Acoustic emissions and electric signal recordings , when cement mortar beams are subjected to three-point bending under various loading protocols

Two experimental techniques are used study the response of cement mortar beams subjected to three-point bending under various loading protocols. The techniques used are the detection of weak electric current emissions known as Pressure Stimulated Currents and the Acoustic Emissions (in particular, the cumulative AE energy and the b-value analysis). Patterns are detected that can be used to predict upcoming fracture, regardless of the adopted loading protocol in each experiment. The experimental results of the AE and PSC techniques lead to the conclusion that when the calculated Ib values decrease, the PSC starts increasing strongly.


INTRODUCTION
he mechanical behavior and the evolution of damage in heterogeneous materials under compressive loading are of great interest in a wide range of application fields.Particularly, the Service Life of cement-based structures deteriorates due to heavy loads, fatigue, aging and natural disasters.Accordingly, diagnostic methods for damage assessment have been developed and implemented, aiming to assess the impending failure.To this end, the Acoustic Emissions (AE) technique was developed and is being improved continuously in order to provide a useful tool for monitoring and understanding the mechanisms of dynamic processes, but also for warning of upcoming failure [1].Acoustic Emission events occur due to sudden release of mechanical energy in the form of short mechanical vibrations due to the fact that a material is under mechanical loading.This mechanical vibration travels in the form of spherical waveforms through the material's bulk.The AE event analysis is used to estimate the damage degree when brittle materials such as concrete and rocks are subjected to mechanical loading [2][3][4][5].Acoustic Emission (AE) due to crack growth in brittle materials is usually observed in the high frequency range, typically between 50 kHz and 800 kHz.AE are recorded even from the early stages of the damage process.During the fracture process in quasi-brittle materials, the micro cracks formation and growth are manifested by a number of AE events released at different amplitudes.It has been shown that as a specimen under mechanical loading approaches failure, there is T an increase in the AE activity rate, as a result of the deterioration of the mechanical properties.Particularly in cement based materials and concrete structures the AE testing is one of the most widely used methods for monitoring crack growth [1].Another phenomenon related to the micro-crack growth in quasi-brittle nonmetallic materials is the production of electric charges that shape electric dipoles forming a rather complicated charge system [6][7][8].Such electric dipoles produce an electric potential across a crack resulting in the appearance of an electric current [9].Such currents have been measured in both laboratory [7,10] and at a geodynamic scale [11] and their detection may be useful as a precursor of a fracture.These electrical signals (weak electrical current emissions) are detected using a novel experimental technique, called Pressure-Stimulated Currents Technique, and the recorded electrical currents are described by the term Pressure Stimulated Currents (PSC) [12].The PSC's are weak electric currents detected with sensitive electrometers when a pair of electrodes is attached at proper locations on the specimen that is subjected to mechanical stress.Initially, the PSC technique was applied when rock specimens like marble [6,13] and amphibolite [14] were subjected to compressive axial stress increasing up to failure.Consequently, it was successfully applied to cement-based materials [15,16].The PSC technique was also tested during laboratory experiments of three-point bending (3PB) tests on marble [17] and cementbased specimens [5].The PSC technique has been adopted by several researchers [9,18], while others use similar techniques [19][20][21].Both AE and PSC signals provide important information about the damage processes occurring in specimens under compression or under bending tests.In particular, the PSCs show a considerable increase when the applied load reaches the vicinity of failure and attain their peak value shortly before failure [12,13,15,16].In Acoustic Emissions one of the statistical parameters, which is often used to estimate a critical situation, is the b-value [1,22,23], which exhibits systematic variations during the different stages of fracture processes.The AE based b-value analysis and the variation of the b-value, have attracted researchers working in the engineering field [23,24].Other AE statistical parameters that have been used include the event and energy release rates, the cumulative energy and the ring down counts.In this paper attention is focused on the parallel presentation of AE and PSC detected during 3PB of cement mortar specimens.The main difference between the three experiments is the loading mode.Specifically, one test was conducted under a constant loading rate (i.e.linear load increase) up to fracture while at the other two experiments the load was increasing according to a non-linear mode.

EXPERIMENTAL DETAILS
he specimens used for the experiments were prismatic cement mortar beams with dimensions 250x50x50 mm 3 .Their bending strength (L f ) varied from 3.5kN to 4.0kN.Details regarding the specimens, the preparation process and the experimental apparatus can be found in a previous work [25].Contrary to previous publications [25], the electrodes that were used to capture the PSC emissions were placed as shown in Fig. 1, i.e. on the lower side of the beam (tension zone) and at the left and right sides, symmetrically with respect to the specimen's central cross section where the load is applied.This topology was decided after several experiments conducted in order to estimate the best installation process that ensures the recording of strong PSCs and limits the influence of electric noise.The best electrode distance (  ) was also investigated and it was empirically found that for the specific type of experiments it should be    / 5 , where  is the distance between the two rigid metallic cylindrical rods used to support the cement beam.The PSC was captured by the electrodes and measured using a high sensitivity electrometer (Keithley, model 6517).The data were recorded in real time and stored on a hard disk through a GPIB interface.The mechanical load applied was recorded with the use of an analog-to-digital (A/D DAQ) data acquisition device (Keithley model KUSB-3108).The whole setup was placed in a Faraday shield in order to avoid interference from external electrical noise.
The system that was used to detect and record the AE is the 2-channel PCI-2 AE acquisition system (Physical Acoustics Corp).The R15a sensor (manufactured by PAC, resonant at 75 kHz) was placed in the middle of the beam (see Fig. 1) in order to focus on the region of the crack development processes that take place due to the externally applied bending load.The sensor was coupled to the test specimen using vacuum grease.Preamplifier was used along with the sensor with gain set at 40dB.The signals were band-pass filtered between 20-400 kHz using the software control of the data acquisition system.To set the threshold value for recording and to ensure that sensors were correctly performing, pencil lead breaks (5mm, HB leads) were carried out near the crack tip and the recorded signals were observed.The value of 40dB was selected since it was found to prevent the recording of lower amplitude reflected signals from the pencil lead break tests.For the detected AE data processing the Physical Acoustics Corp. Noesis software was used. T

RESULTS AND DISCUSSION
ab. 1 includes all the details regarding the loading protocol of each specimen for the three experiments implemented, i.e. load (L) vs. time (t), considering that the function is best described as: where 1 C and 2 C are constants.During the first experiment (i.e.specimen CB01) a linear increase of the load was conducted at a rate of 100N/s while the two following experiments were conducted following a non-linear increase of the mechanical load (see Tab. 1).  1.The characteristics of the loading protocol followed during each of the three experiments conducted and the main quantities studied regarding the PSC and AE recordings.

     
Fig. 2 shows the temporal variation of the emitted PSC from the three experiments, as well as the corresponding behavior of the mechanical load.Shortly before failure of the specimens the emitted PSC tends to show a peak.The recorded peak value (PSCpeak) is similar in all three experimental loading protocols (see Tab. 1).The fact that the PSC is maximized before the failure of the cement mortar beam has been systematically observed during compressive stress tests on both cement based [15] and rock [26,27] materials.This behavior is best demonstrated by the specimen CB02.This is attributed to the fact that the specimen approaches failure at the lowest load rate (see blue lines in Fig. 2).The level of the bending load, for which the PSC signal begins to show an intense and continuous increase, irrespectively of the loading mode, is estimated to be at approximately 60% of the ultimate 3PB strength (L f ) (see Fig. 3), in all three experiments.As a next step, it is interesting to consider the total electric charge ( T Q ), released during the three experiments up to failure.The T Q was calculated using the relation ( f t is the failure time).
It is worth mentioning that the total electric charge for the three different loading modes reaches almost the same value (see Tab. 1).In a previous work [13], during compressive stress application on marble specimens it was theoretically supported and experimentally proven [28] that the total electric charge value does not depend on the applied mechanical stress rate.This behavior is verified experimentally herein when cement mortar specimens are subjected to 3PB loading.Regarding the AE recordings, during the three experiments, a different number of AE hits was recorded (see Tab. 1).The micro-cracks generated during loading are of different sizes and therefore different AE amplitudes are recorded.A common way of obtaining quantitative information from the above statements is by computing the b-value of AE using the methods adopted in seismology [29,30].The b-value is defined as the log-linear slope of the frequency-magnitude distribution of AE [31].In the present study, the b-values were calculated by applying a method that is known as improved b-value (Ib) analysis after Shiotani et al. [32].The Ib-value is defined as: where,  is the standard deviation of the magnitude distribution in one group of events,  is the mean value of the magnitude distribution in the same group of events, and  1 as well as  2 are constants, for which usually a value equal to 1 is adopted [1].In order to highlight the variability of the b I -value, during the three experiments, the b-value analysis of ΑΕ is, in general, applied to a certain number of events, as follows: from event 1 to n, then from event 2 to n+1, and so on.The determination of the value of n has been considered by Shiotani et al. [32] and Colombo et al. [22], since it constitutes a parameter that can influence the results.A value exceeding n=50 is assumed to be a satisfactory one.In the present study, n=70 has been selected.Each value of b I was calculated from a group of the events 1 to 70, consequently from 2 to 71, and so on.

Ib -value
It is obvious that during the early stages (L<0.6Lfapproximately), the -values show an almost constant increase indicating the prevalence of micro-cracking [33].In these stages the PSC signal is low, showing a slight increase (see Fig. 4).Then, at the following stages (L>0.6Lf)-values show an intense reduction combined with a sharp increase of the PSC.This behavioral change in the variability of the -value is observed at a critical normalized load value [L/Lf]C ≈ 0.6, for all three experiments.As it is marked in Fig. 4 (dotted arrows), the exact value [L/L f ] C is 0.55, 0.62 and 0.70 for specimens CB01, CB02 and CB03, respectively.It seems that increasing the rate of the applied bending load the increase of the PSC signal and the corresponding reduction of -values are recorded earlier.It should be pointed out that thevalue fluctuation, which is observed in all three experiments at the early stage, is related to the fact that the micro-cracks begin to form randomly.Moreover, in all three experiments, attains values around 1 during the last seconds before the failure of the specimens, which is indicative of the final formation of a localized macro-crack.The ring down count (rdc) data was also used, since it is a measure of the number of AE waveform oscillations through a preset voltage threshold [1].In order to highlight the identification and assessment of important stages of micro-crack damage in a cement mortar material, a common practice has been followed to examine the cumulative ring down count Σ(rdc) and energy (ΣE).These results are illustrated in Figs. 5 and 6 with respect to the normalized load (L/L f ).In all three experiments an almost constant increase in Σ(rdc) was initially observed, related to the initiation of new microcracks.For L/Lf > [L/Lf]C the Σ(rdc) increases at a rate indicating the development of severe damage in the specimen.Concurrently, an abrupt increase of the PSC signal is observed.

Figure 1 :
Figure 1: The experimental installation and the location of the AE and PSC sensors.

Figure 2 :
Figure 2: The time variation of the applied 3PB load and the corresponding time variation of the PSC, during all three experiments conducted.

Figure 3 :
Figure 3: PSC versus normalized load (L/L f ) for all three experiments conducted.
Each calculated b I value is associated to the time of the 70th event of each group and to that of the corresponding value of normalized (L/L f ) load.The variability of the b I -value during the three experiments with respect to the normalized (L/L f ) load is shown in Fig. 4. In the same figure the corresponding variation of PSC signals is also presented.

Figure 4 :
Figure 4: Variation of b I -values and PSC signals vs. normalized load (L/L f ) for the three conducted experiments.

Figure 5 :
Figure 5: The cumulative ring down count of ΑΕ hits vs. normalized load (L/L f ) for the three conducted experiments.

Figure 6 :
Figure 6: (a) The cumulative AE energy vs. normalized load (L/Lf) for the three experiments and (b) a detailed view close to the fracture load.