CN104636612A - Karst tunnel water outburst and mud outburst overall process gradual dynamic risk assessment method - Google Patents

Abstract

The invention discloses a karst tunnel water outburst and mud outburst overall process gradual dynamic risk assessment method. The method includes the steps that 1, in the tunnel investigation stage, the tunnel address and the hydrogeological information of the surrounding rock near the tunnel are obtained, that is, the hidden danger environment for water outburst of the tunnel is obtained, and the risk states of the geological conditions of all segments of the tunnel are known; 2, the risk rating value is worked out according to an expert evaluation vector and a factor weight vector, and consistency check is conducted; 3, danger factors are introduced to influence factors of risk evaluation, the hidden danger environment and the danger factors are considered comprehensively, water outburst risk evaluation is conducted, and the segment distribution feature with the tunnel water outburst risk is divided; 4. The values of all indexes are corrected in real time in combination with actual field construction so as to achieve dynamic evaluation of the water outburst and mud outburst risks. The purposes of optimizing construction organization design and avoiding medium and large water outburst geological disasters are achieved through dynamic construction risk evaluation and control, and important data are provided for later-period tunnel operation risks, and the method has very wide application prospects.

Description

Karst Tunnel gushing water is dashed forward the gradual risk dynamic assessment method of mud overall process

Technical field

The present invention relates to a kind of Karst Tunnel gushing water to dash forward the gradual risk dynamic assessment method of mud overall process.

Background technology

Tunnels and underground engineering relates to traffic engineering (railway, vcehicular tunnel), Hydraulic and Hydro-Power Engineering (water-conveyance tunnel, underground power house) etc. important engineering field, become the important component part that national major infrastructure project is built gradually, and along with the planning of country's strategic development in science and technology such as " 12 ", the center of gravity of key project construction is just gradually to topographic and geologic extreme complicated western mountainous areas transfer, the karst faced in process of construction water burst of dashing forward is one of Geological Hazards in the tunnel construction of karst area, thus be the key of analysis and solutions problem for the dash forward risk assessment of mud overall process of tunnel, karst area gushing water.

But, domestic also not too perfect about research in this respect at present, in risk assessment processes, often have ignored the complicacy of underground works geologic condition, and most of appraisal procedure can only provide qualitative or semiquantitative analysis result, the value of each index cannot be corrected in real time in conjunction with on-the-spot practice of construction situation, to realize the dynamic evaluation of water bursting factor risk.

Summary of the invention

Object of the present invention is exactly to solve the problem, a kind of Karst Tunnel gushing water is provided to dash forward the gradual risk dynamic assessment method of mud overall process, the value that it has risk dynamic evaluation index is corrected in real time in conjunction with on-the-spot practice of construction situation, can realize the advantage of the dynamic evaluation of water bursting factor risk.

To achieve these goals, the present invention adopts following technical scheme:

A kind of Karst Tunnel gushing water is dashed forward the gradual risk dynamic assessment method of mud overall process, and step is as follows:

, there is from tunnel the hydrogeological information obtaining tunnel location and neighbouring country rock thereof the pregnant dangerous environment of prominent water burst in step (1): in the tunnel surveying stage; The hydrogeological information of described tunnel location and neighbouring country rock thereof comprise unfavorable geology, formation lithology, underground water table, can lava and non-can lava contact zones, topography and geomorphology, the attitude of rocks, aspect and crack in layer and Grades of Surrounding Rock;

Step (2): obtain target layers table according to step (1), according to target layers table and factor weight vector, calculates risk stratification value, and carries out consistency check;

Step (3): risk factors is introduced in the influence factor of risk assessment, consider the tunnel location of step (1) and the hydrogeological information of neighbouring country rock thereof and risk factors, carry out the estimation of prominent water burst risk, divide tunnel and to dash forward the section distribution characteristics of water burst risk;

Step (4): hydrogeological information step (1) obtained is corrected in real time in conjunction with on-the-spot practice of construction situation, to realize the dynamic evaluation of water bursting factor risk.

The risk factors of described step (3) refers to the direct factor causing analysis of possibility of water inrush to occur, as arrangement and method for construction, operating technique and personnel activity.

Described step (2) according to target layers table and factor weight vector, calculate risk stratification value, calculating risk stratification value step is:

B＝[b ₁b ₂…b _i…b _n] (1)

ω＝[ω ₁ω ₂…ω _i…ω _n] ^T(2)

S＝B·ω (3)

Wherein, B is expert analysis mode vector, and ω is factor weight vector, and S is risk stratification value.

The step of described step (2) is as follows:

Step (2-1): build target layers table according to hydrogeological information;

Step (2-2): utilize 1-9 scaling law Judgement Matricies;

Step (2-3): utilize the judgment matrix of step (2-2) to calculate weighted value;

Step (2-4): utilize the weighted value of step (2-3) to calculate the eigenvalue of maximum λ of judgment matrix _max:

${\mathrm{\λ}}_{\max }=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{{\left(\mathrm{AW}\right)}_i}{n{W}_i}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{a}_{\mathrm{ij}}{W}_j}{W_i}---\left(7\right)$

A _ijmeaning of parameters be each element in judgment matrix.

Step (2-5): consistency check:

Whether the weighted value eventually through each factor calculated can reflect its importance truly objectively, is decided by the authenticity of judgment matrix; Therefore, the Conformance Assessment of satisfaction must be carried out to judgment matrix:

$\mathrm{CI}=\frac{{\mathrm{\λ}}_{\max }-n}{n-1}---\left(8\right)$

$\mathrm{CR}=\frac{\mathrm{CI}}{\mathrm{RI}}---\left(9\right)$

In formula: CI: coincident indicator; CR: Consistency Ratio; RI: Aver-age Random Consistency Index, is checked in by table 3.

Table 3

Because of subnumber (n)	1	2	3	4	5	6	7	8	9
										Corresponding value	0.00	0.00	0.58	0.90	1.12	1.24	1.32	1.41	1.45

If Consistency Ratio CR < 0.1, judgment matrix meets consistance, otherwise, need adjust judgment matrix, till judgment matrix meets consistance.

The target layers table of described step (2-1) is in table 1:

Table 1 target layers table

The judgment matrix of described step (2-2), as shown in table 2:

Table 2 judgment matrix

The step of described step (2-3) is as follows:

Step (2-3-1): the M calculating each row element of judgment matrix _i

$M_i=\underset{i=1}{\overset{n}{\mathrm{\&Pi;}}}{a}_{\mathrm{ij}}(i,j=\mathrm{1,2},...,n)---\left(4\right)$

Step (2-3-2): calculate M _in root

$W_i=\sqrt[n]{M}---\left(5\right)$

Step (2-3-3): to vector normalized

$W=\frac{\stackrel{\&OverBar;}{W_i}}{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{W_i}}(i=\mathrm{1,2,3},...,n)---\left(6\right)$

The ω obtained=[ω ₁ω ₂ω _iω _n] ^tbe required factor weight vector.

The step that risk factors is introduced in the influence factor of risk assessment by described step (3) is:

In hydrogeological information, if descend water level to be quantitative description only, adopt the method for structure membership function to determine the degree of membership of each grade, the whether objective validity determining evaluation result of membership function, is constructed as follows function:

${\mathrm{\&mu;}}_1=\left\{\begin{array}{c}1(x>x_1)\\ \frac{x-x_2}{x_1-x_2}(x_2<x<x_1)\\ 0(x<x_2)\end{array}\right.---\left(10\right)$

${\mathrm{\&mu;}}_3=\left\{\begin{array}{c}0(x>x_2)\\ \frac{x_2-x}{x_2-x_3}(x_3<x<x_2)\\ \frac{a}{x_2-x}(x<x_3)\end{array}\right.---\left(12\right)$

${\mathrm{\&mu;}}_4=\left\{\begin{array}{c}0(x>x_3)\\ \frac{x_2-x}{x_2-x_3}(x_3<x<x_2)\\ \frac{x_2-x-a}{x_2-x}(x<x_3)\end{array}\right.---\left(13\right)$

In formula (10) ~ (13), μ ₁, μ ₂, μ ₃, μ ₄the subordinate function of underground water four risk levels respectively, and x ₁get 60, x ₂get 30, x ₃get 0 and a get 10;

Can not the analysis of possibility of water inrush factor of quantificational description for other, adopt Karwowski algorithm to propose membership function, in table 4:

Table 4

Each factor has k grade, and each factor grade has corresponding impact for the evaluation index evaluating collection, and its influence degree degree of membership represents, sets up Evaluations matrix B:

In formula: b _ikrepresent the degree of membership of i-th factor kth grade relative to factor.

The risk estimation steps of described step (3) is:

Combination weights method solves factor weight vector w and to dash forward the objective weight vector P that mud example obtains and the subjective weight vector Q weighted sum obtained by analytical hierarchy process and obtain by adding up typical gushing water:

ω＝P·Ψ _P+Q·Ψ _Q(15)

P＝[p ₁p ₂…p _i…p _n] (16)

Q＝[q ₁q ₂…q _i…q _n] (17)

In formula, P is objective weight vector, Q be subjective weight vector respectively with, Ψ _pfor the distribution weights of objective weight, Ψ _qfor the distribution weights of subjective weight, Ψ _pwith Ψ _qthen obtained by the confidence index method based on expert investigation.

The dash forward step of section distribution characteristics of water burst risk of described division tunnel is:

Step (3a-1): set up set of factors;

With fuzzy language, classification statement is carried out to analysis of possibility of water inrush probability of happening evaluation objective, sets up and evaluate collection:

V＝(v ₁，v ₂，v ₃，v ₄) (18)

In formula: v ₁for primary risk; v ₂for secondary risk; v ₃for tertiary risk; v ₄for level Four risk, gushing water geologic hazard degree weakens successively, each element v _irepresenting various possible total evaluation result, namely considering all gushing water influence factors, drawing best evaluation result from evaluating to concentrate.

Step (3a-2): risk assessment and sequence;

Application blurring mapping principle carries out comprehensive evaluation to relevant each factor, obtains the comprehensive evaluation collection P of analysis of possibility of water inrush probability of happening _f:

A is construction factor judgment matrix.A _irepresent each construction factor (advance geologic prediction, monitoring measurement, excavation supporting etc.), B is Evaluations matrix, b _ikrepresent the degree of membership of i-th factor kth grade relative to factor

According to maximum membership grade principle, in formula (20) result of calculation, in last subordinated-degree matrix, overall target υ _lthe degree of membership c of corresponding opinion rating _lsoprano, is assessment of w ater inrush risk grade;

Be formulated as:

${\mathrm{\&upsi;}}_s=\{c_l/{\mathrm{\&upsi;}}_l\&RightArrow;\underset{i=1}{\overset{k}{\max \left(c_i\right)}}\}---\left(20\right);$

C _ifor risk class membership vector, v _sfor gushing water is dashed forward mud risk class, v _lfor overall target.

The step of described step (4) is:

With research object in water bursting factor risk assessment Matter element system for things, prediction index is feature, and standardize criteria value is eigenwert, and water bursting factor risk assessment matter-element following formula describes:

Get a certain forecasting object in water bursting factor risk assessment Matter element system, only predict according to the prediction index that adjacent lower is corresponding, be called point prediction, it is the basis of water bursting factor risk class assessment.

Wherein the process of point prediction is as follows:

Step (4-1): calculate Classical field matter-element, joint territory matter-element and present situation matter-element respectively;

Step (4-2): the matter-element grade degree of association in analytical procedure (4-1);

Step (4-3): utilize the degree of association in step (4-2) to carry out forecasting object classification Comprehensive Evaluation;

Step (4-4): the index weights of forecasting object is analyzed.

The step of described step (4-1) is:

Step (4-1-1): Classical field matter-element

According to prominent water burst risk assessment classification standard, the Classical field matter-element of forecasting object can be obtained:

$R_{0j}=\left[\begin{array}{ccc}{N}_{0j},& c_1,& v_{01j}\\ & c_1,& v_{02j}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{0\mathrm{nj}}\end{array}\right]=\left[\begin{array}{ccc}{N}_{0j},& c_1,& <a_{01j},b_{01j}>\\ & c_2,& <a_{02j},b_{02j}>\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& <a_{0\mathrm{nj}},b_{0\mathrm{nj}}>\end{array}\right]---\left(22\right)$

In formula:

R _0jfor jth is individual with levying matter-element,

N _0jfor forecasting object jth rank (j=1,2 ..., m), m is institute's divided rank number;

C _ifor forecasting object i-th feature (i=1,2 ..., n), n is the characteristic number of forecasting object;

V _0ij=﹤ a _0ij, b _0ij﹥ is rank j i-th feature c _ivalue territory after standardization, the i.e. Classical field got about the index of forecasting object of each rank.

Step (4-1-2): joint territory matter-element

$R_P=\left[\begin{array}{ccc}{N}_p,& c_1,& v_{P1}\\ & c_2,& v_{P2}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{\mathrm{Pn}}\end{array}\right]=\left[\begin{array}{ccc}{N}_P,& c_1,& <a_{P1},b_{P1}>\\ & c_2,& <a_{P2},b_{P2}>\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& <a_{\mathrm{Pn}},b_{\mathrm{Pn}}>\end{array}\right]---\left(23\right)$

In formula:

N _pfor forecasting object rank is all;

V _pi=﹤ a _pi, b _pi﹥ is that P is about feature c _ithe codomain got, i.e. N _pjoint territory.Above formula is designated as R _p(P, C, V _p), because each feature codomain after standardization is ﹤ 0,1 ﹥, so must save territory R _p=[N _p, c _i, < 0,1 >].Obviously have, (i=1,2 ..., n; J=1,2 ..., m).

Step (4-1-3): present situation matter-element

According to actual conditions to forecasting object N _keach eigenwert carry out statistics and analysis, and to process according to given standard, thus obtain present situation matter-element;

$R_k=\left[\begin{array}{ccc}{N}_k,& c_1,& v_{k1}\\ & c_2,& v_{k2}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{\mathrm{kn}}\end{array}\right]---\left(24\right)$

In formula, N _kfor forecasting object; v _kifor i-th feature c of forecasting object _istandardization value.

The step of described step (4-2) is:

Step (4-2-1): the single index degree of association

To water bursting factor risk assessment object N _ki-th feature c _icorrelation function about a jth rank is designated as K _j(v _ki), can be calculated by following formula:

$K_j\left(v_{\mathrm{ki}}\right)=\left\{\begin{array}{c}\frac{-\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})}{\left|v_{0\mathrm{ij}}\right|},v_{\mathrm{ki}}\&Element;v_{0\mathrm{ij}}\\ \frac{\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})}{\mathrm{\&rho;}(v_{\mathrm{ki}},v_{\mathrm{Pi}})-\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})},v_{\mathrm{ki}}\&NotElement;v_{0\mathrm{ij}}\end{array}\right.---\left(25\right)$

$\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})=|v_{\mathrm{ki}}-\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}|-\frac{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}}{2}---\left(26\right)$

|v _0ij|＝|b _0ij-a _0ij| (27)

$\mathrm{\&rho;}(v_{\mathrm{ki}},v_{\mathrm{pi}})=|v_{\mathrm{ki}}-\frac{a_{\mathrm{pi}}+b_{\mathrm{pi}}}{2}|-\frac{b_{\mathrm{pi}}-a_{\mathrm{pi}}}{2}---\left(28\right)$

Step (4-2-2): the comprehensive multi-index degree of association

Water bursting factor risk assessment object N _kthe degree of association about grade j is

$K_j\left(N_k\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{w}_i{K}_j\left(v_{\mathrm{ki}}\right)---\left(29\right)$

In formula: w _ifor prediction index weight coefficient.To water bursting factor risk assessment object N _ki-th feature c _icorrelation function about a jth rank is designated as K _j(v _ki)

The step of described step (4-3) is:

If

K _j0(N _k)＝max{K _j(N _k)| _j＝1，2，… _，m} (30)

Then water bursting factor risk class is j ₀level.

Order

${\stackrel{\&OverBar;}{K}}_j\left(N_k\right)=\frac{K_j\left(N_k\right)-\underset{j}{\min }{K}_j\left(N_k\right)}{\underset{j}{\max }{K}_j\left(N_k\right)-\underset{j}{\min }{K}_j\left(N_k\right)}---\left(31\right)$

$j^{*}=\frac{\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}j\&CenterDot;{\stackrel{\&OverBar;}{K}}_j\left(N\right)k}{\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{K}}_j\left(N_k\right)}---\left(32\right)$

Then claim j ^*for forecasting object N _kgrade variable eigenwert.According to j ^*find out the degree of another grade of deflection.

The step of described step (4-4) is:

Adopt the weight of simple correlation function method determination prediction index, its computing method are as follows:

If $r_{\mathrm{ij}}=\left\{\begin{array}{cc}\frac{2(v_{\mathrm{ki}}-a_{0\mathrm{ij}})}{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}},& v_{\mathrm{ki}}\&le;\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}\\ \frac{2(b_{0\mathrm{ij}}-v_{\mathrm{ki}})}{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}},& v_{\mathrm{ki}}\&GreaterEqual;\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}\end{array}\right.---\left(33\right)$

If v _ki∈ v _pi, then

$r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})=\underset{j=1}{\overset{m}{\max }}\left\{r_{\mathrm{ij}}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\right\}---\left(34\right)$

If prediction index c _ithe rank that falls into of data larger, this index should give larger weight

$r_i=\left\{\begin{array}{c}{j}_{\max }\&times;(1+r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})),\\ j_{\max }\&times;0.5,\begin{array}{c}{r}_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\&GreaterEqual;-0.5\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})<-0.5\end{array}\end{array}\right.---\left(35\right)$

Otherwise, if prediction index c _ithe rank that falls into of data larger, this index should give less weight

$r_i=\left\{\begin{array}{c}(m-j_{\max }+1)\&times;(1+r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})),\\ ({m-j}_{\max }+1)\&times;0.5,\begin{array}{c}{r}_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\&GreaterEqual;-0.5\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})<-0.5\end{array}\end{array}\right.---\left(36\right)$

Then prediction index c _iweight be

$w_i=\frac{r_i}{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{r}_i}---\left(37\right).$

Beneficial effect of the present invention: the invention solves waterpower boundary conditions inside water-bearing bed in similar model test control problem, have the following advantages:

1, constructing tunnel phase all potential risk factors causing water bursting disaster to occur by inquiry, and carry out classified finishing and the screening process of system, the identification of analysis of possibility of water inrush factor can be carried out from pregnant dangerous environment and risk factors two angles;

2, the value of risk dynamic evaluation index is corrected in real time in conjunction with on-the-spot practice of construction situation, can realize the dynamic evaluation of water bursting factor risk.

Accompanying drawing explanation

Fig. 1 is evaluation procedure process flow diagram of the present invention.

Embodiment

Below in conjunction with accompanying drawing and embodiment, the invention will be further described.

(1) in the tunnel surveying stage, obtain the hydrogeological information of tunnel location and neighbouring country rock thereof, namely there is the pregnant dangerous environment of prominent water burst in tunnel, understands the risk status of geologic condition residing for each section of tunnel;

(2) calculate risk stratification value according to expert analysis mode vector with factor weight vector and carry out consistency check;

B＝[b ₁b ₂…b _i…b _n] (1)

ω＝[ω ₁ω ₂…ω _i…ω _n] ^T(2)

S＝B·ω (3)

Wherein, B and ω is respectively expert analysis mode vector and factor weight vector, and S is risk stratification value.

Analytical hierarchy process (AHP) is selected in the calculating of weight.The main calculation procedure of AHP method is as follows:

A step builds target layers figure, as shown in table 1;

Table 1

B Judgement Matricies, utilizes 1-9 degree method, as shown in table 2;

Table 2

C calculates weighted value

1) M of each row element of judgment matrix is calculated _i

$M_i=\underset{i=1}{\overset{n}{\mathrm{\&Pi;}}}{a}_{\mathrm{ij}}(i,j=\mathrm{1,2},...,n)---\left(4\right)$

2) M is calculated _in root

$W_i=\sqrt[n]{M}---\left(5\right)$

3) to vector $\stackrel{\&OverBar;}{W}=[\stackrel{\&OverBar;}{W_1},\stackrel{\&OverBar;}{W_2},\stackrel{\&OverBar;}{W_3},...,\stackrel{\&OverBar;}{W_n}]$ Normalized

$W=\frac{\stackrel{\&OverBar;}{W_i}}{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{W_i}}(i=\mathrm{1,2,3},...,n)---\left(6\right)$

W=[the W obtained ₁, W ₂, W ₃..., W _n] ^tbe required weight vector.

D. the eigenvalue of maximum λ of judgment matrix is calculated _max

After having calculated above, just can calculate the eigenvalue of maximum λ of judgment matrix _max, there is eigenvalue of maximum by known its of matrix theory:

${\mathrm{\&lambda;}}_{\max }=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\frac{{\left(\mathrm{AW}\right)}_i}{n{W}_i}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\frac{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{a}_{\mathrm{ij}}{W}_j}{W_i}---\left(7\right)$

E. consistency check

As can be seen from above computation process, whether the weighted value eventually through each factor calculated can reflect its importance truly objectively, depends mainly on the authenticity of judgment matrix.Therefore, the Conformance Assessment of satisfaction must be carried out to judgment matrix.

$\mathrm{CI}=\frac{{\mathrm{\&lambda;}}_{\max }-n}{n-1}---\left(8\right)$

$\mathrm{CR}=\frac{\mathrm{CI}}{\mathrm{RI}}---\left(9\right)$

In formula: CI: coincident indicator; CR: Consistency Ratio; RI: Aver-age Random Consistency Index, can be checked in by table 3.

Table 3

Because of subnumber (n)	1	2	3	4	5	6	7	8	9
										Corresponding value	0.00	0.00	0.58	0.90	1.12	1.24	1.32	1.41	1.45

If CR < 0.1, judgment matrix meets consistance, otherwise, need adjust judgment matrix, till satisfied.(3) risk factors is introduced in the influence factor of risk assessment, consider pregnant dangerous environment and risk factors, carry out prominent water burst risk assessment, divide tunnel and to dash forward the section distribution characteristics of water burst risk;

In analysis of possibility of water inrush factor, if descend water level to be quantitative description only, the method for structure membership function can be adopted to determine the degree of membership of each grade, and the whether objective validity determining evaluation result of membership function, is constructed as follows function:

${\mathrm{\&mu;}}_1=\left\{\begin{array}{c}1(x>x_1)\\ \frac{x-x_2}{x_1-x_2}(x_2<x<x_1)\\ 0(x<x_2)\end{array}\right.---\left(10\right)$

${\mathrm{\&mu;}}_3=\left\{\begin{array}{c}0(x>x_2)\\ \frac{x_2-x}{x_2-x_3}(x_3<x<x_2)\\ \frac{a}{x_2-x}(x<x_3)\end{array}\right.---\left(12\right)$

${\mathrm{\&mu;}}_4=\left\{\begin{array}{c}0(x>x_3)\\ \frac{x_2-x}{x_2-x_3}(x_3<x<x_2)\\ \frac{x_2-x-a}{x_2-x}(x<x_3)\end{array}\right.---\left(13\right)$

In formula (10) ~ (13), μ ₁, μ ₂, μ ₃, μ ₄the subordinate function of underground water four risk levels respectively, and x ₁, x ₂, x ₃60,30,0 and 10 are got respectively with a.

Can not the analysis of possibility of water inrush factor of quantificational description for other, the membership function that Karwowski conventional in engineering proposes can be adopted, in table 4.

Table 4

Each factor has k grade, and each factor grade has corresponding impact for the evaluation index evaluating collection, and its influence degree degree of membership represents, sets up Evaluations matrix B:

In formula: b _1krepresent the degree of membership of i-th factor kth grade relative to factor.

Risk estimation methods

1. combination weights method solves factor weight vector

Combination weights method factor weight vector is obtained by objective weight vector and subjective weight vector weighted sum:

ω＝P·Ψ _P+Q·Ψ _Q(15)

P＝[p ₁p ₂…p _i…p _n] (16)

Q＝[q ₁q ₂…q _i…q _n] (17)

In formula, P and Q is respectively objective weight vector and subjective weight vector, Ψ _pwith Ψ _qbe respectively the distribution weights of objective and subjective weight.

P to dash forward mud example by adding up typical gushing water, and Q is obtained by analytical hierarchy process, Ψ _pwith Ψ _qthen obtained by expert's confidence index method.

To be dashed forward mud accident by meter system domestic over one hundred example typical gushing water, obtain gushing water each factor weight in mud risk assessment set of factors of dashing forward and be respectively: unfavorable geology p ₁=0.252, formation lithology p ₂=0.089, underground water table p ₃=0.104, can lava and non-can lava contact zones p ₄=0.081, topography and geomorphology p ₅=0.067, attitude of rocks p ₆=0.022, aspect and crack in layer p ₇=0.037, Grades of Surrounding Rock p ₈=0.030, advance geologic prediction p ₉=0.141, monitoring measurement p ₁₀=0.095, excavation supporting p ₁₁=0.082.

Analytical hierarchy process is selected in the calculating of weight, distributes weights and adopts confidence index method to obtain.

Risk section partition

1) set of factors is set up

With fuzzy language, classification statement is carried out to analysis of possibility of water inrush probability of happening evaluation objective, sets up and evaluate collection:

V＝(v ₁，v ₂，v ₃，v ₄)

(18)

In formula: v ₁for primary risk; v ₂for secondary risk; v ₃for tertiary risk; v ₄for level Four risk, gushing water geologic hazard degree weakens successively, each element v _irepresenting various possible total evaluation result, namely considering all gushing water influence factors, drawing best evaluation result from evaluating to concentrate.

2) risk assessment and sequence

Application blurring mapping principle carries out comprehensive evaluation to relevant each factor, can obtain the comprehensive evaluation collection of analysis of possibility of water inrush probability of happening:

According to maximum membership grade principle, in above formula result of calculation, in last subordinated-degree matrix, overall target υ _lthe degree of membership c of corresponding opinion rating _lsoprano, is assessment of w ater inrush risk grade.Be formulated as:

${\mathrm{\&upsi;}}_s=\{c_l/{\mathrm{\&upsi;}}_l\&RightArrow;\underset{i=1}{\overset{k}{\max \left(c_i\right)}}\}---\left(20\right)$

(4) value of each index is corrected in real time in conjunction with on-the-spot practice of construction situation, to realize the dynamic evaluation of water bursting factor risk.

With research object in water bursting factor risk assessment Matter element system for things, prediction index is feature, and standardize criteria value is eigenwert, and water bursting factor risk assessment matter-element can describe with following formula:

Get a certain forecasting object in water bursting factor risk assessment Matter element system, only predict according to the prediction index that adjacent lower is corresponding, be called point prediction, it is the basis of water bursting factor risk class assessment.

(1) Classical field, joint territory, present situation matter-element

1) Classical field matter-element

According to prominent water burst risk assessment classification standard, the Classical field matter-element of forecasting object can be obtained:

$R_{0j}=\left[\begin{array}{ccc}{N}_{0j},& c_1,& v_{01j}\\ & c_1,& v_{02j}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{0\mathrm{nj}}\end{array}\right]=\left[\begin{array}{ccc}{N}_{0j},& c_1,& <a_{01j},b_{01j}>\\ & c_2,& <a_{02j},b_{02j}>\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& <a_{0\mathrm{nj}},b_{0\mathrm{nj}}>\end{array}\right]---\left(22\right)$

In formula: R _0jfor jth is individual with levying matter-element, N _0jfor forecasting object jth rank (j=1,2 ..., m), m is institute's divided rank number; c _ifor forecasting object i-th feature (i=1,2 ..., n), n is the characteristic number of forecasting object; v _0ij=﹤ a _0ij, b _0ij﹥ is rank j i-th feature c _ivalue territory after standardization, the i.e. Classical field got about the index of forecasting object of each rank.

2) territory matter-element is saved

$R_P=\left[\begin{array}{ccc}{N}_p,& c_1,& v_{P1}\\ & c_2,& v_{P2}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{\mathrm{Pn}}\end{array}\right]=\left[\begin{array}{ccc}{N}_P,& c_1,& <a_{P1},b_{P1}>\\ & c_2,& <a_{P2},b_{P2}>\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& <a_{\mathrm{Pn}},b_{\mathrm{Pn}}>\end{array}\right]---\left(23\right)$

In formula: N _pfor forecasting object rank is all; v _pi=﹤ a _pi, b _pi﹥ is that P is about feature c _ithe codomain got, i.e. N _pjoint territory.Above formula is designated as R _p(P, C, V _p), because each feature codomain after standardization is ﹤ 0,1 ﹥, so can save territory R _p=[N _p, c _i, < 0,1 >].Obviously have, (i=1,2 ..., n; J=1,2 ..., m).

3) present situation matter-element

According to actual conditions to forecasting object N _keach eigenwert carry out statistics and analysis, and to process according to given standard, thus obtain present situation matter-element (matter-element to be evaluated).

$R_k=\left[\begin{array}{ccc}{N}_k,& c_1,& v_{k1}\\ & c_2,& v_{k2}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{\mathrm{kn}}\end{array}\right]---\left(24\right)$

In formula, N _kfor forecasting object; v _kifor i-th feature c of forecasting object _istandardization value.

(2) matter-element grade correlation analysis

1) the single index degree of association

To water bursting factor risk assessment object N _ki-th feature c _icorrelation function about a jth rank is designated as K _j(v _ki), can be calculated by following formula:

$K_j\left(v_{\mathrm{ki}}\right)=\left\{\begin{array}{c}\frac{-\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})}{\left|v_{0\mathrm{ij}}\right|},v_{\mathrm{ki}}\&Element;v_{0\mathrm{ij}}\\ \frac{\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})}{\mathrm{\&rho;}(v_{\mathrm{ki}},v_{\mathrm{Pi}})-\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})},v_{\mathrm{ki}}\&NotElement;v_{0\mathrm{ij}}\end{array}\right.---\left(25\right)$

$\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})=|v_{\mathrm{ki}}-\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}|-\frac{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}}{2}---\left(26\right)$

|v _0ij|＝|b _0ij-a _0ij| (27)

$\mathrm{\&rho;}(v_{\mathrm{ki}},v_{\mathrm{pi}})=|v_{\mathrm{ki}}-\frac{a_{\mathrm{pi}}+b_{\mathrm{pi}}}{2}|-\frac{b_{\mathrm{pi}}-a_{\mathrm{pi}}}{2}---\left(28\right)$

2) the comprehensive multi-index degree of association

Water bursting factor risk assessment object N _kthe degree of association about grade j is

$K_j\left(N_k\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{w}_i{K}_j\left(v_{\mathrm{ki}}\right)---\left(29\right)$

In formula: w _ifor prediction index weight coefficient.

(3) forecasting object classification Comprehensive Evaluation

If

K _j0(N _k)＝max{K _j(N _k)| _{j＝1，2，…，m}} (30)

Then water bursting factor risk class is j ₀level.

Order

${\stackrel{\&OverBar;}{K}}_j\left(N_k\right)=\frac{K_j\left(N_k\right)-\underset{j}{\min }{K}_j\left(N_k\right)}{\underset{j}{\max }{K}_j\left(N_k\right)-\underset{j}{\min }{K}_j\left(N_k\right)}---\left(31\right)$

$j^{*}=\frac{\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}j\&CenterDot;{\stackrel{\&OverBar;}{K}}_j\left(N\right)k}{\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{K}}_j\left(N_k\right)}---\left(32\right)$

Then claim j ^*for forecasting object N _kgrade variable eigenwert.According to j ^*the degree of another grade of deflection can be found out.Such as, j ₀=2, j ^*=2.6, represent forecasting object N _krisk class is II level deflection III level.

(4) prediction index weight analysis

Prediction index weighing computation method can be divided three classes: a class is subjective weighting method, rule of thumb judges to obtain, as Delphi method, analytical hierarchy process etc. primarily of expert; Equations of The Second Kind is objective weighted model, as frequency number analysis, range method, Information Entropy etc.; 3rd class is combination weights method, also known as information set connection, is obtained by subjective weight and objective weight COMPREHENSIVE CALCULATING.Adopt the weight of simple correlation function method determination prediction index herein, its computing method are as follows:

If $r_{\mathrm{ij}}=\left\{\begin{array}{cc}\frac{2(v_{\mathrm{ki}}-a_{0\mathrm{ij}})}{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}},& v_{\mathrm{ki}}\&le;\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}\\ \frac{2(b_{0\mathrm{ij}}-v_{\mathrm{ki}})}{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}},& v_{\mathrm{ki}}\&GreaterEqual;\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}\end{array}\right.---\left(33\right)$

If v _ki∈ v _pi, then

$r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})=\underset{j=1}{\overset{m}{\max }}\left\{r_{\mathrm{ij}}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\right\}---\left(34\right)$

If prediction index c _ithe rank that falls into of data larger, this index should give larger weight

$r_i=\left\{\begin{array}{c}{j}_{\max }\&times;(1+r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})),\\ j_{\max }\&times;0.5,\begin{array}{c}{r}_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\&GreaterEqual;-0.5\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})<-0.5\end{array}\end{array}\right.---\left(35\right)$

Otherwise, if prediction index c _ithe rank that falls into of data larger, this index should give less weight

$r_i=\left\{\begin{array}{c}(m-j_{\max }+1)\&times;(1+r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})),\\ ({m-j}_{\max }+1)\&times;0.5,\begin{array}{c}{r}_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\&GreaterEqual;-0.5\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})<-0.5\end{array}\end{array}\right.---\left(36\right)$

Then prediction index c _iweight be

$w_i=\frac{r_i}{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{r}_i}---\left(37\right)$

By reference to the accompanying drawings the specific embodiment of the present invention is described although above-mentioned; but not limiting the scope of the invention; one of ordinary skill in the art should be understood that; on the basis of technical scheme of the present invention, those skilled in the art do not need to pay various amendment or distortion that creative work can make still within protection scope of the present invention.

Claims (10)

1. Karst Tunnel gushing water is dashed forward the gradual risk dynamic assessment method of mud overall process, and it is characterized in that, step is as follows:

, there is from tunnel the hydrogeological information obtaining tunnel location and neighbouring country rock thereof the pregnant dangerous environment of prominent water burst in step (1): in the tunnel surveying stage; The hydrogeological information of described tunnel location and neighbouring country rock thereof comprise unfavorable geology, formation lithology, underground water table, can lava and non-can lava contact zones, topography and geomorphology, the attitude of rocks, aspect and crack in layer and Grades of Surrounding Rock;

Step (2): obtain target layers table according to step (1), according to target layers table and factor weight vector, calculates risk stratification value, and carries out consistency check;

Step (3): risk factors is introduced in the influence factor of risk assessment, consider the tunnel location of step (1) and the hydrogeological information of neighbouring country rock thereof and risk factors, carry out the estimation of prominent water burst risk, divide tunnel and to dash forward the section distribution characteristics of water burst risk;

Step (4): hydrogeological information step (1) obtained is corrected in real time in conjunction with on-the-spot practice of construction situation, to realize the dynamic evaluation of water bursting factor risk.

2. a kind of Karst Tunnel gushing water as claimed in claim 1 is dashed forward the gradual risk dynamic assessment method of mud overall process, it is characterized in that, described step (2) according to target layers table and factor weight vector, calculate risk stratification value, calculating risk stratification value step is:

B＝[b ₁b ₂…b _i…b _n] (1)

ω＝[ω ₁ω ₂…ω _i…ω _n] ^T(2)

S＝B·ω (3)

Wherein, B is expert analysis mode vector, and ω is factor weight vector, and S is risk stratification value.

3. a kind of Karst Tunnel gushing water as claimed in claim 1 is dashed forward the gradual risk dynamic assessment method of mud overall process, and it is characterized in that, the step of described step (2) is as follows:

Step (2-1): build target layers table according to hydrogeological information;

Step (2-2): utilize 1-9 scaling law Judgement Matricies;

Step (2-3): utilize the judgment matrix of step (2-2) to calculate weighted value;

Step (2-4): utilize the weighted value of step (2-3) to calculate the eigenvalue of maximum λ of judgment matrix _max:

${\mathrm{\&lambda;}}_{\max }=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\frac{{\left(\mathrm{AW}\right)}_i}{{\mathrm{nW}}_i}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\frac{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{a}_{\mathrm{ij}}{W}_j}{W_i}---\left(7\right)$

A _ijmeaning of parameters be each element in judgment matrix;

Step (2-5): consistency check:

Whether the weighted value eventually through each factor calculated can reflect its importance truly objectively, is decided by the authenticity of judgment matrix; Therefore, the Conformance Assessment of satisfaction must be carried out to judgment matrix:

$\mathrm{CI}=\frac{{\mathrm{\&lambda;}}_{\max }-n}{n-1}---\left(8\right)$

$\mathrm{CR}=\frac{\mathrm{CI}}{\mathrm{RI}}---\left(9\right)$

In formula: CI: coincident indicator; CR: Consistency Ratio; RI: Aver-age Random Consistency Index, is checked in by table 3;

Table 3

Because of subnumber (n) 1 2 3 4 5 6 7 8 9 Corresponding value 0.00 0.00 0.58 0.90 1.12 1.24 1.32 1.41 1.45

If Consistency Ratio CR < 0.1, judgment matrix meets consistance, otherwise, need adjust judgment matrix, till judgment matrix meets consistance.

4. a kind of Karst Tunnel gushing water as claimed in claim 3 is dashed forward the gradual risk dynamic assessment method of mud overall process, it is characterized in that,

The step of described step (2-3) is as follows:

Step (2-3-1): the M calculating each row element of judgment matrix _i

$M_i=\underset{i=1}{\overset{n}{\mathrm{\&Pi;}}}{a}_{\mathrm{ij}}(i,j=\mathrm{1,2},...,n)---\left(4\right)$

Step (2-3-2): calculate M _in root

$W_i=\sqrt[n]{M}---\left(5\right)$

Step (2-3-3): to vector normalized

$W=\frac{{\stackrel{\&OverBar;}{W}}_i}{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{W}}_i}(i=\mathrm{1,2,3},...,n)---\left(6\right)$

The ω obtained=[ω ₁ω ₂ω _iω _n] ^tbe required factor weight vector.

5. a kind of Karst Tunnel gushing water as claimed in claim 1 is dashed forward the gradual risk dynamic assessment method of mud overall process, and it is characterized in that, the step that risk factors is introduced in the influence factor of risk assessment by described step (3) is:

In hydrogeological information, if descend water level to be quantitative description only, adopt the method for structure membership function to determine the degree of membership of each grade, the whether objective validity determining evaluation result of membership function, is constructed as follows function:

${\mathrm{\&mu;}}_1=\left\{\begin{array}{c}1\left({x>x}_1\right)\\ \frac{x-x_2}{x_1-x_2}(x_2<x<x_1)\\ 0(x<x_2)\end{array}\right.---\left(10\right)$

${\mathrm{\&mu;}}_3=\left\{\begin{array}{c}0(x>x_2)\\ \frac{x_2-x}{x_2-x_3}(x_3<x<x_2)\\ \frac{a}{x_2-x}(x<x_3)\end{array}\right.---\left(12\right)$

${\mathrm{\&mu;}}_4=\left\{\begin{array}{c}0(x>x_3)\\ \frac{x_2-x}{x_2-x_3}(x_3<x<x_2)\\ \frac{x_2-x-a}{x_2-x}(x<x_3)\end{array}\right.---\left(13\right)$

In formula (10) ~ (13), μ ₁, μ ₂, μ ₃, μ ₄the subordinate function of underground water four risk levels respectively, and x ₁get 60, x ₂get 30, x ₃get 0 and a get 10;

Can not the analysis of possibility of water inrush factor of quantificational description for other, adopt Karwowski algorithm to propose membership function, in table 4:

Table 4

Each factor has k grade, and each factor grade has corresponding impact for the evaluation index evaluating collection, and its influence degree degree of membership represents, sets up Evaluations matrix B:

In formula: b _ikrepresent the degree of membership of i-th factor kth grade relative to factor.

6. a kind of Karst Tunnel gushing water as claimed in claim 1 is dashed forward the gradual risk dynamic assessment method of mud overall process, it is characterized in that,

The risk estimation steps of described step (3) is:

Combination weights method solves factor weight vector w and to dash forward the objective weight vector P that mud example obtains and the subjective weight vector Q weighted sum obtained by analytical hierarchy process and obtain by adding up typical gushing water:

ω＝P·Ψ _P+Q·Ψ _Q(15)

P＝[p ₁p ₂…p _i…p _n] (16)

Q＝[q ₁q ₂…q _i…q _n] (17)

In formula, P is objective weight vector, Q be subjective weight vector respectively with, Ψ _pfor the distribution weights of objective weight, Ψ _qfor the distribution weights of subjective weight, Ψ _pwith Ψ _qthen obtained by the confidence index method based on expert investigation.

7. a kind of Karst Tunnel gushing water as claimed in claim 1 is dashed forward the gradual risk dynamic assessment method of mud overall process, it is characterized in that, the dash forward step of section distribution characteristics of water burst risk of described division tunnel is:

Step (3a-1): set up set of factors;

With fuzzy language, classification statement is carried out to analysis of possibility of water inrush probability of happening evaluation objective, sets up and evaluate collection:

V＝(v ₁，v ₂，v ₃，v ₄) (18)

In formula: v ₁for primary risk; v ₂for secondary risk; v ₃for tertiary risk; v ₄for level Four risk, gushing water geologic hazard degree weakens successively, each element v _irepresenting various possible total evaluation result, namely considering all gushing water influence factors, drawing best evaluation result from evaluating to concentrate;

Step (3a-2): risk assessment and sequence;

Application blurring mapping principle carries out comprehensive evaluation to relevant each factor, obtains the comprehensive evaluation collection P of analysis of possibility of water inrush probability of happening _f:

A is construction factor judgment matrix; a _irepresent each construction factor, B is Evaluations matrix, b _ikrepresent the degree of membership of i-th factor kth grade relative to factor

According to maximum membership grade principle, in formula (20) result of calculation, in last subordinated-degree matrix, overall target υ _lthe degree of membership c of corresponding opinion rating _lsoprano, is assessment of w ater inrush risk grade;

Be formulated as:

$v_s=\{c_l/v_l\&RightArrow;\underset{i=1}{\overset{k}{\max \left(c_i\right)}}\}---\left(20\right);$

C _ifor risk class membership vector, v _sfor gushing water is dashed forward mud risk class, v _lfor overall target.

8. a kind of Karst Tunnel gushing water as claimed in claim 1 is dashed forward the gradual risk dynamic assessment method of mud overall process, and it is characterized in that, the step of described step (4) is:

With research object in water bursting factor risk assessment Matter element system for things, prediction index is feature, and standardize criteria value is eigenwert, and water bursting factor risk assessment matter-element following formula describes:

Get a certain forecasting object in water bursting factor risk assessment Matter element system, only predict according to the prediction index that adjacent lower is corresponding, be called point prediction, it is the basis of water bursting factor risk class assessment;

Wherein the process of point prediction is as follows:

Step (4-1): calculate Classical field matter-element, joint territory matter-element and present situation matter-element respectively;

Step (4-2): the matter-element grade degree of association in analytical procedure (4-1);

Step (4-3): utilize the degree of association in step (4-2) to carry out forecasting object classification Comprehensive Evaluation;

Step (4-4): the index weights of forecasting object is analyzed.

9. a kind of Karst Tunnel gushing water as claimed in claim 8 is dashed forward the gradual risk dynamic assessment method of mud overall process, and it is characterized in that, the step of described step (4-1) is:

Step (4-1-1): Classical field matter-element

According to prominent water burst risk assessment classification standard, the Classical field matter-element of forecasting object can be obtained:

$R_{0j}=\left[\begin{array}{ccc}{N}_{0j},& c_1,& v_{01j}\\ & c_2,& v_{02j}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{0\mathrm{nj}}\end{array}\right]=---\left(22\right)\left[\begin{array}{ccc}{N}_{0j},& c_1,& <a_{01j},b_{01j}>\\ & c_2,& <a_{02j},b_{02j}>\\ & & .\\ & & .\\ & & .\\ & c_n,& <a_{0\mathrm{nj}},b_{0\mathrm{nj}}>\end{array}\right]---\left(22\right)$

In formula:

R _0jfor jth is individual with levying matter-element,

N _0jfor the jth rank of forecasting object, j=1,2 ..., m, m are institute's divided rank number;

C _ifor i-th feature of forecasting object, i=1,2 ..., n, n are the characteristic number of forecasting object;

V _0ij=﹤ a _0ij, b _0ij﹥ is rank j i-th feature c _ivalue territory after standardization, the i.e. Classical field got about the index of forecasting object of each rank;

Step (4-1-2): joint territory matter-element

$R_P=\left[\begin{array}{ccc}{N}_P,& c_1,& v_{P1}\\ & c_2,& v_{P2}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{\mathrm{Pn}}\end{array}\right]=\left[\begin{array}{ccc}{N}_P,& c_1& <a_{P1},b_{P1}>\\ & c_2,& <a_{P2},b_{P2}>\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& <a_{\mathrm{Pn}},b_{\mathrm{Pn}}>\end{array}\right]---\left(23\right)$

In formula:

N _pfor forecasting object rank is all;

V _pi=﹤ a _pi, b _pi﹥ is that P is about feature c _ithe codomain got, i.e. N _pjoint territory; Above formula is designated as R _p(P, C, V _p), because each feature codomain after standardization is ﹤ 0,1 ﹥, so must save territory R _p=[N _p, c _i, < 0,1 >]; Obviously have, i=1,2 ..., n; J=1,2 ..., m;

Step (4-1-3): present situation matter-element

According to actual conditions to forecasting object N _keach eigenwert carry out statistics and analysis, and to process according to given standard, thus obtain present situation matter-element;

$R_k=\left[\begin{array}{ccc}{N}_k,& c_1,& v_{k1}\\ & c_2,& v_{k2}\\ & .& .\\ & .& .\\ & .& .\\ & c_n,& v_{\mathrm{kn}}\end{array}\right]---\left(24\right)$

In formula, N _kfor forecasting object; v _kifor i-th feature c of forecasting object _istandardization value.

10. a kind of Karst Tunnel gushing water as claimed in claim 8 is dashed forward the gradual risk dynamic assessment method of mud overall process, and it is characterized in that, the step of described step (4-2) is:

Step (4-2-1): the single index degree of association

To water bursting factor risk assessment object N _ki-th feature c _icorrelation function about a jth rank is designated as K _j(v _ki), can be calculated by following formula:

$K_j\left(v_{\mathrm{ki}}\right)=\left\{\begin{array}{c}\frac{-\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})}{\left|v_{0\mathrm{ij}}\right|},v_{\mathrm{ki}}\&Element;v_{0\mathrm{ij}}\\ \frac{\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})}{\mathrm{\&rho;}(v_{\mathrm{ki}},v_{\mathrm{Pi}})-\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})},v_{\mathrm{ki}}\&Element;v_{0\mathrm{ij}}\end{array}\right.---\left(25\right)$

$\mathrm{\&rho;}(v_{\mathrm{ki}},v_{0\mathrm{ij}})=|v_{\mathrm{ki}}-\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}|-\frac{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}}{2}---\left(26\right)$

|v _0ij|＝|b _0ij-a _0ij| (27)

$\mathrm{\&rho;}(v_{\mathrm{ki}},v_{\mathrm{pi}})=|v_{\mathrm{ki}}-\frac{a_{\mathrm{pi}}+b_{\mathrm{pi}}}{2}|-\frac{b_{\mathrm{pi}}-a_{\mathrm{pi}}}{2}---\left(28\right)$

Step (4-2-2): the comprehensive multi-index degree of association

Water bursting factor risk assessment object N _kthe degree of association about grade j is

$K_j\left(N_k\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{w}_i{K}_j\left(v_{\mathrm{ki}}\right)---\left(29\right)$

In formula: w _ifor prediction index weight coefficient; To water bursting factor risk assessment object N _ki-th feature c _icorrelation function about a jth rank is designated as K _j(v _ki);

The step of described step (4-4) is:

Adopt the weight of simple correlation function method determination prediction index, its computing method are as follows:

If $r_{\mathrm{ij}}=\left\{\begin{array}{cc}\frac{2(v_{\mathrm{ki}}-a_{0\mathrm{ij}})}{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}},& v_{\mathrm{ki}}\&le;\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}\\ \frac{2(b_{0\mathrm{ij}}-v_{\mathrm{ki}})}{b_{0\mathrm{ij}}-a_{0\mathrm{ij}}},& v_{\mathrm{ki}}\&GreaterEqual;\frac{a_{0\mathrm{ij}}+b_{0\mathrm{ij}}}{2}\end{array}\right.---\left(33\right)$

If v _ki∈ v _pi, then

$r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})=\underset{j=1}{\overset{m}{\max \left\{r_{\mathrm{ij}}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\right\}---\left(34\right)}}$

If prediction index c _ithe rank that falls into of data larger, this index should give larger weight

$r_i=\left\{\begin{array}{c}{j}_{\max }\&times;(1+r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})),\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\&GreaterEqual;-0.5\\ j_{\max }\&times;0.5\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})<-0.5\end{array}\right.---\left(35\right)$

Otherwise, if prediction index c _ithe rank that falls into of data larger, this index should give less weight

$r_i=\left\{\begin{array}{c}({m-j}_{\max }+1)\&times;(1+r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}}))\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})\&GreaterEqual;-0.5\\ (m-j_{\max }+1)\&times;0.5\\ r_{{\mathrm{ij}}_{\max }}(v_{\mathrm{ki}},v_{0\mathrm{ij}})<-0.5\end{array}\right.---\left(36\right)$

Then prediction index c _iweight be

$w_i=\frac{r_i}{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{r}_i}---\left(37\right).$