Innovative technical solutions for improving performance of wooden floors

Abstract

This paper presents an innovative technology of timber floor whose function is both the replacement of horizontal elements in historical buildings, not subject to constraints of protection, and the utilization in new buildings. The new system responds to the market demand that requires high performance characteristics in terms of stiffness in the plane, correct transmission of the horizontal actions, lightness and low thicknesses. The floor at issue is composed of two plywood panels glued in the intrados and extrados to a moulded grid panel in GFRP with a single warping of glass fibres. We want here to present the stages of study, research and design, that helped define an element considerably rigid and at the same time lightweight, trying to contain the thickness of the section. Laboratory testing phase has highlighted the increase in stiffness and flexural limitation that the coupling of the moulded grid panel with plywood panels undergoes in the different configurations.
Sandwich wooden panel slab, GFRP core moulded grating, Stiffness improvement, Cross section optimization

STATE OF THE ART
In historical construction, intermediate horizontal structures offer several recurring typologies. The most widespread, for upper floors' slabs, is a wooden structure with fitted boards, or tiles or wicker mats ("cannicciati").
The constructive typology is usually made up of beams in simple support on masonry walls, with less than 5 meters span. Above these stands directly wooden planks or the substructure on which the slab is placed.
Wooden beams static failure is a recurring aspect, largely due to poor availability in money and local materials. When the situation described above involves primary structural elements, without any particular architectural qualities, the possible intervention of renovation is inappropriate for both economic and operational reasons. This allows for greater stiffness in the plane, proper connection between structural components, suitable cross sections and elimination of the most common degradation causes. [7-8-10] In the last century, in non-monumental historic buildings without safeguard restrictions, replacing wooden floors with concrete and masonry flooring systems was normal practice, even compulsory when operating in seismic areas, until promulgation of "Ordinanza No 3227/2003"into Italian regulation. Adopting this system shown many problems over time, especially following seismic events; floors and roofs, made up of reinforced concrete lying on ancient walls built in brick, stone or mixed techniques, amplified the earthquake effects. Because of excessive weight and extreme stiffness, they caused greater push on the walls, enough to produce, in some cases, crushing and overturning. [1][2][3][4][5][6][7][8][9][10][11] In this regard, construction market proposes many innovative systems designed to ensure stiffness in the floor plane together with a proper transmission of horizontal actions, with significantly lower weights. Most of the solutions however present the application of wood-concrete mixed floors. [16-17-18-19-20-21-22-23-24-25-26-27-28-29-30-31-32-33-34-35]

WOODEN SLAB'S FEATURES
The component, subject matter of experimentation, well integrable into nowadays' market, has to face floors ordinary functions: loads' support and distribution to vertical structures, as well as connection between the walls to ensure greater stability. One key feature is lightness, to avoid, in case of earthquake, the onset of high dynamic actions on the walls, the main cause of their collapse outside the plane for thrust and hammering. [5] The floor at issue consists of a core made of a grating printed in polyester resin, reinforced with a single frame of glass fibres. This is an innovative structural element with high load-bearing capacity, low thickness and low weight.
In order to increase its stiffness, the structural core has been coupled with plywood panels both at the intrados and at the extrados. The coupling of GFRP gratings and wood multilayer panels is done by a gluing system made with a high performance adhesive and fiberglass mesh.
This solution minimizes relative sliding problems between various layers and punching phenomena associated with the use of mechanical connectors, which will be employed only for connections to the bearing structure and to the perimetric bond-beam. [2][3] Glass fibre mesh, which acts as a "matrix" for bonding, is functional in restoring the discontinuities determined by joints

CARATTERISTICHE DELL'IMPALCATO
L'elemento di fabbrica oggetto di sperimentazione, che ben si inserisce nel mercato odierno, è chiamato ad assolvere le funzioni ordinarie degli orizzontamenti: sostegno dei carichi e ripartizione degli stessi alla struttura in elevazione, nonché collegamento tra le pareti garantendone una maggiore stabilità. Caratteristica fondamentale è la leggerezza per scongiurare, in caso di sisma, l'insorgenza di elevate azioni dinamiche sulle pareti, causa principale del collasso fuori dal piano delle stesse per spinta e between gratings. The whole package has a significant reduction in weight and thickness compared to a traditional intervention with cooperating r.c. slab: the plywood panels have a weight of about 5 kg/m 2 for 9 mm thick while the moulded grating weighs about 30 kg/m 2 for 42 mm thick. Total thickness is therefore limited to 62 mm and the entire slab weighs about 40 kg/m 2 . (Fig. 1)

FIRST TESTING PHASE
Experimental work main purpose is to highlight stiffness increase and bending limitation generated by coupling moulded grating panels with plywood panels.
The experimentation involves a comparison of two adhesives to determine which one ensures better cohesion between materials and highest mechanical properties to the entire panel. The two adhesives compared are a water based one, IPN, and an epoxy resin.
All types of tests defined match up to different configurations of the specimens.
In first place for presence or absence of structural layer discontinuities, related to the need to test slab behaviour in presence of disconnections dictated by standard moulded panels sizes (usually 1220x2000 mm). On second hand for the use of the two bonding systems in order to decree which one ensures the best performances. The specimens tested consider, in fact, both nondiscontinuous configurations and discontinuous ones, with joints in the moulded panels near the maximum stress zones, concerning bending moment (mid-point) and shearing force (support).
This choice was made in order to obtain a condition close to that of the actual laying configuration, in which wooden beams, on which lies the slab, have 1000 mm span. The load is supplied by a press through a splitter that transfers load to two loading pins symmetrically positioned at 350 mm from the external supports (Fig. 2). The test machine was calibrated using two different data tracking systems: two electromagnetic displacement transducers (LVDT) and a specific DIC 3D camera (Digital Image Correlation) able to determine three-dimensional displacements and deformations by capturing stereoscopic images. [14] Monotonic and quasi-static cyclic tests were performed on different sandwich panel configurations. Standalone tests were also performed both on moulded grating panels and on plywood panels.
The testing and analysis procedure used for single specimens was the same for all configurations. In monotonic tests, the sample was loaded until reaching crisis and subsequently discharged. In cyclic tests, subsequent loading steps were followed (10, 20, 30 kN) unloading the specimen every time to check the onset of plasticity residue.
For sandwich panels, plywood panel failure was considered as slab's crisis: once occurred, the specimen was unloaded. Subsequently, results obtained from different tests were compared to determine the real increase in stiffness of the sandwich panel compared to the single moulded grating, as well as for gluing system choice.

DATA ANALYSIS
Thanks to results extrapolation, after processing the images captured with the DIC system, intrados vertical displacements were identified to define most stressed areas. Five checkpoints were selected on the specimen centre line to evaluate panel homogeneity in response to loading. (Fig. 3

MOULDED GRATING
The addition of plywood panels resulted in a remarkable increase in stiffness, but caused a decrease in the maximum load corresponding to the specimen crisis, when compared to the standalone moulded grating one. This is true if the specimen crisis is meant as the moment when one of its components breaks. In fact, only the underlying plywood panel cracked, subjected to tensile stress, at a load lower than the one reached by the single moulded grating. This assumes that, if the specimen were still loaded after the plywood panel failure, the load-displacement curve would be positioned close to that of the single moulded grating, ensuring a maximum load closer to that of the test in question, or even greater. In this case, stiffness would have been affected, being subjected to a drastic drop caused by the intrados panel loss of collaboration.
In regards to the two sandwich panels with different adhesives, results do not deviate too much from each other, leaving great freedom in gluing system choice.

THE SUPPORT
In the presence of the central discontinuity in the grating, the behaviour of the two specimens with IPN gluing system and epoxy resin was extremely similar, with a slight decrease in performance in the IPN system. Conversely, in case of a joint near the support, the performance of the panel glued with water adhesive (IPN) was lower than that of the epoxy resin. In case of the IPN, the plywood panel unglued without cracking. This is related to a lack of strength of the adhesive in shearing force: once it became too high, it generated the detachment of the two surfaces in the point where strains and fragility were greater.

TESTS AND THE PREVIOUS ONES
Since a previous test campaign has been carried out on this sandwich panel, using a different adhesive applied without glass fibre mesh, it is possible to compare the results obtained from tests with the same typological characteristics. In particular, a comparison was made between packages presenting the new gluing systems, IPN or epoxy resin, and the previous epoxy adhesive.

C O N F R O N T O PROVE CON DISCONTINUITÀ IN MEZZERIA E IN PROSSIMITÀ DELL'APPOGGIO
In presenza della discontinuità centrale del grigliato, il comportamento dei due provini, rispettivamente con incollaggio in IPN e resina epossidica, è risultato estremamente simile, con una leggera diminuzione delle prestazioni nel sistema con l'IPN. Al contrario nel caso di giunto in prossimità dell'appoggio le caratteristiche prestazionali del pacchetto che prevede l'utilizzo dell'adesivo ad acqua (IPN) sono risultate inferiori a quelle The presence of glass fibre mesh in the new configurations increases the panel stiffness, to the detriment of ultimate load, which is lower. Stiffness increase is clear from the first load values, where the slope of the curve tangent line is significantly greater, providing smaller deflections. Early failure phenomena, however, can be attributed to the over-stiffness of the interface. This causes greater stress on plywood panels, therefore they crack first, without compromising the resistance reserve determined by the grating, which remains intact. The aspects that lead to prefer resin are related to a factor of practicality in laying and gripping times. The workability of the IPN has a direct dependence on its preparation, which is strongly modified by the addition of the thixotropic agent. This adhesive showed signs of collapse in the test with separation between the gratings placed near the support, resulting in an interface detachment, similar to a lack of resistance in case of high shear stresses. The addition of fiberglass mesh further contributes to increase stiffness and limit flexural displacements. Through monotonous tests, it was also verified that ultimate loads are extremely high, well above those used for residential slabs.

PANEL'S STIFFNESS
After the experimental phase, which has shown excellent results in terms of stiffness and flexural limitation, it seems appropriate to make a critical analysis of obtained results. This in order to be able to define whether the placement of the grating between the two plywood panels is effective, ensuring a maximized contribution. The goal is to obtain a slab not only effective and able to bear considerable size loads, but equally efficient, ensuring to obtain these results in the best possible way. After experimentally determining the elastic modulus E g of the grating, total stiffness of the entire slab has been calculated, given the elastic modulus E l of plywood panels and the moments of inertia J of the resistant cross sections.
From this result, the percentage contribution of the grating can be calculated: It is evident from this parameter that the grating is not exploited at its fullest mechanical potential, as the percentage of maximum bore bending moment does not reach high values. [9] The incorrect placement of the grating, which implies an ineffective exploitation, is evidenced by the fact that just after failure, cracking occurred only in the lower plywood panel, subjected to tensile stress, whereas in the composite material inside the core of the sandwich panel failure did not occur.
Given these experimental data, the slab cross section with grating inserted between two plywood panels does not optimize these two materials performance. The grating contribution is significantly lower than the plywood, less performing material. This is theoretically the result of internal stress distribution, generated by the application of a bending moment. [15] Therefore it is obvious, even at a theoretical level, that the grating, located near the neutral axis, is placed in a position that does not allow its maximum exploitation while the wood is placed in the most stressed areas. [13] Trying to ensure a more rational use of the composite material, more performant to tensile stress, it was decided to search for other design solutions redistributing the cross section components in order to place the most performant in the most stressed position.
Bearing this in mind, the new design hypotheses, studied analytically in terms of total stiffness and percentage contribution of the grating, provide for a different distribution of the elements. In the first instance it was supposed to remove the underlying plywood panel; this solution shows an increase in grating contribution accompanied by a marked decrease in total stiffness. In view of the poor use of the grating in the initial configuration, the possibility of replacing it with a material with lower mechanical characteristics, such as polystyrene, was also investigated.
Various configurations have been hypothesized introducing XPS replacing the GFRP grating. The analytical results are reported in terms of stiffness, in order to be compared with that of the sandwich panel, object of the experimentation first phase. In analytical terms only the XPS of 60 mm thickness, if interposed between the wooden panels, would be able to guarantee a stiffness comparable to the one of sandwich panel containing the grating. (Tab. 2) sollecitate. [ Avendo ottenuto risultati migliori con la resina epossidica durante la prima campagna sperimentale, si è utilizzato solo questo sistema di incollaggio escludendo l'IPN. La macchina di prova presenta la medesima configurazione della prima sperimentazione in modo da ottenere dei risultati comparabili. Per determinare gli abbassamenti dell'elemento di prova sono stati utilizzati soltanto dei trasduttori di spostamento in quanto in base ai risultati delle prove precedenti si è ritenuta l'acquisizione con la DIC ridondante. Anche in questa fase, la rottura del pannello di legno superiore è stata considerata come crisi dell'impalcato. Ottenuta la stessa si è infatti deciso di scaricare il provino benché si possa ipotizzare che il multistrato inferiore avrebbe raggiunto la rottura per trazione con carichi superiori.
The first ones would have shown a stiffness close to that obtainable with fibre-reinforced material. This hypothesis has been refuted by results of the experimentation, in which polystyrene and grating curves do not exhibit similar slopes.
In the previous analytical phase, the flexural stiffness of the symmetrical configurations with 9 mm thick plywood panels and variable XPS thicknesses were calculated.
To determine the stiffness obtained experimentally, we rely on the formulation of the deflection generated by the load pattern given by the testing machine. From this formulation, it is possible to determine stiffness, known the maximum load and the maximum displacement, using the following formula: As the load increases, the cross section modifies its shape, since the XPS is a compressible material, so stiffness is continuously changing. Stiffness, being equal to the product between elastic module and moment of inertia, directly depends on the cross section geometry.
The graph shows a maximum point corresponding to no load applied; while loads keep increasing, there is a drastic decrease in EJ. (Fig. 9) The specimen reaches expected stiffness only when the cross section is not subject to loads and then the maximum value reached in the graph corresponds, in order of magnitude, with that determined analytically. This is due to the fact that polystyrene, being a compressible material, allows the cross section to modify its geometry as the load increases because of local punching effects. At the al grigliato e il nuovo oggetto di sperimentazione, dove il grigliato viene sostituito da pannelli di XPS di diverso spessore. Il grafico evidenzia che la rigidezza e la resistenza dell'impalcato contenente pannelli di polistirene sono nettamente inferiori a quelle dell'ipotesi progettuale originaria. Le rigidezze inoltre non raggiungono i valori calcolati teoricamente. Questa seconda fase sperimentale, al contrario di quanto definito per via analitica, ha dimostrato l'inefficacia dell'utilizzo di un materiale meno performante nella realizzazione del sandwich. Ne segue la necessità di mantenere il gigliato in PRFV quale elemento di interposizione tra i pannelli di legno multistrato. Dal confronto tra le curve caricospostamento dei provini con l'XPS con quello contenente il grigliato si nota una mancanza di corrispondenza tra i valori di rigidezza teorici e quelli pratici relativi alle configurazioni oggetto della seconda fase sperimentale. I primi avevano evidenziato una rigidezza prossima a quella ottenibile con il materiale fibrorinforzato, ipotesi confutata dai risultati della sperimentazione in cui le curve di polistirene e grigliato non presentano pendenze simili. Nella precedente fase analitica si erano calcolate le rigidezze flessionali delle configurazioni simmetriche con pannelli multistrato lignei di spessore 9 mm e XPS di spessore variabile. Per determinare la rigidezza ottenuta sperimentalmente ci si basa sulla formulazione della freccia generata dallo schema di carico impartito dalla macchina di prova. Da questa formula è possibile determinare la rigidezza, noti il carico massimo e lo spostamento massimo, attraverso la seguente formulazione. All'aumento del carico la sezione modifica la sua conformazione, essendo l'XPS un materiale comprimibile, per tanto la rigidezza subisce delle modifiche. Questa essendo pari al prodotto tra il modulo elastico e il momento di inerzia dipende direttamente dalla geometria della sezione. Il grafico presenta un punto di massimo in corrispondenza del carico nullo, con l'applicazione della forza Figure 9. Load-stiffness graph of sandwich panel with 9 mm XPS core.

BENDING RESISTANCE VERIFICATION OF THE SAND-WICH PANEL SLAB
Experimentally determined the effectiveness of using the grating as structural core of the entire sandwich panel, the goal became how to use collected data in order to generalize the considerations carried out so far. The objective is to set up a basic calculation and verification method that allows the designed sandwich panel to be used in real situations that may widely differ from the geometric configuration of the testing machine.
In order to make considerations regarding future use of the sandwich panel slab it is necessary to define the loads weighing on the analysed structure and the floor deflection limits imposed in residential buildings by Italian regulations (D.M. January the 14th, 2008).
With this aim, for the mere purpose of calculations, a traditional completion (screed, soundproofing, subflooring and flooring) has been hypothesized in order to provide higher security standards in terms of non-structural permanent loads. Once known the loads weighing on the structure, it is then possible to determine the maximum load corresponding to normal operating conditions and thus determine the maximum deflection reached by the slab.
The theoretically calculated deflection is confirmed by experimental data and is always significantly lower than the maximum deflection allowed by regulations.
The load displayed on the diagram with a red line is equal to 1.621 kN and is applied by the piston under testing conditions. It corresponds to the load that same time, there is a decrease in the moment of inertia, which causes a parallel decrease in total stiffness to values that strongly diverge from theoretical ones.
si ha una drastica diminuzione di EJ. (Fig. 9  tensile strength can be determined by Navier's formula [15]: where W is the elastic section modulus depending on its inertia (J) and on the distance from the neutral axis (y) to the most stressed fibres: the mixed cross section in elastic phase was calculated. The strain distribution in this case will depend on the presence of two different materials and in particular on the relationship between the two elastic modules.
Having witnessed during the experimental test the failure of the lower plywood panel due to tensile stress, the outermost fibre has then to bear a stress equal to the ultimate tensile strength previously calculated. [6][7][8][9][10][11][12] Until reaching the interface with the grating stresses follow the typical stress distribution of a homogeneous cross section. Since no ungluing have ever appeared during the tests and given the intrinsic stiffness of the epoxy resin, the two materials can be considered free of horizontal mutual sliding. Such hypotheses ensure the plane sections remain plane and constant curvature. These conditions ensure that plywood and moulded grating have the same strain at their interface. By knowing this strain, value is possible to determine stresses inside the grating, according to the ratio between the two elastic modules.
Once defined stresses on the cross section, it is possible to determine internal stresses that generate the maximum bending moment (M Rd ) that the cross section can withstand during the elastic phase. The internal moment is generated by two contributions, given by both different materials. The bending resistance verification method can be set once determined the highest bending moments the cross section can bear. For a variable-size slab, the verification must be carried out for both the most stressed and the less resistant cross sections. The first ones coincide with those exposed to maximum moment; usually in such positions, following a suitable arrangement of the panels, there will be no discontinuities in sandwich slab layers and the ultimate moment of resistance is the sum of the contributions of both materials. (Fig. 11) Figure 11. Stress distribution diagram in sandwich panel's cross section. elastica. L'andamento delle tensioni in questo caso dipenderà dalla presenza di due materiali differenti e in particolare dal rapporto tra i due moduli elastici. Avendo assistito nel corso della prova sperimentale alla rottura del pannello in legno inferiore per trazione, si impone che la fibra più esterna sia soggetta a una tensione pari alla resistenza ultima a trazione per flessione calcolata precedentemente. [6][7][8][9][10][11][12] Fino all'interfaccia con il grigliato le tensioni seguono l'andamento a farfalla tipico di una sezione omogenea. Poiché non si sono mai manifestati scollamenti durante le prove e data la rigidezza intrinseca della resina epossidica, si possono considerare i due materiali privi di scorrimenti orizzontali reciproci. Tali ipotesi garantisce la conservazione delle sezioni piane e una curvatura costante.
Queste condizioni assicurano che all'interfaccia il legno e il grigliato presentino la medesima deformazione, da cui è possibile determinare, in base al rapporto tra i moduli elastici, anche la tensione agente nel grigliato. Definite le tensioni sulla sezione è possibile determinare le sollecitazioni interne che generano il momento flettente massimo che la sezione è in grado di sopportare nella fase elastica M Rd . Il momento interno prevede due contributi, dati dai diversi materiali. Determinati i momenti flettenti che la sezione è in grado di riprendere si può impostare il metodo di verifica a flessione. Per un solaio di dimensioni variabili la verifica dovrà essere effettuata sia per le sezioni più sollecitate sia per quelle meno resistenti. Le prime coincidono con quelle sottoposte a momento massimo; di norma in tali posizioni, seguendo un'opportuna disposizione dei pannelli, non sono presenti discontinuità negli strati componenti l'impalcato e il momento resistente è somma dei contributi dei due materiali. (Fig. 11 In case of maximum deflection the formula is: To maximize allowable span, the maximum deflection (f max ) reachable according to regulations must be taken into account: Substituting this formula in the previous one, the equation becomes: From which the maximum span can be calculated: In case of maximum stress instead the formula is: By equating the calculated moment to the ultimate one, maximum allowable span becomes: As expected, the most demanding condition is the one in terms of maximum admissible deflection. The excellent mechanical properties of the moulded GFRP grating are not enough to ensure its use as a slab for residential buildings.
The coupling with plywood panels, though still improvable, is thus effective in ensuring a sharp decrease in the deformability of the entire sandwich slab.

FUTURE DEVELOPMENTS
In the light of the encouraging results obtained experimentally, and in order to follow up the analytically developed theoretical considerations two different developments can be hypothesized. First, an experimental campaign, which tests the sandwich slab configurations that have been studied only through a theoretical analysis and which have an optimized cross section in terms of resistance. However, it is considered mandatory, once the optimal configuration has been established, to study the behaviour of the grating under horizontal stresses and to test its bi-directional warping also for loads acting in the plane, parallel to the element.