Hrvatski

Doctoral, master's and specialist theses written and/or mentored by the staff of the Chair for Metal Structures

Fuštar, Boris: FATIGUE LIFE ASSESSMENT OF HIGH FREQUENCY MECHANICAL IMPACT TREATED WELDED STEEL STRUCTURAL DETAILS, doctoral thesis, mentor: Asst. Prof. Ivan Lukačević, PhD., Faculty of Civil Engineering, University of Zagreb, 2023.

Fatigue is a progressive and localised process of damage accumulation in the material due to cyclic stresses. In the case of welded steel structures, fatigue damage occurs within the welded details, representing geometric and material discontinuities. Sudden changes in geometry cause high-stress concentrations, which reduces the fatigue life of welded steel details. The welding process causes tensile residual stresses in the weld area, which can further reduce fatigue life. Despite numerous design guidelines and recommendations, the satisfactory fatigue life of welded steel details is not always possible to achieve. Therefore, Post Weld Treatment methods, such as the High-Frequency Mechanical Impact treatment, have been developed, extending their fatigue life. This is achieved by improving the weld geometry, reducing the stress concentration, introducing compressive residual stresses, and increasing the hardness at the location of the treatment. Within the scope of this research, deterministic and stochastic calculation models were developed and calibrated, which enable reliable estimation of the fatigue life of untreated and treated welded details. Laboratory tests and numerical analysis were used to determine fatigue life improvement parameters and the model calibration was performed based on the laboratory cyclic tests.

Janjić, Mario: OPTIMAL FIRE PROTECTION SYSTEM FOR STEEL STRUCTURE OF WAREHOUSE, specialist thesis, mentor: Prof. Davor Skejić, PhD., Faculty of Civil Engineering, University of Zagreb, 2022.

Specialist thesis carried out analysis of conditions for construction of warehouses prescribed by Croatian Fire Safety Regulation in Warehouses and it was compared to conditions given in American Standard NFPA 1. The key differences relate to the fire resistance of the structure, the obligation to install a smoke and heat removal system and a fire alarm system. The NFPA does not prescribe the fire resistance of the structure, while the Croatian Regulation for warehouses without sprinkler installations requires it for a duration of 30 minutes. The load-bearing steel structure of the analysed warehouse is made of seven portal frames with a span of 14,30 m placed at a distance of 4,90 m. The height of the structure in the eave is approx. 4,90 m, while the height in the ridge is approx. 5,80 m. The columns and span structure are made of S355 quality steel, while the rest of the structure is made of S235 quality steel. It was analysed storing of five different types of goods inside the same steel warehouse. For every type of good were calculated fire loads, the speed of fire development was determined and a realistic fire curve temperature-time was modelled. The load-bearing capacity check of the previously dimensioned steel elements of the building in “cold state” was conducted in relation to ISO 834 standard fire curve and to the realistic fire curves as well, depending on the expected fire load, that is the purpose of the warehouse itself. In four out of five realistic fire scenarios, the load-bearing structure satisfied the load-bearing condition for 30 minutes. Parts of the structure failed in the load-bearing capacity of the standard fire, while in the case of a textile goods fire, the entire structure failed. The analysis of fire scenarios was determined the factors, which affect the most on the fire resistance of the steel structure without necessary additional protection. For steel structure that in certain fire scenarios was not satisfied load-bearing analysis, it is proposed several types of structure protection. Active measures include fire sprinkler system, while passive measures predict fireproof coatings, fire resistant boards and heat resistant plaster. The advantages and disadvantages of certain types of protection, suitability for installation in storage areas were determined and an economic analysis was carried out. The cheapest but least suitable protection is heat resistant plasters, and as an optimal solution, protection with fireproof coatings is adopted.

Ivan Ćurković: BEHAVIOUR OF COMPOSITE PLATE SHEAR WALLS UNDER EARTHQUAKE, doctoral thesis, mentor: Assoc. Prof. Davor Skejić, PhD., Faculty of Civil Engineering, University of Zagreb, 2017.

In addition to the traditional vertical stabilisation systems used, existing knowledge, related to joint action of steel and concrete, allowed development of shear walls composed of frames and composite plate infills. Such systems enable better structural performance and in the same time provide economic viability due to less damage occurrence as a consequence to seismic response. Due to lack of insight into the behaviour of these systems design provisions, if provided at all, are quite limited or reference provisions provided for similar systems with steel infills. This is the case with requirements on column stiffness of composite steel plate shear walls which are designed for the formation of complete tension field within the infill. Due to fact that load-bearing mechanism of composite steel plate shear walls assumes utilisation of infill plate shear strength, there is reasonable ground to suspect that such requirement is conservative. Therefore, in this thesis an experimental study was conducted in order to establish the actual behavior of single story shear walls either with steel or composite infill depending on three various column stiffness values. The obtained results served as a basis for the development of numerical models. Using calibrated models, numerical parametric analyses were performed to investigate the influence of column bending stiffness as well as the quality of the steel material of the infill panel for various geometrical configurations, i.e. L/H ratios. Based on the obtained results, proposition to reduce the currently valid criterion for column bending stiffness is given. The proposals are given separately for shear walls with steel infills and with composite infills. Finally, modification of expression for the column bending stiffness was proposed in order to include dependence on the material quality of the steel infill plate.

Ivan Lukačević: FATIGUE RELIABILITY EQUALIZATION OF STRUCTURAL STEEL DETAIL CATEGORIES, doctoral thesis, mentor: Prof. Darko Dujmović, PhD., Faculty of Civil Engineering, University of Zagreb, 2014.

Modern structural standards provide engineers in practice a quick and simple design of structures subjected to fatigue. However, studies have shown that such a design approach results in uneven reliability levels of structural details. Research in this thesis deals with nominal stress approach and S-N curves. It is primarily focused on the equalising the reliability levels of various structural details of new steel structures subjected to fatigue. Generally, uncertainties regarding loads and resistance are very significant for fatigue assessment. Uncertainties associated with resistance were analysed using stochastic procedures on existing fatigue test data. As a result, uncertainties regarding resistance are quantified and categorised in specific groups depending on detail categories. Uncertainties of loads were investigated based on field measurements and simulations using numerical models. As a result, the assessment procedure of reliability and durability of existing steel structures subjected to fatigue were established. This procedure enables the reliability assessment of structural details for which is not possible to conduct field measurements. Stochastic models based on quantified uncertainties of the resistance and loads were established. By calibration procedure, considering the differentiation of consequences classes associated with fatigue assessment methods recommended in European standards, the optimal sets of partial factors for fatigue resistance and fatigue loads were calculated. Based on the research conducted in this thesis, refined assessment procedure of reliability, i.e. durability of existing steel structures subjected to fatigue and a new method for the design of steel structures subjected to fatigue which provides uniform reliability levels of detail categories has been proposed.

Anka Javora: RESISTANCE OF BOLTED JOINTS WITH DIFFERENT BLIND BOLT SYSTEM, specialist thesis, mentor: Asst. Prof. Davor Skejić, PhD., Faculty of Civil Engineering, University of Zagreb, 2013.

Numerous structural and architectural reasons favour the use of hollow sections compared to open section columns in structural applications. Cold rolled tubes still have limitations regarding welding, and the limited ability of connecting hollow section and open section elements have been so far an obstacle to their wider application. Blind bolted connections enable jointing of I-beams and hollow section columns. Blind bolts utilise specially developed mechanisms allowing them to tighten on only one side of a connection. Today's most widely used blind bolts are: Huck bolts, Hollo-Bolt, Blind bolts, Ajax ONESIDE. In addition to the listed, one can use a Flowdrill system enabling single sided connecting of elements with the use of standard bolts. Blind bolts are construction products, therefore designer should use bolt tensile resistances given on the basis of tests carried out by the manufacturer. Behaviour of blind bolted connections does not significantly differ from behaviour of connections using standard bolts and connection's resistance does not depend on the hole drilling technique. The tests carried out so far on the beam-to-column joints made with blind bolts show that the most common cause of failure is flange bending of the hollow column.The most accurately assessment of the column face flexibility can be made based on the equation using an elliptical yield line mechanism. Membrane action increases column flange capacity up to 30%. In the first part of a thesis, an analysis of the blind bolted beam to column connection was performed with different blind bolted systems, using nominal bolt diameter of 20 mm, while in the second part of thesis, a parametric analysis of the blind bolted beam to column connection was performed using square hollow section column SHS 200x200 and beam IPE 300. Yielding of column flange was most common failure mode (63%) of 1155 combinations of parameters. Parametric analyses have shown that yielding of column flange is most common failure mode, where bolt hole size, required for the installation, has important impact to connection capacity. In connections using same nominal size of bolts but different required hole sizes, connection will have lower bending resistance if the bigger hole size is required (Hollo-Bolt and Ajax ONESIDE). By applying elliptical yield line model and taking into account the impact of membrane actions, capacity of connection is increased and becomes closer to the experimental values. To gain better understanding of the connection's behaviour it would be necessary to: further examine and verify the share of the membrane action in the total capacity of the connection, verify the compliance of test results with the elliptical yield line model with respect to the membrane actions and further investigate the connection's behaviour in the areas where the change of failure modes occurs.

Ana Komorski: RELIABILITY OF SUSPENSION TRANSMISSION POWER LINE TOWERS, master's thesis, mentor: Asst. Prof. Davor Skejić, PhD., Faculty of Civil Engineering, University of Zagreb, 2011.

The development of the regulations and the standards for the overhead transmission lines in the Republic of Croatia started in the year 1949. Until today, regulations and standards for overhead transmission lines have undergone several changes. The last major change of the ''Regulation on technical standards for construction of overhead transmission lines voltage 1 kV to 400 kV'', was in year 1988. and it is still in application. In the process of joining the European Union, Croatia has committed itself to adopting European standards, among which is the standard for the design of overhead transmission lines ''EN 50341-1 Overhead electrical lines exceeding AC 45 kV''. The same standard was adopted in Croatia in the year 2008. under the heading HRN EN 50341-1. Croatian normative annex is necessary for the same to be applicable and should take into account the experience in designing the overhead transmission lines in our country. As with all constructions, overhead transmission line towers are exposed to various weather conditions that can be one of the potential causes of damage. It is a fact that the overhead transmission line is a line object that often exceeds hundreds of kilometers, so the likelihood of extreme weather conditions affecting it is far greater than on the rest, spot-located, constructions. Over the years, the damages on overhead transmission lines have been reported in some areas of Croatia. However, systematic analysis of possible critical areas has not been done. It is the fact that the map of the damages of the overhead transmission lines provides an overview of the behaviour of the towers in real load conditions. This article presents an originally made map of the damages covering all detected rehabilitation projects of overhead transmission lines of 35 kV to 400 kV voltage in which the damage on overhead transmission line towers was noted. The map covers the period from 1979. to 2010. and it is the result of systematic research of the Dalekovod Project Ltd.’s archive. The background on which the map of the damages is made is a map of the Croatian electric power system for 110, 220 and 400 kV overhead transmission lines. The map was produced by HEP OPS Ltd. and was allowed to be used for the research. Damages were classified according to the causes of damage - damage caused by the wind load, ice load and the combined load of wind and ice loads. Damages from other causes (landslides, tree falls, war, etc.) are not shown on this map. Given the almost passing the Croatian national normative annex necessary for the application of the EN 50341-1, it is interesting to compare the calculations of transmission line towers built in different areas of the Croatia when analysed according to the current Regulation, according to the basic European standard EN 50341 -1, and by using the German normative annex EN 50341-3-4, (herein after only the EN 50341-3-4). Therefore, this paper presents the results of analysis of six supporting towers used in various areas of Croatia. Three towers were selected with lines that have experienced damage, and each of them was damaged by a different load, or a combination of loads. It is assumed that the most reliable calculation is the one in which the simulated behaviour of the overhead transmission line towers coincides with the recorded (set) mode of the collapsed (damaged) tower. On the basis of the analysis, at the end of the paper are given guidelines for the development of Croatian normative annex .

Davor Skejić: RELIABILITY ASSESSMENT OF BEAM-TO-COLUMN JOINTS WITH CLEATS, doctoral thesis, mentors: Prof. Darko Dujmović, PhD.; Prof. Mladenko Rak, PhD., Faculty of Civil Engineering, University of Zagreb, 2009.

The customary design practice shows that the analysis of steel frame structures is still conducted on the assumption that the beam-to-column joints are rigid and full strength. The shortcomings of this approach show themselves in unnecessarily expensive structure which is evident during the design and execution of such details. However, the benefits resulting from the partial strength/semi-rigid design are increasingly recognised today. Such design approach consisting in the use of semi-rigid joints with partial strength is getting more and more attainable and popular. It requires the knowledge of the full nonlinear moment-rotation behaviour which is also one of the parameters in such design approach. Typical and to date insufficiently researched type of joints having the stated properties is the bolted beam-to-column joint executed by flange cleats. Furthermore, the assembly of this type of joints is simple, quick and inexpensive. Nevertheless, although there are several theoretical models for the characterisation of behaviour of this type of joints so far there are no clear and probabilistically evaluated recommendations for the use of optimum model. The research that is shown in this thesis treats this issue and pays particular attention to the probabilistic evaluation of theoretical models. Moreover, the research has been extended to include the so far unresearched types of flange cleats with stiffener and their contribution to the total behaviour of joints with flange cleats. The results of the conducted experimental testing and numerical analyses are shown for twelve separated pairs of cleats in tension. Accordingly, the F- Delta curves for four different types of cleats in bending have been recorded and characterised. Experimental values of resistance, stiffness and ductility have been obtained. The results have been used for the comparison of the existing models of resistance and stiffness and for the development of models for the assessment of the stiffened cleat behaviour. Furthermore, the results of the conducted experimental testing of eight beam-to-column joints with flange cleats are described. In doing so four configurations have been studied considering the application of the stiffened upper and/or lower flange cleat. This way the M-Phi curves of the analysed configurations of joints with regard to their resistance, stiffness and rotation capacity have been recorded and characterised. Most of the existing theoretical models for the assessment of resistance and stiffness of cleats in bending are based on the T-stub models. By upgrading the existing models within this research, new original models for the behaviour assessment of the stiffened cleat in bending have been developed. An effective length was proposed taking into consideration the stiffening of angle leg in bending. All those models have been evaluated considering the performed experimental tests and comprehensive numerical simulations for four different cleat types. The proposed new models for the assessment of stiffened cleat give results of a satisfactory level of accuracy. This research of separated pairs of cleats in tension is the base for the research of upper flange cleat in the tension zone of the joint exposed to bending. In order to research the behaviour of flange cleat in the tension zone of joint, the tests were extended to the experimental testing of joints with flange cleats. Methodology for the behaviour assessment of four tested configurations of joints regarding resistance and stiffness is presented. This methodology is based on the models developed on the component (cleat in bending) with a proposed expression for the lever arm in joints with stiffened lower compression flange cleat. By comparing with the experimental results, the accuracy of the developed models for the characterisation of stiffness and resistance of the tested joint configurations has been confirmed. At the end of this research probabilistic analysis has been conducted for all considered resistance and stiffness models for the tested types of cleat pairs in tension and for the tested joint configurations. Taking into consideration the results of the conducted experimental testing through the uncertainty resistance and stiffness model factor probabilistic assessment of the considered models has been conducted. This way, by using scientific arguments, the specific optimum models for the characterisation of behaviour of the analysed types of cleats in bending (as a component) have been determined, as well as of the analysed configurations of joints with flange cleats.
 

Davor Skejić: RELIABILITY OF WELDED SEMI-RIGID BEAM TO COLUMN JOINTS, master's thesis, mentors: Prof. Darko Dujmović, PhD.; Prof. Mladenko Rak, PhD., Faculty of Civil Enginering, University of Zagreb, 2005.

In this paper a probabilistic evaluation of the component method for design of welded semi-rigid beam-to-column joints, as a probabilistic evaluation of carried experimental tests, is presented. The experimental tests are carried out on six nominally identical welded beam-to-column (IPE 240 – HE 200 A) joint specimens, made of steel S 235. Each specimen's element is obtained from different European producers. For qualitative preparation of the experimental tests and data processing of experimental results, numerical analysis based on Finite Element Method provided by computer program COSMOS-M, as well as, an extensive semi-probabilistic analysis according to prEN 1993-1-8 have been used. Obtained experimental results are finally described with M-phi curves which represent real behaviour of tested type of joint. For considered component method, ultimate limit states are defined and expressed as functions of n-dimensional vectors of stochastic (basic) variables. The stochastic model, which describes properties of basic variables with mean value, standard deviation and type of probability distribution, is formed. Statistical data required for the stochastic model configuration are obtained from own measurements and results of the extensive experimental tests on behaviour of considered type of welded joints. Probabilistic, actually numerical, analysis of reliability on level II is conducted using recent reliability methods implemented in the module COMREL-TI which is a part of the software package STRUREL. Calculations are performed on the simple design model which represents tested laboratory specimen of welded beam-to-column joint. Obtained reliability indices' values represent values for objective and quantitative evaluation of the reliability level of each joint component, as well as, possibility of the quantitative comparison between reliability levels of other structural elements in the frame steel structure, in which the analysed joint is constituent part. Additional probabilistic evaluation of component method using laboratory test results, through resistance model factor, provides the ground for scientifically based decision about the acceptability of certain resistance models in order to find experimental resistance of this type of joint. This kind of analysis provides objective evaluation of the real reliability level for these amusing structural systems.