Structural Durability & Health Monitoring

---

DOI: 10.32604/sdhm.2022.013141

ARTICLE

Dynamic Risk-Warning of Center Diaphragm and Bench Composite Method During Construction

Xiaozhong Li* and Caiyun Sun

School of Civil Engineering, Lanzhou Jiaotong University, Lanzhou, 730070, China
*Corresponding Author: Xiaozhong Li. Email: kiki50257@163.com
Received: 27 July 2020; Accepted: 14 October 2020

1  Introduction

As a very complex system engineering, subway has the characteristics of large investment, large number of project participants, complex technology, long period, strong concealment and various geological conditions and so on, which brings great risks to construction. Therefore, scholars at home and abroad have carried out a lot of research on tunnel construction safety. There are many factors affecting construction safety and their contribution to risk value is different, so, in order to evaluate the construction safety scientifically and systematically, scholars like Wang et al. [1] To solve the randomness and fuzziness of water inrush risk evaluation, a comprehensive evaluation model was established based on normal cloud theory; Yuan et al. [2] established a Catastrophe theory model of tunnel collapse risk assessment based on the analysis of risk pregnancy environment and inducement of Hongyansi tunnel; Alireza et al. [3] and others combined the game theory with the interactive decision structure model of the hierarchical process of fuzzy analysis for risk management in tunnel engineering design, construction and operation; Sousa et al. [4] proposed a systematic approach to assess and manage tunnel construction-related risks based on Bayesian network; Ye et al. [5] developed an intelligent risk assessment system for deep-base pit precipitation; Abdolreza [6] proposed a risk assessment model based on fuzzy set theory to assess risk events during tunnel construction; Bai et al. [7] proposed a multi-stage risk management approach to better carry out risk management and optimize risk mitigation; Cagatay [8] selected event tree analysis method to analyze the operating risk of tunnel construction TBM (tunnel tunneling machine); Han et al. [9] proposed a risk assessment method to evaluate the pre-existing circumferential cracks of shield tunnel in order to ensure the safety of deep-digging tunnel during construction; Xia et al. [10] based on fuzzy set theory and similarity measurement theory, proposed a risk assessment model which is difficult to quantify fuzzy characteristics; Lin et al. [11] applied the improved cloud model to tunnel construction risk assessment; Xue et al. [12] rediction model for subway tunnel collapse risk based on Delphi-Ideal point method and geological forecast. Some scholars use monitoring techniques to predict the possibility of risk occurrence, scholars like Mahdi et al. [13] combined monitoring with mathematical evaluation methods according to the safety impacts of subway construction on adjacent existing bridges, and established the safety risk assessment and control system of the existing bridges including four aspects: pre-work detection, pre-work assessment, in-work dynamic control, post-work evaluation and recovery; Su et al. [14] used PSO-ANN model to predict the feasibility of surface settlement caused by tunnel excavation; Liu et al. [15] introduced microseismic monitoring and risk management theory in tunnel blasting excavation, which can obtain the probability of occurrence, potential consequences and risk grade of rock burst in real time; Liu et al. [16] proposed the reliability analysis of subway tunnel operation based on dynamic bayesian Copula model; Li et al. [17] evaluated the safety of slope with efficient Bayesian network on the strength of monitoring technology including sensors and wireless; Kang et al. [18] based on the real-time monitoring data of water inrush in Zoumaling tunnel excavation section, used fuzzy data analysis method to analyze and predict the content of water inrush.

In summary, in the aspect of risk research, on the one hand, some scholars can comprehensively identify the risk evaluation index and use scientific mathematical model to calculate risk values. However, the qualitative index is mostly and difficult to quantify, and their value are subjective only based on expert scoring. On the other hand, some scholars also use monitoring technology to carry out risk warning according to deformation value, but do not consider the different contribution of each index deformation to the overall risk. In this paper combines these two methods, it will rely on the Changping section tunnel of Guanhui Intercity Metro, and establish a dynamic risk-warning model for the construction process of subway tunnel with the CD-Bench composite method. Not only does the each risk evaluation index adopt a combination weighting method to give different weights, and uses scientific mathematical models to calculate their risk value, but also greatly reduces the difficulty of quantifying indicators.

The rest of this paper is organized as follows. In Section 2, the monitoring scheme is descripted. In Section 3, the construction risk evaluation is presented in detail. In Section 4 the information feedback of early-warning is descripted. In Section 5, a case analysis is performed to demonstrate the effectiveness of the proposed method. Conclusions and future work of the research are discussed in Section 6.

2  Monitoring Scheme

2.1 Engineering Situation

The Guanhui Intercity Rail Transit runs from Hongmei Station in Dongguan City to Huizhou Railway Station, This article mainly focuses on the Changping section of the tunnel as the research object. This part of the tunnel has such characteristics: the surrounding environment is complex, close to factory buildings, residential areas, etc., and it needs to pass through the artificial lake of the railway park, the HanViRiver, and cross the broken zone; and its ground elevation of this section is 7.31∼14.32 m, and its buried depth is about 22–36 m. The engineering geological conditions are poor, mainly quaternary strata and broken zones; and need to pass through the artificial lake and HanViRiver in the railway park, so the surface water is abundant, and the groundwater is mainly bedrock fissure water and pour water, which has a greater impact on construction and is locally corrosive. In order to control tunnel deformation and ensure the safety of the construction, it is particularly important to establish a risk warning model.

In order to effectively prevent surface subsidence and large deformation of surrounding rock, the Changping section of the Guanhui intercity adopts the CD-Bench composite construction method, which means that the tunnel section is divided into two parts for excavation: The upper part is constructed by the CD method and the lower part is constructed by the bench method. This makes full use of the respective advantages of these two construction methods: the upper part is provided with vertical support by the partition wall, which effectively controls the settlement of the surrounding rock vault; and the lower part adopts the bench cut method for construction and facilitates the access of large machinery. In addition, due to the addition of temporary inverted arches, closed steel arches and other measures, the tunnel is less deformed during construction than the traditional single construction method, so has higher safety. At the same time, the construction period is shortened due to staggered construction. Although the construction cost of this method is high, compared with other treatment measures, the “performance-to-price ratio” is still the highest. The on-site construction drawing is shown in Fig. 1.

Figure 1: Site construction drawings

2.2 Monitoring Content and Frequency

The surface of this section of subway tunnel is roads with high traffic volume, HanViRiver Bridge, residential houses and levees, etc., and it needs to pass through the broken zone. Therefore, accurate and timely monitoring and measurement play an important role in the protection of ground buildings (structures) and the safety of tunnel construction. According to the Technical Specification for Monitoring and Measurement of Railway Tunnels (Q/CR 9218-2015) [19] stipulate the important and selective items for monitoring and measurement of railway tunnel construction, as shown in Table 1.

To sum up, with reference to the scale, characteristics and design requirements of this project, the monitoring content and frequency are shown in Table 2.

2.3 Point Layout

2.3.1 Point Layout Principles

(1)The measurement points are arranged in accordance with the construction monitoring measurement plan. If the actual terrain does not allow, the measuring points can be set near the design points, but the surface points, the convergence points and the settlement points of the vault should corresponding to the mileage of the same section, which should be stable, obvious and simple, which can reflect the actual state and development trend of the monitoring object;

(2)The measuring points should be arranged a certain time in advance, and the initial value should be read early;

(3)Arrange the points for verifying the design parameters at the most unfavorable position of the section and the design. the point is set up for guiding the construction to be arranged at the first construction position under the same working condition, which purpose is to feedback information in time to facilitate the modification of the design and guide the construction;

(4)In the course of construction, the measuring points should avoid the influence of the construction. Once damaged, the measuring point should be added as soon as possible in the original position or as close as possible to the original position to ensure the continuity of the monitoring data of the measuring point;

(5)Three points should be taken into consideration for the location of the measuring points of surface deformation and building settlement: firstly, it can fully reflect the deformation characteristics of the target object; secondly, it is easy to use instruments to monitor and measure; thirdly, it is necessary to ensure that the measuring points are not easily damaged as much as possible, the measuring points of surrounding buildings are mainly arranged at the corners;

(6)Select representative crack monitoring, with at least 2 measuring points for each crack.

2.3.2 Layout of Measuring Points

According to the principle of measuring point arrangement, the arrangement of measuring points for crown settlement and clearance convergence of this project is shown in Fig. 2, in which measuring points 1 and 4 are used to measure crown settlement, and the measuring lines A, B, C are used to measure the horizontal convergence of measuring clearance.

Figure 2: Point layout

Embedment of surface settlement measurement points: Before the tunnel excavation, the surface settlement observation points are buried in time. The specific measurement point arrangement rules are: the ground surface is set up a section every 20 meters along the longitudinal direction, each section is arranged for 10 measurement points, which are set on the tunnel axis and both sides and can be properly encrypted according to the needs. When burying, 20∼50 cm deep hole is drilled on the surface by using the core-drilling machine, and the steel bar Φ 22∼25 is vertically placed, and cement mortar can be filled between the steel bar and the hole wall, and the steel bars should be about 1 cm above the ground. As shown in Figs. 3a, 3b.

Figure 3: Survey points of surface subsidence

2.4 Setting of Alarm Value for Tunnel Construction

For the setting of the alarm value, according to the “Technical Regulations for Monitoring and Measurement of Railway Tunnels” (QCR 9218-2015) and the research results of Zhang et al. [20], combined with the actual situation of the site, he modified control value of the surrounding rock deformation is obtained, as shown in Table 3.

2.5 Safeguard Measures

(1)   Organizational preparation. Organizational preparation refers to the organizational safeguard activities for the implementation of risk early warning activities, which includes not only aquatinting emergency countermeasures for construction risk accidents, but also establishing and improving corresponding safety rules, regulations and standards to provide a secure organizational environment for the entire early warning management.

(2)   Daily monitoring. The management activity of special third-party personnel monitoring the indicator system determined by the early warning analysis activity, which is the core of the early warning activities. These indicators are dangerous. Once a safety risk accident is induced, other disasters associated with it are likely to quickly cause uncontrollable dangerous situations, causing huge casualties and property losses. In addition, In addition, in daily monitoring, it is necessary to predict the possible outcome of the crisis and the severity of the accident after the spread of the accident as much as possible, so as to prevent problems before they occur.

(3)   Emergency plan. The emergency plan mainly involves the emergency leading group, emergency plan, special rescue plan and emergency response measures, etc., which are the key to early warning activities. Once a risk accident occurs during tunnel construction and can be restored to a controllable state as soon as possible, the task of emergency management plan will be declared to be accomplished.

3  Construction Risk Evaluation

3.1 Risk Evaluation Index System

Based on the Technical Specification for Monitoring and Measurement of Railway Tunnels, and combined with the construction monitoring and measurement design plan of this project, the construction risk evaluation index system of CD-Bench composite method based on the monitoring method rely on the Changping section tunnel of Guanhui Intercity Subwayrisk warning indicator system for the CD-bench composite construction method in the Changping section of the Guanhui intercity is equationted, it contains the following five first-level indicators: geological and supporting condition, crown settlement, clearance convergence, ground settlement and building settlement, The complete index system is shown in Fig. 4. From Fig. 4, the following observations can be stated as:

(1)   The geological and supporting conditions will seriously affect the construction safety. By monitoring the risk factors such as groundwater, initial branch cracks, surrounding rock characteristics, state of seepage and fault fracture zone, and their quantitative values are determined according to the quantitative rules of qualitative indicators, in order to the risk level of geological and supporting condition can be evaluated.

(2)   Crown settlement, refers to the absolute settlement (quantity) of the tunnel crown measuring point, is a key indicator that characterizes the level of risk during the construction phase. Through the measurement of the absolute settlement and settlement rate of the vault measurement points, it will pave the way for the overall risk assessment.

(3)   The change of the relative position between two points in the tunnel is characterized by the clearance convergence value. Therefore, by measuring the absolute convergence value and convergence rate of this indicator, which also is a key indicator to evaluate the risk status.

(4)   As urban subway tunnel construction generally causes surface settlement, the measurement of the absolute settlement and settlement rate of the surface measurement points, which can reflect the level of surrounding rock, the quality of the construction level, and the magnitude of the risk to a certain extent.

(5)   During underground construction, it is inevitable that the displacement and settlement of the building (construction) within a certain range, which may cause danger to existing buildings (structures), so by measuring their absolute settlement and settlement rate can also reflect the risks in the construction phase.

Figure 4: Construction risk-warning index system of CD-bench composite method

3.2 Extension Theory

Researcher Cai Wen, a Chinese scholar, created a new subject—Extension, which assumes that the name of the thing is denoted as N, the feature is denoted as C, and the value is denoted as V, and the ordered triple group R = (N, C, V) is used to describe the basic matter element of things, the specific calculation steps [21] are as follows.

3.2.1 Determine of Classical Domain

According to the extension theory, classical domain is expressed as in Eq. (1):

Rj=(Nj,Ci,Vij)=[NjC1v1jC2v2j⋮⋮Cnvnj]=[N0jC1(a1j,b1j)C2(a2j,b2j)⋮⋮Cn(anj,bnj)] (1)

where: Rj represents an matter element, Nj represents the jth evaluation index, namely the event described in the classical domain, Ci represents the i th discriminant index, and Vij = (aij, bij) represents the classical domain, that is, the range of values specified by the category Nj on the discriminant index Ci.

3.2.2 Determine the Controlled Domain

The controlled domain is expressed as in Eq. (2):

Rp=(P,Ci,Vpi)=[PC1vp1C2vp2⋮⋮Cnvpn]=[PC1(ap1,bp1)C2(ap2,bp2)⋮⋮Cn(apn,bpn)] (2)

where: P represents the whole of the evaluation category; Vpi represents the controlled domain of P, that is, the range of all values of P with respect to its discriminant index Ci.

3.2.3 Determine of Matter Element to be Evaluated

For things T to be evaluated, the collected data is expressed in the form of objects, that is, the matter element to be evaluated R, which is expressed as:

R=(T,Ci,Xi)=[TC1X1C2X2⋮⋮CnXn] (3)

where: The Xi is the value of the evaluation index T on the discrimination index the Ci, that is, the actual data of the matter element to be evaluated.

3.2.4 Determination of Correlation Degree

The correlation degree of the i-th single evaluation index to the j-th category level of the matter element to be evaluated can be expressed as:

Kj(Xi)={ −ρ(Xi, Vij)|Vij|, Xi∈Vijρ(Xi, Vij)ρ(Xi, Vpi)−ρ(Xi, Vij), Xi∉Vij(4)

where:

ρ(Xi,Vij)=|xi−aij+bij2|−bij−aij2;ρ(Xi,Vpi)=|xi−api+bpi2|−bpi−api2 (5)

3.3 Determination of Weights

The scientific level of weight directly affects the objectivity of the calculation results of the whole evaluation model. At present, the method of subjective and objective combination weighting is channeled into solving the weight problem, which leads to the favorable results have been achieved in practical applications. Therefore, this paper uses entropy method and improved analytic hierarchy process to determine the subjective and objective weight separately, and then introduce the distance function to combine the subjective and objective weights, and then determine the weight value of each evaluation index.

3.3.1 Entropy Method

Set m evaluation grade and n evaluation index to form a matrix R = (Kj(Xi))m×n, where the Kj(Xi) is the correlation degree of the i-th evaluation index under the j-th evaluation grade, and the weight of each index as follows:

wi=1−k∑j=1mPj(Xi)⋅lnPj(Xi)∑(1−k∑i=1mPj(Xi)⋅lnPj(Xi)) (6)

where:

k=1ln⁡m;Pj(Xi)=1−Kj(Xi)∑j=1m(1+Kj(Xi)) (7)

3.3.2 Improved Analytic Hierarchy Process (IAHP)

In order to facilitate the comparison of indicators, this paper selects the improved three-scale analytic hierarchical process [22].

(1)   Establishment of a cluster judgment matrix

Based on the construction risk-warning index system of CD-Bench composite method which has been established, First, according to the “0”, “1”, and “2” three-level scale method, the relevant experts are asked to compare the importance of the same level indicators in pairs, and the feedback situation is handled by the cluster method; Secondly, To report back the processed results to the experts and ask the experts to re-mark. Finally, we re-feedback the index importance scale to the experts. After three rounds of consultation, if the opinions of the experts tend to be unified, we will take the last round group judgment matrix as the final result of the evaluation index system and enter the next analysis process.

Where, the cluster judgement matrix is to process the scores of the experts on the same index according to the following mathematical treatment:

xij=H+4m+L6 (8)

where: m is the average value of all experts’ scores for a certain evaluation index; and H is the average value of scores higher than m; and L is the average value of scores lower than m.

Finally, after mathematical processing, the group judgment matrix A can be obtained:

AC1C2C3CmC1x11x12⋯x1mC2x21x22⋯x2m⋮⋮⋮⋮⋮Cmxm1xm2⋯xmm (9)

(2)   Converting comparison matrix A into judgement matrix D

dij=9(ri−rj)/R (10)

R=max{r1, r2, …, rn}−min{r1, r2…rn}(11)

ri=∑j=1nxiji=1,2,…,n (12)

where: the ri is the importance coefficient of each element of the matrix A.

(3)   Calculate the arithmetic mean of all elements in each row of the judgment matrix D and normalize to obtain the weight

wi¯=1m∑j=1mdij (13)

wi=wi¯∑i=1mwi¯ (14)

(4)   Consistency testing

1)

Calculate the consistency index CI:

CI=λmax−mm−1 (15)

λmax=∑i=1m(Dw)imwi (16)

where: (Dw)i is the ith index of vector DW, and m is the order of the judgment matrix.

2)   Find the corresponding average random consistency index RI (Table 4)

3)   Calculation consistency ratio

CR=CIRI (17)

When CR=x< 0.10 , the judgment matrix has satisfactory consistency; Otherwise, the judgment matrix needs to be re-calculated until the consistency requirement is met.

3.3.3 Combined Weight

Suppose the subjective weight is wsj, the objective weight is woj, the combined weight is wzj, its expression [23] as:

wzj=αwsj+βwoj (18)

where: α,  β represent the distribution coefficients of the two weights, respectively, which can be obtained from Eq. (19):

{d(wsj,woj)=[12∑j=1n(wsj−woj)2]12d(wsj,woj)2=(α−β)2α+β=1 (19)

3.4 Comprehensive Evaluation

3.4.1 Membership Matrix of First-Degree Indicators

By multiplying the comprehensive weight vector wij of the second level index with the correlation degree Kj(Xi) of each risk grade, and it can be obtained the membership degree matrix Ak of the k th first-degree index for each risk grade, as shown in Eq. (20):

Akj=∑i=1nwijKj(Xi) (20)

where

∑i=1nwij=1 (21)

3.4.2 Membership Matrix of Target Layer

Same as solving the first-level index risk membership matrix, the membership matrix of the target layer is shown as below:

Aj=∑wk⋅Akj,∑k=1nwk=1 (22)

3.4.3 Identification of Risk Levels

If Kj = max j∈(1,2,⋅⋅⋅,m)Kj(T), then the warning level of the warning object is j, So Then, let j be the characteristic value of the risk level, as shown in the following equation:

Kj∗(T)=Kj(T)−minj∈(1,2,⋯,m)Kj(T)maxj∈(1,2,⋯,m)Kj(T)−minj∈(1,2,⋯,m)Kj(T) (23)

j∗=∑j=1mj⋅Kj∗(T)∑j=1mKj∗(T) (24)

4  Information Feedback of Early-Warning

4.1 Classification of Risk Levels

According to the principle of setting the alarm level and the construction risk characteristics of this type of weak surrounding rock section, the construction risk of this project is divided into five levels, which are shown in Table 5.

4.2 Five-Color Risk-Warning System

The five-color method is applied to set the alarm signal [24], as shown in Table 6.

4.3 Mapping Relationship

Let the value range of the dynamic risk index and risk-warning signal be the Ux and Uy respectively, f(x) represents a mapping from Ux to Uy, recorded as Y = f (x), and the mapping relationship between dynamic risk index and risk-warning signal is shown in Eq. (25).

5  Case Analysis

5.1 Criteria of Construction Risk Assessment

Refer to the research results of Zhang et al. [20], Liu et al. [25] and others, as well as the Technical Specification for Monitoring and Measurement of Railway Tunnels, and in order to facilitate the evaluation, the indicators are dimensionless processed: Crown settlement, clearance convergence, ground settlement, building settlement on the basis of domain measured data, and the cumulative value is the ratio of the cumulative measured value to the allowable value, and the rate is the ratio of the rate change value to the allowable value. The specific risk evaluation standards for each indicator are shown in Table 7. Then According to the actual situation of the project and consulting relevant experts with Site construction management personnel, the evaluation value of each indicator is shown in the last column of Table 7.

UxUy([4,5][3,4)[2,3)[1,2)[0,1))⟶f(x)(Red pre−warningOrange pre−warningYellow pre−warningBlue pre−warningGreen pre−warning) (25)

5.2 Determine the Extension Evaluation Matrix

(1)   Taking the secondary index of the A1 as an example, the classical domain R1-R5, the controlled domain RP, and the matter element to be evaluated Xi be obtained by Eqs. (1) and (3), as shown in the following matrix:

[CiR1R2R3R4R5RPXiB1(0,0.15)(0.15,0.35)(0.35,0.55)(0.55,0.75)(0.75,1)(0,1)0.32B2(0,0.15)(0.15,0.35)(0.35,0.55)(0.55,0.75)(0.75,1)(0,1)0.34B3(0,0.15)(0.15,0.35)(0.35,0.55)(0.55,0.75)(0.75,1)(0,1)0.17B4(0,0.15)(0.15,0.35)(0.35,0.55)(0.55,0.75)(0.75,1)(0,1)0.30B5(0,0.15)(0.15,0.35)(0.35,0.55)0.55,0.75(0.75,1)(0,1)0.14] (26)

(2)   Determine the degree of risk association

According to Eq. (4), the risk correlation degree of index B1-B5 is shown in matrix A1:

A1=[−0.350.10−0.09−0.42−0.57−0.360.03−0.03−0.38−0.55−0.110.13−0.51−0.69−0.77−0.330.20−0.14−0.45−0.600.08−0.07−0.60−0.75−0.81] (27)

5.3 Determine the Index Weight

5.3.1 Determination of Objective Weight Based on Entropy Weight

It can be obtain according to Eq. (5):

m = 5, k = 0.621, R = A1 (as above)

Then the objective weight of B1–B5 is calculated as:

W1 = 0.20…W5 = 0.19. The specific calculation results are shown in Table 8.

5.3.2 Determination of Subjective Weights Based on IAHP

A=[A1B1B2B3B4B5B112222B201022B302122B400011B500011] (28)

Take the first-level indicator layer as an example, according to Eqs. (8)–(17), its comparison matrix A is shown above, and the judgment matrix D is as follows:

D=[A1B1B2B3B4B5wi¯wiB113.5101.873994.8770.49B20.28510.5342.5642.5641.3890.14B30.5341.87314.8044.8042.6030.26B40.1110.3900.208110.5420.05B50.1110.3900.208110.5420.06] (29)

Find out λmax = 5.461, CI = 0.115, CR = 0.100, so it meets the consistency check requirements.

5.3.3 Determining the Combination Weight

The obtained objective weight and subjective weight are brought into Eq. (19), get α = 0.523,  β = 0.477, and then take these two combination coefficient into Eq. (18) to obtain the combined weight of each indicator, as shown in Table 8.

It can be observed from Table 8 that the weight values of the five first-order indicators affecting construction risk of CD-Bench composite method are as follows: geological and supporting condition observation > clearance convergence > Crown settlement > building settlement > ground settlement, and the contribution to the overall risk decreases sequentially. In the secondary indicators, the weight values of groundwater, initial bracing crack, wall rock (soil) Characteristics, state of seepage and fault fracture zone are approximately equal, which indicates that their contribution to the risk of its upper level indicator is equal. And their second-level index, the weight value of rate value is much larger than the accumulative value, which indicates that it is essential to increase the monitoring frequency appropriately. It also show that the calculation results of weight value is consistent with objective reality.

5.4 Identification of Risk Grade

5.4.1 Determine the Membership Degree of the First-Level Indicators

According to Eq. (20), the membership matrix is:

A1=∑i=15W1jKj(Xi)=(0.20,0.21,0.22,0.17,0.20)⋅[−0.350.1−0.09−0.42−0.57−0.360.03−0.03−0.38−0.55−0.110.13−0.51−0.69−0.77−0.330.20−0.14−0.45−0.600.08−0.07−0.60−0.75−0.81]=(−0.2099,0.0749,−0.2803,−0.5421,−0.6629) (30)

The membership degree of other first-level indicators can be obtained in the same method.

5.4.2 Determine the Membership Degree of the Target Layer

According to Eq. (22), the membership matrix of the target layer is:

A=(0.46,0.17,0.24,0.06,0.08)⋅[−0.20990.0749−0.2803−0.5421−0.6629−0.4150−0.26300.0090−0.0030−0.3710−0.4280−0.23400.1540−0.0964−0.4424−0.4396−0.2144−0.0120−0.0236−0.4608−0.26680.1304−0.1900−0.4260−0.6184]=(−0.3175,0.0688,−0.1064,−0.3085,−0.5513) (31)

Get: K1(T) = −0.3175; K2(T) = −0.0688; K3(T) = −0.1064; K4(T) = −0.3085; K5(T) = −0.5513.

5.4.3 Identification of Risk Grade

It can be get accord to Eq. (23):

K1∗(T)=−0.3175−(−0.5513)−0.0688−(−0.5513)=0.4846K2∗(T)=1.0000;K3∗(T)=0.9221;K4∗(T)=0.5032;K5∗(T)=0; (32)

The risk grade characteristic value ( j∗ ):

j∗=1×0.4846+2×1+3×0.9221+4×0.5032+5×00.4846+1+0.9221+0.5032+0=2.50 (33)

By substituting the daily measurement data into the construction risk-warning model established in this paper, it can be getting that risk grade characteristic value which reflects the construction risks at this stage was 2.50. Thus the risk-warning signal of this evaluation is “yellow”, namely the construction risk is level 3, indicating that the phenomenon of local collapse and water and mud inrush have occurred at this time, which will affect the tunnel construction, but will not cause major accidents. Among the crown settlement rate, clearance convergence accumulation and ground subsidence accumulation are the higher risk sources, which are classified grade 4. And the vault settlement accumulation, clearance convergence rate, ground subsidence rate and building settlement accumulation are medium risks, which are classified grade 3. Therefore, the construction party should take corresponding emergency measures and safety schemes based on the above judgment.

5.5 Control Measures

According to the risk-warning results, the relevant units negotiated to change the original design and construction plan, as follows:

(1)   Long anchors are added to the crown. Due to the relatively complex construction force at the crown, and the phenomenon of deformation exceeding the alarm value is particularly frequent. In order to control the further development of the risk unit at this location, it is recommended to add long anchors to the top of the tunnel arch in time.

(2)   Increase the amount of deformation reserved. On the one hand, it can offset the large displacement generated by the initial support, release the in-situ stress to a large extent, and effectively prevent the problem of the exceeding the limit of initial support after the large deformation. On the other hand, it can reduce the load acting on the secondary lining, which is beneficial to the safety and stability of the tunnel structure. Therefore, it is recommended to adjust the reserved deformation of individual dangerous areas to 30–40 cm.

(3)   Improve the rigidity of the supporting structure. According to the characteristics of deeper plastic zone and large damage range of soft rock tunnel, Measures should be taken to enlarge the steel frame, such as: the enlarged steel frame is adopted, the I-beam is adjusted from I20b to H175, longitudinal spacing is adjusted from the current 0.6 m to 0.5 m, and the large stiffness support system is sprayed with 30 cm-thick C25 concrete.

6  Conclusions

This paper relies on the Changping section tunnel of Guanhui Intercity Metro, a dynamic risk-warning model is established for the construction process of subway tunnel with the CD-Bench composite method. The following conclusions can be drawn:

(1)   Due to the particularity and complexity of subway tunnel construction, most of the index systems established in previous studies are qualitative indexes and difficult to quantify. In the actual construction process, the deformation value can reflect the construction risk simply and directly. Therefore, based on this feature, this article equationtes a monitoring plan based on the monitoring specifications and the project characteristics, and establishes a risk evaluation index system on this basis, avoiding the difficulty of quantifying qualitative indicators. In addition, assigning different combination weights to each indicator and the mathematical model is used to calculate the risk value, which improves the one-sidedness and irrationality of evaluating construction risks based on deformation alone, and promotes scientific research on dynamic risk control.

(2)   Among the four indexes of crown settlement, clearance convergence, ground settlement, and building settlement, clearance convergence has the greatest impact on construction safety risks. For their two secondary indicators, the rate of change is much greater than the cumulative change value contributing to the risk value of the upper indicator. Therefore, in actual construction, it is necessary to appropriately increase the frequency of monitoring for poor geological conditions or abnormalities in monitoring data, and focus on the settlement value of clearance convergence.

(3)   Through analysis and calculation, the initial construction risk characteristic value of this stage is 2.50, and the early warning signal is yellow, indicating that a major accident may occur at present, and control or rectification measures must be taken. For this reason, the construction unit recommends measures such as long anchors are added to the vault, increase the amount of deformation reserved and the stiffness of supporting structures and so on. After the calculation is performed again, the warning signal turns green, showing that the construction risk status at this stage is safe after taking these measures, indicating that these technologies can reduce the construction risk value and provide technical reference for similar projects.

In the future, we will continue to explore the applicability of dynamic risk-warning; the comparative analysis will also be conducted on different theoretical methods to achieve higher prediction accuracy.

Funding Statement: The work is support by the National Natural Science Foundation of China: 51668037.

Conflicts of Interest: The authors declare that they have no conflicts of interest to report regarding the present study.

References

 1.  Wang, Y. C., Yin, X., Jing, H. W., Liu, R., Sul, H. (2016). A novel cloud model for risk analysis of water inrush in karst tunnels. Environmental Earth Sciences, 75(22), 1–13. DOI 10.1007/s12665-016-6260-7. [Google Scholar] [CrossRef]

 2.  Yuan, Y. C., Li, Z. C., Li, L. P., Lei, T., Wang, S. et al. (2016). Evaluation theory and methods of collapse risk in shanling tunnel. Journal of Central South University (Natural Science Edition), 47(7), 2406–2414. DOI 10.11817/j.issn.1672-7207.2016.07.031. [Google Scholar] [CrossRef]

 3.  Alireza, A., Jafar, S. S., Meisam, J. E. (2011). Design a new intelligence expert decision making using game theory and fuzzy AHP to risk management in design, construction, and operation of tunnel projects (case studies: Resalat tunnel). International Journal of Advanced Manufacturing Technology, 53(5–8), 705–715. DOI 10.1007/s00170-010-2852-7. [Google Scholar] [CrossRef]

 4.  Sousa, R. L., Einstein, H. H. (2011). Risk analysis during tunnel construction using Bayesian networks: Porto metro case study. Tunnelling and Underground Space Technology, 27(1), 86–100. DOI 10.1016/j.tust.2011.07.003. [Google Scholar] [CrossRef]

 5.  Ye, X. W., Ran, L., Dong, X. B. (2012). Intelligent risk assessment for ddewatering of metro-tunnel deep excavations. Mathematical Problems in Engineering, 2012, 618979. DOI 10.1155/2012/618979. [Google Scholar] [CrossRef]

 6.  Abdolreza, Y. C. (2014). Proposing a new methodology based on fuzzy logic for tunnelling risk assessment. Journal of Civil Engineering and Management, 20(1), 82–94. DOI 10.3846/13923730.2013.843583. [Google Scholar] [CrossRef]

 7.  Bai, Y., Dai, Z. R., Zhu, W. J. (2014). Multiphase risk-management method and its application in tunnel engineering. Natural Hazards Review, 15(2), 140–149. DOI 10.1061/(ASCE)NH.1527-6996.0000124. [Google Scholar] [CrossRef]

 8.  Cagatay, P. (2015). Analysis and management of risks experienced in tunnel construction. Acta Montanistica Slovaca, 20(4), 271–281. DOI 10.3390/ams20040271. [Google Scholar] [CrossRef]

 9.  Han, J. Y., Zhao, W., Jia, P. J., Guan, Y. P., Chen, Y. et al. (2018). Risk analysis of the opening of shield-tunnel circumferential joints induced by adjacent deep excavation. Journal of Performance of Constructed Facilities, 32(1), 04017123. DOI 10.1061/(ASCE)CF.1943-5509.0001122. [Google Scholar] [CrossRef]

10. Xia, Y. P., Xiong, Z. M., Wen, Z., Ma, C. (2018). Fuzzy risk assessment of a deeply buried tunnel under incomplete information. Royal Society Open Science, 5(10), 180305. DOI 10.1098/rsos.180305. [Google Scholar] [CrossRef]

11. Lin, C. J., Zhang, M., Li, L. P., Zhou, Z. Q., Liu, S. et al. (2020). Risk assessment of tunnel construction based on improved cloud model. Journal of Performance of Constructed Facilities, 34(3), 04020028. DOI10.1061/(ASCE)CFC1943-5509.0001421. [Google Scholar] [CrossRef]

12. Xue, Y. G., Li, Z. Q., Qiu, D. H., Yang, W. M., Zhang, L. W. et al. (2019). Prediction model for subway tunnel collapse risk based on delphi-ideal point method and geological forecast. Soil Mechanics and Foundation Engineering, 56(3), 191–199. DOI 10.1007/s11204-019-09589-4. [Google Scholar] [CrossRef]

13. Mahdi, H., Majid, N. B., Jahed, A. D. (2016). Feasibility of PSO-ANN model for predicting surface settlement caused by tunneling. Engineering with Computers, 32(4), 705–715. DOI 10.1007/s00366-016-0447-0. [Google Scholar] [CrossRef]

14. Su, J., Zhang, D. L., Zhou, Z. Y., Niu, X. K., Tai, Q. M. (2015). Assessment and control of safety risk of subway tunnel crossing existing bridges. Journal of Rock Mechanics and Engineering, 34(S1), 3188–3195. DOI 10.13722/j.cnki.jrme.2014.0230. [Google Scholar] [CrossRef]

15. Liu, G. F., Feng, X. T., Feng, G. L., Chen, B. R., Chen, D. F. et al. (2016). A method for dynamic risk assessment and management of rockbursts in drill and blast tunnels. Rock Mechanics and Rock Engineering, 49(8), 3257–3279. DOI 10.1007/s00603-016-0949-5. [Google Scholar] [CrossRef]

16. Liu, W. L., Cai, L. X., Chen, J., Wang, Y. Y., Wu, H. (2020). Reliability analysis of operational metro tunnel based on a dynamic Bayesian copula model. Journal of Computing in Civil Engineering, 34(3), 05020002. DOI 10.1061/(ASCE)CP.1943-5487.0000886. [Google Scholar] [CrossRef]

17. Li, X. Y., Zhang, L. M., Zhang, S. (2018). Efficient Bayesian networks for slope safety evaluation with large quantity monitoring information. Geoscience Frontiers, 9(6), 1679–1687. DOI 10.1016/j.gsf.2017.09.009. [Google Scholar] [CrossRef]

18. Kang, X. B., Luo, S., Xu, M., Zhang, Q., Yang, Y. N. et al. (2019). Dynamic estimating the karst tunnel water inrush based on monitoring date during excavation. Acta Carsologica, 48(1), 117–127. DOI 10.3986/ac.v48i1.4654. [Google Scholar] [CrossRef]

19. China Railway Corporation (2014). Technical specification for monitoring and measurement of railway tunnels (Q/CR 9218-2015). [Google Scholar]

20. Zhang, M. Q., Huang, H. J., Wu, C. (2014). A study and application of the deformation safety grade of surrounding rock in railway tunnel construction. Journal of Railway Engineering, (11), 87–93. [Google Scholar]

21. Guo, Z. W. (2020). Evaluation of financial ability of port listed companies based on entropy weight TOPSIS model. Journal of Coastal Research, 103(sp1), 182–185. DOI 10.2112/SI103-039.1. [Google Scholar] [CrossRef]

22. Wu, Y. N., Bian, Q. (2013). The research on bidding decision-making for electric power construction enterprises—based on the improved grey minkowski-TOPSIS model. Journal of Technical Economics & Management, 27(4), 12–16. DOI CNKI:SUN:JXJG.0.2013-04-003. [Google Scholar]

23. Tang, H., Duan, Z., Tang, S. L., Wang, D. P., Zeng, P. (2019). Combined weight cloud model of rock mass quality classification. Journal of Xi’an University of Science and Technology, 39(1), 79–87. DOI 10.13800/j.cnki.xakjdxxb.2019.0112. [Google Scholar] [CrossRef]

24. Wu, X. G., Chen, Y. Q., Zhang, L. M., Yao, C. J. (2013). A study on early warning management and evaluation standard for construction safety of subway engineering. Journal of Railway Engineering, 30(5), 107–111. [Google Scholar]

25. Liu, J., Ai, Z. Y., Su, H. (2012). Dynamic risk assessment of mountain tunnel during NATM construction. Journal of Tongji University (Natural Science Edition), 40(8), 1142–1146. DOI 10.3969/j.issn.0253-374x.2012.08.004. [Google Scholar] [CrossRef]

This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.