Brick masonry response to wind driven rain

Engineering Structures 204 (2020) 110080

Contents lists available at ScienceDirect

Engineering Structures journal homepage: www.elsevier.com/locate/engstruct

Brick masonry response to wind driven rain

T

R. Cacciotti Institute of Theoretical and Applied Mechanics AS CR, v. v. i, Prosecká 809/76, 190 00 Prague, Czech Republic

A R T I C LE I N FO

A B S T R A C T

Keywords: Wind driven rain Brick masonry Damage Wind tunnel simulation Rain penetration test Performance

This study develops an innovative wind tunnel testing methodology for assessing the hygro-mechanical response of brick masonry to wind driven rain. The results highlight that such response depends largely on material properties, rain load characteristics, saturation level and wetting history. Additionally, rain penetration, even during milder rainfall events, induces the degradation in mechanical properties, stress redistribution and curtailing of damage and failure stress thresholds in bricks. Finally, synergies with additional adverse conditions, such as temperature and salts, produce a combined damaging action that can undermine the structural integrity of building elements. Within the research framework, future work is also proposed.

1. Introduction Wind driven rain (WDR) represents one of the most significant sources of moisture in buildings [1,2]. It is the result of the interaction between airflow and rainfall and it can be simply described as rain that is given a horizontal velocity component by wind [3]. Rain penetration can be defined as water ingress through the exposed surface of building components induced by the interplay between material properties and conditions, raindrop velocity and wind forces. Brick degradation due to moisture can raise a series of issues ranging from inadequate usability to material durability and structural integrity. In combination with other factors, it may significantly decrease safety in buildings and lead to defects or sudden failures [4]. There exists a vast amount of research concentrating on moisture movement in masonry and its effects on buildings. In particular, moisture-related problems deriving from damp conditions have been closely investigated in relation to the durability of the material [5,6], occupants’ health and structural issues [7,8]. The influence on moisture transport in masonry of material properties [9], interfaces between units and mortar joints [10] and workmanship quality [11] has been thoroughly analysed. Considerable effort has been put also in the study of wind driven rain simulation concentrating mainly on the raindrop size and rain intensity distributions, the deposition rate and impinging patterns on physical models and the effects on buildings in terms of rain penetration and induced damages. For example, studies by Flower and Lawson [12] and by Rayment and Hilton [13] have focussed on the analysis of raindrop trajectories and rain. The ability to reproduce in laboratory the effects of natural rain has also been explored in past research [14–16] providing particular insights on the characteristics of the simulated WDR including investigation of raindrop size distribution [17] and spatial uniformity of

rain as a function of nozzle positioning [18]. In an unpublished report, Blocken et al. have tested the effects of WDR on full-size buildings, concentrating on the calibration of the wind tunnel for rain simulation as well as on the measurement of the impinging rain on building models. Finally, another interesting research [19] has produced data on wind driven rain intrusion and interior damages on single-storey residential buildings. The assessment of water penetration into a wall assembly is of vital importance to predict the vulnerability of the components prior to construction. This is usually done observing the performance of the wall assembly exposed to wind driven rain till its failure i.e. when leakage occurs on the back side. Although several testing standards for water ingress assessment are available [20–22], these employ test parameters, such as the spray rate and the pressure differences across the wall assembly, which are not representative of the weather characteristics at a specific location. Methods for a far more realistic and suitable rain simulation, proposed by Sahal and Lacasse [23], aim at adjusting the testing conditions to local driving rain intensities and wind pressures. This is based on the Straube and Burnett approach [24] to estimate the quantities of rainfall deposited on the wall assembly and on Choi method [25] to calculate the short duration driving rain intensity and driving rain wind pressures. Although past research uncovers basic theoretical clues related to WDR testing and monitoring of the response of building components to rain penetration, many practical challenges still remain to be addressed. Evaluation methods commonly employed in this field are too general and often based on simple failure assessment (e.g. building component passing or failing a water penetration test). WDR simulations require the introduction of standardised calibration procedures as well as the development of consistent testing criteria and protocol. Moisture measurement necessitates repeatable, minimally invasive and

E-mail address: [email protected]. https://doi.org/10.1016/j.engstruct.2019.110080 Received 4 June 2019; Received in revised form 3 October 2019; Accepted 9 December 2019 0141-0296/ © 2019 Elsevier Ltd. All rights reserved.

Engineering Structures 204 (2020) 110080

R. Cacciotti

2.1. Climatic wind tunnel simulation

inexpensive solutions with regulated data gathering and processing practices. The lack of appropriate and sustainable methodologies for assessing the performance of brick masonry components postulates that experimental data necessary for further validating theoretical studies and field observations are insufficient. These are strongly required for the sake of generating innovative outcomes and ground-breaking research advancements in this field. This study attempts a systematic analysis of brick masonry constructions which undergo continuous moisture loading from wind driven rain. The primary aim is to develop a comprehensive testing methodology for assessing the response of brick masonry to wind driven rain and to evaluate appropriately the results in terms of the hygric and mechanical behaviour of different brick types. More specifically, the objectives of the research include the following:

In this study the Vincenc Strouhal climatic tunnel located in the Centre of Excellence Telč, Czech Republic is used (Fig. 1). It is a small scale tunnel intended for civil engineering purposes, designed as a closed circuit composed of two sections, the climatic and the aerodynamic one. For the experimental investigation presented in this research the former section has been used (Fig. 2). This presents a rectangular cross section 2.5 m × 3.9 m with length of 9 m. Different weather conditions with combinations of wind, rain, snow, freeze and radiant heat can be reproduced. Wind speed can vary between 0.02 m/s and 20 m/s. A radiation system with four infrared lamps with total power of 8 kW and a maximal incidence of 60° to the floor is available. The sprinkler system consists of a 2.3 m × 4.2 m moveable panel (functioning also as the ceiling of this section) to which a spray rack is connected. The rack features three longitudinal lines of sprinklers with a set spacing of 0.65 m. Individual sprinkler supports can be located and fixed by means of screws at different distances along such lines depending on the density of spraying required during testing. The system allows a maximum of 40 sprinklers to be activated. The simulation of wind driven rain in the wind tunnel should ensure a realistic representation of rainfall events in a laboratory environment. It necessitates understanding how the governing input parameters of the tunnel relate to its outputs. In this perspective, a preliminary assessment should be carried out in order to grant the reproduction of a homogeneous volume of rainwater and a natural raindrop size distribution. The assessment performed concentrates mainly on the calibration of wind speed and rainwater intensity inside the tunnel: the first is obtained by measuring with a vane anemometer wind speed against different engine power of the wind generating fan in kW; the second is achieved by monitoring the variation in parameters of the sprinkler system, such as pressure at the nozzle tip (assessed by internal sensors), nozzle type, number of sprinklers active, spray rate and spacing between nozzles, in relation to the rain intensity simulated and its distribution inside the tunnel (mapped using a laser precipitation monitor and a tipping bucket rain gauge). Based on the results of this assessment, it is possible to transform the rainfall characteristics provided by weather data, such as rain intensity and wind speed, into input parameters for the tunnel, such as number of active sprinklers, spacing, pressure, fan power etc. Wind driven rain simulation in the wind tunnel is usually carried out in rain pulses i.e. by activating cyclically the sprinkler system for a determined duration of time. In this perspective, it is suggested to maintain the magnitude of such rain pulses constant throughout the tests while varying only their distribution and density in time. This permits to reproduce a wide range of rain intensities while keeping the same sprinkler layout (i.e. nozzle type, number of

• To establish an adequate testing strategy for the investigation of wind driven rain penetration in masonry elements. • To study a suitable methodology for monitoring the response of masonry specimens subjected to wind driven rain loads. • To establish the characteristics of the impact of wind driven rain on the hygric and mechanical behaviour of brick masonry.

The paper presents the following structure: the first part describes the methods used in the study; the second part outlines the experimental investigation with insights on the test wall and set-up; the following section summarizes the main findings of the research while the conclusive section presents suggestions for future work.

2. Research methodology This research concentrates on three fundamental aspects related to rain penetration testing of brick masonry components: (1) how to test – the feasibility of wind driven rain simulation in the climatic wind tunnel is investigated and a correct testing procedure and test wall specifications are established; (2) how to measure rain penetration – the changes in moisture distribution inside brick masonry components is measured by employing an electrical resistance-based moisture monitoring system and a procedure is developed to convert the results to moisture content values; (3) how to evaluate the response – the impact of rain penetration on the hygric and mechanical behaviour of brick masonry is analysed using gravimetric methods and mechanical testing of brick specimens. Methods employed in this research are further described in the sections below.

Fig. 1. Plan Vincenc Strouhal climatic tunnel. 2

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 2. View inside the climatic section of the tunnel.

the following conditions: firstly it should allow for testing different specimens at once; the specimens composing the model must present a similar exposure to rain loads; as a consequence to similar exposure, the model should prevent corner effects at its sides; the wind blocking effect of the model should avoid building up of strong pressures in the tunnel; finally the model must allow for easy handling and installation inside the climatic tunnel. The model is validated in the tunnel and the above assumptions are also verified, by measuring with a wind driven rain gauge the WDR intensities at different locations on the prototype (Fig. 3, right) and calculating the rain admittance factor.

sprinklers, spacing and pressure). It is clear, in fact, that in operational conditions changing nozzle types and sprinkler distances in the tunnel is not feasible. Hence, an optimal test set-up which can be kept fixed throughout the performance of the rain simulation is individuated at the end of the preliminary assessment of the tunnel. In this study, the optimal set-up for rain simulation is determined as follows: the height of the sprinkler system panel is 2.3 m from ground level, a TG-2.8W nozzle type is used, 6 active sprinklers with 0.65 m × 0.8 m grid size and a pressure at the nozzle tip equal to 120 kPa (200 kPa is also used). The spray rate for such set-up is 0.939 L/min and the rain intensity is equal to 0.6 mm/min. Once defined, the optimal set-up is verified in the tunnel for the actual magnitude and distribution of the rain intensity by carrying out additional measurements. Rain intensity maps are produced by employing a tipping bucket rain gauge at different locations in the wind tunnel. Raindrop size is validated by comparing the actual distribution, taken with a laser precipitation monitor, with simplified theoretical models (such as Dingle and Lee’s). Details concerning the wind tunnel calibration procedure and the rain simulation validation are here omitted for the sake of simplification. In order to support the reliability of the experimental investigation carried out in this research, the extended results from the wind tunnel assessment can be found in a dedicated study published by the author [26]. In addition to the wind tunnel assessment and as part of an initial phase necessary in preparation to the experimental investigation, a model prototype has been also developed and validated in the tunnel. This enables establishing the geometry and size of the test wall (Section 3.1.2) as well as the arrangement of the masonry specimens to be tested inside the tunnel. The proposed prototype (Fig. 3, left) has dimensions 1.9 m × 1.1 m and is composed of 3 solid panels (which represent the locations of the masonry pillars to be tested) and 4 framed ones. Frame panels are made of a wooden frame and a layer of plastic net. The purpose of these components is to prevent large pressures to build-up on the model due to the wind-blocking effect and at the same time to catch a percentage of applied rain. This avoids extensive wetting on the sides of the masonry pillars and it ensures to a certain extent the simulation of the effect of a continuous wall. The prototype must satisfy

2.2. Moisture monitoring system Brick masonry response to WDR is evaluated by monitoring, among other parameters, moisture movements. An innovative moisture monitoring system is proposed and employed in this study. It is based on the measurement of the changes in electrical resistance due to the presence of water. The monitoring system is shown in Fig. 4. It is composed of an electronic board for data measurement applying an alternating 5 V and a brick unit instrumented with epoxy resin and graphite electrodecouple sensors drilled directly into the material. The electrode is composed of a solid copper wire inserted into 6 mm diameter, 30 mm deep hole in the brick and injected with a mixture of Araldite 2020 epoxy resin and graphite powder (43% content by mass). Each electrode presents a measured resistance equal to 170 kΩ at T = 25 °C, which is a reasonable value considering that the range of resistance to be measured varies between 150 MΩ (dry brick) and hundreds of kΩ (wet brick). Further details related to the instrumented bricks used in this study are presented in Section 3.1.3. Each board record simultaneously the resistance detected by seven sensors. It is provided with a USB plug in order to be easily connected to a laptop for data logging and power supply. The monitoring system works in the range 0–300 MΩ with accuracy of ± 1 kΩ. Although the validation of the system, available in another study published by the author [27], is here omitted for the sake of conciseness, its main steps can be summarised as follows: the electronic board 3

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 3. Left: model prototype; right: WDR gauge.

is firstly calibrated with known resistors for ensuring the correctness of the measurement. It is defined empirically 0.5 kHz as the optimal frequency for which the measured resistance is very close to the nominal one of the resistors. It is also determined that at such frequency, the measured resistance values should be numerically corrected for a

parallel impedance of 500 MΩ, due to the capacitance provided by the traces of its input channels, in order to report the correct values of the resistors being measured. Similarly, the adequacy of the electrodecouple sensor is tested by measuring a 10 MΩ resistor, a piece of stranded copper wire and samples of distilled water: the results confirm

Fig. 4. Moisture monitoring system: left electronic board; right instrumented brick and data collection via PC. 4

Engineering Structures 204 (2020) 110080

R. Cacciotti

3.1.1. Moisture content calibration curves are obtained by firstly carrying out a gravimetric analysis of the sorption isotherm of brick prisms during which both the weight and the electrical resistance are recorded for different levels of relative humidity (RH). The geometry factor is calculated for the brick prisms and used to compute the electrical resistivity. Then, temperature coefficients of resistivity are empirically determined in order to report the resistance values to a reference temperature (in this study 20 °C). By plotting the resistivity of brick prisms measured against their moisture content it is possible to obtain the curves. However, these may need one more adjustment before being used for calibration of the measurements taken during the WDR simulation in the wind tunnel. In fact, as it is for this study, the waters used for the sorption isotherm laboratory test (e.g. distilled water) might be different to that used in the WDR simulation. In such case pore water resistivity for both cases should be determined using a water extraction method. By using Archie’s model [28] and substituting in it the pore water resistivity of the wind tunnel, the initial moisture functions plotted can be corrected to moisture calibration curves which can be finally used to convert electrical resistivity results from wind tunnel test.

its ability to report correctly the electrical resistance of resistors and that of distilled water. Thirdly, an experimental validation is carried out in order to investigate the correct activation of the sensors in presence of moisture as well as the ability of the monitoring system, once individuated its presence, to adequately describe the changes in moisture content. The latter objective is achieved by gravimetric method, observing the correspondence between the mass increase of a brick specimen fully immersed in water with the decrease in electrical resistance; the former objective is investigated by a series of capillary uptake tests of full-size bricks, recording visually the water rise and comparing data with the activation time of the sensors (using image analysis and x-ray plus ultrasound method for detecting the rising water level). It is shown that the sensors successfully monitor the increase in moisture content also distinguishing between different moisture regimes inside the pores as well as between changing transport mechanisms (by noticing a change in the rate of resistance decrease). It is observed also a correct performance of the system in detecting, during capillary uptake tests, the arrival times of water at a certain position in the brick by the activation of the sensors. Finally a numerical simulation of the capillary absorption tests is also carried out. This adds to the validation of the system by backing the results achieved during the experimental tests: the activation times at the different positions along the brick calculated by the numerical model are in line with those measured with the system during experiments. This evidences the correct functioning of the proposed moisture monitoring system as well as its ability to describe the moisture content distribution at different times for different positions in the brick. Full results of the system validation are available in [27].

2.3. Response evaluation 2.3.1. Hygric response The method used for the evaluation of the hygric performance of brick masonry during wind tunnel tests is developed by Straube [29]. It focuses on the individuation of fractions of applied water which is shed, absorbed or transmitted across the building element. This method is directly related to the hygric performance of the wall: the amount of shed water reveals the absorption capacity of the material; the amount of water that is stored points out the storage capacity of the material signalling if this falls within a threatening range that may lead to deterioration; the amount of rain that is transmitted through the wall can suggest its sensitivity to water penetration and reveal if any internal layer may also be damaged or degraded. The total amount of rain is known from the simulated rain loads (i.e. the testing criteria employed, see Section 3.3); the rain absorbed is measured gravimetrically by continuously weighting the masonry specimens during testing (pillars are hanged to force transducers, see Section 3.1.3); rain shed is computed as the total rain minus the absorbed water; the rain transmitted to

2.2.1. Conversion of electrical resistance to moisture content The electrical resistance measurements necessitate a number of laboratory activities in order to be transformed into moisture content values, required for the evaluation of the results from the experimental investigation. Electrical resistance values are first converted to electrical resistivity considering different factors which can be empirically quantified: pore water resistivity, geometry factor and temperature effects. Electrical resistivity is finally converted to moisture content by using moisture calibration curves. Fig. 5 presents the moisture calibration curves at a reference temperature of 20 °C obtained for three types of brick (P20, P40 and C) selected for the experimental investigation, whose properties and descriptions are outlined in Section

Fig. 5. Moisture calibration curves for the three types of bricks employed in the experimental investigation. 5

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 6. Mechanical testing set-up for brick specimens: left, compression test; right, flexural test.

atmospheric conditions. The moisture storage functions of the investigated brick types are presented in Section 3.1.1. The exponential functions governing the reduction in the mechanical characteristics are hence determined (see Section 4.3). By monitoring the moisture content distribution with the electronic system presented in Section 2.2, these functions are then used to identify the mechanical effects of rain penetration into the brick masonry specimens tested in the climatic wind tunnel. Hilsdorf’s model, which provides a simple failure criterion for brick under combined compression- bilateral tension [31], is used to determine the reduction in damage and failure allowable stresses in the bricks due to rain penetration (see Section 4.3). Assuming that the deformations of brick-mortar composite at the joints are compatible, the lateral stresses can be determined using Eq. (1) [32]:

deeper layers of the material is assessed using the monitoring system. It is determined as the water which penetrates into the brick up to a depth of 3 cm from its wet face, i.e. considering the relative increase in moisture content at position R2 (see Section 3.1.3 for sensor layout). The transmitted rain in one brick is computed transforming the change in moisture content into the increase of water mass by knowing the total absorption capacity of the bricks (properties presented in Section 3.1.1) and the fraction of brick volume interested by such increase. It assumes a uniform distribution of moisture throughout this deeper layer of the brick. Results of the evaluation of the hygric response of selected brick masonry pillars are shown in Section 4.2. 2.3.2. Mechanical response The change in the compressive and flexural strength as well as in the elastic modulus in compression with increasing moisture content for the brick types are investigated in the laboratory by carrying out mechanical tests of brick units according to the standards ČSN EN 12372 (721145) (Fig. 6). Cube and prismatic specimens are cut from each brick type, described in Section 3.1.1. Cubes of 65 mm side are tested in compression while 25 mm × 30 mm × 290 mm prisms are tested in bending. The compression test is performed with an MTS 250 kN hydraulic load unit with a constant loading speed of 0.45 mm per minute. The load is applied perpendicularly to the bedding face. The compressive strength is calculated as the ratio between the maximum applied force and the net area of the loaded face of the cube. The three-point bending test, with 200 mm spacing between supports, is carried out using an MTS100kN force transducer applying a constant load at a rate of 0.15 mm per minute. The flexural strength σf is instead computed as σf=3Fl/2bh2 where F is the maximum applied force, l is the distance between supports and b and h are respectively the width and height of the prism cross-section. The elastic modulus in compression has been determined by measuring the full field strain over the specimen surface by digital image correlation [30]. This is a novel strain measurement technique which allows capturing a series of images throughout the test and processing the results by means of dedicated software. A 14 × 14 grid of control points has been determined over the cut surface of the specimen in the direction of loading. The relative displacement between two rows of control points falling in the middle of the cube specimen is measured for computing its deformation during testing (rows 5 and 9). Four sets of five brick specimens are tested: the 1st set includes brick specimens in the dry state; the 2nd set is constituted by brick specimens in equilibrium with a 40% RH environment, which represents the humidity level at which adsorbed water molecules begin accumulating in the material (i.e. the hygroscopic regime); the 3rd set comprises brick specimens exposed to a 97% RH atmosphere, which is considered as the humidity level at which the critical moisture content is achieved (i.e. the shift between the hygroscopic and capillary regimes); the 4th set encompasses brick specimens that are saturated with water under

σL =

Eb ·μ Em m h Eb · (1 t Em

− μb

− μm ) + (1 − μb )

σx (1)

where σL is the lateral stress in the brick, σx is the vertical compressive stress, Eb and Em are the elastic moduli of respectively brick and mortar, μb and μm are the corresponding Poisson’s ratios for brick and mortar, t is the thickness of the bed mortar joint while h is the height of the brick. Damage, in the form of vertical cracks, appears when the following criterion Eq. (2) is satisfied:

σL σ + x =1 fbt fb

(2)

where fb is the uniaxial compressive strength of the brick and fbt is the strength of the brick under biaxial tension. Such linear criterion set by Hilsdorf defines that when the bi-lateral tension is zero, failure occurs as the vertical compressive stress reaches the uniaxial compressive strength of the brick while when the vertical compressive stress is zero failure takes place for the exceedance of the biaxial tensile strength in the brick. Whenever the developing internal stresses in the brick increase due to external compression and intersect the line representing the failure function, a local crack occurs. According to Hisldorf the actual point of failure of brick units is reached only when these can no longer provide the minimum lateral tension necessary to restrain the deformation of the mortar joint. Such minimum lateral tension can be determined by the following Eq. (3), based on concrete failure theory:

σL =

1 t · (σx − fm ) 4.1 h

(3)

where fm is the uniaxial compressive strength of mortar. The damage and failure stresses of the investigated brick types P20, P40 and C (see Section 4.3) are determined, following the Hilsdorf’s model, for the different WDR scenarios considered in the experimental investigation. 6

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 7. Brick types employed in the experimental investigation.

and ΦC = 0.279. The vacuum and atmospheric saturations highlight a similar trend: P20 presents a wsat vacuum = 0.239 and wsat atm = 0.204; C has the second largest absorptions with wsat vacuum = 0.141 and a wsat atm = 0.134; P40 shows the lowest moisture storage capacity having wsat vacuum = 0.128 and wsat atm = 0.123. A MIP analysis [34] has been carried out on brick specimens with a MICROMERITICS AutoPore IV 9500 (Figs. 8 and 9). The maximum pressure applied is 228 MPa (33,000 psi). For the P20 brick, the pore size distribution ranges between 0.01 μm and 4 μm, the average total volume of intrusion is 0.208 mL/g, the average median pore diameter is about 1 μm and the average porosity is 33.40%. P40 brick has a pore size distribution varying between 0.02 and 1 μm, while its average total volume of intrusion is 0.122 mL/g, the average median pore diameter is 0.360 μm and its average porosity is found to be 24.50%. Similarly, the C brick presents a pore size distribution ranging between 0.02 and 1 μm, an average total volume of intrusion is 0.133 mL/g, the average median pore diameter is 0.241 μm and the average porosity is 26.2%. It is possible to notice that P20 has the largest volume of pores as well as a wider range of pore diameters: for this brick type, in fact, roughly 90% of the total pore volume falls within 0.1 μm and 3 μm, while the P40 and C samples present a narrower pore distribution with about 88% of volume of pores sized 0.1–1 μm. It should be highlighted that the porosity measured by MIP is usually lower than the one obtained from the vacuum saturation as the former excludes fissures and cavities. Furthermore, the MIP results allow giving a general indication of the susceptibility of bricks to driving rain penetration: while smaller pores induce higher capillary action and hence suction of water, large capillaries exert less resistance to liquid flow inside the media and hence constitute the main route which rain penetration undertakes when entering the brick. The coefficient of capillary absorption A and the water content at capillary saturation wcap are determined using available standards [33]. The amount of water absorbed in time by an initially dry sample of

3. Experimental investigation 3.1. Test wall 3.1.1. Materials Three types of bricks and a premixed mortar are used for building the test wall (Fig. 7). The selection is grounded in the need to simulate typical brick masonry constructions with varying hygric and mechanical parameters. A damaged brick type (brick C) is also selected in order to investigate the effect of faults on the mechanical and absorption response of the material to WDR. Bricks selected for testing, with dimensions 290 mm × 140 mm × 65 mm, include the following types: P20 (nominal compressive strength 20 N/mm2) brick produced by Cihelna Vysoké Mýto (Czech Republic); P40 (nominal compressive strength 40 N/mm2) brick produced by Cihelna Štěrboholy in Prague (Czech Republic); C brick (from a batch of P40 bricks discarded after firing) also produced by Cihelna Štěrboholy, with initial defects such as cracks and fissures. The physical properties of the selected bricks are summarised in Table 1. These are determined using standardised procedures available in the literature [33]. Brick P20 presents the lowest density (ρP20 = 1619 kg/m3) while the P40 and C bricks show higher values (respectively ρP40 = 2025 kg/m3 ρC = 1978 kg/m3). Consequently, P20 results to be the most porous brick among the investigated types, with ΦP20 = 0.388, while the other two porosities found are ΦP40 = 0.260 Table 1 Density, porosity and water absorption of bricks.

P20avg P40avg Cavg

Density (kg/m3)

Φ (%)

wsat

1618.92 ± 7.28 2025.41 ± 11.03 1977.78 ± 29.84

38.83 ± 0.15 26.00 ± 0.50 27.91 ± 1.06

23.94 ± 0.10 12.81 ± 0.32 14.09 ± 0.75

vacuum

(%)

wsat

atm

(%)

20.43 ± 0.39 12.26 ± 0.28 13.38 ± 0.75

7

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 8. Pore size distribution for bricks.

material in contact with water follows the relation m=A√t where m is the mass of water in kg/m2 and time t is in seconds and A indicates the coefficient of capillary absorption. The average A coefficients found are (Table 2): A = 0.119 kg/m2 s0.5 for the P20 brick, A = 0.090 kg/m2 s0.5 for type P40 and A = 0.121 kg/m2 s0.5 for brick C. The water content at capillary saturation (Table 3) constitutes the equilibrium content of specimens in capillary contact with water. Capillary uptake till full saturation is a slow process which is mainly governed by porosity. Having the highest porosity, P20 records a wcap = 0.188 kg/kg followed by brick C with a wcap = 0.132 kg/kg and P40 with a water content at capillary saturation of wcap = 0.117 kg/kg. By measuring the sorption isotherm of brick specimens the moisture storage functions are determined (Fig. 10). These are put in a controlled environment at constant temperature T = 23 °C and changing relative humidity: RH = 97, 83, 68 and 37%. After reaching equilibrium at each RH value, the mass of the specimens is recorded. Following the last

Table 2 Coefficients of capillary absorption. Coefficient capillary absorption A (kg/m2 s0.5): P20

P40

C

0.124 0.121 0.112 0.119

0.096 0.088 0.086 0.090

0.149 0.126 0.089 0.121 Avg

equilibrium point, specimens are dried in the oven at 105 °C and the dry mass is measured. The moisture content is calculated as the difference between the equilibrium water mass content at a specific RH and the dry mass. Bricks have a very low hygroscopic behaviour when compared to other building materials. This is clearly observable from the

Fig. 9. Cumulative intrusion volume for bricks. 8

Engineering Structures 204 (2020) 110080

R. Cacciotti

Table 3 Water content at capillary saturation.

Table 4 Moisture storage and transport properties.

Water content at capillary saturation wcap (kg/kg): P20

P40

C

0.185 0.182 0.197 Avg 0.188

0.117 0.114 0.121 0.117

0.141 0.128 0.127 0.132

3

w80 (kg/m ) wcap (kg/m3) Wvacuum (kg/m3) A (kg/m2 s0.5) Dws 80 (m2/s) Dww 80 (m2/s) Dws cap (m2/s) Dww cap (m2/s)

functions found for the P20, P40 and C types. However, some distinctions can be made. Taking 97% RH as the theoretical critical point between the hygroscopic and the capillary regimes, the P20 brick shows a moisture content equal to about 14% of its vacuum absorption while the P40 and C specimens present equilibrium moisture content of respectively 5 and 6.5%. Although the capillary transport remains the predominant mechanism in all of the brick types analysed, brick P20 shows a higher capacity to transport moisture in its adsorbed form having a larger pore internal surface to which water molecules can attach to. The liquid transport coefficient Dw defines the capacity of materials to transport water per unit time (Table 4). The liquid transport coefficient for suction Dws describes the capillary uptake of water when the surface is fully wet. The transport is dominated by larger capillaries that present lower flow resistance; once the supply of water is removed, i.e. when the surface of the material starts drying, the liquid transport coefficient for redistribution Dww comes into play. A mathematical commonly used method for approximating the values of Dw with respect to the moisture content of the material is given in the literature [35–37]. The Dws curves in Fig. 11 are obtained considering the first point occurring at w80 (for RH 80%), taken as the theoretical moisture content at which liquid transport occurs (i.e. no liquid transport is observed in the hygroscopic regime). The function increases linearly till the capillary saturation content wcap. For higher moisture contents up to wvacuum the value of the liquid transport coefficient equals to the one for capillary saturation. Similarly, the Dww curves are obtained using as a starting point the same coefficient value Dws at moisture content w80 and the redistribution liquid transport coefficient at capillary saturation, calculated as Dww=10Dws [35], which is kept constant for higher moisture contents. The highest rate of increase in Dws value with respect to moisture content is found for brick P20, which having larger

P20

P40

C

23 304 387 0.119 9.85E−10 9.85E−10 5.82E−07 5.82E−08

4 238 259 0.090 6.17E−10 6.17E−10 5.43E−07 5.43E−08

6 261 279 0.121 9.65E−10 9.65E−10 8.17E−07 8.17E−08

capillaries opposes less flow resistance. On the other hand, for the same reason of presenting larger pores, P20 has a liquid transport coefficient at capillary saturation (5.82E−07 m2/s) similar or lower than those calculated for P40 and C bricks (respectively 5.43E−07 and 8.17E−07 m2/s). As previously mentioned the redistribution coefficients Dww are an order of magnitude lower than the Dws. It is possible to notice that the slope of the curves for this parameter is less steep compared to the suction coefficients narrowing down the range of Dww achievable as the moisture content increases in the material. Finally, the mortar used for the construction of the pillars is the PCI Pecicret K01 produced by BASF s.r.o in the Czech Republic. This is a premixed two binder material commonly used for building masonry structural elements. The selection of this material represents a trade-off between the modern cement based materials which offer better mechanical characteristics and 28 days hardening time and the traditional lime mortars which confers a more porous and permeable pore structure. Mortar properties are presented in Table 5. The standardized nature of the product and the knowledge of its material properties ensures a controlled performance of the mortar joints during the simulated rain testing. The material shows mechanical and hygric properties compatible with the brick types used. The next section outlines the characteristics of the test wall and the different phases of its construction. 3.1.2. Geometry and specifications The geometry and composition of the test wall, shown in Fig. 12, follow the design specifications set for the model prototype in Section 2.1. The pillars are separated by framed panels with similar height and width. The aim of the framed panels is to protect the exposed sides of

Fig. 10. Moisture storage functions. 9

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 11. Liquid transport coefficients.

1 cm head and bedding mortar joints. One course is made of two bricks laid in parallel which are rotated at 90° with respect to the course below and above. As shown in Fig. 13, pillars are built on a metal plate which is provided with three ring bolts for lifting the construction. The instrumented bricks are embedded in the masonry element and the wires of the sensors are bent toward the back side of the test wall (i.e. windward). Once the construction of the pillar is finished and adequately cured, a plastic layer is put on its side and top faces in order to ensure unidirectional rain penetration. The pillar is finally provided with a metal structure composed of rods connecting the metal plate at the bottom to a T frame at the top which is then finished with a hook bolt. Such assembly permits the pillars to be hanged to a force transducer for monitoring the weight increase due to rain absorption. More details are provided in the following sections.

Table 5 Mortar properties. Mortar properties Strength class A (kg/m2 s0.5)

M5 0.039

Bulk density dry (kg/dm3) Bulk density wet (kg/dm3)

1.5 1.8

3.1.3. Sensor layout The sensor layout is presented in Fig. 14. Pillars are hanged to a force transducer which continuously measures their weight in order to establish the mass of water absorbed by each pillar (GTM 5 kN series K transducers are used for pillars P40 and C while a LUKAS S-22 10 kN for P20). A strip sensor, 15 cm in length, is used for monitoring temperature and is placed during construction in the bedding joint at midheight of the middle pillar (i.e. the P20 pillar). Due to unknown reasons the sensor failed before testing. Hence, temperature has been measured at different times during testing using an infrared camera on both the rain and back sides of the test wall. Also temperature sensors internal to the tunnel have been employed to monitor temperature changes inside the climatic section. Two sets of instrumented bricks have been embedded in each masonry pillar. One set of two bricks (a rain side brick and a back side one) is laid, counting from the bottom, at the second course of the masonry elements (made of 15 courses) and another set at the 14th course. The sensor layout for the instrumented brick section, depicted in Fig. 15, shows a denser presence of sensors in the rain side brick with a finer grid close to the wet surface. The bricks used for the monitoring of moisture are provided with a number of sensors distributed along their width and depth aligned perpendicularly to the direction of rain penetration. This permits comparing the difference between rain penetration at the bottom and at the top of the elements (due to higher WDR at the top and greater run-off at the bottom, for example). The orientation of the instrumented bricks, with the long edge perpendicular to the direction of the water flow, is chosen for the

Fig. 12. Test wall.

the pillars and to restrain their movement when subjected to torsion due to wind turbulence. The smaller panels at the extremities of the test wall allow avoiding corner effects influencing the measurement at the lateral pillars, ensuring hence a similar WDR distribution on the masonry elements. It should be underlined that, in order to allow for relative vertical displacement, the connection between the framed panels and the masonry elements is pinned. Each masonry pillars have a cross-section of 29 cm × 29 cm and a height of about 110 cm. They are composed of 15 brick courses with 10

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 13. Side waterproofing and lifting arrangement of the pillar.

Fig. 15. Instrumented bricks.

masonry section, easing the unbalance in moisture distribution due to the WDR hitting one side of the pillar. The sensors named from R1 to R4 are located on the rain side brick, while those labelled from R5 to R7 are on the back side brick. Starting from the wet face of the rain side brick, the sensors are found at the following depths: R1 at 1.5 cm, R2 at 3 cm, R3 at 7 cm, R4 at 12.5 cm, R5 at 16.5 cm, R6 at 22 cm and R7 at 27.5 cm. 3.2. Test set-up The test wall is installed inside the climatic section of the Vincenc Strouhal tunnel (Fig. 16) at a distance of about 1.45 m from the first row of sprinklers. It is symmetrically aligned with respect to the tunnel centreline, having the axis of the middle pillar corresponding to the centre of the tunnel section. The three brick masonry elements are hanged by means of steel rods to a 2.3 m × 3 m steel frame, at about 10 cm height above its bottom beam. One of the framed panels located between the pillars is made detachable, in order to allow for inspecting the back of the setup during testing. Wooden beams are provided at the rear side of the test wall in order to keep it in position within the plane of the frame and avoid any dangerous oscillatory movements during the

Fig. 14. Sensor layout.

presence of the mortar joint between the two units that acts as a barrier for moisture transport. This implies that critical moisture profiles, i.e. larger differentials in moisture contents within the pillar section, are more likely to occur with such orientation. On the contrary, the orientation featuring the short side of the brick exposed on the wetting face of the pillar would be strongly influenced by the presence of the vertical joint. This would rapidly drive moisture deep inside the 11

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 16. Test wall installed inside the wind tunnel.

Fig. 17. Backside of the test wall. Left: wooden beam and plexiglass plate at the top of the pillar; right: waterproofing of electronic components.

average duration of the observed rainfall at the selected location, as well as on operational requirements for the simulation in the tunnel, the duration of each event-based test is set to 3 h. Similarly, the average amount of rainfall per event is estimated to be 7.2 mm. The performance-based testing criteria are purposely designed to be representative of an extreme rainfall event in Prague with a return period of 10 years, grounded on the Straube and Burnett approach [24] and on the Choi method to determine the short duration driving rain intensity and driving rain wind pressures [25]. The calculated parameters for the performance-based test include a spray rate of 2.9 L/min with a wind speed varying between 8 and 11 m/s (Table 9). The performance-based test is carried out in cycles of 15 min each, for a total duration of 10 cycles (in line with the 3 h limit set for the climatic tunnel). The testing protocol used is as follows: starting from dry conditions, the test wall is exposed to simulated wind driven rain as per event 1 for duration of 3 h. A period of no rain IET (Inter Event Time) [38], not shorter than 16.5 h, follows. This is usually taken as 1 day in order to better fit the daily working hours at the facility. After the IET, the test wall undergoes the wind driven rain load designed for event 2 for duration of 3 h. The simulation is once more followed by a 24 h stop of the climatic tunnel. Event 3 is then simulated for another 3 h followed by the IET. The penetration test is finally performed. This features ten 15-minute cycles (plus 30 s at the beginning of each cycle). At the end

test. The top of the pillars presents a shielding plexiglass plate which prevents water runoff along its back face (Fig. 17, left). This construction detail keeps the back side of the masonry elements dry, thus ensuring a unidirectional flow of rain penetration, starting at the rain side brick. Electronic boards, sensors and transducers are waterproofed by using a plastic wrapping (Fig. 17, right).

3.3. Testing criteria and protocol Two sets of testing criteria are proposed in this research for the rain penetration testing of brick masonry specimens in the wind tunnel, namely the event-based (Fig. 18) and the performance-based test criteria (Fig. 19). Weather data collected at the UTAM station in Prague, Czech Republic have been processed, with the intent to establish the predominant rainfall trends. From such observations, three main eventbased tests are proposed: event 1 (Table 6), characterizing a storm in which 80% of the rain is discharged in just 15% duration of the event, followed by a long period of drizzling (remaining 20% of rain falling in 85% of the event duration); event 2 (Table 7), presenting a stepped trend in which the rate of rainfall remains more or less constant during the whole duration of the storm; event 3 (Table 8), defining the trend of events inverted with respect to event 1 meaning long light rain at the beginning and heavy shower at the end of the event. Based on the 12

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 18. Event-based test criteria: main trends (thicker lines) and recorded weather data (thin lines).

4.1. Moisture content results

of each load, 10–15 min are required in order to perform technical checks. Following the application of the 10 cycles, the test is finished. The weight of the pillars, their electrical resistance at instrumented positions and temperature changes are continuously monitored throughout the whole sequence of tests i.e. from the beginning of event 1 simulation to the end of the performance based test. The wind tunnel set-up (optimal set-up) for the simulation is outlined in Section 2.1. Rain reproduction is usually carried out in rain pulses i.e. by activating cyclically the sprinkler system for a determined duration of time. Rain pulses specifications are provided in the tables (Tables 6–9).

As introduced in Section 2.2, the electrical resistance values monitored during testing need to be first corrected numerically for the electronic board parallel impedance of 500 MΩ; secondly they are converted into resistivity values by multiplying it by the geometry factors and finally the temperature correction coefficients (empirically determined for each brick type) are applied to the actual temperature measurements (Fig. 20), to report the resistivity results to the reference temperature T = 20 °C. The moisture content is finally determined by transforming the monitored resistivity using moisture content calibration curves (Section 2.2.1, Fig. 5). Important observations related to the performance of the masonry pillars can be drawn from the WDR tests: the characteristics of rain penetration are found to significantly vary depending on the porosity and damage state of the brick units; top and bottom sections of the pillars experience a divergent moisture distribution during rain

4. Results and discussion The wind tunnel simulations carried out in this study provide interesting insights on the performance of brick masonry subjected to wind driven rain. The following sections present the moisture content results together with the hygric and mechanical response evaluation.

Fig. 19. Performance-based test criteria. 13

Engineering Structures 204 (2020) 110080

R. Cacciotti

Table 6 Event 1 test specifications. Event 1 duration:

3h 180 min

Time (%)

(min)

Rainfall (%)

(mm/h)

(mm/min)

6S @120 kPa (0.6 mm/min)

Wind speed ranges (m/s)

0–15 15–30 30–45 45–80 80–100

0–27 27–54 54–80 80–144 144–180

0–85 85–85 85–90 90–90 90–100

13,8 0 0,9 0 1,25

0,23 0 0,015 0 0,02

11 × 1 min pulse – 1 × 40 sec pulse – 1 × 70 sec pulse

7–3,5/5,5–3/3–1,5 1,5–3 4,5–2/6–3 1,5–3 4–2/2,5–1,5

observations can be summarised:

simulations; near surface, corner positions of the sensors present a distinctive response to rain loadings with respect to deeper locations in the middle of the width of the pillars; dramatic differences are observed in the moisture content distribution between the rain side brick and the back side one. From the moisture content results, pillar P20 appears to be the most sensitive to WDR in terms of rainwater penetration and wind drying (Figs. 21 and 22). Its bottom part shows higher moisture content when compared to the top section due to a greater availability of water for penetrating the brick (i.e. run-off and splashing of raindrops). It is noticed that the middle position in the pillar (i.e. R2, see layout in Section 3.1.3) presents greater values of moisture content as opposed to the point R1. This is due to the susceptibility of the R1 location (at the corner of the pillar and near to the surface) to wind drying as well as to the less amount of deposited rain at this point due to the wind flow deflection of raindrop trajectories. Pillar P40 is, as expected, far more watertight than the P20 pillar being less porous (Figs. 23 and 24). The denser brick structure and the different pore distribution reduce in fact its sensitivity to the dynamic effects observed in pillar P20, inducing thus a slower response to WDR. Moisture content values at the top and bottom of the P40 pillar do not show big differences with the exception of point R2 at the bottom part which is mostly influenced by the presence of a crack (as well as R3 in the same section). This indicates that even if more water is available to ingress the pillar at the bottom section (due to run-off and splashing) this is not absorbed. In pillar C it is also noticeable, both at the top and bottom sections (Figs. 25 and 26), the lower values of R1 as opposed to R2: the moisture content at R1 is smaller as being at the corner less incident rain hits the exposed face and, at the same time, the moisture content at the top (e.g. at R2 and R3) is much higher due to the presence of cracks. Pillar C is composed of initially damaged brick units. This is the reason why this pillar absorbs more moisture than the P40 pillar and its distribution is less predictable: moisture is transported deep into the section (for example at positions R3 and R4) as well as no significant difference is observed between the moisture content at the top and bottom sections. From the moisture content results the following general

– Porous bricks, such as P20, appear to be more sensitive to WDR loads, experiencing considerable wetting from impinging rain as well as drying effects from wind at the surface. Less permeable and denser units, such as P40 and C, present instead a slower response to rain loads, with the exception of localised faults in the material which could significantly increase the amount and rate of penetration of rainwater. In fact, cracks are observed to have a predominant role in the transport of moisture across the brick units (as for example seen for brick type C): if exposed to strong WDR loads even less permeable bricks can experience rain penetration deep inside the material due to fissures and faults, inducing, therefore, a less predictable moisture distribution within the masonry section. – For a strong rainfall, the saturation occurring at the surface of the exposed face of the pillars can have a strong influence on the penetration of rainwater, by creating a barrier effect which prevents further moisture to ingress the masonry elements. This induces a variable amount of water run-off at the masonry surface, producing hence a higher rain load at the bottom of the pillars with respect to the top. If the masonry units are capable of absorbing such increased load of water, larger penetration takes place at the bottom. This effect is, of course, more accentuated in porous brick units (i.e. P20), whose residual absorption capacity at the bottom allows for further rainwater penetration, while it is hardly observed in denser materials (P40 and C). – WDR induces a heterogeneous distribution of moisture not only along the depth of the masonry section but also along its width with lower moisture content being monitored at the corners of the pillar as opposed to deeper positions in the middle of its width. This is attributable to the susceptibility of the brick, at the former position, to wind drying as well as to the minor quantity of deposited rain due to the wind flow deflection of raindrop trajectories. More porous bricks, such as P20, react more dynamically to these effects with respect to the other brick types investigated which show instead no particular response.

Table 7 Event 2 test specifications. Event 2 duration:

3h 180 min

Time (%) 0–10 10–15 15–25 25–30 30–40 40–80 80–85 85–100

(min)

Rainfall (%)

(mm/h)

0–18 18–28 28–47 47–56 56–75 75–148 148–157 157–180

0–25 25–25 25–50 50–50 50–75 75–75 75–90 90–100

6,1 0 6,1 0 6,1 0 7,3 1,9

14

(mm/min)

6S @120 kPa (0.6 mm/min)

Wind speed ranges (m/s)

0,1 0 0,1 0 0,1 0 0,12 0,03

3 – 3 – 3 – 2 1

4–2 1,5–3 4–2 1,5–3 4–2 1,5–3 4–2 1–2,5

× 1 min pulse × 1 min pulse × 1 min pulse × 1 min pulse × 1 min pulse

Engineering Structures 204 (2020) 110080

R. Cacciotti

Table 8 Event 3 test specifications. Event3 duration:

3h 180 min

Time (%)

(min)

Rainfall (%)

(mm/h)

(mm/min)

6S @120 kPa (0.6 mm/min)

Wind speed ranges (m/s)

0–20 20–55 55–70 70–85 85–100

0–36 36–100 100–126 126–153 153–180

0–10 10–10 10–15 15–15 15–100

1,25 0 0,9 0 13,8

0,02 0 0,015 0 0,23

1 × 70 sec pulse – 1 × 40 sec pulse – 11 × 1 min pulse

4–2/2,5–1,5 1,5–3 4,5–2/6–3 1,5–3 7–3,5/5,5–3/3–1,5

The main factors influencing rain penetration discussed above are further discussed in Section 4.2.

Table 9 Performance test specifications. Cycle duration:

15 Spray rate (l/m2/ min)

2.9

min Sprinkler

6S coupled @ 200 kPa

Wind speed (m/s) Min

max

8

11

4.2. Hygric response results As presented in Section 2.3.1, the hygric performance of the test wall is evaluated by assessing the fractions of applied water which is shed, absorbed or transmitted across the pillars. Result examples from such analysis are shown in Figs. 27 and 28. Considering the activation times of the sensor response to rain loads, determined at a depth of 3 cm from the exposed face (i.e. position R2, see Section 3.1.3), interesting effects can be observed. Fig. 29 shows the response activation times versus the rain load rate for the three brick types studied. The general trend is that these reduce with increasing load rates (calculated from rain pulses in Section 3.3). This is true in particular for initially dry conditions when rain is firstly absorbed and then driven into the masonry. It can be observed that the rate at which activation time decreases is a function of the brick type: in the case of P20 the activation time reduces at a rate which is double that found for P40 and 50% more the rate noticed for brick C. This

– Moisture penetrated inside the pillars concentrates mainly at the rain side of the building elements leaving the back side units in much drier conditions. This induces relevant differentials in moisture content inside the masonry sections implying distinct processes to occur (e.g. differential degradation, redistribution of stresses, etc.). The magnitude of the variation in the moisture content within the same masonry section is shown to be dependent on the porosity of the brick units and on the position considered (e.g. top or bottom, corner or middle).

Fig. 20. Example of temperature measurement results with thermography: Event 3. 15

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 21. Moisture content results at section P20 top.

the larger is the amount of absorbed water and its rate. Whenever one of these properties is exceeded by the intensity and duration of the rainfall, shedding or leakage occur. If the storage capacity of the brick is not backed by an ability of the material to dry at a certain rate or by design details of the wall assembly to enable draining of the absorbed water, the building up of moisture inside the porous material can induce damages. Furthermore, the porosity of bricks along with the pore structure and connectivity represent a good indicator of the permeability of the material to water penetration. The larger the volume of interconnected capillaries available for the water flow, the faster is the rate of penetration and its magnitude into the bricks. It should be underlined that additionally to the absorption, storage and transport properties also the surface texture can partially influence the response to rain penetration. In addition, it is observed a strong influence between the rain load

indicates how the porosity of the former brick type, composed of larger capillaries, enables faster rain penetration into the deeper layer. Nevertheless it can be also highlighted that in the case of the C brick, even if falling at a slower rate, activation times are considerably lower than the other two bricks. This effect can be better understood by analysing the rain penetration speed in time of the different bricks (Fig. 30). After an initial speed lower than P20, the one recorded for brick C spikes up to a maximum of over 35 mm/min followed by a steep fall down to a value similar to the one observed for brick P40. This peak represents the effect of cracks which transmit rapidly water into the brick for a short period of time after which it becomes saturated and the effect is almost nullified. The results from the response assessment indicate that the absorbance and storage capacity of the bricks play a crucial role in the way they react to wind-driven rain. The higher the value of these properties,

Fig. 22. Moisture content results at section P20 bottom. 16

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 23. Moisture content results at section P40 top.

well as by the speed at which the brick can absorb water (which reduces with time). Concerning the penetration of rain into the bricks it can be said that the magnitude of the rain loads is directly proportional to the amount of water penetrated, however this reduces with increasing rate of application of the load due to shedding at the surface. Summarising the results, it can be said that when the bricks are subjected to a higher magnitude of rain loads, they store and transmit a larger fraction of water; if the bricks are exposed to a higher rain load rate, these shed a larger fraction of applied rain (i.e. they store less), they absorb faster and transmit less. It should be considered that the effects of rain loads here discussed are very variable in time depending on the saturation at the surface as well as on the wetting history of the wall assembly. The effect of the wetting history observed during the WDR tests is dependent on the specific testing protocol employed for the tunnel simulations. In particular, the comparison between the response of the investigated bricks to the effects of event 1 and event 3 allow

characteristics and the response of the masonry elements. In terms of the response magnitude of the rain penetrated into the masonry (assessed as kilograms of water absorbed), it is shown that there is a linear relationship with the rain load magnitude (Fig. 31). The amount of water which penetrates the deeper layer of the bricks in fact increases with the amount of rain deposited. This does not consider the rate of loading but only the amount of water to which the brick is subjected. Conversely when the time distribution of the rain load is taken into consideration (Fig. 32), it can be observed that the response magnitude decreases with the rain load rate of application as probably more water is likely to be lost by shedding at the surface. It can be concluded that the characteristics of the rain loads are very relevant in the way masonry absorbs and transmits rain. A larger fraction of milder rainfall is more likely to be absorbed by the bricks if compared with heavier precipitation. This is governed by the larger bouncing and shedding of rainwater during high rainfall intensities as

Fig. 24. Moisture content results at section P40 bottom. 17

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 25. Moisture content results at section C top.

previous wetting and drying cycles it is possible to evaluate the actual and expected performance of building elements. Surface saturation affects the way rainwater is absorbed and transmitted across the masonry elements. As moisture content increases at the surface, it creates a barrier effect which prevents further rain to enter the material while at the same time triggering the penetration of stored water deeper into the unit. This represents a behaviour typically observed in the wetting of porous material in which to the initial saturation of larger pores follows a redistribution of moisture in smaller capillaries. Another important effect of the saturation of the surface on the response of the brick concerns its relation to the transmission of water. It is in fact clear that once the surface becomes saturated, i.e. when the brick has the brick storage capacity has been reached at that location, water penetration occurs. From Fig. 34 it is possible to observe how the response magnitude at the deeper layer in the bricks increases

concluding on such effect. The two events are constituted by same rain loads inversely distributed in time. Event 1 is performed on initially dry masonry specimens while event 3 is tested on the same specimens with different initial conditions, having been exposed already to event 1 and event 2. The effect of the wetting history on the absorption and transmission of rain is observed to be very much related to the properties of the brick discussed in Section 3.1.1. Fig. 33 shows the speed of rain penetration (which is calculated as the activation time of the response over the distance of the deeper layer considered from the wet face, i.e. 3 cm) versus the rain load rate for the three brick types and the two events considered (i.e. event 1 and event 3). It is noticeable a relaxation of activation times among the events due to the higher moisture content recorded at the beginning of event 3. The results highlight that the wetting and drying history is therefore fundamental in the assessment of the response of masonry to wind-driven rain. By understanding the

Fig. 26. Moisture content results at section C bottom. 18

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 27. Example of event based test results (rain load 0.135 kg): absorbed, shed and transmitted rain for brick.

degradation of strength and stiffness parameters with increasing moisture content (expressed as a percentage of the saturated moisture). For brick P20 it is determined that in dry and saturated conditions the compressive strength varies between 12.02 and 5.31 MPa (Table 10) corresponding to a 56% decrease, the flexural strength ranges between 4.3 and 2.9 MPa (Table 11) revealing a 33% reduction, while the elastic modulus in compression changes from 2.6 (dry) to 1.2 GPa (saturated) underlining the occurrence of a 54% decrease (Table 12). Similarly brick P40 is shown to change its mechanical parameters with increasing moisture content as follows: in dry and saturated states the compressive strength fluctuates between 26.18 and 20.61 MPa (Table 10) corresponding to a 21% reduction, the flexural strength instead varies between 7 and 5.7 MPa (Table 11) denoting a 19% decrease and finally the elastic modulus ranges between 12.5 and 10.3 GPa (Table 12) showing a reduction of 18%. For brick C the measured variation in mechanical parameters in dry and saturated conditions includes: the compressive strength changes between 25.93 and 19.46 MPa (Table 10) corresponding to a 25% decrease, the flexural strength also reduces by

with time: such increase is more pronounced during the second load cycle and slowly reduces during the following ones. The occurrence of surface saturation depends mainly on the rate of the rain load applied and on the absorption and storage capacity of the bricks. This defines how fast and how much rain can be taken by the material before the process is halted. It is possible to conclude that as the surface of the brick becomes saturated, the absorption of the rainwater slows down and eventually stops. The fraction of water that is shed increases noticeably. The penetration of rain into the deeper layer of the bricks usually occurs immediately after the slowing down of the absorption phase. Once rain penetration occurs, the rate of transmission of water increases significantly with time up to a maximum point at which saturation is locally achieved. 4.3. Mechanical response results The results from the mechanical characterisation of the brick specimens (Figs. 35–37), discussed in Section 2.3.2, shows an explicit

Fig. 28. Example of performance based test results (rain load 16.278 kg): absorbed, shed and transmitted rain. 19

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 29. Sensor activation time at depth 3 cm at different rainfall rates.

Fig. 30. Peak rain penetration speed following saturation at surface.

parameters in line with the lower bound of the ranges outlined for standard brick units. Brick types P40 and C instead reach higher strength and stiffness values which still fall within the standard ranges set. However, it can be noticed that lower results for the mechanical properties of brick C with respect to brick P40 are obtained, due to the presence of cracks in the tested specimens. In particular, the flexural strength of this brick type is the lowest among the investigated masonry units. It is interesting to notice that, for all types of brick studied, the most significant reduction in mechanical parameters occurs for moisture contents between the dry state and roughly 20%. More specifically, it is possible to notice that 80% of the overall reduction in compressive strength and elastic modulus determined for brick P20 takes place in a range of moisture contents up to 15%; 60% and 73% of the overall reduction in respectively compressive strength and elastic modulus for brick P40 occurs for moisture contents up to roughly 6%; similarly 87%

25% ranging between 2.16 and 1.62 MPa (Table 11) while the elastic modulus in compression varies between 11.5 and 9 GPa presenting a reduction of 22% (Table 12). The calculated coefficients of variation show that dispersion of data is more consistent for results related to flexural strength and elastic modulus with respect to those concerning compressive strength. In fact, flexural strength test results are very sensitive to initial defects and inhomogeneity of test specimens. In this perspective, it possible to observe how the CV calculated for brick type C (selected with initial defects) can be significantly high. The values of the mechanical parameters for the tested brick types are in line with the ranges provided in the literature for standard brick units [39–41], which are: 3–21 MPa for compressive strength (up to 100 MPa for particular brick types), 1–5 MPa for flexural strength and 3–34 GPa for the modulus of elasticity. Brick P20 presents the lowest compressive strength and elastic modulus among the brick types investigated; being a porous, soft brick it attains values of mechanical 20

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 31. Response versus load magnitude.

Fig. 32. Response magnitude versus rain load rate.

to WDR as the operational moisture contents of brick masonry building components are usually found within the upper hygroscopic and lower capillary regimes. As shown by the moisture content results from the WDR tests (Section 4.1), rain penetration generates a highly unbalanced moisture distribution inside the brick masonry units with larger divergence observed between the wetting face and the deeper dry area of the material. A 3-D moisture content surface inside the brick unit is obtained by spreading the data recorded at different sensor positions using the model shown in Fig. 38. The corresponding distribution of mechanical properties is then determined as a function of the moisture content (Fig. 39), using the relationships individuated by the mechanical testing of the brick types (Figs. 35–37). A number of profiles induced by WDR testing in the investigated pillars are established. For each masonry pillar tested, three moisture content profiles at the mid-width section of

of the total reduction in compressive strength and 72% of the overall decrease in elastic modulus found for brick C develops for moisture content ranging from dry state to 7%. As highlighted by Cherblanc et al., the strength of porous materials is very sensitive to the water content and an increase in water content of as little as 1% from the dry state can have a marked effect on strength. In fact, the authors notice that the loss of mechanical strengths is almost maximal at 97% RH [42]. Considering the equilibrium at 97% RH, the critical moisture content for brick P20, P40 and C is respectively 14.5%, 5% and 6.5% (see Section 3.1.1). All these values are very close to the moisture contents at which the bricks experience the most significant reduction in mechanical parameters. Therefore this indicates that such considerable decrease in strength and stiffness begins in the lower bound of the hygroscopic regime up to the shift to the capillary regime. This finding is of extreme relevance in the performance of real structures subjected 21

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 33. Rain penetration speed versus rain load rate for two consecutive event-based tests.

Figs. 40 and 41 present respectively the damage and failure envelopes due to rain penetration in the selected brick types. The upper bound of the envelopes is determined by horizontal solid lines, referring to the stress distribution along the mid-width section of a uniformly dry unit; the lower bound is delimited by the solid lines representing the worst stress conditions induced by the critical profiles; the dashed curves falling within the envelope define intermediate stress states produced by other critical profiles as individuated in Table 13. The damage envelope (Fig. 40) describes the compressive and lateral tensile stresses necessary to reach local cracking in the brick unit. The damaging stresses found vary between 4.6–7 MPa compression and 0.4–0.8 MPa lateral tension for brick P20, between 13.3–16.5 MPa compression and 2–2.4 MPa lateral tension for brick P40 and between 7.4–9.3 MPa compression and 1.1–1.3 MPa lateral tension for brick C. Considering the simulation of real rainfall characteristics (i.e. event

the rain side brick unit are considered (Table 13). Such profiles represent the extreme differentials observed in the brick types at the end of either event-based or performance-based rain simulations. Such extreme distributions might induce a sufficient distortion of the mechanical behaviour of the masonry units, eventually leading to cracking or failure. The differential in moisture content implies a correspondent reduction in the mechanical properties producing the redistribution of compressive stress along the cross-section of bricks and the lowering of the thresholds of damage and failure stresses (determined as in Section 2.3.2) with respect to dry brick units leading, in particular scenarios, to damage or even failure. Based on the failure criteria of each brick type, it is possible to determine the changing failure conditions along the mid-width section of the units, due to the critical moisture profiles induced by WDR testing.

Fig. 34. Rainwater storage with time. 22

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 35. Compressive strength versus moisture content.

constant up to a brick depth of 6 cm and then increasing exponentially till the end of the unit section; in brick P40, at brick depth equal to zero, damaging stresses are found to be 13.3 MPa (81% of dry). and 2 MPa (82% of dry) for respectively the compressive and the lateral tensile stress, growing exponentially from the wetting face of the brick up to the end of the section depth; for brick C, at brick depth equal to zero, damage takes place for a compressive stress of around 7.4 MPa (79% of dry) and for a lateral tension around 1.1 MPa (79% of dry). The failure envelope (Fig. 41) defines the stress conditions, both compressive and tensile, which induce the loss of the capacity of the units to restraint mortar deformations, leading therefore to brick failure. The stresses determined for the investigated brick types range between 5.2–8.5 MPa compression and 0.1–0.2 MPa lateral tension for brick P20, between 19.5–23 MPa compression and 0.6–0.7 MPa lateral tension for brick P40 and between 14.8–17.3 MPa compression and 0.4–0.5 MPa lateral tension for brick C. Event-based tests induce the following failure stress conditions in the units: for brick P20 (denser

based-test) the following is observed: for brick P20 (denser dotted curve), at brick depth equal to zero (i.e. wetting face), damage occurs for a compressive stress close to 5.8 MPa (83% of dry) and for a lateral tension around 0.6 MPa (77% of dry), increasing linearly up to a brick depth of 14 cm; in brick P40 (dashed curve), at brick depth equal to zero, damaging stresses equal to 14.3 MPa (87% of dry) and 2.1 MPa (88% of dry) for respectively the compressive and the lateral tensile stress, growing at a constant rate up to the end of the section depth (i.e. 14 cm); for brick C (denser dotted curve), at brick depth equal to zero, damage takes place for a compressive stress of around 8.4 MPa (90% of dry) and for a lateral tension around 1.2 MPa (91% of dry), increasing linearly up to a brick depth of 14 cm. If the heavier rainfall simulation is analysed (i.e. performance-based test) the damaging stresses are identified, for all brick types, by the lower bound of the envelopes: for brick P20, at brick depth equal to zero, damage occurs for a compressive stress close to 4.6 MPa (66% of dry) and for a lateral tension around 0.4 MPa (54% of dry), remaining

Fig. 36. Flexural strength versus moisture content. 23

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 37. Elastic modulus in compression versus moisture content. Table 10 Results for compressive strength.

Table 11 Results for flexural strength.

Compression

Flexural

P20 Spec

CP40

C

P20

Strength (MPa)

Spec

Strength (MPa)

Spec

Strength (MPa)

SET 1 (DRY) P20-a P20-b P20-c P20-d P20-e Average CV (%)

12.130 11.604 10.661 11.432 14.294 12.024 11

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

25.843 26.384 25.132 27.367

C-D-a C-D-b C-D-c C-D-d C-D-e

28.955 27.650 29.866 19.971 23.191 25.927 16

SET 2 (40% RH) P20-a P20-b P20-c P20-d P20-e Average CV (%)

8.340 9.575 10.278 9.839 11.193 9.845 11

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

SET 3 (97% RH) P20-a P20-b P20-c P20-d P20-e Average CV (%)

8.891 4.177 7.738 6.660 6.209 6.735 26

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

SET 4 (SATURATED) P20-a P20-b P20-c P20-d P20-e Average CV (%)

6.134 6.072 3.392 5.551 5.390 5.308 21

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

26.182 4

26.419 27.010 24.570 25.999 5 23.513 21.160 22.122 24.740

C-D-a C-D-b C-D-c C-D-d C-D-e

C-D-a C-D-b C-D-c C-D-d C-D-e

22.884 7

18.996 25.317 17.503 20.605 20

C-D-a C-D-b C-D-c C-D-d C-D-e

Spec

SET 1 (DRY) P20-a P20-b P20-c P20-d P20-e Average CV (%)

P40 Strength (MPa)

Spec

Strength (MPa)

Spec

Strength (MPa)

3.055 1.980 4.341 4.326 3.413 3.423 29

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

8.242

C-D-a C-D-b C-D-c C-D-d C-D-e

1.867

SET 2 (40% RH) P20-a 3.654 P20-b P20-c 3.107 P20-d 3.257 P20-e 2.800 Average 3.204 CV (%) 11

19.952 26.431 21.165 21.161 22.177 13

17.159 20.309 22

SET 3 (97% RH) P20-a 2.850 P20-b P20-c P20-d 3.092 P20-e 3.120 Average 3.020 CV (%) 5

20.407 22.622 14.002 20.267 20 19.459 17

SET 4 (SATURATED) P20-a P20-b P20-c 3.039 P20-d 2.900 P20-e Average 2.970 CV (%) 3

23.458

24

C

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

7.273 6.898 6.625 7.259 10 6.206

7.391 6.799 12 7.939 3.805

5.668 5.804 36

5.520 5 5.124 4.478 5.031 6

2.156 2.448 2.157 14

C-D-a C-D-b C-D-c C-D-d C-D-e

2.630 1.958 3.095 3.185 2.717 21

C-D-a C-D-b C-D-c C-D-d C-D-e

1.772 1.105

1.438 33 C-D-a C-D-b C-D-c C-D-d C-D-e

2.730 0.501 1.6163 98

Engineering Structures 204 (2020) 110080

R. Cacciotti

dotted curve), at brick depth equal to zero (i.e. wetting face), failure occurs for a compressive stress of 6.8 MPa (80% of dry) and a lateral tensile stress of 0.1 MPa (79% of dry), increasing linearly up to a brick depth of 14 cm; in brick P40 (dashed curve), at brick depth equal to zero, failure stresses equal to 20.6 MPa (90% of dry) and 0.7 MPa (96% of dry) for respectively the compressive and the lateral tensile stress, growing at a constant rate up to the end of the section depth (i.e. 14 cm); for brick C (denser dotted curve), at brick depth equal to zero, failure takes place for a compressive stress of around 16.2 MPa (94% of dry) and for a lateral tension around 0.5 MPa (96% of dry), increasing linearly up to a brick depth of 14 cm. Concerning the simulation of performance based tests, for all brick types, failure stresses present the worst scenario, coinciding with the lower bound of the envelopes: for brick P20, at brick depth equal to zero, failure occurs for a compressive stress close to 5.2 MPa (62% of dry) and for a lateral tension around 0.1 MPa (63% of dry), remaining constant up to a brick depth of 6 cm and then increasing exponentially till the end of the unit section; in brick P40, at brick depth equal to zero, failure stresses are found to be 19.5 MPa (85% of dry) and 0.6 MPa (94% of dry) for respectively the compressive and the lateral tensile stress, growing exponentially from the wetting face of the brick up to the end of the section depth; for brick C, at brick depth equal to zero, failure takes place for a compressive stress of around 14.8 MPa (86% of dry) and for a lateral tension around 0.4 MPa (91% of dry). From the results outlined above, it can be observed that for both damage and failure stresses rain penetration imposes a reduction with respect to the dry brick conditions. The wet side of the masonry units present in fact much lower capacity to carry compressive and tensile stresses: as low as 60% of the allowable stresses in dry units can induce cracking in brick P20 and just 62% of dry stress conditions are sufficient to reach failure of the unit; 82% of stresses producing damage in a dry P40 brick can do so in the maximum moisture content determined for this brick type while 90% of dry stress state is required to induce failure of the unit; in brick C 79% of stresses required in dry units can form damage while 88% of dry stresses would force the brick to fail. Rain penetration induced by common summer storms (event-based test) can increase the probability of brick units to fail, depending on their hygric and mechanical properties: softer and porous bricks, such as P20, are observed to diminish their damaging and failure stresses to around 80% of those assessed for a dry unit, which in combination with other degradation processes (and accompanied by a significant stress redistribution in the section) could indeed lead to serious damage. Conversely denser brick units, such as P40 and C, for the effects of event-based rain penetration only show no such issues, presenting a reduction in the damaging and failure stresses well above 90% of the dry brick. During heavy rainfall events instead rain penetration is much deeper and stronger and depending on the material characteristics, damaging and failure stresses can deteriorate considerably with respect to dry conditions: porous units, such as P20, lower their safety thresholds in terms of cracking and failure by as little as 60% of a correspondent dry unit while less porous bricks, such as P40 and C, experience a reduction in their failure stresses of around 90% with respect to a dry unit. These observations highlight the importance of rain penetration effects on the durability and structural integrity on brick constructions. In fact, considering the synergy occurring in real life situations among degradation processes due to rain penetration, such as the reduction in mechanical properties, the redistribution of compressive stress due to moisture differentials within the unit section and the lowering of damaging and failure stresses with respect to dry units, together with additional adverse actions, such as volumetric expansion, salts and freezing temperature can pose serious threats to existing structures.

Table 12 Results for elastic modulus in compression. Younǵs modulus P20

P40

C

Spec

E (GPa)

Spec

E (GPa)

Spec

E (GPa)

SET 1 (DRY) P20-a P20-b P20-c P20-d P20-e Average CV (%)

2.439 1.886 3.050 1.428 4.502 2.661 45

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

17.124 10.090

C-D-a C-D-b C-D-c C-D-d C-D-e

14.953

SET 2 (40% RH) P20-a P20-b P20-c P20-d P20-e Average CV (%)

2.408 1.772 3.090 3.218 1.271 2.352 36

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

SET 3 (97% RH) P20-a P20-b P20-c P20-d P20-e Average CV (%)

2.398 0.536 1.202 2.217 1.051 1.481 53

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

SET 4 (SATURATED) P20-a P20-b 1.520 P20-c 0.386 P20-d 1.487 P20-e 1.409 Average 1.200 CV (%) 45

P40-D-a P40-D-b P40-D-c P40-D-d P40-D-e

10.257 12.491 32 16.704 12.172 10.122 13.438 5.917 11.670 34

C-D-a C-D-b C-D-c C-D-d C-D-e

8.272 12.619

C-D-a C-D-b C-D-c C-D-d C-D-e

9.231 13.757 10.970 24

12.342 7.359 11.202 10.301 25

7.557 12.058 11.523 32 10.014 10.210 9.521

9.915 4

C-D-a C-D-b C-D-c C-D-d C-D-e

11.365 8.590 9.977 20 9.732 4.487 12.886 8.920 9.006 38

Fig. 38. Top: sensor positions; bottom: moisture distribution model.

25

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 39. Top: example of moisture content surface; bottom: example of correspondent elastic modulus distribution.

contextualised to a specific geographical scenario providing a more accurate investigation of localised moisture related issues. – Test wall specifications and test set-up are proposed following the validation of a physical model in the wind tunnel. These specifications enable to test different specimens at once by ensuring similar exposure to rain loads and the prevention of wind blocking effect. They also foster easy handling and installation inside the climatic tunnel.

Table 13 Critical moisture content profiles selected. Profile name

During test

Occurring at minute

4P20B 4P20T 2P20T 3P40B 2P40B 1P40T 4CT 3CB 1CB

Performance Performance Event 2 Performance Event 3 Event 1 Performance Performance Event 1

750 750 360 750 540 50 750 750 20

Furthermore, the results from this study provide relevant advancements in the understanding of brick masonry response in terms of mechanical and hygric behaviour. Wind driven rain simulations highlight the main elements which govern rain penetration: one of the most relevant factors is found to be the properties of the bricks, namely the water absorbance and moisture storage capacity; the other fundamental element affecting the response is the rain load applied to the bricks, characterised by its magnitude and duration; the saturation level at the brick surface plays an important role in terms of water absorption and transmission; lastly, the wetting history of the bricks proves to be relevant, as it changes the properties of the brick with time. Other factors that concur in defining the brick masonry response to WDR include wind only effects, construction details, crack and faults in the material

5. Conclusions and future work The main aspects of innovation stemming out from this research include: – New testing methodology for assessing the response of brick masonry to wind driven rain and for the evaluation of experimental results. – Compared to available standards, rain testing criteria are 26

Engineering Structures 204 (2020) 110080

R. Cacciotti

Fig. 40. Reduction of allowable damage stress due to rain penetration.

Fig. 41. Reduction of allowable failure stress due to rain penetration.

combination with salts might produce sufficient levels of salt concentration in the pore system of bricks further reducing the allowable strength of the material and eventually leading, by wetting and drying cycles, to salt crystallization. Synergic effects should therefore be considered as a key element in the assessment of the durability and structural performance of brick masonry constructions. Indeed, the evaluation of hygric and mechanical effects of wind driven rain only, carried out in this research, has been fundamental in order to understand which factors might be manipulated by professionals in the perspective of defining an optimal design, treatment or retrofitting strategy for brick building components. Following the main findings of this study, a framework for desirable future work can be drafted. It appears to be particularly crucial the necessity to obtain additional wind tunnel measurements for further validating the proposed testing criteria and protocol as well as the test

or run-off. Although rain penetration, even during milder rainfall events, can induce durability problems, such as the formation of local damage (i.e. initiation of cracks close to the exposed face of bricks), it is unlikely that its deterioration action alone would be sufficient to undermine the structural integrity of constructions. In real structures, however, the synergy between moisture-induced degradation and additional adverse conditions, such as temperature and salt actions, produces a combined damaging effect greater than the sum of their separate effects which could be indeed structurally dangerous. In fact, while moisture profiles generated by the intrusion of atmospheric precipitation are observed to induce the degradation in mechanical properties, stress redistribution and lowering of damaging and failure stresses in bricks, the combined effect with freezing temperature might lead to volumetric changes in the units as well as to freeze and thaw cycling. Furthermore moisture in 27

Engineering Structures 204 (2020) 110080

R. Cacciotti

wall specifications and its experimental set-up. Future work should be also focused on the experimental analysis of the influence of wind speed on the rate of rainwater penetration in porous materials and how this relate to capillary absorption forces. Finally, concerning the mechanical degradation in porous materials induced by moisture, the effects of over-pressure in pores needs further research in order to identify its relevance in real moisture content conditions, i.e. in operational environmental conditions usually found between the upper hygroscopic and lower capillary regimes for masonry structures subjected to wind driven rain. Additionally, testing conditions, such as the length of saturation of specimens, are also found to have an important role in mechanical degradation and further research should be concentrated to the understating of their impact on the experimental results.

[16] Inculet D. The design of cladding against wind-driven rain doctoral thesis London: The University of Western Ontario, Canada; 2001 [17] Bitsuamlak GT, Chowdhury AG, Sambare D. Application of a full-scale testing facility for assessing wind-driven-rain intrusion. Build Environ 2009;44:2430–41. [18] Knasiak K, Schick RJ, Kalata W. Multiscale design of rain simulator. ILASS–Americas, 20th annual conference on liquid atomization and spray systems, Chicago, USA. 2007. [19] Baheru T. Development of test-based wind-driven rain intrusion model for hurricane-induced building interior and contents damage FIU Electronic Theses and Dissertations Paper 1127 Miami: Florida International University, USA; 2014. [20] ASTM E331. Standard test method for water penetration of exterior windows, skylights, doors and curtain walls by uniform static air pressure difference. USA: American Society for Testing and Materials; 2000. [21] BS EN 12155. Curtain walling–watertightness–laboratory test under static pressure. UK: British Standards Institution; 2000. [22] Van Den Bossche N, Lacasse MA, Janssens A. A uniform methodology to establish test parameters for watertightness testing Part I: a critical review. Build Environ 2013;63:145–56. [23] Sahala N, Lacasse MA. Proposed method for calculating water penetration test parameters of wall assemblies as applied to Istanbul, Turkey. Build Environ 2008;43:1250–60. [24] Straube JF, Burnett EFP. Simplified prediction of driving rain on buildings. The international building physics conference, Eindhoven, the Netherlands. 2000. p. 375–82. [25] Choi ECC. Criteria for water penetration testing. Water leakage through building facades. ASTM STP 1998;1314:3–16. [26] Cacciotti R, Pospíšil S, Kuznetsov S, Trush A. A proposed calibration procedure for the simulation of wind-driven rain in small-scale wind tunnel. Exp Techn 2019;4:369–84. [27] Cacciotti R, Valach J, Wolf B. Innovative and easy-to-implement moisture monitoring system for brick units. Constr Build Mater 2018;186:598–614. [28] Archie GE. The electrical resistivity log as an aid in determining some reservoir characteristics. Trans Am Inst Min Metall Pet Eng 1942;146:54–67. [29] Straube JF, Burnett EFP. Vents, ventilation and masonry wall systems. 8th Canadian masonry symposium, Jasper, Canada. 1998. p. 194–207. [30] McCormick N, Lord J. Digital image correlation. Mater Today 2010;13:52–4. [31] Hilsdorf HK. An investigation into the failure mechanism of brick masonry loaded in axial compression. In: Johnson FB, editor. Designing, engineering/engineering and constructing with masonry products. Houston (Texas): Gulf Publishing; 1969. [32] Bakker T. Advanced chapters of structural design-failure models of masonry under compression. Germany: Dresden TU; 2012. [33] Wilson MA, Carter MA, Hoff WD. British Standard and RILEM water absorption tests: a critical evaluation. Mater Struct 1999;32:571–8. [34] Webb PA. An introduction to the physical characterization of materials by mercury intrusion porosimetry with emphasis on reduction and presentation of experimental data. Norcross (Georgia): Micromeritics Instrument Corp; 2001. [35] Van Belleghem M, Steeman M, Janssen H, Janssens A, De Paepe M. Validation of a coupled heat, vapour and liquid moisture transport model for porous materials implemented in CFD. Build Environ 2014;81:340–53. [36] Krus M, Holm A. Simple methods to approximate the liquid transport coefficients describing the absorption and drying process. The 5th symposium building physics in the Nordic countries, Goteborg, Sweden. 1999. p. 241–8. [37] Nikitsin VI, Backiel-Brzozowska B. Methods of determination of liquid transfer coefficient in building materials. Int J Heat Mass Tran 2012;55:4318–22. [38] Joo JG, Lee JH, Jun HD, Kim JH, Jo DJ. Inter-event time definition setting procedure for urban drainage systems. Water 2014;6:45–58. [39] Narayanan SP, Sirajuddin M. Properties of brick masonry for FE modelling. AJER 2013;1:6–11. [40] Bošnjak-Klečina M, Lozančić S. Testing of physical and mechanical properties of bricks and mortar in historic structures. Tehni ki vjesnik 2010;2:209–15. [41] Brooks JJ, Ajmad MA. Elasticity and strength of clay brickwork test units. 8th IB2Mac, Dublin, Ireland. 1988. p. 342–9. [42] Cherblanc F, Berthonneau J, Bromblet P, Huon V. Influence of water content on the mechanical behavior of limesonte: role of clay minerals content. Rock Mech Rock Eng 2016;49:2033–42.

Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References [1] Straube J. Moisture control and enclosure control systems doctoral thesis Waterloo: The University of Waterloo, Canada; 1998 [2] Gentilini C, Franzoni E, Graziani G, Bandini S. Mechanical properties of fired-clay brick masonry models in moist and dry conditions. Key Eng Mat 2015;624:307–12. [3] Blocken B, Abuku M, Roels S, Carmeliet J. Wind-driven rain on building facades: some perspectives. EACWE 5, Florence, Italy. 2009. p. 19–23. [4] Drdácký M, Cacciotti R, Slížková Z. Dangerous synergies causing failures of historical structures. IABSE symposium vancouver 2017: engineering the future, Vancouver, Canada. 2017. p. 1058–65. [5] Straube JF, Burnett EFP. Rain control and screened wall systems rain. The 7th conference on building science and technology, Toronto, Canada. 1997. p. 17–37. [6] Castellazzi G, Colla C, de Miranda S, Formica G, Gabrielli E, Molari L, et al. A coupled multiphase model for hygrothermal analysis of masonry structures and prediction of stress induced by salt crystallization. Constr Build Mater 2013;41:717–31. [7] Pavlíková M, Pavlík Z, Černý R. Hygric and thermal properties of materials used in historical masonry. The 8th symposium on building physics in the Nordic countries, Lyngby, Denmark. 2008. p. 903–10. [8] Witzany J, Cejka T, Zigler R. The effect of moisture on significant mechanical characteristics of masonry. Eng Struct Technol 2010;2:79–85. [9] Brocken HJP, Adan OCG, Pel L. Moisture transport properties of mortar and mortar joints: a NMR study. Heron 1997;42:55–69. [10] Derluyn H, Janssen H, Carmeliet J. Influence of the nature of interfaces on the capillary transport in layered materials. Constr Buil Mater 2011;25:3685–93. [11] Groot CJWP, Gunneweg JTM. The influence of materials characteristics and workmanship on rain penetration in historic fired clay brick masonry. Heron 2010;55:141–53. [12] Flower JW, Lawson TV. On the laboratory representation of rain impingement on buildings. Atmos Environ 1972;6:55–60. [13] Rayment R, Hilton M. The use of bubbles in a wind tunnel for flow-visualisation and the possible representation of raindrops. J Ind Aerodyn 1977;2:149–57. [14] Aaron KM, Hernan M, Parikh P, Sarohia V. Simulation and analysis of natural rain in a wind tunnel via digital image processing techniques. AlAA 24th aerospace sciences meeting, Reno, Nevada, USA. 1986. [15] Surry D, Inculet DR, Skerlj PF, Lin JX, Davenport AG. Wind rain and the building envelope: a status report of ongoing research at the University of Western Ontario. J Wind Eng Ind Aerodyn 1994;53:19–36.

28