Nut group cooperative tightening control method and system

Abstract

The application discloses a nut group cooperative tightening control method and system. The method obtains configuration information containing the number of nuts, topological relations and group-level quality targets, and initializes a group-level state; determines a next target nut according to the group-level state, generates an adaptive tightening target, controls a tightening device to execute after limiting the execution target through group-level constraints; extracts a single-point representation and updates the group-level state after completing current tightening, and the cycle continues until the group-level termination. The system includes configuration management, group-level decision, execution control and state update modules. The application realizes group-level sequence and target closed-loop cooperation, and improves pretightening uniformity.

Description

Technical Field

[0001] This application relates to the field of intelligent assembly and fastening process control technology, specifically to a method for controlling the coordinated tightening of nut groups, and also to a control system for the coordinated tightening of nut groups. Background Technology

[0002] In engineering equipment and high-end manufacturing, many critical connections are not single-point fasteners, but rather groups of nuts or bolts arranged in a certain spatial distribution to jointly undertake the clamping and sealing functions. For example, structures such as pipe flanges, gearbox end covers, compressor cylinder heads, and new energy motor housings typically have 8, 12, 16, or even more fastening points. The assembly quality at the group level depends on the synergy of the pre-tightening states at each point: even if the final tightening torque at each point is nominally the same, the actual clamping force may still be significantly uneven due to stiffness coupling, gasket compression nonlinearity, and workpiece flatness errors, leading to local leakage, warping deformation, or decreased reliability of the mating surfaces.

[0003] In existing technologies, a multi-point tightening scheme has been proposed that executes a fixed process script point by point, supplemented by single-point quality judgment or single-point process self-adaptation. At the group level, the tightening is usually carried out in a fixed sequence such as diagonal cross, star-shaped radiation, and clockwise rotation, with a unified target for each pass. At the single-point level, anomalies can be judged based on curve characteristics after tightening, or the locking stage and locking status can be identified based on a sliding time window during a single locking process, and the parameters of that point can be adjusted online. This type of scheme has engineering effectiveness under standard working conditions and is beneficial to single-point pass rate and traceability. However, its group-level sequence and target for each pass are usually fixed at once, and the closed loop of single-point judgment and process self-adaptation is limited to a single fastener. When there is mechanical coupling at multiple points, even if each point is qualified individually, it is still difficult to make group-level linkage adjustments to the order and target of subsequent points based on the feedback of tightened points, and uneven pre-tightening may still occur in the whole group.

[0004] In summary, existing technologies have limited adaptability to group-level fluctuations in incoming materials and coupling; single-point qualification does not equate to group-level qualification, and it is difficult to adjust the sequence of subsequent points and tightening targets based on process feedback from completed points during group-level assembly. Therefore, how to improve the group-level coordination and preload uniformity of nut group tightening without adding new nut-side sensing hardware is a pressing technical problem to be solved in this field. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for coordinated tightening control of nut groups. By acquiring configuration information containing topological relationships and group-level quality targets and initializing the group-level state, the next target nut is determined based on the group-level state and an adaptive tightening target is generated. After obtaining the execution target by limiting it with preset group-level constraints, the tightening device is controlled to execute. After the current tightening of the nut is completed, a single-point characterization is extracted and the group-level state is updated. This process is repeated until the group terminates, thereby improving the preload uniformity and assembly consistency of the entire group and overcoming the problems of insufficient adaptability to fixed sequence and uniform target torque, as well as the fact that single-point qualification does not equal group-level qualification.

[0006] To achieve the above objectives, the technical solution provided in this application is as follows: This application provides a method for coordinated tightening control of a nut group, comprising the following steps: acquiring configuration information of the nut group to be tightened and initializing the group-level state, wherein the configuration information includes at least the number of nuts, the topological relationship between nuts, and the group-level quality target; determining the next target nut based on the group-level state and generating an adaptive tightening target for the next target nut; constraining and limiting the adaptive tightening target according to preset group-level constraints to obtain an execution target; controlling the tightening device to tighten the next target nut according to the execution target; after the next target nut completes the current tightening action, extracting a single-point representation based on the torque-angle process data of the nut, and updating the group-level state based on the single-point representation; repeating the steps of determining the next target nut, generating the adaptive tightening target, constraining and limiting, executing tightening, and updating the group-level state until the group-level termination condition is met. In the above technical solution, the next target nut is determined from the nuts that have not yet completed their target; the group-level state is maintained by extracting single-point representations from the torque-angle process data of the tightening device and then converging them. Initially, it can be empty or preset, and it is updated as each nut completes its current tightening; the decision occurs at the inter-point group level. First, the next target nut and its tightening target are determined and executed, and then the group-level state is written back, rather than changing the stage and state parameters within a single-point process. This upgrades the fixed script to a group-level closed-loop collaborative control based on the feedback of completed tightening, and facilitates software deployment in existing tightening stations.

[0007] As a preferred embodiment, after the next target nut completes its current tightening action, a single-point representation is extracted based on the torque-angle process data of that nut, and the group-level state is updated based on the single-point representation. This includes: extracting a single-point representation based on the torque-angle process data of that nut; constructing a clamping contribution proxy amount for the corresponding nut based on the single-point representation; and updating the group-level state based on the single-point representation and the clamping contribution proxy amount. The single-point representation includes at least three of the following: final tightening torque, contact point angle, stiffness slope after contact, torque-angle energy integral, and average torque during the screw-in segment. In the above technical solution, the single-point representation and clamping contribution proxy amount can be determined by the torque-angle process data on the tightening device side, eliminating the need for a separate preload sensor embedded in the nut, which facilitates production line deployment.

[0008] As a preferred embodiment, the group-level state further includes a coupled propagation estimate of the clamping contribution proxy for nuts that have not yet been tightened; the coupled propagation estimate is obtained through neighborhood interpolation or message passing aggregation based on the topological relationship. In the above technical solution, the coupled propagation estimation enables the group-level state to simultaneously characterize the predicted distribution of observed points and incomplete points, providing a unified input for determining the next target nut and generating an adaptive tightening target.

[0009] As a preferred embodiment, determining the next target nut includes: for each candidate nut whose target has not yet been achieved, calculating a selection score based on its expected improvement to the group-level quality target and the path switching cost, and determining the candidate nut with the highest selection score as the next target nut. In the above technical solution, the tightening sequence is determined online based on the group-level status and path switching cost, instead of being fixed as a diagonal or star-shaped script.

[0010] As a preferred embodiment, generating the adaptive tightening target includes: making a bounded correction to the nominal process target based on the deviation of the clamping contribution surrogate amount of the next target nut from the mean within the group. In the above technical solution, the target allocation is boundedly corrected within the neighborhood of the nominal process target to reduce the group-level loss function, taking into account both uniformity improvement and process constraints; the group-level loss function may include at least one of the preload surrogate amount dispersion term and deformation or warpage surrogate term.

[0011] As a preferred embodiment, the method employs multiple tightening passes. Within the same pass, the tightening ratio coefficient for each nut is set according to the group-level state differences, and the difference in the target performance of each nut within the same pass satisfies the inter-point target range constraint. In the above technical solution, the tightening ratio coefficient within the same pass allows for controlled differences to compensate for local looseness or tightness, and is limited by the inter-point target range constraint.

[0012] As a preferred solution, when the initial tightening is completed but the group-level quality prediction result does not meet the group-level quality target, supplementary tightening is performed. This supplementary tightening includes: determining a subset of nuts to be tightened based on the marginal benefit of each nut in reducing group-level losses, generating a bounded supplementary tightening increment, and controlling the tightening device to perform supplementary tightening according to the bounded supplementary tightening increment. In the above technical solution, supplementary tightening is implemented on the subset of nuts with higher marginal benefits, which is beneficial for improving group-level indicators within a limited cycle time and suppressing over-tightening.

[0013] This application also provides a nut group collaborative tightening control system, comprising: a configuration management module, used to acquire configuration information of the nut group to be tightened and initialize the group-level state, wherein the configuration information includes at least the number of nuts, the topological relationship between nuts, and the group-level quality target; a group-level decision module, used to determine the next target nut based on the group-level state, generate an adaptive tightening target for the next target nut, and constrain the adaptive tightening target according to preset group-level constraints to obtain an execution target; an execution control module, used to control the tightening device to tighten the next target nut according to the execution target; and a state update module, used to extract a single-point representation based on the torque-angle process data of the next target nut after the current tightening action is completed, and update the group-level state based on the single-point representation; the group-level decision module, the execution control module, and the state update module cooperate to repeat the execution until the group-level termination condition is met. In the above technical solution, each module of the system corresponds to a method step; the group-level decision module can be further divided into sequential decision and target allocation functions, and the status update module can be further divided into single-point representation extraction and group status write-back functions; the group-level status is maintained by the process data on the device side through single-point representation aggregation. In the loop, after completing the current tightening, the single-point representation is extracted and the group-level status is written back, thereby supporting configuration, sequential and target joint decision-making and execution closed loop.

[0014] As a preferred embodiment, the system further includes a supplementary tightening planning module. This module is used to determine a subset of nuts to be tightened based on the marginal benefit of each nut to the reduction of group-level losses when the initial tightening is completed and the group-level quality prediction result does not meet the group-level quality target, and to generate a bounded supplementary tightening increment. In the above technical solution, initial tightening and supplementary tightening are incorporated into the same group-level framework, adapting to high-requirement scenarios such as sealing applications.

[0015] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 The diagram shown is a flowchart of an embodiment of a nut group coordinated tightening control method provided in this application.

[0018] Figure 2 The diagram shown is a schematic representation of an embodiment of a nut group coordinated tightening control system provided in this application.

[0019] Figure 3 The diagram shown is a structural schematic of an embodiment of the arrangement of twelve-point circumferential flange nuts provided in this application.

[0020] Figure 4 The diagram shown is a structural schematic of an embodiment of the arrangement of eight rectangular end cap nuts provided in this application.

[0021] Figure labeling: 10, Configuration Management Module; 20, Single Point Representation Extraction Module; 30, Group Status Update Module; 40, Sequential Decision Module; 50, Target Allocation Module; 60, Execution Control Module; 70, Tightening Planning Module; 80, Fault Rollback Module; 90, Tightening Device. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0023] In the description of this application, it should be noted that the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance. Unless otherwise expressly specified and limited, the term "connection" should be interpreted broadly; for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0024] In this embodiment, the symbols have the following meanings: N —Number of nuts, a positive integer; G=( V , E — The topology diagram corresponding to the topological relationships between nuts. V For a set of nodes, E Let it be the set of edges; T nom —Nominal final tightening torque, in units of N · m ; l r ——No. r The multiple tightening ratio is dimensionless. T end,i ——nut i Final tightening torque, in units of N · m ; i snug,i ——nut i The contact point angle, in degrees; i end,i ——nut i The turning angle at the end of this pass, in degrees; k i ——nut i The stiffness slope after bonding, in units of N · m / °; E i ——nut i Torque-angle energy integral, in units of N · m •° or as per unified agreement; T̄ run,i ——nut i The average torque of the screw-in section, in units of N · m ; F̂ i ——nut i The clamping contribution agent quantity is dimensionless or according to the normalization convention; S —Group-level status; L —Group-level loss function; α , β , c —Group-level loss weighting coefficient; Dl —Bounded correction amount for the proportional coefficient of multiple tightening cycles; d —Bounded adjustment range upper limit; or — Bounded correction gain. Example

[0025] This embodiment corresponds to the 12-point circumferential flange nut assembly; the method flow is as follows. Figure 1 For the topological relationship between the points and the nuts, please refer to [reference needed]. Figure 3 .

[0026] In this embodiment, the measured torque and measured rotation angle are process quantities measured by sensors on the tightening device side during single-point tightening, used to extract single-point characterization; the nominal torque, nominal rotation angle, or nominal multi-pass ratio are the nominal benchmarks in the group-level formula, kept fixed within a single tightening task and used as the benchmark for the nominal process target; the target torque, target rotation angle, and multi-pass tightening ratio coefficient are control setpoints issued at the current point, which can be corrected within the neighborhood of the nominal benchmark and constrained to form the execution target. The single-point execution layer can adopt conventional torque method, rotation angle method, or composite control, and can also nest and call adaptive control within the single-point process.

[0027] (1) Configuration information and group-level state initialization Configuration information includes the number of nuts, the topological relationship between nuts, group-level quality objectives, and nominal process objectives. Among these, group-level quality objectives include one or more of the following: preload uniformity parameters, point-to-point consistency parameters, deformation or warpage parameters, and cycle time parameters; in this embodiment, the configuration is as follows: Number of nuts N =12, the points are evenly distributed along the flange circumference and numbered 1 to 12; Topological relationship between nuts: adjacent edges i — i +1 (modulo 12, weight 1.0), diagonal side i — i +6 (weight 0.4); Group-level quality objective: Final clamping contribution rate F̂i The coefficient of variation does not exceed 0.10; the difference in final tightening torque between adjacent points does not exceed 8. N · m Single-group beat limit T cycle,max =180 s ; Nominal process target: Nominal final tightening torque T nom =80 N · m Multiple tightening passes are used, with each pass having a specific tightening ratio coefficient. l 1 / l 2 / l 3 =0.3 / 0.7 / 1.0.

[0028] In this embodiment, the final tightening torque is selected from the following: final tightening torque, contact point angle, stiffness slope after contact, torque-angle energy integral, and average torque of the screw-in section. T end,i , corner of the bonding point i snug,i Stiffness slope after bonding who Three terms as a single-point representation: T end,i This is the measured torque at the end of this pass; i snug,i The rotation angle corresponding to the peak value of the second-order difference of the measured torque is used to divide the screw-in section and the bonding section. who In order to [ i snug,i , i end,i The slope of the linear regression of torque–angle. Clamping contribution agent. F̂ i =0.6· T end,i / T nom +0.4 k i / k̄ (Used for relative comparison within a group, not equivalent to the true value of preload); the baseline is used when there is no completion point. k̄ =2.0 N · m / °, after a completion point k̄ Take the average value of the stiffness slope after the points have been properly bonded.

[0029] Preset group-level constraints include the upper limit of the target for each single point in each pass, the target range constraint between points within the same pass, and the maximum number of additional tightening attempts; the adaptive tightening target uses the target torque. Group-level status. S It is maintained by the aggregation of single-point representations of the nuts that have been tightened, initialized as an empty set, and may contain incomplete tightening masks (all initial values ​​are 1) and coupled propagation estimates of the clamping contribution surrogate quantity for the incomplete tightened nuts.

[0030] (2) First time The first pass is used to establish contact and update the group-level status. The bounded correction range for the multi-pass tightening ratio coefficient is then used. Dl |Maximum value 0.02; In the selection of rating, the expected improvement amount is weighted. wq =1.0, Path switching cost weight toilet =0.20 (manual gun holding). Determine the next target nut based on the group-level status. The complete sequence for this pass is: 1→7→4→10→2→8→5→11→3→9→6→12 The nominal process target torque at each point is l 1 · T nom =24.0 N · m After constraint limiting, the difference in the target execution range of each nut within the same pass satisfies the inter-point target range constraint (the range in this pass does not exceed 4).N · m The control tightening device tightens according to the execution target; after each point completes the current tightening action, the above three single-point characteristics are extracted based on the torque-angle process data, the clamping contribution agent quantity is constructed, and the group-level status is updated. After all 12 points are completed, the second round begins.

[0031] (3) Second time Tighten the second and subsequent tightening ratios. l 2 =0.7. Take | Dl |Maximum value 0.05 w q =1.0、 w c =0.15; In the preset group-level constraints, the upper limit for a single-point target is 0.75. T nom =60 N · m The objective range constraint between points is that the range does not exceed 6. N · m .

[0032] The following provides numerical values ​​for the process of determining the next target nut and generating the execution target in one step.

[0033] step A Update group-level status. Assume points 4 and 6 have completed their current tightening. Final tightening torque at point 4. T end,4 =56.0 N · m Stiffness slope after bonding k 4 =1.25 N · m / °, k̄ =2.00 N · m / °, then clamp the contribution agent volume F̂ 4 =0.6×(56.0 / 80)+0.4×(1.25 / 2.00)=0.420+0.250=0.670 Point 6 clamping contribution agent volume F̂ 6 =0.580. Within-group mean F̄ = (0.670 + 0.580) / 2 = 0.625.

[0034] step BEstimate the clamping contribution surrogate for candidate nuts whose objectives have not yet been achieved. For candidate nut 5, its neighborhood includes completed points 4 and 6 (adjacent edges, weight 1.0) and incomplete point 11 (diagonal edge, weight 0.4; the prior clamping contribution surrogate is 0.25). The surrogate is obtained by weighting the values ​​according to the topological relationships between nuts. F̂ pred (5) = (1.0 × 0.670 + 1.0 × 0.580 + 0.4 × 0.25) / (1.0 + 1.0 + 0.4) = 1.35 / 2.4 = 0.5625 step C : Determine the next target nut. The current tightening device is at point 6, and the candidate nuts that have not yet achieved their target are points 5 and 10. Calculate a selection score for each candidate nut: The selection score is determined based on the expected improvement of the candidate nut to the group-level quality target and the path switching cost. In this example, we take... s j = w q · ΔL̂ j - w c · C path ( j ),in ΔL̂ j For the expected improvement, C path The cost of path switching is calculated (adjacent points are counted as 1, and for points spanning multiple points, the shortest hop count is accumulated). Take... ΔL̂ 5 =0.040 C path (5) = 1; ΔL̂ 10 =0.045 C path (10) = 4, then s 5 =1.0×0.040-0.15×1=-0.110 s 10 =1.0 × 0.045 - 0.15 × 4 = -0.555 Select the candidate nut with the highest score as the next target nut, i.e., point 5.

[0035] step D Generate an adaptive tightening target and constrain it to obtain the execution target. Based on the deviation of the clamping contribution of the next target nut from the group mean, make a bounded correction to the nominal process target: take or =0.5 d =0.05, Dl = clip ( or ·( F̄ - F̂ pred (5)),- d , d )= clip (0.5×(0.625-0.5625),-0.05,0.05)=0.03125 Obtain the torque corresponding to the adaptive tightening target ( l 2 + Dl )· T nom = (0.7 + 0.03125) × 80 = 58.5 N · m Furthermore, based on preset group-level constraints, this value is limited to 60 units below the single-point target upper limit. N · m Furthermore, since the objective range constraint between points is satisfied, the objective value is set to 58.5. N · m The tightening device is controlled to tighten point 5 according to the execution target.

[0036] Following the same rolling decision-making rules, the complete sequence for the second round in this embodiment is as follows: 4→6→5→3→2→1→12→11→10→9→8→7 Group-level loss L = α · Var ({ F̂ i})+ β · WarpProxy ( S )+ c · CyclePenalty , this calculation example α =1.0、 β =0.3 c =0.1, used for group-level quality prediction.

[0037] (4) Third pass, group-level termination and additional tightening Tighten the ratio coefficient multiple times for the third time. l 3 =1.0, | Dl |≤0.03, single-point target upper limit 1.05· T nom =84 N · m The cost weight of path switching can be reduced to toilet =0.08. A complete sequence is: 1→2→3→4→5→6→7→8→9→10→11→12 After three passes, if the group-level quality target is met (coefficient of variation ≤ 0.10 and difference in torque between adjacent final tightening passes ≤ 8), the result is satisfactory. N · m If the initial tightening is completed but the group-level quality prediction result does not meet the group-level quality target (in this example)... L > L th =0.012), then perform additional tightening: determine the subset of nuts to be tightened based on the marginal benefit of each nut to the reduction of group-level loss. The marginal benefits of points 3, 5, and 7 are 0.0048, 0.0031, and 0.0045, respectively. Take the first... K = 3 subsets of nuts to be tightened {3,7,5}; generate bounded tightening increments. ΔT rework =2.0 N · m Maximum number of tightening cycles N rework,max =3, and control the tightening device to perform supplementary tightening according to the bounded supplementary tightening increment. Example

[0038] This embodiment corresponds to an 8-point rectangular end cap nut assembly; the method flow is as follows. Figure 1 For the topological relationship between the points and the nuts, please refer to [reference needed]. Figure 4 .

[0039] In this embodiment, the measured torque and measured rotation angle are process quantities measured by sensors on the tightening device side during single-point tightening, used to extract single-point characterization; the nominal torque, nominal rotation angle, or nominal multi-pass ratio are the nominal benchmarks in the group-level formula, kept fixed within a single tightening task and used as the benchmark for the nominal process target; the target torque, target rotation angle, and multi-pass tightening ratio coefficient are control setpoints issued at the current point, which can be corrected within the neighborhood of the nominal benchmark and constrained to form the execution target. The single-point execution layer can adopt conventional torque method, rotation angle method, or composite control, and can also nest and call adaptive control within the single-point process.

[0040] (1) Configuration information and group-level state initialization The points are numbered in two rows and four columns as follows: 1234 5678 Configuration information includes the number of nuts, the topological relationship between nuts, group-level quality objectives, and nominal process objectives. Among these, group-level quality objectives include one or more of the following: preload uniformity parameters, point-to-point consistency parameters, deformation or warpage parameters, and cycle time parameters; in this embodiment, the configuration is as follows: Number of nuts N =8; Topological relationships between nuts: adjacent sides in the same row / column (e.g., 1-2, 1-5) have a weight of 1.0; diagonal sides in the same sealing zone (e.g., 1-6, 2-5) have a weight of 0.5; Group-level quality objective: Clamp contribution agent volume F̂i The coefficient of variation does not exceed 0.12; the difference in final tightening torque between adjacent torques in the same row does not exceed 5. N · m ; T cycle,max =90 s ; Nominal process objectives: T nom =45 N · m Multiple tightening ratio coefficients l 1 / l 2 =0.5 / 1.0.

[0041] In this embodiment, the final tightening torque is selected from the following: final tightening torque, contact point angle, stiffness slope after contact, torque-angle energy integral, and average torque of the screw-in section. T end,i Torque-angle energy integral Hey Average torque of the screw-in section T̄ run,i Three terms as a single-point representation: T end,i This is the measured torque at the end of this pass; E i = ΣTk · Dthk Spin into the starting point and accumulate to the end point of this cycle; T̄ run,i To match the average torque of the front section. Clamping contributes to the agent's output. F̂ i =0.6· T end,i / T nom +0.3 E i / E +0.1 T̄ run,i / T ̄ run (Used for relative comparison within a group, not equivalent to the true value of preload); the baseline is used when there is no completion point. E 0 =200 N · m ·°、 T ̄ run,0 =4.0 N· m After there is a completion point E , T̄ run Take the average of the quantities corresponding to the completed points.

[0042] Preset group-level constraints include single-point target upper limit, inter-point target range constraint, and maximum number of tightening passes; the adaptive tightening target uses the target torque. Group-level status. S It is maintained by the aggregation of single-point representations of the nuts that have been tightened, initialized as an empty set, and may contain incomplete tightening masks and coupled propagation estimates of the clamping contribution surrogate for the incomplete tightening nuts.

[0043] (2) First time Automated workstation, retrieve | Dl |Maximum value 0.02, w q =1.0、 w c =0.10. Determine the next target nut based on the group-level status. The complete sequence for this pass is: 1→5→4→8→2→6→3→7 The nominal process target torque at each point corresponds to 0.5 × 45 = 22.5. N · m After constraints are applied, the difference in the target performance within the same pass satisfies the inter-point target range constraint (the range in this pass does not exceed 3). N · m After each tightening action is completed, the final tightening torque, torque-angle energy integral, and average torque of the screw-in segment are extracted to construct the clamping contribution agent and update the group-level status.

[0044] Example (point 1 has just been completed): T end,1 =22.0 N · m , E 1 =210 N · m ·°, T̄ run,1 =3.6 N · m There are currently no other completion points, so take... E =200 N · m ·°、 T̄ run =4.0 N · m ,but F̂ 1=0.6×(22.0 / 45)+0.3×(210 / 200)+0.1×(3.6 / 4.0)=0.293+0.315+0.090=0.698 (3) Second time Tighten the second and subsequent tightening ratios. l 2 =1.0. Take or =0.4 d =0.04; Single-point target upper limit 1.05· T nom =47.25 N · m The target range constraint between points is that the range does not exceed 4. N · m .

[0045] step A Assume points 1 and 5 have completed the current tightening, and the clamping contributes to the agent volume. F̂ 1 =0.698 F̂ 5 =0.640, within-group mean F̄ = (0.698 + 0.640) / 2 = 0.669.

[0046] step B The neighborhood of candidate nut 6, whose objective has not yet been achieved, includes point 5 (adjacent edge, weight 1.0) and point 1 (diagonal edge, weight 0.5), and is obtained by weighting the neighborhoods according to the topological relationships between the nuts. F̂ pred (6) = (1.0 × 0.640 + 0.5 × 0.698) / (1.0 + 0.5) = 0.989 / 1.5 = 0.659 step C The current tightening device is at point 5, and the candidate nuts are at points 6 and 8. For each candidate nut, a score is calculated based on the expected improvement and the cost of switching paths: [The score is not provided in the original text]. ΔL̂ 6 =0.028 C path (6) = 1; ΔL̂ 8 =0.030, C path (8) = 3, then s 6 =1.0 × 0.028 - 0.10 × 1 = -0.072 s 8 =1.0 × 0.030 - 0.10 × 3 = -0.270 The one with the highest score will be selected as the next target nut, i.e., point 6.

[0047] step D Based on the deviation of the clamping contribution of the next target nut from the group mean, a bounded correction is made to the nominal process target. Dl = clip (0.4×(0.669-0.659),-0.04,0.04)=0.0040, the adaptive tightening target torque is (1.0+0.0040)×45=45.18 N · m After further constraints at the preset group level, the execution target is set to 45.2. N · m The tightening device is controlled to tighten according to the target.

[0048] Following the same rolling decision-making rules, the second complete sequence in this embodiment is: 5→6→7→8→4→3→2→1 (4) Group-level termination and additional tightening After two passes, if the group-level quality target is met (coefficient of variation ≤ 0.12 and difference in final torque between adjacent rows ≤ 5), the result is satisfactory. N · m If the initial tightening is completed but the group-level quality prediction result does not meet the group-level quality target (in this example)... L > L th =0.015), then perform additional tightening: determine the subset of nuts to be tightened based on the marginal benefit of each nut to the reduction of group-level loss. The marginal benefits of points 2, 3, and 7 are 0.0035, 0.0020, and 0.0032, respectively. K =2 yields the subset {2,7} of nuts to be tightened; generates a bounded tightening increment. ΔT rework =1.5 N · m Maximum number of tightening cycles N rework,max =2, and control the tightening device to perform supplementary tightening according to the bounded supplementary tightening increment. Example

[0049] This embodiment provides a specific implementation of a nut group coordinated tightening control system. (See also...) Figure 2 .

[0050] The system includes: a configuration management module 10, a group-level decision-making module, an execution control module 60, and a status update module; wherein the group-level decision-making module can be divided into a sequential decision-making module 40 and a target allocation module 50, and the status update module can be divided into a single-point representation extraction module 20 and a group status update module 30; preferably, it also includes a supplementary tightening planning module 70 and a fault rollback module 80. The execution control module 60 communicates with the tightening device 90.

[0051] In this embodiment, the configuration information read by the configuration management module 10 includes the number of nuts, the topological relationship between nuts, the nominal process target, the group-level quality target, and the preset group-level constraints. The group-level quality target includes one or more of the following: preload uniformity parameters, point consistency parameters, deformation or warping parameters, and cycle time parameters. The preset group-level constraints include the upper limit of single-point targets for each pass, the range constraint of inter-point targets, and the maximum number of additional tightening cycles. The adaptive tightening target can be one or more of the following: target torque, target rotation angle, or multi-pass tightening ratio coefficient, which are constrained to form the execution target. The measured torque and measured rotation angle are provided by sensors on the tightening device side, which are used by the single-point characterization extraction module 20 to extract single-point characterizations and construct clamping contribution surrogate quantities. The group-level state maintained by the group state update module 30 is formed by the aggregation of single-point characterizations of nuts that have been tightened, and may include the incomplete tightening mask and the coupled propagation estimate of the clamping contribution surrogate quantities of incomplete points. The single-point execution layer can use the torque method, the rotation angle method, or a composite control, or it can be nested with adaptive control within the single-point process.

[0052] (1) Module division of labor and data flow Configuration management module 10 reads configuration information (number of nuts) N Nominal process target T nom (Including multiple tightening ratio coefficients, topological relationship edge tables between nuts, group-level quality targets and preset group-level constraints), and initialize the group-level state. S The sequential decision-making module 40 outputs the next target nut number based on the selected score. The target allocation module 50 generates an adaptive tightening target and determines the execution target based on preset group-level constraints, encapsulating it into a sent message. The execution control module 60 sends "point identifier + execution target for this pass" to the tightening device 90 and receives the tightening completion event. The single-point characterization extraction module 20 uses the device-side torque-angle process data (sampling rate 1) to extract the target nut number. kHz Single-point process cache no less than 8 s Extract single-point representations from the current configuration of single-point representation combinations and construct clamping contribution surrogate quantities. F̂ i The process is recorded during writing. After the event is completed, the group state update module 30 writes back the group-level state based on the single-point representation and the clamping contribution agent quantity. S (Industrial computer control cycle 10) msStatus write-back time limit ≤ 50 ms When the initial tightening is completed but the group-level quality prediction result does not meet the group-level quality target, the supplementary tightening planning module 70 outputs a subset of nuts to be tightened and a bounded supplementary tightening increment. Each module works in coordination to repeat this process until the group-level termination condition is met.

[0053] Taking point 5 of the second pass of the 12-point circumferential flange configuration as an example, the execution target 58.5 is issued. N · m When running with the 8-point rectangular end cap configuration, the corresponding execution target of the configuration is issued.

[0054] (2) Workstation configuration and path cost determination Manual gun-holding station: The interface highlights the next nut. toilet A value of 0.20 to 0.25 is acceptable. For automatic or multi-axis workstations: movement is based on a point coordinate table. toilet A value of 0.08 to 0.15 can be used; the positioning time for adjacent points should be 0.8. s Diagonal or cross-regional calculations are based on 1.6. s Include cycle time. Cycle time budget varies depending on configuration (12-point circumferential flange configuration is 180). s The 8-point rectangular end cap configuration is 90. s ).

[0055] (3) Engineering parameters for constraints and rollback The target allocation module has built-in preset group-level constraints, including the upper limit of single-point targets for each pass, the target range constraint between points, and the maximum number of additional tightening attempts. When the confidence score of the sequential decision for selecting all candidate points is continuously below 0.20 for 3 decision steps, or the target allocation touches the constraint boundary 5 times consecutively, the fault rollback module 80 switches the nuts for which the target has not yet been completed to the pre-stored fixed sequence and unified multi-pass tightening ratio coefficient of the configuration, and marks the rollback operation. Example of the pre-stored rollback sequence for a 12-point circumferential flange configuration: 1-7-4-10-2-8-5-11-3-9-6-12; Example of the pre-stored rollback sequence for an 8-point rectangular end cap configuration: 1-5-3-7-2-6-4-8.

[0056] (4) Process Record It can save group-level decision trajectories: group identifier, complete sequence of each pass, execution target and measured curve summary for each point and each pass. F̂ i , L , whether to tighten again, whether to retract.

[0057] Those skilled in the art will understand that the system structure shown in the accompanying drawings is merely illustrative, and each module can be implemented through software, hardware, or a combination of both; the module division does not preclude further merging or splitting. All modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for controlling the coordinated tightening of a nut assembly, characterized in that, Includes the following steps: Acquire the configuration information of the group of nuts to be tightened and initialize the group-level state. The configuration information includes at least the number of nuts, the topological relationship between the nuts, and the group-level quality target. Based on the group-level status, determine the next target nut and generate an adaptive tightening target for the next target nut; The adaptive tightening target is constrained and limited according to preset group-level constraints to obtain the execution target; The control tightening device tightens the next target nut according to the execution target; After the next target nut completes its current tightening action, a single-point representation is extracted based on the torque-angle process data of that nut, and the group-level state is updated based on the single-point representation. Repeat the steps of determining the next target nut, generating an adaptive tightening target, setting constraints, performing tightening, and updating the group-level state until the group-level termination condition is met.

2. The method according to claim 1, characterized in that, The update of the group-level state further includes: constructing the clamping contribution proxy of the corresponding nut based on the single-point characterization, and updating the group-level state according to the single-point characterization and the clamping contribution proxy; the single-point characterization includes at least three of the following: final tightening torque, contact point rotation angle, stiffness slope after contact, torque-rotation angle energy integral, and average torque of the screw-in segment.

3. The method according to claim 1, characterized in that, The determination of the next target nut further includes: for each candidate nut whose target has not yet been achieved, calculating a selection score based on its expected improvement to the group-level quality target and the path switching cost, and determining the candidate nut with the highest selection score as the next target nut.

4. The method according to claim 1 or 2, characterized in that, The generation of the adaptive tightening target includes: making a bounded correction to the nominal process target based on the deviation of the clamping contribution agent of the next target nut from the mean of the group.

5. The method according to claim 1, characterized in that, The method employs multiple tightening passes; within the same pass, the multiple tightening ratio coefficient of each nut is set according to the group-level state differentiation, and the difference in the target of each nut within the same pass satisfies the inter-point target range constraint.

6. The method according to claim 1, characterized in that, When the initial tightening is completed and the group-level quality prediction result does not meet the group-level quality target, supplementary tightening is performed. The supplementary tightening includes: determining a subset of nuts to be tightened based on the marginal benefit of each nut to the reduction of group-level loss, generating a bounded supplementary tightening increment, and controlling the tightening device to perform supplementary tightening according to the bounded supplementary tightening increment.

7. A nut group cooperative tightening control system, characterized in that, include: The configuration management module is used to acquire the configuration information of the nut group to be tightened and initialize the group-level state. The configuration information includes at least the number of nuts, the topological relationship between the nuts, and the group-level quality target. The group-level decision module is used to determine the next target nut based on the group-level state, generate an adaptive tightening target for the next target nut, and constrain the adaptive tightening target according to preset group-level constraints to obtain the execution target. The execution control module is used to control the tightening device to tighten the next target nut according to the execution target; The state update module is used to extract a single-point representation based on the torque-angle process data of the next target nut after the current tightening action is completed, and update the group-level state according to the single-point representation. The group-level decision-making module, the execution control module, and the status update module work together to execute repeatedly until the group-level termination condition is met.

8. The system according to claim 7, characterized in that, The system also includes a supplementary tightening planning module, which is used to determine a subset of nuts to be tightened based on the marginal benefit of each nut to the reduction of group-level loss when the initial tightening is completed and the group-level quality prediction result does not meet the group-level quality target, and to generate a bounded supplementary tightening increment.

Images