Training method of detection model and detection method of bolt tightening quality

Abstract

The embodiment of the invention provides a training method of a detection model and a detection method of bolt tightening quality. The training method of the detection model comprises the following steps: generating a curve sample set based on historical tightening curves of different bolt stations; wherein each historical tightening curve corresponds to one quality category; determining the corresponding accuracy of the detection model under the curve sample set; wherein the accuracy rate represents the correctness degree of the quality category corresponding to the tightening curve identified by the detection model; and adjusting parameters of the detection model through an objective function constructed by the accuracy rate, so that the accuracy rate of the detection model when identifying the quality category corresponding to the tightening curve meets a preset target accuracy rate. According to the technical scheme provided by the embodiment of the invention, the trained detection model can be utilized to output the corresponding quality category according to the to-be-detected tightening curve, so that the purpose of identifying the tightening defect generated in the tightening operation of the bolt is achieved.

Description

Technical Field

[0001] The present application relates to the field of detection technology, and in particular to a method for training a detection model and a method for detecting bolt tightening quality. Background Art

[0002] With the continuous development of production technology, bolts acting on the connectors of various components in fixed equipment have become a frequently used detachable connection method.

[0003] In the related art, in order to determine the performance of a bolt during a tightening operation, it is usually judged whether the final torque output by the bolt station used to perform the bolt tightening operation reaches a set value. This makes it difficult to detect tightening defects caused by bolts and nuts during the tightening operation. For example, impurities in the thread will cause the bolt thread to deform during the tightening operation, thereby failing to maintain the ideal clamping force for a long time, and other potential faults.

[0004] Therefore, how to identify the tightening defects generated during the bolt tightening operation has become a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] To solve the above technical problems, the embodiments of the present application provide a method for training a detection model and a method for detecting the quality of bolt tightening, a computer-readable storage medium, and an electronic device.

[0006] According to one aspect of an embodiment of the present application, a method for training a detection model is provided, the training method comprising: generating a curve sample set based on historical tightening curves of different bolt stations; wherein each historical tightening curve corresponds to a quality category; determining the accuracy of the detection model corresponding to the curve sample set; wherein the accuracy represents the degree of correctness of the detection model in identifying the quality category corresponding to the tightening curve; and adjusting the parameters of the detection model through an objective function constructed by the accuracy, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets a preset target accuracy.

[0007] According to one aspect of an embodiment of the present application, a method for detecting bolt tightening quality is provided, the detection method comprising: obtaining a tightening curve to be detected generated by a target bolt station; determining a target quality category corresponding to the tightening curve to be detected based on a preset basic tightening curve and the tightening curve to be detected through a detection model, wherein the preset basic tightening curve corresponds to a quality category, and the detection model is obtained using the training method of the detection model in the above embodiment.

[0008] According to one aspect of an embodiment of the present application, a training device for a detection model is provided, comprising: a sample acquisition module configured to generate a curve sample set based on historical tightening curves of different bolt stations; wherein each historical tightening curve corresponds to a quality category; a sample identification module configured to determine the accuracy of the detection model under the curve sample set; wherein the accuracy represents the correctness of the detection model in identifying the quality category corresponding to the tightening curve; and a parameter adjustment module configured to adjust the parameters of the detection model through an objective function constructed based on the accuracy, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets a preset target accuracy.

[0009] In some embodiments of the present application, based on the aforementioned scheme, the sample recognition module is further configured to: output multiple basic tightening curves and tightening curves to be detected according to the curve sample set; input the multiple basic tightening curves and the tightening curves to be detected into the detection model to obtain curve features corresponding to the multiple basic tightening curves and the tightening curves to be detected; determine the similarity between the tightening curve to be detected and each basic tightening curve based on the curve features of the multiple basic tightening curves and the curve features of the tightening curve to be detected; determine the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve based on the similarity between the tightening curve to be detected and each basic tightening curve; determine the accuracy of the detection model based on the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve and the quality category corresponding to the tightening curve to be detected.

[0010] In some embodiments of the present application, based on the aforementioned scheme, the parameter adjustment module is further configured as follows: step S1, calculating the current gradient direction based on the objective function; step S2, adjusting the parameters of the detection model according to the current gradient direction and the preset learning rate; step S3, looping step S1 and step S2 until the training stop condition is met.

[0011] According to one aspect of an embodiment of the present application, a device for detecting the quality of bolt tightening is provided, comprising: a module for acquiring a sample to be detected, configured to obtain a tightening curve to be detected generated by a target bolt station; and an identification and processing module, configured to determine, through a detection model, a target quality category corresponding to the tightening curve to be detected based on a preset basic tightening curve and the tightening curve to be detected, wherein the preset basic tightening curve corresponds to a quality category, and the detection model is obtained by using the training method of the detection model in the above-mentioned embodiment.

[0012] In some embodiments of the present application, based on the aforementioned scheme, the identification and processing module is further configured to: determine a test tightening curve from the historical tightening records of the target bolt station before determining the target quality category corresponding to the tightening curve to be detected based on the historical tightening curve and the tightening curve to be detected through the detection model; wherein the test tightening curve corresponds to a quality category; input the preset basic tightening curve and the test tightening curve into the detection model to determine the accuracy of the detection model; and adjust the parameters of the detection model according to the objective function constructed by the accuracy to maximize the accuracy of the detection model in identifying the quality category corresponding to the tightening curve generated by the target bolt station.

[0013] In some embodiments of the present application, based on the aforementioned scheme, the recognition and processing module is further configured to: input the preset basic tightening curve and the tightening curve to be detected into the detection model respectively to obtain curve features corresponding to the preset basic tightening curve and the tightening curve to be detected; determine the similarity between the tightening curve to be detected and the preset basic tightening curve based on the curve features of the preset basic tightening curve and the curve features of the tightening curve to be detected; if the similarity reaches a preset similarity threshold, then use the quality category of the preset basic tightening curve as the target quality category.

[0014] In some embodiments of the present application, based on the aforementioned scheme, under the condition that the preset basic tightening curves include multiple ones and each preset basic tightening curve corresponds to a different quality category, the recognition processing module is further configured to: input the multiple preset basic tightening curves and the tightening curve to be detected into the target detection model respectively to obtain curve features corresponding to each preset basic tightening curve and the tightening curve to be detected; determine the similarity between the tightening curve to be detected and each preset basic tightening curve based on the curve features of each preset basic tightening curve and the curve features of the tightening curve to be detected; and take the quality category corresponding to the preset basic tightening curve with the highest similarity between the tightening curve to be detected and each preset basic tightening curve as the target quality category.

[0015] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer executes the training method of the detection model or the method for detecting the bolt tightening quality as described in the above embodiments.

[0016] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device implements the training method of the detection model or the method for detecting the bolt tightening quality as described in the above embodiments.

[0017] In the technical solution of the embodiment of the present application, a curve sample set is first generated based on the historical tightening curves of different bolt stations, and each historical tightening curve corresponds to a quality category. Then, the accuracy of the detection model under the curve sample set is determined, and then the parameters of the detection model are adjusted through the objective function constructed by the accuracy, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets the preset target accuracy. Subsequently, the trained detection model can be used to output the corresponding quality category according to the tightening curve to be detected, thereby achieving the purpose of identifying the tightening defects generated by the bolts during the tightening operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, serving to explain the principles of the present application. It is obvious that the drawings described below are merely some embodiments of the present application, and a person of ordinary skill in the art can derive other drawings based on these drawings without inventive effort. In the drawings:

[0019] Figure 1 is a flowchart of a method for training a detection model shown in an exemplary embodiment of the present application;

[0020] Figure 2 yes Figure 1 The flowchart of step S130 in the illustrated embodiment in an exemplary embodiment;

[0021] Figure 3 This is a flow chart of a method for detecting bolt tightening quality shown in an exemplary embodiment of the present application;

[0022] Figure 4 yes Figure 3 A flowchart in an exemplary embodiment before step S220 in the illustrated embodiment;

[0023] Figure 5 is a block diagram of a detection model training device shown in an exemplary embodiment of the present application;

[0024] Figure 6 is a block diagram of a device for detecting bolt tightening quality shown in an exemplary embodiment of the present application;

[0025] Figure 7 It is a structural diagram of an electronic device shown in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0026] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art.

[0027] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner.In the following description, many specific details are provided so as to provide a full understanding of the embodiments of the present application. However, it will be appreciated by those skilled in the art that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps etc. can be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0028] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0029] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0030] It should be noted that the term "plurality" used in this document refers to two or more. "And / or" describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. The character " / " generally indicates an "or" relationship between the associated objects.

[0031] The technical solution of the embodiment of the present application proposes a method for training a detection model. The execution subject of the method can be a training device or equipment for the detection model, such as a personal computer or a computer cluster, or a terminal with a model training function, such as a laptop, a smart phone, a tablet computer, etc. Figure 1 The method includes at least steps S110 to S130, which are described in detail as follows:

[0032] In step S110 , a curve sample set is generated based on historical tightening curves of different bolt stations.

[0033] It's important to note that bolt stations are used to tighten bolts. During this process, the bolts must be screwed into the holes and tightened to a certain torque to prevent them from falling out. A tightening curve characterizes the change in torque during the tightening process, indicating whether the tightening has been successful.

[0034] Each historical tightening curve corresponds to a quality category. This quality category can be divided based on the bolt's performance during the tightening operation, for example, into categories such as qualified torque, abnormal torque fluctuation, and bolt seizure.

[0035] The method of generating a curve sample set based on the historical tightening curves of different bolt stations can be flexibly set as needed. In one example, the historical tightening curve of each bolt station can be directly added to the curve sample set to achieve the purpose of generating a curve sample set.

[0036] In another example, the target bolt station can be first determined from different bolt stations, and then the historical tightening curve of the target bolt station can be added to the curve sample set to achieve the purpose of generating the curve sample set. At the same time, the targetedness of the curve sample set can be improved, that is, the historical tightening curves of the bolt stations in certain areas can be collected.

[0037] In step S120 , the accuracy of the detection model corresponding to the curve sample set is determined.

[0038] Among them, the accuracy rate represents the correctness of the detection model in identifying the quality category corresponding to the tightening curve.

[0039] In an embodiment of the present application, after the curve sample set is generated, the accuracy of the detection model corresponding to the curve sample set can be determined, that is, the correctness of the detection model's current identification of the tightening curve into its corresponding quality category can be determined.

[0040] In order to determine the accuracy of the detection model under the curve sample set, in some embodiments of the present application, multiple base tightening curves and tightening curves to be tested can be output based on the curve sample set. For example, a historical tightening curve can be determined from the curve sample set as the tightening curve to be tested, and then other historical tightening curves in the curve sample set can be used as base tightening curves. Alternatively, a historical tightening curve can be determined from the curve sample set as the tightening curve to be tested, and then other historical tightening curves in the curve sample set, other than the historical tightening curve, from the same bolt station as the historical tightening curve can be used as base tightening curves.

[0041] Then, multiple basic tightening curves and the tightening curve to be tested are input into the detection model to obtain the curve features corresponding to the multiple basic tightening curves and the tightening curve to be tested. Then, the similarity between the tightening curve to be tested and each basic tightening curve is determined based on the curve features of the multiple basic tightening curves and the curve features of the tightening curve to be tested.

[0042] The mathematical expression for determining the similarity between the tightening curve to be tested and each basic tightening curve includes:

[0043]

[0044] c is the cosine similarity, is the curve feature output by the detection model based on the tightening curve to be detected, and h(xi) is the curve feature output by the detection model based on the i-th basic tightening curve.

[0045] Then, based on the similarity between the tightening curve to be detected and each basic tightening curve, the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve is determined.

[0046] The mathematical expression for determining the probability that the quality category of the tightening curve to be tested is the quality category corresponding to each basic tightening curve includes:

[0047]

[0048] is the probability that the quality category of the tightening curve to be detected is the quality category corresponding to the i-th basic tightening curve, and k is the number of basic tightening curves.

[0049] Finally, the accuracy of the detection model is determined according to the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve and the quality category corresponding to the tightening curve to be detected.

[0050] The mathematical expressions for determining the accuracy of the detection model include:

[0051]

[0052] is the accuracy of the detection model, S is a set of multiple basic tightening curves, y is the quality category corresponding to the tightening curve to be detected, and y i is an m-dimensional one-hot encoding vector consisting of the quality categories corresponding to each basic tightening curve, m is the total number of quality categories, y i The value of is 1 when the quality category indicated by it corresponds to the quality category of the tightening curve to be tested, and 0 in other cases.

[0053] In some embodiments of the present application, in order to determine the accuracy of the detection model corresponding to the curve sample set, multiple basic tightening curves and multiple tightening curves to be detected can be output according to the curve sample set, wherein the multiple basic tightening curves and the multiple tightening curves to be detected can come from the same bolt station or from different bolt stations; then the accuracy of the detection model is determined for each tightening curve to be detected according to the above embodiment, thereby determining the accuracy of the detection model corresponding to each tightening curve to be detected, so as to further understand the correctness of the detection model.

[0054] Based on this, the mathematical expressions for determining the similarity between the tightening curve to be detected and each basic tightening curve, determining the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve, and determining the accuracy of the detection model can be adjusted according to the following formula:

[0055]

[0056] in, is the lth tightening curve to be tested, is the accuracy of the detection model under the lth tightening curve to be detected.

[0057] In some embodiments of the present application, multiple support sets and multiple query sets can also be output based on the curve sample set, wherein each support set includes multiple historical tightening curves as basic tightening curves, and each query set includes multiple historical tightening curves as tightening curves to be detected, wherein the multiple support sets and multiple query sets can correspond to each other in pairs, and the two corresponding support sets and query sets are both from the same bolt station, and then for each tightening curve to be detected in each query set, the accuracy of the detection model is determined according to the above embodiment under different support sets, thereby determining the accuracy of the detection model in different support sets under each tightening curve to be detected, so as to further understand the correctness of the detection model.

[0058] Based on this, the mathematical expressions for determining the similarity between the tightening curve to be detected and each basic tightening curve, determining the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve, and determining the accuracy of the detection model can be adjusted according to the following formula:

[0059]

[0060] in, is the lth tightening curve to be tested in the bth query set, x si is the tightening curve of the ith foundation in the sth support set, is the accuracy of the detection model in the u-th support set S under the l-th tightening curve to be detected in the b-th query set.

[0061] In some embodiments of the present application, after outputting a support set and a query set based on a curve sample set, each basic tightening curve in the support set can be input into the first neural network of a detection model to obtain curve features for each basic tightening curve. Simultaneously, each tightening curve to be tested in the query set is input into the second neural network of the detection model to obtain curve features for each tightening curve to be tested. By simultaneously processing the basic tightening curves and the tightening curves to be tested via the first and second neural networks, the time required to obtain the corresponding curve features for each basic tightening curve and the tightening curve to be tested is shortened.

[0062] The first and second neural networks can be flexibly selected as needed, for example, a multi-layer perceptron, a dilated causal convolutional network, various LSTM-based models, and various transformer-based models. Correspondingly, different neural networks can employ different extraction methods when extracting curve features from the tightening curve, such as temporal features or statistical features, thereby enhancing the recognition capabilities of the detection model by increasing the differentiation between curve features.

[0063] Based on this, the mathematical expressions for determining the similarity between the tightening curve to be detected and each basic tightening curve and determining the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve can be adjusted according to the following formula:

[0064]

[0065] in, is the curve feature output by the second neural network based on the lth tightening curve to be detected in the bth query set, g(x si ) is the curve feature output by the first neural network based on the i-th tightening curve to be detected in the s-th support set.

[0066] In step S130 , the parameters of the detection model are adjusted through the objective function constructed by the accuracy, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets the preset target accuracy.

[0067] In the embodiments of the present application, after determining the accuracy of the detection model, the parameters of the detection model can be adjusted using an objective function constructed based on the accuracy. This is done so that, after training the detection model, the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets a preset target accuracy. The objective function is used to guide the adjustment direction of the parameters to be adjusted in the detection model.

[0068] In some embodiments of the present application, under the condition of having a tightening curve to be detected and multiple basic tightening curves, the mathematical expression of the objective function constructed by the accuracy rate includes:

[0069]

[0070] Among them, θ is the parameter to be adjusted of the detection model.

[0071] In some embodiments of the present application, under the condition of having multiple support sets and multiple query sets, the mathematical expression of the objective function constructed by accuracy can be adjusted to:

[0072]

[0073] Among them, E S～L To extract the expectation of support set from the task, the task consists of multiple support sets and multiple query sets, E B～L is the expectation of extracting the query set from the task.

[0074] In addition, in order to further ensure that the accuracy of the detection model can be improved after the parameters of the detection model are adjusted, the amount of data can be further expanded. In some embodiments of the present application, a task set consisting of multiple tasks can be constructed. Based on this, the mathematical expression of the objective function constructed by the accuracy rate can be adjusted to:

[0075]

[0076] Among them, E L～T is the expectation of extracting tasks from the task set.

[0077] The method of adjusting the parameters of the detection model through the objective function constructed by accuracy can be flexibly set as needed. In one example, the current gradient direction can be calculated based on the objective function, and then the parameters of the detection model can be adjusted according to the current gradient direction and the preset learning rate, wherein the gradient direction represents the direction of parameter update, and the preset learning rate represents the step size of the parameter update, so as to achieve the purpose of adjusting the parameters of the detection model.

[0078] In another example, the process of adjusting the parameters of the detection model by the objective function constructed by accuracy can be referred to Figure 2 The detailed description is as follows:

[0079] Step S1, calculating the current gradient direction based on the objective function;

[0080] Step S2, adjusting the parameters of the detection model according to the current gradient direction and the preset learning rate;

[0081] Step S3, looping through steps S1 and S2 until the training stop condition is met.

[0082] In the above process, steps S1 and S2 are executed in a loop so that the parameters of the detection model can be continuously adjusted according to the gradient direction and the preset learning rate, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve can be gradually improved and optimized until the accuracy of the detection model meets the preset target accuracy.

[0083] Among them, in addition to using the accuracy reaching the preset target accuracy as the training stop condition, it can also be flexibly set according to needs, for example, the change in accuracy in several consecutive cycles is lower than the preset threshold, the number of cycles reaches a preset number, etc.

[0084] Through the above implementation method, a curve sample set is first generated based on the historical tightening curves of different bolt stations, and each historical tightening curve corresponds to a quality category. Then, the accuracy of the detection model under the curve sample set is determined, and then the parameters of the detection model are adjusted through the objective function constructed by the accuracy, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets the preset target accuracy. In this way, the trained detection model can be used to output the corresponding quality category according to the tightening curve to be detected, thereby achieving the purpose of identifying the tightening defects generated by the bolts during the tightening operation.

[0085] In the above embodiment, the training method of the detection model is mainly described. The following embodiment describes the application method of the detection model. Figure 3 As shown, Figure 3 This is a flow chart of a method for detecting bolt tightening quality proposed by the technical solution of the embodiment of the present application. The method includes at least steps S210 to S220, which are described in detail as follows:

[0086] In step S210, a tightening curve to be detected generated by the target bolt station is obtained.

[0087] The target bolt station is the bolt station whose bolt tightening performance needs to be tested. The tightening curve to be tested is a tightening curve generated by the torque change of the bolt when the target bolt station performs the bolt tightening operation.

[0088] In step S220 , the target quality category corresponding to the tightening curve to be detected is determined by the detection model based on the preset basic tightening curve and the tightening curve to be detected.

[0089] Therein, a preset basic tightening curve corresponds to a quality category.

[0090] It's important to note that during equipment assembly, the torque required by different bolting stations during bolt tightening operations will vary as the torque required for different bolt holes in the equipment changes. This means that the tightening curves used to determine whether a bolt tightening operation is acceptable will vary depending on the station.

[0091] In an embodiment of the present application, in addition to directly presetting a universal tightening curve as a preset basic tightening curve, a preset basic tightening curve can also be determined based on the historical tightening records of the target bolt station to improve the adaptability between the preset basic tightening curve and the target bolt station.

[0092] The method of determining the preset basic tightening curve according to the historical tightening records of the target bolt station can be flexibly set as needed. In one example, the tightening curve with the latest recording time in the historical tightening records can be used as the preset basic tightening curve.

[0093] In another example, the operating parameters of the target bolt station may be obtained first, and then a tightening curve associated with the operating parameters may be filtered out from historical tightening records, and the filtered tightening curve may be used as a preset basic tightening curve.

[0094] Secondly, the detection model is obtained by using the training method of the detection model in any of the above embodiments.

[0095] In an embodiment of the present application, after the tightening curve to be detected is obtained, the target quality category corresponding to the tightening curve to be detected can be determined by a detection model based on a preset basic tightening curve and the tightening curve to be detected.

[0096] In some embodiments of the present application, in order to determine the target quality category corresponding to the tightening curve to be detected, the preset basic tightening curve and the tightening curve to be detected can be respectively input into the detection model to obtain the curve characteristics corresponding to the preset basic tightening curve and the tightening curve to be detected, and then the similarity between the tightening curve to be detected and the preset basic tightening curve is determined based on the curve characteristics of the preset basic tightening curve and the curve characteristics of the tightening curve to be detected. Thereafter, the target quality category corresponding to the tightening curve to be detected can be determined by judging whether the similarity reaches a preset similarity threshold.

[0097] Specifically, if the similarity reaches a preset similarity threshold, indicating that the tightening curve to be tested is similar to the preset basic tightening curve, the quality category of the preset basic tightening curve is used as the target quality category. That is, the performance of the bolt in the tightening operation represented by the target quality category of the tightening curve to be tested is consistent with the quality category of the preset basic tightening curve. Conversely, if the similarity does not reach the preset similarity threshold, indicating that the tightening curve to be tested is not similar to the preset basic tightening curve, the quality category corresponding to the quality category of the preset basic tightening curve is used as the target quality category. That is, if the quality category of the preset basic tightening curve represents that the bolt's performance in the tightening operation is qualified, the target quality category of the tightening curve to be tested represents that the bolt's performance in the tightening operation is unqualified.

[0098] In addition, to finely classify the performance of bolts during tightening operations, in some embodiments of the present application, multiple preset basic tightening curves may be included, and each preset basic tightening curve corresponds to a different quality category. Accordingly, in order to determine the target quality category corresponding to the tightening curve to be detected, the multiple preset basic tightening curves and the tightening curve to be detected can be first input into the target detection model to obtain the curve features corresponding to each preset basic tightening curve and the tightening curve to be detected. Then, based on the curve features of each preset basic tightening curve and the curve features of the tightening curve to be detected, the similarity between the tightening curve to be detected and each preset basic tightening curve is determined. Then, the quality category corresponding to the preset basic tightening curve with the highest similarity between the tightening curve to be detected and each preset basic tightening curve is used as the target quality category. In this way, while determining the target quality category corresponding to the tightening curve to be detected, the range of quality categories that can be identified by the detection model is expanded, thereby achieving the purpose of finely classifying the performance of the target bolt station when performing the bolt tightening operation.

[0099] Through the above implementation, the tightening curve to be detected generated by the target bolt station is first obtained, and then the target quality category corresponding to the tightening curve to be detected is determined based on the preset basic tightening curve and the tightening curve to be detected by the detection model, thereby determining the performance of the bolt during the tightening operation.

[0100] See also Figure 4 , Figure 4 FIG. 1 is a flow chart showing a method for detecting the quality of bolt tightening according to another exemplary embodiment. Figure 4 As shown, in Figure 3 Before step S220 in the illustrated embodiment, the method may further include steps S310 to S330, which are described in detail as follows:

[0101] In step S310 , a test tightening curve is determined from historical tightening records of the target bolt station.

[0102] In an embodiment of the present application, before determining the target quality category corresponding to the tightening curve to be detected based on a preset basic tightening curve and the tightening curve to be detected through a detection model, a test tightening curve can also be determined from the historical tightening records of the target bolt station, wherein the test tightening curve corresponds to a quality category.

[0103] The method of determining the test tightening curve from the historical tightening records can be flexibly set as needed. In one example, any tightening curve in the historical tightening records can be used as the test tightening curve.

[0104] In another example, the operating parameters corresponding to the preset basic tightening curve may be determined first, and then tightening curves associated with the operating parameters other than the preset basic tightening curve may be screened out from historical tightening records, and the screened tightening curves may be used as test tightening curves.

[0105] In step S320 , the preset basic tightening curve and the test tightening curve are input into the detection model to determine the accuracy of the detection model.

[0106] In an embodiment of the present application, after the test tightening curve is determined, the preset basic tightening curve and the test tightening curve may be input into the detection model to determine the accuracy of the detection model.

[0107] In particular, under the condition that the preset basic tightening curve and the test tightening curve are used as inputs to the detection model, the method for determining the accuracy of the detection model can refer to the embodiment described in the above-mentioned step S120. Specifically, the preset basic tightening curve and the test tightening curve can be first input into the detection model to obtain the curve characteristics corresponding to the preset basic tightening curve and the test tightening curve, respectively. Then, the similarity between the test tightening curve and the preset basic tightening curve is determined based on the curve characteristics of the preset basic tightening curve and the curve characteristics of the test tightening curve. Then, based on the similarity between the test tightening curve and the preset basic tightening curve, the probability that the quality category of the test tightening curve is the quality category corresponding to the preset basic tightening curve is determined. Finally, the accuracy of the detection model is determined based on the probability that the quality category of the test tightening curve is the quality category corresponding to the preset basic tightening curve and the quality category corresponding to the test tightening curve.

[0108] In addition, in the process of inputting the preset basic tightening curve and the test tightening curve into the detection model to determine the accuracy of the detection model, multiple preset basic tightening curves or multiple test tightening curves can be selected as needed to participate in the calculation of the detection model accuracy to improve the authenticity of the accuracy of the determined detection model.

[0109] In step S330, the parameters of the detection model are adjusted by the objective function constructed by the accuracy, so as to maximize the accuracy of the detection model in identifying the quality category corresponding to the tightening curve generated by the target bolt station.

[0110] In an embodiment of the present application, after determining the accuracy of the detection model under the condition that the preset basic tightening curve and the test tightening curve are used as the input of the detection model, the parameters of the detection model can be adjusted through the objective function constructed by the accuracy, so as to maximize the accuracy of the detection model in identifying the quality category corresponding to the tightening curve generated by the target bolt station, thereby facilitating the improvement of the accuracy of the target quality category subsequently determined by the detection model for the tightening curve to be detected.

[0111] The method of adjusting the parameters of the detection model through the objective function constructed by the accuracy rate can refer to the embodiment recorded in the above step S130, and will not be repeated here.

[0112] The following describes an embodiment of a detection model training device in this application, which can be used to execute the detection model training method in the above-mentioned embodiment of this application. For details not disclosed in the embodiment of the device of this application, please refer to the embodiment of the detection model training method in the above-mentioned embodiment of this application.

[0113] Figure 5 FIG. 1 is a block diagram of a detection model training device 100 according to an embodiment of the present application.

[0114] Reference Figure 5 As shown, a training device 100 for a detection model according to an embodiment of the present application includes: a sample acquisition module 110, configured to generate a curve sample set based on historical tightening curves of different bolt stations; wherein each historical tightening curve corresponds to a quality category; a sample identification module 120, configured to determine the accuracy of the detection model under the curve sample set; wherein the accuracy represents the correctness of the detection model in identifying the quality category corresponding to the tightening curve; a parameter adjustment module 130, configured to adjust the parameters of the detection model through an objective function constructed by the accuracy, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets the preset target accuracy.

[0115] In some embodiments of the present application, based on the aforementioned scheme, the sample identification module 120 is further configured to: output a plurality of basic tightening curves and a tightening curve to be detected according to the curve sample set; input the plurality of basic tightening curves and the tightening curve to be detected into the detection model to obtain curve features corresponding to the plurality of basic tightening curves and the tightening curve to be detected; determine the similarity between the tightening curve to be detected and each basic tightening curve based on the curve features of the plurality of basic tightening curves and the curve features of the tightening curve to be detected; determine the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve based on the similarity between the tightening curve to be detected and each basic tightening curve; determine the accuracy of the detection model based on the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each basic tightening curve and the quality category corresponding to the tightening curve to be detected.

[0116] In some embodiments of the present application, based on the aforementioned scheme, the parameter adjustment module 130 is further configured as follows: step S1, calculating the current gradient direction based on the objective function; step S2, adjusting the parameters of the detection model according to the current gradient direction and the preset learning rate; step S3, looping through steps S1 and S2 until the training stop condition is met.

[0117] It should be noted that the detection model training device 100 provided in the above embodiment and the detection model training method provided in the above embodiment belong to the same concept, and the specific way in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here.

[0118] The following describes an embodiment of a bolt tightening quality detection device in this application, which can be used to perform the bolt tightening quality detection method described in the above embodiments of this application. For details not disclosed in the embodiment of the device in this application, please refer to the embodiment of the bolt tightening quality detection method described in the above embodiments of this application.

[0119] Figure 6 FIG. 4 is a block diagram of a device 200 for detecting bolt tightening quality according to an embodiment of the present application.

[0120] Reference Figure 6 As shown, a device 200 for detecting the quality of bolt tightening according to an embodiment of the present application includes: a sample acquisition module 210 to be detected, configured to obtain a tightening curve to be detected generated by a target bolt station; an identification processing module 220, configured to determine a target quality category corresponding to the tightening curve to be detected based on a preset basic tightening curve and the tightening curve to be detected through a detection model, wherein the preset basic tightening curve corresponds to a quality category, and the detection model is obtained by using the training method of the detection model in the above embodiment.

[0121] In some embodiments of the present application, based on the aforementioned scheme, the identification processing module 220 is further configured to: determine a test tightening curve from the historical tightening records of the target bolt station before determining the target quality category corresponding to the tightening curve to be detected based on the historical tightening curve and the tightening curve to be detected through the detection model; wherein the test tightening curve corresponds to a quality category; input the preset basic tightening curve and the test tightening curve into the detection model to determine the accuracy of the detection model; and adjust the parameters of the detection model according to the objective function constructed by the accuracy to maximize the accuracy of the detection model in identifying the quality category corresponding to the tightening curve generated by the target bolt station.

[0122] In some embodiments of the present application, based on the aforementioned scheme, the identification processing module 220 is further configured to: input the preset basic tightening curve and the tightening curve to be detected into the detection model respectively to obtain curve features corresponding to the preset basic tightening curve and the tightening curve to be detected; determine the similarity between the tightening curve to be detected and the preset basic tightening curve based on the curve features of the preset basic tightening curve and the curve features of the tightening curve to be detected; if the similarity reaches a preset similarity threshold, then use the quality category of the preset basic tightening curve as the target quality category.

[0123] In some embodiments of the present application, based on the aforementioned scheme, under the condition that the preset basic tightening curves include multiple ones and each preset basic tightening curve corresponds to a different quality category, the recognition processing module 220 is further configured to: input the multiple preset basic tightening curves and the tightening curve to be detected into the target detection model respectively to obtain curve features corresponding to each preset basic tightening curve and the tightening curve to be detected; determine the similarity between the tightening curve to be detected and each preset basic tightening curve based on the curve features of each preset basic tightening curve and the curve features of the tightening curve to be detected; and take the quality category corresponding to the preset basic tightening curve with the highest similarity between the tightening curve to be detected and each preset basic tightening curve as the target quality category.

[0124] It should be noted that the bolt tightening quality detection device 200 provided in the above embodiment and the bolt tightening quality detection method provided in the above embodiment belong to the same concept, and the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here.

[0125] An embodiment of the present application also provides an electronic device, including a processor and a memory, wherein the memory stores computer-readable instructions, which, when executed by the processor, implement the aforementioned detection model training method or bolt tightening quality detection method.

[0126] Figure 7A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown.

[0127] It should be noted that Figure 7 The computer system 300 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0128] like Figure 7 As shown, the computer system 300 includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. Various programs and data required for system operation are also stored in the RAM 303. The CPU 301, ROM 302 and RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0129] The following components are connected to the I / O interface 305: an input section 306 including a keyboard, a mouse, and the like; an output section 307 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 308 including a hard disk and the like; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. Removable media 311, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 310 as needed, so that computer programs read therefrom can be installed into the storage section 308 as needed.

[0130] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, the various functions defined in the system of the present application are executed.

[0131] It should be noted that the computer-readable medium shown in the embodiments of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable computer program. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0132] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0133] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0134] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments, or may exist independently without being incorporated into the electronic device. The computer-readable storage medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device implements the method described in the above embodiments.

[0135] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the application, the features and functions of two or more modules or units described above can be concretized in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0136] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of the present application.

[0137] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of this application and include common knowledge or customary techniques in the art that are not disclosed herein.

[0138] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for training a detection model, characterized in that: The method comprises: Generate a curve sample set based on historical tightening curves of different bolt stations; wherein each of the historical tightening curves corresponds to a quality category; Determining the accuracy of the detection model corresponding to the curve sample set; wherein the accuracy represents the degree to which the detection model correctly identifies the quality category corresponding to the tightening curve; The parameters of the detection model are adjusted by the objective function constructed through the accuracy, so that the accuracy of the detection model in identifying the quality category corresponding to the tightening curve meets the preset target accuracy.

2. The training method according to claim 1, characterized in that Determining the accuracy of the detection model corresponding to the curve sample set includes: Outputting a plurality of basic tightening curves and tightening curves to be tested according to the curve sample set; Inputting the plurality of basic tightening curves and the tightening curve to be detected into the detection model to obtain curve features corresponding to the plurality of basic tightening curves and the tightening curve to be detected; determining the similarity between the tightening curve to be detected and each of the basic tightening curves according to the curve characteristics of the multiple basic tightening curves and the curve characteristics of the tightening curve to be detected; Determining, based on the similarity between the tightening curve to be detected and each of the basic tightening curves, a probability that the quality category of the tightening curve to be detected is the quality category corresponding to each of the basic tightening curves; The accuracy of the detection model is determined according to the probability that the quality category of the tightening curve to be detected is the quality category corresponding to each of the basic tightening curves and the quality category corresponding to the tightening curve to be detected.

3. The training method according to claim 1, characterized in that The objective function constructed by the accuracy rate is used to adjust the parameters of the detection model, comprising the following steps: Step S1, calculating the current gradient direction based on the objective function; Step S2, adjusting the parameters of the detection model according to the current gradient direction and the preset learning rate; Step S3, looping through steps S1 and S2 until the training stop condition is met.

4. A method for detecting the quality of bolt tightening, characterized in that: The method comprises: Obtain the tightening curve to be tested generated by the target bolt station; The target quality category corresponding to the tightening curve to be detected is determined by a detection model based on a preset basic tightening curve and the tightening curve to be detected, wherein the preset basic tightening curve corresponds to a quality category, and the detection model is obtained by using the training method of the detection model described in any one of claims 1-3.

5. The detection method according to claim 4, characterized in that Before determining the target quality category corresponding to the tightening curve to be detected based on the historical tightening curve and the tightening curve to be detected by using a detection model, the detection method includes: Determining a test tightening curve from the historical tightening records of the target bolt station; wherein the test tightening curve corresponds to a quality category; Inputting the preset basic tightening curve and the test tightening curve into the detection model to determine the accuracy of the detection model; The parameters of the detection model are adjusted by the objective function constructed through the accuracy, so as to maximize the accuracy of the detection model when identifying the quality category corresponding to the tightening curve generated by the target bolt station.

6. The detection method according to claim 4, characterized in that The determining, by using a detection model based on a preset basic tightening curve and the tightening curve to be detected, a target quality category corresponding to the tightening curve to be detected includes: Inputting the preset basic tightening curve and the tightening curve to be detected into the detection model respectively to obtain curve features corresponding to the preset basic tightening curve and the tightening curve to be detected; determining a similarity between the tightening curve to be detected and the preset basic tightening curve based on a curve feature of the preset basic tightening curve and a curve feature of the tightening curve to be detected; If the similarity reaches a preset similarity threshold, the quality category of the preset basic tightening curve is used as the target quality category.

7. The detection method according to claim 4, characterized in that The preset basic tightening curves include a plurality of them, each of which corresponds to a different quality type; The determining, by using a detection model based on the preset basic tightening curve and the tightening curve to be detected, a target quality category corresponding to the tightening curve to be detected includes: Inputting the plurality of preset basic tightening curves and the tightening curves to be detected into the target detection model respectively to obtain curve features corresponding to each of the preset basic tightening curves and the tightening curves to be detected; determining a similarity between the tightening curve to be detected and each of the preset basic tightening curves based on a curve feature of each of the preset basic tightening curves and a curve feature of the tightening curve to be detected; The quality category corresponding to the preset basic tightening curve having the highest similarity between the tightening curve to be detected and each of the preset basic tightening curves is used as the target quality category.

8. A training device for a detection model, characterized in that: include: A sample acquisition module is configured to generate a curve sample set based on historical tightening curves of different bolt stations; wherein each of the historical tightening curves corresponds to a quality category; A sample identification module is configured to determine the accuracy of the detection model corresponding to the curve sample set; wherein the accuracy represents the degree of accuracy of the detection model in identifying the quality category corresponding to the tightening curve; The parameter adjustment module is configured to adjust the parameters of the detection model through the objective function constructed by the accuracy rate, so that the accuracy rate of the detection model in identifying the quality category corresponding to the tightening curve meets the preset target accuracy rate.

9. A device for detecting the quality of bolt tightening, characterized in that: include: A sample acquisition module to be tested is configured to obtain a tightening curve to be tested generated by a target bolt station; An identification processing module is configured to determine, through a detection model, a target quality category corresponding to the tightening curve to be detected based on a preset basic tightening curve and the tightening curve to be detected, wherein the preset basic tightening curve corresponds to a quality category, and the detection model is obtained by using the training method of the detection model described in any one of claims 1 to 3.

10. A computer-readable storage medium, characterized in that Computer-readable instructions are stored thereon, and when the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the method for training a detection model described in any one of claims 1 to 3, or the method for detecting the bolt tightening quality described in any one of claims 4 to 7.

Images