Acoustic Emission Test of Marble Powder Concrete

Abstract

Uniaxial compression tests were carried out based on acoustic emission technology to study the rule and process of the damage evolution of marble powder concrete (MPC) under compression. The fracture characteristics of MPC are examined by analyzing the variation characteristics of acoustic emission signal parameters. The damage evolution model of MPC is established based on the characteristic parameters. The results showed that the strength loss rate of MPC decreased with the curing time, and the maximum strength loss rate of 28 d was 6.2% when the substitution percentage was 15%. According to the analysis of acoustic emission parameters, the compression failure process of MPC can be divided into three stages, the acoustic emission activity is higher in the early stage of loading, the relative stress during critical instability failure is reduced, and the substitution percentage is reduced by 4.2% at 15%. MPC is mainly fractured by tensile failure. With the increase in marble powder substitution percentage, the proportion of tensile mode cracks in stages I and II decreased. The proportion in stage III showed an increasing trend, and the failure characteristics gradually changed from brittleness to ductility. The fluctuation amplitude of the b-value increased with the substitution percentage, and the highest volatility was 18.5% when the substitution percentage was 15%. The crack propagation behavior gradually changed from stable growth to unstable growth. When the substitution percentage is lower than 15%, the damage development of MPC in the middle of stress (relative stress 20~70%) is relatively slow, and the damage development is accelerated in the late stress stage.

1. Introduction

Marble is found in many parts of the world, such as China, Egypt, Turkey, and other countries. For example, the marble powder reserves in Hezhou, Guangxi alone amount to 2.6 billion cubic meters, with an annual output of 1 million tons of solid waste [1]. A large amount of solid waste is piled up in unused vacant land or discharged into bodies of water, causing a waste of land resources and seriously endangering the environment. Recycling marble waste powder has received widespread attention around the world.

Using marble powder in concrete production has the advantage of high dosage and low cost. Bacarji [2] et al. believed that marble powder was a relatively stable, inert material and had a weak influence on the hydration process. Aliabdo [3] et al. found that the replacement of 15% cement with marble powder had a negligible effect on the compressive strength of concrete and had a filling impact at low water-to-ash ratios. Rodrigues [4] et al. found that concrete’s splitting tensile strength and compressive strength decreased slightly with increasing marble powder substitution, while water-reducing agents could counteract the adverse effects. Xiao [5] et al. found that marble powder can effectively improve mortar fluidity, and the greater the dosage, the more pronounced the effect. Singh [6] et al. found that adding 15% marble powder can improve the mechanical properties of concrete, and the particle size distribution has little influence on the compressive strength of concrete. Zhang [7] et al. found that when marble powder was used to replace 10% of cement, the number of pores in concrete could be reduced. The above study analyzes the impact characteristics of marble powder on concrete while revealing the strength performance pattern of MPC and verifying the feasibility of marble powder for the concrete field. However, at this stage, there are very few research results on the damage mechanism of MPC under pressure. To better apply it in practical engineering, one needs to understand the mechanical properties of MPC systematically, and clarify its compressive fracture process, fracture mode, and damage development trend are of great significance, and so it is urgent to supplement the relevant experimental research.

Applying acoustic emission monitoring technology in the concrete sector is becoming increasingly sophisticated. Watanabe [8] et al. revealed that recycled coarse aggregate concrete enters the instability damage stage earlier than plain concrete, based on the acoustic emission technique. Haneef [9] et al. found that adding fly ash to concrete can reduce acoustic emission activity because fly ash can reduce the original defects of concrete. Han [10] et al. found that rubber concrete fractured to a greater extent than plain concrete before ultimate compressive strength, using a typical analysis of acoustic emission. Bhosale [11] et al. applied acoustic emission techniques and concluded that steel fiber concrete and mixed fiber concrete had more shear modes than synthetic fiber concrete in flexural tests. Chen [12] et al. studied the damage evolution mechanism of steamed concrete after experiencing freeze–thaw cycles through acoustic emission tests. They found that the freeze–thaw environment could increase the initial damage of steamed concrete. The above studies show that, unlike other detection methods, acoustic emission technology can provide real-time, non-destructive dynamic monitoring of the internal damage process of concrete materials and has the advantages of synchronization and accuracy [13,14]; so, this paper adopts acoustic emission technology to study the damage mechanism of MPC under pressure.

In this paper, based on acoustic emission technology, uniaxial compression tests were carried out on MPC specimens with different substitution percentages, and the development process and law of MPC fracture were analyzed by acoustic emission (AE) signal parameters, and the compression fracture form and crack propagation behavior of MPC were discussed. The research results can provide a reference for theoretical research and the practical engineering application of MPC.

2. Materials and Methods

The test cement is P·O 42.5 ordinary Portland cement produced by Anhui Conch Cement Company. The fine aggregate is natural river sand with a fineness modulus of 2.85. The coarse aggregate is limestone, with a maximum particle size of 20 mm and continuous gradation. The water-reducing agent is a naphthalene series superplasticizer, and the water-reducing rate is about 22–28%. The marble powder was taken from the cut waste powder of the Gangshi Processing Plant in Hezhou City, Guangxi Province, and was dried before the test. The average particle size of marble powder was 16.87 μm, the average particle size of the surface area was 4.18 μm, the apparent density was 2.58 g·cm⁻³, and the specific surface area was 388 m²·kg⁻¹. Its chemical composition is shown in

Table 1. Chemical composition of marble powder.

CaO	MgO	SiO2	Al2O3	Na2O	K2O	Fe2O3	Loss
47.937	5.492	1.974	0.045	0.377	0.069	0.003	44.103

, and its particle size distribution curve is shown in Figure 1. Combined with the existing research results, C30 was taken as the benchmark strength, and the mix proportion calculation was carried out according to the Design Regulations of Ordinary Concrete Mix Proportion (JGJ55-2011). The cement was replaced with marble powder at 0%, 5%, 10%, 15%, 20%, and 25%, and the mix proportion is shown in

Table 2. Design of MPC specimen mix proportion.

Sample	W/CM	Dosage per Volume/kg·m−3
		Water	Cement	Fine Aggregate	Coarse Aggregate	Water Reducer	Marble
							Powder
MPC-0	0.55	198	360	788	1087	1.8	-
MPC-5			342				18
MPC-10			324				36
MPC-15			306				54
MPC-20			288				72
MPC-25			270				90

. Three standard cubic compression test blocks of 150 mm × 150 mm × 150 mm were made into each group, and the test results were averaged. The specimen is demolded after molding and the specimen is put into the standard curing box to be cured to the age to be tested. The test was carried out after wiping the surface moisture of the specimen.

The test was conducted using Sensor-Highway III acoustic emission equipment from Physical Acoustics, West Windsor Township, NJ, USA, with a PK15I built-in pre-amplified sensor. To accurately locate the location of the fracturing source, four sensors were arranged diagonally on both surfaces of the specimen to collect AE signals (Figure 2); each sensor was 30 mm from the edge of the specimen. The sensor contact surface is coated with a thin layer of vacuum grease, and the surface of the specimen is polished with sandpaper and then pasted with hot melt adhesive. The acoustic emission parameters were calibrated before the test. The AE wave speed was calculated using a fractured lead core test collection and averaged. The calibrated parameter settings are shown in

Table 3. Acoustic emission parameter settings.

Parameter	Threshold/dB	PDT/μs	HDT/μs	HLT/μs	Wave Speed/m·s−1
Value	45	150	300	500	5300

. The loading system is a WHY-1000 type microcomputer-controlled pressure tester produced by Shanghai Hualong Testing Instrument Company, with a maximum range of 1000 KN, which can automatically record the test data by using the load control loading method, setting the loading rate to 5 KN/s, and starting two devices at the same time during the test to ensure that the data can be collected simultaneously.

3. Results

3.1. The Compressive Strength

Figure 3 shows the variation of the mean compressive strength of MPC in the uniaxial compression test with the curing time. It can be observed that the MPC strength growth pattern is similar to that of plain concrete, with a faster growth rate in the early stages of curing and a slower growth rate after reaching 7 d of age. Its average compressive strength at all ages is less than that of plain concrete and decreases with an increasing substitution percentage. Figure 4 shows the MPC compressive strength loss rate variation with curing time. The strength loss rate is defined as the percentage of the strength difference between MPC and plain concrete in the strength of plain concrete under the same age conditions. As can be seen from the graph, the strength loss rate increases gradually with the rise of marble powder substitution percentage, the maximum strength loss rate does not exceed 15% when the substitution percentage is less than 15%, and the strength loss rate increases steeply when the substitution percentage is more than 15%, reaching 26.3%. According to the literature [15], the influence of marble powder on the cement hydration process is weak, and the reduction in MPC strength is mainly attributed to the reduction in cement content in the concrete mixture. At the same time, it can be found that MPC has the highest strength loss rate in the early stage of curing and then decreases with the extension of curing time. The decline is particularly significant when the substitution percentage is high. When the marble powder substitution percentage increased from 5% to 25%, the strength loss rate from 3 d to 28 d decreased by 3.1%, 6.9%, 8.6%, 11.6%, and 12.1%, respectively, so the strength growth of MPC was more dependent on the curing time. From the above test results, the effect of marble powder on the strength of concrete is small when the substitution percentage is less than 15%.

3.2. Acoustic Emission Ringing Counts

Figure 5 shows the variation curves of AE counts and cumulative AE counts with the stress level of the MPC specimen at 28 days of age during uniaxial compression. Based on the different characteristics of AE counts, this paper divides the MPC compression damage process into three stages.

(I) Compacting stage. In the early loading stage, the AE signal is active for a short time because MPC is a heterogeneous material, the original defects in the specimen are closed at low stress, and its internal microscopic pore structure changes. The specimen is in an elastic state at this time, and no microcracks occur. (II) Stable development stage. The load increases, the AE counts fluctuate smoothly in a small range, and the AE cumulative counts rise approximately linearly. Micro cracks occurred gradually at the tip of the original defect and continued to extend along the binding weakness in the MPC specimen. The aggregate particles are also micro-broken, and the crack development is stable. (III) Unstable and destructive phase. When approaching the ultimate compressive strength, the AE counts increase sharply, the internal fracture behavior of MPC develops rapidly, the bearing capacity decreases sharply, the damage is close to the peak, and soon expands to the surface of the specimen to form macroscopic cracks, as the macroscopic cracks continue to converge and extend, concrete debris begins to fall, and the MPC specimen finally loses bearing capacity and sudden fracture failure.

In Figure 5, the acoustic emission activity of MPC specimens in stage I is significantly higher. The analysis of the AE signal generation mechanism in this stage shows that the original defects of MPC are more than those of plain concrete and pore structure deterioration. The original defects are continuously compacted (or opened) at low stress, resulting in more AE signals. When the ultimate compressive strength is approached, the MPC changes from the stable development stage to the unstable failure stage, and the AE counts performance increases sharply, indicating that the specimen enters the instability stage, which is a precursor to the aggravation of damage and imminent strength failure. It can be seen from

Table 4. Statistics of MPC specimen AE parameters.

Sample	Relative Peak Stress	AE Cumulative Count Peak/105
MPC-0	90.5%	3.8
MPC-5	86.8%	3.8
MPC-10	86.4%	3.6
MPC-15	86.3%	3.3
MPC-20	84.3%	2.8
MPC-25	80.4%	2.7

that the relative stress value of MPC at the demarcation points of stages II and III is lower than that of plain concrete and decreases with the increase of substitution percentage, indicating that MPC is easier to enter into instability and failure than plain concrete. The difference with plain concrete is less than 5% when the substitution percentage is less than 15%. The cumulative peak AE count decreases with the increase in the marble powder substitution percentage, since the more original defects in the MPC make it more prone to secondary cracks during uniaxial compression, the microcracks in the specimen keep expanding and converging, the fracture degree is severe, the strain energy is gradually released earlier, the overall decrease of acoustic emission activity, and the shortening of uniaxial compression time due to the decrease of load capacity, so the amount of AE signal that can be collected decreases. At the same time, it can be found that when the marble powder substitution percentage increases from 15% to 20%, the peak cumulative AE count decreases by 15.2%. The reason for this is that as the marble powder substitution percentage increases, the filling effect is no longer sufficient to offset the drastic reduction in the cementitious material content [6], and the compressive strength of MPC decreases rapidly. Then the compressive damage behavior changes dramatically at a high substitution percentage.

3.3. Fracture Mode

Concrete materials often accompany burst type acoustic emission signals when fractured, and the primary signal parameters are shown in Figure 6. According to the research, the following characteristic parameters of RA and AF can be defined [16]:

RA = rise time/amplitude

AF = count/duration

AE signals with low RA and high AF values are tensile mode cracks, whereas AE signals with high RA and low AF values are shear mode cracks [17]. The fracture mode and properties of concrete materials can be qualitatively determined by analyzing the characteristic parameters of AE signals (Figure 7).

Due to the influence of the sensor type, AE test parameters, RA-AF coordinate axis size, and other factors, the critical values of the two fracture modes are not specified. To overcome subjective selection’s adverse impact on test results’ accuracy, the Gaussian mixture model (GMM) was applied to cluster analysis of the feature parameter RA-AF [18]. GMM clustering is a label-free machine learning algorithm that can calculate the Gaussian probability of each sample in the sample space to achieve classification. For the D-dimensional sample space, if each element x_i follows a multivariate Gaussian distribution with mean μ and covariance ∑, its probability density function is:

$$p\left(x|\mu ,\sum \right)=\frac{1}{{\left(2\pi \right)}^{2n}{\left|\sum \right|}^{\frac{1}{2}}}{e}^{-\frac{1}{2}{\left(x-\mu \right)}^T{\sum }^{-1}\left(x-\mu \right)}$$

where μ is the D-dimensional mean vector, and ∑ is a D × D covariance matrix.

Considering the superposition of K Gaussian distributions, the probability density function of Gaussian mixture distribution is:

$$p\left(x|\lambda \right)={\displaystyle \sum }_{k=1}^K{\omega }_{k}p(x|{\mu }_k,{\sum }_k)$$

${\omega }_k$ is the mixed weights, $0\ll {\omega }_k\le 1$; $\lambda =\left({\omega }_k,{\mu }_k,{\sum }_k\right)$; K is the number of categories in the Gaussian mixture model; and $p(x|{\mu }_k,{\sum }_k)$ is the K-th Gaussian distribution.

The EM algorithm is used to iteratively calculate each parameter in GMM until the convergence condition is met (the number of iterations reaches the maximum or the likelihood function does not change significantly). Accordingly, the element xi is classified into the category with the maximum probability. Figure 8 shows the schematic diagram of the effect of GMM clustering on the RA-AF classification of the MPC-0 specimen in this paper. The scatter points in the figure represent microcracks, and there are clear dividing lines between the two types of point positions, so GMM is suitable for the classification of microcracks.

When the concrete is compressed, microcracks propagate and converge to form macroscopic cracks. Therefore, the statistical analysis of the proportion of the two types of microcracks can reveal the fracture mode of concrete at the microscopic level. Figure 9 shows the percentage of tensile mode cracks in MPC in different stages. In stages I and II, the compression fracture behavior of MPC is similar to that of plain concrete, with tensile mode crack as the main fracture form, accounting for more than 75%, but gradually decreases with the increase in substitution percentage. The reasons for this suggest that as the marble powder substitution percentage increases, the C₃A and C₂S content required for the MPC hydration process decreases, and the structural system becomes less bonded [6], thus, increasing the number of shear mode cracks when subjected to compression. It can also be found that the percentage of tensile mode cracks in MPC in stage I decreases rapidly from 82.16% to 77.46% at a substitution percentage higher than 15%, indicating that a high substitution percentage causes a sharp increase in original defects in concrete. In Stage III, the number of unstable microcracks in the MPC specimens increased dramatically as the load grew, gradually losing load-carrying capacity, with the percentage of tensile mode cracks rapidly decreasing to less than 75% and the number of shear mode cracks increasing [19]. At the same time, it can be found that the percentage of tensile mode cracks tends to increase with the increase of marble powder substitution percentage. When the marble powder substitution percentage was increased from 0% to 25%, the percentage of tensile mode cracks in MPC specimens decreased from Stage II to Stage III by 24.62%, 21.25%, 15.41%, 14.54%, 14.3%, and 8.47%, respectively, indicating that MPC would shift to shear mode damage earlier than plain concrete. Tensile mode cracks are mainly formed in the microcrack initiation of concrete under pressure. Shear mode cracks are due to the mutual extrusion friction between the micro-damage surfaces caused by microcrack propagation, the energy is higher [20,21]. Analysis shows that the number of shear mode cracks in MPC in the first two stages increases with the increase in substitution percentage, the accumulated energy is continuously released in advance, and the failure characteristics gradually change from brittleness to ductility so that the proportion of tensile mode cracks in the final failure is relatively high.

Figure 10 shows the 28 d scanning electron microscope (SEM) microstructure images of marbled concrete specimens. To ensure the simultaneity of the test results, the sample was taken from a fragment of the mortar mixture produced during the compressive strength test, approximately 5 mm in diameter, which was glued to the test rig with conductive glue, vacuumed and sprayed with gold, and then observed in an electron microscope, using a Japanese Hitachi SM8020 field emission scanning electron microscope. It can be observed that the calcium silicate hydrate gel (C-S-H) on the surface of plain concrete samples has a dense distribution, and the acicular ettringite (AFt) has a high density. It overlaps with each other, and the porosity is lower. However, with the increase in substitution percentage, the distribution of calcium silicate hydrate gel on the surface of marbled concrete samples gradually became looser. Many pores (V) weakened the bonding properties at the interfaces between cement paste and fine aggregate, mortar, and coarse aggregate and reduced their compressive strength. At the same time, in the process of compression deformation, more microcracks will be generated and spread continuously, which will cause extrusion friction between each micro-damaged surface. More shear mode cracks will be generated. Figure 11 records the macroscopic crack image of the marbled concrete specimen after losing its bearing capacity. As shown in Figure 11, the main cracks of the MPC-0 specimen are mostly vertical, with fewer branch cracks and low overall density. The fracture area is relatively complete, and the concrete debris on both sides falls off in blocks. After the failure of MPC-10, the main cracks were distributed vertically and diagonally, the density of branch cracks was higher, tiny debris fell off the surface, and the damage degree was more serious. MPC-20 showed transverse and vertical microcracks interleaved on the surface of the specimen, branch cracks distributed radially and densely, and a large amount of debris fell off the surface of the specimen, with a severe degree of damage. The above observations support the results of the RA-AF analysis of the acoustic emission signal characteristics, which indicate that the MPC has a lower percentage of tensile mode cracks and an increased number of shear mode cracks in stages I and II, while the percentage increases in stage III and the damage characteristics gradually change from brittle to ductile.

3.4. b-Value Analysis

b-value analysis was first used in seismology. Seismologists generally believe that the b-value can quantitatively describe the level of seismic activity. An AE event can be regarded as a microearthquake because of the release of strain energy when a material breaks. Now b-value has become an effective method to evaluate the crack propagation behavior of concrete materials. In 1944, Gutenberg and Richter proposed that the G–R formula to measure seismic activity could be used to analyze the characteristics of the concrete fracture process [22,23]:

$$lgN=a-b(\frac{A_{dB}}{20})$$

where N is the number of AE events with amplitude greater than A_dB, and A_dB is the peak amplitude of acoustic emission event. a and b are constants, where b is the b-value of an AE event at a certain amplitude level.

This paper uses the least square method to calculate the b-value. Figure 12 shows the dynamic change curves of the b-value during uniaxial compression of MPC specimens. Before the unstable failure stage, the b-value of plain concrete fluctuates in a small range, with a maximum fluctuation amplitude of 7.7% (dotted box in Figure 12a). At this time, the proportion of microcracks of different scales in the specimen is unchanged; that is, the fractured state of the material develops gradually and slowly [24]. When the load increases to about 90% of the relative stress, the accumulated strain energy in the middle and early stages of the stress can cause more microcracks in the concrete. The specimen entered the unstable failure stage, the b-value suddenly changed from 1.07 to 1.31, and the proportion of small-scale events increased significantly quickly. Then the b-value suddenly decreased, and the proportion of high-energy AE events increased. Large-scale microcracks in the specimen extended to the surface of the specimen, and, finally, the bearing capacity was lost. Compared with plain concrete, when the substitution percentage of MPC is less than 15% before the unstable stage, the fluctuation of the b-value is relatively stable. In the middle of loading, there will be a slight step fluctuation, with a maximum fluctuation of 18.5% (dotted box in Figure 12d). This indicates that the internal microdamage of the material is stable before reaching the ultimate compressive strength but is also accompanied by a small number of unstable microcracks. The b-value of MPC specimens with a substitution percentage greater than 15% fluctuates by leaps and bounds, with fluctuations of up to 27.8% (dotted box in Figure 12f). The proportion of microcracks at different scales varies considerably, and the number of unstable cracks surges. The b-value points are denser than plain concrete when approaching ultimate compressive strength, indicating that the fracture magnitude before visible cracks appears much higher than that of plain concrete. The sudden, disorderly expansion of microcracking leads to a significant reduction in concrete strength, so the high substitution percentage significantly impacts microcracking development in concrete under compression. The reason for this is that with the increase of marble powder substitution percentage, the pores in the MPC specimen system grow significantly, the pore structure development tends to be unfavorable, the compactness decreases, the bearing capacity deteriorates [7], and the microcrack expansion behavior gradually changes from stable growth to non-stable growth.

3.5. Acoustic Emission Cumulative Count Damage Evolution Model

In continuous damage mechanics, material damage degree is expressed by damage factor D, which is defined by Kachanov as:

$$D=\frac{A_d}{A}$$

(1)

where $A_d$ is the total area of fracture surface microdamage, and $A$ is the total area of the material in the non-destructive state.

When the initial non-destructive material is completely broken, the AE cumulative counts is $N_a$, and then the acoustic emission rate of microdamage per unit area is:

$$n_a=\frac{N_a}{A}$$

(2)

When the cumulative damage area of initial non-destructive material during loading is $A_d$, the AE cumulative counts is:

$$N=n_a{A}_d=\frac{N_a}{A}{A}_d=N_a\frac{A_d}{A}$$

(3)

Substitute Equation (3) into (1) to obtain the damage factor D expressed by AE cumulative count:

$$D=\frac{N}{N_a}$$

(4)

Ohtsu [25] et al. found that the cumulative AE counts N at a given moment, and the stress σ at that moment can be expressed by the following equation:

$$N=c_1{\sigma }^{a_1}{e}^{b_1\sigma }$$

(5)

where $a_1,b_1,c_1$ is the test constant; σ is the stress value at a loading moment; and N is the accumulated counts of AE at that time.

Substitute Equation (5) into Equation (4), and the damage evolution equation of the MPC specimen can be obtained as follows:

$$D=\frac{1}{N_a}{c}_1{\sigma }^{a_1}{e}^{b_1\sigma }$$

(6)

The stress σ in uniaxial compression of the MPC specimen and the corresponding cumulative AE counts N are substituted into Equation (5) to obtain the values of the test constants $a_1,b_1,c_1$ by fitting the analysis (

Table 5. Fitting results of damage evolution model for MPC specimens.

Sample	${\mathit{a}}_1$	${\mathit{b}}_1$	${\mathit{c}}_1$	R2
MPC-0	0.7602	0.01274	11,280	97.27%
MPC-5	0.3151	0.03937	23,780	96.52%
MPC-10	0.2891	0.04026	22,460	97.53%
MPC-15	0.132	0.04934	17,070	93.74%
MPC-20	0.3396	0.02281	31,780	96.08%
MPC-25	0.6844	0.01208	32,544	98.64%

).

By substituting the values of $a_1,b_1,c_1$ and the peak cumulative AE counts $N_a$ above Equation (6), a model for the evolution of MPC damage at different substitution percentages can be developed (Figure 13). As shown in the figure, the damage degree of plain concrete under uniaxial compression is relatively stable with the development of stress levels. Its internal microcracks gradually expand slowly and burst to fracture when approaching the ultimate stress. When the relative stress is less than 20%, compared with plain concrete, the damage development of MPC is aggravated with the increase in the substitution percentage, which reflects that the addition of marble powder will form more original defects, and concrete material is constantly cracked or closed under the action of external forces due to their internal original defects, and it finally evolves into unstable failure. Therefore, more original defects provide an important reason for the decrease in MPC carrying capacity. The damage development of MPC with a replacement ratio lower than 15% is relatively slow when the relative stress is about 20–70%, and the damage development is accelerated when the relative stress reaches 70%. Still, the damage value at the ultimate load is slightly lower than that of plain concrete. The damage development of MPC with a substitution percentage higher than 15% is always greater than plain concrete during loading. The above analysis shows that the damage evolution model expressed by the AE cumulative count can better describe the damage degree change with MPC specimens’ stress level.

4. Discussion

In summary, acoustic emission technology can be effectively used to study the compression fracture characteristics of MPC. The test results show that MPC has increased porosity and original defects due to decreased cementitious material content. At the same time, because marble powder does not have the volcanic ash effect of fly ash, the crack propagation behavior gradually changes to unstable growth when it is compressed, which accelerates the occurrence of cracks in shear mode. This is different from the literature [9] on the performance of fly ash concrete under pressure cracks, but has little effect when the substitution percentage is 15%. The application of marble powder solid waste is beneficial to solve environmental problems, the topic is essential and systematic, and the performance of components made of MPC needs to be further studied.

5. Conclusions

(1)

The compressive strength of MPC specimens decreases with increasing substitution percentages at all ages. The rate of strength loss is highest in the early stages of curing and decreases with curing time after that, with a maximum of 6.2% at age 28 d for substitution percentages below 15%.

(2)

The compression failure process of MPC can be divided into three stages based on the characteristics of the AE count changes. With the increase of marble powder substitution percentage, the peak of acoustic emission accumulation count decreased. Still, the acoustic emission activity at the beginning of loading increased, indicating that the original defects of MPC were more. At the same time, the relative stress at critical instability failure decreases with the substitution percentage, and the substitution percentage decreases by 4.2% at 15%.

(3)

The RA-AF correlation analysis showed that the tensile mode damage was the main fracture form of MPC. As the marble powder substitution percentage increases, the percentage of tensile mode cracks decreases in stages I and II, the number of shear mode cracks increases, and the percentage in stage III tends to increase. The damage characteristics gradually change from brittle to ductile. The SEM images and the macroscopic cracking behavior of the failure support this result. The b-value analysis shows that the fluctuation of b-value increases with the increase of marble powder substitution percentage; the highest fluctuation is 18.5% at 15% substitution percentage, the number of unstable microcracks under pressure increases, and the crack expansion behavior gradually changes from stable growth to non-stable growth.

(4)

The MPC damage evolution model was established with cumulative AE counts. The results showed that when the substitution percentage was lower than 15%, the damage development of MPC was relatively slow in the middle stress (relative stress 20–70%) and accelerated after reaching 70%.