Bolt residual torque intelligent detection method and system

Abstract

The invention relates to the related technical field of bolt detection, in particular to an intelligent detection method and system for residual torque of a bolt, and the method comprises the steps: positioning the head of the bolt, introducing model information to make a map, dividing detection partitions, determining a detection sequence according to a request, associating a wrench driving head, inserting a switching node, and carrying out segmented synchronous detection. The technical problems that the importance degree difference of bolts at different positions in equipment operation is difficult to distinguish, the looseness of the bolts at key positions cannot be detected in time, the switching of driving heads is frequent, and the detection efficiency of the residual torque of the bolts is low are solved, and the detection efficiency of the residual torque of the bolts is low through bolt distribution map partitioning, detection priority evaluation and driving head switching path optimization. The bolt distribution map is associated and matched with the wrench driving head, the driving head switching node is inserted into the detection sequence, the detection segments are divided according to the type of the driving head, the invalid switching of the driving head is reduced, the automatic planning of the detection sequence is realized, the detection coverage rate of the bolt at the core part is ensured, and the detection efficiency of the residual torque of the bolt is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of bolt detection, and particularly to an intelligent detection method and system for bolt residual torque. Background Art

[0002] With the rapid development of industrial automation and intelligent manufacturing, the requirements for the detection of bolt connection quality in industries such as mechanical equipment, bridge construction, and aerospace are becoming increasingly stringent. Accurately detecting the bolt residual torque can effectively prevent equipment failures and safety accidents caused by bolt loosening, and has become an important link to ensure the reliable operation of complex systems.

[0003] However, the existing automated detection methods lack systematic planning and often ignore the differences in bolt models, installation positions, and equipment operating conditions, resulting in chaotic detection paths and frequent switching of driving heads, causing waste of resources and low detection efficiency. In addition, most detection means can only measure the torque value once, and cannot dynamically adjust the detection strategy in combination with the bolt installation environment, making it difficult to achieve accurate evaluation and abnormal warning of the residual torque and unable to adapt to the detection tasks under complex working conditions.

[0004] In summary, there are technical problems in the prior art that it is difficult to distinguish the importance differences of bolts at different positions during equipment operation, the loosening of bolts at key positions cannot be detected in time, the driving head switches frequently, and the detection efficiency of bolt residual torque is low. Summary of the Invention

[0005] The present application provides an intelligent detection method and system for bolt residual torque, aiming to solve the technical problems in the prior art that it is difficult to distinguish the importance differences of bolts at different positions during equipment operation, the loosening of bolts at key positions cannot be detected in time, the driving head switches frequently, and the detection efficiency of bolt residual torque is low.

[0006] In view of the above problems, the technical solution of the present application is as follows: On the one hand, the present application provides an intelligent detection method for bolt residual torque, wherein the method includes: In the area to be measured, locate the bolt head, introduce the bolt model information, and draw up a bolt distribution map with model markings; divide the bolt distribution map into multiple detection priority zones; receive a residual torque detection request; according to the multiple detection priority zones, compare with the residual torque detection request to determine the residual torque detection sequence; connect the residual torque detection wrench, and perform associated matching through the bolt distribution map and the wrench drive head, and insert a drive head switching node in the residual torque detection sequence, where the wrench drive head of the residual torque detection wrench can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick-change head; through the drive head switching node, determine the ratchet square drive and multiple first residual torque detection segments, the open head and multiple second residual torque detection segments, the internal hexagon and multiple third residual torque detection segments, the quick-change head and multiple fourth residual torque detection segments, and synchronize them to the residual torque detection wrench for automated segmented detection.

[0007] Preferably, define a key connection area, a secondary connection area, and an auxiliary fixing area through the functional importance of the area to be measured under equipment association; locate the high-vibration zones according to the vibration frequency distribution during equipment operation; perform an overlapping comparison of the key connection area, the secondary connection area, and the auxiliary fixing area with the high-vibration zones, and evaluate the priority score to determine the multiple detection priority zones.

[0008] Preferably, identify the geometric features of the bolt head, establish a mapping database between the bolt model information and the wrench drive head, and the mapping database is used to store the types of wrench drive heads adapted to different bolt models; at the same time, verify the drive head compatibility through the contact pressure distribution characteristics feedback by the torque sensor.

[0009] Preferably, perform spatial position clustering on the multiple first residual torque detection segments to determine the initial population; use the drive head switching time and the total path length under the drive head switching node as penalty factors, and perform local search and enhancement in the initial population to determine the first preferred detection path set corresponding to the ratchet square drive; through the first preferred detection path set corresponding to the ratchet square drive, the second preferred detection path set corresponding to the open head, the third preferred detection path set corresponding to the internal hexagon, and the fourth preferred detection path set corresponding to the quick-change head, determine the drive head switching order.

[0010] Preferably, set the objective function through the detection accuracy loss rate; select the solutions with objective function values lower than the accuracy loss threshold from the initial population as the target solutions, and the target solutions are used to execute the update decision of the multiple first residual torque detection segments.

[0011] Preferably, task decomposition is performed according to the multiple first residual torque detection segments to generate an operation sequence including a pre-tightening force measurement subtask, a dynamic torque monitoring subtask, and an angle compensation subtask; the operation sequence is used to schedule detection resources, and at the same time, a real-time feedback mechanism between the operation sequence and the state of the residual torque detection wrench is established.

[0012] Preferably, the pre-tightening force measurement subtask in the operation sequence adopts a first sampling period to synchronously trigger the sliding window length evaluation of the dynamic torque monitoring subtask; when the angle compensation subtask detects a sudden change in angular acceleration, an event-driven priority hybrid promotion strategy is activated, and the priority hybrid promotion strategy is used to compensate the first sampling period to be promoted to a second sampling period.

[0013] Preferably, axial strain data is collected at the root of the bolt thread; a long short-term memory network is used to extract the temporal features of the axial strain data to obtain the predicted value of the bolt residual torque under the operation sequence; the deviation value between the predicted value of the bolt residual torque and the wrench measurement value is compared with the deviation threshold for error comparison to determine whether to trigger the real-time feedback mechanism.

[0014] On the other hand, the present application provides an intelligent bolt residual torque detection system, wherein the system includes: A bolt positioning module for positioning the bolt head in the area to be measured, introducing bolt model information, and drawing up a bolt distribution map with model markings; a request receiving module for dividing the bolt distribution map into multiple detection priority zones; receiving a residual torque detection request; and determining a residual torque detection sequence according to the multiple detection priority zones in comparison with the residual torque detection request; an association matching module for connecting a residual torque detection wrench and performing association matching between the bolt distribution map and the wrench drive head, and inserting a drive head switching node into the residual torque detection sequence, wherein the wrench drive head of the residual torque detection wrench can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick change head; a segmented detection module for determining, through the drive head switching node, a ratchet square drive and multiple first residual torque detection segments, an open head and multiple second residual torque detection segments, an internal hexagon and multiple third residual torque detection segments, and a quick change head and multiple fourth residual torque detection segments, and synchronizing them to the residual torque detection wrench for automated segmented detection.

[0015] In summary, in one or more technical solutions provided in this application, through bolt distribution map zoning, detection priority assessment, and drive head switching path optimization, by associating and matching the bolt distribution map with the wrench drive head, drive head switching nodes are inserted into the detection sequence, and the detection segments are divided according to the drive head type, reducing ineffective drive head switching, realizing the automated planning of the detection sequence, ensuring the detection coverage rate of bolts in the core parts, and improving the technical effect of the bolt residual torque detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic flowchart of a method for intelligent detection of bolt residual torque provided in this application; Figure 2 is a schematic structural diagram of a system for intelligent detection of bolt residual torque provided in this application.

[0017] Description of reference numerals: bolt positioning module M100, request receiving module M200, association and matching module M300, segmented detection module M400. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] Embodiment 1 The following specifically describes this application with reference to the drawings. As Figure 1 shown, this application provides a method for intelligent detection of bolt residual torque, where the method includes: S1: In the area to be measured, locate the bolt head, introduce bolt model information, and draw up a bolt distribution map with model markings.

[0019] Specifically, using machine vision technology, the image data of the bolts in the area to be measured is obtained through an image acquisition device, and the specific position coordinates of the bolt heads are identified through an image processing algorithm. At the same time, combined with bolt model recognition technology, the model parameters (such as specifications, dimensions, etc.) of each bolt are obtained; these information are integrated to generate a distribution map marked with the model and position of each bolt. For example, in the area to be measured of a mechanical equipment, images are collected through an industrial camera, the positions of 10 bolts are identified, and the model of each bolt is determined to be M12, M16, etc., and then a two-dimensional or three-dimensional distribution map containing this information is drawn.

[0020] In a feasible implementation manner, first, the bolts are positioned using a machine vision system, and its positioning accuracy can reach the millimeter level; then, through a bolt model recognition algorithm (such as an object detection model based on deep learning), the bolt model information is accurately obtained, and then these data are integrated into a bolt distribution map, providing a basis for subsequent detection planning, ensuring the accurate identification and positioning of the detection object, enabling the subsequent detection process to be carried out targeted, and improving the detection efficiency.

[0021] S2: Divide the bolt distribution atlas into multiple detection priority partitions; receive a residual torque detection request; and determine a residual torque detection sequence according to the multiple detection priority partitions and in comparison with the residual torque detection request.

[0022] Specifically, dividing the bolt distribution atlas into multiple detection priority partitions means dividing the generated bolt distribution atlas into different regions according to certain rules, and each region has a different detection priority.

[0023] In a feasible implementation manner, divide the area to be measured into a key connection area, a secondary connection area, and an auxiliary fixing area according to the functional importance under equipment association. For example, for the wing structure of an aircraft, the key connection area is the connection part that bears the main load, the secondary connection area is the auxiliary connection part, and the auxiliary fixing area is the connection part for fixing some secondary components. At the same time, locate the high-vibration partition according to the vibration frequency distribution during equipment operation; evaluate the priority scores of each region by overlapping and comparing these regions with the high-vibration partition, so as to determine multiple detection priority partitions; in the whole solution, this step can ensure that the bolts in the key parts are detected first, effectively improving the detection efficiency and resource utilization rate, and reducing unnecessary detection workload.

[0024] S3: Connect the residual torque detection wrench, perform associated matching between the bolt distribution atlas and the wrench drive head, and insert a drive head switching node into the residual torque detection sequence, where the wrench drive head of the residual torque detection wrench can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick-change head.

[0025] Specifically, connect the residual torque detection wrench to the bolt distribution atlas, and match the wrench drive head according to the bolt model information in the bolt distribution atlas; the drive head switching node refers to the position where the wrench drive head needs to be switched according to the different bolt models during the detection process. For example, during the detection process, when the bolt model to be detected changes from M12 to M16, the wrench drive head needs to be switched from the drive head adapted to M12 to the drive head adapted to M16, and this switching position is the drive head switching node; for the wrench drive head of the residual torque detection wrench, it can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick-change head, which means that there are multiple types of wrench drive heads available for selection, including a ratchet square drive, an open head, an internal hexagon, and a quick-change head. Different drive heads are suitable for different models of bolts. For example, a ratchet square drive is suitable for larger-sized bolts, an open head is suitable for hexagon head bolts, an internal hexagon is suitable for internal hexagon bolts, and a quick-change head is convenient for quick switching.

[0026] In a feasible implementation, first, connect the residual torque detection wrench to the bolt distribution map. Through the bolt model information in the bolt distribution map, retrieve the appropriate wrench drive head type from the mapping database. For example, when detecting an area containing M12 and M16 bolts, according to the bolt distribution map, it will be recognized that two drive heads, namely an open-end head and an Allen head, are required. Then, insert a drive head switching node into the residual torque detection sequence. When the detection sequence switches from M12 bolts to M16 bolts, the wrench will automatically switch between the open-end head and the Allen head. By intelligently matching the drive head, the detection efficiency is improved.

[0027] S4: Through the drive head switching node, determine the ratchet square drive and multiple first residual torque detection segments, the open-end head and multiple second residual torque detection segments, the Allen head and multiple third residual torque detection segments, and the quick-change head and multiple fourth residual torque detection segments, and synchronize them to the residual torque detection wrench for automated segmented detection.

[0028] Specifically, according to the drive head switching node, divide the entire detection process into multiple segments. Each segment corresponds to a drive head type and its corresponding residual torque detection task. For example, the first residual torque detection segment using the ratchet square drive mainly detects large bolts; the second residual torque detection segment using the open-end head mainly detects hex head bolts, etc.

[0029] In a feasible implementation, before the detection starts, according to the bolt distribution map and the drive head switching node, pre-plan the detection path and divide the detection task into four segments, corresponding to four drive head types respectively. Specifically, in a detection task of a mechanical equipment, the system divides the detection task into a ratchet square drive segment (detecting large bolts), an open-end head segment (detecting hex head bolts), an Allen head segment (detecting Allen bolts), and a quick-change head segment (for quickly switching to detect a small number of bolts of different specifications); then, synchronize these segments to the residual torque detection wrench, and the wrench automatically executes the detection task according to the segment information, realizing the modular management and automated execution of the detection process, improving the detection accuracy and efficiency, and enhancing the maintainability and scalability of the system in a segmented detection manner.

[0030] Furthermore, divide the bolt distribution map into multiple detection priority partitions. The method of the present application includes: Define a key connection area, a secondary connection area, and an auxiliary fixing area through the functional importance of the area to be measured under equipment association; locate the high-vibration partition according to the vibration frequency distribution during equipment operation; perform an overlapping comparison of the key connection area, the secondary connection area, and the auxiliary fixing area with the high-vibration partition, and evaluate the priority score to determine the multiple detection priority partitions.

[0031] Specifically, the functional importance under equipment association refers to the critical degree of the functions undertaken by different areas of the equipment in the overall equipment operation. For example, in the aerospace field, the connection area between the aircraft wing and the fuselage is crucial for flight safety, and its functional importance is very high; the key connection area is the area in the equipment that plays a major connection role and has the greatest impact on the operation stability of the equipment. Taking a large bridge as an example, the connection part between the pier and the bridge deck is the key connection area; the secondary connection area is the connection part that has a certain impact on the equipment operation but is not the core stress-bearing or key functional area. For example, in mechanical equipment, the connection area between some auxiliary components and the main equipment; the auxiliary fixing area is mainly used to assist in fixing equipment components and has a relatively small impact on the main functions and operation stability of the equipment. For example, some fixing screw areas on the equipment shell; the high vibration partition refers to the area with a relatively high vibration frequency during equipment operation. Around the engine, due to high-frequency vibrations generated during operation, it belongs to the high vibration partition.

[0032] In a feasible implementation manner, according to the functional importance under equipment association, combining data such as the design drawings and operating conditions of the equipment, the area to be measured is divided into a key connection area, a secondary connection area, and an auxiliary fixing area. Vibration sensors are used to collect the vibration frequency data of each area during equipment operation. Through signal processing and analysis, the high vibration partition is located; the key connection area, the secondary connection area, and the auxiliary fixing area are compared with the high vibration partition for overlap. For example, if it is found that a certain key connection area overlaps partially with the high vibration partition, it means that the bolts in this area need to bear both the stress of key connection and the influence of high-frequency vibrations, and the risk of loosening is relatively high.

[0033] Then, according to factors such as the overlap situation and functional importance of each area, the priority scores are evaluated to determine multiple detection priority partitions. Generally speaking, the area that is both in the key connection area and in the high vibration partition will obtain a higher priority score and be divided into the partition to be detected first. Preferably, this division method can ensure that the bolts at key parts are detected first, effectively improving the detection efficiency.

[0034] Furthermore, for the connection residual torque detection wrench, by associating and matching the bolt distribution map with the wrench drive head, drive head switching nodes are inserted into the residual torque detection sequence. The method of this application includes: Identifying the geometric features of the bolt head, establishing a mapping database between the bolt model information and the wrench drive head. The mapping database is used to store the types of compatible wrench drive heads under different bolt models; at the same time, the compatibility of the drive head is verified through the contact pressure distribution characteristics fed back by the torque sensor.

[0035] Specifically, the geometric features of the bolt head refer to geometric parameters such as the shape, size, and contour of the bolt head. For example, the head of a hexagon head bolt has a hexagonal contour, and its dimensions include the across flats distance, height, etc.; the mapping database is a data structure used to store and manage the correspondence between different bolt models and the types of wrench drive heads that fit them. For example, for a hexagon head bolt with M10×1.5mm, the mapping database will record that its corresponding wrench drive head is an open-end head; the contact pressure distribution feature refers to the spatial distribution of the pressure on the surface of the torque sensor when the torque sensor contacts the bolt. For example, when the wrench drive head contacts the bolt head and applies torque, the pressure distribution in the contact area will show a certain pattern, and different bolt models and drive head types will have different contact pressure distribution features.

[0036] In a feasible implementation manner, first, use image recognition technology or laser scanning technology to identify the geometric features of the bolt head. For example, take an image of the bolt head with an industrial camera, and then use an image processing algorithm to identify that the shape of the bolt head is hexagonal; then, according to the identified bolt model information, look up the corresponding wrench drive head type in the mapping database.

[0037] At the same time, verify the compatibility between the drive head and the bolt through the contact pressure distribution feature feedback by the torque sensor. Specifically, when the open-end drive head contacts the M12×1.75mm hexagon head bolt and applies a certain torque, the torque sensor will detect whether the contact pressure distribution is uniform; if the pressure distribution is uniform, it indicates that the drive head is compatible with the bolt; if the pressure distribution is not uniform, there may be a problem that the shape of the drive head does not match the bolt head. The accuracy rate of compatibility verification through the contact pressure distribution feature ensures that the selected wrench drive head matches the bolt, improves the accuracy and reliability of detection, and reduces the detection error and equipment damage risk caused by the mismatch of the drive head.

[0038] Furthermore, through the drive head switching node, determine the ratchet square drive and multiple first residual torque detection segments. The method of the present application further includes: Perform spatial position clustering on the multiple first residual torque detection segments to determine the initial population; use the drive head switching time and the total path length under the drive head switching node as penalty factors, and perform local search and reinforcement in the initial population to determine the first preferred detection path set corresponding to the ratchet square drive; determine the drive head switching order through the first preferred detection path set corresponding to the ratchet square drive, the second preferred detection path set corresponding to the open-end head, the third preferred detection path set corresponding to the hexagon socket head, and the fourth preferred detection path set corresponding to the quick-change head.

[0039] Specifically, spatial position clustering refers to dividing adjacent detection segments into the same cluster according to the positional relationship of multiple first residual torque detection segments in space. For example, within a region to be measured of a mechanical structure, if multiple bolts are in the same area and have similar spatial positions, the detection segments corresponding to these bolts can be divided into the same cluster through spatial position clustering; the initial population means that each cluster obtained through spatial position clustering can be regarded as an individual in the initial population, and these individuals contain the characteristic information of different clusters, providing a basis for subsequent operations such as path optimization.

[0040] In a feasible implementation, the penalty factor is used to penalize certain factors that do not meet the constraint conditions or need to be optimized, so that they are restricted during the optimization process, thereby guiding the algorithm to find a better solution. Taking the drive head switching time and the total path length as the penalty factors is because these two factors directly affect the detection efficiency. For example, a longer drive head switching time and total path length will lead to an increase in detection time and a decrease in efficiency; during the optimization process, penalizing these two factors can prompt the algorithm to find a detection path with a short switching time and a small total path length.

[0041] Local search enhancement is to deeply search the local area of the solution space based on the initial population to find a better solution. During this process, the penalty factor is used to penalize the solutions that do not meet the requirements, while retaining and optimizing the better solutions, so as to gradually find a better detection path within the local area. For example, for each individual (i.e., cluster) in the initial population, with the current cluster as the center, search for a better detection path within a certain range, while considering the constraint of the penalty factor on the path.

[0042] The first preferred detection path set is a set of the optimal detection paths corresponding to the ratchet square drive obtained after local search enhancement. These paths are obtained through the screening and optimization of the optimization algorithm under the condition of meeting the penalty factor constraint, and have a short drive head switching time and a small total path length, which can improve the detection efficiency; the preferred detection path set means that for different types of drive heads (ratchet square drive, open head, hexagon socket head, quick-change head), the respective preferred detection path sets are obtained according to their corresponding detection segments and penalty factor constraints. The paths in these sets are the optimal paths processed by the optimization algorithm and can ensure a high detection efficiency when using the corresponding drive head.

[0043] The driving head switching order is determined according to factors such as the path characteristics in each preferred detection path set and the priority of the detection task. For example, in the detection task, if the detection task corresponding to the ratchet square drive is relatively urgent and the priority of the detection path is high, then it can be switched to the ratchet square drive for detection first, and then switched to the open head, internal hexagon or quick-change head for detection as needed; by reasonably determining the driving head switching order, the detection process can be further optimized, unnecessary switching operations can be reduced, and the efficient progress of the detection work can be ensured.

[0044] Furthermore, taking the driving head switching time and the total path length under the driving head switching node as penalty factors, and locally searching and strengthening in the initial population, the method of the present application further includes: Setting an objective function through the detection accuracy loss rate; selecting, from the initial population, solutions with objective function values lower than the accuracy loss threshold as target solutions, and the target solutions are used to execute the update decision of the multiple first residual torque detection segments.

[0045] Specifically, the detection accuracy loss rate refers to the degree of deviation between the detection result and the true value during the detection process due to various factors (such as detection methods, equipment accuracy, environmental interference, etc.). For example, in the detection of bolt residual torque, if the accuracy of the detection equipment is limited or the detection method is not optimized enough, there may be a difference between the detected torque value and the actual residual torque value of the bolt, and this difference degree is the detection accuracy loss rate; the objective function is set according to the detection accuracy loss rate and is used to quantify the influence of factors such as the detection path and driving head switching on the detection accuracy. For example, the objective function can be defined as the sum of the weights of the detection accuracy loss rate, driving head switching time, total path length, etc. The purpose is to minimize the value of this function during the optimization process to find the best detection scheme.

[0046] The initial population can be different detection path sets obtained by spatial position clustering. These path sets contain various possible detection path combinations and provide a basis for subsequent optimization operations; the target solutions are solutions that meet certain accuracy requirements screened from the initial population through an optimization algorithm. The accuracy requirements are limited by the accuracy loss threshold. The target solutions are those solutions that can minimize the objective function value while meeting the accuracy requirements and are used to execute the update decision of the multiple first residual torque detection segments, that is, to determine the optimal detection path and driving head switching strategy, etc.

[0047] In a feasible implementation, by analyzing various error sources during the detection process, such as sensor accuracy, limitations of detection methods, etc., a calculation method for the detection accuracy loss rate is determined. For example, through multiple experiments and data statistics, it is found that the detection accuracy loss rate is mainly related to factors such as the measurement error of the sensor, the adaptability between the driving head and the bolt, and a corresponding mathematical model is established to quantify the influence of these factors on the accuracy loss rate; then, according to the detection accuracy loss rate, an objective function is set. Preferably, the objective function can be expressed as a weighted fusion of the detection accuracy loss rate, the driving head switching time, and the total path length according to their importance. From the initial population, solutions with objective function values lower than the accuracy loss threshold are selected as target solutions. For example, in the initial population, there are 100 different detection paths, and the objective function values corresponding to each path are calculated respectively. If the accuracy loss threshold is set to 5%, then the detection paths with objective function values lower than this threshold are selected as target solutions. After screening, eligible target solutions will be obtained, and these target solutions represent relatively optimized detection paths under the condition of meeting the accuracy requirements.

[0048] These target solutions are used to execute update decisions for multiple first residual torque detection segments. For example, according to the detection path information in the target solutions, the detection sequence and driving head switching strategy of the detection device are updated, so that during the actual detection process, the detection is carried out according to the optimized path and strategy, improving the detection efficiency and accuracy. In this way, it can be ensured that the detection work is carried out in a more efficient manner on the premise of meeting the accuracy requirements, reducing unnecessary detection operations and time waste.

[0049] Furthermore, the method of the present application includes: Task decomposition is performed according to the multiple first residual torque detection segments to generate an operation sequence including a pre-tightening force measurement sub-task, a dynamic torque monitoring sub-task, and an angle compensation sub-task; the operation sequence is used to schedule detection resources, and at the same time, a real-time feedback mechanism between the operation sequence and the state of the residual torque detection wrench is established.

[0050] Specifically, the residual torque detection segmentation refers to dividing the entire bolt residual torque detection task into multiple small parts or stages, with each part corresponding to a certain number of bolt detection tasks. For example, the detection task of a large mechanical equipment may be divided into multiple detection segments, and each segment contains a certain number of bolts to facilitate the management and optimization of the detection process; task decomposition refers to the process of splitting a complex task or a large task into several relatively simple and specific subtasks, that is, further decomposing each residual torque detection segment into three specific subtasks: pre-tightening force measurement, dynamic torque monitoring, and angle compensation, to facilitate separate processing and optimization; the operation sequence refers to arranging each subtask in a certain logical order to form a complete detection operation process for guiding the execution of the detection equipment; the real-time feedback mechanism refers to establishing real-time data interaction between the execution process of the operation sequence and the state of the residual torque detection wrench to ensure the dynamic optimization and efficient execution of the detection process.

[0051] In a feasible implementation manner, task decomposition is performed on multiple first residual torque detection segments. For example, assuming that a detection segment contains the detection task of 10 bolts, then this segment can be decomposed into 10 pre-tightening force measurement subtasks, 10 dynamic torque monitoring subtasks, and 10 angle compensation subtasks. Each subtask performs corresponding detection operations for a specific bolt. This decomposition method makes the detection task more detailed and clear, which helps with subsequent precise scheduling and execution.

[0052] According to the results of the task decomposition, an operation sequence including pre-tightening force measurement subtasks, dynamic torque monitoring subtasks, and angle compensation subtasks is generated. For example, these subtasks can be arranged into an ordered operation sequence according to the position order of the bolts or the detection priority. For example, first perform pre-tightening force measurement on the first bolt, then dynamic torque monitoring, and finally angle compensation; then repeat the same operation for the second bolt, and so on. Such an operation sequence can ensure the systematicness and integrity of the detection work and avoid missing any key detection steps.

[0053] Use this operation sequence to schedule detection resources. For example, detection resources include detection wrenches, sensors, data acquisition devices, etc. According to the requirements of each subtask in the operation sequence, these resources are reasonably allocated. For example, when performing the pre-tightening force measurement subtask, schedule the corresponding force sensor and data acquisition device; when performing the dynamic torque monitoring subtask, schedule the torque sensor and signal processing unit, etc.; in this way, the efficient utilization of detection resources can be ensured, and the idle and waste of resources can be avoided.

[0054] Meanwhile, a real-time feedback mechanism for the operation sequence and the status of the residual torque detection wrench is established. Specifically, during the detection process, the detection wrench will provide real-time feedback on its working status information, such as the current operation being performed, the measured torque value, the device operation status, etc. These feedback messages will be used to dynamically adjust the execution of the operation sequence. If an abnormal torque value of a certain bolt is detected, the subsequent operation sequence can be adjusted in a timely manner according to the feedback information, increasing the re-inspection times for that bolt or adjusting the detection strategy. This real-time feedback mechanism can significantly improve the flexibility and adaptability of the detection, ensuring the efficiency and accuracy of the detection work.

[0055] Furthermore, the method of the present application includes: The pre-tightening force measurement subtask in the operation sequence adopts a first sampling period to synchronously trigger the sliding window length evaluation of the dynamic torque monitoring subtask; when the angular acceleration mutation is detected by the angle compensation subtask, an event-driven priority hybrid promotion strategy is activated, and the priority hybrid promotion strategy is used to compensate the first sampling period to be promoted to a second sampling period.

[0056] Specifically, the first sampling period refers to the time interval used to collect pre-tightening force data in the pre-tightening force measurement subtask. For example, setting the first sampling period to 0.1 second means collecting pre-tightening force data every 0.1 second; the sliding window length evaluation refers to the process of evaluating and determining the sliding window length in the dynamic torque monitoring subtask. The sliding window length refers to the time window size used to analyze torque data in the dynamic torque monitoring. The appropriate window length is determined through evaluation to better capture the torque change characteristics; the angular acceleration mutation refers to the situation where the angular acceleration of the bolt suddenly changes in the angle compensation subtask, usually indicating that the force state or motion state of the bolt has changed significantly, which may affect the accuracy of the detection; the event-driven priority hybrid promotion strategy refers to when a specific event (such as angular acceleration mutation) is detected, the detection process is optimized by promoting the priority of related tasks. For example, when an angular acceleration mutation is detected, this strategy will increase the priority of the pre-tightening force measurement subtask to collect data more frequently, so as to capture the changes more accurately.

[0057] In a feasible implementation, first, the pre-tightening force measurement subtask collects data using the first sampling period. For example, the first sampling period is set to 0.1 second, and the detection device collects the pre-tightening force data of the bolt every 0.1 second to ensure that the change of the pre-tightening force can be obtained in a timely manner. At the same time, the evaluation of the sliding window length of the dynamic torque monitoring subtask is carried out synchronously. For example, according to the change characteristics of the pre-tightening force data and the requirements of dynamic torque monitoring, the sliding window length is evaluated and determined to be 1 second, that is, the torque data within 1 second is analyzed each time to capture the short-term change trend of the torque. In this way, the execution of the two subtasks of pre-tightening force measurement and dynamic torque monitoring can be coordinated, and the detection efficiency can be improved.

[0058] When the angle compensation subtask detects a sudden change in angular acceleration, for example, it is detected by an angular velocity sensor that the angular acceleration of the bolt increases by more than 50% in a short time, which indicates that the stress state of the bolt has changed significantly and may affect the accuracy of the pre-tightening force. At this time, the event-driven priority hybrid promotion strategy is activated, and the priority of the pre-tightening force measurement subtask is promoted from the original medium priority to the high priority. At the same time, the first sampling period is increased from 0.1 second to the second sampling period of 0.05 second, so that the pre-tightening force data can be collected more frequently, and the change of the pre-tightening force caused by the sudden change in angular acceleration can be captured in a timely manner, thereby improving the detection accuracy.

[0059] Furthermore, a real-time feedback mechanism between the operation sequence and the state of the residual torque detection wrench is established. The method of this application further includes: At the root of the bolt thread, axial strain data is collected; a long short-term memory network is used to extract the temporal characteristics of the axial strain data to obtain the predicted value of the bolt residual torque under the operation sequence; the deviation value between the predicted value of the bolt residual torque and the wrench measurement value is compared with the deviation threshold to determine whether to trigger the real-time feedback mechanism.

[0060] Specifically, the axial strain data refers to the data of the deformation degree of the bolt in the axial direction (i.e., the length direction of the bolt). When the bolt is subjected to tensile or compressive force, it will produce elongation or shortening deformation in the axial direction. The deformation degree can be measured by sensors such as strain gauges to obtain the axial strain data; the long short-term memory network (LSTM) is a special type of recurrent neural network (RNN) that can learn long-term dependencies and performs excellently in the processing of time series data. It is suitable for analyzing data with a time sequence such as axial strain data. For example, for the axial strain data sequence of the bolt over a period of time, the LSTM network can capture the long-term change trend and short-term fluctuation characteristics in the data.

[0061] Temporal feature extraction refers to extracting information that can represent the characteristics of data from time-series data. For the temporal feature extraction of axial strain data, it is to analyze the strain conditions of the bolt at different time points and find out the characteristics such as change rules, trends, and periodicity. For example, by analyzing the axial strain data, characteristics such as the strain fluctuation frequency and amplitude change of the bolt during operation can be extracted; the deviation value refers to the degree of difference between the predicted value of the bolt residual torque and the measured value by the wrench. For example, if the predicted value is 50 N·m and the measured value by the wrench is 45 N·m, then the deviation value is 5 N·m, which reflects the gap between the predicted value and the actual measured value; the deviation threshold is a preset judgment criterion used to determine whether the deviation value is within an acceptable range. For example, if the deviation threshold is set to 10 N·m, when the deviation value exceeds this value, it is considered that the difference between the predicted value and the actual measured value is too large, and a real-time feedback mechanism needs to be triggered for adjustment.

[0062] In a feasible implementation manner, a strain sensor is installed at the root of the bolt thread to collect axial strain data. For example, a high-precision strain gauge is selected and pasted at the root of the bolt thread, and the strain data is obtained through a data acquisition device at a certain sampling frequency (such as 100 times per second). These data can reflect the axial force condition of the bolt during operation and provide a basis for subsequent analysis; a long short-term memory network is used to perform temporal feature extraction on the collected axial strain data. For example, the collected strain data is input into a pre-trained LSTM network, and the network will extract feature information such as the strain fluctuation frequency and amplitude change of the bolt according to the time-series characteristics of the data. Then, through the established model, this feature information is converted into a predicted value of the bolt residual torque. For example, according to the trained model, the residual torque of the current bolt is predicted to be 50 N·m, and the residual torque of the bolt is predicted in real time to provide a reference for detection.

[0063] The predicted value of the bolt residual torque is compared with the measured value by the wrench, the deviation value is calculated, and compared with the deviation threshold. For example, the measured value by the wrench is 45 N·m, the deviation value is 5 N·m, and the preset deviation threshold is 10 N·m. Since the deviation value is less than the deviation threshold, the real-time feedback mechanism is not triggered. If the deviation value exceeds the deviation threshold, for example, the deviation value reaches 12 N·m, then the real-time feedback mechanism will be triggered, and the detection strategy will be adjusted according to the deviation situation, such as increasing the detection frequency, recalibrating the detection device, etc., to ensure the accuracy of detection. In this way, errors in the detection process can be discovered and corrected in time, and the reliability and accuracy of detection can be improved.

[0064] In summary, the beneficial effects of the embodiments of this application are: By adopting the method of positioning the bolt head in the area to be measured, introducing the bolt model information, and drawing up a bolt distribution map with model markings; dividing the bolt distribution map into multiple detection priority zones; receiving a residual torque detection request; determining a residual torque detection sequence according to the multiple detection priority zones and in comparison with the residual torque detection request; connecting a residual torque detection wrench, associating and matching it with the wrench driving head through the bolt distribution map, and inserting a driving head switching node into the residual torque detection sequence, wherein the wrench driving head of the residual torque detection wrench can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick-change head; and determining, through the driving head switching node, the ratchet square drive and multiple first residual torque detection segments, the open head and multiple second residual torque detection segments, the internal hexagon and multiple third residual torque detection segments, and the quick-change head and multiple fourth residual torque detection segments, and synchronizing them to the residual torque detection wrench for automated segmented detection. By providing a bolt residual torque intelligent detection method and system, through bolt distribution map zoning, detection priority evaluation, and driving head switching path optimization, using the association and matching between the bolt distribution map and the wrench driving head, inserting a driving head switching node into the detection sequence, and dividing the detection segments according to the driving head type, the invention reduces the ineffective switching of the driving head, realizes the automated planning of the detection sequence, ensures the detection coverage rate of the bolts in the core parts, and improves the technical effect of the bolt residual torque detection efficiency.

[0065] Embodiment 2 Based on the same inventive concept as a bolt residual torque intelligent detection method in the foregoing embodiment, as Figure 2 shown, the embodiment of the present application provides a bolt residual torque intelligent detection system, wherein the system includes: A bolt positioning module M100, configured to position the bolt head in the area to be measured, introduce the bolt model information, and draw up a bolt distribution map with model markings.

[0066] A request receiving module M200, configured to divide the bolt distribution map into multiple detection priority zones; receive a residual torque detection request; and determine a residual torque detection sequence according to the multiple detection priority zones and in comparison with the residual torque detection request.

[0067] An association and matching module M300, configured to connect a residual torque detection wrench, perform association and matching between the bolt distribution map and the wrench driving head, and insert a driving head switching node into the residual torque detection sequence, wherein the wrench driving head of the residual torque detection wrench can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick-change head.

[0068] The segmented detection module M400 is used to determine, through the driving head switching node, the ratchet square drive and multiple first residual torque detection segments, the open head and multiple second residual torque detection segments, the hexagon socket and multiple third residual torque detection segments, and the quick-change head and multiple fourth residual torque detection segments, and synchronize them to the residual torque detection wrench for automated segmented detection.

[0069] Further, the receiving request module M200 is used to execute the following method: Define the key connection area, the secondary connection area, and the auxiliary fixing area according to the functional importance of the area to be measured under equipment association; locate the high-vibration partition according to the vibration frequency distribution during equipment operation; perform an overlapping comparison of the key connection area, the secondary connection area, and the auxiliary fixing area with the high-vibration partition, and evaluate the priority score to determine the multiple detection priority partitions.

[0070] Further, the association matching module M300 is used to execute the following method: Identify the geometric features of the bolt head, establish a mapping database between the bolt model information and the wrench driving head, and the mapping database is used to store the compatible wrench driving head types under different bolt models; at the same time, verify the driving head compatibility through the contact pressure distribution characteristics fed back by the torque sensor.

[0071] Further, the segmented detection module M400 is used to execute the following method: Perform spatial position clustering on the multiple first residual torque detection segments to determine the initial population; use the driving head switching time and the total path length under the driving head switching node as penalty factors, and perform local search and reinforcement in the initial population to determine the first preferred detection path set corresponding to the ratchet square drive; determine the driving head switching order through the first preferred detection path set corresponding to the ratchet square drive, the second preferred detection path set corresponding to the open head, the third preferred detection path set corresponding to the hexagon socket, and the fourth preferred detection path set corresponding to the quick-change head.

[0072] Further, the segmented detection module M400 is also used to execute the following method: Set the objective function through the detection accuracy loss rate; select the solutions with objective function values lower than the accuracy loss threshold from the initial population as the target solutions, and the target solutions are used to execute the update decision of the multiple first residual torque detection segments.

[0073] Further, the segmented detection module M400 is also used to execute the following method: Perform task decomposition according to the multiple first residual torque detection segments, generate an operation sequence including a pre-tightening force measurement subtask, a dynamic torque monitoring subtask, and an angle compensation subtask; use the operation sequence to schedule detection resources, and at the same time, establish a real-time feedback mechanism between the operation sequence and the state of the residual torque detection wrench.

[0074] Further, the segmented detection module M400 is further configured to execute the following method: The pre-tightening force measurement subtask in the operation sequence adopts a first sampling period to synchronously trigger the sliding window length evaluation of the dynamic torque monitoring subtask; when the angle compensation subtask detects a sudden change in angular acceleration, activate an event-driven priority hybrid boosting strategy, and the priority hybrid boosting strategy is used to compensate the first sampling period to be boosted to a second sampling period.

[0075] Further, the segmented detection module M400 is further configured to execute the following method: At the root of the bolt thread, collect axial strain data; use a long short-term memory network to extract the time series characteristics of the axial strain data to obtain the predicted value of the bolt residual torque under the operation sequence; compare the deviation value between the predicted value of the bolt residual torque and the wrench measurement value with a deviation threshold to determine whether to trigger the real-time feedback mechanism.

[0076] In summary, any step can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor, and no redundant restrictions are made here.

[0077] Furthermore, the above technical solutions only reflect the preferred technical solutions of the technical solutions of the embodiments of the present application. Some changes that those skilled in the art may make to some parts thereof all reflect the principles of the novel embodiments of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. An intelligent detection method for bolt residual torque, characterized in that, The method includes: In the area to be measured, locate the bolt head, introduce bolt model information, and draw up a bolt distribution map with model markings; Divide the bolt distribution map into multiple detection priority zones; receive a residual torque detection request; according to the multiple detection priority zones, compare with the residual torque detection request to determine the residual torque detection sequence; Connect the residual torque detection wrench, perform associated matching between the bolt distribution map and the wrench drive head, and insert a drive head switching node in the residual torque detection sequence, where the wrench drive head of the residual torque detection wrench can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick-change head; Through the drive head switching node, determine the ratchet square drive and multiple first residual torque detection segments, the open head and multiple second residual torque detection segments, the internal hexagon and multiple third residual torque detection segments, the quick-change head and multiple fourth residual torque detection segments, and synchronize them to the residual torque detection wrench for automated segmented detection.

2. The intelligent detection method for bolt residual torque according to claim 1, characterized in that Divide the bolt distribution map into multiple detection priority zones, the method includes: Define a key connection area, a secondary connection area, and an auxiliary fixing area through the functional importance of the area to be measured under equipment association; Locate the high-vibration zones according to the vibration frequency distribution during equipment operation; Perform overlapping comparison between the key connection area, the secondary connection area, the auxiliary fixing area and the high-vibration zones, and evaluate the priority scores to determine the multiple detection priority zones.

3. The intelligent detection method for the residual torque of a bolt according to claim 2, characterized in that, Connect the residual torque detection wrench, perform associated matching between the bolt distribution map and the wrench drive head, and insert a drive head switching node in the residual torque detection sequence, the method includes: Identify the geometric features of the bolt head, establish a mapping database between the bolt model information and the wrench drive head, and the mapping database is used to store the types of compatible wrench drive heads for different bolt models; At the same time, perform drive head compatibility verification through the contact pressure distribution characteristics feedback by the torque sensor.

4. The intelligent detection method for the residual torque of a bolt according to claim 3, wherein Through the drive head switching node, determine the ratchet square drive and multiple first residual torque detection segments, the method further includes: Perform spatial position clustering on the multiple first residual torque detection segments to determine the initial population; Take the drive head switching time and the total path length under the drive head switching node as penalty factors, and perform local search and enhancement in the initial population to determine the first preferred detection path set corresponding to the ratchet square drive; Determine the drive head switching order through the first preferred detection path set corresponding to the ratchet square drive, the second preferred detection path set corresponding to the open head, the third preferred detection path set corresponding to the internal hexagon, and the fourth preferred detection path set corresponding to the quick-change head.

5. The intelligent detection method for the residual torque of a bolt according to claim 4, wherein, Take the drive head switching time and the total path length under the drive head switching node as penalty factors, and perform local search and enhancement in the initial population, the method further includes: Set an objective function through the detection accuracy loss rate; Select the solutions with objective function values lower than the accuracy loss threshold from the initial population as the target solutions, and the target solutions are used to execute the update decision of the multiple first residual torque detection segments.

6. The intelligent detection method for bolt residual torque according to claim 5, wherein The method includes: Perform task decomposition according to the multiple first residual torque detection segments, and generate an operation sequence including a pre-tightening force measurement subtask, a dynamic torque monitoring subtask, and an angle compensation subtask; Use the operation sequence to schedule detection resources. At the same time, establish a real-time feedback mechanism between the operation sequence and the state of the residual torque detection wrench.

7. The intelligent detection method for bolt residual torque according to claim 6, wherein, The pre-tightening force measurement subtask in the operation sequence adopts a first sampling period to synchronously trigger the sliding window length evaluation of the dynamic torque monitoring subtask; When the angle compensation subtask detects a sudden change in angular acceleration, activate an event-driven priority hybrid boosting strategy, which is used to compensate the first sampling period to be boosted to a second sampling period.

8. The intelligent detection method for the residual torque of a bolt according to claim 6, wherein Establish a real-time feedback mechanism between the operation sequence and the state of the residual torque detection wrench. The method further includes: Collect axial strain data at the root of the bolt thread; Use a long short-term memory network to extract temporal features from the axial strain data to obtain a predicted value of the bolt residual torque under the operation sequence; Compare the deviation value between the predicted value of the bolt residual torque and the wrench measurement value with a deviation threshold to determine whether to trigger the real-time feedback mechanism.

9. An intelligent detection system for bolt residual torque, characterized in that, For implementing an intelligent detection method for bolt residual torque according to any one of claims 1-8, the system includes: A bolt positioning module, configured to position the bolt head in the area to be measured, introduce bolt model information, and draw up a bolt distribution map with model markings; A request receiving module, configured to divide the bolt distribution map into multiple detection priority partitions; receive a residual torque detection request; and determine a residual torque detection sequence according to the multiple detection priority partitions in comparison with the residual torque detection request; An association matching module, configured to connect a residual torque detection wrench, perform association matching between the bolt distribution map and the wrench drive head, and insert a drive head switching node into the residual torque detection sequence, where the wrench drive head of the residual torque detection wrench can be selected from a ratchet square drive, an open head, an internal hexagon, and a quick change head; A segmented detection module, configured to determine, through the drive head switching node, the ratchet square drive and multiple first residual torque detection segments, the open head and multiple second residual torque detection segments, the internal hexagon and multiple third residual torque detection segments, and the quick change head and multiple fourth residual torque detection segments, and synchronize them to the residual torque detection wrench for automated segmented detection.

Images