Study on the Influence of Mud Properties on the Stability of Excavated Face of Slurry Shield and the Quality of Filter Cake Form

Journal of Marine Science and Engineering

Article Study on the Inﬂuence of Mud Properties on the Stability of Excavated Face of Slurry Shield and the Quality of Filter Cake Formation

Hu Wang 1, Jian Chen 1,2, Hongjun Liu 1, Lei Guo 3,*, Yao Lu 1, Xiuting Su 1, Yongmao Zhu 1 and Tao Liu 1,4,5,*

1 College of Environmental Science and Engineering, Ocean University of China, Qingdao 266100, China; [email protected] (H.W.); [email protected] (J.C.); [email protected] (H.L.); [email protected] (Y.L.); [email protected] (X.S.); [email protected] (Y.Z.) 2 China Railway 14 Bureau Group Co., LTD., Jinan 250101, China 3 Institute of Marine Science and Technology, Shandong University, Qingdao 266000, China 4 Key Laboratory of Shandong Province for Marine Environment and Geological Engineering, Ocean University of China, Qingdao 266100, China 5 Laboratory for Marine Geology, Qingdao National Laboratory, Qingdao 266000, Shandong, China * Correspondence: [email protected] (L.G.); [email protected] (T.L.); Tel.: +86-0532-6678-1020 (T.L.)

Received: 26 March 2020; Accepted: 15 April 2020; Published: 18 April 2020

Abstract: In the construction of subsea tunnels, the stability and control of the excavation surface are the main concerns of the engineering community. In this paper, the Xiamen Metro Line 2 is used as the study case. The filter cake formation of mud shields is experimentally studied, and the excavation surface is numerically simulated. It is found that the formation of filter cake does not require a large pressure difference, and can be formed under 0.06 MPa. With the increase of pressure, the quality of filter cake is further improved, and a small amount of seawater (volume ratio less than 3%) also has a significant effect on the viscosity of mud. Under different cross-section geological conditions, with the decrease of the support pressure of the excavation face, the vertical displacement and vertical (Y-direction) displacement of the excavation face dome gradually increase, the maximum longitudinal displacement is 9.7 mm, the maximum longitudinal displacement can reach 23.9 mm, and the growth trend is nonlinear. According to different stratum conditions, during the excavation of the tunnel, the plastic area of the excavation face is different.

Keywords: shield; ﬁlter cake; excavation surface; stability

1. Introduction China has become one of the countries with the largest application of shield technology in the world [1]. In the past 10 years, the number of tunnels constructed by the shield method has increased significantly, especially in the construction of urban subways, river crossings and undersea tunnels [2], such as the Nanjing Yangtze River Tunnel, Qingdao Jiaozhou Bay Subsea Tunnel, Xiamen Rail Transit Line 2 and Cross-sea Section of Line 3 and other projects [3,4]. Line 2 from Xiamen Haicang Avenue Station to Dongdu Road Station is China’s first cross-sea subsea tunnel constructed with slurry shield machine. The pumps send the slurry prepared on the ground to the pressure chamber. A dynamic filter cake is formed on the excavation surface when the mud is seeping into the formation. The mud pressure of the pressure chamber is balanced by the filter cake and the earth pressure and water pressure on the excavation surface to maintain the stability of the excavation surface. During the construction of long-distance shield tunnel, it is necessary to open warehouses, repair and replace worn cutter discs and knives [5]. Under the adverse conditions of underground seepage field and

J. Mar. Sci. Eng. 2020, 8, 291; doi:10.3390/jmse8040291 www.mdpi.com/journal/jmse J. Mar. Sci. Eng. 2020, 8, 291 2 of 17 seawater intrusion, the excavation surface often collapses, water gushing accident, mud deterioration and other accidents, when it is bad, it will cause personal safety and even the entire tunnel to be scrapped [6]. Therefore, in the construction of subsea shield tunnels, the stability and control of the excavation face of the shield tunnel into the warehouse has become a major topic of concern in the engineering community. Some examples of forced shutdown and opening of shield tunnels have been reported in the world. Martin and Bapple reported an example of the opening of the fourth tunnel of the Elbe in Germany at the bottom of the river. Heijboer [7] introduced the example of the active opening of the Westerschelde tunnel in the Netherlands [8]. Although there are many examples of successful cabin opening in China and abroad [9–12], most of them are organized by engineers and technicians based on experience, and they are biased towards the construction process, and there is still a lack of theoretical research on the stability guarantee of the excavation face during cabin opening. The invention of the shield machine started with maintaining the stability of the excavation face. The research on the theory and method of evaluating the stability of excavation face is continuously deepened. The current research methods mainly include theoretical analysis and research, indoor physical test and computer numerical simulation. The theories used to analyze the active instability of excavation faces include limit analysis method and limit equilibrium method. The limit analysis method includes the upper limit analysis method and the lower limit analysis method. Davis et al. gave the upper limit solution of four failure modes under the assumption of the failure surface [13]. Leca used energy dissipation theory to give upper bound solutions for three failure modes [14]. These two analysis methods have the same point in that they both assume a failure mode, that is, the form of the failure surface. Lee et al. applied this upper limit solution to the analysis of the seepage force of the excavated face [15]. Davis gave the elastic lower bound solution of the supporting force without considering gravity. Leca applied this solution to friction-type materials. In addition, many scholars have verified and optimized the classic theory of active instability of shield excavation face through a large number of scale model tests and centrifugal model tests. Takano et al. introduced CT technology to a shield excavation face model test [16]; Kirsch first applied PIV technology to the excavation face stability model test [17]. Mair used a three-dimensional centrifugal model test to study the stability of the excavation face of a shield tunnel in clay [18]. With the continuous development of computer technology, many scholars have used finite element or boundary element technology to establish two-dimensional or three-dimensional models that simulate tunnel excavation, and have achieved considerable results. Zhu et al. established a two-dimensional finite element calculation model. Materials and segments are simulated to provide a method basis for numerical simulation of shield construction [19]. Ding et al. established a two-dimensional finite element model to analyze the influence of in-situ stress on the strength and deformation of the surrounding rock of the tunnel. [20], Yang et al. established a finite difference model to simulate tunnel digging, and provided ideas for the numerical simulation of the stability of the excavation face [21]. In addition, Zhang et al. [22], Jiang et al. [23], Li et al. [24], Zhang et al. [25] and others established three-dimensional numerical calculation models to simulate the excavation of shield tunnels, and studied the ground deformation caused by shield construction. These studies provide ideas for numerical simulation of excavation stability. During the excavation process of the slurry shield, the pressurized mud forms an air-tight filter cake on the excavation face [26], which converts the mud pressure to resist water and earth pressure in front of the excavation face, ensuring the stability of the excavation face. This is especially the case when entering warehouses [27]. The conversion of mud pressure is related to many factors, such as mud properties, mud pressure, formation characteristics, and the form and quality of filter cake formation, etc. [28–30] However, it is difficult to observe the formation process of filter cake in actual engineering. Therefore, it is possible to simulate the permeation process of pressurized mud in the formation by means of laboratory tests, analyze whether the filter cake can form and form regularity, and change the test conditions to study the influence factors of filter cake formation. Watanabe et al. studied the effect of mud density on mud loss in J. Mar. Sci. Eng. 2020, 8, 291 3 of 17 experiments on the formation of mud films in high-permeability formations and found that the sand content of mud plays an important role in reducing mud loss [31]. Min et al. conducted filter cake 4 formation tests of different muds in formations with a permeability coefficient of 10− m/s, and it was J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 3 of 18 clear that the clay content in the mud has an important effect on the filter cake quality [32]. Thefilter above cake formation studies are tests based of different on ordinary muds slurryin formations shields with as a a research permeability object, coefficient and the of lack 10−4 of m/s, studies on theand excavation it was clear face that of the subsea clay content tunnel in shieldsthe mud is has insu anffi importantcient to fullyeffect explainon the filter the cake mechanism quality [32]. of filter cake formation.The above In thisstudies paper, are thebased Xiamen on ordinary Metro slurry Line 2shields is used as asa research the study object, case, and and the the lack filter of cake formationstudies of on the the slurry excavation shield face is experimentally of subsea tunnel studied, shields andis insufficient the numerical to fully simulation explain the of mechanism the excavation face isof used filter tocake further formation. supplement In this paper, and the enrich Xiamen the Metro stability Line of 2 theis used excavation as the study face case, and and the the law filter of filter cake formation.cake formation of the slurry shield is experimentally studied, and the numerical simulation of the excavation face is used to further supplement and enrich the stability of the excavation face and the 2. Overviewlaw of filter of Shieldcake formation. Project

The2. Overview Xiamen of Metro Shield sea-crossing Project tunnel crosses the sea for 2760 m. During construction, it passes through highlyThe Xiamen permeable Metro strata sea-crossing such as tunnel sand crosses layers. the Among sea for them, 2760 m. the During vertical construction, permeability it passes coeffi cient in the residual sandy clay soil layer is up to 6.09 10 4 m/s. The vertical permeability coefficient in through highly permeable strata such as sand× layers.− Among them, the vertical permeability metamorphiccoefficient sandstone in the residual is up sandy to 3.58 clay 10soil4 layerm/s, is and up the to 6.09 vertical × 10− permeability4 m/s. The vertical coeffi permeabilitycient in granite × − formationscoefficient is up in tometamorphic 0.73 10 4samndstone/s. is up to 3.58 × 10−4 m/s, and the vertical permeability coefficient × − Thein granite area formations of construction is up to works 0.73 × 10 is−4 shown m/s. in Figure1. The shield section adopts two sets of ф6700 mmThe composite area of construction slurry shields. works The is shown overall in lineFigure is a1. “V”,The shield first descending section adopts to two the lowestsets of ф point6700 and then goingmm composite up. The slurry maximum shields. slope The overall of downhill line is a is“V 2.8%,”, first and descending the maximum to the lowest slope point of uphill and then is 2.9%. going up. The maximum slope of downhill is 2.8%, and the maximum slope of uphill is 2.9%. The The thickness of the covering soil ranges from 8.7 m~65.7 m, the maximum distance from the tide level thickness of the covering soil ranges from 8.7 m~65.7 m, the maximum distance from the tide level to to thethe tunnel tunnel is is about about 55 55 m,m, andand the the maximum maximum soil soil and and water water pressure pressure is about is about 6 bar. 6 bar.

Figure 1. Area of construction works. Figure 1. Area of construction works.

J. Mar. Sci. Eng. 2020, 8, 291 4 of 17

J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 4 of 18 3. Experimental Study on Filter Cake Formation Mechanism 3. Experimental Study on Filter Cake Formation Mechanism 3.1. Test Materials and Procedures The3.1. Test density, Materials viscosity, and Procedures clay content, and particle size of the slurry have important effects on filter cake qualityThe [ density,33,34]. Theviscosity, infiltration clay content, process and of particle the pressurized size of the slurry mud have in the important formation effects was on simulated filter by meanscake quality of laboratory [33,34]. The tests, infiltration the formation process of of the the filterpressurized cake were mud analyzed,in the formation and thewas test simulated conditions wereby changed means toof studylaboratory the influencingtests, the formation factors of of the filter filter cake cake formation. were analyzed, According and the to test the conditions stratum with high permeabilitywere changed to of study Xiamen the Metroinfluencing Line factors 2, diff erentof filter muds cake wereformation. prepared According to study to the the stratum morphological with high permeability of Xiamen Metro Line 2, different muds were prepared to study the morphological characteristics and formation rules of filter cake formation. characteristics and formation rules of filter cake formation.

Figure 2. Filter cake air Penetration test system. (a) drainage apparatus and valve, (b) different Figure 2. Filter cake air Penetration test system. (a) drainage apparatus and valve, (b) different measuring instruments adopted for the test. measuring instruments adopted for the test. The testing system is mainly divided into drainage apparatus, mud infiltration devices and The testing system is mainly divided into drainage apparatus, mud infiltration devices and measuringmeasuring instruments. instruments. The The filter filter cake cake air air Penetration Penetration test system system is is shown shown in inFigure Figure 2.The2. The pressure pressure devicedevice can becan usedbe used to applyto apply pressure pressure to to the the mud mud andand formation pore pore water water pressure. pressure. The Themud mud infiltrationinfiltration device device enables enables the the pressure pressure mud mud to to form form a filtera filter cake cake in in the the simulated simulated formation.formation. There There are threeare specifications three specifications for di ffforerent different diameters, diameters, namely namely 15 15 cm, cm, 30 30cm cm and 50 50 cm. cm. The The data data of filtration of filtration volumevolume and formationand formation pore pore water water pressure pressure are are tested tested byby measuring instruments. instruments. In theIn experiment, the experiment, ordinary ordinary quartz quartz sand aftersandscreening after screening is used is to used prepare to soilprepare sample. soil Consideringsample. the effConsideringect of different the effect permeability of different formations permeability on form theations formation on the of formation filter cake, of filter three cake, types three of formationtypes of formation were prepared in the test. The formation parameters are shown in the Table 1. The were prepared in the test. The formation parameters are shown in the Table1. The formation formation permeability was measured using constant head permeability test. The constant head permeability was measured using constant head permeability test. The constant head permeability test permeability test is to use the constant head permeability device to measure the seepage flow and the is to usehead the height constant of different head permeabilitypoints, so as to devicecalculat toe the measure seepage the velocity seepage and flowhydraulic and thegradient, head so height as of differentto calculate points, the so permeability as to calculate coefficient. the seepage velocity and hydraulic gradient, so as to calculate the permeability coefficient. Mud preparationTable 1. materials Mass percentage include of G1, water, G2 and granular G3 stratum materials, in different and particle additives. size ranges. Granular materials are mostly clay, bentonite, stoneStratum powder, silt, and fine sand, and G1 additives(%) G2 are (%) mostly G3 chemical (%) reagents. See Table2 for formation test mud parameters.1.6 mm~2.36 mm - - 35–45 During the test, referring to the engineering1.43 mm~1.6 situation mm of -Xiamen Metro - Line 15–20 2, the maximum loading pressure was set to 0.5 MPa, and1.18 adjustments mm~1.43 mm were made - according - to 15–25 the test conditions 1 mm~1.18 mm - - 10–15 during the test.Mass In percentage this process, in different a reasonable pressure loading step needs to be set, and the pressure 0.8 mm~1 mm - 15–25 - of each stage is setparticle to 0.02 size MPa ranges in the initial loading stage, and the loading pressure of each stage is 0.6 mm~0.8 mm - 40–50 - adjusted according to the change of the filtered water amount in the later stage. The loading time of 0.5 mm~0.6 mm - 30–35 - each stage of the mud is controlled within 10~20 s, and the amount of filtered water during filter cake 0.4 mm~0.5 mm 30–35 - - formation is collected at the same time. 0.3 mm~0.4 mm 25–30 - -

J. Mar. Sci. Eng. 2020, 8, 291 5 of 17

Table 1. Mass percentage of G1, G2 and G3 stratum in diﬀerent particle size ranges.

Stratum G1 (%) G2 (%) G3 (%) 1.6 mm~2.36 mm - - 35–45 1.43 mm~1.6 mm - - 15–20 1.18 mm~1.43 mm - - 15–25 1 mm~1.18 mm - - 10–15 Mass percentage in diﬀerent particle 0.8 mm~1 mm - 15–25 - size ranges 0.6 mm~0.8 mm - 40–50 - 0.5 mm~0.6 mm - 30–35 - 0.4 mm~0.5 mm 30–35 - - 0.3 mm~0.4 mm 25–30 - - 0.25 mm~0.3 mm 25–30 - - <0.25 2.5–5 2.5–7.5 2.5–5 Permeability coeﬃcient (m/s) 3.2 10 4 3.3 10 3 1.93 10 2 × − × − × −

Table 2. Stratum parameter.

Bentonite Clay Sand Slurry Mud CMC Density Viscosity Stratum Particle Particle 3 Number Content Content Content (mL/1000 g) (g/cm ) (Pa s) Size Size · (%) (%) (%) (mm) (mm) 1 95 <0.25 5 - - 10 1.07 1.15 10 2 × − 2 97.5 <0.25 2.5 - - - 1.06 1.06 10 2 G1 × − 3 95 <0.25 5 - - - 1.07 1.12 10 2 × − 4 97.5 <0.25 2.5 - - 10 1.06 1.15 10 2 × − 1 94.5 <0.25 2.5 0.3~0.25 3 7 1.07 1.26 10 2 × − 2 91.5 <0.25 2.5 0.3~0.25 6 7 1.08 1.36 10 2 × − G2 3 88.5 <0.25 2.5 0.3~0.25 9 7 1.11 1.43 10 2 × − 4 89 <0.25 5 0.3~0.25 6 7 1.1 1.38 10 2 × − 5 92.5 <0.25 7.5 - - 7 1.09 1.23 10 2 × − 1 91.5 <0.25 2.5 0.4~0.3 6 30 1.1 1.57 10 2 × − G3 2 91.5 <0.25 2.5 0.6~0.5 6 30 1.1 1.53 10 2 × − 3 85 <0.25 5 0.8~0.6 10 20 1.12 1.75 10 2 × − Note: The CMC is sodium carboxymethyl cellulose.

3.2. Filter Cake Formation State

3.2.1. G1 Formation The permeability coefficient of G1 formation is 3.2 10 4 m/s, which is a highly water-permeable × − formation. The pressure loading of G1 formation mud is divided into 21 levels of loading, and the loading pressure of each stage increases by 0.02 MPa, and the maximum loading pressure is 0.42 MPa. During the loading process, the filtered water quantity is read every 5 s. The formation and morphology of filter cake are shown in Figure3. The mud pressure is greater than the pore water pressure in the formation, and the water and fine components in the mud will penetrate into the formation through the pores of the formation. The mud particles fill the pores of the formation, making the pores of the formation smaller and the permeability coefficient smaller. Infiltration of water increases pore water pressure in the formation. Mud particles gather on the surface of the formation, forming a filter cake of mud skin type. A small amount of mud particles penetrate into the formation, blocking the pores of the formation. J. Mar. Sci. Eng. 2020, 8, 291 6 of 17 J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 6 of 18

FigureFigure 3. Morphology3. Morphology of of filter ﬁlter cake cake (not (not in scale). in scale).

The fourThe types four oftypes mud of mud form form the the “mud “mud skin + +permeablepermeable zone” zone”filter cake filter in the cake G1 formation. in the G1 A formation. layer of dense mud with a thickness of about 1 mm is formed on the surface of the formation. Under A layer ofthe dense same clay mud content, with such a thickness as 1# mud ofand about 3# mud, 1 mm4# mud is and formed 2# mud, on the the distance surface of the of mud the formation. Under thepenetrating same clay into content, the formation such becomes as 1# mud larger and as th 3#e viscosity mud,4# decreases. mud and Increasing 2# mud, the content the distance of of the mud penetratingclay can reduce into thethe distance formation that mud becomes penetrates larger into as the the formation, viscosity such decreases. as 1# mud and Increasing 4# mud, 2# the content of clay canmud reduce and 3# the mud. distance that mud penetrates into the formation, such as 1# mud and 4# mud, The formation of the filter cake has been completed in the initial stage of loading. As the loading 2# mud andprogresses, 3# mud. the quality of the filter cake continues to improve. Figure 4 shows the change in the The formationamount of mud of thewater filter filtration cake with has time been and completed loading steps, in thewhich initial acts at stage a smaller of loading. initial loading As the loading progresses,pressure the quality (0.02 MPa). of Then, the filterthe amount cake of continues filtered water to increased improve. sharply, Figure and4 then shows the growth the change rate in the amount ofof mudthe filtered water water filtration volume slowed with time down and significan loadingtly. The steps, entire which process actswas completed at a smaller within initial 30 loading s, and the slowed growth of the filtered water volume indicated the formation of a filter cake. In the pressure (0.02subsequent MPa). loading Then, process, the amount the amount of filtered of filtered water water increasedincreases rapidly sharply, after andeach thenstep of the loading growth rate of the filteredis completed, water volume and the slowed growth rate down is much significantly. smaller than the The initial entire loading process stage, wasand then completed the growth within 30 s, and the slowedof filtered growth water slows of the down. filtered After the water loading volume pressure indicated is greater than the 0.04 formation MPa, if the of effect a filter of each cake. In the subsequentstage loading of loading process, on the increase the amount in the amount of filtered of filtered water water increases is ignored, rapidly the increase after eachin the amount step of loading is of filtered water changes linearly, indicating that the quality of the filter cake continues to improve completed,after and each the stage growth of loading. rate is much smaller than the initial loading stage, and then the growth of filtered water slows down. After the loading pressure is greater than 0.04 MPa, if the effect of each stage of loading on the increase in the amount of filtered water is ignored, the increase in the amount of filtered water changes linearly, indicating that the quality of the filter cake continues to improve after eachJ. stage Mar. Sci. of Eng. loading. 2020, 8, x FOR PEER REVIEW 7 of 18

250 0.5 Amount of filtered water(g) Loading pressure(MPa) 200 0.4

150 0.3

100 0.2

50 0.1 pressure(MPa) Loading Amount of filtered water(g) offiltered Amount

0 0.0 0 200 400 600 800 1000 1200 1400 t(s)

Figure 4.FigureThe 4. amount The amount of ﬁltered of filtered water water varies with with loading loading and time and in time stratum instratum G1. G1.

3.2.2. G2 Formation The filter cake formed in the G2 stratum is in the form of “mud skin + permeable zone”(see Figure 5). Figure 6 shows the changes in the amount of filtered water with the loading process in stratum G2. Among them, the 5# mud failed to form a filter cake.With the increase of clay content and sand content, the thickness of filter cake gradually becomes thicker, the maximum thickness is 3 mm, but the variation is within 2 mm. Only a small amount of sand exists when the filter cake is dissolved in water, indicating that the filter cake is mainly composed of clay particles. From Table 3, at the same clay content, when the sand content increases from 3% to 9%, the mud invasion distance decreases from 5.5 cm to 2.8 cm; when the sand content is 6%, the clay increases from 2.5% to 5%, and the mud invasion distance decreases by 0.4 mm (see Table 3).

Table 3. Mud invasion distance (thickness of permeable zone).

Mud Number 1 2 3 4 5 Mud invasion distance 5.5 3.7 2.8 3.3 -- (length of permeable zone) (cm)

Figure 5. Mud intrusion into the formation in stratum G2. (a)3# mud Intrusion Formation; (b)5# mud Intrusion Formation.

J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 7 of 18

250 0.5 Amount of filtered water(g) Loading pressure(MPa) 200 0.4

150 0.3

100 0.2

50 0.1 pressure(MPa) Loading Amount of filtered water(g) offiltered Amount

0 0.0 0 200 400 600 800 1000 1200 1400 t(s) J. Mar. Sci. Eng. 2020, 8, 291 7 of 17

Figure 4. The amount of filtered water varies with loading and time in stratum G1. 3.2.2. G2 Formation 3.2.2. G2 Formation The filter cake formed in the G2 stratum is in the form of “mud skin + permeable zone” (see FigureThe5). filter Figure cake6 shows formed the in changes the G2 instratum the amount is in the of filteredform of water“mud withskin the+ permeable loading process zone”( insee stratumFigure 5 G2.). Figure Among 6 shows them, thethe 5# changes mud failed in the to formamount a filter of filtered cake. With water the with increase the loading of clay content process and in sandstratum content, G2. Among the thickness them, ofthe filter 5# mud cake failed gradually to form becomes a filter thicker, cake.With the maximum the increase thickness of clay iscontent 3 mm, butand thesand variation content, is the within thickness 2 mm. of Only filter a cake small gr amountadually ofbecomes sand exists thicker, when the the maximum filter cake thickness is dissolved is 3 inmm, water, but indicatingthe variation that is the within filter 2 cake mm. is mainlyOnly a composedsmall amount of clay of particles.sand exists From when Table the3 ,filter at the cake same is claydissolved content, in water, when indicating the sand content that the increases filter cake from is mainly 3% to 9%,composed the mud of invasionclay particles. distance From decreases Table 3, fromat the 5.5same cm clay to 2.8 content, cm; when when the the sand sand content content is 6%,increases the clay from increases 3% to 9%, from the 2.5% mud to invasion 5%, and thedistance mud invasiondecreases distance from 5.5 decreases cm to 2.8 bycm; 0.4 when mm the (see sand Table content3). is 6%, the clay increases from 2.5% to 5%, and the mud invasion distance decreases by 0.4 mm (see Table 3). Table 3. Mud invasion distance (thickness of permeable zone). Table 3. Mud invasion distance (thickness of permeable zone). Mud Number 1 2 3 4 5 Mud Number 1 2 3 4 5 Mud invasion distance Mud invasion distance 5.5 3.7 2.8 3.3 – (length of permeable zone) (cm) 5.5 3.7 2.8 3.3 -- (length of permeable zone) (cm)

J. Mar.Figure Sci. Eng. 5. Mud 2020 ,intrusion 8, x FOR PEER into REVIEW the formation inin stratumstratum G2.G2. ((aa))3# 3# mud Intrusion Formation;Formation; ((bb))5# 5# mud 8 of 18 Intrusion Formation.

400 ） g （ 300

200

1# 100 2# 3# Amount of filteredwater 4# 5# 0 0 200 400 600 800 1000 Time(s)

Figure 6. Changes in the amount of ﬁltered water with the loading process in stratum G2. Figure 6. Changes in the amount of filtered water with the loading process in stratum G2.

It can be seen from the three stages of the change in the amount of water filtration that the filter cake is basically formed at the end of the first stage. At this stage, the filter cake mainly forms a permeable zone, and the formation still has a certain permeability. In the second stage, due to the pores of the formation are small due to the filling of mud particles, and the small clay particles in the mud gradually accumulate on the surface of the formation to form mud crusts. Because the mud skin is relatively dense, the permeability of the formation is further reduced. In the third stage, the permeability of the formation has become very small. The loading pressure at each stage is relatively high, mud particles continue to accumulate, and the mud skin is compressed and consolidated under pressure, resulting in the permeability of the formation becoming very small. Because the permeability of the G3 stratum is too large and the permeability coefficient is 1.9 cm/s, the three muds prepared all penetrated into the stratum during initial loading (0.02 MPa), and none of them formed a filter cake in the stratum(see Figure 7).

Figure 7. Mud invasion state in G3 stratum. (a)1#Mud intrusion formation; (b)3#Mud intrusion formation.

3.3. Experimental Study on the Effects of Seawater Erosion on Mud Properties Because the project is located on the sea floor, and most of the surrounding rocks are fragmented and affected by seawater, the groundwater contains a large amount of Ca2+, Mg2+, and Na+ ions. After

J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 8 of 18

400 ） g （ 300

200

1# 100 2# 3# Amount of filteredwater 4# 5# 0 0 200 400 600 800 1000 Time(s)

J. Mar. Sci. Eng. 2020, 8, 291 8 of 17 Figure 6. Changes in the amount of filtered water with the loading process in stratum G2.

It cancan bebe seenseen from from the the three three stages stages of of the the change change in thein the amount amount of water of water filtration filtration that thethat filter the filter cake iscake basically is basically formed formed at the endat the of theend first of the stage. first At stag thise. stage, At this the stage, filter cakethe filter mainly cake forms mainly a permeable forms a zone,permeable and thezone, formation and the stillformation has a certain still has permeability. a certain permeability. In the second In stage,the second due to stage, the pores due to of the formationpores of the are formation small due are to thesmall filling due of to mud the filling particles, of mud and theparticles, small clayand particlesthe small in clay the particles mud gradually in the accumulatemud gradually on theaccumulate surface ofon thethe formationsurface of tothe form formation mud crusts.to form Because mud crusts. the mudBecause skin the is relativelymud skin dense,is relatively the permeability dense, the ofpermeability the formation of isthe further format reduced.ion is further In the third reduced. stage, In the the permeability third stage, of the formationpermeability has of become the formation very small. has become The loading very pressuresmall. The at eachloading stage pressure is relatively at each high, stage mud is relatively particles continuehigh, mud to particles accumulate, continue and the to accumulate, mud skin is compressedand the mud and skin consolidated is compressed under and pressure, consolidated resulting under in thepressure, permeability resulting of in the the formation permeability becoming of the very formation small. becoming very small. Because thethe permeabilitypermeability of of the the G3 G3 stratum stratum is toois too large large and and the permeabilitythe permeability coeffi coefficientcient is 1.9 is cm 1.9/s, thecm/s, three the mudsthree muds prepared prepared all penetrated all penetrated into the into stratum the stratum during during initial loading initial loading (0.02 MPa), (0.02 and MPa), none and of themnone of formed them aformed filter cake a filter in thecake stratum in the stratum(see (see Figure7 ).Figure 7).

Figure 7. Mud invasion state in G3 stratum. (a) 1# Mud intrusion formation; (b) 3# Mud intrusion Figure 7. Mud invasion state in G3 stratum. (a)1#Mud intrusion formation; (b)3#Mud intrusion formation. formation. 3.3. Experimental Study on the Effects of Seawater Erosion on Mud Properties 3.3. Experimental Study on the Effects of Seawater Erosion on Mud Properties Because the project is located on the sea floor, and most of the surrounding rocks are fragmented and aBecauseffected bythe seawater,project is located the groundwater on the sea floor, contains and amost large of amount the surrounding of Ca2+, rocks Mg2+ ,are and fragmented Na+ ions. Afterand affected the mud by and seawater, groundwater the groundwater are mixed, contains the stability a large of the amount mud isof a Caffected.2+, Mg When2+, and these Na+ ions. positively After charged ions change, they change from a suspended state to an agglomerated state, which reduces the particles involved in the formation of the filter cake, reduces the viscosity of the mud, and reduces the ability of the mud to form a filter cake [35,36]. When the slurry shield works, after the salt is mixed with the slurry, the slurry penetrates into the stratum to form a filter cake. A thorough understanding of the mixing characteristics of mud and seawater and the characteristics of filter cake formed by salt cement slurry is the key to mud shield construction. The test mud was 1# mud of G1. 150 mL of seawater was added to 1500 mL of mud one by one, and the amount of each addition was 15 mL. The test was performed 10 times. In order to eliminate the influence of water in the seawater on the viscosity of the mud, a control group experiment was designed in which the viscosity of the mud changed when tap water was added.

4. Numerical Simulation of Excavation Surface Stability

4.1. Calculation Model Under different stratum conditions, the degree of secondary disturbance caused by shield tunnel construction is not the same, and the stability of the tunnel face after excavation is also different. Under complex engineering conditions, it is not enough to rely on simple engineering analogies. It is J. Mar. Sci. Eng. 2020, 8, 291 9 of 17 necessary to use numerical calculation methods to determine the optimal excavation and support process, and to analyze the dynamic behavior of the tunnel. It is necessary to judge the stability of the caverns [37–40]. In order to evaluate and analyze the stability of the excavation face of shield construction in this project, unfavorable sections with poor geological conditions are selected according to the geological map of the project interval. The FLAC3D finite difference software is used to simulate the construction process of a shield tunnel, and the stability of the shield excavation face under different stratum conditions is analyzed [41]. The numerical model of shield tunneling is established on the platform of FLAC3D. In order to minimize the “boundary effect” in the calculation, combined with the excavation diameter of the shield, the overall size of the model is 90 m (X axis) 60 m (Y axis) 55.36 m (Z axis); the inner diameter × × of the segment is 6.0 m, the outer diameter is 6.7 m, the thickness is 0.35 m, and the segment length is 1.2 m. Mohr-Coulomb constitutive model can effectively simulate the mechanical behavior of the formation when it enters the plastic stage; elastic constitutive model is used for segment lining structure. In order to simplify the calculation, different buried depth is simulated by applying different loads on the model. See Table4 for the calculation parameters of related strata involved in the calculation. It mainly includes physical property test, compression-consolidation test and triaxial shear test of soil. Meanwhile, the mechanical parameters of shield segment lining during model excavation are shown in Table5.

Table 4. The calculation parameters of related strata involved in the calculation.

Elastic Internal K (coeﬃcient Density Poisson’s Cohesion 0 Section Strata Situation Modulus Friction of Horizontal (kg/m3) Ratio (MPa) (MPa) (◦) Earth Pressure) plain ﬁll 1870 7.68 0.46 20 15 0.45 silt 1620 7.68 0.35 10 15 0.56 silty clay 1960 30 0.25 10 4 0.54 1 residual sandy clay 1840 50 0.33 22 20 0.52 fully weathered granite 1870 80 0.3 25 25 0.45 scattered strongly weathered 1920 150 0.25 30 27 0.40 granite Fragmented strong 2210 1300 0.2 40 28 0.35 weathered granite silt 1620 7.68 0.35 10 15 0.62 fully weathered granite 1870 80 0.3 25 25 0.45 2 scattered strongly weathered 1920 150 0.25 30 27 0.42 granite moderately weathered 2640 2000 0.2 50 30 0.35 granite silty sandstone 1780 7.68 0.38 8 12 0.45 silt 1620 7.68 0.35 10 15 0.60 3 silty clay 1960 30 0.25 30 20 0.54 silty clay 1750 7.68 0.35 12 5 0.62 strongly sandstone 1960 140 0.25 28 25 0.42

Table 5. Calculation parameters of shield segment lining.

Length Thickness Elastic Modulus Related Parameters Poisson’s Ratio Density (kg/m3) (m) (m) (MPa) Shield segment 1.2 0.35 34.5 103 0.17 2460 ×

4.2. Basic Calculation Assumptions The assumptions in the calculation are as follows: (1)The ground surface and each soil layer are assumed to be in a uniform horizontal layered distribution; (2) The squeezing and shearing eﬀects of the shield shell itself and the soil body are not considered; (3) The change of soil physical and mechanical parameters during construction is not considered.; (4) The size of the shield tail gap, the degree of J. Mar. Sci. Eng. 2020, 8, 291 10 of 17 grouting and ﬁlling compaction, and the degree and range of disturbance of the soil around the tunnel are not taken into account during shield construction.

4.3. Numerical Simulation Calculation Process This paper mainly focuses on the mechanical behavior and stability of the tunnel excavation surface during shield construction. Therefore, the axial excavation section of the tunnel is set as the target surface in the calculation model to analyze the behavior of the target surface during construction. Therefore, in the process of construction simulation, one-time excavation and support are taken into account before and after the target, i.e., 30 m (25 rings) before the target, 1 excavation step is set at the middle target section, 1.2 m (1 ring segment width) is excavated, and 28.8 m (24 rings) is excavated after the target surface; During each excavation step, shield excavation is simulated and shield segments are assembled.

4.4. Calculation of Silo Pressure Simulation Adjustment Strategy Shield excavation is a progressive process. The focus of this paper is the adjustment strategy of the support pressure of the excavation surface during the construction process. Therefore, the support pressure ratio is used to adjust the support pressure. The calculation formula is as follows: σ λ = s (1) σ0

In the above formula, σs is the support stress of the pressure chamber bulkhead node, and σ0 is the lateral static earth pressure of the initial formation. According to Rankine’s earth pressure formula, shield position at depth of about 19 m, σh0 at this point is 480 kPa. In the calculation, the support stress ratio of the left and right line construction simulation was adjusted to achieve the simulation of the support pressure of the excavated tunnel face. For diﬀerent calculation sections, the stability of the shield tunnel excavation face under diﬀerent support pressure ratios (factors) is analyzed during the calculation; according to the construction process of the shield tunnel, the support pressure ratios (factors) used in the calculation of each section. There are 10 calculation conditions of 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1, and 0.05 respectively (Support stress of excavation face/initial formation lateral static earth pressure). The stability of the excavation of the target surface under each calculation condition is compared and analyzed.

4.5. Boundary Conditions Because the shield construction speed is fast and the segments can be assembled in time, the change of groundwater level is not obvious at the ﬁrst time, so the seepage of water can not be considered during construction, that is, there is no seepage boundary condition. Displacement boundary conditions: lateral displacement is constrained, longitudinal displacement is constrained on the front and back, vertical displacement is constrained on the bottom, and the top of the model is a free surface.

5. Discussion

5.1. Analysis of Filter Cake Formation Mechanism In the G1 formation, the higher the clay content and the higher the viscosity in the mud, the slower the growth rate of the filtered water volume and the smaller the final filtered water volume (see Figure8). It can be seen from Figure9 that, because the clay content and viscosity of 1# mud is the largest, the growth rate of the water filtration capacity and the final filtration volume are the smallest. As the clay content in the mud decreases and the viscosity decreases, the growth rate of filtered water and the final filtered water volume increase. The change in the amount of filtered water of 2# mud and 3# mud is basically the same, indicating that the problem of insufficient particulate matter can be overcome by adjusting the viscosity of the mud. J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 11 of 18 J. Mar. Sci. Eng. 2020, 8, 291 11 of 17

4.5. Boundary Conditions ItBecause can also the be shield seen from construction Figure 10 thatspeed the is formation fast and the of the segments filter cake can can be beassembled completed in at time, a lower the pressurechange of (less groundwater than 0.06 MPa). level As is thenot pressure obviousincreases, at the first the time, quality so ofthe the seepage filter cake of furtherwater can improves, not be andconsidered the amount during of filtered construction, water is weakenedthat is, there by the is increaseno seepage in pressure. boundary During condition. the second Displacement half of the loading,boundary the conditions: amount of lateral water filtereddisplacement showed is aconstrained, linear change longitudinal with time, indicatingdisplacement that is the constrained filter cake canon the improve front theand quality back, vertical in a short displacement time and offset is co thenstrained effect of increasedon the bottom, pressure. and At the the top same of the time, model it is notedis a free that surface. the slopes of the second half of the four muds are basically the same, indicating that at the same pressure In the following, the permeability of the filter cake formed by using the same material to prepare5. Discussion the mud is the same. The properties of the filter cake are related to the mud material, but the J.ratio Mar. ofSci. material Eng. 2020, use8, x FOR is not PEER significant. REVIEW 12 of 18 5.1. Analysis of Filter Cake Formation Mechanism 700 In the G1 formation, the higher the clay content and the higher the viscosity 1# in the mud, the slower the growth600 rate of the filtered water volume and the smaller the final filtered 2#water volume(see Figure 9). It can be seen from Figure 8 that, because the clay content and viscosity 3# of 1# mud is the largest, the growth500 rate of the water filtration capacity and the final filtration volume 4# are the smallest. As the clay content in the mud decreases and the viscosity decreases, the growth rate of filtered water and the final filtered400 water volume increase. The change in the amount of filtered water of 2# mud and 3# mud is basically the same, indicating that the problem of insufficient particulate matter can be overcome by adjusting300 the viscosity of the mud. It can also be seen from Figure 10 that the formation of the filter cake can be completed at a lower pressure (less than200 0.06 MPa). As the pressure increases, the quality of the filter cake further improves, and the amount of filtered water is weakened by the increase in pressure. During the second half of

Amount of filtered water(g) filtered of Amount 100 the loading, the amount of water filtered showed a linear change with time, indicating that the filter cake can improve the quality in a short time and offset the effect of increased pressure. At the same 0 time, it is noted that 0the slopes 200 of the second 400 half 600of the four 800 muds are 1000 basically 1200 the same, indicating that at the same pressure In the following, the permeability of the filter cake formed by using the Time(s) same material to prepare the mud is the same. The properties of the filter cake are related to the mud material, but theFigureFigure ratio 9. 8. ofComparisonComparison material useof of different diisﬀ noterent significant. mud mud water water f ﬁltrationiltration capacity in stratum G1.

No filter cake was formed in G3 formation. There are three main reasons for the analysis: First, although the particle size of sand particles in the mud can block the pores in the formation, but because the particle size distribution of the sand particles in the mud is uniform, it still exists after blocking the pores in the formation. Mud-permeable channels. Secondly, for highly permeable formations, due to the large porosity of the formation, the mud needs to provide enough particulate matter to block the seepage channels. The content of particulate matter in the mud prepared in the test is not enough to form a filter cake. Lastly, due to the large formation pores, the viscosity of the mud prepared in the test has not yet affected the formation of filter cake.

5.2. Effect of Seawater on Mud Viscosity As can be seen from Figure 10, the addition of seawater reduces the viscosity of the slurry. When the volume ratio of seawater is less than 3%, the viscosity decreases quickly, decreasing by 0.15 Pa·s. After 3%, the viscosity decreases slowly. When the volume ratio of seawater is 10%, the viscosity decreases by 0.15 Pa·s. In the control group test in which tap water was added, when the tap water was added in a volume ratio of less than 7%, the viscosity of the mud decreased substantially linearly, and then the decrease was reduced. Comparing the effect of adding seawater and tap water on the viscosity of the mud, it can be found that after adding more than 7% of the volume, the viscosity of Figure 9. The soil proﬁle along Section3. the mud is basically the same, indicatingFigure 8. thatThe soilthe profileinfluence along of Sectionions in 3.the seawater is weakened at this time, and the water content in the mud is mainly affected at this time. It can be seen that even a small No ﬁlter cake was formed in G3 formation. There are three main reasons for the analysis: amount of seawater (less than 3% by volume) has a significant effect on mud viscosity. First, although the particle size of sand particles in the mud can block the pores in the formation, but because the particle size distribution of the sand particles in the mud is uniform, it still exists

J. Mar. Sci. Eng. 2020, 8, 291 12 of 17 after blocking the pores in the formation. Mud-permeable channels. Secondly, for highly permeable formations, due to the large porosity of the formation, the mud needs to provide enough particulate matter to block the seepage channels. The content of particulate matter in the mud prepared in the test is not enough to form a filter cake. Lastly, due to the large formation pores, the viscosity of the mud prepared in the test has not yet affected the formation of filter cake.

5.2. Effect of Seawater on Mud Viscosity As can be seen from Figure 10, the addition of seawater reduces the viscosity of the slurry. When the volume ratio of seawater is less than 3%, the viscosity decreases quickly, decreasing by 0.15 Pa s. After 3%, the viscosity decreases slowly. When the volume ratio of seawater is 10%, · the viscosity decreases by 0.15 Pa s. In the control group test in which tap water was added, when the · tap water was added in a volume ratio of less than 7%, the viscosity of the mud decreased substantially linearly, and then the decrease was reduced. Comparing the effect of adding seawater and tap water on the viscosity of the mud, it can be found that after adding more than 7% of the volume, the viscosity of the mud is basically the same, indicating that the influence of ions in the seawater is weakened at thisJ. Mar. time, Sci. and Eng. the2020 water, 8, x FOR content PEER REVIEW in the mud is mainly affected at this time. It can be seen that even 13 of a 18 small amount of seawater (less than 3% by volume) has a significant effect on mud viscosity.

Figure 10. Effect of seawater on mud viscosity. Figure 10. Effect of seawater on mud viscosity. 5.3. Analysis of Excavation Surface Stability 5.3.Under Analysis di ffoferent Excavation geological Surface conditions, Stability as the support pressure of the excavation face decreases, the verticalUnder displacement different geological and longitudinal conditions, (y-direction) as the suppo displacementrt pressure ofof thethe excavationexcavation face face gradually decreases, increase,the vertical and thedisplacement growth trend and appears longitudinal non-linear. (y-direction) The maximum displacement vertical of displacement the excavation of theface excavationgradually surfaceincrease, occurs and the at thegrowth vault trend position, appears and no then-linear. longitudinal The maximum maximum vertical displacement displacement of the of excavationthe excavation surface surface occurs occurs at the at center the vault position positi ofon, the and tunnel the excavation longitudinal surface maximum (see Figure displacement 11). of theAccording excavation to surface the data occurs of Section at the3 incenter the Figureposition 12 ,of during the tunnel the excavation excavation of surface shield tunnel,(see Figurethere 11). is a certainAccording settlement to the at thedata top of ofSection the excavation 3 in the Figure face and12, during a certain the upliftexcavation at the of bottom shield oftunnel, the arch; there theis verticala certain displacement settlement at of the the top excavation of the excavation face is di fffaceerent and under a certain different uplift support at the bottom pressure. of the arch; the vertical displacement of the excavation face is different under different support pressure.

Support pressure coefficient 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 ） 0 mm （

-2

-4

-6

-8 Section 1 Section 2 Section 3 -10 Vertical displacement of vault position of excavation face

Figure 11. Vertical Displacement of Excavation Face under Different Supporting Pressure Coefficients.

J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 13 of 18

Figure 10. Effect of seawater on mud viscosity.

5.3. Analysis of Excavation Surface Stability Under different geological conditions, as the support pressure of the excavation face decreases, the vertical displacement and longitudinal (y-direction) displacement of the excavation face gradually increase, and the growth trend appears non-linear. The maximum vertical displacement of the excavation surface occurs at the vault position, and the longitudinal maximum displacement of the excavation surface occurs at the center position of the tunnel excavation surface (see Figure 11). According to the data of Section 3 in the Figure 12, during the excavation of shield tunnel, there isJ. Mar.a certain Sci. Eng. settlement2020, 8, 291 at the top of the excavation face and a certain uplift at the bottom of the13 arch; of 17 the vertical displacement of the excavation face is different under different support pressure.

Support pressure coefficient 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 ） 0 mm （

-2

-4

-6

-8 Section 1 Section 2 Section 3 -10 Vertical displacement of vault position of excavation face Figure 11. Vertical Displacement of Excavation Face under Diﬀerent Supporting Pressure Coeﬃcients. Figure 11. Vertical Displacement of Excavation Face under Different Supporting Pressure J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 14 of 18 Coefficients.

Figure 12. Vertical displacement cloud diagram of excavation face under various conditions of Section3. Figure 12. Vertical displacement cloud diagram of excavation face under various conditions of Section (The numerical value in the figure represents the support pressure coefficient). 3. (The numerical value in the figure represents the support pressure coefficient). Figure 13 shows variation curve of vault vertical displacement of excavation face under different support pressures. Under different geological conditions, as the support pressure of the excavation Section1 face decreases, the plastic25 area of the excavation face gradually develops; where the support pressure Section2 is large, the plastic area of the excavation face is distributed around the tunnel cavity (see Figure 14). Section3 Under the condition20 of small support pressure, the plastic area of the excavation face is relatively large;

Maximum longitudinal displacement of excavation face(mm) excavation of displacement longitudinal Maximum 0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Support pressure coefficient

Figure 13. Variation curve of vault vertical displacement of excavation face under different support pressures.

Figure 13 shows variation curve of vault vertical displacement of excavation face under different support pressures. Under different geological conditions, as the support pressure of the excavation face decreases, the plastic area of the excavation face gradually develops; where the support pressure is large, the plastic area of the excavation face is distributed around the tunnel cavity(see Figure 14). Under the condition of small support pressure, the plastic area of the excavation face is relatively

J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 14 of 18

J. Mar. Sci. Eng. 2020, 8, 291 14 of 17

Figure 12. Vertical displacement cloud diagram of excavation face under various conditions of Section according to diﬀerent stratum conditions, the plastic area of the excavation face has certain diﬀerences 3. (The numerical value in the figure represents the support pressure coefficient). during the tunnel excavation process.

Section1 25 Section2 Section3 20

Maximum longitudinal displacement of excavation face(mm) excavation of displacement longitudinal Maximum 0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 15 of 18 Support pressure coefficient large;Figure according 13. Variationto different curve stratum of vault conditions, vertical displacementthe plastic area of excavationof the excavation face under face di hasﬀerent certain Figure 13. Variation curve of vault vertical displacement of excavation face under different support differencessupport during pressures. the tunnel excavation process. pressures.

Figure 14. DistributionDistribution of of plastic plastic area area in in excavation face face with with different different support force coefficients coefficients in Section 33.. (The (The numerical numerical value value in in the the figure figure represents represents the the support support pressure pressure coe coefficient).fficient).

6. Conclusions In this paper, the Xiamen Metro Line 2 is used as the engineering study case, the formation of filter cake in slurry shield was studied experimentally. By preparing mud with different properties and carrying out filter cake formation tests in three different strata, the formation patterns and formation laws of filter cakes are tested. The effects of stratum and mud properties on filter cake formation were studied, and numerical simulations of the stability of the excavation face were carried out. The following conclusions can be drawn: (1) The filter cake has been formed in the initial loading stage, the formation of a filter cake does not require a large pressure difference. As the loading pressure increases, the quality of the filter cake continuously improves; (2) For strongly permeable formations, filter cakes mainly in the form of “mud skin + permeable zone” are formed, in which the mud skin is thin and the permeable zone is thick; (3) By increasing the content of clay in the mud, the particle size and content of sand, and the viscosity of the mud, a better-quality filter cake can be formed, reducing the penetration of mud into the formation, and reducing the amount of mud needed to form a filter cake; (4) Filter cake can be formed only by bentonite base slurry and a small amount of clay in formations with low permeability, but as the permeability of the formation becomes larger and the pores in the formation increase and the pore size increases, the particles in the mud need to be increased. The content of the substance and the viscosity of the mud usually need to ensure a certain sand content in the mud and increase the viscosity of the mud in the project. However, if the

J. Mar. Sci. Eng. 2020, 8, 291 15 of 17

6. Conclusions In this paper, the Xiamen Metro Line 2 is used as the engineering study case, the formation of filter cake in slurry shield was studied experimentally. By preparing mud with different properties and carrying out filter cake formation tests in three different strata, the formation patterns and formation laws of filter cakes are tested. The effects of stratum and mud properties on filter cake formation were studied, and numerical simulations of the stability of the excavation face were carried out. The following conclusions can be drawn: (1) The filter cake has been formed in the initial loading stage, the formation of a filter cake does not require a large pressure difference. As the loading pressure increases, the quality of the filter cake continuously improves; (2) For strongly permeable formations, filter cakes mainly in the form of “mud skin + permeable zone” are formed, in which the mud skin is thin and the permeable zone is thick; (3) By increasing the content of clay in the mud, the particle size and content of sand, and the viscosity of the mud, a better-quality filter cake can be formed, reducing the penetration of mud into the formation, and reducing the amount of mud needed to form a filter cake; (4) Filter cake can be formed only by bentonite base slurry and a small amount of clay in formations with low permeability, but as the permeability of the formation becomes larger and the pores in the formation increase and the pore size increases, the particles in the mud need to be increased. The content of the substance and the viscosity of the mud usually need to ensure a certain sand content in the mud and increase the viscosity of the mud in the project. However, if the permeability of the formation is too large, the gradation of large particles in the mud needs to be considered in order to form a filter cake. (2) The stratum traversed by the tunnel has good self-stabilizing ability. The stratum that is prone to instability is mainly medium-sand-bearing stratum. However, considering that this project is a subsea shield tunnel, it is affected by high water pressure, fissure development, and shield tunneling. The excavation disturbs the stratum stress field and seepage field, and there is uncertainty in the stability of the excavation face during the opening of the warehouse, which threatens the safety of the construction. How to use and improve the stability of the stratum has become the focus of the project.

Author Contributions: Conceptualization, H.W.; methodology, J.C.; software, T.L.; validation, Y.L.; formal analysis, L.G.; investigation, T.L.; resources, Y.Z.; data curation, X.S.; writing—original draft preparation, H.W.; writing—review and editing, H.L.; visualization, H.L.; supervision, L.G.; project administration, J.C.; funding acquisition, T.L. All authors have read and agreed to the published version of the manuscript. Funding: This study was supported by the National Natural Science Foundation of China (Project Nos. 41672272 and 41806075), the Fundamental Research Funds for the Central Universities (201962011), and the foundation for Shandong Province Key Research and Development Plan (Grant No. 201809260098). We also thank Southwest Jiaotong University for providing experimental conditions. Acknowledgments: We appreciate anonymous reviewers who gave comments to revise the paper. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Min, F.; Zhu, X.; Lin, C.; Guo, X. Opening the excavation chamber of the large-Diameter size slurry shield: A case study in Nanjing Yangtze River Tunnel in China. Tunn. Undergr. Space Technol. 2015, 46, 18–27. [CrossRef] 2. Jun, S. On the Technical Advantages and Problems of Constructing Tunnels Across Rivers and Seas. Tunn. Constr. 2013, 33, 337–342. 3. Shuming, Z. Overall Design and Construction of Qingdao Jiaozhou Bay Subsea Tunnel. Tunn. Constr. 2013, 33, 38–44. 4. Jingeng, L. Research on Xiamen Urban Rail Transit Development Strategy. Master’s Thesis, Huaqiao University, Quanzhou, China, 2015. J. Mar. Sci. Eng. 2020, 8, 291 16 of 17

5. Zhicehng, W. Procedure of Replacing of Cutting Teeth under Atmospheric Pressure and Its Optimization. Tunn. Constr. 2010, 3, 339–343. 6. Chuangmian, L. Technical gist of entering earth cabin with press in shield construction. Constr. Mech. 2008, 4, 51–53. [CrossRef] 7. Herrenknecht, E.M.; Bäppler, K. Compressed air work with tunnel boring machines. In Underground Space-the 4th Dimension of Metropoises; Taylor & Francis Group: London, UK, 2007; pp. 1175–1181. 8. Heijboer, J.; Van Den Hoonaard, J.; van de Linde, W. The Westerschelde tunnel: Approaching limits; CRC Press: Boca Raton, FL, USA, 2003; pp. 1–36. 9. Zhongshan, W.; Zhong, X.; Yang, Z. Calculation of the minimum excavation face support pressure for opening chamber in slurry shield in Nanjing Weisan Road River Tunnel. China Harb. Eng. 2014, 2, 33–36. 10. Fanlu, M.; Teng, J.; Daizhou, W.; Yazhou, Z. Experimental Study on Preparation and Filter Membrane Formation of Slurry during Hyperbaric Operation in Tunneling by Slurry Shield. Tunn. Constr. 2014, 34, 857–861. [CrossRef] 11. Ning, Z.; Zhu, W.; Min, F.; Xu, J. Microscopic pores of filter membranes and permeability during chamber opening under high pressure in slurry shield. Chin. J. Geotech. Eng. 2017, 39, 495–500. [CrossRef] 12. Mingliang, C.; He, F.; Chuantian, L. Tools repairing and welding techniques with high pressure of slurry shield under deep water-rich sandy gravel stratum. Eng. Sci. 2010, 12, 46–50. 13. Davis, E.H.; Gunn, M.J.; Mair, R.J.; Seneviratne, H.N. The stability of shallow tunnels and underground openings in cohesive material. Geotechnique 1980, 30, 397–416. [CrossRef] 14. Leca, E.; Dormieux, L. Upper and lower bound solutions for the face stability of shallow circular tunnels in frictional material. Geotechnique 1990, 40, 581–606. [CrossRef] 15. Lee, I.; Lee, J.; Nam, S. Effect of seepage force on tunnel face stability reinforced with multi-Step pipe grouting. Tunn. Undergr. Space Technol. 2004, 19, 551–565. [CrossRef] 16. Takano, D.; Otani, J.; Nagatani, H.; Mukunoki. Application of x-ray CT on boundary value problems in geotechnical engineering: Research on tunnel face failure. In Proceedings of the GeoCongress 2006: Geotechnical Engineering in the Information Technology Age, Atlanta, GA, USA, 26 February–1 March 2006; pp. 1–6. 17. Kirsch, A. Experimental investigation of the face stability of shallow tunnels in sand. Acta Geotech. 2010, 5, 43–62. [CrossRef] 18. Mair, R.J. Centrifugal Modelling of Tunnel Construction in Soft Clay. Ph.D. Thesis, University of Cambridge, Cambridge, UK, 1978. 19. Hehua, Z.; Wenqi, D.; Li, X. Construction simulation for the mechanical behavior of shield tunnel and its application. China Civ. Eng. J. 2000, 3, 98–103. [CrossRef] 20. Chunlin, D.; Shiyou, Z.; Zhou, S. Influence of in-Situ stress release on the stability of surrounding rock masses and ground surface settlement for shield-driven tunnel. Chin. J. Rock Mech. Eng. 2002, 11, 1633–1638. 21. Hongjie, Y.; Deming, F.; Ge, X. Experimental study and numerical simulation of earth pressure around shield machine. Chin. J. Rock Mech. Eng. 2006, 8, 1652–1657. [CrossRef] 22. Zhang, H.; Yin, Z.; Zhu, J. 3d finite element simulation on deformation of soil mass during shield tunneling. Chin. J. Rock Mech. Eng. 2005, 5, 755–760. 23. Xinliang, J.; Cui, Y.; Li, Y.; Zm, Z. Measurement and simulation of ground settlements of Tianjin subway shield tunnel construction. Rock Soil Mech. 2005, 10, 91–95. 24. Xiaoqing, L.; Zhu, C. Numerical analysis on the ground settlement induced by shield tunnel construction. J. Highw. Transp. Res. Dev. 2007, 6, 86–91. [CrossRef] 25. Yintao, Z.; Tao, L.; Bian, J. Analysis of the numerical simulation of ground movement induced by shield tunneling. J. Beijing Univ. Technol. 2006, 4, 332–337. 26. Lei, H. Research on Technology of Replacing Cutting Tool with Pressure of EPB Shield in Water-Rich and Pebbly Sand Stratum. Ph.D. Thesis, Southwest Jiaotong University, Sichuan, China, 2009. 27. Anagnostou, G.; Kovári, K. The face stability of slurry-Shield-Driven tunnels. Tunn. Undergr. Space Technol. 1994, 9, 165–174. [CrossRef] 28. Harding, P.G. High density slurry shield extends Japanese technology. Tunn. Tunn. Int. 1982, 14, 55–56. 29. Broere, W. Tunnel Face Stability and New CPT Applications; Delft University of Technology Press: Delft, The Netherlands, 2002. J. Mar. Sci. Eng. 2020, 8, 291 17 of 17

30. Kim, S.H.; Tonon, F. Face stability and required support pressure for TBM driven tunnels with ideal face membrane–Drained case. Tunn. Undergr. Space Technol. 2010, 25, 526–542. [CrossRef] 31. Watanabe, T.; Yamazaki, H. Giant size slurry shield is a success in Tokyo. Tunn. Tunn. Int. 1981, 13, 13–17. 32. Min, F.L.; Zhu, W.; Han, X.R.; Zhong, X.C. The effect of clay content on filter-Cake formation in highly permeable gravel. In Proceedings of the GeoShanghai International Conference 2010, Shanghai, China, 3 June 2010; pp. 201–215. 33. Xiaorui, H.; Zhu, W.; Liu, Q.; Zhong, X.; Min, F. Influence of slurry property on filter-Cake quality on working face of slurry shield. Rock Soil Mech. 2008, 29, 288–292. [CrossRef] 34. Xu, T.; Bezuijen, A. Bentonite slurry infiltration into sand, filter cake formation under various conditions. Géotechnique 2019, 69, 1–36. [CrossRef] 35. Jiarui, D.; Min, F.; Yao, Z. Experimental study of property change of slurry and filter cake of slurry shield under seawater intrusion. Tunn. Constr. 2018, 38, 1182–1188. 36. Kelessidis, V.C.; Tsamantaki, C.; Dalamarinis, P. Effect of pH and electrolyte on the rheology of aqueous Wyoming bentonite dispersions. Appl. Clay Sci. 2007, 38, 86–96. [CrossRef] 37. Su, G. Study on Stability Analysis and Intelligent Optimization for Large Underground Caverns under High Geostress Condition. Ph.D. Thesis, Chinese Academy of Sciences, Beijing, China, 2006. 38. Jun, S.; Zhu, H. Mechanical simulation and analysis of behaviour of soft and weak rocks in the construction of a tunnel opening. Rock Soil Mech. 1994, 4, 20–33. 39. Xiao, M. Three-Dimensional numerical model of construction process for underground opening. Chin. J. Geotech. Eng. 2000, 4, 421–425. 40. Xiao, H. Study on Stability of Surrounding Rock and Optimization of Bolt Parameters for TBM Assembly Cavern in High Stress. Ph.D. Thesis, Southwest Jiaotong University, Chengdu, China, 2010. 41. Li, Y.; Emeriault, F.; Kastner, R.; Zhang, Z.X. Stability analysis of large slurry shield-Driven tunnel in soft clay. Tunn. Undergr. Space Technol. 2009, 24, 472–481. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).