A method for analyzing the cumulative deformation curve of a natural gas pipeline landslide and predicting the landslide

Abstract

The present disclosure provides a method for analyzing the cumulative deformation curve of a natural gas pipeline landslide and predicting the landslide, belonging to the field of prediction and prevention of geological disasters of natural gas pipelines, including: obtaining the monitoring data of the cumulative deformation of the natural gas pipeline landslide, and preprocessing the monitoring data sequence and analyzing the trend of the deformation curve; establishing the grey differential equation of the grey theory model GM(1,1), obtaining the time response sequence equation of the monitoring data of the cumulative deformation of the natural gas pipeline landslide, and conducting the residual test of the grey model GM(1,1); outputting the predicted value of the grey model GM(1,1), assigning weights to each point of the single prediction model in the order of high to low accuracy, obtaining the minimum value reached by the sum of the squares of the reciprocals of the prediction data errors of the nth period, establishing a dynamic grey deformation model based on the induced ordered weighted harmonic (IOWHA) operator, and predicting the deformation of the natural gas pipeline landslide on the basis of the trend analysis of the cumulative deformation curve of the landslide, so as to achieve the effect of efficiently and scientifically improving the prediction accuracy of the deformation of the natural gas pipeline landslide.

Description

Technical Field

[0001] The present invention relates to the field of geological disaster prediction and prevention of natural gas pipelines, and specifically to a method for analyzing the cumulative deformation curve of natural gas pipeline landslides and predicting landslides. Background Art

[0002] Currently, for the methods of analyzing and predicting the cumulative deformation curve and deformation trend of natural gas pipeline landslides, it has developed from the qualitative analysis of each stage of pipeline landslide deformation based on the displacement-time curve to the quantitative intelligent prediction and analysis using mathematical models, and the intelligent mathematical models are gradually increasing, such as the grey prediction model, time series prediction model, genetic algorithm prediction model, BP neural network prediction model, etc.

[0003] However, most of the intelligent mathematical models are mainly single prediction models, which have certain limitations, and the calculation process and steps are cumbersome and complex, and they cannot reflect the particularity of natural gas pipeline landslides. Therefore, the application of the combined prediction method can not only make up for the limitations of the single prediction model, but also improve the prediction accuracy and calculation efficiency. Summary of the Invention

[0004] The purpose of the present disclosure is to address the limitations of single prediction models in the prior art and the problems such as large uncertainties in grey prediction theory. By introducing the induced ordered weighted harmonic (IOWHA) operator, a combined prediction model for natural gas pipeline landslide deformation based on the induced ordered weighted harmonic (IOWHA) operator and dynamic grey time series is established, which effectively improves the prediction accuracy of natural gas pipeline landslide deformation and provides technical support for natural gas pipeline landslide deformation prediction and disaster prevention.

[0005] To achieve the above object, the present disclosure is implemented through the following technical solutions: A method for analyzing the cumulative deformation curve of natural gas pipeline landslides and predicting landslides mainly includes the following steps:

[0006] Step 1: Obtain the monitoring data of natural gas pipeline landslide deformation, and perform preprocessing and cumulative deformation curve analysis on the monitoring data sequence;

[0007] Step 2: Establish the grey differential equation of the grey theory model GM(1,1), obtain the time response sequence equation of the natural gas pipeline landslide deformation monitoring data, and perform residual test on the grey model GM(1,1);

[0008] Step 3: Output the predicted values of the grey model GM(1,1), introduce the induced ordered weighted harmonic (IOWHA) operator, and assign weights to each point of the single prediction model according to the order of accuracy from high to low;

[0009] Step 4: Obtain the minimum value reached by the sum of the squares of the reciprocals of the prediction data errors in the nth period, establish a dynamic grey deformation model based on the induced ordered weighted harmonic (IOWHA) operator, and predict the deformation of the natural gas pipeline landslide;

[0010] Preferably, in step 1, a series of random original data obtained from deformation monitoring is acquired by arranging surface deformation monitoring sensors for the natural gas pipeline landslide;

[0011] Preferably, the preprocessing of the monitored natural gas pipeline landslide deformation data sequence is carried out according to formulas (1) and (2), and the specific implementation method is as follows:

[0012] Assume that the original data sequence of the natural gas pipeline landslide deformation monitoring is:

[0013] X (0) ={x (0) (1),x (0) (2),...,x (0) (n)} (1)

[0014] By performing cumulative generation on formula 1, the following can be obtained after preprocessing the monitoring data sequence:

[0015] X (1) ={x (1) (1),x (1) (2),...,x (1) (n)} (2)

[0016] In the above formula, x(0) is the original value of the natural gas pipeline landslide deformation monitoring, and x(1) is the data sequence generated after one cumulative addition;

[0017] Preferably, step 2 is specifically as follows: First, for the grey number sequence generated after preprocessing the original data of the natural gas pipeline landslide deformation, establish a grey differential equation of a single grey theory model GM(1,1); then whiten the grey differential equation to obtain the time response sequence equation of the coloured differential equation; further, conduct a one-to-one residual test on the predicted value of the grey theory model and the on-site actual monitoring value of the natural gas pipeline landslide deformation, and the specific implementation is as follows:

[0018] Assume that Z(1) is the grey number sequence generated from X(1) after preprocessing:

[0019]

[0020] In the above formula: k is a natural number 1, 2, 3…

[0021] According to the grey number sequence after preprocessing the natural gas pipeline landslide deformation monitoring data, establish a grey differential equation of a single grey theory model GM(1,1):

[0022] x (0) (k)+az (1) (k)=b (4)

[0023] In the above formula: a is the development coefficient of the grey prediction model for the landslide deformation monitoring of the natural gas pipeline, and b is its acting strain variable.

[0024] By whitening the grey differential equation, since and letting x (1) (0)=x (0) (1), the time response sequence equation of the grey differential equation for the landslide deformation prediction of the natural gas pipeline can be obtained:

[0025]

[0026] Perform a one-to-one residual test on the difference between the predicted value of the grey theory model for the landslide deformation of the natural gas pipeline and the actual on-site monitoring value, and calculate the absolute residual of the original data x (0) (i) and :

[0027]

[0028] Preferably, the specific implementation of step three is as follows: Through the prediction of the grey theory GM(1,1) model for the landslide deformation of the natural gas pipeline, introduce the induced ordered weighted harmonic (IOWHA) operator, and assign weights to each point of the single grey theory GM(1,1) model according to the accuracy order. The specific implementation is as follows:

[0029] At a certain moment, the landslide deformation prediction of the natural gas pipeline based on the combined model of the IOWHA operator is:

[0030]

[0031] In the above formula, xit (i = 1, 2) is the i-th prediction model, Pit represents the prediction accuracy value of the i-th prediction model at the t-th moment, and p-index(it) is the subscript of the larger value of the prediction accuracy of the single prediction method.

[0032] At the t-th moment, a larger weight coefficient is assigned to the method with higher prediction accuracy. Assume

[0033] e a-index(it) =1 / x t -1 / x p-index(it) (8)

[0034] In the above formula, a is the noise sequence;

[0035] Preferably, the specific implementation of step four is as follows:

[0036] According to the IOWHA combined prediction theory, the minimum value of the sum of the squares of the reciprocals of the prediction data errors in the nth period of the landslide deformation of the natural gas pipeline is obtained:

[0037]

[0038] In the above formula, \(x^t\) is the weighted harmonic average at time \(t\), and \(l=(l_1, l_2)^T\) is the combined prediction weighting coefficient;

[0039] A dynamic grey time series prediction model for the landslide deformation of oil and gas pipelines based on the induced ordered weighted harmonic (IOWHA) operator is established as:

[0040]

[0041] According to the prediction model established by the above formula, based on the landslide deformation data of the natural gas pipeline obtained by monitoring, the landslide deformation is predicted.

[0042] In summary, the beneficial technical effects of the present invention are as follows:

[0043] Aiming at the fact that it is difficult for the single grey model GM(1,1) prediction method to simultaneously consider the geological complexity of the landslide of the natural gas pipeline and the high consequences of the occurrence of disasters, the induced ordered weighted harmonic (IOWHA) operator is introduced, and the landslide prediction values of the single method are combined in a weighted average manner to establish an IOWHA combined prediction model, which can realize the dynamic prediction of the landslide deformation of the natural gas pipeline, and can effectively, efficiently and scientifically improve the prediction accuracy of the landslide deformation of the natural gas pipeline. Description of the Drawings

[0044] Figure 1 It is a flowchart of the combined prediction method for the landslide deformation of the natural gas pipeline provided by the embodiments of the present disclosure;

[0045] Figure 2 It is a comparison chart of the single prediction model, the combined prediction model and the actual monitoring values in the specific implementation of the present disclosure;

[0046] Figure 3 It is an error process chart in the specific embodiments of the present disclosure. Detailed Embodiments

[0047] Next, the technical solutions in the embodiments of the present disclosure will be clearly and completely described in conjunction with the drawings.

[0048] A combined prediction method for the landslide deformation of a natural gas pipeline includes the following steps:

[0049] Step 1: Deploy deformation monitoring sensors on the landslide of the natural gas pipeline, obtain the original monitoring data within a certain period, and preprocess the monitoring data sequence:

[0050] The preprocessing of the natural gas pipeline landslide deformation data series obtained by monitoring is carried out according to Formulas (1) and (2), and the specific implementation method is as follows:

[0051] Assume that the original data series of the natural gas pipeline landslide deformation monitoring is:

[0052] X (0) ={x (0) (1),x (0) (2),...,x (0) (n)} (1)

[0053] By cumulatively generating Formula 1, the following can be obtained after preprocessing the monitoring data series:

[0054] X (1) ={x (1) (1),x (1) (2),...,x (1) (n)} (2)

[0055] In the above formula, x(0) is the original value of the natural gas pipeline landslide deformation monitoring, and x(1) is the data series generated after one cumulative addition.

[0056] Step 2: Establish the grey differential equation of the grey theory model GM(1,1), obtain the time response sequence equation of the natural gas pipeline landslide deformation monitoring data, and conduct the residual test of the grey model GM(1,1);

[0057] For the grey number sequence generated after preprocessing the original data of the natural gas pipeline landslide deformation, establish the grey differential equation of a single grey theory model GM(1,1); then whiten the grey differential equation to obtain the time response sequence equation of the coloured differential equation; further conduct a one-to-one residual test on the predicted value of the grey theory model and the on-site actual monitoring value of the natural gas pipeline landslide deformation. The specific implementation is as follows:

[0058] Assume that Z(1) is the grey number sequence generated from X(1) after preprocessing:

[0059]

[0060] In the above formula: k is a natural number 1, 2, 3…

[0061] According to the grey number sequence after preprocessing the natural gas pipeline landslide deformation monitoring data, establish the grey differential equation of a single grey theory model GM(1,1):

[0062] x (0) (k)+az (1) (k)=b (4)

[0063] In the above formula: a is the development coefficient of the grey prediction model for the landslide deformation monitoring of natural gas pipelines, and b is its acting response variable.

[0064] By whitening the grey differential equation, since and letting x (1) (0) = x (0) (1), the time response sequence equation of the grey differential equation for the landslide deformation prediction of natural gas pipelines can be obtained:

[0065]

[0066] Perform a one-to-one residual test on the difference between the predicted values of the grey theory model for the landslide deformation of natural gas pipelines and the actual on-site monitoring values, and calculate the absolute residual of the original data x (0) (i) and :

[0067]

[0068] Step 3: Output the predicted values of the grey model GM(1,1), introduce the induced ordered weighted harmonic (IOWHA) operator, and assign weights to each point of the single prediction model according to the order of accuracy;

[0069] Through the grey theory GM(1,1) model prediction of the landslide deformation of natural gas pipelines, introduce the induced ordered weighted harmonic (IOWHA) operator, and assign weights to each point of the single grey theory GM(1,1) model according to the accuracy order. The specific implementation is as follows:

[0070] At a certain moment, the landslide deformation prediction of the natural gas pipeline based on the combined model of the IOWHA operator is:

[0071]

[0072] In the above formula, xit (i = 1, 2) is the i-th prediction model, Pit represents the prediction accuracy value of the i-th prediction model at the t-th moment, and p-index(it) is the subscript of the larger value of the prediction accuracy of the single prediction method.

[0073] At the t-th moment, a larger weight coefficient is assigned to the method with higher prediction accuracy. Assume

[0074] e a-index(it) = 1 / x t - 1 / x p-index(it) (8)

[0075] In the above formula, a is the noise sequence.

[0076] Step 4: Obtain the minimum value reached by the sum of the squares of the reciprocals of the prediction data errors in the nth period, establish a dynamic grey deformation model based on the induced ordered weighted harmonic (IOWHA) operator, and predict the deformation of the natural gas pipeline landslide;

[0077] According to the IOWHA combined prediction theory, obtain the minimum value of the sum of the squares of the reciprocals of the prediction data errors of the natural gas pipeline landslide deformation in the nth period:

[0078]

[0079] In the above formula, \(x^t\) is the weighted harmonic mean at time \(t\), and \(l=(l_1, l_2)^T\) is the combined prediction weighting coefficient.

[0080] Establish a dynamic grey time series oil and gas pipeline landslide deformation prediction model based on the induced ordered weighted harmonic (IOWHA) operator as:

[0081]

[0082] According to the prediction model established by the above formula, based on the monitored natural gas pipeline landslide deformation data, predict the deformation of the landslide.

[0083] Select an error evaluation index according to the prediction results, as shown in Table 1, establish a prediction error system, and quantitatively analyze the accuracy of the prediction.

[0084] Table 1 Error evaluation table

[0085]

[0086] The above-mentioned combined prediction method for natural gas pipeline landslide deformation proposed in this disclosure is illustrated by the following examples:

[0087] Example 1

[0088] Use the method proposed in this disclosure to conduct an example analysis of combined deformation prediction on the Liujiaao landslide in Longzhouping Town along a certain natural gas pipeline.

[0089] The Liujiaao Landslide is located in the third group of Liujiaao, Longzhouping Town, Yichang City. The slope is a soil slope (crushed stone soil) with a predicted volume of 1000 m³. The pipeline is located below the slope and runs through the slope. The pipeline alignment is 170°, with a buried depth of 2.0 m. There is a masonry retaining wall at the bottom, and there are local cracks in the shape of an "eight". The overlying Quaternary cover layer on the slope is relatively thick, with a thickness of 3 - 8 m. The surface layer is mainly composed of loose and soft Quaternary residual slope deposits (Q), which are composed of yellowish-brown clay, silty clay, and crushed stones. Secondly, there are artificial accumulations, mainly distributed on the slope surface. The soil is loose and soft, with poor drainage and easy to be saturated and deformed. Geological disasters such as landslides are likely to occur during heavy rainfall. In 2019, a large-scale distributed potential deformation displacement meter was installed on the landslide to monitor the surface deformation of the pipeline landslide. One month of monitoring data from August 6, 2019, to September 7, 2019, was selected as the original sample, as shown in Table 2. The deformation after September 7, 2019, was selected for prediction. The predicted values and prediction errors of the landslide deformation are shown in Tables 3 and 4 respectively. The comparison of the single prediction model, combined prediction model, and actual monitoring values is as Figure 2 shown, and the absolute error process is as Figure 3 shown.

[0090] Table 2 Selected Original Sample Data

[0091]

[0092] Table 3 Measured Values and Predicted Values of Landslide Deformation

[0093]

[0094]

[0095] Table 4 Prediction Errors of Landslide Deformation

[0096]

[0097] As can be seen from Table 4, the maximum absolute prediction error of the combined prediction model established in this paper is 0.54, and the maximum absolute prediction error of the single gray theory GM(1,1) model is 0.72. Obviously, the prediction error established in this paper is relatively small and closer to the actual monitoring value.

[0098] It should be understood that those skilled in the art can make improvements based on the above description, and all such improvements should fall within the protection scope of the appended claims of this disclosure.

Claims

1. A method for analyzing the cumulative deformation curve of a natural gas pipeline landslide and predicting the landslide, characterized in that: It includes the following steps: Step 1: Obtain the cumulative deformation monitoring data of the natural gas pipeline landslide, and preprocess the monitoring data sequence and analyze the trend of the deformation curve. Step 2: Establish the grey differential equation of the grey theory model GM(1,1), obtain the time response sequence equation of the natural gas pipeline landslide deformation monitoring data, and conduct the residual test of the grey theory model GM(1,1). The specific process is to obtain the minimum value of the sum of the squares of the reciprocals of the prediction data errors of the nth period of the natural gas pipeline landslide deformation according to the IOWHA combined prediction theory: In the above formula, x^t is the weighted harmonic mean at time t, and l=(l1,l2)T is the combined prediction weighting coefficient. Establish a dynamic grey time series oil and gas pipeline landslide deformation prediction model based on the induced ordered weighted harmonic (IOWHA) operator: ； Step 3: Output the prediction value of the grey theory model GM(1,1), introduce the induced ordered weighted harmonic (IOWHA) operator, and assign weights to each point of the single prediction model according to the order of accuracy. The specific process is to predict through the grey theory model GM(1,1) of the natural gas pipeline landslide deformation, introduce the induced ordered weighted harmonic (IOWHA) operator, and assign weights to each point of the single grey theory model GM(1,1) according to the accuracy order. The specific implementation is as follows: At a certain moment, the prediction of the natural gas pipeline landslide deformation by the combined model based on the IOWHA operator is: In the above formula, xit (i = 1,2) is the ith prediction model, Pit represents the prediction accuracy value of the ith prediction model at the tth moment, and p-index(it) is the subscript of the larger value of the prediction accuracy of the single prediction method. At time t, a larger weight coefficient is assigned to the method with higher prediction accuracy. Assume e a-index(it) = 1 / xt - 1 / x p-index(it) (8) In the above formula, a is the noise sequence. Step 4: Obtain the minimum value reached by the sum of the squares of the reciprocals of the prediction data errors of the nth period, establish a dynamic grey deformation model based on the induced ordered weighted harmonic (IOWHA) operator, and predict the deformation of the natural gas pipeline landslide on the basis of the trend analysis of the cumulative deformation curve.

2. The method for analyzing the cumulative deformation curve of a natural gas pipeline landslide and predicting the landslide according to claim 1, wherein: In Step 1, a series of random original data obtained from deformation monitoring need to be obtained through the surface deformation monitoring sensors of the natural gas pipeline landslide, and the monitoring data sequence of the natural gas pipeline landslide deformation is preprocessed and the trend of the deformation curve is analyzed according to Formulas (1) and (2). The specific implementation is as follows: Assume that the original data sequence of the natural gas pipeline landslide deformation monitoring is: X(0) = {x(0)(1),x(0)(2),...,x(0)(n)} (1) After performing the cumulative generation on Formula 1 and preprocessing the monitoring data sequence, the following can be obtained: X(1) = {x(1)(1),x(1)(2),...,x(1)(n)} (2) In the above formula, x(0) is the original value of the natural gas pipeline landslide deformation monitoring, and x(1) is the data sequence generated after one cumulative addition.

3. A method for analyzing the cumulative deformation curve of a natural gas pipeline landslide and predicting the landslide according to claim 1, characterized in that: The specific process of the second step is as follows: First, for the grey number sequence generated after the preprocessing of the original data of the landslide deformation of the natural gas pipeline, a single grey differential equation of the grey theory model GM(1,1) is established; then the grey differential equation is whitened to obtain the time response sequence equation of the coloured differential equation; further, a one-to-one residual test is carried out on the predicted value of the grey theory model and the actual on-site monitoring value of the landslide deformation of the natural gas pipeline. The specific implementation is as follows: Assume that Z(1) is the grey number sequence generated by X(1) after preprocessing: In the above formula: k is a natural number 1, 2, 3... According to the grey number sequence after the preprocessing of the landslide deformation monitoring data of the natural gas pipeline, establish a single grey differential equation of the grey theory model GM(1,1): x(0)(k) + az(1)(k) = b(4) In the above formula: a is the development coefficient of the grey prediction model for the landslide deformation monitoring of the natural gas pipeline, and b is its acting response variable; Next, whiten the grey differential equation. Since and let \(x^{(1)}(0)=x(0)(1)\), the time response sequence equation of the grey differential equation for predicting the landslide deformation of the natural gas pipeline can be obtained: Carry out a one-to-one residual test on the difference between the predicted value of the grey theory model of the landslide deformation of the natural gas pipeline and the actual on-site monitoring value, and calculate the absolute residual between the original data x(0)(i) of the landslide deformation monitoring of the pipeline and 。

Images