An optimized method for assessing the connection quality of two tubular elements
A two-model system using first-order numeric variables and machine learning evaluates tubular element connections, addressing reliability issues in existing methods, ensuring accurate and safe connections in oil and gas production.
Patent Information
- Application Number
- JP2024519032
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-09-28
- Filing Date
- 2022-09-15
- Publication Date
- 2024-11-07
- Estimated Expiration
- 2042-09-15
AI Technical Summary
Existing methods for evaluating the quality of connections between tubular elements in oil and gas production are unreliable, prone to human error, and can lead to improper connections with safety and environmental consequences.
A method involving a two-model system: a first model based on first-order numeric variables and a second model driven by machine learning, to assess connection quality by analyzing tightening graphs, ensuring accurate and reliable evaluation of tubular element connections.
The method provides a robust and automated evaluation of connection quality, reducing human error and improving the integrity of tubular element assemblies by accurately identifying compliant or non-compliant states, thereby enhancing safety and reliability.
Smart Images

Figure 2024541130000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates generally to tubular threaded elements, and more particularly to a method for connecting a threaded portion of a first tubular element to a threaded portion of a second tubular element.
[0002] More particularly, the present invention relates to a method for assessing the suitability of a connection between a threaded portion of a first tubular element and a threaded portion of a second tubular element. [Background technology]
[0003] In the field of oil and gas production, whether at an offshore or onshore facility performing well drilling and production operations, the operations carried out include connecting tubular elements to each other and lowering them into the well to form a drilling string, or an oil or gas production string.
[0004] A male or female threaded portion disposed at one end of a first tubular element may be directly connected to a complementary threaded portion of a second tubular element.
[0005] In another scenario, the first and second tubular elements may be indirectly connected by an intermediate tubular element, such as a joint.
[0006] The tubular elements are assembled within defined constraints to meet the tightening and sealing requirements imposed by the working conditions and to ensure the integrity of the assembly in service throughout its service life.
[0007] However, connections can sometimes be made incorrectly, resulting in poor sealing in the lines, damage to the tubular elements, and even premature separation.
[0008] Therefore, the quality of the fastening directly affects the sealing property and life of the assembly of the tubular elements, and therefore the quality of the connection must be evaluated according to whether the fastening is acceptable or not.
[0009] Conventionally, tools for connecting tubular elements are equipped with sensors configured to determine the torque applied during tightening and the number of turns of a first tubular element relative to a second tubular element. These tools make it possible to plot a graph (called a "tightening graph") showing the change in the value of the torque as a function of the number of turns during assembly.
[0010] A solution for assessing the quality of the connection of two tubular elements is the creation of a tightening graph by an expert.
[0011] However, this person may be exposed to risks associated with being on the platform where the connection is made, some criteria may not be clear to the worker, and the above type of assessment may not be satisfactorily reliable as it depends on the worker's skills.
[0012] Other known methods for assessing the connection quality of two tubular elements include automated ones based on machine learning. They rely on a number of parameters of the obtained tightening graph.
[0013] Based on the values of these parameters, a connection status of the tubular elements, representing a compliant or non-compliant state, is associated with the tightening graph, which makes it possible to define whether the connection complies or does not comply with the required specifications.
[0014] However, the probability of an unsatisfactory evaluation remains large. This probability needs to be reduced. Some mismatched connections may be erroneously determined to be properly connected and vice versa. Improper connections can have dramatic safety and environmental consequences. Techniques for evaluating tubular element connections need to be improved to improve the integrity of the formed strings. Summary of the Invention [Problem to be solved by the invention]
[0015] Therefore, the present invention aims to overcome the above mentioned drawbacks and relates to a method for connecting tubular threaded elements, which allows the quality of the connection to be evaluated accurately and reliably. [Means for solving the problem]
[0016] Therefore, a method is proposed for connecting a first threaded portion of a first tubular element with a second threaded portion of a second tubular element, where the first threaded portion and the second threaded portion have a predetermined optimum torque corresponding to the torque to be reached in the final position of the connection. engaging the first threaded portion with the second threaded portion; rotating the first tubular element relative to the second tubular element to tighten the threaded portion; obtaining a tightening graph showing the torque applied during tightening of the first threaded portion against the second threaded portion to a final position as a function of the amount of relative rotation between the first tubular element and the second tubular element, e.g. as a function of the number of rotations of the first tubular element relative to the second tubular element; Includes.
[0017] The method further comprises a step of evaluating the connection quality of the first and second tubular elements based on the first and second models by a pass / fail to the obtained tightening graph and by assigning a connection status representing a conforming or non-conforming state of the connection of the first and second tubular elements, respectively. Here, the first model is configured to reject the tightening graph when at least one primary numerical variable of the obtained tightening graph is outside a range of reference values associated with the at least one primary numerical variable. The range of reference values represents a conforming state of the connection of the first and second tubular elements. Also, the second model is based on an algorithm driven by machine learning based on a reference variable of a reference tightening graph. The reference tightening graph is stored, for example, in a database. The second model is configured to evaluate the connection quality of the first and second tubular elements as a reference variable when the obtained tightening graph was previously passed by the first model.
[0018] The first model, based on primary numerical variables, allows a complete interpretation of the graph failures and allows precise identification of the causes of failures. The second model, based on a machine-learned algorithm, provides a historical summary of tightening curves that were deemed conforming or non-conforming by expert technical judgment. Typically, the second model can be a decision algorithm based on a learned defined numerical model.
[0019] To be able to drive the model, the tightening graph and the tightening parameters need to be reduced to a list of numerical variables, called criterion variables, that correlate to the success or failure of the connection. Thus, the second model does not, strictly speaking, use explicit criteria, but rather criterion variables, each of which describes a property of the graph (e.g., area under the curve, torque difference between two points, slope, etc.).
[0020] In this way, the combination of the first and second models can provide sufficient performance to assess the suitability of the obtained connection.
[0021] Preferably, the learning is performed by segmenting a population of initial reference graphs, the suitability of which is determined, for example, by the technical judgment of an expert. A first part of this population is used for the learning itself, for example by an algorithm, such as a random forest algorithm. Another part of this population is used to measure the validity of the predictions of the algorithm thus trained. This segmentation can be performed repeatedly to statistically check the performance of the model. In this way, the second model, as a complement to the first model, is able to detect with good accuracy almost all graphs that have been rejected in the past by the technical judgment of an expert. The second model detects more past rejections, since it does not only apply the rules as applied by the first model, but also explicitly tries to reproduce past labelings.
[0022] However, unlike the first model, the decision of the second model is difficult to interpret, so it is difficult to know the reason why the graph failed. Therefore, preferably, the tightening graph is evaluated by the first model first, and if the tightening graph obtained by the first model is classified as conforming, then the tightening graph obtained by the second model is evaluated. This strategy can provide an optimal balance between detection performance, interpretability of the decision, and robustness of the model, empirically.
[0023] Advantageously, the primary numerical variables include one or more of the following variables: torque at the final position, torque at a shoulder position where the respective shoulders of the first and second tubular elements contact, the amount of relative rotation between the first and second tubular elements between the shoulder position and the final position (e.g. the number of rotations between the shoulder position and the final position), the slope of the graph between the shoulder position and the final position, torque at a sealing position where the respective sealing seats of the first and second tubular elements contact, and / or the amount of relative rotation between the first and second tubular elements between the sealing position and the shoulder position (e.g. the number of rotations between the sealing position and the shoulder position).
[0024] Preferably, some of the reference tightening graphs (e.g., reference graphs of the database before learning with the second model) are associated with a conforming or non-conforming state of the connections according to the technical judgment of the expert. Furthermore, some of the reference tightening graphs (e.g., reference graphs of the database obtained from learning with the second model) are associated with a conforming or non-conforming state of the connections according to learning without the technical judgment of the expert.
[0025] According to one embodiment, the reference variables include one or more secondary numerical variables. The second model evaluates the connection quality of the first tubular element and the second tubular element as a function of the one or more secondary numerical variables. The one or more secondary numerical variables are calculated based on the respective primary numerical variables and minimum and maximum reference values. The minimum and maximum reference values define a range of reference values associated with the primary numerical variables. The one or more secondary numerical variables are calculated according to the following formula:
[0026]
number
[0027] Preferably, the reference variables include one or more standardized variables. Preferably, the second model evaluates the connection quality of the first tubular element and the second tubular element based on one or more standardized variables calculated as a function of the respective primary numerical variables. The primary numerical variables represent torque. The one or more standardized variables are equal to the ratio of the corresponding primary numerical variables to an optimal torque.
[0028] Advantageously, the reference variable comprises the sum of the torque losses between two successive points of the graph. Preferably, the second model is capable of evaluating the connection quality of the first tubular element and the second tubular element as a function of a standardized value of the sum of the torque losses between two successive points of the obtained tightening graph. The standardized value is equal to the ratio of the sum of the torque losses to the optimal torque.
[0029] Advantageously, the reference variable comprises a gradient between the shoulder position and the final position. Preferably, the second model is capable of evaluating the connection quality of the first tubular element and the second tubular element as a function of the variation of the gradient between the shoulder position and the final position of the obtained tightening graph.
[0030] Advantageously, the reference variable comprises a maximum torque loss value between two successive points of the graph, the two successive points being located between the sealing position and the shoulder position and / or between the shoulder position and the final position. Preferably, the second model is capable of evaluating the connection quality of the first tubular element and the second tubular element as a function of a standardized value of the maximum torque loss value between two successive points of the obtained tightening graph. The standardized value is equal to the ratio of the maximum torque loss value to the optimal torque.
[0031] Advantageously, the primary numerical variables include a tightening speed. Preferably, the first model is capable of evaluating the connection quality of the first and second tubular elements as a function of the tightening speed during the connection of the first and second tubular elements.
[0032] According to one embodiment, the primary numerical variable comprises a loss of linearity of the graph between the shoulder position and the final position. Preferably, the first model is capable of evaluating the connection quality of the first tubular element and the second tubular element as a function of the loss of linearity obtained between the shoulder position and the final position for the obtained tightening graph.
[0033] Advantageously, the primary numerical variable comprises a maximum torque loss value between two successive points of the graph, the two successive points being located between the shoulder position and the final position. Preferably, the first model is capable of evaluating the connection quality of the first tubular element and the second tubular element as a function of a maximum torque loss value between two successive points of the obtained tightening graph, the two successive points being located between the shoulder position and the final position.
[0034] Advantageously, the primary numerical variable comprises the amount of relative rotation between the first and second tubular elements, for example as a number of rotations, during the torque loss occurring between the shoulder position and the final position. Preferably, the first model is capable of evaluating the connection quality of the first and second tubular elements as a function of the amount of relative rotation between the first and second tubular elements during the torque loss occurring between the shoulder position and the final position.
[0035] Advantageously, the primary numerical variable comprises the amount of relative rotation between the first and second tubular elements, for example as a number of rotations, between the engaged position and the final position. The engaged position is a position before the relative rotation between the first and second tubular elements. Preferably, the first model evaluates the connection quality of the first and second tubular elements as a function of the amount of relative rotation between the first and second tubular elements, between the engaged position and the final position.
[0036] Advantageously, the primary numerical variables include a maximum torque before the sealing position. Preferably, the first model is capable of evaluating the connection quality of the first and second tubular elements as a function of the maximum torque value of the tightening graph obtained before the sealing position. The tightening graph obtained is rejected by the first model when the maximum torque value is above 10% of the optimal torque.
[0037] Preferably, the second model includes one or more of the failure criteria of the first model, i.e. the criterion variables include one or more primary numerical variables. [Brief description of the drawings]
[0038] Further objects, advantages and features of the present invention will become apparent from the following description taken in conjunction with the accompanying drawings. [Figure 1A] FIG. 2 is a cross-sectional view of a first tubular element and a second tubular element to be connected in a clamped position. [Figure 1B] FIG. 2 is a cross-sectional view of a first tubular element and a second tubular element to be connected in a sealing position. [Figure 1C] FIG. 2 is a cross-sectional view of a first tubular element and a second tubular element to be connected in a shoulder position. [Diagram 2] 4 is a tightening graph showing the torque applied during connection of a first tubular element and a second tubular element as a function of the number of turns by the first tubular element relative to the second tubular element. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0039] 1A, 1B and 1C show different steps in the connection of a first tubular element 1 and a second tubular element 2. FIG.
[0040] In the embodiment shown, the tubular element 1 is a sleeve-type junction adapted to connect a second tubular element 2 and a third tubular element (not shown).
[0041] The first tubular element 1 and the second tubular element 2 each comprise a threaded portion 3 and 4, advantageously arranged at one of their ends. The thread 5 of the threaded portion 3 and the thread 6 of the threaded portion 4 are configured to interact.
[0042] Also, the first tubular element 1 is provided with a first sealing seat 7 and the second tubular element 2 is provided with a second sealing seat 8. The sealing seats are formed from surfaces intended to seal the assembly when the tubular elements 1 and 2 are connected.
[0043] Furthermore, the first tubular element 1 comprises a first shoulder 9 and the second tubular element 2 comprises a second shoulder 10. The shoulders 9 and 10 form a stop for fastening.
[0044] The sealing seats 7 and 8 and the shoulders 9 and 10, respectively, are arranged to interact with each other.
[0045] There are tubular elements without stops. There are also tubular elements without sealing seats. There are also tubular elements without stops or sealing seats. The invention is suitable to be applied partially to the connections of these types of tubular elements and can be applied to the graph part from the start of the connection to the contact between the sealing seats, to the graph part from the contact between the sealing seats to the end of the connection, or to the graph part from the start of the connection to the end of the connection without any contact or stop between the sealing seats during the connection.
[0046] To connect two tubular elements 1 and 2, the connection method comprises the step of engaging a first tubular element 1 with a second tubular element 2. More specifically, the connection method comprises the step of engaging a first threaded portion 3 with a second threaded portion 4.
[0047] To tighten the threaded portions 3 and 4 , the first tubular element 1 is rotated relative to the second tubular element 2 .
[0048] The tool used for the connection is a tightening wrench. This tightening wrench is provided with a gripper and a motor for rotating the first tubular element 1 and the second tubular element 2 relative to each other. The tightening wrench is also provided with sensors for measuring the number of revolutions and the tightening torque. These sensors are connected to an electronic unit. This allows data on the torque and the revolutions as well as data on the assembly in progress to be stored. The electronic unit is connected to a processing unit which contains processing algorithms. The processing unit is also provided with a user interface which displays the evaluation results and / or tightening graphs obtained during the connection.
[0049] The method therefore includes a step of obtaining a set of points constituting a tightening graph representing the torque applied during tightening of the first tubular element 1 into its final position as a function of the number of rotations of the first tubular element 1 relative to the second tubular element 2.
[0050] A typical profile of a graph showing the torque applied during tightening as a function of the number of revolutions, called a torque / revolution graph, is shown in Figure 2. It can be seen that the graph obtained contains three parts with different slopes.
[0051] The first portion 11 corresponds to the engagement of the first tubular element 1 and the second tubular element 2 and the tightening of the threaded portions 3 and 4, as shown in Figure 1A. The threads 5 and 6 gradually come into contact and a higher torque is applied.
[0052] At the contact point Rl, the first sealing seat 7 of the first tubular element 1 comes into contact with the second sealing seat 8 of the second tubular element 2. The tubular elements 1 and 2 are then in what is called the sealing position, as shown in FIG. 1B.
[0053] The high friction caused by the contact of the sealing seats 7 and 8 causes a change in the slope, in particular an increase in the applied torque per revolution, resulting in the second portion 12 of the graph.
[0054] With continued tightening of the threaded portions 3 and 4, the rotation of the first tubular element 1 relative to the second tubular element 2 reaches the shoulder point Rs. The tubular elements 1 and 2 are now in what is called the shoulder position, as shown in FIG. 1C, where the first shoulder 9 of the first tubular element 1 and the second shoulder 10 of the second tubular element 2 are in contact. The increased friction between the respective surfaces of the shoulders 9 and 10 is added to the friction resulting from the contact between the threads 5 and 6 and the friction between the sealing seats 7 and 8, so that the slope changes again and the applied torque increases significantly, resulting in a third portion 13 of the graph that extends to the final point Rf where the tubular elements reach their final position.
[0055] In the present invention, final position means the position of the first tubular element 1 and the second tubular element 2 when the maximum tightening torque is applied and the connection is completed.
[0056] The contact point Rl, the shoulder point Rs, the final point Rf of the tightening graph and the slope coefficient S between the shoulder point Rs and the final point Rf are parameters of the tightening graph that can be taken into account to determine the suitability of the connection of the first tubular element 1 and the second tubular element 2. The value of each of these parameters depends on the type of tubular elements to be connected.
[0057] The slope coefficient S is the ratio of the torque at the shoulder point Ts, the torque at the final point Tf, and the optimal torque T * The slope coefficient can be calculated based on the slope between the shoulder position and the final position, expressed as the optimal torque T * This is defined as dividing by:
[0058]
number
[0059] Optimal Torque T * means a predefined torque that is reached in the final position, which is specific to the model of the tubular elements and the models of the joints to be connected.
[0060] The connection method comprises a step of evaluating a connection quality of the first tubular element 1 and the second tubular element 2 based on a failure model of the first and second tightening graphs.
[0061] The first and second models pass or fail the resulting tightening graph as a function of the fail criterion. If the resulting tightening graph is passed, a connection status representative of a conforming state of the connection of the first tubular element 1 and the second tubular element 2 is assigned to the tightening graph. If the resulting tightening graph is failed, a connection status representative of a non-conforming state of the connection of the first tubular element 1 and the second tubular element 2 is assigned to the tightening graph.
[0062] The automation resulting from the assessment of the connection quality of the tubular elements by the first and second models makes it possible to eliminate the human factor during the fitness assessment of the connection and thus improves the accuracy of the assessment.
[0063] The first model is configured to reject the tightening graph when the value of at least one primary numerical variable A of the obtained tightening graph deviates from a range of reference values representing a compatibility state of the connection of the first tubular element 1 and the second tubular element 2.
[0064] The first model can rely on an algorithm.
[0065] The second model is based on an algorithm driven by machine learning or artificial intelligence. In particular, the algorithm is driven based on reference variables of a reference tightening graph stored in a database.
[0066] Each of the reference variables considered in the algorithm of the second model represents a characteristic of the tightening graph.
[0067] When the obtained tightening graph has previously been passed by the first model, the second model evaluates the connection quality of the first tubular element 1 and the second tubular element 2 to confirm or deny the passing of the tightening graph by the first model.
[0068] That is, the evaluation of the connection quality of the first tubular element 1 and the second tubular element 2 is performed in a first step by a first failing model and in a second step by a second model when the tightening graph has previously been passed by the first model.
[0069] The primary numerical variable A may be determined on the tightening graph based on parameters including the contact point Rl, the shoulder point Rs, the end point Rf, and / or the slope coefficient S between the shoulder point Rs and the end point Rf of the tightening graph.
[0070] In the illustrated embodiment, the primary numerical variables A taken into account by the first model are the torque at the final position Tf, the torque at the shoulder position Ts, the number of rotations ΔR between the shoulder position and the final position. s-f , and a slope coefficient S.
[0071] According to one embodiment, the torque Tl at the sealing position and the number of rotations ΔR between the sealing position and the shoulder position are l-s Other primary numerical variables A can be considered, such as:
[0072] For each primary numeric variable A considered, a minimum criterion value A min and maximum reference value A max A range of reference values is determined, which is defined by: The range of reference values represents a compatibility state of the connection of the first tubular element 1 and the second tubular element 2.
[0073] According to one embodiment, the minimum reference value A min and maximum reference value A maxcan be determined based on a reference graph that is stored in a database and associated, for each type of tubular element and model of the joint to be connected, with a connection status representing a matched or non-matched state of the connection of a reference first tubular element and a reference second tubular element.
[0074] A successfully connected reference graph is associated with a conforming state, and conversely, a failed connection reference graph is associated with an unconformable state.
[0075] Preferably, the reference tightening graphs are stored in a database. These reference graphs in the database are associated, preferably according to the technical judgement of an expert, with a conforming or non-conforming state of the connection. If the connection proves to be successful, the expert can associate a "conforming" status with the obtained reference graph and pass the connection of the reference tubular element. Conversely, if the connection proves to be unsuccessful, the expert can associate a "non-conforming" status with the obtained reference graph and fail the connection of the reference tubular element. In this way, a reliable and extensive database can be obtained.
[0076] In the illustrated embodiment, the first model is a function of the torque at the final position Tf, the torque at the shoulder position Ts, and the number of rotations ΔR between the shoulder position and the final position. s-f , and the slope coefficient S are associated with the minimum criterion A min or the associated maximum reference value A max is configured to fail the tightening graph if
[0077] To improve the accuracy of the assessment of the connection, additional rejection criteria can be incorporated into the first model.
[0078] Advantageously, the first model is capable of evaluating the connection quality of the first tubular element 1 and the second tubular element 2 as a function of the tightening speed. A too high speed, especially at the shoulder location, may indicate a connection failure.
[0079] Preferably, the tightening graph is rejected when the tightening speed exceeds a predetermined threshold (eg, 5 revolutions per minute).
[0080] According to one embodiment, the rejection criteria for the first model can be based on determining a loss of straightness between the shoulder position and the final position, in particular, a loss of straightness can be indicative of plastic deformation and slippage.
[0081] Linear interpolation can be performed by plotting a straight line between the shoulder point Rs and the final point Rf, passing through the 25% index and the 75% index. Then, the maximum distance between the tightening graph obtained during the connection method and the straight line is calculated.
[0082] The loss of linearity is the average deviation of the graph from the linear interpolation of the optimal torque T * Preferably, the absolute value of the mean deviation is calculated by an algorithm driven by machine learning, which results in a high coefficient of oscillation due to linear interpolation, making it easier to determine.
[0083] When the distance gained by this loss of linearity exceeds a predetermined threshold, the tightening graph is rejected.
[0084] Advantageously, the first model is capable of evaluating the connection quality of the first tubular element 1 and the second tubular element 2 as a function of the maximum torque loss value between two consecutive points of the tightening graph obtained between the shoulder position and the final position.
[0085] In particular, the determination of the maximum torque loss value is advantageous for the detection of slip-related connection defects and can generally indicate the appearance of noise in the power cable.
[0086] The maximum torque loss value means the maximum torque loss determined on the tightening graph.
[0087] Advantageously, the rejection criterion of the first model may be based on the number of revolutions made during the torque loss that occurs between the shoulder position and the final position.
[0088] A brief torque loss may simply correspond to an artifact, however, if the number of revolutions made during the torque loss exceeds a certain threshold, the torque loss may be interpreted as the occurrence of a fault resulting in an incompatible connection.
[0089] Advantageously, the first model is capable of evaluating the connection quality of the first tubular element 1 and the second tubular element 2 as a function of the number of rotations Rf at the final position in order to reject graphs that are too short, indicating an incompatible state of the connection.
[0090] For example, a tightening graph is rejected when the number of revolutions Rf at the final position is below a predefined threshold (eg, equal to one revolution).
[0091] Advantageously, the first model evaluates the connection quality of the first tubular element and the second tubular element as a function of the value of the maximum torque before the sealing position. The tightening graph obtained is characterized by the maximum torque value being the optimum torque T * When the difference is greater than 10% of the first model, it is rejected by the first model.
[0092] The first model is a combination of numerical criteria that represent distinct causes for failing a tightening graph. Each of these fail criteria is calculated based on a numerical criterion and a minimum or maximum threshold value in this numerical criterion. This first model is adaptive and is able to identify the fail scenarios that correspond to the most frequent problems that occur during the connection of two tubular elements 1, 2. It is also able to identify the causes of the fail and the levels at which the thresholds are exceeded and is fully interpretable.
[0093] Preferably, when the first model fails the tightening graph, the failure criteria on which the assignment of a status representative of a non-conforming condition was based is displayed to inform the operator of the exact cause.
[0094] Also preferably, the second model is capable of evaluating the connection quality of the first tubular element and the second tubular element as a function of one or more secondary numerical variables B. Each of the secondary numerical variables B is determined by a primary numerical variable A and a minimum reference value A that defines a range of reference values according to the following formula: min and maximum reference value A max It is calculated based on:
[0095]
number
[0096] The calculated secondary numeric variable B determines the minimum reference value A. min and maximum reference value A max The deviation of the primary numerical variable A from the range of reference values defined by can be measured more easily and with higher sensitivity. If the calculated secondary numerical variable B is below 0 or above 1, this means that the primary numerical variable A is not included within the range of reference values. The algorithm rejects the tightening graph, which is associated with a non-conforming state of the connection.
[0097] As a result, finer and more reliable discrimination regarding fastening quality is possible among various joint models.
[0098] In the illustrated embodiment, the secondary numerical variable B is a torque Tf at the final position, a torque Ts at the shoulder position, and a rotation speed ΔR between the shoulder position and the final position. s-f , and a slope coefficient S.
[0099] For example, the secondary numerical variable Bf related to the final torque is calculated according to the following formula:
[0100]
number
[0101] Preferably, the second model is capable of evaluating the connection quality of the first tubular element 1 and the second tubular element 2 on the basis of one or more standardized variables C. The standardized variables C are calculated as a function of a primary numerical variable A representing a torque. The standardized variables are then used to estimate the optimum torque T * It is equal to the ratio of the primary numeric variable A to . It is defined by the following formula:
[0102]
number
[0103] For example, the normalized variable related to the final torque is calculated according to the following formula:
[0104]
number
[0105] In the illustrated embodiment, the normalized variable C is the torque at the final position Tf, the torque at the shoulder position Ts, the torque at the sealing position Tl, the delta torque ΔT between the sealing position and the shoulder position l-s and the delta torque ΔT between the shoulder position and the final position s-f It is calculated based on:
[0106] Delta torque ΔT between the sealing position and the shoulder position l-scorresponds to the difference in torque value measured between the sealing position and the shoulder position.
[0107] Preferably, the second model is capable of rejecting a tightening graph as a function of the sum of torque losses between two successive points of the obtained tightening graph, which makes it possible in particular to detect thread interference. In this context, the normalized value is the sum of the calculated torque losses relative to a predetermined optimum torque T * It is calculated and defined by dividing by.
[0108] Advantageously, the second model is capable of evaluating the connection quality of the first tubular element 1 and the second tubular element 2 as a function of the variation of the slope after the shoulder position of the obtained tightening graph.
[0109] The slope variation is the deviation from the list defined by the slope between the shoulder position and the final position of the obtained tightening graph, which is the optimal torque T * It is equal to divided by.
[0110] Advantageously, the second model is capable of evaluating the connection quality of the first tubular element 1 and the second tubular element 2 as the maximum torque loss value between two consecutive points of the tightening graph obtained between the sealing position and the shoulder position and / or between the shoulder position and the final position.
[0111] More advantageously, the rejection criterion based on maximum torque loss is based on the optimum torque T * The torque loss can be evaluated based on a normalized value equal to the ratio of the maximum torque loss value to the maximum torque loss value.
[0112] Additionally, the rejection criteria based on maximum torque loss may be evaluated based on a normalized value equal to the ratio of the maximum torque loss value to the shoulder torque Ts.
[0113] Furthermore, the second model can evaluate the connection quality of the first tubular element 1 and the second tubular element 2 as a function of the maximum torque value before the sealing position and / or between the sealing position and the shoulder position. The tightening graph can be calculated based on the maximum torque value at which the optimal torque T * is rejected by the second model when it exceeds 10% of the
[0114] Preferably, the normalized value is defined as the maximum torque divided by the shoulder torque Ts, with high values indicating joint problems.
[0115] Furthermore, the rejection criterion for the second model is the value of the area defined between the graph and the X-axis, i.e., the number of revolutions of the last two revolutions is the optimum torque T * This can be based on dividing the contact point Rl or shoulder point Rs, among others, to reject the tightening graph.
[0116] The second model is based on a machine learning driven algorithm. To drive the model, the tightening graph and tightening parameters are reduced to a list of criterion variables that correlate with tightening success or failure. Each criterion variable represents a feature of the tightening graph.
[0117] Preferably, from a performance point of view, the second model includes one or more of the failure criteria of the first model. Increasing the number of variables describing the graph makes it easier for the algorithm to classify the graph. Learning is then performed on a database of reference tightening graphs.
[0118] The method of connecting a first tubular element 1 and a second tubular element 2 may comprise a step of evaluating the connection quality as a function of a rejection criterion specific thereto, and then establishing a score by means of a second model.
[0119] For example, when the score of the second model is above a predefined threshold, the tightening graph is rejected, and conversely, when the score is below the threshold, the tightening graph is passed.
[0120] This second machine learning model is able to detect with very high accuracy almost all of the tightening graphs that were rejected by the first model.
[0121] The first and second models are complementary and are able to detect approximately 70% and 99% of failure scenarios, respectively. This strategy of evaluating the connection quality of the two tubular elements allows us to empirically provide an optimal balance between detection performance, interpretability of the decisions, and robustness of the overall model.
[0122] This combination allows for the detection of almost all non-compliant connections in an automated manner, without human intervention.
Claims
1. A method for connecting a first threaded portion (3) of a first tubular element (1) with a second threaded portion (4) of a second tubular element (2), said first threaded portion (3) and said second threaded portion (4) are threaded to a predetermined optimum torque (T * ), the method comprising: Engaging the first threaded portion (3) with the second threaded portion (4); rotating the first tubular element (1) relative to the second tubular element (2) to tighten the first threaded portion (3) and the second threaded portion (4); obtaining a tightening graph showing the torque applied during tightening of the first threaded portion (3) against the second threaded portion (4) to the final position as a function of the amount of relative rotation between the first tubular element (1) and the second tubular element (2); Including, evaluating a connection quality of the first tubular element (1) and the second tubular element (2) based on a first model and a second model by pass / fail of the obtained tightening graph and by assignment of a connection status representative of a conforming or non-conforming state of the connection of the first tubular element (1) and the second tubular element (2), the first model is configured to reject the tightening graph when at least one primary numerical variable (A) of the obtained tightening graph is outside a range of reference values associated with at least one primary numerical variable (A), the range of reference values representing a conformity state of the connection of the first tubular element (1) and the second tubular element (2); characterised in that the second model is based on an algorithm driven by machine learning on the basis of reference variables of a reference tightening graph, the second model being configured to evaluate the connection quality of the first tubular element (1) and the second tubular element (2) as a function of the reference variables when the obtained tightening graph has previously been passed by the first model. method.
2. The primary numerical variable (A) is a torque at the final position (Tf), a torque at a shoulder position where the shoulders (9, 10) of the first tubular element (1) and the second tubular element (2) contact (Ts), a relative rotation amount (ΔR) between the first tubular element (1) and the second tubular element (2) between the shoulder position and the final position. s-f ), the slope of the tightening graph between the shoulder position and the final position, the torque (Tl) at a sealing position where the sealing seats (7, 8) of the first tubular element and the second tubular element respectively come into contact, and / or the amount of relative rotation (ΔR l-s 2. The method of claim 1 , further comprising one or more variables of:
3. 3. The method of claim 2, wherein the reference variable comprises a gradient between the shoulder position and the final position, and the second model evaluates the connection quality of the first tubular element and the second tubular element as a function of the variation of the gradient between the shoulder position and the final position of the obtained tightening graph.
4. The reference variables include a maximum torque loss value between two consecutive points of the tightening graph, the two consecutive points being located between the sealing position and the shoulder position and / or between the shoulder position and the final position, and the second model evaluates the connection quality of the first tubular element (1) and the second tubular element (2) as a function of a standardized value of the maximum torque loss value between the two consecutive points of the obtained tightening graph, the standardized value being a function of the optimum torque (T * 4. The method of claim 2 or 3, wherein the ratio of the maximum torque loss value to the maximum torque loss value is equal to the ratio of the maximum torque loss value to the maximum torque loss value.
5. 5. The method according to claim 2, wherein the primary numerical variable (A) comprises a loss of linearity of the tightening graph between the shoulder position and the final position, and the first model evaluates, for the obtained tightening graph, the connection quality of the first tubular element (1) and the second tubular element (2) as a function of the loss of linearity obtained between the shoulder position and the final position.
6. 6. The method according to claim 2, wherein the primary numerical variable (A) comprises a relative rotation amount between the first tubular element (1) and the second tubular element (2) during a torque loss occurring between the shoulder position and the final position, and the first model evaluates the connection quality of the first tubular element (1) and the second tubular element (2) as a function of the relative rotation amount between the first tubular element (1) and the second tubular element (2) during a torque loss occurring between the shoulder position and the final position.
7. 7. The method according to claim 2, wherein the primary numerical variable (A) comprises an amount of relative rotation between the first tubular element (1) and the second tubular element (2) between an engagement position and the final position, the engagement position being a position before relative rotation between the first tubular element (1) and the second tubular element (2), and the first model evaluates a connection quality of the first tubular element (1) and the second tubular element (2) as a function of the amount of relative rotation between the first tubular element (1) and the second tubular element (2) between the engagement position and the final position.
8. The primary numerical variable (A) comprises a maximum torque value before the sealing position, and the first model evaluates the connection quality of the first tubular element (1) and the second tubular element (2) as a function of a maximum torque value of the tightening graph obtained before the sealing position, the tightening graph obtained being such that the maximum torque value is equal to the optimal torque (T * 8. The method of claim 2, wherein the first model is rejected when the first model exceeds 10% of the first model.
9. 9. The method according to any one of claims 1 to 8, wherein the primary numerical variable (A) comprises a maximum torque loss value between two consecutive points of the tightening graph, the two consecutive points being located between the shoulder position and the final position, and the first model evaluates the connection quality of the first tubular element (1) and the second tubular element (2) as a function of the maximum torque loss value between two consecutive points of the obtained tightening graph.
10. The method of any one of claims 1 to 9, wherein a portion of the baseline tightening graph is associated with a compliant or uncompliant state of the connection according to expert engineering judgement.
11. The criterion variables include one or more secondary numerical variables (B), and the second model evaluates the connection quality of the first tubular element and the second tubular element as a function of the one or more secondary numerical variables (B), the one or more secondary numerical variables (B) being determined based on the respective primary numerical variables (A) and a minimum criterion value (A min ) and maximum reference value (A max ) is calculated based on the minimum standard value (A min ) and maximum reference value (A max ) defines a range of reference values associated with said primary numerical variable (A), and said one or more secondary numerical variables (B) are calculated according to the formula: [0070] Here, B is the secondary numerical variable, A is the primary numerical variable, A min is the minimum reference value equal to the lower limit of the range of reference values associated with the primary numerical variable, A max A method according to any one of claims 1 to 10, wherein σ represents the maximum reference value equal to the upper limit of the range of reference values associated with the primary numerical variable.
12. The reference variables include one or more standardized variables (C), and the second model evaluates the connection quality of the first tubular element and the second tubular element based on the one or more standardized variables (C) calculated as a function of each of the primary numerical variables (A), the primary numerical variables (A) representing torque, and the one or more standardized variables (C) representing the optimal torque (T * 12. The method according to claim 1, wherein the first-order numerical variable (A) is equal to the ratio of the first-order numerical variable (A) to the corresponding first-order numerical variable (A).
13. The reference variable comprises a sum of torque losses between two successive points of the tightening graph, and the second model evaluates the connection quality of the first tubular element (1) and the second tubular element (2) as a function of a standardized value of the sum of torque losses between two successive points of the obtained tightening graph, the standardized value being a function of the optimum torque (T * The method according to any one of claims 1 to 12, wherein the torque loss is equal to the ratio of the sum of the calculated torque losses to the torque loss ratio (Tw).
14. 14. The method according to any one of claims 1 to 13, wherein the primary numerical variable (A) comprises a tightening speed and the first model evaluates a connection quality of the first tubular element (1) and the second tubular element (2) as a function of the tightening speed during connection of the first tubular element (1) and the second tubular element (2).
15. The method of any one of claims 1 to 14, wherein the second model includes one or more failure criteria of the first model.
Citation Information
Patent Citations
Autonomous connection evaluation and automated shoulder detection for tubular makeup
US20180224029A1
Method for assessing the quality of the coupling of two tubular components
US20210088396A1
