1. The utility model provides a load determination method of underground utility tunnel gas leakage explosion which characterized in that: the realization process is as follows:

1. integrated process

In the discretized risk indicator space, each element R integrates a certain dangerous scene and the information carried by the dangerous scene,

R＝{<S_i，L_i，C_i>} (1)

wherein R is the risk degree, S_iFor the ith dangerous scene, C_iIs S_iAs a result of (A), L_iIs C_iThe possibility of the occurrence of the above-mentioned problems,

defining 'natural gas leakage explosion of the comprehensive pipe gallery' as a dangerous scene set S, and respectively using 'explosion overpressure peak value and impulse' as a consequence set C_PAnd C_I；

2. Failure frequency model

Selecting a long-distance natural gas pipeline accident database of an European gas pipeline accident data organization EGIG as a reference, defining an accident causing one-time accidental gas leakage in the EGIG as one-time failure, dividing failure types into three types of small holes (d <20mm), large holes (20mm < d <300mm) and breakage (d ═ pipe diameter) according to different equivalent leakage diameters d, and finally calculating the failure frequency according to EGIG statistical data to be used as the failure frequency of leakage of the natural gas pipeline in the comprehensive pipe rack;

according to the stipulation of the design and construction of the gas pipeline cabin supporting facilities of the comprehensive pipe gallery (18GL502), an emergency cut-off system is required to be arranged when gas enters the gallery, and the leakage flow Q (t) is calculated according to a leakage model of the open literature;

3. sample selection method

Determining a certain dangerous scene Si, at least needing three parameters of leakage occurrence position, leakage direction and leakage flow, and recording as S_i(X1,X2,X3)；

In a fire zone 200m long, the location of the leak was random, assuming X1 was at [0,200 ]]Subject to uniform distribution, the direct use of the Monte Carlo method may cause the sampling result to be very different from the original distribution, so according to the layered sampling principle, 0,200]Divided into five sections from central axis to two sides of fire-proof partitionCalled the median, interior and end regions, leakage can theoretically occur in any direction, but for simplicity, it is believed that in a rectangular spatial coordinate system leakage can only occur in the + -X, + -Y and + -Z directions, where the X direction is used to denote alpha<Under the condition of 45 degrees, X2 is a discrete random variable and takes values randomly in { +/-, Y, +/-Z }, the probability of being selected in each direction is 1/6, the ratio of the small hole, the large hole and the fracture failure frequency is approximate to 14:4:1, so X3 is the random variable capable of reflecting the proportional relation, when large hole leakage is considered, the difference of the leakage flow corresponding to 20mm and 300mm can approach two orders of magnitude, so the value range can be divided into a plurality of small intervals according to the equal ratio principle, and then selection is carried out in each small interval, so S_iThe method can be regarded as a ternary random variable, the dimensions are independent and obey different distributions;

4. cumulative ignition frequency

It is assumed that the probability of an ignition source occurring at various locations in the piping lane is the same, and as the leaking gas gradually diffuses, the volume of the flammable gas cloud continues to evolve. In the relatively closed space of the pipe gallery, ignition takes place while the ignition source is present and the mixture gas concentration there is within the explosive limits, so that the basic probability of ignition p is:

in the formula, S_pThe possibility of fire sources; FLAM is the volume of the combustible gas cloud (m)³) (ii) a V is total volume (m) of fire-retardant subareas³)；

Constructing an ignition event tree, judging whether to ignite at the end of each second, and calculating the ignition probability at a certain moment by using conditional probability, namely the probability P of being ignited at the very tth second_0tComprises the following steps:

in general, the probability of ignition P is accumulated over t seconds_tIs composed of

The accumulated ignition probability constructed by the method has only one empirical parameter s_pThe method has practicability, and simultaneously considers the characteristics of the real combustible gas cloud development process and the time growth of the real combustible gas cloud development process;

5. joint simulation of leakage and explosion

And performing combined simulation of gas leakage diffusion and explosion according to the calculated leakage flow and the selected simulation working conditions with proper quantity, and corresponding the explosion consequences under each leakage amount one to one.

2. The underground utility tunnel gas leakage explosion load determination method according to claim 1, characterized in that: in the step 5, the specific process is as follows:

5.1 cloud cumulative frequency curve for combustible gas

After FLACS leakage diffusion simulation calculation is finished, in an rt.fuel file, except for generating a combustible region volume FLAM, an equivalent combustible gas cloud volume Q9 is generated, Q9 is used for equivalently obtaining an obtained non-uniform combustible gas cloud volume into a uniform combustible gas cloud volume capable of being completely combusted according to a certain rule, and Q9 is:

in the formula, V_i(i ═ 1,2, … n) is the volume of each control grid occupied by the flammable region in the flow field (m 3); ER_iIs where the mixed gas equivalence ratio of the ith control grid; ERfac (ER)_i) Reflects the influence of the combustion rate of the laminar flame; ve (ER)_i) Reflects the influence of the expansion capacity of the gas;

5.2 correlation of equivalent combustible gas cloud and explosion consequence

Randomly extracting a certain quantity from a data result file generated by the leakage diffusion simulation to perform real gas cloud explosion simulation, aiming at establishing a correlation between combustible gas cloud and explosion consequences;

5.3 overrun probability curve of explosion overpressure peak value and impulse

Comprehensively considering the failure frequency of the natural gas pipeline, the cumulative frequency curve of the equivalent combustible gas cloud Q9, the cumulative ignition probability curve of each sample and the explosion result C of natural gas leakage in the comprehensive pipe gallery_iThe possibilities of (a) are:

L_i(C_i)＝F_f×Cum(Q9_i)×P_ti (6)

when mixing C_iDefined as the explosion overpressure peak P_maxTime, ignition source parameter s_pAll (P) were calculated according to formula (6) at 0.01, 0.05 and 0.1_max,L_i) And extracting an upper envelope curve of the sample point, drawing an over-voltage probability curve of an over-voltage peak value, representing the maximum probability of causing a certain over-voltage peak value, and similarly, calculating the over-voltage probability curve of the impulse.

3. The underground utility tunnel gas leakage explosion load determination method according to claim 1, characterized in that: in the failure frequency model in the step 2, the following three factors are considered to reasonably select and modify the EGIG statistical data for use:

(1) frequency of failure F_fThe statistical data of the last ten years from 2007 to 2016 are selected as a reference because the statistical data decline year by year and are stabilized at a lower level since the beginning of the century;

(2) the diameter d of the natural gas pipeline of the comprehensive pipe gallery is usually 300-600 mm, so 279.4-584.2 mm of statistical data are selected as a reference;

(3) after the gas enters the corridor, the failure caused by the damage of the third party is basically avoided, and the effect of natural disasters can be reduced to a great extent, so that the influences of the damage of the third party and the natural disasters are respectively reduced by 90% and 60% and then calculated.

Background

At present, China is vigorously pushing the infrastructure construction of the comprehensive pipe gallery, encouraging gas to enter the gallery, and facilitating unified management. Because utility tunnel has been equipped with modern control, early warning and control system, can reduce the quantity of gas incident to a certain extent. However, in the case of a gas leakage explosion in the closed gas compartment, the consequences can be even more serious. Therefore, the natural gas leakage explosion of the comprehensive pipe gallery has the characteristics of low accident risk value and serious accident consequence, and when the anti-explosion design is carried out, a load value taking method related to gas explosion in the specification is not applicable any more, and related research needs to be carried out.

As a commonly used load-valuation method, Quantitative Risk Assessment (QRA) uses a limited number of calculations to quantitatively describe the impact of an event and its likely extent. The risk level calculation facing various facilities in the chemical industry is explained in detail in the specification, and the CFD result is recommended to be used for supporting when explosive load risk evaluation is carried out. Hansen et al first propose to use CFD technology to carry out risk analysis of gas explosion; yet, etc. use FLACS gas explosion software to carry out quantitative risk evaluation on heat, overpressure peak value and impulse of the petrochemical storage tank area and give personal risk values of different areas. Thereafter, as the computer level develops, more related researches are spread around various factors influencing the risk evaluation result, such as the number of simulation scenes and the representativeness of the simulation scenes. And the method aims to improve the refinement degree of the existing risk evaluation system and enhance the robustness of the prediction result. However, the related research is limited to the field of process industry, and the conclusion that the related research can be directly applied to the antiknock design of the comprehensive pipe gallery is less. And the leakage diffusion and explosion of gas are a whole process, and the leakage diffusion and explosion of gas are combined together and are only considered, so that the method has certain limitation on the application of a new scene, and a gas leakage explosion load value taking method aiming at the comprehensive pipe gallery needs to be proposed urgently.

Disclosure of Invention

The invention aims to overcome the defects of the technology and provide a method for determining the load of gas leakage and explosion of an underground comprehensive pipe gallery.

In order to achieve the purpose, the invention adopts the technical scheme that the comprehensive pipe gallery gas leakage explosion load value taking method based on risk assessment is implemented by the following steps:

1. integrated process

Risk assessment is a statistical method to quantitatively assess the impact of an event and its likely extent. In the discretized risk indicator space, each element R integrates a certain dangerous scene and the information it carries.

R＝{<S_i,L_i,C_i>} (1)

Wherein R is the risk degree, S_iFor the ith dangerous scene, C_iIs S_iAs a result of (A), L_iIs C_iThe possibility of the occurrence of the above-mentioned problems,

and defining the natural gas leakage explosion of the comprehensive pipe gallery as a dangerous scene set S. The "explosion overpressure peak and impulse" are taken as the outcome sets CP and CI, respectively.

2. Failure frequency model

Because no data about the failure and leakage of the natural gas pipeline in the comprehensive pipe rack exists at present, a long-distance natural gas pipeline accident database of the European gas pipeline accident data organization (EGIG) is selected as a reference. An EGIG defines an accident that results in an accidental gas leak as a failure, and the failure types are classified into three types, i.e., small hole (d <20mm), large hole (20mm < d <300mm), and fracture (d ═ tube diameter), according to the equivalent leak diameter d. The following three factors are considered for reasonably selecting and correcting the EGIG statistical data for use:

(1) frequency of failure F_fThe statistics of the last decade from 2007 to 2016 are selected as a reference because of the decline year by year and the stability at a lower level since the beginning of the century.

(2) The diameter d of the natural gas pipeline of the comprehensive pipe rack is usually between 300 and 600mm, so statistics of 279.4 to 584.2mm are selected as a reference.

(3) After the gas enters the corridor, the failure caused by the damage of the third party is basically avoided, and the effect of natural disasters can be reduced to a great extent, so that the influences of the damage of the third party and the natural disasters are respectively reduced by 90% and 60% and then calculated.

And combining the consideration, and finally calculating the failure frequency according to the EGIG statistical data to be used as the failure frequency of the leakage of the natural gas pipeline in the comprehensive pipe gallery.

According to the regulations of the design and construction of gas pipeline and cabin supporting facilities of the comprehensive pipe gallery (18GL502), an emergency cut-off system is required to be equipped when gas enters the gallery. Thus, when the emergency shut-off system is in operation, an unsteady leak will occur at the leak orifice and, starting from the upstream shut-off valve position, the leaking pipe section will be reduced to a fixed volume rigid vessel, and the leak flow Q (t) can be calculated according to the leak model in the document Montiel H, V i lchez J A, Casal J, et al.

3. Sample selection method

Theoretically, the number of elements in the dangerous scene set S is infinite, and it is impossible to calculate and analyze all dangerous scenes. Therefore, a limited number of dangerous scenes need to be scientifically selected for calculation and analysis, so as to obtain reliable risk assessment results. Determining a certain danger scenario S_iAt least three parameters of the leakage position, the leakage direction and the leakage flow are required and are recorded as S_i(X1,X2,X3)。

Due to the long and narrow characteristic of the comprehensive pipe gallery, the leakage occurrence position can be defined only by the X coordinate along the length direction. In a fire zone 200m long, the location of the leak was random, assuming X1 was at [0,200 ]]And uniformly distributed. The direct use of the Monte Carlo method may cause a large difference between the sampling result and the original distribution, so that according to the hierarchical sampling principle, 0,200 is first used]The leakage can only occur along the directions of +/-X, +/-Y and +/-Z in a rectangular space coordinate system for simplifying the problem. Wherein the X direction is taken to mean alpha<45 deg. so that X2 is a discrete random variable, randomly valued in { + -X, + -Y, + -Z } and chosen for each directionThe rates are all 1/6. The ratio of small pores, large pores and failure frequency at break was approximately 14:4: 1. Therefore, X3 is a random variable that reflects the proportional relationship described above. When large hole leakage is considered, the difference of the leakage flow rates corresponding to 20mm and 300mm is close to two orders of magnitude, so the value range can be divided into a plurality of small intervals according to the equal ratio principle, and then selection is performed from each small interval. Thus, S_iCan be regarded as a ternary random variable, the dimensions are independent of each other and obey different distributions.

4. Cumulative ignition frequency

It is assumed that the probability of an ignition source occurring at various locations in the piping lane is the same, and as the leaking gas gradually diffuses, the volume of the flammable gas cloud continues to evolve. In the relatively closed space of the pipe gallery, ignition takes place while the ignition source is present and the mixture gas concentration there is within the explosive limits, so that the basic probability of ignition p is:

in the formula, s_pThe possibility of fire sources; FLAM is the volume of the combustible gas cloud (m)³) (ii) a V is total volume (m) of fire-retardant subareas³)。

Constructing an ignition event tree, judging whether to ignite at the end of each second, and calculating the ignition probability at a certain moment by using conditional probability, namely the probability P of being ignited at the very tth second_0tComprises the following steps:

in general, we are more concerned about the cumulative probability of ignition P within t seconds_tIs composed of

The accumulated ignition probability constructed by the method has only one empirical parameterNumber s_pIt has practicability. And simultaneously considers the development process of the real combustible gas cloud and the characteristics of the real combustible gas cloud which grows along with the time.

5. Joint simulation of leakage and explosion

And performing combined simulation of gas leakage diffusion and explosion according to the calculated leakage flow and the selected simulation working conditions with proper quantity, and corresponding the explosion consequences under each leakage amount one to one.

5.1 cloud cumulative frequency curve for combustible gas

After the FLACS leakage diffusion simulation calculation is finished, in the rt.fuel file, an equivalent combustible gas cloud volume Q9 is generated in addition to the combustible region volume FLAM. Q9 becomes the even combustible gas cloud volume that can burn completely with the equivalence of the even combustible gas cloud volume that obtains according to certain rule, and Q9 is:

in the formula, V_i(i ═ 1,2, … n) is the volume (m) of each control grid occupied by a combustible region in the flow field³)；ER_iIs where the mixed gas equivalence ratio of the ith control grid; ERfac (ER)_i) Reflects the influence of the combustion rate of the laminar flame; ve (ER)_i) Reflecting the effect of the gas expansion capacity.

5.2 correlation of equivalent combustible gas cloud and explosion consequence

In a data result file generated by the leakage diffusion simulation, a certain quantity of data is randomly extracted to carry out real gas cloud explosion simulation, and the aim is to establish the correlation between combustible gas cloud and explosion consequences. Such a treatment has two advantages: firstly, because a plurality of factors such as the size of the gas cloud, the shape of the gas cloud, the position of the gas cloud, the ignition position, the equivalence ratio and the like can influence the result of gas explosion, a large number of samples are needed for carrying out quantitative risk assessment directly through gas explosion simulation; second, in a new environment where the cumulative frequency distribution of the combustible gas cloud is not substantially grasped, directly assuming the distribution of the combustible gas cloud sizes may lead to erroneous evaluation results.

5.3 overrun probability curve of explosion overpressure peak value and impulse

Comprehensively considering the failure frequency of the natural gas pipeline, the cumulative frequency curve of the equivalent combustible gas cloud Q9, the cumulative ignition probability curve of each sample and the explosion result C of natural gas leakage in the comprehensive pipe gallery_iThe possibilities of (a) are:

L_i(C_i)＝F_f×Cum(Q9_i)×P_ti (6)

when mixing C_iDefined as the explosion overpressure peak P_maxTime, ignition source parameter s_pAll (P) were calculated according to formula (6) at 0.01, 0.05 and 0.1_max,L_i) And extracting an upper envelope curve of the sample point, drawing an over-voltage probability curve of an over-voltage peak value, representing the maximum probability of causing a certain over-voltage peak value, and similarly, calculating the over-voltage probability curve of the impulse.

Compared with the prior art, the method has the advantages that the method is based on the computational fluid dynamics software FLACS simulation, considers the real scene and is more accurate in determination of the explosive load compared with the traditional method; the invention considers the influence of random factors such as gas leakage position, leakage amount, ignition position and the like on the explosion load, and the calculation result is more accurate.

Drawings

FIG. 1 is a main flow diagram of the present invention;

FIG. 2 is a graph of different types of failure frequencies before and after correction of EGIG statistical data;

FIG. 3 is a leak location zone diagram;

FIG. 4 is a leakage direction selection diagram;

FIG. 5 is a diagram of a FLACS numerical simulation complete flow and file information;

FIG. 6 is a diagram of a gas explosion simulation ignition position selection method;

FIG. 7a is a utility tunnel layout;

FIG. 7b is a diagram of a collection model;

FIG. 8a is a plot of unsteady flow time course of orifice leakage;

FIG. 8b is a plot of unsteady flow time course of large hole 1 leakage;

FIG. 8c is a plot of unsteady flow time course for large pore 2 leakage;

FIG. 8d is a plot of unsteady state leakage flow time history at a pipe break;

FIG. 9 is a calculated operating condition selection map;

FIG. 10 is a graph of cumulative frequency of a cloud of combustible gas;

FIG. 11 is a graph of combustible gas cloud values at a certain cumulative frequency;

FIG. 12 is a Q9 diagram of a different leakage hole;

FIG. 13a is a graph relating Q9 to an overpressure peak;

FIG. 13b is a graph of Q9 versus detonation impulse;

FIG. 14a is an overrun probability plot of an explosion overpressure spike

Fig. 14b is a transcendental probability plot of the explosion impulse.

Detailed Description

The following detailed description of the preferred embodiments will be made with reference to the accompanying drawings. A comprehensive pipe gallery gas leakage explosion load value taking method based on risk assessment comprises the following implementation processes:

1. integrated process

Risk assessment is a statistical method to quantitatively assess the impact of an event and its likely extent. In the discretized risk indicator space, each element R integrates a certain dangerous scene and the information it carries.

R＝{<S_i,L_i,C_i>} (1)

Wherein R is the risk degree, S_iFor the ith dangerous scene, C_iIs S_iAs a result of (A), L_iIs C_iThe possibility of the occurrence of the above-mentioned problems,

and defining the natural gas leakage explosion of the comprehensive pipe gallery as a dangerous scene set S. Respectively using 'explosion overpressure peak value and impulse' as result set C_PAnd C_I. The overall flow of the comprehensive pipe gallery natural gas leakage explosion quantitative risk assessment is shown in figure 1.

2. Failure frequency model

Because no data about the failure and leakage of the natural gas pipeline in the comprehensive pipe rack exists at present, a long-distance natural gas pipeline accident database of the European gas pipeline accident data organization (EGIG) is selected as a reference. An EGIG defines an accident that results in an accidental gas leak as a failure, and the failure types are classified into three types, i.e., small hole (d <20mm), large hole (20mm < d <300mm), and fracture (d ═ tube diameter), according to the equivalent leak diameter d. The following three factors are considered for reasonably selecting and correcting the EGIG statistical data for use:

(1) frequency of failure F_fThe statistics of the last decade from 2007 to 2016 are selected as a reference because of the decline year by year and the stability at a lower level since the beginning of the century.

(2) The diameter d of the natural gas pipeline of the comprehensive pipe rack is usually between 300 and 600mm, so statistics of 279.4 to 584.2mm are selected as a reference.

(3) After the gas enters the corridor, the failure caused by the damage of the third party is basically avoided, and the effect of natural disasters can be reduced to a great extent, so that the influences of the damage of the third party and the natural disasters are respectively reduced by 90% and 60% and then calculated.

In combination with the above considerations, the failure frequency finally calculated from the EGIG statistical data is shown in fig. 2, and is taken as the failure frequency of the natural gas pipeline in the utility tunnel with leakage.

According to the regulations of the design and construction of gas pipeline and cabin supporting facilities of the comprehensive pipe gallery (18GL502), an emergency cut-off system is required to be equipped when gas enters the gallery. Thus, when the emergency shut-off system is in operation, an unsteady leak will occur at the leak orifice and, starting from the upstream shut-off valve position, the leaking pipe section will be reduced to a fixed volume rigid vessel, and the leak flow Q (t) can be calculated according to the leak model in the document Montiel H, V i lchez J A, Casal J, et al.

3. Sample selection method

Theoretically, the number of elements in the dangerous scene set S is infinite, and it is impossible to calculate and analyze all dangerous scenes. Due to the fact thatTherefore, a limited number of dangerous scenes need to be scientifically selected for calculation and analysis, so as to obtain reliable risk assessment results. Determining a certain danger scenario S_iAt least three parameters of the leakage position, the leakage direction and the leakage flow are required and are recorded as S_i(X1,X2,X3)。

Due to the long and narrow characteristic of the comprehensive pipe gallery, the leakage occurrence position can be defined only by the X coordinate along the length direction. In a fire zone 200m long, the location of the leak was random, assuming X1 was at [0,200 ]]And uniformly distributed. The direct use of the Monte Carlo method may cause a large difference between the sampling result and the original distribution, so that according to the hierarchical sampling principle, 0,200 is first used]The fire-proof partition is divided into five sections, namely a middle section, an inner section and an end section from the central axis to two sides of the fire-proof partition in sequence, as shown in figure 3. Theoretically, leakage can occur in any direction, but for simplicity, it is believed that leakage can only occur in the + -X, + -Y and + -Z directions in a spatial rectangular coordinate system. Wherein the X direction is taken to mean alpha<45 deg., as shown in fig. 4. Thus, X2 is a discrete random variable that takes on values randomly in { + -X, + -Y, + -Z } and the probability of being chosen per direction is 1/6. The ratio of small pores, large pores and failure frequency at break was approximately 14:4: 1. Therefore, X3 is a random variable that reflects the proportional relationship described above. When large hole leakage is considered, the difference of the leakage flow rates corresponding to 20mm and 300mm is close to two orders of magnitude, so the value range can be divided into a plurality of small intervals according to the equal ratio principle, and then selection is performed from each small interval. Thus, S_iCan be regarded as a ternary random variable, the dimensions are independent of each other and obey different distributions.

4. Cumulative ignition frequency

It is assumed that the probability of an ignition source occurring at various locations in the piping lane is the same, and as the leaking gas gradually diffuses, the volume of the flammable gas cloud continues to evolve. In the relatively closed space of the pipe gallery, ignition takes place while the ignition source is present and the mixture gas concentration there is within the explosive limits, so that the basic probability of ignition p is:

in the formula, s_pThe possibility of fire sources; FLAM is the volume of the combustible gas cloud (m)³) (ii) a V is total volume (m) of fire-retardant subareas³)。

Constructing an ignition event tree, judging whether to ignite at the end of each second, and calculating the ignition probability at a certain moment by using conditional probability, namely the probability P of being ignited at the very tth second_0tComprises the following steps:

in general, we are more concerned about the cumulative probability of ignition P within t seconds_tIs composed of

The accumulated ignition probability constructed by the method has only one empirical parameter s_pIt has practicability. And simultaneously considers the development process of the real combustible gas cloud and the characteristics of the real combustible gas cloud which grows along with the time.

5. Joint simulation of leakage and explosion

And performing combined simulation of gas leakage diffusion and explosion according to the calculated leakage flow and the selected simulation working conditions with proper quantity, wherein the explosion consequences under each leakage amount are in one-to-one correspondence, and the specific flow is shown in fig. 5.

5.1 cloud cumulative frequency curve for combustible gas

After the FLACS leakage diffusion simulation calculation is finished, in the rt.fuel file, an equivalent combustible gas cloud volume Q9 is generated in addition to the combustible region volume FLAM. Q9 becomes the even combustible gas cloud volume that can burn completely with the equivalence of the even combustible gas cloud volume that obtains according to certain rule, and Q9 is:

in the formula, V_i(i ═ 1,2, … n) is the volume (m) of each control grid occupied by a combustible region in the flow field³)；ER_iIs where the mixed gas equivalence ratio of the ith control grid; ERfac (ER)_i) Reflects the influence of the combustion rate of the laminar flame; ve (ER)_i) Reflecting the effect of the gas expansion capacity.

5.2 correlation of equivalent combustible gas cloud and explosion consequence

In a data result file generated by the leakage diffusion simulation, a certain number of data are randomly extracted to perform real gas cloud explosion simulation, and the flow is shown in fig. 6, aiming at establishing the correlation between combustible gas cloud and explosion consequences. Such a treatment has two advantages: firstly, because a plurality of factors such as the size of the gas cloud, the shape of the gas cloud, the position of the gas cloud, the ignition position, the equivalence ratio and the like can influence the result of gas explosion, a large number of samples are needed for carrying out quantitative risk assessment directly through gas explosion simulation; second, in a new environment where the cumulative frequency distribution of the combustible gas cloud is not substantially grasped, directly assuming the distribution of the combustible gas cloud sizes may lead to erroneous evaluation results.

5.3 overrun probability curve of explosion overpressure peak value and impulse

Comprehensively considering the failure frequency of the natural gas pipeline, the cumulative frequency curve of the equivalent combustible gas cloud Q9, the cumulative ignition probability curve of each sample and the explosion result C of natural gas leakage in the comprehensive pipe gallery_iThe possibilities of (a) are:

L_i(C_i)＝F_f×Cum(Q9_i)×P_ti (6)

when mixing C_iDefined as the explosion overpressure peak P_maxTime, ignition source parameter s_pAll (P) were calculated according to formula (6) at 0.01, 0.05 and 0.1_max,L_i) The upper envelope of the sample points is extracted and the transcendental probability curve of an overpressure peak is plotted as shown in fig. 12, representing the maximum probability of causing a certain overpressure peak, and similarly the overpressure probability curve of the impulse can be calculated.

According to the comprehensive pipe gallery design diagram shown in fig. 7(a), a geometric model shown in fig. 7(b) is established, in order to make the numerical model more accurate, escape openings at two ends of a gas cabin (200 x2 x 4m), self-service and standby ventilation openings (0.8 x 0.8m) and a wind pavilion at the upper part are considered, and a DN300 medium-pressure (0.4MPa) gas pipeline, buttresses and a ladder stand are included inside the comprehensive pipe gallery as typical conventional facilities.

And the steady-state leakage flow range of the large holes calculated according to the leakage model is between 0.26 and 13.60kg/s, and 4 large hole leakage flow representative values are selected in the range according to an equal ratio principle in order to reduce the number of times of grid re-division in the simulation process. The 2-hole leakage flow rate was selected in the range of less than 0.26kg/s, and 13.60kg/s was used as the fracture leakage flow rate. Considering emergency brake valve operation at 1km intervals to isolate the leaking pipe section, the resulting unsteady leakage flow used in the simulation is shown in figures 8a-8 d. In order to keep the proportion of different leakage flows in the sample consistent with the proportion of failure frequency, the simulation results (combustible gas cloud) of the two types of small hole leakage flows are subjected to equal ratio interpolation. Since the + Y and-Y leakage directions have little difference, only the + Y direction is calculated in the simulation. The total number of simulation conditions and the total number of samples are shown in fig. 9.

The maximum values of FLAM and Q9 of each sample for risk assessment are counted, and cumulative frequency curves are respectively drawn as shown in fig. 10, and values under certain cumulative frequency are shown in fig. 11. It can be seen that the maximum value of Q9 is 903.8m within 180s of calculation time³Only 5% of the probability will exceed 300m³Half of the probability is not more than 84.5m³Compared with 1600m for 200 × 2 × 4³The size of the fire-proof subareas is not large, the value of Q9 is not large generally, a stable flow field of 0.5-1 m/s can be formed in the cabin already at the accident ventilation rate of 5m/s, and the combustible gas cloud is further limited along with the time. The volume of the combustible gas cloud that can be formed is therefore limited, given the complete safety system. In the limited space of the comprehensive pipe gallery, the leakage flow and the size of the combustible gas cloud are not in positive correlation, and the equivalent combustible gas cloud generated by the leakage of the large hole of 20-42 mm is the largest, as shown in fig. 12.

Randomly extracting 700 data result files generated in the leakage diffusion simulationAnd performing real gas cloud explosion simulation on 420 gas clouds, and aiming at establishing a correlation between combustible gas clouds and explosion consequences. 294 samples successfully ignited at the final ignition position of 420 real gas clouds selected randomly. Overpressure peaks in all sample calculations were extracted and plotted against the equivalent combustible gas cloud Q9 as shown in fig. 11. Due to the influence of the factors on the explosion result, phenomena such as two pressure peaks or incomplete consumption of fuel can occur in the explosion simulation of the real combustible gas cloud, and it can also be observed through the simulation result that the explosion overpressure peak value when the middle part of the massive combustible gas cloud is ignited is generally large, so that the correlation between Q9 and the overpressure peak value is established by selecting the upper bound 31 results of the sample points, which is conservative, as shown in FIG. 13 (a). The fitting result is (R)²＝0.96)：

P_max＝3.749×(Q9)^0.719 (7)

Q9 dependence on detonation impulse As shown in FIG. 13(b), all samples were fitted and the result was (R)²＝0.88)：

I＝1.624×(Q9)^0.746 (8)

The transcendental probability curve of an overpressure peak is plotted according to equation 7, representing the maximum probability of causing a certain overpressure peak, as shown in fig. 14 (a). It can be seen that with s_pGradually increasing, and gradually slowing the rising speed of the transcendental probability curve. When the control of the fire source in the pipe gallery is poor(s)_p0.1), the transcendental probability curve approaches a certain upper bound. When the ignition factor can be effectively controlled, sp may be taken to 0.01, otherwise, sp may be set conservatively to a larger value. The transcendental probability curve of the explosion impulse obtained according to the same principle is shown in fig. 14 (b).

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.