1. Introduction

2. Materials and Methods

2.4. Preparation and bio-stimulated healing mortar

Initial sample characterization was followed by the preparation of a new type of mortar and its characterization by examining the strengths, chemical composition and setting time.

The new mortar system was prepared with a 10% slag replacement fraction instead of alite cement using the method described in SRPS EN 196-1:2017. The sample prisms were of standard dimensions 40 x 40 x 160 mm and laboratory dimensions of 10×10×60 mm. Flexural and compressive strengths were tested after 2, 28 and 240 days of water curing on hydraulic presses by Tinius Olsen, working force up to 231 kN and VEB Thuringer Industriewerk Rauenstein (standard prisms), working force 0-5 kN and 25-50 kN (laboratory prisms). For this type of mortar, the setting time was also determined using Vicat apparatus, according to SRPS EN 196-3:2017. The mortar system composition is given in Table 1.

Table 1. Mortar system composition.

Table 1. Mortar system composition.

System/ composition	CEM I, g	Three-fraction sand, g	Granulated blast furnace slag, g	Water, ml
Az10	405	1350	45	225

* Legend: Az₁₀ -CEM I with the addition of 10% slag.

Determination of the strengths, setting time and chemical composition of the new mortar type was followed by the preparation of laboratory sample prisms for the bacterial bio-stimulated healing experiment. After preparation and water curing for 28 days, sample prisms of laboratory dimensions were cut with a diamond knife (Figure 1(a)) to 10x10 mm for easier manipulation, and a groove was artificially made on each sample as a real crack simulation (Figure 1(b)). In this way, enough samples were prepared for the needs of further experiments.

Figure 1. Sample preparation: (a) applied laboratory equipment (diamond knife); (b) appearance of the prepared cracked sample.

Figure 1. Sample preparation: (a) applied laboratory equipment (diamond knife); (b) appearance of the prepared cracked sample.

Considering that the selected bacterial strain S. pasteurii DSM 33 is highly sensitive to environmental conditions [29], the pH value had to be reduced to a value below 10, which was achieved by alternating immersion in distilled water and drying in a dryer for 2 hours each. After 20 washing cycles, the pH value was lowered from 14 to 9.

According to the established experiment program, 45 samples were singled out. The system was divided into five groups that were treated under different conditions, as shown in Table 2. Conditions 2 to 4 are control.

Table 2. Conditions for systems.

Table 2. Conditions for systems.

Groups	Conditions	Days
1	bacterial suspension, nutrient medium, sterile demineralized water	7,14,28
2	nutrient medium, sterile demineralized water	7,14,28
3	sterile demineralized water	7,14,28
4	Danube water	7,14,28
5	bacterial suspension, nutrient medium, Danube water	7,14,28

The samples were placed in Petri dishes so that the crack was on the upper side of the sample, for easier healing agent application, with three repetitions in each group. Petri dishes with samples were sterilized for 1 h at 160 °C to eliminate any possibility of contamination. After sterilization, bacteria were applied in a laboratory at constant temperature of 25 °C, which favors bacterial growth.

The bacterial suspension was freshly prepared with sterile demineralized water. The nutrient medium was also freshly prepared from urea, NaHCO₃, NH₄Cl and sterile demineralized water.

The nutrient medium and bacterial suspension were applied using a sterile pipette in the crack mid-line. Each sample of the first, second and fifth group was first given 50 μl of nutrient medium. After absorbing of nutrient medium, 50 μl of bacterial suspension was applied to each sample of the first and fifth group (Figure 2). Finally, the sterile demineralized water was added with a sterile pipette up to 1/3 of the sample height in the first, second and third group, while the fresh Danube water was added also up to 1/3 of the sample height in the fourth and fifth group (10 ml for smaller 6 cm diameter Petri dishes and 15 ml for larger 10 cm diameter Petri dishes). This amount of water was optimal for maintaining system humidity. After setting up the system, the Petri dishes with the samples were transferred to the Binder Climate Chamber KBWF 240. In the climate chamber, the systems were kept under controlled conditions of a temperature of 30 °C and humidity of 70% until testing.

The tests duration was 7, 14 and 28 days after the experiments setup. Non-destructive methods were selected to monitor changes in the bio-stimulated healing process, portable microscope imaging (described below – chapter 2.5) and SEM analysis, already described in chapter 2.1.

Figure 2. Setup of bio-stimulated healing process.

Figure 2. Setup of bio-stimulated healing process.

2.6. Image classification

Image classification was performed in the MATLAB 2023A programming language using the convolutional neural networks methodology - ResNet 50, which has already been pretrained to extract features from digital images based on the rgb model.

Basic characteristics of the Convolutional neural networks (CNN) model

Convolutional neural networks are an extended model of multi-layer artificial neural networks, which were developed by adding a new type of layer used for image analysis (recognition and classification) [30].

Convolutional neural networks mimic the human visual system and may recognize complicated image features gradually. In the first layers of this network, simple attributes such as edges are detected. Based on what was detected in the previous, contours are recognized in the next layer. Contour detection is followed by the recognition of specific parts of the object so that the object can be classified in the final layer [31].

The principle behind convolutional neural networks is reflected in the direct input of image data in raw pixels since the image consists of pixels as the smallest element in a digital image (shown as the image having, for example, a width of 32, a height of 32 and a depth of 3). The first task of a convolutional neural network model is to transform the image it receives into a computer-understandable format. More precisely, the input data is represented as a two-dimensional pixel intensity matrix in the case of a single-channel image or as a multi-dimensional matrix in the case of a multi-channel image. So, in the input layer, the image data is entered into the network, and each entered pixel represents an input characteristic. Each pixel is described by its color as a color imaging system. There are many color spaces in which images exist and one of the most often used is RGD with three color channels: red, green and blue. Thus, for example, white is presented as (0,0,0), while black is (255,255,255) [32,33].

Data entry is followed by several alternating convolutional layers and pooling layers, which alternation reduces their dimensions. The end result is images very small in size, each representing one matrix. The matrix values are arranged into a vector that represents the input to the fully connected network [30,31,34,35]. A schematic representation of the basic convolutional neural network on the example of a cement prism crack image is shown in Figure 3.

Figure 3. Simplified schematic representation of a convolutional neural network on the example of a crack image [adapted from 36].

Figure 3. Simplified schematic representation of a convolutional neural network on the example of a crack image [adapted from 36].

Convolutional layer

The convolutional layer serves to extract the features of the input data using the filters located inside the convolutional layer. The input image size is larger than the filter size, but their depth is the same. Convolutional neural networks understand images in parts, so it is necessary to move the filter several times to complete the entire processing and obtain the value for the whole image [34].

The filter window moving starts from the upper left corner and goes to the right by the same value. When we reach the edge, we move it downward, repeating the process. The filter is moved to multiply and add up each position of the starting pixel. That is, the filter scans subsamples of the initial images to detect localized features. The size of the subsample depends on the size of the filter. Each convolutional layer has a filter displayed as a two-dimensional matrix 3 × 3 (an example of m × n is shown in Figure 4) with an increment of 1. The output volume is obtained based on several filters that make up this layer and are grouped along the axis [34,35,36,37,38].

Figure 4. Simplified schematic representation of the convolutional layer operation [adapted from 35].

Figure 4. Simplified schematic representation of the convolutional layer operation [adapted from 35].

ReLu layer

The ReLu layer is introduced after the convolutional layer as an additional layer to increase image non-linearity and reduce calculation time. This layer serves to break the linearity imposed by passing the image through the convolutional layer [34,35].

Pooling layer

The convolutional layer is followed by the pooling layer, in which the data dimension is reduced. This layer also has filters, but they have no weight here, unlike in the convolutional layer. Here, the role of the filter is to select the pixel that is within the dimensions covered by this layer in a given way. The maximum is often used as it works better in practice, although medium values are also acceptable (Figure 5). The pooling layer reduces the dimension of the feature maps while retaining the most important sample information. In this way, excessive overlaps are prevented. For maximum pooling, a 2 × 2 matrix is adopted. After pooling, the image feature map is aligned in one column, resulting in a vector. The long vector obtained in this way serves as the input to the artificial neural network, which is also the final step. An artificial neural network consists of an input, output and fully connected layer (network) [36,37,39].

Figure 5. Simplified representation of the pooling layer operation [35].

Figure 5. Simplified representation of the pooling layer operation [35].

The purpose of the convolutional neural network is to combine features from the input data to form a wide range of attributes in image classification [31].

Validation success can also be shown computationally through a precision expression adapted from[40]:

$$precision=\frac{TP}{TP+FP}$$

(1)

$$recall=\frac{TP}{TP+FN}$$

(2)

$$F1-score=\frac{2\ast recall\ast precision}{recall+precision}$$

(3)

$$accuracy=\frac{TP+TN}{TP+TN+FN+FP}$$

(4)

wherein:

TP – True Positive

FP – False Negative

TN – True Negative

FN – False Negative

F1-score and accuracy are used to precisely assess model performance, where F1-score takes into account precision and recall, and is calculated for both classes while accuracy represents the proportion of correct recognition of the entire sample.

3. Results and Discussion

3.1. Physical - chemical characteristics of slag

By control particle distribution check in the microsizer, 72.15% of the -0.045 mm size class was found for the slag sample after crushing, which corresponds to the standard requirements and can be used for mortar preparation. The results are shown in Figure 6.

Figure 6. Particle size distribution in the slag sample graph.

Figure 6. Particle size distribution in the slag sample graph.

The chemical composition is given in Figure 7.

Figure 7. Chemical composition of the slag sample.

Figure 7. Chemical composition of the slag sample.

The slag sample is dominated by the contents of SiO₂ and CaO, which affects their chemical activity and volume constancy. The alkali content is low, which is advantageous as their presence is undesirable in cement production. The share of MgO in the slag is increased compared to the allowed 5%. A higher quantity of MgO can affect, in addition to volume instability, the appearance of cracks [41] in mixtures with cement, which should be paid special attention to in experimental work. The measured slag pH value is 8.

In addition to the high quantities of non-crystallized material, a smaller amount of akermanite (Ak) mineral was identified. The powder diffractogram is shown in Figure 8.

Figure 8. Diffractogram of the slag sample.

Figure 8. Diffractogram of the slag sample.

SEM microphotographs of the initial slag sample are shown in Figure 9. A typical slag appearance can be explained by a porous structure with open pores, on the surfaces of which certain crystals can be observed, which, according to the available literature, can be classified as akermanites [42,43].

Figure 9. SEM image of the slag sample (a) magnification 50x, (b) magnification 200x, (c) magnification 500x and (d) magnification 1000x.

Figure 9. SEM image of the slag sample (a) magnification 50x, (b) magnification 200x, (c) magnification 500x and (d) magnification 1000x.

Based on the detailed sample characterization and the obtained results, it was concluded that the tested slag sample can be used as an additive for cement mortars after micronization.

3.2. Physical - chemical characteristics of mortar

A new type of cement was prepared with added 10 mas% slag. After the preparation and curing, with and without slag, in a period of 2, 28 and 240 days, the mortar samples were tested for flexural and compression strength. In addition, the setting time of both cement mixtures was determined and their chemical composition was analyzed. The results of geomechanical tests of standard and laboratory sample prisms of the new mortar system are given in Table 3 and Figure 10.

Table 3. Mechanical characteristics of standard and laboratory prisms.

Table 3. Mechanical characteristics of standard and laboratory prisms.

System/ strength	Flexural 2 days, MPa	Compressive 2 days, MPa	Flexural 28 days, MPa	Compressive 28 days, MPa	Flexural 240 days, MPa	Compressive 240 days, MPa
A* standard	5	30.63	6.47	59.33	8.01	61.42
Az10 standard	3.44	24.22	6.39	58.61	8.23	59.71
A laboratory			11.44	68.90	9.58	66.29
Az10 laboratory			7.94	41.08	9.10	53.07

* Legend: A - CEM I, Az10 -CEM I with the addition of 10 mas% slag, standard 40x40x160 mm and laboratory 10x10x60 mm.

Figure 10. Mechanical characteristics of standard and laboratory prisms.

Figure 10. Mechanical characteristics of standard and laboratory prisms.

The standard does not prescribe flexural strength values, but they are always given in reports, while compressive strength values are an important parameter that is tested and defined by the standard or based on the Rulebook on Cement Quality [44]. The compressive strength decreases with the addition of slag, but in both systems, the strength increases with the hydration time. By comparing the values obtained for standard and laboratory prisms, it can be concluded that the compressive strength results for standard prisms fully follow the compressive strength values for the laboratory prisms. Thus, the results are comparable. This conclusion is a good indicator that the following experiments can fully rely on laboratory samples and raw materials saving.

The setting time is shown in Table 4.

Table 4. Alite systems setting time.

Table 4. Alite systems setting time.

System/ Setting time	Setting to start time, min	Setting to finish time, min	Total setting time, min
A	86	212	126
Az10	115	223	108

The new slag mortar system has a total setting time of 108 minutes which is shorter than the setting time of CEM I (alite) mortar which is 126 minutes, but takes longer for the setting to start.

The chemical composition of the alite and the new mortar system is shown through silicate analysis given in Table 5.

Table 5. Chemical composition of the alite and the new mortar system.

Table 5. Chemical composition of the alite and the new mortar system.

	Content, %
System	SiO2	Fe2O3	Al2O3	CaO	MgO	SO3	P2O5	TiO2	Na2O	K2O	Annealing loss
A	22.10	1.57	4.45	36.16	2.11	3.15	0.10	0.28	0.55	0.65	3.85
Az10	22.80	1.36	4.66	50.37	2.73	3.62	0.09	0.26	0.46	0.82	3.63

The chemical composition of the alite and the new mortar system (Az₁₀) is shown based on silicate analysis given in Table 5. The results of the silicate analysis show that the annealing loss in both systems is below the value prescribed by the Rulebook [44]. The sulfate content (as SO₃) is below the limit value for both types of cement, so there is no possibility of subsequent sulfate corrosion. The content of CaO is much higher in the slag system, which originates from the slag itself in which it is present with 40.82% and can negatively affect the volume stability of this cement system. The total alkali content expressed as Na₂O+K₂O is rather high, but these values do not exceed the limit (<1.5). The presence of alkali in the cement can lead to a later reaction with the aggregate in the concrete, causing a decrease in the structure strength [45].

3.3. Chemical characteristics of Danube river water

Danube water used in the experiments was tested for pH, total nitrogen, total phosphorus, total organic carbon, ammonia, nitrates, nitrites, chlorides, COD (Chemical Oxigen Demand), BOD₅ (Biochemical Oxigen Demand), sulfates, phosphates, dissolved oxygen, fluorides, hexavalent chromium, calcium, magnesium, manganese, lead, zinc, total chromium, cadmium, copper, mercury, arsenic, iron (Table 6) [46].

Table 6. Analysis of Danube river water.

Table 6. Analysis of Danube river water.

parameter	pH	total nitrogen	total phosphorus		total organic carbon	ammonia	nitrates
measured value	8.47	1.35	0.047		1.89	<0.078	1.317
reference value	6.5-8.5	2.0	0.2		5.0	0.3	3.0
parameter	nitrites	COD	BOD5		sulfates	phosphates	dissolved oxygen
measured value	0.019	8.0	1.2		37.38	0.044	8.86
reference value	0.03	15	5.0		100	0.1	7.0
parameter	fluorides	hexavalent chromium	calcium		manganese	manganese	lead
measured value	<0.5	<0.1	53.0		12.86	0.01	<0.01
reference value	-	0.1	-		-	0.1	0.05
parameter	zinc	total chromium	cadmium	copper	mercury	arsenic	iron
measured value	<0.03	<0.006	0.0009	<0.02	<0.0003	<0.01	0.122
reference value	0.7	0.05	0.005	0.022	0.001	0.01	0.5

The analysis determined the presence of all tested parameters in values below the maximum allowed, i.e. reference values, so that no negative influence of water was expected in further experiments of the bacterial healing process.

3.4. Bio-stimulated healing of cracks

After setting up the bacterial experiments, the changes in the cracks were monitored with a portable microscope after 7, 14 and 28 days (Table 7 shows the changes after 7 days under three characteristic conditions). In Table 7, all mean values of the change in cracks are shown tabularly (Table 8) and graphicaly (Figure 11).

Table 7. Changes in the new mortar sample after 7 days under controlled conditions, optical microscopy.

Table 7. Changes in the new mortar sample after 7 days under controlled conditions, optical microscopy.

Conditions/days	0 days	7 days
First group/condition
Third group/condition
Fifth group/condition

Table 8. Mean crack width of the new mortar system with healing efficiency by conditions.

Table 8. Mean crack width of the new mortar system with healing efficiency by conditions.

	Group/condition
Healing	First	Second	Third	Fourth	Fifth
Initial width, mm	0.504	0.466	0.567	0.567	0.569
Width after 7 days, mm	0.361	0.336	0.466	0.508	0.445
Efficiency, %	28.12	27.37	16.75	9.96	22.88
Initial width, mm	0.542	0.659	0.638	0.487	0.588
Width after 14 days, mm	0.347	0.468	0.529	0.357	0.330
Efficiency, %	32.02	29.04	16.90	26.73	42.74
initial width, mm	0.447	0.722	0.517	0.571	0.506
Width after 28 days, mm	0.346	0.483	0.407	0.399	0.357
Efficiency, %	22.39	33.68	20.99	29.99	27.80

Figure 11. Efficiency of the new mortar system by days and conditions.

Figure 11. Efficiency of the new mortar system by days and conditions.

Although, according to literature data, bacteria reach their precipitation maximum after 7 days [29], when they can create a sufficient amount of calcium carbonate to close a 0.5-0.8 mm wide crack [47], the most favorable healing values were recorded with the new slag mortar system after 14 days in both waters, with efficiency being 32.02% and 42.74% for sterile demineralized and Danube water, respectively.

Bio-stimulated healing rates after 7 and 28 days range between 22.39% and 28.12%, respectively. Healing was also recorded under bacteria-free conditions, namely, in sterile demineralized water 20.99% and Danube water 29.99% with initial cracks of about 0.5 mm, which can be explained by the process of autogenous healing due to unreacted mortar and slag grains [13,15].

Figure 12 (a) and (b) shows SEM images of the sample with bacteria in sterile demineralized water at 1000 x magnifications and the characteristic morphology of calcite crystals that started the process of bio-stimulated healing in the cracks as the result of bacterial precipitation. The appearance of the crystals is confirmed by the previous studies [48]. The EDS spectrum in Figure 12 (c) and (d) confirms the presence of calcium, oxygen and carbon, which once again indicates that it is calcium carbonate. Previously published studies show that the main healing product is also calcium carbonate [17], which was confirmed in this research.

Figure 12. SEM, EDS results of the sample with bacteria in sterile demineralized water (a) and (b) 1000x magnification, (c) and (d) EDS spectrum at 1000x magnification.

Figure 12. SEM, EDS results of the sample with bacteria in sterile demineralized water (a) and (b) 1000x magnification, (c) and (d) EDS spectrum at 1000x magnification.

SEM images of the sample (Figure 13 (a)) in the sterile demineralized water without nutritient and bacterial culture show that the crack has not self-healed and the crystals at 500x and 2000x magnifications (Figure 13 (b) and (c)) are characteristic of slag mortars and are in line with previously published studies [49].

Figure 13. SEM, EDS images of the sample without bacteria in sterile demineralized water (a) 50x magnification, (b) 500x magnification, (c) 2000x magnification, (d) and (e) EDS spectrum at 2000x magnification.

Figure 13. SEM, EDS images of the sample without bacteria in sterile demineralized water (a) 50x magnification, (b) 500x magnification, (c) 2000x magnification, (d) and (e) EDS spectrum at 2000x magnification.

The SEM images and EDS of the sample treated with bacteria in Danube water (Figure 14) also confirmed the precipitation of calcium carbonate as the main compound that fills the cracks and leads to healing process. Calcium carbonate crystals are larger with clear rosettes, unlike crystals formed in sterile demineralized water where the crystals are smaller and more uniform.

Figure 14. SEM, EDS results of the sample with bacteria in Danube water a) 500x magnification, b) and c) EDS spectrum at 3000x magnification.

Figure 14. SEM, EDS results of the sample with bacteria in Danube water a) 500x magnification, b) and c) EDS spectrum at 3000x magnification.

3.5. The success of crack image classification using CNN

ResNet50 program package within Matlab2023A was used to classify images of cracks in the new slag mortar system after a period of bacterial treatment in sterile demineralized and Danube water.

The microscopic images showing bacterial healing in sterile demineralized water and in Danube water are visually very similar and can be classified into the same data set (Class 1). For the second data set (Class 2), images from mortar samples with demineralized water without nutritient and bacterial culture, were selected.

Each group had 102 input images. After entering the input data into the network, the program extracts one image from each group as shown in Figure 15.

Figure 15. Typical image of a sample from the group left) with bacteria and right) without bacteria.

Figure 15. Typical image of a sample from the group left) with bacteria and right) without bacteria.

The CNN had 177 nodes with 192 connections classified into input, convolutional, pooling, fully connected, ReLu and output layers. Figure 16 shows the simplified architecture of convolutional neural network 50-layer ResNet while the network architecture is broken down by layers as given in Table 9.

Figure 16. Simplified architecture of convolutional neural network 50-layer ResNet, showing convolutional layers (C1,1, C2,1, … C5,9), pooling layers (P1 and P2) and a fully connected layer (FC1).

Figure 16. Simplified architecture of convolutional neural network 50-layer ResNet, showing convolutional layers (C1,1, C2,1, … C5,9), pooling layers (P1 and P2) and a fully connected layer (FC1).

Table 9. Architecture of the resulting network.

Table 9. Architecture of the resulting network.

For network training and for the test, the program selected 62 (30 %) and 142 (70 %) input images, respectively. The first covolutional matrix with the number of images obtained by the test is:

59	12
11	60

After reprocessing the data through the network, a generated matrix was obtained, shown with accuracy in relation to the number of images from the test (Figure 17):

Figure 17. Generated matrix.

Figure 17. Generated matrix.

The final classification accuracy using this deep learning model is 91.55%.

To validate the classification results, a new data set with 18 images for each group was selected. After validation, a deviation was obtained in the group with bacteria for one input image (5.5 %), while in the group without bacteria the deviation for two input images was (11 %).

Validation success is also obtained from Equations (1) to (4) and the results are given in table 10.

Table 10. Indicators for each class.

Table 10. Indicators for each class.

Indicators	Precision	Recall	F1 - score	Accuracy
Class 1	0.94029	0.88732	0.91303	0.91549
Class 2	0.89333	0.94366	0.91781

The characteristics obtained in this way confirmed the obtained validation values and amount to 91.55%.

Based on the obtained CNN network data, it can be concluded that the model is suitable for the classification of images of changes in mortar sample cracks.

4. Conclusion

References

1. Mahmod, A. K.; Al-Jabbar, L.A.; Salman M. M. Bacteria Based Self-Healing Concrete : A Review, 2 nd online Scientific conference for Graduate Engineering Students, Journal of Engineering and Sustainable Development, 2021, 43-56.
2. Suleiman, A. R.; Zhang, L. V.; Nehdi, M. L. Quantifying Crack Self-Healing in Concrete with Superabsorbent Polymers under Varying Temperature and Relative Humidity, Sustainability 2021, 13, 13999. [CrossRef]
3. Termkhajornkit, P.; Nawa, T.; Yamashiro, Y.; Saito, T. Self-healing ability of fly ash–cement systems, Cement and Concrete Composites 31, 2009, 195–203. [CrossRef]
4. Van Breugel, K. Is there a market for self-healing cement-based materials, Proceedings of the First International Conference on Self Healing Materials, Noordwijk aan Zee, The Netherlands, 2007. 1-9.
5. Van Tittelboom, K.; De Belie, N. Self-Healing in Cementitious Materials—A Review, Materials, 2013, 6, 2182-2217. [CrossRef]
6. Mors, R. M.; Jonkers, H. M. Bacteria-Based Self-Healing Concrete – Introduction, Second International Conference on Microstructural-related Durability of Cementitious Composites, Amsterdam, The Netherlands, 2012, 1-7.
7. Mauludin, L. F.; Oucif, C. Modeling of Self-Healing Concrete: A Review, Journal of Applied Computational Mechanics, 5, 3, 2019, 526-539.
8. Zhuang, X.; Zhou, S. The Prediction of Self-Healing Capacity of Bacteria-Based Concrete Using Machine Learning Approaches CMC, 2019, vol. 59, no. 1, 57-77.
9. Huang, X.; Sresakoolchai, J.; Qin, X.; Fan Ho, Y.; Kaewunruen, S. Self-Healing Performance Assessment of Bacterial-Based Concrete Using Machine Learning Approaches, Materials, 2022, 15, 4436. [CrossRef]
10. Zhutovsky, S., Nayman, S. Modeling of crack-healing by hydration products of residual cement in concrete, Construction and Building Materials, 340, 2022, 127682. [CrossRef]
11. Van der Bergh, J. M.; Miljević, B.; Šovljanski, O.; Vučetić, S.; Markov, S.; Ranogajec, J.; Bras, A. Preliminary approach to bio-based surface healing of structural repair cement mortars, Construction and Building Materials, 248, 2020, 118557.
12. Roy, R. Bacteria-based self-healing mortar with bio-plastic healing agents, Kth Royal Institute of Technology School of Architecture And The Built Environment, 2020, 176.
13. Grubišić, I. Samozacjeljivanje betona autogenim i autonomnim postupkom s naglaskom na metodu bakterija, Diplomski rad, Sveučilište u Splitu, Fakultet građevinarstva, arhitekture i geodezije, 2020. in Serbian.
14. Zhang, W.; Zheng, Q.; Ashour, A.; Han, B. Self-healing cement concrete composites for resilient infrastructures: A review, Composites Part B 189, 2020, 107892. [CrossRef]
15. Guzlena, S.; Sakale, G. Self-healing concrete with crystalline admixture – a review, IOP Conf. Series: Materials Science and Engineering 660, 2019, 012057. [CrossRef]
16. Van der Bergh, J. M.; Miljević, B.; Vučetić, S.; Šovljanski, O.; Markov, S.; Riley, M.; Ranogajec, J.; Bras, A. Comparison of Microbially Induced Healing Solutions for Crack Repairs of Cement-Based Infrastructure, Sustainability 2021, 13, 4287.
17. Li, G.; Liu, S., Niu, M.; Liu, Q.; Yang, X.; Deng, M. Effect of granulated blast furnace slag on the self-healing capability of mortar incorporating crystalline admixture, construction and building materials 239, 2020, 117818. [CrossRef]
18. Šovljanski, O.; Tomić, A.; Markov, S. Relationship between Bacterial Contribution and Self-Healing Effect of Cement-Based Materials, Microorganisms 2022, 10, 1399. [CrossRef]
19. Sagripanti, J. L.; Bonifacino, A. Comparative Sporicidal Effects of Liquid Chemical Agents, Applied and Environmental Microbiology, Vol. 62, No. 2, 1996, 545–551. [CrossRef]
20. Algaifi, H. A.; Abu Bakar, S.; Sam, A. R. M.; Abidin, A. R. Z.; Shahir, S.; AL-Towayti, W. A. H. Numerical modeling for crack self-healing concrete by microbial calcium carbonate, Construction and Building Materials, 189, 2018, 816-824. [CrossRef]
21. Althoey, F.; Amin, M. N.; Khan, K.; Usman, M. M.; Ali Khan, M.; Javed, M. F.; Sabri, M. M. S.; Alrowais, R.; Maglad, A. M. Machine learning based computational approach for crack width detection of self-healing concrete, Case Studies in Construction Materials, 17, 2022, e01610. [CrossRef]
22. Technical documentation of the Department of Carbohydrate Foods of the Faculty of Technology in Novi Sad.
23. Documentation of the Mining Institute.
24. Documentation of the Faculty of Mining and Geology in Belgrade.
25. Documentation of the University Center for Electron Microscopy in Novi Sad (Department of Biology and Ecology, Faculty of Science).
26. Standard SRPS EN 196-3:2017.
27. Standard SRPS EN 196-1:2017.
28. Žarković, D. B. Jonska hromatografija – razvoj metode za analizu i kontrolu kvaliteta vode u proizvodnji papira, doktorska disertacija, Tehnološko–metalurški fakultet, Univerzitet u Beogradu, 2011. in Serbian.
29. Šovljanski, O. Mikrobiološka precipitacija karbonata –od odabira induktora do ispitivanja bioprocesnih parametara, doktorska disertacija, Tehnološki fakultet Novi Sad, 2021. in Serbian.
30. Ševo, I. Specijalizovana neuronska mreža za klasifikaciju i segmentaciju aero-snimaka, doktorska disertacija, Elektrotehnički fakultet, Univerzitet u Banjoj Luci, 2020. in Serbian.
31. Neuronske mreže i duboko učenje Mašinsko učenje 2020/21. Matematički fakultet Univerzitet u Beogradu, in Serbian.
32. Matoš, I. Klasifikacija prometnih znakova korištenjem konvolucijskih neuronskih mreža, diplomski rad, Fakultet elektrotehnike, računarstva i informacijskih tehnologija, Sveučilište Josipa Jurja Strossmayera u Osijeku, 2020. in Croatian.
33. Shang, K. FSA, CFA, PRM, SCJP, Applying Image Recognition to Insurance, Society of Actuaries Research Expanding Boundaries Pool, 2018.
34. Repić, M.; Ralević, N. Primena konvolucionih neuralnih mreža kod prepoznavanja slika u osiguranju, Zbornik radova Fakulteta tehničkih nauka, Novi Sad, UDK: 517.98 DOI: https://doi.org/10.24867/13JV01Repic. [CrossRef]
35. Zariea, M.; Jahedsaravania, A.; Massinaeib, M. Flotation froth image classification using convolutional neural networks Minerals Engineering, 155, 2020, 106443.
36. Lu, F.; Liang, Y.; Wang, X.; Gao, T.; Chen, Q.; Liu, Y.; Zhou, Y.; Yuan, Y.; Liu, Y. Prediction of amorphous forming ability based on artificial neural network and convolutional neural network, Computational Materials Science, 210, 2022, 111464. [CrossRef]
37. Nakhaei, F.; Rahimi, S.; Fathi, M. Prediction of Sulfur Removal from Iron Concentrate Using Column Flotation Froth Features: Comparison of k-Means Clustering, Regression, Backpropagation Neural Network, and Convolutional Neural Networ, Minerals 2022, 12, 1434. [CrossRef]
38. https://hashdork.com/sr/neural-network/.
39. Dabović, M. M.; Tartalja, I. I. Duboke konvolucijske neuronske mreže – koncepti i aktuelna istraživanja, Zbornik 61. Konferencije za elektroniku, telekomunikacije, računarstvo, automatiku i nuklearnu tehniku, ETRAN 2017, ISBN 978-86-7466-692-0, VI1.1.1-6. in Serbian.
40. Gao, X.; Tang, Z.; Xie, Y.; Zhang, H.; Gui, W. A layered working condition perception integrating handcrafted with deep features for froth flotation, Minerals Engineering 170, 2021, 107059. [CrossRef]
41. Bušatlić, I.; Bušatlić, N.; Merdić, N.; Haračić, N. Osnove hemije i tehnologije portland cementa, Zenica, 2020, ISBN 978-9958-17-168-0, p. 385. in Bosnian.
42. Duana, P.; Shuia, Z.; Chena, W.; Shenb, C. E nhancing microstructure and durability of concrete from ground granulated blast furnace slag and metakaolin as cement replacement materials, J. Mater. Res. Technol. 2013, 2, 1, 52-59.
43. Han, Z.; Feng, P.; Gao, C.; Shen, Y.; Shuai, C.; Peng, S. Microstructure, mechanical properties and in vitro bioactivity of akermanite scaffolds fabricated by laser sintering, Materials Science, DOI:10.3233/BME-141017,. [CrossRef]
44. Pravilnik o kvalitetu cementa ("Sl. glasnik RS", br. 34/2013 i 44/2014) in Serbian.
45. Ilić, I. Mogućnost primene mikroniziranog i klasiranog elektrofilterskog pepela kao aditiva za proizvodnju građevinskih materijala, Magistarska teza, Rudarsko–geološki fakultet, Univerzitet u Beogradu, 2009. in Serbian.
46. Archive of technical documentation of Thermal Power Plant Kostolac.
47. Grubišić, I. Samozacjeljivanje betona autogenim i autonomnim postupkom s naglaskom na metodu bakterija, Diplomski rad, Sveučilište u Splitu, Fakultet građevinarstva, arhitekture i geodezije, 2020. in Croatian.
48. Vučetić, S.; Čjepa, D.; Miljević, B.; van der Bergh, J. M.; Šovljanski, O.; Tomić, A.; Nikolić, E.; Markov, S.; Hiršenberger, H.; Ranogajec, J. Bio-Stimulated Surface Healing of Historical and Compatible Conservation Mortars, Materials 2023, 16, 642.
49. Schneider, N.; Stephan, D. Reactivation of a Retarded Suspension of Ground Granulated Blast-Furnace Slag, Materials 2016, 9, 174. [CrossRef]