Model for predicting five-factor double-competitive strain of high-steel-grade pipeline under action of landslide

Abstract

The invention belongs to the field of pipeline safety assessment, and relates to a five-factor double-competitive strain prediction model for a high-steel-grade pipeline under the action of landslide. According to the method, a three-dimensional finite element model of the high-steel-grade pipeline under the landslide action is established through a finite element method, the maximum axial strain of the pipeline under different pipe diameters, wall thicknesses, internal pressures, landslide widths and landslide sectional areas is simulated and calculated, and a five-factor double-competition strain prediction model of the high-steel-grade pipeline under the landslide action is established according to a simulation result; according to the model, the coupling effect among the pipe diameter, the wall thickness, the landslide width and the landslide sectional area is considered, the maximum axial strain of the left side and the right side of the pipeline can be calculated at the same time, and therefore the maximum axial strain of the pipeline is determined. Through verification, the five-factor double-competition strain prediction model for the high-steel-grade pipeline under the landslide effect is high in prediction precision, an effective mathematical calculation method is provided for strain prediction of the high-steel-grade pipeline under the landslide effect, and practicability is high.

Description

Technical Field

[0001] The invention belongs to the field of pipeline safety assessment and relates to a five-factor double-competition strain prediction model for high-grade steel pipelines under landslide action. Background Art

[0002] With the continuous development of oil and natural gas, the total mileage of oil and gas pipelines continues to increase, among which high-grade steel materials are widely used in long-distance natural gas pipelines. During the laying process of long-distance natural gas pipelines, it is inevitable that they will pass through mountainous areas prone to landslides. The risk of fracture failure of high-grade steel pipelines under the action of landslides increases. If the pipeline breaks and fails, it will not only bring serious economic losses, but also cause environmental pollution. Therefore, it is necessary to study the prediction method of the mechanical behavior of high-grade steel pipelines under the action of landslides.

[0003] At present, the oil and gas industry generally studies the stress changes of pipelines under the action of landslides. However, high-grade steel materials have good ductility and elongation of more than 10%. The stress cannot fully reflect the mechanical behavior of high-grade steel pipelines. The mechanical behavior of high-grade steel pipelines should be studied based on the strain.

[0004] The strain of high-grade steel pipelines under landslides is affected by multiple parameters, and some factors interact with each other, which makes strain prediction very difficult. At present, many scholars have studied strain prediction models. He et al. used genetic-back propagation neural network to establish a stress prediction model for pipelines with circumferential welds under landslides in the three-point direction. Xiang et al. established a mathematical model for strain prediction in the three-point direction of pipelines based on the strain simulation results of pipelines with landslides under different influencing factors. Liu Peng et al. combined pipeline landslide tests and numerical simulations and used machine learning methods to establish a pipeline mechanical behavior prediction model.

[0005] At present, most prediction models of pipeline mechanical behavior under landslides rely on machine learning algorithms, but mathematical models of strain are relatively scarce. Therefore, based on the simulation results of the three-dimensional finite element model of high-grade steel pipelines under landslides, the present invention establishes a five-factor dual-competition strain prediction model for high-grade steel pipelines under landslides, providing a new method for the field of pipeline strain prediction models. Summary of the invention

[0006] Aiming at the deficiencies of the existing strain prediction models, the present invention proposes a five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action.

[0007] The establishment process of a five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action mainly includes the following four steps:

[0008] S101: Considering the nonlinear discontinuous effect between the landslide and the pipeline, a three-dimensional finite element model of the high-grade steel pipeline under the action of the landslide is established by the finite element method;

[0009] S102: According to the main influencing factors of the axial strain of the landslide pipeline, the maximum axial strain of the pipeline at 3 points (close to the landslide direction) and 9 points (away from the landslide direction) under different pipe diameters, wall thicknesses, internal pressures, landslide widths and landslide cross-sectional areas is calculated respectively;

[0010] S103: Establish a five-factor maximum axial strain prediction function model for high-grade steel pipelines under landslide action. ε=a·f(D)+b·f(t)+c·f(L)+d·f(A)+e·f(P)+C

[0011] In the formula: ε—maximum axial strain of the pipeline; a~e—coefficients; f(x)—functional relationship between x factor and strain; C—constant term;

[0012] S104: According to the finite element simulation results, a nonlinear regression algorithm is used to determine the coefficient term and constant term of the maximum axial strain prediction function model.

[0013] Where: ε3—maximum axial strain at 3 o'clock in the pipeline, dimensionless; ε9—maximum axial strain at 9 o'clock in the pipeline, dimensionless; D—pipeline diameter, mm; t—pipeline wall thickness, mm; L—landslide width, m; A—landslide cross-sectional area, m 2 ; P—pipeline internal pressure, MPa.

[0014] The finite element simulation and prediction model considered the effect of the landslide cross-sectional area on the maximum axial strain of the pipeline.

[0015] The prediction model can calculate the maximum axial strain of the pipeline at 3 and 9 points respectively, and determine the true maximum axial strain of the pipeline by comparing the two results.

[0016] A five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action takes into account the coupling effect between pipe diameter, wall thickness, landslide width and landslide cross-sectional area, and establishes a coupling model between pipe diameter, wall thickness, landslide width and landslide cross-sectional area BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a linear fitting diagram of the predicted values ​​and simulated values ​​in the three-point directions of the pipeline in a five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action.

[0018] Figure 2This is a linear fitting diagram of the predicted values ​​and simulated values ​​in the 9-point direction of the pipeline in a five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action. DETAILED DESCRIPTION

[0019] A specific implementation of a five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action is as follows:

[0020] Step 1: Considering the nonlinear discontinuous effect between the landslide and the pipeline, a three-dimensional finite element model of the high-grade steel pipeline under the action of the landslide is established by the finite element method.

[0021] Step 2: According to the main influencing factors of the axial strain of the landslide pipeline, the finite element simulation and prediction model both consider the influence of the landslide cross-sectional area on the maximum axial strain of the pipeline, and calculate the different pipe diameters (609, 800, 1016, 1219, 1422mm), wall thickness (10.6, 13.6, 15.6, 18.4, 21.8mm), internal pressure (4, 6, 8, 10, 12MPa), landslide width (10, 20, 30, 40, 50m) and landslide cross-sectional area (28, 55, 64, 96, 114m 2 )The maximum axial strain of the pipeline at 3 points (close to the landslide direction) and 9 points (away from the landslide direction).

[0022] Step 3: Establish a five-factor maximum axial strain prediction function model for high-grade steel pipelines under landslide action. ε=a·f(D)+b·f(t)+c·f(L)+d·f(A)+e·f(P)+C

[0023] In the formula: ε is the maximum axial strain of the pipeline; a~e are coefficients; f(x) is the functional relationship between the x factor and the strain; C is the constant term.

[0024] Step 4: Consider the coupling effect between pipe diameter, wall thickness, landslide width and landslide cross-sectional area, and establish a coupling model between pipe diameter, wall thickness, landslide width and landslide cross-sectional area; according to the finite element simulation results, use the nonlinear regression algorithm to determine the coefficient term and constant term of the maximum axial strain prediction function model. The prediction model can calculate the maximum axial strain at the 3-point and 9-point directions of the pipeline respectively. By comparing the two results, the true maximum axial strain of the pipeline is determined.

[0025] Where: ε3—maximum axial strain at 3 o'clock in the pipeline, dimensionless; ε9—maximum axial strain at 9 o'clock in the pipeline, dimensionless; D—pipeline diameter, mm; t—pipeline wall thickness, mm; L—landslide width, m; A—landslide cross-sectional area, m 2 ; P—pipeline internal pressure, MPa.

[0026] The linear fitting diagrams of the predicted values ​​and simulated values ​​at 3 and 9 points of the pipeline in a five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide are shown in Figure 2. Figure 1 and Figure 2 , the slopes of the fitting lines are 0.954 and 0.969, respectively, indicating that the prediction model has high accuracy.

[0027] The practicality of the five-factor dual-competition strain prediction model for a high-grade steel pipeline under landslide was verified through field application. According to the field landslide pipeline parameters in Table 1, the maximum relative error of the model strain prediction is 11.36%, and the average relative error is 9.78%, indicating that the model can accurately predict the maximum axial strain and its position of the landslide pipeline, and has good practicality.

[0028] Table 1 Parameters of on-site landslide pipeline Pipe diameter / mm Wall thickness / mm Landslide width / m <![CDATA[Landslide cross-sectional area / m 2 > Internal pressure 1016 15.6 25 35 6

[0029] Table 2 Strain detection and prediction results of on-site landslide pipeline

Claims

1. A five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action, characterized in that: The process includes the following: S101: Considering the nonlinear discontinuous effect between the landslide and the pipeline, a three-dimensional finite element model of the high-grade steel pipeline under the action of the landslide is established by the finite element method; S102: According to the main influencing factors of the axial strain of the landslide pipeline, the maximum axial strain of the pipeline at 3 points (close to the landslide direction) and 9 points (away from the landslide direction) under different pipe diameters, wall thicknesses, internal pressures, landslide widths and landslide cross-sectional areas is calculated respectively; S103: Establish a five-factor maximum axial strain prediction function model for high-grade steel pipelines under landslide action. ε=a·f(D)+b·f(t)+c·f(L)+d·f(A)+e·f(P)+C In the formula: ε—maximum axial strain of the pipeline; a~e—coefficients; f(x)—functional relationship between x factor and strain; C—constant term; S104: According to the finite element simulation results, a nonlinear regression algorithm is used to determine the coefficient term and constant term of the maximum axial strain prediction function model. Where: ε3—maximum axial strain at 3 points of the pipeline, dimensionless; ε9—maximum axial strain at 9 o'clock in the pipeline, dimensionless; D—pipeline diameter, mm; t—pipeline wall thickness, mm; L—landslide width, m; A—landslide cross-sectional area, m 2 ; P—pipeline internal pressure, MPa.

2. A five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action as described in claim 1, characterized in that: The finite element simulation and prediction model considered the effect of the landslide cross-sectional area on the maximum axial strain of the pipeline.

3. The five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action as described in claim 1 is characterized in that: The prediction model can calculate the maximum axial strain of the pipeline at 3 and 9 points respectively, and determine the true maximum axial strain of the pipeline by comparing the two results.

4. The five-factor dual-competition strain prediction model for high-grade steel pipelines under landslide action as described in claim 1 is characterized in that: The prediction model takes into account the coupling effects among pipe diameter, wall thickness, landslide width and landslide cross-sectional area: A coupling model between pipe diameter, wall thickness, landslide width and landslide cross-sectional area was established

Images