Structural assessment of existing timber roof structures for green roofs in Rotterdam

Structural assessment of existing timber roof structures for green roofs in Rotterdam

Rovers, L.H.C.J.

Van de Kuilen, J.W.G. (mentor)
Ravenshorst, G.J.P. (mentor)
Blom, C.B.M. (mentor)
Gard, W.F. (mentor)

Civil Engineering and Geosciences

Structural Engineering

2015-11-24

Green roofs, also known as vegetation roofs, are becoming very popular for residents of houses or apartments and for good reasons. Vegetation on rooftops have many benefits on global and local level. Expectations show that the climate change will lead to heavier storms and causes the sewers in Rotterdam to overflow. Hence, the municipality of Rotterdam wishes to apply this special type of roof on a large scale to buffer rainwater which can gradually be discharged. Although the concept of a green roof is nothing new, applying them on this large scale to solve an urban problem makes it an interesting topic. Two types can be distinguished: intensive and extensive green roofs. Both types are able to retain rainwater but can be distinguished in their function. Intensive green roofs allow for recreation and gardening while extensive green roofs have an aesthetical function. The municipality of Rotterdam and its citizens both want green roofs instead of bitumen roofs, however they neither have the time nor the knowledge to determine whether their timber roof structure is suitable for this extra ballast. A first simplified calculation indicated that there was not enough strength to resists this load. If no research is done then Rotterdam will only be able to have vegetation on structures of steel or concrete while timber roofs are commonly present. In an ideal future every roof in Rotterdam is a green roof. This thesis researches the residual capacity of flat timber roofs by reducing the uncertainties associated with the strength. The main goal is therefore to be able to predict, and if necessary increase, the true strength capacity without demolishing the roof structure. The past The first step is to identify the size of the problem. A multi-criteria analysis was already performed by the municipality and translated into a potential map. However the criterion “year of construction” has a high level of uncertainty because there is a lack of knowledge about older timber roof structures. The map distinguished five groups with different ranges of construction years which was based on experience. A logical step is an archival research to the history and typologies of houses in Rotterdam. This investigation revealed that most houses were built before 1940 (pre-war) but a large amount of roof structures are renovated or renewed in the 80’s. Timber was the main building material for roof structures but after World War II the focus was on speed and efficiency which resulted into more concrete roofs. Another consequence of World War II was the destruction of the city archive. This resulted into the loss of information about the present timber properties and dimensions. The structural geometry of flat roofs did not change over the years. During archival and literature research it was found that a beam supported by two masonry walls is standard practice. The main consequence of a green roof is than an increase of the bending moment of the existing timber elements. This increase may lead to collapse of the roof structure. Before extensive research towards the true strength of timber elements was performed, a more general investigation to gaining strength and possible weak spots was done. The idea was that some design norms throughout the years might have used too conservative values and thus strength could be gained by recalculating the structure with the current regulations. It was found that the values for roof structures stayed practically the same in the building codes. However most structures also satisfy the deflection requirement but this is not legally established. Beams that are designed on this requirement have extra strength. The visual grading norms, which determines the strength of a timber element based on visual characteristic, have become more flexible over the years. Also the strength classes changed, before 1933 no strength value was used but dimensions were based on experience. Afterwards two main strength classes were defined as standard building wood and construction wood which are more or less equal to C18 and C24 in modern times. Wood is an organic material which is sensitive to time dependent processes that reduce the strength. Age is not necessarily a strength-reducing factor but is associated with strength-reducing processes. Four degradation mechanisms can be distinguished for timber: mechanical, physical, chemical and biological. The latter is the largest problem for roofs because insulation or treating the wood was not always done (correctly). Table 2-5 gives an overview of positive and negative aspects for roof structures. The present Current approaches are based on identifying the state of the structure and calculating the extra load according to the active regulations for new buildings. The building code refers to the NEN8700 for coping with existing structures. This norm gives five solutions when the strength is not sufficient: reduce the reference period, values are based on actual use, adjust the use, adjust the safety margin or adjust the strength. This thesis focused on the first and last option. Reducing the reference period is discussed and not recommended unless the engineer can determine and control the load with high precision. It is not legally determined if the change of the safety level is allowed. This will lead to discussions with “construction and housing inspection” in the future because roofs are less safe. The extra capacity must thus be found in adjusting the strength. The idea is as follows, freshly sawn structural timber is graded into a strength class. This means that a small amount of the graded timber does not have to meet a certain limit strength. Nowadays the 5% lower probability value is chosen as the limit value. This way of strength grading allows for beams to be stronger than the characteristic value. The experiments were aiming to predict the actual strength without demolishing the roof. During the research, thirteen beams were obtained from an ongoing demolishment. Ten of these members are of a renovated roof structure from 1983 while the original structure was from 1923. The other three members are from another building where the original structure of 1923 was still present. A strength prediction model was used that required the density and dynamic modulus of elasticity (MOE). The density can be measured with aid of a resistograph. Here it is important to drill in radial direction. Next a vibration meter was used to measure the wave speed which can be combined with the density to gain the dynamic MOE. Five sub-experiments were conducted to determine the difference between in-situ situations and free vibrations. As it turns out, a screw is the best way to introduce the wave and the surrounding increase the wave speed. This needs to be corrected with a certain coefficient on the frequency. A 6 to 10 percent increase of the frequency was found during testing. At last the true bending strength was checked with a four point bending test. The true strength was 20 to 140 percent higher than the characteristic value of the initial grade. Four strategies for future assessments are proposed: calculate as new structure according the current regulations, reduce the reference period (not recommended), visual upgrading and non-destructive tests. Each step requires more work but will, most likely, lead to extra strength. Two case studies were worked out following the different strategies. As expected, non-destructive tests is the most beneficial strategy because information about the actual strength is attained. With this information the strength class could be upgraded. In this case study it becomes clear that low weight green roofs (1 kN/m²) can be applied while a heavy green roof (3,4 kN/m²) needs more attention. A solution between these two extreme is also possible. Furthermore in one case the characteristic bending strength was increased with a factor of 1,5. Time dependent factors seem to be the main problem in all strategies. The duration of load may cause excessive deflections or even creep rupture. Limits to the deflections are not legally established and can be concealed with a lowered ceiling. The future The choice for a method of reinforcing an existing structure depends on a number of criteria. An engineer and user should discuss the possibilities that satisfies both. The most optimal solution will depend on the existing timber structure because every situation is unique. For the case studies the best solution is to increase the cross section of specific individual beams with timber elements. This method is easy, fast and cheap. At last an action plan was made for future assessments. Following the different steps in this protocol can make reinforcement and extra costs unnecessary and is thus the first step towards a Rotterdam with only green roofs.

green roof
existing roof
vegetation roof
timber
wood
existing structure

http://resolver.tudelft.nl/uuid:12b346f8-e6d3-421a-9f62-e3e28eef0d18

Student theses

master thesis

(c) 2015 Rovers, L.H.C.J.