Damage identification in concrete under impact loading at varying temperatures using voltage strain relations technique: an experimental and numerical study

Impact-loaded concrete structures cause severe and rapid damage, resulting in significant property and human life loss. As the temperature rises, the damage caused by impact loading becomes increasingly severe. Concrete structures need structural health monitoring (SHM) to avoid this damage and loss. In this study, the voltage strain relation technique was used to identify the damaged state of concrete under impact loads at various temperature conditions experimentally and numerically. For this purpose, an experimental study was performed on concrete cube specimens in which different piezo configurations (surface bonded, non-bonded, and jacketed) were installed to acquire the voltage data. Before applying an impact load to the top surface of the concrete specimen, it was preheated at 50 °C, 100 °C, and 150 °C to provide the temperature effect, and then a free-falling iron ball was dropped from 3 m heights on the top of the specimens. Furthermore, finite element analysis has been carried out to validate the experimental results with analytical results. The experimental results show that the voltage strain relation technique is well capable of detecting the damage in concrete under the temperature and impact loading conditions. The maximum absolute voltage value (Vp) of 17.11 V was recorded for the jacketed sensors under an impact height of 3 m at 100 °C. All the piezo sensor configurations are capable of finding the damage. Jacketed sensors are more efficient in the health assessment of concrete in terms of voltage strain relations. In terms of strain values, the analytical results are in good agreement with the experimental results.


Abstract
Impact-loaded concrete structures cause severe and rapid damage, resulting in significant property and human life loss. As the temperature rises, the damage caused by impact loading becomes increasingly severe. Concrete structures need structural health monitoring (SHM) to avoid this damage and loss. In this study, the voltage strain relation technique was used to identify the damaged state of concrete under impact loads at various temperature conditions experimentally and numerically. For this purpose, an experimental study was performed on concrete cube specimens in which different piezo configurations (surface bonded, non-bonded, and jacketed) were installed to acquire the voltage data. Before applying an impact load to the top surface of the concrete specimen, it was preheated at 50°C, 100°C, and 150°C to provide the temperature effect, and then a free-falling iron ball was dropped from 3 m heights on the top of the specimens. Furthermore, finite element analysis has been carried out to validate the experimental results with analytical results. The experimental results show that the voltage strain relation technique is well capable of detecting the damage in concrete under the temperature and impact loading conditions. The maximum absolute voltage value (Vp) of 17.11 V was recorded for the jacketed sensors under an impact height of 3 m at 100°C. All the piezo sensor configurations are capable of finding the damage. Jacketed sensors are more efficient in the health assessment of concrete in terms of voltage strain relations. In terms of strain values, the analytical results are in good agreement with the experimental results.

Introduction
In recent years, structures have become more vulnerable to terrorist attacks, blast waves, Earthquakes, industrial accidents, etc High loading rates often cause fire outbreaks and spread, exposing structures to harsh conditions. Limited studies have demonstrated the assessment of concrete damage caused by an impact load at various temperatures. The majority of studies only show the individual impact and temperature loading effects on concrete structures. Many areas of damage identification for pre and post-heated concrete under impact loading remain unexplored. The voltage strain relation technique has emerged as an advanced technique in structural health monitoring (SHM) in recent years due to its wide application in the field of piezoelectric materials. When exposed to impact loads, concrete structures primarily fail in two ways: globally and locally. Local substantial failure is commonly caused by crushing, plug development, and scabbing, whereas global failure is caused by bending. The best method for examining the impact behaviour of concrete materials was the drop impact test [1]. Bischoff et al [2] examined and explained concrete's behaviour under both static and impact loading conditions. The responses of different loading conditions were compared using strain value, strength and energy-absorption capacity. The behaviour of the slab, when subjected to impact loading has been investigated by several researchers [3][4][5][6][7][8][9][10]. Youmn Al Rawi et al [11] investigated the behaviour of post-tensioned concrete slabs under the effect of impact load. For the purpose of analyzing the damage behaviour, two different grades of post-tensioned concrete slabs were tested at a height of 20 m and a weight of 605 kg. Many researchers [12][13][14][15][16] investigated the impact load analysis of beams. Jin et al [17] tested 12 different types of simply supported beams with various stirrup ratios and steel fibre volume fractions to examine the impact resistance behaviour of steel fibre reinforced concrete beams. Additionally, a three-dimensional finite element simulation was performed, and it was found that the addition of steel fibres improved the steel fibre beam's capacity to withstand impacts.
Many researchers have already studied the effect of low and high temperatures on concrete [18][19][20][21][22][23][24][25]. One of the most significant advantages of concrete as a building material is its ability to withstand high temperatures. The disturbance caused by high temperatures in concrete manifests itself as surface cracking and spalling [26]. According to Wang et al [27], because of anisotropy and porosity, thermal distribution and transmission in concrete are typically non-uniform and time-dependent, causing a negative correlation between temperatures and concrete Young's modulus/natural frequencies. Under conditions of higher temperatures, structural elements like beams and slabs show greater signs of deterioration [28,29]. Researchers have recently become increasingly focused on detecting the many sorts of strength behaviour and damage situations in concrete, such as Sher. [30] , cracking pattern [31,32], fracture performance [33], reduction due to explosion [34], corrosion monitoring using piezo sensors [35][36][37][38][39] and strength parameters [40] under the static and dynamic loading conditions.
Researchers have recently emphasized the behaviour of concrete structures element (beam and slab) under both loading condition impact and temperatures. The behaviour of concrete structures was studied by varying some key parameters of impact loading and temperature, such as the mass of the impactor, velocity of the impactor, geometry of the impactor, and at higher degrees of temperature. The complete literature review of the concrete structures subjected to both impact and temperatures is shown in table 1.
1.1. Finite element analysis (FEM) of concrete structures under temperature and impact loading FEM analysis has emerged as a superior technique for analyzing concrete structures in recent years because of its accurate interpretation and faster results. Due to these advantages, most researchers now employ the FEM technique to analyze concrete structures subjected to impact loading, thermal loading, and the combined effect of impact and thermal.
This section shows the literature review for the FEM analysis of a concrete structure subjected to thermal loading conditions. Many researchers have demonstrated that FEM analysis produces significant results when analyzing concrete under thermal loading conditions [44,45]. The strengthening effect of glass fibre and polypropylene fibre-based engineered cementitious composites (GFPPECC) on fire-damaged reinforced concrete short exterior columns was explained by Bhuvaneshwari et al [46], and the FEM analysis performed by ANSYS was compared with the experimental results and found that there was no deviation in both the experimental and numerical results. Song et al [47] simulated the mass concrete freezing shaft lining and frozen soil wall under the influence of low temperatures by using ANSYS finite element software.  With more steel fibres, SFRC beams fail in flexure under impact loads, not shear.Steel fibre content has a minimal impact on the thermal and mechanical behaviour of damaged beams, limited to low-energy preimpact.
Velocity of impact = 4 m s −1 , Three different temperature variations = 200°C, 500°C and 800°C 80% degradation of concrete strength with an increase in temperature 80% degradation of concrete strength with an increase in temperature [43] Ganesan et al [48] used ANSYS explicit dynamics software to perform FEM analysis of the RC and prestressed RC slab at low-velocity impact under two different boundary conditions, the opposite side simply supported and the adjacent side simply supported. Their analysis demonstrated that the boundary condition significantly impacted the slab's deflection. Singh et al [10] performed the FEM analysis of the conventional and geopolymer reinforced concrete beam and slab under the low and high-velocity impact using ANSYS explicit dynamics software; the results shows that the geopolymer concrete performs better impact resistance behaviour in comparison to the conventional reinforced concrete.
FEM behaviour of a concrete structure subjected to thermal and impact loading was comprehensively studied [49]. The author also performed a three-dimensional finite element study of the RC slab when it was destroyed by fire and then loaded by impact. The fire effect was simulated using transient 3D FE thermomechanical analysis, whereas the impact loading was simulated using explicit multi-body dynamic analysis. According to the study's findings, the simulation can match the experimental results. Jin et al [50] investigated the effect of elevated temperature on dynamic compressive properties of heterogeneous concrete using a mesoscale numerical study. They divided the study into two steps, first simulating heat conduction behaviour and then simulating the dynamic mechanical behaviour of concrete using the first stage's output as the initial conditions. Previous researchers employed dynamic strain measurement utilizing the voltage strain relation approach to locate damage in concrete structures under impact load conditions and discovered that this voltage strain relation technique is quite capable of finding damage under impact load situations [51,52].
The purpose of this research is to use various piezo configuration sensors to detect damage in concrete underimpact loading conditions at varying temperatures. For this purpose, experimental studies were carried out to extract the voltage data from these various sensors. Further, to validate the experimental results, finite element analysis was performed on a concrete sample under free boundary conditions (FBC).

Background information for the research
This section discusses the research background, beginning with the dynamic strain measurement using PZT sensors and progressing to an introduction to the finite element method (FEM) analysis for the impact loading condition.

Dynamic strain measurement using PZT sensors
Materials with piezoelectric properties PZT has the ability to sense changes in its environment, such as changes in geometry, and mechanical or electrical properties, and respond quickly, which can be easily measured. The PZT operates under the effect of piezoelectricity, which has the ability to convert electrical responses into mechanical disturbances and mechanical disturbances into electric pulses. These effects are represented mathematically, with equation (1) demonstrating the sensing behaviour of the PZT and equation (2) demonstrating the actuation behaviour. One dimensional interaction between PZT Patch and the host structure is shown in figure 1.
Where S 1 represents the strain value in direction '1'. D 3 is the electric displacement over the PZT patch, the piezoelectric strain coefficient is d 31 , E 3 is the external electric field and T 1 is the axial stress in direction '1'. Y E is the complex young's modulus of elasticity of the PZT patch at a constant electric field as shown in equation Where η and δ represent the mechanical loss factor and the dielectric loss tangent respectively. As shown in equation (5), Shanker et al [67] developed a relationship between the voltage across the PZT patch and strain in the host structure at the point of attachment.
Where V represents the voltage generated across the PZT patch of the thickness h. The output voltage can be measured using the instrument oscilloscope and with the help of this measured voltage, strain-induced can be calculated using the above equation (5).

Experimental setup
This section covers the entire methodology starting with the fabrication and installation of different piezo configurations such as jacketed piezo sensors (JKTPS), surface-bonded piezo sensors (SBPS), and non-bonded piezo sensors (NBPS) in concrete specimens. Further, the experimental setup consisting of testing age, and collection of data procedures have been explained. SBPS sensors, as shown in figure 2(b) were directly bonded on the concrete using epoxy at the centre on the side face of the concrete adjacent to the impactor face, and the upper face of the bonded PZT was again covered with epoxy to protect it from the environment. The NBPS, as shown in figure 2(a), was fabricated by using a surface-cleaned aluminium strip (100 mm × 10 mm × 3 mm) and a PZT patch (10 mm × 10 mm × 0.2 mm) fixed at the middle of the aluminium strip. Epoxy was used as a binding material to bond aluminium sheets and PZT, and special care was taken to ensure that they were adequately bonded. Furthermore, after connecting with the aluminium strip, the PZT patch was soldered with the coaxial wire on the positive and negative electrodes, and an epoxy layer was applied above the soldered PZT after soldering to protect the PZT patch from environmental conditions. On the concrete side face adjacent to the impactor face, NBPS sensors were installed using adhesive epoxy.

Experimentation details and procedure
In this study, a standard 150 × 150 × 150 mm concrete cube with M30 grade was used, which has a 28-day strength of 30 N mm −2 . J.K Super ordinary Portland cement (OPC) grade 43 was employed as a binding material. Two varieties of coarse aggregate (10 mm and 20 mm) were used, as well as locally available wellgraded fine aggregate that comes within Zone II of the Indian standard code (IS: 2386 (Part I)-1963) [53]. The casting of the concrete cube was done in the ratio of 1:1.84:3.48(Cement: Fine aggregate: Coarse aggregate) in a standard mould with proper mixing in the pan mixture as shown in figure 3(a) and compaction is done by vibrating table and left the mould for 24 h as shown in figure 3(b). After 24 h, all of the concrete moulds were demoulded and placed in a curing tank as shown in figure 3(c) for 28 days for normal curing. Furthermore, until the samples were tested, they were all kept in a laboratory. The entire concrete cube sample is cast in the laboratory, which includes the embedded and non-bonded sensors type as well as the standard cube. Impact loads were applied from a 3 m height onto a concrete cube using a 5.45 kg free-falling spherical cast iron ball. The cast iron ball with a diameter of 11.5 cm falls through a PVC pipe with a thickness of 4 mm and an internal diameter of 15 cm. The PVC pipe is provided to guide the spherical ball for proper impact positioning on the concrete cube. All the concrete cubes were pre heated in the rectangular oven provided by Khera Instruments Pvt Ltd of size 605 × 605 × 910 mm before the impact. The experimental setup for testing the concrete cube under impact loading is depicted in figure 4, which also includes a description of the experimental setup.
In this study, the damage was detected in the reference to voltage data recorded at the time of testing the cube. The first cube was subjected to three different temperatures, ranging from 50°C, 100°C and 150°C, heated in the oven for 120 min then cooled for 60 min at room temperature. Furthermore, cooled cube samples were tested at the 3 m impact height and voltage data were recorded after each impact up to failure.

Finite element modelling of the concrete sample
Two bodies are involved in finite element analysis, the concrete cube and the steel impactor. In the first stage, a concrete cube is subjected to thermal loading, which includes heating and cooling at room temperature. The transient thermal module in ANSYS was used to perform the thermal analysis, and the impact load analysis was performed using the explicit dynamics module in ANSYS.  Table 2 shows the material and thermal properties at ambient temperature that were used in the finite element analysis in detail. Steel impactors with a specific gravity of 6839 kg/m 3 were used, the same as in the experiment. Steel impactors were assumed to be linear elastic materials in the FEM analysis, with Young's modulus Es = 210 GPa and Poisson's ratio μ = 0.30.

Description of elements and meshing
For the modelling of the concrete cube in this FEM analysis, hexahedral solid elements were used. Frictionaltype body interactions were used in the explicit dynamics module between the host structure (concrete cube) and the impactor (spherical ball).In this study, a 15 mm mesh size was used for both the concrete and the   impactor, with the element mid-nodes dropped at the time of meshing. Figure 5 depicts the finite element model of a concrete sample with and without meshing.

Boundary condition
In this study, concrete cubes were simulated with free boundary conditions (FBC), which were provided as same in the experiment. Figure 6 depicts the FBC applied to a concrete cube with just the bottom portion of the cube fixed and the other remaining faces of the cube free.

Loading condition
As mentioned in the previous section, concrete cubes that had previously been tested experimentally were simulated for a first drop of impact under thermomechanical and explicit dynamics loading conditions. In thermomechanical analysis, the initial thermal loading is given in terms of an initial temperature of 22°C to be applied to the entire body of a concrete cube. There were two steps of analysis performed in thermomechanical analysis. In the first step, the concrete cube was heated for two hours using the convection method and then cooled down for one hour in which a linear decrement of ambient temperature occurs as in the experiment. The concrete cube was heated at three different temperature conditions of 50, 100 and 150°C same as in the experiment.
After thermal exposure, concrete cubes were subjected to the spherical steel ball impact loading at the top surface of the concrete cube at the centre. Impact loading simulations were performed for the same height of 3 m as the experiment. The impact load is generated by free fall dropping of the steel ball of weight 5.45 kg mass from the velocity of 7.67 m s −1 .

Experimental results and discussion
This section is divided into two parts. First, discuss the experimental results of the voltage-time history and then go over the results of the finite element analysis (FEM) in detail.

Quantification of impact load using voltage-time history
In general, the impact load time histories are very minute, and hence, the loading's distinguishing features are highly evasive. A high acquisition rate digital oscilloscope was used to solve this problem (model Tektronix TBS 2104B series), which captured the impact value for intervals as short as 1 μs. The voltage generated by the PZT patch attached to the host structure was essentially captured by the oscilloscope. Voltage measurements were taken at a high acquisition rate, with acquisition times ranging from 500 ms to 2 s. The voltage-time histories were thoroughly examined to correlate the accumulating damage with strain. A total of typical 12 sensors were used to determine the effect of voltage with the impact loading under temperature variation for different sensors configuration. Peak voltage was recorded for each impact loading and then absolute peak voltage (V p ) was identified for every sample of concrete cube. Furthermore, this obtained V p were used to calculate the accumulated value of strain for each sensor. Figure 7 shows a typical variation of the voltage vs time graph for four different cube conditions. Figure 7(a) depicts the voltage variation for cube C9, SBPS at ambient temperature conditions following the second impact, in which the voltage value first increases to 0.104 V and then decreases to 0.080 V. The graph in figure 7(b) shows the same pattern as the previous one, which depicts the voltage vs time graph for cube no. C5 after the second impact. Figures 7(c) and (d) depicts the voltage variation for cube no. C12 and under C8 after the 3rd and 2nd impact respectively for SBPS and JKTPS sensors at the temperature variation of 150°C, in both of them the voltage value first decreased and then increased.
Quantification of impact force is represented with the help of obtained peak voltage from each sensor at the time of impact and it was found that the peak voltage follows a continuously decreasing trend for the surface bonded and non-bonded sensors with the increase in the number of impact loads on each cube. This decreasing trend of peak voltage is due to the increase in damping of the cube. Peak voltage has followed the opposite trend for the JKTPS because they have been placed in the line of impact loading as a result strain development is more in this case. Table 3 displays all of the typical 12 cubes, which include the representation number, number of impacts up to failure and Vp value. Furthermore, strain values were calculated using equation (6) with the help of PZT parameters and a Vp value. Table 3 depicts the Vp value for each cube, which is the maximum peak voltage recorded against the impact loading. Thevalue Vp of 17.11 V was recorded for the JKTPS sensors under an impact height of 3 m at 100°C, this is due to the Placing of JKTPS sensors in the line of impact. Furthermore, after the first impact, the peak voltage and strain value of each PZT installed in the cube were calculated as shown in table 4.

FEM result and discussion
The results of the thermomechanical and dynamic simulations are shown and discussed in this section. Figures 8(a)-(d) illustrates the distribution of elastic strain at a 3 m height of impact with temperatures ranging from ambient to 150°C. Maximum strain values of 0.0074913, 0.0077198, 0.0081335, and 0.0085585 were observed at different variable temperatures of ambient, 50,100, and 150°C, respectively, and they clearly illustrate that the strain value was temperature-dependent and that it goes up with the increase in temperature. The distribution area of elastic strain became similar from ambient to 100°C temperature. Figure 8(d) shows the distribution of elastic strain at 150°C temperature, it was clearly indicated that the area of distribution of elastic strain was affected by the temperature, as the temperature is high elastic strain distribution was also in the larger size and if the temperature was low the area of distribution of elastic strain was found to be less. This was due to the loss of the stiffness of the concrete behaviour at the higher temperature.
The elastic strain time history curve of the concrete cube was also obtained for 3 m height of impact at temperatures ranging from ambient temperature to 150°C. Figures 9(a)-(d) depicts the variation of the elastic strain time history curve under the height of 3 m impact loading from ambient to 150°C. It shows that, as the temperature rises, the elastic strain values for the fixed height of impact loading at 3 m rise as well. For a better understanding of the elastic strain deviation with the increase in temperature, a combined graph of elastic strain  value was shown in figure 10, which clearly shows the higher peak value of elastic strain for the higher temperature. Table 5 shows all the maximum values of elastic strain for varying temperatures under 3 m height of impact. The maximum elastic strain values obtained for ambient, 50°C , 100°C and 150°C temperatures were 0.0074913, 0.0077198, 0.0081335 and 0.0085585 respectively. As the temperature increases the elastic strain value gets increased i.e. higher for the 150°C temperatures.
To validate the experimental results and investigate the efficacy of numerical simulation analysis for concrete cubes under impact loading at various temperatures, the experimental results of the concrete cube were compared with the finite element analysis.
The elastic strain value of each concrete cube obtained from the experiment and FEM analysis were compared. A typical example of the strain values for both experimental and numerical analysis were shown in table 6. The strain value calculated experimentally using voltage data from an oscilloscope gives the strain around the vicinity of the sensors. Furthermore, this obtained experimental value of strain from the oscilloscope was compared with the FEM analysis. According to table 6, the variation in the experimental and numerical values of elastic strain is less than 10%. The difference in experimental and numerical values was due to the inability of providing the exact fixity at the base of the concrete cube in experimental conditions as compared to the numerical analysis. Based on the comparison of numerical and experimental results, the FEM analysis was  found to be highly consistent with the experimental results, indicating that the proposed modelling methodology is useful for studying the impact-resistant behaviour of the concrete cube under varying temperatures.

Conclusion
This research described both the experimental and numerical simulations of a concrete cube subjected to the combined effects of temperature and impact loading. The voltage strain measurement technique was used to identify the damage caused by the combined effect of impact and temperature in concrete for all three piezo sensors of different configurations. Furthermore, the ANSYS explicit dynamics and transient thermal FEM software was used for the numerical simulation and the conclusions discussed below are based on the outcomes of the experimental and numerical simulation.
• All three sensors embedded, surface bonded, and non-bonded are capable of detecting damage in concrete caused by the combined effect of impact loading and temperatures.
• Peak voltage follows a continuously decreasing trend for the surface bonded and non-bonded sensors with an increase in the number of impact loads on each cube. This decreasing trend of peak voltage is due to the increase in damping of the cube.
• Peak voltage has followed the opposite trend for the JKTPS because they have been placed in the line of impact loading as a result strain development is more in this case.
• In terms of elastic strain value, the numerical simulation results based on explicit dynamics and a transient thermal programme agree well with the measured experimental results.
• This indicates that the experimental and numerical methodology proposed in this study was suitable and showed good agreement between them for detecting damages in concrete structures subjected to impact loading at varying temperatures.