Method for predicting corrosion rate of 13Cr pipe in carbon dioxide environment

Through the combination of orthogonal matrix method and decision tree model, the problem of insufficient corrosion rate prediction of 13Cr pipes in the CO2 environment in the prior art is solved, and the high-accurate corrosion rate prediction in complex environments is achieved, which is suitable for petroleum Industry's CO2 oil wells and other scenarios.

CN119069048AActive Publication Date: 2024-12-03HAINAN BRANCH OF CHINA NATIONAL OFFSHORE OIL (CHINA) CO LTD

Patent Information

Application Number
CN202411228362.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-12-03
Estimated Expiration
2044-09-03

AI Technical Summary

Technical Problem

The prior art has low accuracy in predicting corrosion rate of 13Cr pipes in CO2 environment, especially in gas-phase corrosion environments and complex fluid flow conditions, and the combined effect of multiple corrosion factors is not considered.

Method used

The orthogonal matrix method is used to conduct gas phase corrosion indoor experiments with multi-factor data matching combinations, establish a decision tree model, and realize the corrosion rate prediction of 13Cr pipes under gas phase dynamic corrosion conditions through training and testing data sets, taking into account the comprehensive role of multiple corrosion factors.

Benefits of technology

It improves the accuracy of the prediction of the corrosion rate of 13Cr pipes in complex environments, and can be closer to the pipeline corrosion rules in complex environments such as CO2 oil wells in the oil industry, reducing prediction errors.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119069048A_ABST
    Figure CN119069048A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of corrosion prediction of oil and gas pipelines, in particular to a method for predicting the corrosion rate of a 13Cr pipe in a carbon dioxide environment, which adopts an orthogonal matrix method to carry out a gas phase corrosion indoor experiment of the 13Cr pipe under multiple corrosion factors. A sample data set established by experimental multi-factor data and corrosion rate is imported into the decision tree model for training and testing, a target decision tree model is obtained, and corrosion rate prediction of the 13Cr pipe under the gas phase dynamic corrosion condition is achieved. The influence of the comprehensive action of multiple corrosion factors on the corrosion rate of the pipe is considered, the method is closer to the corrosion rule of the pipe in the complex environment containing CO2 in a shaft, such as a CO2 flooding well, a CO2 storage well and a CO2 reverse-spitting production well, in the petroleum industry, and the accuracy of the corrosion rate prediction result of the 13Cr pipe is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas pipeline corrosion prediction, and more specifically, to a method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment. Background Art

[0002] CCUS (Carbon Capture Utilization and Storage) technology is one of the important technical means to achieve the "dual carbon" goal. The application of this technology is mainly reflected in the continuous supercritical CO2 fracturing transformation of unconventional reservoirs in various oil fields, CO2 drive to improve crude oil recovery, and CO2 geological storage in offshore or onshore depleted oil and gas reservoirs. Since CO2 is acidic, it is easy to corrode downhole pipes when flowing in the wellbore, causing the wall thickness of the pipe to be thin, resulting in reduced pressure bearing capacity of the wellbore and seal leakage. In severe cases, downhole accidents such as shedding occur, resulting in a decrease in oil well production or even shutdown, which seriously affects the normal production of the oil field. At the same time, with the blind and frequent replacement of oil casing, the cost of well repair continues to increase, and the development cost rises. Therefore, it is necessary to predict the corrosion rate of pipes.

[0003] At present, 13Cr pipes are widely used in CO2 injection wells in oil fields. The prediction models of 13Cr pipe corrosion rate in CO2 environment have the following main problems: (1) The existing corrosion rate prediction methods are mainly empirical and semi-empirical models, which are suitable for general carbon steel. Carbon steel is prone to carbonic acid corrosion and electrochemical corrosion, and the corrosion product film formed is mainly composed of thick and loose FeCO3. Compared with low-Cr pipes such as 1Cr and 3Cr, 13Cr pipes have a higher Cr content and are more easily oxidized to form a dense Cr2O3 film. The differences in the properties of the corrosion product film and its formation mechanism lead to the inability of existing models to predict the corrosion rate of 13Cr pipes. There will be a big difference; (2) The existing corrosion rate prediction methods are all derived under liquid phase corrosion conditions. Liquid phase corrosion mainly occurs in the part of the pipe string below the packer in the reservoir that is in contact with the formation water, while the rest of the pipe string is mainly corroded by mixed gases such as CO2 gas and condensate water, which is a gas phase corrosion environment. At present, there is a lack of 13Cr pipe corrosion rate prediction model that meets the gas phase corrosion environment; (3) The existing corrosion rate prediction method does not consider the effect of fluid flow on the corrosion rate, and cannot achieve dynamic prediction of corrosion rate under the flow conditions of corrosive media and over time. When ordinary carbon steel pipes or anti-corrosion pipes are in service underground, corresponding corrosion products will be formed on the surface. When the corrosion product morphology is unstable or there is a high fluid flow rate, the corrosion product will be in a state of equilibrium between generation and dissolution. At this time, the corrosion mode is uniform corrosion; when the corrosion product structure gradually becomes dense, the corrosion product can effectively prevent the contact between the corrosive medium and the pipe as a protective film. At this time, the corrosion mode is deceleration corrosion.

[0004] The prior art discloses a six-factor local corrosion rate prediction model containing carbon dioxide and hydrogen sulfide, including: establishing a six-factor corrosion rate calculation function model that includes the common corrosion environment of oil and gas pipelines; combining the corrosion environment of oil and gas pipelines, there are situations where the six factors do not appear at the same time. When a factor does not exist, the function can automatically return to zero to calculate the corrosion rate values ​​of other factors; through high-temperature and high-pressure reactor experiments, three-dimensional microscope characterization of local corrosion rate data, determine the local corrosion rate under different working conditions, and use linear regression algorithm to determine the constant term in the calculation model of the local corrosion rate prediction model. This scheme takes into account the interactive corrosion effects of CO2 partial pressure and H2S partial pressure, and can calculate flow rate, Cl - The local corrosion rate of oil and gas pipelines, including concentration, CO2 partial pressure, H2S partial pressure, pH value and temperature. However, each corrosion environment has a coupling effect when affecting the corrosion reaction of the pipe, and it is not a linear superposition relationship. The function model in this scheme is the superposition of each influencing factor multiplied by the corresponding weight coefficient. When the corrosion factors fluctuate over a large range, the prediction result of pipeline corrosion rate is less accurate. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art in that the prediction results of pipeline corrosion rate are of low accuracy, and to provide a method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment, taking into account the influence of the comprehensive effect of various corrosion factors on the corrosion rate of the pipe, being closer to the corrosion law of the pipe in a complex environment, and improving the accuracy of the prediction results of the corrosion rate of the pipe.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment is provided, comprising the following steps:

[0008] S1: Establish an orthogonal experimental matrix of multi-factor data, conduct indoor test of 13Cr pipe gas phase corrosion under multi-factor data matching combination, and obtain the corrosion rate under multi-factor data matching combination;

[0009] S2: using the multi-factor data as an independent variable and the corrosion rate as a dependent variable, establishing a sample data set, and dividing the sample data set into a training set and a test set according to a preset ratio;

[0010] S3: training the pre-established decision tree model based on the training set to obtain a trained initial decision tree model;

[0011] S4: inputting the multi-factor data in the test set into the initial decision tree model, and outputting a test corrosion rate;

[0012] S5: Determine whether the error between the test corrosion rate and the corrosion rate in the test set is less than a preset error; if so, use the initial decision tree model as the target decision tree model and proceed to step S6; if not, modify the parameters of the pre-established decision tree model and return to step S3;

[0013] S6: Input the target multi-factor data into the target decision tree model to obtain the corrosion rate prediction result.

[0014] The method for predicting the corrosion rate of 13Cr pipes under a carbon dioxide environment of the present invention adopts an orthogonal matrix method to conduct an indoor experiment on the gas phase corrosion of 13Cr pipes under multiple corrosion factors, and imports the multi-factor data of the experiment and the sample data set established by the corrosion rate into a decision tree model for training and testing, thereby obtaining a target decision tree model and realizing the prediction of the corrosion rate of 13Cr pipes under gas phase dynamic corrosion conditions. The present invention considers the influence of the comprehensive effect of multiple corrosion factors on the corrosion rate of pipes, and is closer to the corrosion law of pipes in complex environments involving CO2 in wellbores such as CO2 flooding wells, CO2 storage wells, and production wells that spit back CO2 in the petroleum industry, thereby improving the accuracy of the prediction results of the corrosion rate of 13Cr pipes when the corrosion factors fluctuate over a large range.

[0015] Preferably, in step S1, the multi-factor data includes condition data affecting the CO2 corrosion reaction, including temperature, CO2 partial pressure, gas phase moisture content, Cl - concentration, gas flow rate and corrosion time.

[0016] Preferably, in step S1, the 13Cr pipe gas phase corrosion indoor test comprises the following steps:

[0017] S101: Milling the 13Cr pipe provided by the oil field to obtain a hanging piece, pre-treating the hanging piece, and weighing the pre-treated hanging piece to obtain a hanging piece mass G1;

[0018] S102: Put the weighed 13Cr hanging piece into the high temperature and high pressure reactor, and prepare the Cl-containing - Add simulated water into the high-temperature and high-pressure reactor, turn on the heating button to heat the reactor to the preset corrosion temperature, introduce CO2 into the reactor to reach the preset corrosion pressure, set and turn on the fluid flow rate controller in the reactor according to the production well output rate, and conduct 13Cr pipe corrosion experiments according to different preset corrosion cycles;

[0019] S103: After reaching the preset corrosion cycle, the corroded hanging piece is taken out, the hanging piece is post-processed, and the hanging piece is weighed to obtain the mass of the corroded hanging piece as G2;

[0020] S104: Calculate the gas phase corrosion rate of the 13Cr pipe in different corrosion cycles using the weight loss method according to the weighing data G1 of step S101 and the weighing data G2 of step S103.

[0021] Preferably, in step S101, the process of pre-treating the hanging piece is: using emery paper to gradually polish the surface of the hanging piece to reach a mirror state, washing it with deionized water and anhydrous alcohol in turn, and then placing it in a vacuum drying oven to dry.

[0022] Preferably, in step S102, during the process of introducing CO2, CO2 needs to be introduced to expel the air in the reactor, and then the outlet valve is closed, and then CO2 is continuously introduced to a preset corrosion pressure.

[0023] Preferably, in step S103, when taking out the corroded coupons, first turn off the fluid flow rate controller in the reactor, then turn off the heating button, wait until the temperature drops to room temperature, perform pressure relief, then open the reactor, remove the container containing the coupons in the reactor, take out the coupons with tweezers, dry the moisture remaining on the surface of the coupons with absorbent paper, and put them into experimental storage bags one by one.

[0024] Preferably, in step S103, the process of post-processing the coupon is: washing the surface of the coupon with distilled water to remove the corrosion residue Cr2O3 film, removing water with anhydrous alcohol and then drying the coupon.

[0025] Preferably, in step S104, the gas phase corrosion rate is calculated as follows:

[0026]

[0027] ΔG = G1 - G2;

[0028] Where V cor represents the average corrosion rate; ΔG represents the mass loss of the coupon; ρ represents the density of the coupon; S represents the contact area; and t represents the experimental time.

[0029] Preferably, in step S3, the process of pre-establishing a decision tree model is as follows: according to the decision tree machine learning principle, the corrosion rate is used as the root node, the feature information gain of the corrosion influencing factor is calculated, and the feature with the largest information gain is selected as the feature of the corresponding corrosion factor. Sub-nodes are established based on different values ​​of this feature, and the above operation is repeated to select different values ​​of the feature from the remaining corrosion influencing factors to establish the next sub-node until no feature is selected, thereby completing the construction of the decision tree.

[0030] Preferably, in step S6, the standard deviation rate of the corrosion rate at each time point is calculated based on the obtained corrosion rate prediction results. If the standard deviation rate of the corrosion rate prediction value at each time point does not exceed 10%, it is determined to be uniform corrosion, and the corrosion rate does not change with time, which is the prediction result of the decision tree model; if the standard deviation rate of the corrosion rate prediction value at each time point exceeds 10% and has a clear downward trend with time, it is determined to be variable-rate corrosion, and the corrosion rate results at different corrosion times are calculated using the functional relationship between the decision tree model learning result and the corrosion time.

[0031] Compared with the prior art, the method for predicting the corrosion rate of 13Cr pipes under carbon dioxide environment of the present invention has the following beneficial effects:

[0032] The combined effect of multiple corrosion factors on the corrosion rate of pipes is considered to be closer to the corrosion law of pipes in complex environments and to improve the accuracy of the prediction results of pipe corrosion rate when the corrosion factors fluctuate over a large range. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a flow chart of a method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment in an embodiment of the present invention;

[0034] Figure 2 This is a flow chart of the indoor test of 13Cr pipe gas phase corrosion in an embodiment of the present invention;

[0035] Figure 3 This is a graph showing the prediction results of the corrosion rate and cumulative corrosion depth of the CO2 storage well in the historical stages in Example 3 of the present invention. DETAILED DESCRIPTION

[0036] The present invention will be further described below in conjunction with specific implementation modes.

[0037] Embodiment 1

[0038] A method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment. Figure 1 As shown, the following steps are included:

[0039] S1: Establish an orthogonal experimental matrix of multi-factor data, conduct indoor test of 13Cr pipe gas phase corrosion under multi-factor data matching combination, and obtain the corrosion rate under multi-factor data matching combination;

[0040] S2: using the multi-factor data as an independent variable and the corrosion rate as a dependent variable, establishing a sample data set, and dividing the sample data set into a training set and a test set according to a preset ratio;

[0041] S3: training the pre-established decision tree model based on the training set to obtain a trained initial decision tree model;

[0042] S4: inputting the multi-factor data in the test set into the initial decision tree model, and outputting a test corrosion rate;

[0043] S5: Determine whether the error between the test corrosion rate and the corrosion rate in the test set is less than a preset error of 20%; if so, use the initial decision tree model as the target decision tree model and proceed to step S6; if not, modify the parameters of the pre-established decision tree model and return to step S3;

[0044] S6: Input the target multi-factor data into the target decision tree model to obtain the corrosion rate prediction result.

[0045] The above-mentioned method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment uses an orthogonal matrix method to conduct indoor experiments on gas phase corrosion of 13Cr pipes under multiple corrosion factors, and imports the experimental multi-factor data and the sample data set established by the corrosion rate into the decision tree model for training and testing, and obtains the target decision tree model to realize the prediction of the corrosion rate of 13Cr pipes under gas phase dynamic corrosion conditions. This embodiment considers the impact of the combined effect of multiple corrosion factors on the corrosion rate of pipes, and is closer to the corrosion law of pipes in complex environments involving CO2 in wellbores such as CO2 flooding wells, CO2 storage wells, and production wells that spit back CO2 in the petroleum industry, thereby improving the accuracy of the prediction results of the corrosion rate of 13Cr pipes when the corrosion factors fluctuate over a large range.

[0046] In step S1, the multi-factor data includes the condition data that affects the CO2 corrosion reaction, including temperature, CO2 partial pressure, gas phase moisture content, Cl - The data of n influencing factors such as concentration, gas flow rate and corrosion time were used, and a total of N experiments were carried out using the multi-factor orthogonal experimental method.

[0047] In step S1, Figure 2 As shown, the indoor test of 13Cr pipe gas phase corrosion includes the following steps:

[0048] The 13Cr pipe provided by the oil field site was milled to obtain a hanging piece, and the surface of the hanging piece was gradually polished with emery paper to a mirror state, and then washed with deionized water and anhydrous alcohol in turn, and then placed in a vacuum drying oven for drying; the hanging piece after pretreatment was weighed to obtain a hanging piece mass of G1;

[0049] The weighed 13Cr hanging piece was placed in a high temperature and high pressure reactor, and the Cl-containing -Add simulated water into the high-temperature and high-pressure reactor, turn on the heating button to heat the reactor to the preset corrosion temperature, introduce CO2 to remove the air in the reactor, then close the outlet valve, and continue to introduce CO2 until the reactor reaches the preset corrosion pressure. Set and open the fluid flow rate controller in the reactor according to the production well output rate, and carry out 13Cr pipe corrosion experiments for 3d, 5d, 7d, and 15d according to the preset corrosion cycle;

[0050] After reaching the preset corrosion cycle, first turn off the fluid flow rate controller in the reactor, then turn off the heating button, wait until the temperature drops to room temperature, perform pressure relief treatment, then open the reactor, remove the container containing the coupons in the reactor, take out the coupons with tweezers, and use absorbent paper to absorb the residual moisture on the surface of the coupons, and put them into the experimental storage bag one by one; refer to "Removal of Corrosion Products on Corrosion Specimens of Metals and Alloys" (GB / T16545-2015) to remove the corrosion products, rinse the surface of the coupons with distilled water to remove the corrosion residue Cr2O3 film, remove water with anhydrous alcohol and dry the coupons, and then weigh the coupons to obtain the mass of the coupons after corrosion as G2;

[0051] The gas phase corrosion rate of 13Cr pipe in different corrosion cycles is calculated by weight loss method. The calculation of gas phase corrosion rate is as follows:

[0052]

[0053] ΔG = G1 - G2;

[0054] Where V cor represents the average corrosion rate; ΔG represents the mass loss of the coupon; ρ represents the density of the coupon; S represents the contact area; and t represents the experimental time.

[0055] In step S3, the process of pre-establishing a decision tree model is as follows: according to the decision tree machine learning principle, the corrosion rate is used as the root node, the feature information gain of the corrosion influencing factor is calculated, and the feature with the largest information gain is selected as the feature of the corresponding corrosion factor, and sub-nodes are established by different values ​​of this feature, and the above operation is repeated to select different values ​​of the feature from the remaining corrosion influencing factors to establish the next sub-node until no feature is selected, and the construction of the decision tree is completed. In this embodiment, the number of corrosion factors that can be input is not limited, and can be unlimitedly increased under the premise of being realized under indoor experimental conditions, and the applicable corrosion environment is wider.

[0056] Using the decision tree machine learning method, n influencing factors such as temperature and CO2 partial pressure are taken as input variables X i(i=1, 2, 3, ..., n), and the corrosion rate is used as the output variable Y. In the N experiments conducted, n experimental conditions are randomly selected as prediction data, and the remaining (Nn) corrosion rate test results are used as training data to predict the corrosion rate.

[0057] In step S5, the parameters of the decision tree model include the number of repeated permutations, the sample size of the test set, and other parameters.

[0058] The operation process of this embodiment is simple and suitable for the complex working conditions in the oil industry involving CO2 oil recovery wells, CO2 storage wells, production wells that spit back CO2, etc., which contain CO2 in the wellbore. It is highly practical and accurate, and saves the workload and working time of on-site corrosion rate prediction.

[0059] Embodiment 2

[0060] This embodiment is similar to the first embodiment, except that in step S6, the standard deviation rate of the corrosion rate at each time point is calculated based on the corrosion rate prediction results obtained at different corrosion times.

[0061] If the standard deviation rate of the predicted corrosion rate at each time point does not exceed 10%, it is judged as uniform corrosion, and the corrosion rate does not change with time, which is the prediction result of the decision tree model.

[0062] If the standard deviation of the corrosion rate prediction value at each time point exceeds 10% and has a clear downward trend over time, it is determined to be variable-rate corrosion. The corrosion rate results under different corrosion times are calculated using the functional relationship between the decision tree model learning results and the corrosion time. The indoor experimental results with a corrosion time of 3d are used as training data to obtain the corrosion rate Y of the test data. The corrosion rate of 13Cr pipes changing with time is calculated as follows:

[0063] V cor =Yat b ;

[0064]

[0065]

[0066] Where V cor represents the average corrosion rate; a and b represent the functions of corrosion temperature T and CO2 molar concentration respectively; t represents the corrosion time; represents the molar concentration of CO2; T represents the corrosion temperature.

[0067] In this embodiment, the different corrosion modes of 13Cr pipes in a wet phase CO2 corrosion environment are determined by calculating the difference in corrosion rate of the pipes at different time points, without the need to perform corrosion product film analysis, thus saving the high cost of microscopic experiments.

[0068] Embodiment 3

[0069] A CO2 storage well was put into production in 2007, with a vertical depth of 1360.58m, a wellhead temperature of about 60℃, a formation temperature of 90℃, a bottom hole pressure of 6MPa, and the oil pipe used was 13Cr pipe. The corrosion rate of the CO2 storage well was predicted.

[0070] The orthogonal test method was used to determine the temperature, CO2 molar concentration, formation water Cl - The four influencing factors of concentration and flow rate were selected, and three values ​​were taken for each influencing factor to establish a four-factor three-level orthogonal test matrix, as shown in Table 1. A total of 9 CO2 dynamic vapor corrosion tests were carried out.

[0071] Table 1 Indoor test corrosion conditions based on orthogonal test method

[0072]

[0073]

[0074] Three coupons were used in each experiment. The materials used in the experiment were 13Cr pipes provided by the oil field site, which were milled and milled. The coupons were 50 mm long, 15 mm wide, and 5 mm thick. The coupons were polished gradually with 200-1000 mesh emery paper until the surface reached a mirror state. After being cleaned with deionized water and anhydrous alcohol, they were placed in a vacuum drying oven for standby use. Before the CO2 dynamic vapor phase corrosion was carried out, the initial mass of the coupons to be used was weighed with an electronic balance with a sensitivity of 0.1 mg, which was 19.5043 g to 20.4986 g.

[0075] The weighed 13Cr hanging piece was placed in a high temperature and high pressure reactor, and the water content was calculated according to the liquid-gas ratio of the gas produced on site. - Simulate the use of water, record the pressure value in the reactor at this time, then turn on the heating button, and after the temperature is heated to the preset corrosion temperature, open the gas booster system, connect the valves, gas tank valves, etc., pass CO2 and continue for 5 minutes to expel the air in the reactor, close the outlet valve, and continue to pass CO2 to the preset corrosion pressure; calculate the flow rate according to the on-site wellhead gas production and pipe diameter, set and turn on the fluid flow rate controller in the reactor, and conduct corrosion experiments for 3d, 5d, 7d, and 15d respectively according to the preset time.

[0076] After the predetermined corrosion time is reached, first turn off the fluid flow rate controller in the reactor, then turn off the heating button of the reactor. After the temperature drops to room temperature, perform pressure relief, open the reactor, remove the container containing the coupons in the reactor, remove the coupons with tweezers, and use absorbent paper to absorb the remaining moisture on the surface of the coupons, and put them into the experimental holding bag one by one.

[0077] The surface of the coupon was rinsed with distilled water to remove the corrosion residue Cr2O3 film, and the coupon was dried after dehydration with anhydrous alcohol. The mass of the coupon after corrosion was weighed using an electronic balance with a sensitivity of 0.1 mg, and the mass was 19.5023g~20.4924g. The weight loss method was used to calculate the vapor phase corrosion rate of 13Cr pipes in different corrosion cycles.

[0078] The decision tree machine learning method was adopted, with influencing factors such as temperature and CO2 content as input variables, and corrosion rate as output variable. Among the 9 experiments conducted, the corrosion conditions and corrosion rate measurement results of 8 experiments were selected as training data, and the remaining 1 condition was selected as test data for corrosion rate prediction.

[0079] Based on the corrosion rate data at different corrosion times obtained from the orthogonal experiment, it is judged that the standard deviation of the corrosion rate at each time point is 4.6%, and there is no downward trend over time. It is determined that the current corrosion mode is uniform corrosion.

[0080] The prediction method of this embodiment is used to predict the bottom corrosion rate of the well by combining the historical data of the well provided on site with the corrosion rate value obtained by the orthogonal experiment. Since the actual working conditions of the well are constantly changing, the corrosion rate results are predicted by substituting the corrosion conditions year by year with the time unit. Figure 3 As shown, the corrosion rate of 13Cr under the well conditions exhibits a uniform corrosion phenomenon, and the corrosion rate fluctuates with the corrosion environment of different service years. The maximum corrosion rate of the 13Cr oil pipe downhole is 0.03mm / a. The annual corrosion rate is accumulated, and the cumulative corrosion depth is about 0.3609mm. The actual maximum cumulative corrosion depth provided on site is 0.3090mm, and the prediction value error is about 16.7961%, which is highly accurate.

[0081] In the specific contents of the above-mentioned specific implementation methods, the various technical features can be combined in any non-contradictory manner. In order to make the description concise, not all possible combinations of the above-mentioned technical features are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0082] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment, characterized in that: The following steps are involved: S1: Establish an orthogonal experimental matrix of multi-factor data, conduct indoor test of 13Cr pipe gas phase corrosion under multi-factor data matching combination, and obtain the corrosion rate under multi-factor data matching combination; S2: using the multi-factor data as an independent variable and the corrosion rate as a dependent variable, establishing a sample data set, and dividing the sample data set into a training set and a test set according to a preset ratio; S3: training the pre-established decision tree model based on the training set to obtain a trained initial decision tree model; S4: inputting the multi-factor data in the test set into the initial decision tree model, and outputting a test corrosion rate; S5: Determine whether the error between the test corrosion rate and the corrosion rate in the test set is less than a preset error; if so, use the initial decision tree model as the target decision tree model and proceed to step S6; if not, modify the parameters of the pre-established decision tree model and return to step S3; S6: Input the target multi-factor data into the target decision tree model to obtain the corrosion rate prediction result.

2. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 1, characterized in that: In step S1, the multi-factor data includes the condition data that affects the CO2 corrosion reaction, including temperature, CO2 partial pressure, gas phase moisture content, Cl - concentration, gas flow rate and corrosion time.

3. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 1, characterized in that: In step S1, the 13Cr pipe gas phase corrosion indoor test includes the following steps: S101: Milling the 13Cr pipe provided by the oil field to obtain a hanging piece, pre-treating the hanging piece, and weighing the pre-treated hanging piece to obtain a hanging piece mass G1; S102: Put the weighed 13Cr hanging piece into the high temperature and high pressure reactor, and prepare the Cl-containing - Add simulated water into the high-temperature and high-pressure reactor, turn on the heating button to heat the reactor to the preset corrosion temperature, introduce CO2 into the reactor to reach the preset corrosion pressure, set and turn on the fluid flow rate controller in the reactor according to the production well output rate, and conduct 13Cr pipe corrosion experiments according to different preset corrosion cycles; S103: After reaching the preset corrosion cycle, the corroded hanging piece is taken out, the hanging piece is post-processed, and the hanging piece is weighed to obtain the mass of the corroded hanging piece as G2; S104: Calculate the gas phase corrosion rate of the 13Cr pipe in different corrosion cycles using the weight loss method according to the weighing data G1 of step S101 and the weighing data G2 of step S103.

4. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 3, characterized in that: In step S101, the process of pre-treating the hanging piece is: using emery paper to gradually polish the surface of the hanging piece to reach a mirror state, washing it with deionized water and anhydrous alcohol in turn, and then placing it in a vacuum drying oven to dry.

5. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 3, characterized in that: In step S102, during the process of introducing CO2, CO2 needs to be introduced to expel the air in the reactor, and then the outlet valve is closed, and CO2 is continuously introduced to a preset corrosion pressure.

6. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 3, characterized in that: In step S103, when taking out the corroded coupons, first turn off the fluid flow rate controller in the reactor, then turn off the heating button, wait until the temperature drops to room temperature, perform pressure relief, then open the reactor, remove the container containing the coupons in the reactor, take out the coupons with tweezers, and use absorbent paper to absorb the remaining moisture on the surface of the coupons, and put them into the experimental storage bag one by one.

7. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 6, characterized in that: In step S103, the process of post-processing the coupon is as follows: washing the surface of the coupon with distilled water to remove the corrosion residue Cr2O3 film, removing water with anhydrous alcohol and then drying the coupon.

8. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 3, characterized in that: In step S104, the gas phase corrosion rate is calculated as follows: ΔG = G1 - G2; Where V cor represents the average corrosion rate; ΔG represents the mass loss of the coupon; ρ represents the density of the coupon; S represents the contact area; and t represents the experimental time.

9. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to claim 1, characterized in that: In step S3, the process of pre-establishing a decision tree model is as follows: according to the decision tree machine learning principle, the corrosion rate is used as the root node, the feature information gain of the corrosion influencing factor is calculated, and the feature with the largest information gain is selected as the feature of the corresponding corrosion factor. Sub-nodes are established based on different values ​​of this feature, and the above operation is repeated to select different values ​​of the feature from the remaining corrosion influencing factors to establish the next sub-node until no feature is selected, and the construction of the decision tree is completed.

10. The method for predicting the corrosion rate of 13Cr pipes in a carbon dioxide environment according to any one of claims 2 to 9, characterized in that: In step S6, the standard deviation rate of the corrosion rate at each time point is calculated based on the obtained corrosion rate prediction results. If the standard deviation rate of the corrosion rate prediction value at each time point does not exceed 10%, it is judged as uniform corrosion, and the corrosion rate does not change with time, which is the prediction result of the decision tree model; if the standard deviation rate of the corrosion rate prediction value at each time point exceeds 10% and has a clear downward trend with time, it is judged as variable-rate corrosion, and the corrosion rate results at different corrosion times are calculated using the functional relationship between the decision tree model learning result and the corrosion time.

Citation Information

Patent Citations

  • Pipeline corrosion rate prediction method and device and computer equipment

    CN110751339A

  • Water pipe uniform corrosion rate prediction model considering interaction

    CN117074283A

  • System and method for corrosion prediction in oil and gas pipeline

    IN202021003471A

Cited By

  • Method and device for acquiring corrosion state of oil well pipeline

    CN119918432A

  • A method and device for obtaining the corrosion state of an oil well pipeline

    CN119918432B

  • Method and device for testing corrosion resistance of product in ammonia salt gas environment

    CN120708768A