Assessment of Seismic Behavior of Traditional Timber Frame Ottoman Houses: Frame Tests Y. D. Aktaş Erdem Program of Restoration, Middle East Technical University, Ankara, Turkey U. Akyüz, B. Erdil, A. Türer Department of Civil Engineering, Middle East Technical University, Ankara, Turkey N. Şahin Güçhan Department of Restoration, Middle East Technical University, Ankara, Turkey ABSTRACT: There are numerous observations made following a number of historic and contemporary earthquakes, claiming that the Ottoman traditional timber-frame, himis, houses have an inherent seismic resistant property. It is argued that timber frame houses behaved better under earthquake loading, based on the fact that the number of collapsed timber structures is less than the other construction types. In most of the reported cases, however, comparisons made for the seismic performances of himis houses and the others remain to be qualitative and lack quantitative engineering evaluation. The aim of this paper is to present the preliminary results obtained after a number of frame experiments which intended to investigate the seismic resistance of timber-frame Ottoman houses, within the framework of an ongoing research project. 1 INTRODUCTION The term “Ottoman house” is commonly used to define traditional timber frame structures, which have certain architectural and technical features that have profoundly been studied by many scholars (e.g. Cerasi 1998, Günay 1998, Kuban 1995, Eldem 1984, Kafesçioğlu 1955). There are also numerous studies reporting the seismic resistance of such timber frame structures, based mainly on visual investigations made in-situ after many historic and contemporary seismic events in Turkey and abroad (e.g. Şahin Güçhan 2007, Gülkan & Langenbach 2004, Gülhan & Özyörük Güney 2000, Tobriner 2000). In these observations, it is reported that the hımış structures had no or little damage, while most of the reinforced concrete structures collapsed. The most affected parts of the damaged timber structures were masonry sections, such as masonry ground floor walls, chimneys etc., and what’s more, the timber frame remained intact, even in the case that loosening of infill materials occurred. However, there have been numerous collapsed hımış houses in Afyon earthquake (Feb. 3rd, 2002), which raised questions about much favored hımış houses. Lack of maintenance, material degradation, and improper nailing are believed to be some of the factors contributing to the occasional poor performances of timber structures under seismic loading. The aim of this paper is to present the preliminary results obtained from the ongoing TUBITAK (The Scientific and Technological Research Council of Turkey) Research Project (106M499), which intended to investigate the seismic resistance of timber structures. 2 TRADITIONAL OTTOMAN HIMIS HOUSE WITH TIMBER FRAME CONSTRUCTION Traditional timber frame Ottoman house (called as “hımış” in Turkish) is characterized by several features such as a typical construction system, which is mostly composed of a masonry base and timber frames with infill. Apart from slight regional differences, the general form and de- sign principles of the Ottoman n houses h were successfully applied to a vast area, regardless re of drastic differences in climate (Şah ahin Güçhan 2007, Günay 1998, Cerasi 1998, Kuban an 1995). Building methods used for the he construction of a timber Ottoman house has evolved ved as to have very simple details, especially in terms of connections. This brings also along the sp speed and easiness in reconstruction of houses ses after a devastating fire, as frequently occurred thro hroughout the history (Şahin Güçhan 2007, Kub uban 1995). The ground floor can be made de of rubble stone masonry or adobe with timber lintel tels in regular intervals as well as of cut stone sto or alternating layers of stone and brick (El (Eldem 1984, Kafesçioğlu 1955). The ground d ffloor might also be constructed by singular vertical al timber elements, filled with non-load bearin aring masonry or left empty (Kafesçioğlu 1955). In bot oth cases, the upper floors were made of frame mes with timber elements connected to each other inn tthe vertical, horizontal and diagonal direction ions. The gaps between these timber elements can be filled with different materials, depending on the availability in the region, such as brick, ston one, adobe or timber. C RESISTANCE OF TRADITIONAL OTTOMAN H HOUSE 3 ASSESSMENT OF SEISMIC WITH TIMBER FRAME CONST STRUCTION For the aim of assessing the seismic seis performances of traditional Ottoman housess w with timber frame construction, Safranbolu,, which w is a UNESCO World Heritage Site since 199 994, has been selected as the case study (Figure ure 1), for being a settlement, where hımış traditionn hhas successfully been used with hundreds of historical traditional houses registered as culturall eentities, and because the examples of tradition ional dwellings with timber frame construction have to a large extent been conserved until today ay within the settlement tissue. In addition, this is a settlement placed in a seismically active are area (first degree seismic zone). Therefore, the experi erimental and analytical studies can be suppor orted by means of site observations in terms of seism eismic performances of the structures. Figure 1. A Safranbolu House (http:/ tp://en.wikipedia.org/wiki/Safranbolu) Figure 2. Two examples of frame selection sel (Photos by Akarsu, 2003. The upper one is the Ho House of Emin Hocazadeler, while the bottom onee is the number 20 on Akpınar Street in Safranbolu) represent the Within this framework, so far two different frames were selected and tested to re Safranbolu himis houses (Figure re 2). These frames, chosen from Safranbolu, were con constructed of fir, by a local builder in the labor oratory so as to be equivalent to those found in the ori original traditional houses in size, configurati ration and material, and tested under reverse-cyclic lo loading. The tests were conducted with empty pty timber frames initially and then were repeatedd w with timber frames having traditionally prepa epared brick infill; although, other infill materials suc uch as adobe, brick, and stone are also used inn practice. p For this aim, a test set-up that at is basically composed of a foundation part, vertica ical masts and horizontal ones securing them to the wall was designed (Figures 3, 4). The top later teral displacement was measured by meanss of o a Linear Variable Differential Transformer (LVD VDT) of 600 mm, while bottom lateral displac lacements were checked on both sides by two LVDT’s T’s of 50 mm. In addition to these, diagonall displacements d were measured, again, on both sid sides, by two LVDT’s of 200 mm. Figure 3. A frame ready to be tested ed Figure 4. Top portions of test set-up up a frames failed at connections. Nails came out an and drawn in As can be seen in Figure 5,, all again in successive cycles. As the th nails were repeatedly moved in and out, they hav have bent and failed at extreme deformations at large number of cycles. Figure 5. Frames always failed at the nailed connections After testing the frames in emp mpty state, local builders came once again to the labor boratory to repair damaged connections using ng new equivalent nails, and then to infill them with ith traditional brick infill materials (Figure 6). Figure 6. Two examples to traditiona onal brick infill materials The test results shown in Tab able 1 indicate that the existence of infill material improves im the maximum load bearing capacity ty by b about 1.5 to 2 times the original capacity. On the he other hand, the failure mechanism always go governed by the connections. The nails coming offf tthe wood at the connection causes developme ment of a mechanism which would lead to overall fai failure. At the ultimate range, the observed capacity cap was about 10 kN to 15 kN lateral force fo for the given frames. The frame with higher lateral lat strength was about 50% longer than the otherr oone (and actually shorter in height), indicatin ating that the overall capacity might be correlated wit with length of the frame. In a very approximat ate way, the lateral strength of hımış frames was ob obtained as 2 kN/m and 3 kN/m for empty and nd brick infilled test frames, respectively, when two w windows and two diagonal members existed. 4 CONCLUSIONS The studies regarding the evaluat uation of the seismic resistance of traditional Ottoman an house with timber frame construction based ed on quantitative engineering approach lack at all in the t relevant literature. One can find numerou ous papers reporting the in-situ observations regardin ing the structures’ seismic performances in hhistoric and contemporary earthquakes. However,, th these studies fall short of quantitative conclusi sions about the seismic evaluation of traditional Otto ttoman dwelling, which have its own technica ical particularities. Table 1. Two frames tested both in empty state and w with infill Frame 1 Infill Photo 1 No infill Frame #1 - 340x485 cm (H x L) Brick 2 Nails of Frame#1 Capacity Curve 3 No infill Frame #2 - 325x310 cm (H x L) 4 Brick Nails of Frame#2 The quantitative evaluation studies reported here include experimental tests on two timber frames with and without infill. Such an evaluation study should also include analytical approach, which has not started yet. The failure mechanisms were observed and recorded in detail during each experiment. At the end of each experiment the resulting load-displacement relation was also investigated. In this paper, the basic load-deflection behaviors obtained at the end of two frame experiments were presented. According to this, the main conclusions that have been drawn so far are: • The connections usually govern the behavior. The frames fail always at the connections. • The number of nails at each connection and their drive angles are of importance. • The infill brick increases the capacity by 1.5 to 2 times the empty frame capacity • Increase in the total mass is at least two times the empty frame; therefore, strength gain due to infill material is questionable for performance improvement in an earthquake. • The drift ratios are in the order of 6% of height for our study, where the frames could still hold their maximum lateral load capacity. Therefore, the hımış type construction is evaluated to be very ductile compared to reinforced concrete structures. • Improving connection strength is expected to directly improve the overall load bearing capacity of timber frames. At the end of this study, it is aimed to contribute to the preservation of traditional Ottoman timber frame dwellings by supporting the existing body of knowledge with a more quantified approach as well as to develop more established strengthening strategies for the existing ones, and to encourage the construction of further houses using this tradition facing extinction. ACKNOWLEDGEMENT This study is being supported by TUBITAK (The Scientific and Technological Research Council of Turkey) research project no 106M499. REFERENCES Aktaş, Y. D., Erdil, B., Akyüz, U., Şahin Güçhan, N., & Türer, A. (2009). Structural Behavior of Timber Frame Ottoman Traditional Hımış Houses - Frame Loading Experiments. International Symposium on Historical Earthquakes and Conservation of Monuments and Sites in the Eastern Mediterranean Region for the 500th Anniversary Year of the 1509 September 10, Marmara Earthquake, (pp. 364366). Istanbul. Cerasi, M. M. (1998). The Formation of Ottoman House Types: A Comparative Study in Interaction With Neighbouring Cultures. Muqarnas 15 , 116-156. Eldem, S. H. (1984). Türk Evi Osmanlı Dönemi / Turkish Houses Ottoman Period I-III. Istanbul: Türkiye Anıt, Çevre ve Turizm Değerlerini Koruma Vakfı. Gülhan, D., & Özyörük Güney, İ. (2000). The Behaviour of Traditional Building Systems Against Earthquake and Its Comparison to Reinforced Concrete Frame Systems: Experiences of Marmara Earthquake Damage Assessment Studies in Kocalei and Sakarya. Earthquake-Safe:Lessons to be Learned from Traditional Buildings. Istanbul: ICOMOS. Gülkan, P., & Langenbach, R. (2004). The Earthquake Resistance of Traditional Timber and Masonry Dwellings in Turkey. 13th World Conference on Earthquake Engineering. Vancouver, B.C. Canada. Günay, R. (1998). Tradition of the Turkish House and Safranbolu Houses. İstanbul: YEM. Kafesçioğlu, R. (1955). Kuzey-Batı Anadolu'da Ahşap Ev Yapıları. İstanbul: İstanbul Teknik Üniversitesi - Mimarlık Fakültesi Pulhan Matbaası. Kuban, D. (1995). The Turkish Hayat House. İstanbul: T.C. Ziraat Bankası. Şahin Güçhan, N. (2007). Observations on Earthquake Resistance of Traditional Timber Framed Houses in Turkey. Building and Environment 42 , 840-851. Tobriner, S. (2000). Wooden Architecture and Earthquakes in Turkey: A Renaissance Report and Commentary on the Performance of Wooden Structures in the Turkish Earthquakes of 17 August and 12 November 1999. Earthquake-Safe:Lessons to be Learned from Traditional Buildings. Istanbul: ICOMOS.