Behaviour of structural glass at high temperatures

Abstract

This thesis presents a study on the behavior of none-structural and structural glass at high temperatures. The main objective of this master thesis is to investigate the capabilities of different glass types when these are used as a structural element with fire resistant requirements. It is therefore important to develop an understanding of the behaviour of various glass types subjected to fire. A literature study is done on glass properties, glass production and methods for making structural glass. The purpose of this study was to investigate the background information and to gain adequate knowledge about the state of art on the thesis topic. Moreover thermal and mechanical properties of glass were determined. Results of these studies show that all glass properties depend on temperature. Subsequently the finite element model (first DIANA model) was made to simulate a glass test which was performed by others at Empa in Switzerland. A laminated beam model was made in DIANA software. This beam consists of 3 soda lime glass layers and two sentry glass interlayers. This model was used for making a sensitivity analysis about the glass properties which were found during the literature studies. The sensitivity analysis showed that from the material properties uses as input for DIANA models, the heat specific property has the most influence on the glass temperature. As next step, six experiments series were performed at the laboratory of Efectis Netherlands BV in Bleiswijk. The purpose of these tests was to investigate fire behaviour of three different glass types. Annealed, fully tempered and heat strengthened glass were tested horizontally and vertically in different conditions during these experiments. Results of each test were analyzed and discussed before performing the next test. This way, the knowledge about the fire behaviour of the glass and the right method for test performing was developed step by step during this research. Annealed glass performance was very weak at high temperatures due the thermal break phenomenon. It seems to imply that annealed glass (without any improvement) does not have the capabilities to be used as a fire resistant structural element. Based on the different behavior of steel and glass at the high temperature, can be noted that the thermal break phenomenon is occurred at the joint locations (connection between glass and steel support). For this reason joint details need extra attentions if glass needs to be used as a fire resistant structural element. The tests observations shows that 550-600°C is the critical temperature for both fully tempered and heat strengthened glass types. The average furnace temperature was around 615°C when glass temperature reached the critical temperature (550°C). The glass behavior changed to plastic phase at the critical temperature. Therefore the glass was not solid anymore and behaved more like liquid material. Afterwards the glass was not able to carry any loads, even not the self-weight. Through the data analysis and the observations of test series 1 to 6 seems to imply that fully tempered and heat strengthened glass, with some improvements, have the capabilities to be used as a structural material under fire loads. It should be mentioned that the relation between maximum allowable tensile stress and temperature, need to be determined first. This information is essential in order to determine if fully tempered glass can be used as fire resistant element. Test setups and the results of the tests are discussed in the chapter “Tests at Efectis” in more details. Single glass layer is tested during all the experiments series. The results show that one layer was not sufficient. This means that using laminated glass is essential in order to have enough fire resistance. The interlayers and the outside glass layers can keep the temperature of the middle layer(s) (loadbearing part) below the critical temperature for certain amount of time. Second DIANA model was made for one of Efectis test in order to see if the finite element model could give a correct estimation of glass behavior at high temperatures. The model results were close to the ones found during the tests. This means DIANA model has the capabilities to predict glass behavior at high temperature. In order to be able to make a model with more precision, glass properties need to be determined. For this reason several material tests should be performed at the beginning of the next research. Metal wires are used for the deflection measuring. The results show that these wires are extended considerably during the performed tests. For this reason this method is not useful for a correct deflection measuring. A new measuring method should be determined in order to have no physical connection between measurement equipment and glass. This way makes measuring glass deflection directly (without any instrument like metal wire) possible. Thermocouples were used for measuring glass temperature surface. These thermocouples were installed on the glass surface. Thermocouples can be placed between layers of laminated glass during the glass production. This way makes more precision temperature measuring possible.