Electroosmotic strengthening of soft clay under different electrification modes

Abstract

Electrification mode is one of the important factors affecting the energy consumption and drainage effect of the electroosmosis method. In this work, a series of laboratory tests were performed to explore the impact of continuous changes in the electric field on the migration of ions and electrons, as well as the migration of electroosmotic flow. The results show that different electrification modes had a significant impact on the current and water discharge. An electrode reversed from anode to cathode, from one circuit to another, is beneficial to drainage, while an electrode reversed from cathode to anode, from one circuit to another, is unfavorable to drainage. The drainage effect of the specific cyclic and progressive electroosmosis (CPE) is not as good as the conventional electroosmosis method. However, combining the two electrification modes can further enhance the shear strength of the soil near a designated electrode. The generation of transverse cracks in soil is mainly due to the drainage difference caused by the electroosmotic flow, which is caused by the adjacent electric field at that location. When the electroosmotic flow is strong, longitudinal cracks will occur in the soil along the strongest electroosmotic flow path. In practical in situ engineering applications, the electrification mode of CPE will be more energy-efficient than conventional electrification modes with the same processing time.

INTRODUCTION

The electroosmosis method has been attempted on different types of civil engineering processes, such as soil improvement,^1,2 foundation pit dewatering,³ slope stability,⁴ and contaminated soil remediation,⁵ and so forth. However, the high energy consumption and the unsatisfactory drainage effect have made it difficult to promote the practical engineering applications of the electroosmosis method. Scholars have proposed many measures to improve electroosmosis methods by innovating in aspects such as electrification mode, electrodes’ material and configuration, additives, and combining with vacuum preloading or other methods. From the perspective of environmental protection and sustainable development, it is of great significance to further innovate the relevant processes of electroosmosis to break through the existing bottlenecks.

Electric fields applied across a clay soil mass result in electrolysis, transport of species by diffusion, ionic migration, electroosmosis, and electrophoresis.⁶ The transport and adsorption, precipitation, and dissolution reactions are the fundamental mechanisms affecting the electroosmotic reinforcement process. The open flow arrangement and electrolysis reactions at the electrodes significantly influenced the energy consumption and drainage effect. The losses in the applied voltage are found to be dependent on the anode material and are less in metallic anodes (steel or copper) than carbon anodes.⁷ However, metal electrodes are prone to corrosion. This, resultingly, affects their service life, while also posing a risk of introducing heavy metals into the soil. In recent years, electrokinetic geosynthetics (EKGs) with good drainage effect and conductive properties have been vigorously developed and successfully applied to soft clay reinforcement.^3,8,9

Electrification mode is one of the important factors affecting the energy consumption and drainage effect of the electroosmosis method. The electrification modes include constant voltage, constant current, electrode polarity reversal, intermittent current, stepped voltage, and cyclic and progressive electroosmosis method (CPE), and so forth. The constant voltage is often selected between 10 and 80 V and selecting an optimized constant voltage, by the spacing between the electrodes, had a significant effect on the efficiency of electrokinetic treatment. Increasing the voltage did not always lead to an increase in the efficiency of electrokinetic process.^10,11 Chang and Sheen¹² showed that under the step-by-step voltages with seven levels and double relation from 1.5 to 96 V, the soil volume and void ratio decrease along with the increased voltage, while the number of ions prompted by the positive and negative electrodes increased. Hamed and Bhadra¹³ applied different constant currents (10–50 mA) to soil specimens. The electroosmotic flow rate increased and the processing time decreased with the increasing constant current and resulted in a slight increase in the energy expenditure. However, due to the impermeable layer formed as a result of the obstruction created by the stiff cemented soil generated near the anodes, electrode polarity reversal did not produce favorable improvement effects in terms of soil shear strength and treatment time.¹⁴ Current intermittent operation strategies are an alternate method of power application for electroosmosis that has the potential to reduce power consumption and alleviate soil heat accumulation.^15,16

Another innovative electrification mode, CPE, has been proven to make soil reinforcement more uniform by causing the electroosmotic flow of soil between multiple electrodes to move in the same direction.¹⁷ The uniformity of soft soil after treatment is also an important indicator for evaluating the effectiveness of soil reinforcement, because soil spatial variability has significant implications in design and construction of civil engineering.^18–20 The reinforcement uniformity of soil can be evaluated using coefficient of variation.^21,22 Soil treated by electroosmosis usually showed uneven distribution of water content and shear strength between anode and cathode due to the unidirectionality of electroosmotic flow.^23,24 Previous studies have shown that there is a limited water content for the clay soil near the anode during electroosmosis, which will stop the migration of electroosmotic flow between the anode and cathode.^25,26 Salt solution injection,^27,28 electrodes arrangement optimization,²⁹ combining with vacuum preloading,^30,31 polarity reversal and other electrification modes¹⁵ were usually used to improve the reinforcement uniformity of electroosmotically treated soft clay. However, solution injection is not economical and environmentally friendly, and polarity reversal usually reduces drainage volume and increases energy consumption. Sun et al.¹⁷ combined CPE with additives in laboratory experiments. After treatment the coefficient of variation of water content and bearing capacity were relatively lower. However, the electrification mode of CPE that can achieve the convergence of electroosmotic flow is flexible and diverse when used in the rectangular, quincunx, and circular electrode arrangements. Hence, the reinforcement effect and electroosmotic flow migration mechanism of CPE under different basic electrification modes need to be further explored through theoretical and experimental methods to expand its application.

In this study, some specific CPE experiments with different circulating electrification basic modes were carried out to explore the impact of continuous changes in the electric field on the migration of ions and electrons and the migration of electroosmotic flow. Another purpose of the experiments was to determine, by comparison with conventional electroosmosis experiments and mechanism analysis, the optimal circulating electrification basic mode for CPE. The variation of current, electroosmotic flow, energy consumption, soil water content, and shear strength were considered in these tests. This study is of great significance to the full understanding of the mechanism of CPE reinforcement.

MATERIALS AND TEST SCHEMES Materials

An overview of the experimental setup is shown in Figure 1. The internal dimensions of an acrylic model box are 592 mm × 90 mm × 94 mm. Tubular EKGs were chosen as the electrodes. The inner and outer diameters are 17 and 27 mm, respectively. The plastic tubes are made of polyethylene, carbon black and graphite with good conductivity and corrosion resistance.¹⁷ Two strands of copper wires are buried along the length of the tubular body, which are used to connect the EKG with direct current (DC) power supply by rubber-insulated copper electric wires. Several drainage grooves run through the outer wall of the tubular EKG along the length direction and numerous drainage holes are drilled at other positions. Furthermore, the EKG, wrapped with a filter cloth, is used as the cathode as illustrated in Figure 1A. The filter cloth is 20 × 27 cm in size and is made from non-woven fabric material with mesh size of 100 × 100 μm. The DC power supply (RXN-605D) has a digital display feature and a steady output voltage with a maximum output power of 60 V × 5 A.

Figure 1. Experimental setup of electroosmosis: (A) side view and (B) cross sections represented as S1–S9 for monitoring points (unit mm).

For water drainage at the cathode, one end of a transparent polyurethane tube is inserted into the bottom of the cathode tube, the other end is connected with an Erlenmeyer flask and a water circulating multi-purpose vacuum pump (SHB-IIIA). The vacuum pump works intermittently for 30 s every time. The water is pumped out from the soil specimens and collected into the Erlenmeyer flask with 1000 mL volume. It is then poured into a measuring cylinder with a minimum scale of 1 mL to measure the volume. The power, maximum vacuum degree, and single tap air sucking amount of the vacuum pump are 180 W, 0.085 MPa, and 10 L/min, respectively.

The soil samples used in this study were prepared by artificially mixing clay powder with specified amounts of distilled water. The distilled water, produced by a deionized water equipment (DBW-SYS), had a pH range of 5.8–6.0 and electrical conductivity less than 6 μS/cm. The clay powder was obtained from a mining company in Jiangning District, Nanjing, China. To obtain basic geotechnical parameters of the soil powder, common geotechnical tests were performed basing on canonical standard ASTM, as shown in Table 1. The natural water content and plasticity index of the clay powder was 5.78% and 29, respectively. Sieving and hydrometer methods were carried out to analyze the soil particle size. The percentage of particle grain size less than 0.075 and 0.005 mm was 87% and 46%, respectively. According to ASTM D2487-06 classification, the soil samples are classified as clay of low plasticity symbolized by CL. The mineralogical characteristics of the clay powder were determined using X-ray diffraction (Bruker D8 Advance). Only quartz and kaolinite were found. The calculated weight of distilled water was determined on the basis of the natural water content and weight of the clay powder. The clay powder and distilled water were stirred uniformly to obtain soil samples with specified water content. The final actual water contents are determined by the oven-drying method.

Table 1 Properties of kaolin powder used in the experiments.

Soil	Specific gravity	Water content (%)	Liquid limit (%)	Plastic limit (%)	Chemical composition (%)
					SiO2	Al2O3	Fe2O3	K2O, Na2O	CaO, MgO	Others
Clay soil	2.74	5.8	50	21	51	24	1.6	2.8	1.3	19.3

Test schemes

For each test, 6.27 kg of clay soil powder with 5.8% water content was measured and mixed with 2.62 kg of distilled water by a mechanical mixer to reach 50% water content. The actual water content of the soil for each test is shown in Table 2. The weight of each sample in the test was 8.4 kg. To get rid of air bubbles, the soil sample was slowly layered and gently pressed in the test device. The height of the soil sample in each test before treatment was about 94 mm.

Table 2 Test schemes.

[Table omitted. See PDF]

The detailed test conditions are presented in Table 2. A conventional electroosmosis test (E1), three types of CPE tests (E2, E3, and E4), and a conventional electroosmosis combined with CPE test (E5) were carried out to explore the impact of continuous changes in the electric field on the migration of ions and electrons, as well as the migration of electroosmotic flow. The electrification basic mode of each test is shown in Figure 2. Each test has three electrodes, numbered 1, 2, and 3, respectively. Circuit 12 was formed when electrodes 1 and 2 connected with the positive pole and negative pole of the DC power supply, respectively. Similarly, in circuit 13, electrode 1 was used as the anode and electrode 3 was used as the cathode, and in circuit 23, electrode 2 was used as the anode and electrode 3 was used as the cathode. These three circuits were not powered on simultaneously. For the E1 test, electrodes 1 and 3 were used as anodes simultaneously and electrode 2 was used as the cathode (circuit 123). Electroosmotic flow always converged toward electrode 2, similar to the existing electrification mode.^3,32 For E2, circuit 12 was first powered on for an hour, then circuit 23 was powered on for an hour. The above steps were repeated multiple times until the drainage rate was less than 1 mL/h. For E3, circuit 13 was first powered on for an hour, then circuit 23 was powered on for an hour. The above steps were also repeated multiple times until the drainage rate was less than 1 mL/h. For E4, circuit 12 was first powered on for an hour, thereafter circuit 13 was powered on for an hour, and lastly circuit 23 was powered on for an hour. The above steps were repeated multiple times until the drainage rate was less than 1 mL/h. For E5, circuit 123 was powered on for the first 80 h, then, for the remaining time, circuit 13 and circuit 12 were alternated for 5 h each time.

Figure 2. Electrification basic modes: (A) E1, (B) E2, (C) E3, and (D) E4.

The design principle of CPE is that the direction of electroosmotic flow under each circuit is the same, so as to achieve the convergence of electroosmotic flow. This is the main difference between CPE and polarity reversal in the normal electroosmosis method. The basic electrification modes in the E2–E5 tests mentioned above always follow this principle, as shown in Figure 2B–D. This consideration is to realize long-distance transmission of electroosmotic flow in actual in situ engineering, thereby improving the uniformity of the soft clay treated by electroosmosis. In E2–E4, considering the feasibility of manual operation during long-term power on process, the electrification time of each circuit in one cycle was determined to be an hour. In E5, the conventional electroosmosis method in the first 80 h had already discharged a significant amount of water from the soil. Considering that the drainage, during CPE modes, will become relatively difficult, the electrification time of each circuit in one cycle was subsequently extended to 5 h. Through the discussion of these basic electrification modes, the electroosmotic strengthening effect of soft clay under different electrification modes is studied.

There is no doubt that there are many options for the basic cyclic electrification modes and electrification intervals based on the CPE principle. For example, the basic cyclic electrification mode, which is theoretically unfavorable for improving the uniformity of soil reinforcement, of circuit 12 combined with circuit 13 (always uses electrode 1 as the anode). These five experiments were selected representatively to carry out and analyze the test law. Although it cannot be generalized, it is still of certain reference significance as a typical representative.

Stepped voltage from 20 to 60 V was applied for each test. The water that converged at the cathode during the electroosmosis process was promptly extracted using the vacuum pump. During each test the variation of current, discharged water, settlement, and energy consumption were monitored. After treatment the soil water content and shear strength at different points were measured, as shown in Figure 1B. The shear strength tests, conforming to ASTM 2001, were conducted by dynamoelectric vane shear (TT-LVS, manufactured by Zhejiang Geotechnical Instrument). The vane shear apparatus has blades that are 25.4 mm in diameter, 25.4 mm in height, and 0.01 mm in thickness. The values of the shear stress and rotation angle were automatically recorded.

RESULTS Electric current

Each group of experiments adopted the working mode of 8–10 h in the day and intermittent current in the evening. The on–off time of each experiment was consistent. To facilitate the analysis, the current and drainage in Figures 3 and 4 did not count the intermittent time.

Figure 3. Current variation of each test: (A) E1 and E5, (B) E2, (C) E3, and (D) E4.

Figure 4. Water discharge. (A) Total water discharge of each test. (B) Water discharge of each circuit in E2, E3, and E4. (C) Drainage rate variation in different circuits.

Figure 3A shows variation of electric current of E1 and E5. Circuit 123 of E1 shows a trend of first increased current and then gradually decreased current during each continuous working period. In the first 72 h, under the condition of gradually increasing voltage from 20 to 60 V (increased by 10 V every 16 h), the current basically remained above 40 mA. After 72 h, the voltage remained unchanged at 60 V, but the current changed between 20 and 30 mA. Moreover, after each night's intermittent current, the next days’ current increased from a small value to a larger value. During the first 80 h of E5, only circuit 123 worked, which had the same electrification mode as E1. However, at voltages of 20, 30, and 50 V, the current of E5 was lower than E1, while at voltages of 40 and 60 V, the current of E5 exceeded E1. The main reason was that only electrode 2 of E1 was wrapped in filter cloth, while both electrode 2 and electrode 3 of E5 were wrapped in filter cloth. The filter cloth increased the interface resistance between the electrode and the soil, resulting in a decrease in current. When the electrification mode of E5 changed to CPE, the current in circuit 12 was higher than that in circuit 13, but the current in both circuits was lower than E1.

Circuit 12 of E2 had a relatively stable current change of 25.2–28.2 mA after the first hour of being powered on. Whereas, after a cycle, the current in circuit 12 tended to rise with a low starting current in each cycle. For instance, under 20 V voltage, the current of one cycle rose from 12.09 to 27.60 mA, and under 60 V voltage, the current of one cycle rose from 6.13 to 58.10 mA. However, in the case of circuit 23, the electrical current tended to dramatically decrease in each cycle. Moreover, when the voltage is stepped up to 40 V, the initial current of this circuit, in each cycle, exceeds the other experiments. However, the high current is difficult to maintain and, therefore, quickly decreases to a small value. For example, the current of one of the cycles decreased from 27.00 to 5.80 mA at 20 V, the current of one of the cycles decreased from 54.90 to 3.40 mA at 40 V, and the current of one of the cycles decreased from 69.10 to 2.20 mA at 60 V.

The electric current in circuit 13 and circuit 23 of E3 changed steadily during each cycle, as shown in Figure 3C. In the same cycle, the current in circuit 23 was almost more than twice the current in circuit 13. Therefore, the total water discharge of the cathode in circuit 23 is 1.8 times that of circuit 13. Comparing circuit 13 of E3 with circuit 123 of E1, for the same volume of soil at the same voltage (20–50 V), the current in circuit 13 of E3 is significantly lower than that in circuit 123 of E1. This is because the soil in circuit 13 of E3 is a series circuit, while the soil in circuit 123 of E1 is a parallel circuit, which reduces the electrical resistance.³³ Circuit 12 of E4 has a relatively stable current change of 22.2–24.7 mA after the first hour of being powered on at a voltage of 20 V. However, after one cycle, the current in circuit 12 started at a low current and continued to rise in each subsequent cycle. For example, the current in one of the cycles increased from 12.75 to 27.20 mA at 20 V, and the current in another cycle increased from 11.97 to 23.6 mA at 60 V. The current in circuit 13 changed steadily in each cycle, but the maximum current did not exceed 26 mA. Nevertheless, the current in circuit 23 showed a sharp downward trend in each cycle. For example, the current in one of the cycles dropped from 23.90 to 6.23 mA at 20 V, and dropped from 22.4 to 1.70 mA at 60 V in another cycle.

The results of E2 and E4 experiments show that the electrode polarity reversal in different circuits has a significant effect on the current of the CPE method, as shown in Figure 3B,D. When the electrode (electrode 2) reversed from cathode (in circuit 12) to anode (in circuit 23), the current in the present circuit (circuit 23) dramatically decreased from a high value. This situation was similar to that of the polarity reversal in conventional electroosmosis.^34,35 When the electrode (electrode 2) reversed from anode (in circuit 23) to cathode (in circuit 12), the current in the present circuit (circuit 12) increased rapidly from a low value. Electroosmotic treatment resulted in nonnegligible gradients variation of soil physicochemical parameters. The changes of pH values of the pore solution influenced surface electrical charge of the soil particles, and consequently, their electrokinetic potential.³⁶ An electrode reversed from anode to cathode, from one circuit to another, is beneficial to drainage, while an electrode reversed from cathode to anode, from one circuit to another, is unfavorable to drainage.

In a DC electric field, electrons flow from the negative pole of the DC power supply to the cathode, and flow from the anode to the positive pole of the DC power supply. While current flows from the positive pole of the DC power supply to the anode, then flows through the soil liquid phase to the cathode, and then flows from the cathode to the negative pole of the DC power supply. The conduction of current in soil is completed by the directional movement of ions in the liquid phase.³⁷ Under the action of an electric field, the cations and anions in the liquid phase migrate to the cathode and anode, respectively. According to Faraday's law, the quantity of electric charge flowing out of the cathode is equal to the quantity of electric charge flowing into the anode, and is also equal to the total quantity of electric charge (Q) flowing through any section of the circuit. The current in the electrode is completely transferred by electrons, but in soil it is completed by the combination of cations and anions in the liquid phase.³⁸ [Image Omitted. See PDF] [Image Omitted. See PDF]where $${Q}_{+}$$, $${Q}_{-}$$, $${I}_{+}$$, $${I}_{-}$$, and $${I}$$ represent the quantity of electric charge, current, and total current carried by cations and anions, respectively.

When the current passes through the EKG electrode, mainly electrolytic water reaction occurs. However, the electrolytic water reaction cannot consume the electrons supplied by the DC power supply to the cathode, because the rate of electrode reaction is limited. More electrons will accumulate on the cathode than the equilibrium state, which will cause the electrode potential of the cathode to decrease.³⁹ When circuit 12 was converted to circuit 23 in E2, electrode 2 was reversed from cathode to anode. The electrons accumulated on electrode 2, in excess of the equilibrium state, immediately flowed to the positive pole of the DC power supply, which, in a short time, caused a larger current in circuit 23. However, the electrolytic water reaction near electrode 2 cannot provide excessive electrons to the DC power supply, resulting in a sharp decrease in current. In contrast, when circuit 23 was converted to circuit 12, the initial current in the circuit was lower but gradually increased to a higher value. This indicates that when electrode 2 reversed from anode (in circuit 23) to cathode (in circuit 12), the electrons on this electrode could be replenished in a timely manner.

Water discharge

As shown in Figure 4, the total water discharge of E1–E5 is 949, 550, 719, 483, and 786 mL, respectively. The water discharge from the cathode in circuit 12 and circuit 23 of E2 is 326 and 224 mL, respectively, while that of E3, in circuit 13 and circuit 23, is 255 and 464 mL, respectively. Furthermore, in circuit 12, circuit 13, and circuit 23 of E4, the water discharge from the cathode is 118, 198, and 168 mL, respectively. The water discharge of electrode 2 of E1 (circuit 123) is about three times that of electrode 2 of E2 (circuit 12) and eight times that of electrode 2 of E4 (circuit 12). Under three different electrification methods, the water discharge from the cathode in circuit 23 of E2, E3, and E4 is 224, 464, and 168 mL, respectively. In this circuit the water discharge of E3 is about twice that of E2 and three times that of E4.

It can be seen that different electrification modes had a significant impact on the water discharge. This was mainly because there were significant differences in the soil volume, power on/off time, and potential gradient between the electrodes of each circuit under different electrification modes.¹⁵

The five experiment scenarios include four types of circuits: circuit 12, circuit 23, circuit 13, and circuit 123. The soil mass between the electrodes of circuit 13 was the same as that of circuit 123, while the soil mass between the electrodes of circuit 12 was the same as that of circuit 23, with the former being twice as large as the latter. The parallel connection method of the soil between the electrodes of circuit 123 ensured that the soil resistance was lower than that of the other circuits. The power on/off time of the four circuits in each test was different. For example, the power on time of circuit 123 of E1 was twice that of circuit 12 of E2 and three times that of circuit 12 of E4. Moreover, the difference in potential gradient, caused by electrode polarity reversal, affected the current and electric potential in the soil of each circuit. These were the driving forces for the migration of electroosmotic flow in the soil.^10,36,40

During the energization process of circuit 12 of E2, the electrode potential of the anode increased, whereas, due to concentration polarization and electrochemical polarization, the electrode potential of the cathode decreased.³⁹ When circuit 12 was converted to circuit 23, electrode 2 was reversed from cathode to anode. Due to electrode polarization, the potential difference between electrode 2 and electrode 3 was lower than the potential difference between electrodes in circuit 12 under the same applied voltage. This led to the differences in current and water discharge of circuit 12 and circuit 23. When circuit 12 was converted to circuit 13 in E4, the electrode potential of electrode 1 increased in both circuits, while, due to electrode polarization, the electrode potentials of electrode 2 and electrode 3 decreased. After being converted to circuit 23, the potential difference between electrode 2 and electrode 3 was higher than the electrode potential difference directly converted from circuit 12 to circuit 23 in E2. However, because the power on time of circuit 23 in E4 was only 2/3 that of circuit 23 in E2, the water discharge of the former was less than the latter. Circuit 13 and circuit 23 in E3 never changed the polarity of the electrodes, and their total water discharge was greater than E2 and E4. According to the water discharge of each circuit of the E2, E3, and E4 tests, the following relationship can be obtained: [Image Omitted. See PDF] [Image Omitted. See PDF] [Image Omitted. See PDF]where $${\unicode{x02206}U}_{12}$$, $${\unicode{x02206}U}_{23}$$, and $${\unicode{x02206}U}_{13}$$ is the electrode potential difference between circuit 12, circuit 23, and circuit 13. $${l}_{12}$$, $${l}_{23}$$, and $${l}_{13}$$ are the distance between electrodes 1 and 2, electrodes 2 and 3, and electrodes 1 and 3.

In the initial 30 h, E1 exhibited the highest water discharge rate, while circuit 23 of E3 basically outperformed others from 30 to 150 h. At 70 h, the water discharge rate of circuit 12 of E4 dropped below 1 mL/h first, indicating that the circuit can be terminated to save energy. Therefore, the electrification mode should not be consistent, but should be continuously optimized based on various factors.

Water content

After the experiment, the distribution of soil moisture content was tested based on the layout diagram of monitoring points in Figure 1B. Due to the varying degrees of crack propagation in each test during the electroosmotic process, some points deviated from the original test section (S1–S9). The average moisture content of the same cross-section was calculated, and the distribution results of each test were depicted in Figure 5A. From electrode 1 to electrode 2, the average moisture content of each section of soil showed an increasing trend. At sections S1–S5, E5 almost had the lowest moisture content and E3 almost had the highest moisture content. From electrode 2 to electrode 3, there were differences in the variation patterns of the average moisture content of the soil section. E1 showed a continuous decreasing trend, while E3 first showed a decreasing and then increasing trend. At sections S6 and S7, E3 had the lowest moisture content. The moisture content distribution of E2 and E4, with a range of 40%–47%, was similar between electrode 2 to electrode 3, showing a slight upward trend.

Figure 5. (A) Water content distribution. (B) Dividing soil based on the distribution of moisture content.

According to the distribution characteristics of moisture content, the soil sample was divided into nine parts along the length direction, and the length of each part of the soil was recorded as $${L}_{i}$$. The total length of the soil sample was $$L=592$$ mm, as shown in Figure 5B. Due to the initial total mass ($$m$$), water mass ($${m}_{{w}}$$), dry soil mass ($${m}_{{s}}$$), and initial moisture content ($$w$$) of the soil samples prepared for each group of experiments, as well as the average moisture content ($${w}_{i}$$) of each part of the soil, after the experiment, is known. Therefore, the mass of residual water in each part of the soil after treatment ($${m}_{{w}_{{i}}}$$) can be calculated according to the following equation, and then the theoretical water loss of the soil between electrode 1 and electrode 2 ($${Q}_{12}$$), and between electrode 2 and electrode 3 ($${Q}_{23}$$) before and after electroosmotic treatment can be calculated: [Image Omitted. See PDF] [Image Omitted. See PDF] [Image Omitted. See PDF]

The calculation results are shown in Table 3. Due to the fact that the soil samples in each group of experiments were divided into nine parts, each part used the average moisture content with a limited number of test points to represent the moisture content of the soil in that part, which may not sufficiently precise. However, the calculated total water discharge of E1–E5 was higher than the actual collected total water discharge by 24%, 49%, 22%, 26%, and 13%, respectively. The main reason is that part of the water was consumed by electrode electrolysis of water and water evaporation caused by rising soil temperature.^34,37 The electrification mode of E2 included circuit 12 and circuit 23, and the water discharge from the two circuits can be approximately assumed to come from the soil between electrode 1 and electrode 2 and the soil between electrode 2 and electrode 3, respectively. It can be seen that the calculated water discharge of the soil between electrode 1 and electrode 2 differed significantly from the actual water discharge, which was due to the high current of circuit 12 and the high consumption of water by electrode electrolysis and evaporation. After comparing the $${Q}_{12}$$ and $${Q}_{23}$$ of each test, it can be seen $${Q}_{12}$$ and $${Q}_{23}$$ of E1 and E3 were similar, while in E2, E4, and E5 the $${Q}_{12}$$ was more than three times higher than $${Q}_{23}$$. Therefore, the following judgment can be made: In E4, almost all of the water discharge in circuit 12 and circuit 13, and some of the water discharge in circuit 23, came from the soil between electrode 1 and electrode 2. The effect obtained under the conventional electroosmotic method of E5 can be broken by the cyclic progressive method, leading to a redistribution of moisture content and shear strength.

Table 3 Calculated water discharge based on soil water content distribution.

Test number	E1	E2	E3	E4	E5
Water mass in the soil before treatment on one side of electrode 2 (mL)	1405	1409	1380	1396	1400
Calculated water mass in the soil after treatment between electrode 1 and 2 (mL)	806	791	965	935	705
Calculated water discharge $${Q}_{12}$$ of soil between electrode 1 and 2 (mL)	598	618	415	460	695
Actual water discharge of soil between electrode 1and 2 (mL)	—	326	—	—	—
Calculated water mass in the soil after treatment between electrode 2 and 3 (mL)	825	1209	920	1246	1203
Calculated water discharge $${Q}_{23}$$ of soil between electrode 2 and 3 (mL)	579	200	460	150	197
Actual water discharge of soil between electrode 2 and 3 (mL)	—	224	—	—	—
Calculated water discharge of the total soil (mL)	1177	818	875	610	892
Actual water discharge of the total soil (mL)	949	550	719	483	786

Shear strength

After each test, the distribution of shear strength was tested based on the layout diagram of monitoring points in Figure 1B. Due to the varying degrees of crack propagation in each test during the electroosmotic process, some points deviated from the original test section (S1–S9). The average value shear strength at the same cross-section was calculated, and the distribution of shear strength in each test is shown in Figure 6. Soil moisture content has a significant impact on the shear strength. At the same position from a certain electrode, the higher the moisture content, the lower the shear strength, and the lower the moisture content, the higher the shear strength. But when the moisture content drops below 20%, smaller changes in moisture content will result in significant differences in shear strength. For example, the moisture content of E1 and E5 at section S1 was 21% and 16%, respectively, while the shear strength was 223 and 429 kPa, respectively. This is mainly because the chemical cementation near the anode is more conducive to improving the shear strength of the soil under low moisture conditions.^11,41

Figure 6. Shear strength distribution.

Electrification modes had a significant impact on the shear strength of soil. From electrode 1 to electrode 2, the shear strength of the soil of E1–E4 showed a trend of first increase and then decrease, while E5 showed an almost linear decrease. The maximum shear strength was 429 kPa near electrode 1 (S1) in E5. The minimum shear strength of each test, lower than 15 kPa, was near electrode 2 (S5). The shear strength of E1 at sections S1 and S2 was much higher than that of E2, E3, and E4 (E2 > E4 > E3), while E5 was higher than E1, except for sections S2 and S5, where it was close to E1. The above results indicate that combining the conventional with cyclic and progressive electrification mode can further enhance the shear strength of the soil near a designated electrode. This has good application prospects in electroosmotic strengthening of pile foundation bearing capacity engineering.^42–44 From electrode 2 to electrode 3, the soil shear strength of E2, E4, and E5 was basically lower than 10 kPa. Sections S7–S9 of E1 and sections S6–S7 of E3 had higher shear strength, while the other sections had lower shear strength. In these tests the shear strength of the soil between electrode 2 and electrode 3 was lower than between electrode 1 and electrode 2 (except E3).

Although the electrification mode of E1 always had electrodes 1 and 3 as anodes and electrode 2 as a cathode, the soil shear strength between electrode 1 and electrode 2 was significantly higher than between electrode 2 and electrode 3. In particular, the shear strength at sections S1 and S2 was significantly higher than at sections S8 and S9. During the experimental installation process, the uniformity of soil sample laying, slight differences in electrodes spacing, and the connection position of wire clamps and electrodes can all cause current differences between electrodes 1 and 2 and electrodes 3 and 2. In the initial few days of E1, the current of the soil between electrodes 3 and 2 was greater than between electrodes 1 and 2. Cracks first began to develop in the middle of the soil between electrode 2 and electrode 3, and near electrode 3. Subsequently, the width and depth of these cracks were always greater than those near electrode 1 and between electrode 1 and electrode 2. The development of cracks blocked the migration path of electroosmotic flow in the soil, which affected the water discharge as well as the distribution of moisture content and shear strength.⁴⁵

Surface settlement

The surface settlement test data for E1 and E2 were missing; therefore, only the average surface settlement of soil between electrode 1 and electrode 2, and between electrode 2 and electrode 3 for E3, E4, and E5 were provided, as shown in Figure 7. The surface settlement of soft soil strengthened by electroosmosis was mainly related to the water discharge.⁴⁶ The more water discharge corresponds to greater settlement. In the first 60 h of E3, the surface settlement rate of the soil between electrode 2 and electrode 3 was higher than between electrode 1 and electrode 2. Afterwards, the rate of the former slowed down and the latter remained almost unchanged. Finally, the surface settlement results of the two were close. The water discharge of the soil on both sides of electrode 2 calculated in Table 3 was close; thus, their settlements were also close. The final average surface settlement of the soil between electrode 1 and electrode 2 in E4, as well as between electrode 2 and electrode 3, exhibited significant differences of 8.0 and 4.6 mm. This is because the corresponding calculated water discharge of the soils between the electrodes was 460 and 150 mL, respectively. The difference of final average surface settlement of the soils between electrode 1 and electrode 2 and between electrode 2 and electrode 3 in E5 is the largest. The difference is 11.6 and 3.9 mm, respectively. The corresponding calculated water discharge of the soil between electrodes is 695 and 197 mL. Comparing the calculated water discharge and surface settlement of the soils between electrode 1 and electrode 2 and between electrode 2 and electrode 3 in E3, E4, and E5, it can be found that the larger the calculated water discharge, the greater the surface settlement. If the calculated water discharge is close, the surface settlement is close.

Figure 7. Surface settlement.

Energy consumption

The energy consumption of each test has little difference, as shown in Figure 8. This is because the current in each circuit was quite small in the laboratory model test size, and the maximum current did not exceed 80 mA. The energy consumption of each test almost varied linearly with time. The longer the power on time, the larger the energy consumption. However, there was a significant difference in water discharge when consuming the same amount of energy for each test. When recording the water discharge per unit of energy consumption as energy efficiency ($$\eta $$), obviously $${\eta }_{E1}\gt {\eta }_{E5}\gt {\eta }_{E3}\gt {\eta }_{E2}\gt {\eta }_{E4}$$. In practical in situ engineering applications, electrodes were usually inserted into soft clay for several meters, and dozens or hundreds of electrodes would be connected in parallel. At this time, there would be significant differences in energy consumption under different electrification modes. Obviously, the electrification mode of CPE will be more energy efficient than conventional electrification mode with the same processing time.

Figure 8. Energy consumption.

Influence of electric field on cracks

As shown in Figure 9, five tests showed different cracks under the same size effect. The reasons for crack propagation in soil are multifaceted under different electrification modes. First, under the action of electric field forces the directional migration of anions and cations in the soft soil caused electroosmotic flow migration toward the cathode. Soil shrinks when drained.⁴¹ Moreover, the moisture content test results showed that the moisture content of the surface soil was significantly lower than that of the middle and bottom. This resulted in uneven shrinkage with large surface shrinkage and small internal shrinkage. Second, temperature stress occurs when soil temperature changes. Heat is generated in any electrokinetic process. Generally, a higher electric current in the soil corresponds to higher temperature, and vice versa.^35,47 As demonstrated in Figure 3, the electric current of each test varies greatly under different voltages and circuits. From this, it is inferred that it will cause a change in temperature. In particular, there is a large temperature difference between the soil inside and surface soil, and the surface soil is subjected to tensile stress. Finally, electrolytic water reaction occurred at the inert cathode and anode to produce gas. With the continuous generation of gas, if the escape of it is hindered, it will lead to an increase in gas pressure. Due to the increasing gas pressure at the soil–electrode interface, cracks are formed in the vicinity of the electrodes.⁴⁵

Figure 9. Crack development of soil after treatment: (a) E1, (b) E2, (c) E3, (d) E4, and (e) E5.

A crack image processing software (PCAS) was used for the quantitative measurement of the development of cracks in each test. The crack rate was obtained, which is defined as the ratio of the crack area to the total area of soil sample. The crack rate of soil samples in E1–E5 are 11.7%, 3.4%, 4.1%, 4.6%, and 7.4%, respectively. The crack rate of conventional electroosmotic methods is higher than that of CPE, and that of E2, with the basic cyclic electrification mode of circuit 12 combined with circuit 23, is the lowest.

Cracks include longitudinal cracks (in the direction of electrode spacing) and transverse cracks (perpendicular to the direction of electrode spacing). E1 exhibited divergent transverse cracks at all three electrodes, and a transverse crack appeared in the middle of the soil between electrode 1 and electrode 2 and between electrode 2 and electrode 3. E2 only had transverse cracks, and the crack at electrode 2 had the widest width. E3 not only had transverse cracks, but also longitudinal cracks. Transverse cracks mainly occurred near electrode 2 and electrode 3, while longitudinal cracks occurred in the soil between electrode 2 and electrode 3. The transverse cracks of E4 appeared at the three electrodes and in the middle of the soil between electrode 1 and electrode 2, while the longitudinal cracks appeared between the transverse cracks of the soil and electrode 2. The crack propagation position of E5 in the first 80 h (circuit 123) was basically consistent. However, by the end of the test, the cracks were longer and wider.

The reason for the different development of cracks is mainly related to the different electric fields applied in each test. Under the impact of electric field forces, directed migration of the anions and cations in the soil was generated, resulting in electroosmotic flow towards the cathode. For E2, the current and drainage were not particularly good compared to other tests (especially circuit 23). The cracks mainly developed in the soil near electrode 1 and electrode 2. The current and drainage of circuit 23 in E3 were comparatively high, resultingly, forming longitudinal cracks along the direction of electroosmotic flow. Moreover, when electrode 2 was used as an anode in circuit 23, drainage was also generated in partial soil between electrode 1 and electrode 2, resulting in transverse cracks due to soil drainage shrinkage. It can be seen from E1 that the electric field overlapped due to the close electrode spacing. The overlap length, which caused the waste of energy consumption, is the distance between the crack and electrode 2. Therefore, it is recommended that the reasonable electrode spacing for conventional electroosmosis in this model experiment is about 40 cm.

From the above analysis, it can be concluded that the cracks in the soil near the anode are mainly caused by electrolytic water reaction, temperature changes, and drainage, while the cracks near the cathode are mainly caused by electrolytic water reaction. The generation of transverse cracks in soil is mainly due to the drainage difference caused by the electroosmotic flow, which is caused by the adjacent electric field at that location. When the electroosmotic flow is strong, longitudinal cracks will occur in the soil along the strongest electroosmotic flow path.

CONCLUSIONS

Several laboratory tests were conducted using different electrification modes to provide deep insights into the influence of different electric fields on the electroosmotic treatment. They were conducted by analyzing the variation of electric current, water discharge, water content, shear strength, surface settlement, energy consumption, and crack development.

The conventional electrification mode, with two parallel circuits working simultaneously, can obtain a larger electric current, and thus have the best water discharge. For CPE, four different circulating electrification basic modes were conducted to compare to the conventional electrification mode. For each test there were significant differences in the soil volume, power on/off time, and potential gradient between the electrodes of each circuit. The water discharge of E3 (circuit 13 and circuit 23) was the best of the CPE tests, but was 24.2% lower than conventional electroosmosis. The water discharge of other CPE tests was even more unsatisfactory, because of the polarity reversal of the electrodes in different circuits. When the electrode reversed from anode to cathode, from one circuit to another, the current in the circuit increased rapidly from a low value, which is beneficial to drainage. When the electrode reversed from cathode to anode, from one circuit to another, the current in the circuit dramatically decreased from a high value, which is unfavorable to drainage. In addition, the CPE electrification modes of constantly changing the cathode without changing the anode is more conducive to improving the shear strength of the soil near the anode. Therefore, the combination of conventional and CPE electrification modes can further improve the shear strength of the soil near the specified electrode. That is achieved by making the specified electrode act as the anode under both electrification modes. This has good application prospects in electroosmotic strengthening of pile foundation bearing capacity in engineering.

The crack development is different under different electric fields. The cracks in the soil near the anode are mainly caused by electrolytic water reaction, temperature changes, and drainage, while the cracks near the cathode are mainly caused by electrolytic water reaction. The generation of transverse cracks in soil is mainly due to the drainage difference, which is caused by the electroosmotic flow caused by the adjacent electric field at that location. When the electroosmotic flow is strong, longitudinal cracks will occur in the soil along the strongest electroosmotic flow path. In practical in situ engineering applications, the electrification mode of CPE will be more energy-efficient than a conventional electrification mode with the same processing time. Therefore, the electrification mode should not be consistent, but should be continuously optimized based on various factors.

AUTHOR CONTRIBUTIONS

All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Zhaohua Sun, Shuwen Xu, Cheng Zhang, Jingxian Geng, and Yefeng Gu. The first draft of the manuscript was written by Zhaohua Sun and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

ACKNOWLEDGMENTS

The authors would appreciate Beukes Demarscho Eugene from the Nantong University, China, for linguistic assistance during the preparation of this manuscript. This work was supported by the National Natural Science Foundation of China (42207189), Key Laboratory of New Technology for Construction of Cities in Mountain Area (LNTCCMA-20230108), and Postgraduate Research & Practice Innovation Program of Jiangsu Province (SJCX24_2073).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.