretised into algebraic equations using linear upwind differencing scheme. The two-dimensional model’s accuracy is verified through a comparison of the two dimensional solutions to one dimensional solutions. Satisfactory correlation between the one and two-dimensional results is reached. However, there is a slight discrepancy in the solutions, which is mainly due to the assumptions made in one-dimensional model. The effect of tube height, tube pitch, tube width, deluge water mass flow rate, frontal air velocity, steam, and air operating conditions on the heat transfer rate and air-side pressure drop for both wet and dry operating modes are investigated. The long tube height, large tube width, small tube pitch, and high frontal air velocity are found to increase the tube bundle’s performance. However, this performance is associated with a high airside pressure drop. The performance of the deluged flat tube bundle is found to be less sensitive to the changes in the deluge water mass flow rate and air operating conditions. Furthermore, the best configuration of a delugeable flat tube bundle is identified through a comparison to round tube bundle presented by Anderson (2014). The performance of the round tube bundle is found to be around 2 times, and 1.5 times of that of flat tube bundle, when both bundles operate as an evaporative and dry air-cooled condenser respectively. / AFRIKAANSE OPSOMMING: In hierdie studie is een en twee-dimensionele modelle ontwikkel vir die evaluering van die termiese prestasie van 'n benatbare plat buis bundel in die tweede stadium van 'n geïnduseerde ontwerp hibriede (droë / nat ) deflegmator van 'n direkte lugverkoelde stoom kondensator. Beide modelle is aangebied deur 'n stel van differensiaalvergelykings. Die een-dimensionele model is analities ontleed deur die gebruik van drie metodes van analise wat: Poppe, Merkel, en die hitte en massa-oordrag analogie. Die twee-dimensionele model is numeries ontleed deur middel van hitte en massa-oordrag analogie metode van analise waardeur , die regerende differensiaalvergelykings gediskretiseer in algebraïese vergelykings met behulp van lineêre windop differensievorming skema. Die tweedimensionele model se akkuraatheid is geverifieer deur 'n vergelyking van die twee dimensionele oplossings te een dimensionele oplossings. Bevredigende korrelasie tussen die een en twee-dimensionele resultate bereik word. Maar daar is 'n effense verskil in die oplossings, wat is hoofsaaklik te wyte aan die aannames wat gemaak in een-dimensional model. Die effek van buis hoogte, buis toonhoogte, buis breedte, vloed water massa-vloeitempo, frontale lug snelheid, stoom, en in die lug werktoestande op die hitte oordrag snelheid en lug - kant drukval vir beide nat en droë maatskappy modi word ondersoek. Die lang buis hoogte, groot buis breedte, klein buisie toonhoogte, en 'n hoë frontale lug snelheid gevind die buis bundel se prestasie te verhoog. Tog is hierdie prestasie wat verband hou met 'n hoë lug - kant drukval. Die prestasie van die oorstroom plat buis bundel gevind word minder sensitief vir die veranderinge in die vloed water massa-vloeitempo en lug werktoestande. Verder is die beste opset van 'n benatbare plat buis bundel geïdentifiseer deur 'n vergelyking met ronde buis bundel aangebied deur Anderson (2014). Die prestasie van die ronde buis bundel gevind word om 2 keer, en 1.5 keer van daardie plat buis bundel , wanneer beide bundels funksioneer as 'n damp en droë lugverkoelde kondensor onderskeidelik.

O conceito de desempenho de edificações vem sendo discutido pelo mundo, principalmente, após o período pós-guerra na Europa. No Brasil, as primeiras discussões se iniciaram na década de 1980, quando novas tecnologias construtivas foram introduzidas aos edifícios habitacionais. Na época, os profissionais da área perceberam que se tornava imprescindível avaliar o comportamento esperado dessas edificações durante sua vida útil. Assim, um grupo de estudo liderado pela ABNT se formou para elaborar uma série de normas técnicas relacionadas ao conceito de desempenho de edificações. Após décadas de discussões, publicou-se a norma ABNT NBR 15575/2013 - Edificações habitacionais - Desempenho, em vigor desde julho de 2013. A publicação da norma de desempenho significou um grande marco para o setor da construção civil, pois, nela são contextualizados conceitos de vida útil, prazo de garantia, manutenibilidade, incumbência de cada agente envolvido, durabilidade e estabelecidos níveis de desempenho: mínimo (M), intermediário (I) e superior (S). Para avaliar o desempenho térmico, a norma propõe dois procedimentos: simplificado e medição in-loco. O procedimento simplificado tem aspecto normativo de análise, onde é verificado o atendimento aos requisitos e critérios das propriedades térmicas dos sistemas de vedação e cobertura. No caso de resultados insatisfatórios, a análise segue pelo método da simulação computacional. No entanto, alguns estudos já apontaram discrepâncias e incertezas nos resultados de desempenho térmico relacionados ao método de simulação computacional proposto pela norma, principalmente na configuração do dia típico de projeto. A falta de definições de alguns parâmetros importantes na norma de desempenho, por exemplo, o parâmetro \"Dia Típico de Projeto\" (Design Day), tem causado questionamentos sobre a melhor forma de se aplicar a simulação computacional para determinação do nível de desempenho térmico dos edifícios. Tendo em vista esses questionamentos e a incerteza dos resultados devido à desconsideração de algumas variáveis, o presente trabalho objetiva avaliar a influência das variáveis desconsideradas pela norma de desempenho na configuração do parâmetro Design Day, do software EnergyPlus, nos resultados de desempenho térmico final da edificação. Para isso utilizou-se como modelo uma edificação multifamiliar de HIS, localizada na Cidade de Sumaré no Estado de São Paulo. Para o estudo, inicialmente, as variáveis do parâmetro Design Day (do EnergyPlus) e as equivalentes, constantes da norma de desempenho foram interpretadas. Identificaram-se seis variáveis sem definição clara e sem valores atribuídos pela norma de desempenho que foram consideradas na simulação para verificar a influência das mesmas no resultado do desempenho térmico de edificações. Concluiu-se que as variáveis de Solar Model Indicator e Sky Clearness foram as variáveis que influenciaram nos resultados de nível de desempenho térmico. Desse modo, a pesquisa verificou que os valores atribuídos a essas variáveis possibilitam que os resultados sejam favoráveis ou não. / The concept of building performance has been discussed by the whole world, especially after the postwar period in Europe. In Brazil, the first discussions have begun in the 1980s when new constructive technologies were introduced to residential buildings. At the time, the professionals of this area realized that it became essential to evaluate the expected behavior of these buildings during their useful life. Thus, a study group led by ABNT was formed to elaborate a series of technical norms related to the concept of buildings performance. After decades of discussions, the ABNT NBR 15575:2013 Standard - Residential Buildings - Performance, which was published in July 2013. The publication of the building performance standard meant a great landmark for the civil construction sector, since it includes concepts of useful life, warranty period, maintenance, responsibility of each agent involved, durability and established performance levels: Minimum (M), Intermediate (I) and Superior (S). To evaluate the building thermal performance, the Standard proposes two procedures: simplified and in loco measurement. The simplified procedure is a normative analysis aspect where the requirements fulfillment and the thermal properties criteria of the sealing walls and roofing systems are verified. In the case of unsatisfactory results, the analysis follows by the computational simulation method. However, some studies have already pointed out discrepancies and uncertainties in the thermal performance results related to the computational simulation method proposed by the referred Standard, especially in the typical design day configuration. The lack of some important parameters definitions in the Performance Standard, for example, the \"Design Day\" parameter has caused questions about the best way to apply the computer simulation to determine the level of buildings thermal performance. Considering these questions and the uncertainty on results due the lack of consideration of some variables, the present work proposes evaluate the influence of variables disregarded by the Performance Standard in the Design Day parameter configuration on the EnergyPlus software in the final results of the building thermal performance. A multi-family housing building, located in Sumare City, State of Sao Paulo, was used as a model. For the study, first, the Design Day parameter variables (of the EnergyPlus) and the equivalents variables included in the Performance Standard were interpreted. It was identified six variables with no clear definition and no values attributed by the Standard that were considered in the simulation to verify their influence on the buildings thermal performance. It was concluded that Solar Model Indicator and Sky Clearness were the variables that influenced the results of thermal performance level. In this way, the research verified that the assigned values to these variables allow being favorable results or not.

O objetivo dessa pesquisa foi avaliar o desempenho térmico de telhas através do monitoramento de células-teste existentes na cidade de São Carlos, comparar os resultados com análises espectrofotométricas das amostras (das telhas das células teste) para correlacionar os resultados com normas que fazem alguma referência ao desempenho térmico de telhas. Foram realizadas análises comparativas das temperaturas internas e da superfície das telhas em quatro células teste: com telha de aço, fibrocimento, cerâmica e material reciclado (de embalagem longa vida). Em uma segunda etapa as telhas de aço e fibrocimento foram selecionadas para receberem a aplicação de tinta látex acrílico branco. Foi verificada, após análise comparativa, a importância do tratamento superficial externo em coberturas e a necessidade de normas específicas para o desempenho térmico de telhas. / The purpose of this research was to evaluate the thermal performance of roofs by monitoring existing test cells in the city of Sao Carlos, and comparing the results with the spectrophotometric analyses of the samples (the roofs of the test cells) for correlating the results with the standards associated with the thermal performance of tiles. Comparative analyses were carried out of the internal temperature and the temperature of the surface of the roofs in four test cells: roofs made of steel, asbestos cement ceramic and recycled material (long-life packaging). At the second stage a white latex acrylic paint was applied to the steel and asbestos cement roofs. This confirmed, following a comparative analysis, the importance of the treatment of the external surface of the roofs, and the necessity for specific standards for the thermal performance of roofs.