Drill string fatigue life reliability analysis method and device

Abstract

The application discloses a kind of drill string fatigue life reliability analysis method and device.The method includes: the parameter of random influencing factor that the hot spot position of drill string influences drill string fatigue life is preset number of times of Monte Carlo sampling, obtain the multiple samples of the first parameter combination of random influencing factor;The multiple samples are input to drill string fatigue calculation neural network, and the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor is output by the drill string fatigue calculation neural network;According to the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor, generate dataset;The data set is analyzed, and the fatigue life reliability of drill string at the hot spot position is determined.The application can effectively improve the precision of fatigue life reliability prediction and reduce the analysis timeliness of drill string fatigue life, help to reduce the influence of uncertain factors on drill string fatigue life prediction.

Description

Technical Field

[0001] This invention relates to the field of oil drilling tool analysis technology, and in particular to a method and apparatus for analyzing the fatigue life reliability of drill strings. Background Technology

[0002] During drilling, the drill string is subjected to complex forces, and failure in this process poses a challenge to operational safety and reliability. Because the drill string is subjected to cyclic loads over a long period, fatigue damage can easily occur at hot spots, and this cumulative effect will reduce the drill string's safety over time.

[0003] Previous drill string fatigue reliability analysis has focused primarily on mechanical analysis, without comprehensively considering the various factors affecting drill string fatigue life. Furthermore, traditional drill string reliability analysis typically employs the Monte Carlo simulation method, which usually requires multiple random sampling and simulations, resulting in a limited number and quality of samples. This makes drill string reliability assessment using traditional methods time-consuming and inaccurate. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a method and apparatus for reliability analysis of drill string fatigue life that overcomes or at least partially solves the above problems.

[0005] As a first aspect of the present invention, embodiments of the present invention provide a method for reliability analysis of drill string fatigue life, comprising:

[0006] A preset number of Monte Carlo sampling operations are performed on the parameters of random influencing factors affecting the fatigue life of the drill string at hot spot locations, resulting in multiple samples of the first parameter combination of the random influencing factors; the parameters of the random influencing factors include: parameters of the material properties, geometric properties, and load properties of the drill string;

[0007] The multiple samples are input into the drill string fatigue calculation neural network, and the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor is output by the drill string fatigue calculation neural network; the drill string fatigue calculation neural network is pre-trained using a dataset formed by the second parameter combination of random influencing factors and the drill string fatigue life data obtained by finite element analysis.

[0008] A dataset is generated based on the first parameter combination of each random influencing factor and the corresponding drill string fatigue life data.

[0009] The dataset was analyzed to determine the fatigue life reliability of the drill string at the hot spot location.

[0010] In one embodiment, the drill string fatigue calculation neural network is pre-trained in the following manner:

[0011] Determine multiple combinations of parameters for random factors affecting the fatigue life of drill strings;

[0012] Using the Latin hypercube sampling method, a preset number of random samplings are performed on multiple combinations of parameters of the random influencing factors to obtain a sample set containing multiple combinations of second parameters.

[0013] The fatigue life data of the drill string corresponding to each combination of the second parameter in the sample set were obtained by using the finite element analysis method.

[0014] The drill string fatigue life data corresponding to each combination of second parameters obtained by the finite element analysis method are used as a dataset to train the neural network model, thus obtaining the drill string fatigue calculation neural network.

[0015] In one embodiment, the finite element analysis method is used to obtain drill string fatigue life data corresponding to each combination of second parameters in the sample set, including:

[0016] Based on various working conditions of the drill string, a finite element model of drill string mechanics is constructed using the finite element method. According to each combination of second parameters in the sample set, the stress on the drill string under each working condition is calculated for each combination of second parameters in the finite element model.

[0017] Based on the stress on the drill string under each working condition under each combination of the second parameters, the total number of cycles required for the drill string to reach the critical state of fatigue failure at the corresponding stress level is calculated, and this number is used as the fatigue life data of the drill string at the hot spot location under the second parameter combination.

[0018] In one embodiment, based on the stress experienced by the drill string under each combination of second parameters, the total number of cycles required for the drill string to reach the critical state of fatigue failure at the corresponding stress level is calculated and used as the fatigue life data of the drill string under the second combination of parameters, including:

[0019] Calculate the ratio of the number of cycles to the ultimate fatigue life under the corresponding stress level under cyclic load in each working condition; sum the ratios under cyclic load in each working condition, and when the sum reaches the preset total cumulative damage value, determine the total number of cycles under each working condition as the drill string fatigue life data.

[0020] In one embodiment, the drill string fatigue life data corresponding to each combination of second parameters obtained by the finite element analysis method are used as a dataset to train a neural network model, resulting in a drill string fatigue calculation neural network, including:

[0021] Normalize the fatigue life data of each combination of second parameters and the corresponding drill string.

[0022] The normalized data is divided into training set, test set and validation set according to a preset ratio;

[0023] The neural network model is trained by inputting the training set into a preset neural network model, and the optimal parameters of the neural network model are found by combining the validation set. The performance of the trained neural network model is evaluated by using the test set. The neural network model that has been trained and whose performance meets the expected target is the drill string fatigue calculation neural network.

[0024] The input to the drill string fatigue calculation neural network is a normalized combination of parameters, and the output is the corresponding drill string fatigue life data.

[0025] In one embodiment, analyzing the dataset to determine the reliability of the drill string at the hotspot location includes:

[0026] Statistical analysis was performed on the samples obtained from Monte Carlo sampling and the corresponding drill string fatigue life data to determine the distribution pattern of fatigue failure probability.

[0027] The fatigue life reliability of the drill string is calculated based on the distribution of the fatigue failure probability.

[0028] In one embodiment, the drill string fatigue life reliability is calculated based on the distribution of fatigue failure probability, including:

[0029] Based on the distribution form of the fatigue failure probability, determine the expression for the probability density function under this distribution form;

[0030] Statistical analysis was performed on the samples obtained from Monte Carlo sampling and the corresponding drill string fatigue life data. The parameter µ, which reflects the average level of the drill string fatigue life data, and the standard deviation σ, which reflects the degree of deviation between the drill string fatigue life data and the parameter µ, were calculated.

[0031] Substituting the calculated parameters µ and standard deviation σ into the probability density function expression and failure probability formula, the failure probability value of the drill string under the preset N1 cycles is obtained.

[0032] The difference between 1 and the failure probability value is used to obtain the fatigue life reliability of the drill string.

[0033] In one embodiment, the parameters of the random influencing factor specifically include:

[0034] Material properties: density, elastic modulus, and Poisson's ratio; geometric properties: outer diameter and wall thickness; load properties: torque load and axial load.

[0035] As a second aspect of the present invention, embodiments of the present invention provide a drill string fatigue life reliability analysis device, comprising:

[0036] The sampling module is used to perform a preset number of Monte Carlo samplings on the parameters of random influencing factors that affect the fatigue life of the drill string at hot spot locations, to obtain multiple samples of the first parameter combination of the random influencing factors; the parameters of the random influencing factors include the parameters of the drill string's material properties, geometric properties, and load properties.

[0037] The drill string fatigue life determination module is used to input the multiple samples into the drill string fatigue calculation neural network, and output the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor through the drill string fatigue calculation neural network; the drill string fatigue calculation neural network is pre-trained through a dataset formed by the second parameter combination of random influencing factors and the drill string fatigue life data obtained by finite element analysis.

[0038] The dataset generation module is used to generate a dataset based on the first parameter combination of each random influencing factor and the corresponding drill string fatigue life data.

[0039] The reliability determination module is used to analyze the dataset and determine the fatigue life reliability of the drill string at the hot spot location.

[0040] As another aspect of the present invention, an embodiment of the present invention provides a server, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the drill string fatigue life reliability analysis method as described above.

[0041] As a third aspect of the present invention, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the drill string fatigue life reliability analysis method as described above.

[0042] As a fourth aspect of the present invention, an embodiment of the present invention provides a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the drill string fatigue life reliability analysis method as described above.

[0043] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:

[0044] The drill string fatigue life reliability analysis method and apparatus provided in this invention fully consider the influence of various factors affecting drill string fatigue life, such as drill string material properties, geometric properties, and load properties, on the drill string fatigue life. Furthermore, by employing a neural network model with nonlinear mapping characteristics combined with the Monte Carlo simulation method, it can efficiently obtain a large-scale combination of influencing factor parameters that conforms to the distribution, as well as high-quality drill string fatigue life data. This provides a sufficient sample basis for subsequent reliability calculations, effectively improving the accuracy of fatigue life reliability prediction and reducing the analysis time of drill string fatigue life. It also makes up for the shortcomings of traditional Monte Carlo methods, which suffer from low accuracy due to insufficient sample size, and significantly shortens the sampling calculation time.

[0045] The drill string fatigue life data obtained by the method provided in the embodiments of the invention can also intuitively restore the drill string fatigue life law under the distribution characteristics of multiple uncertain factors, which helps to reduce the impact of uncertain factors on the prediction of drill string fatigue life.

[0046] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0047] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0048] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0049] Figure 1 This is a flowchart of a drill string fatigue life reliability analysis method provided in an embodiment of the present invention;

[0050] Figure 2 This is a flowchart of a drill string fatigue life reliability analysis based on a neural network, provided in an embodiment of the present invention.

[0051] Figure 3A This is a frequency distribution diagram of the material density of the drill string provided in an embodiment of the present invention;

[0052] Figure 3B This is a probability distribution diagram of the Young's modulus of the drill string provided in an embodiment of the present invention;

[0053] Figure 4 This is a regression diagram of fatigue life data at a certain hot spot location of the drill string provided in an embodiment of the present invention;

[0054] Figure 5A and Figure 5B The curves showing the relationship between the fatigue life and reliability of the drill string at hotspot locations 1 and 2, respectively, are provided in the embodiments of the present invention.

[0055] Figure 6 This is a structural block diagram of the drill string fatigue life reliability analysis device provided in an embodiment of the present invention. Detailed Implementation

[0056] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0057] To address the problems in the prior art, embodiments of the present invention provide a method for analyzing the fatigue life reliability of drill strings, referring to... Figure 1 As shown, the method includes the following steps:

[0058] S11. Perform Monte Carlo sampling on the parameters of random influencing factors that affect the fatigue life of the drill string at the hot spot location to obtain multiple samples of the first parameter combination of the random influencing factors.

[0059] The parameters of the aforementioned random influencing factors include the parameters of the drill string's material properties, geometric properties, and load properties;

[0060] S12. Input the multiple samples into the drill string fatigue calculation neural network, and output the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor through the drill string fatigue calculation neural network; the drill string fatigue calculation neural network is pre-trained through a dataset formed by the second parameter combination of random influencing factors and the drill string fatigue life data obtained by finite element analysis.

[0061] S13. Generate a dataset based on the first parameter combination of each random influencing factor and the corresponding drill string fatigue life data;

[0062] S14. Analyze the dataset to determine the fatigue life reliability of the drill string at the hot spot location.

[0063] In this embodiment of the invention, for ease of distinction, the parameter combination of random influencing factors obtained by Monte Carlo sampling is referred to as the first parameter combination, and the parameter combination of random influencing factors used in the training process of the drill string fatigue calculation neural network is referred to as the second parameter combination.

[0064] In one embodiment, the parameters of the random influencing factors in step S11 above specifically include: material properties such as density, elastic modulus, and Poisson's ratio; geometric properties such as outer diameter and wall thickness; and load properties such as torque load and axial load.

[0065] The combination of parameters of random influencing factors can be expressed as: , where z1~z n These represent variables such as Young's modulus, Poisson's ratio, density, outer diameter, and wall thickness.

[0066] The above steps are for the reliability calculation of fatigue life at a certain hot spot location. If the drill string involves multiple hot spots, the above steps S11-S14 can be performed separately for each hot spot location.

[0067] In one embodiment, the above-mentioned drill string fatigue calculation neural network can be pre-trained, for example, in the following manner:

[0068] 1.1. Determine multiple combinations of parameters of random influencing factors affecting the fatigue life of drill strings;

[0069] 1.2. Using the Latin Hypercube Sampling (LHS) method, random sampling is performed a preset number of times on multiple combinations of parameters of the random influencing factors to obtain a sample set containing multiple combinations of second parameters;

[0070] For example, the LHS sampling method can be used to conduct more than 100 random samplings of influencing factor variables that conform to the corresponding distribution characteristics.

[0071] LHS sampling is a stratified random sampling method, which is an optimized and improved form of Monte Carlo sampling. By stratifying the probability distribution interval of each random variable, it ensures that the sampling results uniformly cover the entire distribution range. Compared with traditional random sampling (such as simple random sampling), it can balance randomness and uniform distribution coverage with a smaller sample size, and is especially suitable for sampling scenarios with multiple variables and complex distributions.

[0072] Specifically, for each influencing random variable (such as Young's modulus, wall thickness, etc.), according to its preset probability distribution (such as normal distribution, uniform distribution), the probability distribution interval of the variable is divided into non-overlapping sub-intervals, and the probability of each sub-interval is equal (i.e., the probability corresponding to each sub-interval).

[0073] Within each predefined sub-interval, one sample point (i.e., one specific value of the random variable) is randomly selected to ensure that only one point is taken in each sub-interval. This step guarantees that each stratum has sample coverage.

[0074] For all influencing random variables (e.g., 3 variables of material properties, 2 variables of geometric properties, 2 variables of load properties, a total of 7 variables), repeat steps 1-2 to obtain 100 sample points for each variable; then randomly shuffle and combine all the sample points of all variables to form 100 groups of multivariate combined samples (each group of samples contains 1 specific value of all influencing factors).

[0075] 1.3. The finite element method was used to obtain the drill string fatigue life data corresponding to each combination of the second parameter in the sample set;

[0076] 1.4. The drill string fatigue life data corresponding to each combination of second parameters obtained by the finite element analysis method are used as a dataset to train the neural network model, thereby obtaining the drill string fatigue calculation neural network.

[0077] Furthermore, the drill string fatigue life data corresponding to each combination of second parameters obtained by the above-mentioned finite element analysis method can be achieved in the following way:

[0078] 2.1. Based on various working conditions of the drill string, a finite element model of drill string mechanics is constructed using the finite element method. According to each combination of second parameters in the sample set, the stress on the drill string under each working condition is calculated for each combination of second parameters in the finite element model.

[0079] The above-mentioned operating conditions include rotary drilling, sliding drilling, bottom hole idling, reaming, and tripping. Considering the effects of bending stress amplification, buckling load, and vibration, a finite element method is used to construct a mechanical model of the drill string, and the stress on the drill string is calculated based on sampling parameters.

[0080] 2.2. Based on the stress on the drill string under each working condition under each combination of the second parameters, calculate the total number of cycles at which the drill string reaches the critical state of fatigue failure under the corresponding stress level, and use this as the fatigue life data of the drill string under the second combination of parameters.

[0081] In one embodiment, for example, the linear cumulative damage theory (Miner's cumulative damage theory) can be used to calculate the ratio of the number of cycles to the ultimate fatigue life under the cyclic load of each working condition at the corresponding stress level; then, the ratios under the cyclic load of each working condition are summed, and when the sum reaches the preset total cumulative damage value, the total number of cycles under each working condition at this time is determined as the drill string fatigue life data.

[0082] Specifically, the actual stress level on the drill string (corresponding to the stress under different operating conditions) is calculated using a finite element model combined with drilling conditions (rotary drilling, sliding drilling, etc.). At this point, the SN curve (core curve of material fatigue performance, with the horizontal axis representing the number of cycles and the vertical axis representing the stress level) of the drill string needs to be called. For each actual stress level, the corresponding ultimate fatigue life number is found on the SN curve.

[0083] For example, if the drill string stress is under a certain working condition, the corresponding cycle (i.e., failure after 1 million cycles) can be found through the SN curve.

[0084] By combining the actual parameters of drilling operations, the actual number of cycles of the drill string under the actual stress level corresponding to each working condition was statistically analyzed, and multiple sets of corresponding data of "stress level - actual number of cycles - ultimate fatigue life number of cycles" were obtained.

[0085] According to Miner's cumulative damage theory, the fatigue life of the drill string is calculated through the following steps: (1) Calculate the single-cycle damage ratio: For each "AND" group, calculate the damage ratio under that stress level (i.e., the damage contribution of a single cycle to the drill string). (2) Accumulate the total damage: Sum the damage ratios under all stress levels to obtain the total cumulative damage. (3) Determine the fatigue life: When the total cumulative damage is reached, the drill string reaches the fatigue failure state. At this time, the sum of the actual number of cycles under all stress levels (or the corresponding total working time) is the fatigue life of the drill string.

[0086] In one embodiment, the training process of the drill string fatigue calculation neural network in step 1.4 above is as follows:

[0087] Normalize the fatigue life data of each combination of second parameters and the corresponding drill string.

[0088] The normalized data is divided into training set, test set and validation set according to a preset ratio;

[0089] The neural network model is trained by inputting the training set into a preset neural network model, and the optimal parameters of the neural network model are found by combining the validation set. The performance of the trained neural network model is evaluated by using the test set. The neural network model that has been trained and whose performance meets the expected target is the drill string fatigue calculation neural network.

[0090] The input to the drill string fatigue calculation neural network is a normalized combination of parameters (the combination of parameters to be predicted), and the output is the corresponding drill string fatigue life data.

[0091] In one embodiment, in step S11 above, Monte Carlo sampling is performed 10,000 times or more on the trained neural network model. During sampling, based on the preset distribution characteristics of the random variables of influencing factors (such as Young's modulus following a normal distribution (mean 206 GPa, standard deviation 10.3 GPa), outer diameter following a normal distribution (mean 0.0127 m, standard deviation 0.00127 m), etc.), a large number of first parameter combinations of influencing factors that conform to the distribution are randomly generated, and these first parameter combinations are input into the neural network model. The model outputs the corresponding drill string fatigue life data. Through this sampling, a large-scale (10,000 or more sets) of data corresponding to the influencing factor parameter combinations and drill string fatigue life data is obtained, providing a sufficient sample basis for subsequent reliability calculations.

[0092] Because the accuracy of the Monte Carlo method depends on the sample size, the more Monte Carlo sampling is performed (10,000 times or more), the closer the random sampling results are to the true probability distribution, and the smaller the error in the final calculated failure probability and reliability. Combined with the advantage of neural network models that can output a sufficient number of samples, it can effectively make up for the shortcomings of traditional Monte Carlo methods that have low accuracy due to insufficient sample size, and also significantly shorten the sampling calculation time.

[0093] In one embodiment, step S14 above, analyzing the dataset to determine the drill string fatigue life reliability, can be achieved in the following way:

[0094] Statistical analysis was performed on the samples obtained from Monte Carlo sampling and the corresponding drill string fatigue life data to determine the distribution pattern of fatigue failure probability.

[0095] The fatigue life reliability of the drill string is calculated based on the distribution of the fatigue failure probability.

[0096] In one embodiment, the drill string fatigue life reliability is calculated based on the distribution of fatigue failure probability, which can be achieved in the following way:

[0097] Based on the distribution form of the fatigue failure probability, determine the expression for the probability density function under this distribution form;

[0098] Statistical analysis was performed on the samples obtained from Monte Carlo sampling and the corresponding drill string fatigue life data. The parameter µ, which reflects the average level of the drill string fatigue life data, and the standard deviation σ, which reflects the degree of deviation between the drill string fatigue life data and the parameter µ, were calculated.

[0099] Substituting the calculated parameters µ and standard deviation σ into the probability density function expression and failure probability formula, the failure probability value of the drill string under N1 cycles is obtained.

[0100] The difference between 1 and the failure probability value is used to obtain the drill string fatigue life reliability.

[0101] The essence of reliability is the probability that the fatigue life of the drill string is less than or equal to the critical fatigue limit, and the calculation method of the probability is determined by the probability distribution form that the fatigue life follows.

[0102] For example, the probability of failure may be distributed in various ways, such as log-normal distribution, normal distribution, etc.

[0103] For example, when the drill string fatigue life N follows a log-normal distribution, its probability density function is: The probability of the drill string failing after N1 cycles is: .

[0104] The formula for calculating reliability is: .

[0105] In the case of a log-normal distribution, μ is the mean of lnN, and σ is the standard deviation of lnN.

[0106] When the drill string fatigue life N follows a normal distribution, the drill string fatigue life follows a normal distribution (μ N Let σ be the arithmetic mean of N and σ be the standard deviation of N. The probability density function of the normal distribution can be found using existing techniques. Reliability can be directly determined using the cumulative distribution function of the normal distribution. .

[0107] in: The cumulative distribution function of the standard normal distribution; It is the critical fatigue limit. It is the drill string fatigue life, μ N for The arithmetic mean, σ is The standard deviation. The value can be obtained using existing techniques.

[0108] The calculation methods for other distributions are similar and will not be repeated here.

[0109] A flowchart for a drill string fatigue life reliability analysis based on a neural network can be found here. Figure 2 As shown, by performing LHS sampling on random variables of material properties, geometric properties, and loads, the resulting combination of random variables (the second parameter combination) is used to perform finite element analysis on drill string fatigue life, obtaining drill string fatigue life data. Then, the combination of LHS random variables is used as the input to a neural network, and the drill string fatigue life is used as the output of the neural network to train the model. Monte Carlo sampling is then performed on the combination of the output variable and the drill string fatigue life of the trained neural network to determine the drill string fatigue failure probability distribution. Based on the distribution, the reliability result of drill string fatigue life is calculated.

[0110] Let me give an example.

[0111] This example includes the following steps:

[0112] (1) Choosing random variables

[0113] When performing fatigue life neural network simulations at various hot spots in the drill string, random variables are used as neurons in the input layer of the neural network. These random variables are selected from material properties such as density, elastic modulus, and Poisson's ratio; geometric properties such as outer diameter and wall thickness; and load properties such as torque and axial force. The distribution type and standard deviation are determined.

[0114] Table 1 Distribution Attributes of Some Influencing Factors

[0115]

[0116] (2) Collect and process data

[0117] Using LHS sampling, 100 random samples conforming to the corresponding distribution characteristics were conducted for 10 influencing factors. The fatigue life analysis of each hot spot location in the drill string was completed using a finite element model, and fatigue life data under the corresponding influencing factors were obtained.

[0118] The above 10 random parameters were sampled using a distribution. Some sampling results are shown below. Figure 3A and Figure 3B As shown.

[0119] Figure 3A This is a frequency distribution diagram of the material density of the drill string. Figure 3B This is a probability distribution diagram of the Young's modulus of the drill string.

[0120] (3) Establish a neural network model to perform fatigue life data regression.

[0121] Using the sample data of influencing factors as input parameters and the drill string fatigue life data obtained by the finite element method as output parameters, the generated data is first preprocessed. The 100 sets of influencing factor and fatigue life data are divided into training, validation, and test sets in a 7:1.5:1.5 ratio. The neural network is then solved to obtain relevant results reflecting the training effect. The regression of the neural network prediction data is as follows... Figure 4 As shown.

[0122] Figure 4 This is a regression diagram of fatigue life data at a certain hot spot location in the drill string.

[0123] (4) Obtain the reliability analysis results of drill string fatigue life based on neural network.

[0124] According to the reliability solution method, the fatigue reliability of the drill string is calculated by combining the Monte Carlo method and the BP neural network prediction of 10,000 fatigue life data, and the fatigue life reliability of the drill string based on the neural network algorithm is obtained.

[0125] Figure 5A The curve shown represents the relationship between drill string fatigue life and reliability at hotspot location 1. Figure 5B The curve shown is the relationship between drill string fatigue life and reliability at hotspot location 2.

[0126] Based on the same inventive concept, this invention also provides a drill string fatigue life reliability analysis device. Since the principle of the problem solved by this device is similar to the aforementioned drill string fatigue life reliability analysis method, the implementation of this device can refer to the implementation of the aforementioned method, and the repeated parts will not be described again.

[0127] This invention provides a drill string fatigue life reliability analysis device, referring to... Figure 6 As shown, it includes:

[0128] The sampling module 61 is used to perform a preset number of Monte Carlo samplings on the parameters of random influencing factors that affect the fatigue life of the drill string at hot spot locations, to obtain multiple samples of the first parameter combination of the random influencing factors; the parameters of the random influencing factors include the parameters of the material properties, geometric properties and load properties of the drill string.

[0129] The drill string fatigue life determination module 62 is used to input the multiple samples into the drill string fatigue calculation neural network, and output the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor through the drill string fatigue calculation neural network; the drill string fatigue calculation neural network is pre-trained through a dataset formed by the second parameter combination of random influencing factors and the drill string fatigue life data obtained by finite element analysis.

[0130] The dataset generation module 63 is used to generate a dataset based on the first parameter combination of each random influencing factor and the corresponding drill string fatigue life data;

[0131] The reliability determination module 64 is used to analyze the dataset and determine the fatigue life reliability of the drill string at the hot spot location.

[0132] An embodiment of the present invention provides a server comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the drill string fatigue life reliability analysis method as described above.

[0133] This invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned drill string fatigue life reliability analysis method.

[0134] The present invention provides a computer program product, which includes a computer program that, when executed by a processor, implements the aforementioned drill string fatigue life reliability analysis method.

[0135] The drill string fatigue life reliability analysis method and apparatus provided in this invention fully consider the influence of various factors affecting drill string fatigue life, such as drill string material properties, geometric properties, and load properties, on the drill string fatigue life. Furthermore, by employing a neural network model with nonlinear mapping characteristics combined with the Monte Carlo simulation method, it can efficiently obtain a large-scale combination of influencing factor parameters that conforms to the distribution, as well as high-quality drill string fatigue life data. This provides a sufficient sample basis for subsequent reliability calculations, effectively improving the accuracy of fatigue life reliability prediction and reducing the analysis time of drill string fatigue life. It also makes up for the shortcomings of traditional Monte Carlo methods, which suffer from low accuracy due to insufficient sample size, and significantly shortens the sampling calculation time.

[0136] The drill string fatigue life data obtained by using the above-mentioned drill string fatigue life reliability analysis method and device provided in the embodiments of the present invention can also intuitively restore the drill string fatigue life law under the distribution characteristics of multiple sources of uncertainty factors, which helps to reduce the impact of uncertainty factors on the prediction of drill string fatigue life.

[0137] The principles by which the above-mentioned devices, servers, media, etc. in the embodiments of the present invention solve the problem are similar to those of the aforementioned methods. Therefore, their implementation can refer to the implementation of the aforementioned methods, and repeated details will not be repeated.

[0138] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0139] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0140] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0141] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0142] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for reliability analysis of drill string fatigue life, characterized in that, include: Monte Carlo sampling was performed a preset number of times on the parameters of random influencing factors that affect the fatigue life of the drill string at the hot spot location, resulting in multiple samples of the first parameter combination of the random influencing factors. The parameters of the random influencing factors include: parameters of the material properties, geometric properties, and load properties of the drill string; the Monte Carlo sampling includes: randomly generating a large number of first parameter combinations of influencing factors that conform to the distribution based on the preset distribution characteristics of the random variables of the influencing factors; The multiple samples are input into the drill string fatigue calculation neural network, and the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor is output by the drill string fatigue calculation neural network; the drill string fatigue calculation neural network is pre-trained using a dataset formed by the second parameter combination of random influencing factors and the drill string fatigue life data obtained by finite element analysis. A dataset is generated based on the first parameter combination of each random influencing factor and the corresponding drill string fatigue life data. Statistical analysis was performed on the samples obtained by Monte Carlo sampling and the corresponding drill string fatigue life data to determine the distribution form of the fatigue failure probability; the distribution form is a normal distribution or a log-normal distribution. Based on the distribution form of the fatigue failure probability, determine the expression for the probability density function under this distribution form; Statistical analysis was performed on the samples obtained from Monte Carlo sampling and the corresponding drill string fatigue life data. The parameter µ, which reflects the average level of the drill string fatigue life data, and the standard deviation σ, which reflects the degree of deviation between the drill string fatigue life data and the parameter µ, were calculated. Substituting the calculated parameters µ and standard deviation σ into the probability density function expression and failure probability formula, the failure probability value of the drill string under the preset N1 cycles is obtained. The difference between 1 and the failure probability value is used to obtain the drill string fatigue life reliability.

2. The method as described in claim 1, characterized in that, The drill string fatigue calculation neural network is pre-trained in the following manner: Determine multiple combinations of parameters for random factors affecting the fatigue life of drill strings; Using the Latin hypercube sampling method, a preset number of random samplings are performed on multiple combinations of parameters of the random influencing factors to obtain a sample set containing multiple combinations of second parameters. The fatigue life data of the drill string corresponding to each combination of the second parameter in the sample set were obtained by using the finite element analysis method. The drill string fatigue life data corresponding to each combination of second parameters obtained by the finite element analysis method are used as a dataset to train the neural network model, thus obtaining the drill string fatigue calculation neural network.

3. The method as described in claim 2, characterized in that, The finite element method was used to obtain the drill string fatigue life data corresponding to each combination of the second parameter in the sample set, including: Based on various working conditions of the drill string, a finite element model of drill string mechanics is constructed using the finite element method. According to each combination of second parameters in the sample set, the stress on the drill string under each working condition is calculated for each combination of second parameters in the finite element model. Based on the stress on the drill string under each working condition under each combination of the second parameters, the total number of cycles required for the drill string to reach the critical state of fatigue failure at the corresponding stress level is calculated, and this number is used as the fatigue life data of the drill string at the hot spot location under the second parameter combination.

4. The method as described in claim 3, characterized in that, Based on the stress experienced by the drill string under each combination of second parameters, the total number of cycles required for the drill string to reach the critical state of fatigue failure at the corresponding stress level is calculated, and this number serves as the fatigue life data of the drill string under the second combination of parameters, including: Calculate the ratio of the number of cycles to the ultimate fatigue life under the corresponding stress level under cyclic load in each working condition; sum the ratios under cyclic load in each working condition, and when the sum reaches the preset total cumulative damage value, determine the total number of cycles under each working condition as the drill string fatigue life data.

5. The method as described in claim 2, characterized in that, The drill string fatigue life data corresponding to each combination of second parameters obtained by finite element analysis are used as a dataset to train the neural network model, resulting in a drill string fatigue calculation neural network, including: Normalize the fatigue life data of each combination of second parameters and the corresponding drill string. The normalized data is divided into training set, test set and validation set according to a preset ratio; The neural network model is trained by inputting the training set into a preset neural network model, and the optimal parameters of the neural network model are found by combining the validation set. The performance of the trained neural network model is evaluated by using the test set. The neural network model that has been trained and whose performance meets the expected target is the drill string fatigue calculation neural network. The input to the drill string fatigue calculation neural network is a normalized combination of parameters, and the output is the corresponding drill string fatigue life data.

6. The method according to any one of claims 1-5, characterized in that, The parameters of the random influencing factors specifically include: Material properties: density, elastic modulus, and Poisson's ratio; geometric properties: outer diameter and wall thickness; load properties: torque load and axial load.

7. A drill string fatigue life reliability analysis device, characterized in that, include: The sampling module is used to perform a preset number of Monte Carlo samplings on the parameters of random influencing factors that affect the fatigue life of the drill string at hot spot locations, to obtain multiple samples of the first parameter combination of the random influencing factors; the parameters of the random influencing factors include the parameters of the drill string's material properties, geometric properties, and load properties. The drill string fatigue life determination module is used to input the multiple samples into the drill string fatigue calculation neural network, and output the drill string fatigue life data corresponding to the first parameter combination of each random influencing factor through the drill string fatigue calculation neural network. The drill string fatigue calculation neural network is pre-trained using a dataset formed by combining the second parameter of random influencing factors and the drill string fatigue life data obtained from finite element analysis. The dataset generation module is used to generate a dataset based on the first parameter combination of each random influencing factor and the corresponding drill string fatigue life data. The reliability determination module is used to perform statistical analysis on the samples obtained from Monte Carlo sampling and the corresponding drill string fatigue life data to determine the distribution form of the fatigue failure probability; the distribution form is a normal distribution or a log-normal distribution; based on the distribution form of the fatigue failure probability, the probability density function expression under this distribution form is determined; statistical analysis is performed on the samples obtained from Monte Carlo sampling and the corresponding drill string fatigue life data to calculate the parameter µ reflecting the average level of the drill string fatigue life data, and the standard deviation σ reflecting the degree of deviation of the drill string fatigue life data from the parameter µ; the calculated parameter µ and standard deviation σ are substituted into the probability density function expression and the failure probability formula to obtain the failure probability value of the drill string under a preset N1 cycles; the difference between 1 and the failure probability value is calculated to obtain the drill string fatigue life reliability.

8. A server, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the drill string fatigue life reliability analysis method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the drill string fatigue life reliability analysis method as described in any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the drill string fatigue life reliability analysis method as described in any one of claims 1-6.

Images