3.1 Rheology properties of casein solution
The rheological properties of the casein solution were studied to understand their influence on the geotechnical characteristics of casein-treated soil. Factors such as casein content, salt effects, pH effects, temperature, and shear rate (rotated speed) were considered. The viscosity test was conducted with a constant alkaline-to-casein ratio and varying casein-to-soil ratios. The viscosity of the casein solutions was found to decrease with increasing shear rate, regardless of the alkaline type or casein dosage. Lower casein concentrations showed similar viscosity at both low and high rotational speeds (Fig. 1a). It was also observed that casein solutions with concentrations below 12% exhibited low viscosity and behaved as nearly Newtonian fluids[13–14]. The pH values fluctuated due to changes in alkaline concentrations as the casein dosage increased, resulting in similar viscosity for lower casein content. Ca(OH)2-casein showed higher viscosity than NaOH-casein, attributed to the larger molecular weight of Ca(OH)2. NaOH-casein demonstrated greater workability when mixed with soil and did not require additional initial water content to increase soil strength, unlike Ca(OH)2-casein. The viscosity of the casein solution decreased as the initial water content increased. At a 3% casein-to-soil ratio, a power line is a fitting line between initial water content and consistency (Fig. 1b). When the initial water content increased from 14–16% to 18% at a speed of 5 rpm, the viscosity decreased from 623.78 cp to 59.7 cp to 41.35 cp, and this viscosity decreased even more when the speed was increased to 100 rpm.
Similar viscosity trends were observed in a solution with a 5% Ca(OH)2-casein to soil ratio. As the initial water content increased from 14–22% at a speed of 5 rpm, the viscosity of the solution decreased from 90942 cp to 179 cp. The viscosity also decreased as the rotational speed increased. The higher viscosity at a 5% casein concentration can be attributed to the entanglement of biopolymers in the solution. Additionally, the viscosity decreased with increasing initial water content due to lower concentrations of casein and alkaline in the solution. Previous research showed that the viscosity of caseinate increased with pH to a maximum at pH 9.8–10.0 before rapidly dropping. The viscosity of caseinate was also concentration dependent. The presence of swollen, partially solvated caseinate particles led to the significant increase in viscosity with concentration[12, 15, 16]. A higher initial water content was necessary to improve the soil's workability with the solution. Similar results were observed for the 5% NaOH-casein to soil ratio, with increasing viscosity with shear rate. After an hour of solution preparation, the 5% Ca(OH)2-casein solution exhibited gel behavior at 14% initial water content, while the 5% NaOH-casein solution had lower viscosity and required more time to transform into a gel state compared to the Ca(OH)2-casein solution.
3.2 Dry condition strength properties of casein-treated soils
3.2.1 The influence of alkaline type and content on the strength properties of casein-amended soil
The strength properties of casein-treated soils under dry conditions were investigated. The influence of alkaline type and content on the strength properties was studied using a 3% casein to soil ratio. Figures 3a and 3b summarize the values for UCS (Unconfined Compressive Strength), E50, ɛf, and absorbed energy for casein-treated sand and natural soil with different alkaline types and contents.
It was observed that increasing the alkaline ratio from 7–14% resulted in improved strength, with the highest UCS achieved at an alkaline-to-casein ratio of 14% for both sodium hydroxide and calcium hydroxide. However, increasing the alkaline ratio beyond 14% decreased the strength, although it still provided stronger strength compared to an alkaline ratio of 7%. The low strength observed at a 7% alkaline ratio, especially for calcium hydroxide, was attributed to insufficient dissolution of casein in the solution. This led to non-uniform dispersion of casein particles, causing some particles to react individually in the soil.
Past research suggests that casein solutions at pH levels lower than eight exhibit strong intermolecular interactions, resulting in powerful attractive interactions between casein particles [13, 14]. In this study, the 7% Ca(OH)2 to casein ratio caused casein particles to coagulate, leading to a non-homogeneous mixture with sand or natural soil. Calcium caseinate performed better in sandy soil, while sodium caseinate performed better in natural soil. This difference in effectiveness is attributed to the viscosity of the solutions, with sodium caseinate having lower viscosity than calcium caseinate. The lower viscosity of sodium caseinate allows for better and more uniform mixing with natural soil containing fine particles.
Comparing the strengths achieved, sodium caseinate-treated natural soil showed a strength of 5008.215 KPa at a 14% NaOH to casein ratio, while calcium caseinate-treated natural soil reached a strength of 626.755 KPa at the same 14% Ca(OH)2 to casein ratio. This difference in effectiveness can be attributed to the higher solubility of casein in NaOH compared to Ca(OH)2. The strong interaction between the negatively charged sodium caseinate particles and the positively charged fine soil particles contributed to the high strength achieved. Furthermore, increasing the initial water content of calcium caseinate improved its effectiveness in both sandy and natural soil[17, 18].
The failure strain data in Fig. 3b shows that there is no correlation between NaOH levels and the ductility of the soil, as similar failure strains were observed at different NaOH levels. In sandy soil, increasing Ca(OH)2 from 7–14% resulted in an increase in failure strain from 3.88–5.48%. This indicates that raising Ca(OH)2 up to the optimal ratio improves the ductility of the soil. This increase in ductility can be attributed to the presence of thick and high-tensile biopolymer films that dehydrate and position themselves among the sand particles. However, as the Ca(OH)2 concentration further increased to 21%, the ductility decreased. This decrease in ductility is associated with the higher pH of the casein solution, which affects the behavior of casein molecules. The high pH leads to the precipitation of serum calcium phosphate onto casein micelles, reducing the cohesive contact and causing the dissolution and dissociation of casein micelles[19, 20].
In natural soil, increasing Ca(OH)2 resulted in a decrease in failure strain. The increase in alkaline content strengthens the bond between the casein particle and the fine soil due to the increased negative charge of the casein. In natural soil matrices, the chemical interaction between small soil particles and casein dis predominates over the entanglement of the casein biopolymer. Regardless of soil type or alkaline type, increasing alkaline content initially increases soil stiffness up to the optimal ratio, after which it slightly decreases. However, in calcium caseinate-treated sandy soil, the highest stiffness was observed at the highest Ca(OH)2 level. The E50 (elastic modulus) of calcium caseinate increased from 23134.5 kPa to 38634.63 kPa to 75080 kPa as the Ca(OH)2 to casein ratio increased from 7–14% to 21%. Energy absorption also increased monotonically with alkaline content up to the optimal ratio. For example, in sandy soil, energy absorption increased by 2.67 times when the Ca(OH)2 to casein ratio increased from 7–14%. Specimens with optimal alkaline content exhibited higher buffering capacity, enabling them to better handle applied loads. The polymer treatment of soil improves its ductility and reduces brittleness [21]. Energy absorption closely correlates with the soil's unconfined compressive stress, exhibiting a similar pattern. The treatment of soil with the optimal alkaline content allows for greater deformation at higher strain levels due to the mobilization of significant tensile stresses and improved strength.
3.2.2 The influence of casein content on the strength properties
The influence of casein content on the strength properties is examined. Increasing the dosage of both calcium caseinate and sodium caseinate biopolymers leads to an increase in soil strength. The rate of strength enhancement depends on the type of alkaline used and the soil type.
In sandy soil, raising the calcium caseinate concentration from 0.5–3% increases the dry strength from 151.95 KPa to 2933.611 KPa, but the strength decreases to 586.955 KPa at a 5% casein ratio. This decrease in strength at 5% can be attributed to sand particle coagulation and difficulty in compacting the soil due to the change of casein from a fluid to a gel status. Increasing calcium caseinate from 0.5–2% in natural soil raises the dry strength from 120.22 KPa to 1671.024 KPa, but it decreases to 626.755 KPa at a 3% casein ratio. The non-homogeneous mixture of fine particles and casein solution in natural soil requires a higher initial water content for an efficient soil-casein mixture, which may account for the lower strength at a 3% dosage.
Sodium caseinate is more effective than calcium caseinate at various casein dosages in both soil types. The lower viscosity of sodium caseinate improves workability. For example, raising sodium caseinate from 1–5% in sandy soil increases the dry strength from 777.85 KPa to 4077.935 KPa. Similar results are observed in natural soil, where the strength increases from 1166.08 KPa to 6161.366 KPa with an increase in sodium caseinate dosage from 1–5%. The strong chemical bonds already present between the fine particles and sodium caseinate make it more effective in natural soil. Wet mixing is used in the study, which is known to be more successful for treating cohesive soils[22]. Even at lower dosages, the anionic character of both sodium caseinate and calcium caseinate promotes better aggregation and binding, reducing porosity. The increase in strength with an increase in biopolymer dosage can be attributed to increased viscosity of the biopolymer gel, leading to soil particle aggregation and inter-particle interaction[23–25].
Figure 4b shows that the ductility of sandy soil increases with increasing casein concentration up to 3% for both sodium caseinate and calcium caseinate. However, the ductility decreases at 5% concentration. The higher concentration of casein allows it to withstand more stretching before deforming due to increased entanglement and stronger chains. For natural soil, sodium caseinate demonstrates greater ductility compared to calcium caseinate, and there is less variation in failure strains at different casein doses. In natural soil, the chemical bonding between fine particles and casein is more significant than casein entanglement. The elastic modulus (E50) and energy absorption show the same pattern as the unconfined compression strength (UCS) when the casein dosage is increased. The calcium caseinate-treated sand absorbs more energy compared to the sodium caseinate-treated sand due to the higher entanglement and viscosity of calcium caseinate. However, the deformation of sodium caseinate-treated natural soil occurs at its highest failure strain, which is opposite to sandy soil.
3.2.3 The influence of the initial water content on the strength properties
Influence of initial water content on the strength properties is shown in Fig. 5. The UCS initially increases dramatically as the initial moisture content increases from 14–20% for a specific calcium caseinate to sand ratio of 5.0%. However, further increase in water content causes a decrease in UCS. Insufficient water negatively impacts the workability of the biopolymer-soil matrix and its mechanical strength, while excessive water leads to large voids in the matrix, reducing soil strength. However, in the study, adding more water beyond the optima ratio does not significantly reduce the soil's strength. The rate of increase with growing initial water content is substantial in sandy soil and natural soil due to the rheological characteristics of the casein solution and its workability. For sandy soil, the initial increase in UCS is significant, but raising the initial water content to 20% causes the soil mixture to become liquid and difficult to remold. The calcium caseinate level affects the amount of absorbed water required for good mixing efficiency.
Contrarily, raising the initial water content is ineffective for sodium caseinate-treated sand and natural soil, where the increase rate is minimal. The reduced viscosity of sodium caseinate doesn't considerably change the consistency of the soil-biopolymer matrix when the initial water content is increased. Dry unit weight has a significant impact on strength metrics, and the specimens prepared at the optimal water content show the highest strength. It should be noted that compaction and index properties of both soil types are modified with the addition of calcium caseinate, and the highest strength is achieved at the optimum water content. Specimens prepared on the dry side of the optimal water content have lower undrained shear strength values than those prepared at the optimal water content obtained from UCS[26]. In general, Fig. 5 shows that there is no trend between failure strain and initial water content for all alkalinity and soil types. Increasing initial water content to the level that produced maximum UCS (20% for calcium caseinate and 18% for sodium caseinate) resulted in greater strain and ductility in treated sand. The highest failure strain for 5% sodium caseinate treated natural soil was obtained at 16% initial water content, while for 3% and 5% calcium caseinate treated natural soil, it was at 18% and 20% initial water content, which did not correspond to the highest UCS. The entanglement of casein gel and its tensile strength in the soil matrix increase soil flexibility. Raising the initial water content reduces casein solution viscosity, improving workability and allowing the biopolymer to effectively fill soil voids and work under external load.
The UCS and failure strain have a direct impact on energy absorption: Energy absorption increases with initial moisture content and is directly related to the tensile strength of the casein gel. As an example, increasing the initial water content of calcium caseinate-treated sand from 14–20% resulted in an increase in energy absorption from 13.41104 KJ/m3 to 163.3002 KJ/m3. Similar patterns were observed in other types of soil and alkaline in Fig. 5b. For soils treated with calcium caseinate, the elastic modulus (E50) showed an ascending trend with initial water content (in sand and natural soil). However, for soils treated with sodium caseinate, E50 decreased as the initial water content increased (in sand and natural soil). Increasing the initial water content in sodium caseinate-treated soils was not necessary and led to a reduction in stiffness correlated with lower sodium caseinate viscosity and increased water content. Raising the initial water content may weaken stiffness due to the soil's void ratio after drying.
3.2.4 The long-term durability properties
Polymers used in geotechnical engineering, such as polyethylene, starch, and polyester, often face decomposition issues[27, 28]. However, research on xanthan gum-treated Red Yellow soil shows slight or constant stability with a slight increase in elastic modulus and compressive strength[29]. Casein biopolymer has shown high strength compared to other biopolymers during re-immersion, but its strength degrades when temperatures rise above 60°C[6, 7].
The long-term resistance of calcium caseinate-treated sand and natural soil to decomposition was tested by keeping them in a nonstandard laboratory room for 180 days. The UCS of calcium caseinate-treated sand and natural soil after 180 days showed fluctuating changes between decrease and increase, indicating stability under dry conditions (Fig. 6a).
The compressive strength of calcium caseinate-treated sand marginally decreased after 180 days, while the rate of strength reduction decreased with increasing casein concentration. The strength of calcium caseinate-treated natural soil increased after 180 days, except for a slight decrease at 3% casein concentration.
As shown in Fig. 6b, Deformation of calcium caseinate-treated sand and natural soil occurred at lower failure strain values after 180 days, reducing soil ductility compared with short time treatment. The stiffness of the soil cured for a long time compared to a short time showed fluctuation between decrease and increase rates, depending on biopolymer content and soil type. Energy absorption improved for calcium caseinate-treated natural soil after 180 days, except for 3% casein concentration. However, the ability of calcium caseinate-treated sand to absorb energy reduced for all biopolymer contents after 180 days. Casein is resistant to thermos-decomposition at temperatures below 60°C and has adhesive properties.
Further research is needed to confirm the stability of casein under cyclic wetting, drying, and freeze-thaw cycles.
3.3 Strength properties under immersion conditions
While achieving high initial bond strengths with biopolymers in soil is simple, maintaining strong bonds in hostile conditions, especially in the presence of moisture, is challenging. Moisture is the main factor causing degradation of stabilized soil bonds. Bond failures between soil particles, whether in the field or in the laboratory, are primarily caused by moisture[3, 4, 6].
According to our observations, Sodium caseinate-treated soil loses strength after only two hours of re-submergence due to irreversible changes caused by moisture. Moisture can lead to hydrolysis, cracking, crazing, and swelling in the adhesive-treated soil, disrupting secondary bonds at the soil/adhesive interface. Sodium caseinate-treated soil exhibits rapid water absorption, causing swelling and subsequent drainage outside the soil samples, leading to high stresses and crack growth near the bond line. Calcium caseinate behaves differently and maintains better water resistance compared to sodium caseinate. The effects of water immersion on the strength properties of calcium caseinate-treated soil were investigated by resubmerging the specimens for 24 hours and conducting unconfined compressive tests. The results were evaluated based on biopolymer type, short-term curing, initial water content, and long-term durability.
3.3.1 Strength properties of underwater immersion after short-term curing and long-term durability
Figure 7 shows the retained strength of calcium caseinate-treated soil specimens after being resubmerged for 24 hours at different concentrations (0.5%, 1%, 2%, and 3%) and starting water content of 14%. Similarly to our results, previous studies indicated that casein-treated soils have stronger inter-particular binding strength than swelling pressure during saturation [6]. The findings reveal that 0.5% and 1% casein-to-soil ratios did not retain strength after 14 days of drying, while 2% and 3% ratios exhibited the highest wet strength (185.51 KPa and 198.83 KPa).The 5% biopolymer-treated sand, with an initial water content of 14%, lost its water durability over time. The 0.5% and 1% specimens saturated completely and disintegrated upon touch, while the 2% and 3% specimens remained dry on the inside but wet on the exterior as shown in Fig. 8. Calcium caseinate-treated natural soil displayed similar behaviors, with lower wet strength compared to calcium caseinate-treated sand. After 180 days of long-term durability, the wet strength of calcium caseinate-treated sand and natural soil at 2% and 3% ratios degraded, while the 0.5% and 1% ratios maintained some strength. The roughness formed by casein on the outside of the specimens after 180 days of drying slows down water absorption, contributing to their water resistance.
Casein's ductility did not significantly increase when submerged in water, but the failure strain changed after 180 days of drying. The elastic modulus under wet conditions increased with increasing biopolymer content, and samples submerged after 14 days of drying showed higher values than those submerged after 180 days of drying. Energy absorption followed a similar trend, with the samples submerged after 14 days of drying showing higher values. These results align with previous studies on the strength characteristics of casein-treated soils[1, 6, 30].
3.3.2 The influence of the initial water content on the wet strength
The influence of the initial water content on the wet strength of calcium caseinate-treated soil (sand and natural soil) was investigated. After being dried for 14 days, the specimens with varying initial water content were immersed for 24 hours.
It was found that the wet strength of the soil increased with higher initial water content, following a similar trend to dry conditions. For example, samples with 5% casein/sand increased their wet strength from 89.592 KPa to 833.30 KPa when the initial water content increased from 16–20%.
Previous research by Chang et al. (2018) [6] also reported a high wet strength of 650 KPa at a 25% initial water content. However, our study achieved optimal wet strength of 20% with a lower initial water content. Bozyigit et al. [24] found that specimens prepared on the dry side of the ideal water content had higher strength, while strength decreased with increasing water content on the wet side. Uniform distribution of casein throughout the mixture and a 20% initial water content resulted in decreased voids ratio and improved water resistance. Lay et al. [31] indicated that water resistance of calcium caseinate increased when the calcium hydroxide content reached 30%. It is important to note that the type of soil also plays a role in the water content's effect. In the case of natural soil samples, increasing the initial water content from 14–18% led to improved wet strength, from 66.94KPa to 174.87KPa, for a casein to natural soil ratio of 3%. Similarly, increasing the initial water content from 18–22% improved the wet strength of samples with a casein to natural soil ratio of 5% from 206.661KPa to 1256.619KPa. The highest wet strength was achieved when natural soil was prepared with a 22% initial water content, resulting in lower void ratios and stronger chemical bonds between casein and fine particles. The inter-particle bonding and aggregation effect between casein and soil particles contribute to the overall strength of casein-soil mixtures. The development of a smooth, plastic-like substance on soil surfaces at higher casein concentrations can enhance water resistance and increase wet strength. Managing the initial water content is crucial in improving the water resistance of casein in soil.
3.4 SEM results.
SEM results were used to examine the effects of calcium caseinate and sodium caseinate on the microstructure of sand. The natural state of Harbin Sands was observed in Fig. 10a, showing freely moving particulates with a moderately subrounded form, medium sphericity, and equal grain sizes.
Figure 10b displays sand treated with a 5% calcium caseinate to sand ratio prepared with a 20% initial water content. The calcium caseinate formed a three-dimensional gel network that surrounded and coated the soil grains. Additionally, thick biopolymer connection bridges were observed between sand particles that were not directly in contact, along with the presence of a smooth plastic-like substance. Figure 10c depicts sand that underwent processing with calcium caseinate and was prepped with a 14% initial water content. This image reveals larger spaces between soil particles and nonhomogeneous particle morphologies. Thin biopolymer connection bridges can be seen forming between sand particles that were not directly in contact.
These SEM images showcase the structural alterations caused by the presence of calcium caseinate in the sand samples. The gel formed by calcium caseinate binds and coats the dirt particles, while at 5% sodium caseinate-treated sand, a thinner gel layer and clumpy accumulated casein layer are generated on the sand particles. Figure 10E shows the spherical shape of natural soil particles, with fewer bonds between tightly packed particles due to increased pore density. For natural soil treated with 5% calcium caseinate at 18% initial water content, clumpy accumulated casein layers with thick biopolymer connecting bridges are observed. When natural soil is treated with 5% calcium caseinate at 22% initial water content, larger clumpy accumulated casein layers form and link the natural soil particles together. At 5% sodium caseinate-treated soil, a thin casein coating is present on the natural soil particles, with some binding holding the particles together.
3.5 Relationship between secant modulus and UCS.
The secant modulus (E50), which indicates the resistance to elastic and plastic deformations, is closely related to the unconfined compressive strength (UCS). Figure 11 depicts the relationship between E50 and UCS for all samples, both in dry and wet conditions. The development of E50 is consistent with UCS, and equations 1 and 2 describe the relationship between E50 and UCS, with high coefficients of determination (R2). These equations can be used to understand the relationship between UCS and E50 in casein-treated soils.
E50 = 93.633UCS0.8007 R² = 0.9104 (dry conditions) Eq-1
E50 = 23.838UCS0.919 R² = 0.9239 (wet condition) Eq-2
Previous research has focused on additions like lime and cement, without specific equations for casein-treated soils[32, 33]. The derived power functions offer valuable engineering insights into the UCS-E50 relationship in casein-treated soils.
3.6 The rules of viscosity solution in the strength development
Figure 12 illustrates the relationship between casein solution viscosity and soil strength for a specific casein-to-soil ratio. As the viscosity of the solution decreased, the strength increased due to improved dispersion and easier handling of the mixture. Lower viscosity also reduced the tendency for segregation during compaction, resulting in lower void ratios.
A power relationship between calcium caseinate viscosity and strength was observed, independent of soil type or condition (wet or dry). For example, the power relationship between solution viscosity and the strength of sandy soil treated with 5% calcium caseinate was found to be UCS = 72487ƞ^-0.422 (R² = 0.9299) in dry conditions and UCS = 620572ƞ^-1.106 (R² = 0.9071) in wet conditions. Similar power relationships were observed for natural soil treated with 5% calcium caseinate.
For a casein-to-soil ratio of 3%, a similar pattern was observed, and a power relationship was derived, UCS = 49410ƞ^-0.678 (R² = 1) in dry conditions and UCS = 563.94ƞ^-0.333 (R² = 0.9779) in wet conditions. Eq. 3, UCS = Aƞ^-B, can be applied generally to treat soil with calcium caseinate and design the casein solution for optimal strength.
In the case of soil treated with sodium caseinate, a linear relationship between strength and viscosity was found. For example, the linear equation UCS = -12.621ƞ + 6373.6 (R² = 0.9754) was observed for 5% sodium caseinate-treated sand, and UCS = -2.4519ƞ + 6602.6 (R² = 0.9979) for 5% sodium caseinate-treated natural soil. Sodium caseinate did not significantly affect strength with increasing viscosity, making it less essential to alter the viscosity. However, it also did not demonstrate resistance to submerging conditions, making calcium hydroxide combined with casein a preferred choice for soil stabilization.
These findings provide insights into the relationship between soil viscosity and treatment with casein, enabling the design of casein solutions for optimal strength.
3.7 Artificial Neural network
To predict the compressive strength of casein biopolymer-treated soil without relying solely on costly and time-consuming laboratory testing, artificial neural networks (ANNs) can be utilized. ANNs are powerful predictive models that can analyze the relationship between different parameters and strength development[34, 35].
The ANN model consists of input, hidden, and output layers. Each node, or artificial neuron, in the network is connected to every other node. If the output of a node exceeds a selected threshold value, it is activated and sends data to the next layer. This process allows the network to learn and improve its accuracy over time. In this study, the ANN model utilized a backpropagation neural network algorithm and included nine input parameters: type of soil (TS), type of alkaline (TA), water immersion (WI), long-term durability (LD), casein content (CC), alkaline content (CA), initial water content (IWC), and low-speed viscosity (ƞl) and high-speed viscosity (ƞh) of the material. The output layer represented the compressive strength. The ANN model was created using the neural network toolbox in MATLAB R2021a. Backpropagation was used as the training algorithm, and a linear function was chosen. The performance of the model was evaluated using mean square error (MSE) and R2 measurements. The data were divided into training (70%), validation (15%), and testing (15%) sets. This predictive approach using ANNs provides a more cost-effective and efficient method for determining the optimal mix ratio and predicting the compressive strength of casein biopolymer-treated soil. Figure 13 shows the interpretation diagram of the ANN model in the study.
The ANN model achieved a maximum R2 value of 0.9767 with a hidden layer consisting of ten neurons, indicating strong predictive performance. The regression curve between the experimentally tested data and the predicted data from the ANN model showed high agreement with a coefficient of determination (R2) of 0.9737 (Fig. 14a). A regression plot comparing the overall experimental data and the global predicted data from the ANN model demonstrated a high degree of agreement with an R2 value of 0.9262(Fig. 14b). To quantify the significance of input variables, Garson's algorithm and the connection weights technique were used in this study. The relative importance of the input variables was analyzed based on their absolute contribution values. According to Garson's algorithm, the order of importance for the input variables was as follows: Water immersion, initial water content, lower speed viscosity, casein content, type of soil, alkaline type, long-term durability, alkaline ratio to casein, and high-speed viscosity. Water immersion had a significant impact on casein biopolymer, which aligns with its known sensitivity to water. Casein performed better in wet conditions compared to other biopolymers due to differences in hydrogen bonding and hydrogel behavior[4]. Overall, the ANN model demonstrated strong predictive capabilities, and the analysis of input variables highlighted the importance of water immersion and other factors in the strength development of casein biopolymer-treated soil.
The second component influencing strength is the initial water content, which has a greater impact on strength than solution viscosity. Each biopolymer-soil mixture may have a preferred initial moisture content for achieving the highest compressive strength. The third component, viscosity, also significantly influences the outcome. The viscosity obtained with a lower shear rate should be used in solution preparation as it has a larger impact on soil strength. Fluid viscosity in the soil's pore size can alter its geotechnical behavior. Besides water immersion, these three factors (initial water content, viscosity, and casein content) must be considered when developing a soil stabilization strategy. Other input variables contribute similarly to the outcome, and their variations are negligible.
The connection weights method was used to assess the significance of input variables based on their positive or negative relationship with the outcome. Figure 15b shows that viscosity and immersion in water has a significant detrimental effect on strength, consistent with previous findings.
Casein content, alkaline type, soil type, and initial water content are important factors to consider. The significance of initial water content when considering its direction effects, is not significant according to the connection weights method. The pH value and casein concentrations play primary roles in influencing solution viscosity, therefore using the viscosity of the casein solution in preparations is preferable.