A method, system, device and storage medium for monitoring and controlling optimization of a hole diameter of a sheet spring assembly

Abstract

The application discloses a kind of sheet spring assembly's aperture monitoring control optimization method, system, equipment and storage medium, including by the depth of multiple times of guiding bolt penetration and the sequence of propulsive force, standard guiding gap and maximum allowable propulsive force are determined in combination with radial trajectory envelope analysis, and standard aperture monitoring control set is constructed;Further, in parameter space, the control combination is expanded using Latin hypercube sampling, form multidimensional second control set, and it is mapped as sleeve displacement and torque control instruction applied to penetration monitoring process;Through a plurality of monitoring data feature extraction and Q learning training, obtain assembly accessibility incentive value, finally construct three-dimensional response surface and screen optimal control parameter, realize aperture monitoring control adaptive optimization and high-precision path evaluation;Solve the problem that sheet spring lamination aperture monitoring parameter relies on experience setting, assembly accessibility and hole shaft deviation identification are difficult to coordinate balance and detection misjudgment rate is higher.

Description

Technical Field

[0001] This invention relates to the field of aperture monitoring technology, and more specifically, to an aperture monitoring and control optimization method, system, device, and storage medium for sheet spring assembly. Background Technology

[0002] In the existing sheet spring assembly process, during the stack alignment stage, the guide pin feedback hole diameter monitoring technology is usually used to detect the assembly status of the stack hole group. This technology introduces a guide sleeve to establish a unified axis constraint in the free stacking state of the sheets, so that the guide pin passes through the hole structure of multiple sheet springs along a fixed path. During the penetration process, the insertion depth and propulsion force response data are collected. By analyzing the changes in penetration resistance, the obstruction section and the penetration completion status, the throughput capacity, hole axis offset and local deformation of the stack hole group are determined.

[0003] In the monitoring of the guide pin feedback aperture, guide clearance and maximum allowable thrust are two core control parameters. Guide clearance, determined by the difference between the guide sleeve inner diameter and the pin diameter, limits the radial offset range of the pin during penetration. Maximum allowable thrust is the upper limit of the axial force allowed for the pin during its advancement, restricting its advancing behavior. The values ​​of these two parameters jointly affect the mechanical response and path adjustment behavior of the pin during penetration.

[0004] In existing technologies, the values ​​of guide clearance and maximum permissible thrust are usually determined based on operational experience or fixed configurations. However, in practical applications, there is a complex correlation between the values ​​of these two parameters and the geometric state and deviation distribution of the leaf spring lamination hole group. Under different lamination states and hole diameter deviations, fixed parameter configurations cannot guarantee the stability and consistency of detection results, and the problems of insufficient detection sensitivity and high false judgment rate have long existed. No technical solution has yet been proposed that can achieve coordinated configuration of guide clearance and maximum permissible thrust based on the actual lamination state.

[0005] To address the above problems, this invention proposes a solution. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method, system, device, and storage medium for aperture monitoring and control optimization of sheet spring assembly. Through closed-loop iteration of Latin hypercube sampling, reinforcement learning training, and response surface screening, data-driven optimization of guide clearance and maximum allowable propulsion force is achieved, thereby solving the problems of parameter dependence on experience setting, difficulty in coordinating and balancing assembly accessibility and hole axis deviation identification, and high detection misjudgment rate in sheet spring lamination aperture monitoring.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An optimization method for aperture monitoring and control in sheet spring assembly includes the following steps: Collecting the insertion depth and propulsion force sequence of multiple penetrations of the guide pin; after aligning the depths, selecting the penetration process with the smallest propulsion force fluctuation; taking the upper quantile value of its propulsion force as the maximum allowable propulsion force; performing envelope analysis on the maximum radial offset of the pin trajectory to obtain the guide sleeve constraint range and determine the standard guide clearance, forming a standard aperture monitoring and control set; generating several combinations of guide clearance and maximum allowable propulsion force using Latin hypercube sampling within the two-dimensional parameter space of the standard aperture monitoring and control set, obtaining a second aperture monitoring and control set; and analyzing the radial constraint range based on the trajectory acquisition of the sleeve inner wall, converting the guide clearance into a sleeve... The sleeve adjustment displacement is determined based on the relationship between the drive current and the propulsion force response, thereby transforming the aperture monitoring and control set into output torque and sleeve adjustment displacement. Each second aperture monitoring and control set is transformed and applied to the guide pin feedback aperture monitoring process to obtain the actual assembly diameter capability and hole shaft deviation status of the sheet spring lamination hole group. Features are extracted for each monitoring process, and each second aperture monitoring and control set is trained based on the Q-learning algorithm to output the corresponding assembly reachability excitation value. A three-dimensional response surface is constructed based on the second aperture monitoring and control set and the corresponding assembly reachability excitation value, and the optimal aperture monitoring and control set is analyzed and selected for application to guide pin feedback aperture monitoring.

[0008] An aperture monitoring and control optimization system for sheet spring assembly includes a penetration modeling module, a sampling expansion module, a parameter mapping module, a bore detection module, a reinforcement learning module, and an optimization control module. The penetration modeling module collects the insertion depth and propulsion force sequence of multiple penetrations of the guide pin, aligns the depths, selects the penetration process with the smallest propulsion force fluctuation, takes its upper quantile value as the maximum allowable propulsion force, performs envelope analysis on the maximum radial offset of the pin trajectory, obtains the guide sleeve constraint range, and determines the standard guide clearance, forming a standard aperture monitoring and control set. The sampling expansion module generates several combinations of guide clearance and maximum allowable propulsion force using Latin hypercube sampling within the two-dimensional parameter space of the standard aperture monitoring and control set, obtaining a second aperture monitoring and control set. The parameter mapping module is used to sample the trajectory of the sleeve inner wall. The system analyzes the radial constraint range, converts the guide clearance into sleeve adjustment displacement, and determines the output torque based on the relationship between the drive current and the propulsion force response, thus transforming the aperture monitoring and control set into output torque and sleeve adjustment displacement. The bore detection module converts each second aperture monitoring and control set and applies it to the guide pin feedback aperture monitoring process, obtaining the actual assembly bore capacity and bore shaft deviation status of the sheet spring lamination hole group. The reinforcement learning module extracts features from each monitoring process, trains each second aperture monitoring and control set based on the Q-learning algorithm, and outputs the corresponding assembly reachability excitation value. The optimal control module constructs a three-dimensional response surface based on the second aperture monitoring and control set and the corresponding assembly reachability excitation value, and analyzes and selects the optimal aperture monitoring and control set for application to guide pin feedback aperture monitoring.

[0009] An electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform an aperture monitoring and control optimization method for sheet spring assembly.

[0010] A computer-readable storage medium storing a computer program that, when executed by a processor, implements an aperture monitoring and control optimization method for sheet spring assembly.

[0011] The technical effects and advantages of the aperture monitoring and control optimization method, system, equipment and storage medium for sheet spring assembly of the present invention are as follows: 1. This invention establishes benchmark control parameters based on the process of minimizing propulsion fluctuations, and uses Latin hypercube sampling to generate an optimal control set covering different constraint strengths and propulsion levels in the parameter space. This makes aperture monitoring and control no longer dependent on a single parameter set by experience, but rather determines control parameters based on a comprehensive characterization of penetration process data, thereby simultaneously coordinating radial constraints and axial propulsion capabilities during pin penetration.

[0012] 2. This invention generates excitation values ​​for each set of control parameters through reinforcement learning training, and combines three-dimensional response surface to screen the optimal control combination, enabling the system to select the optimal control parameters according to the changes in penetration resistance response and spatial constraints under different assembly states, while ensuring the pin penetration capability and maintaining the ability to identify insufficient hole diameter and hole axis offset. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating the aperture monitoring and control optimization method for sheet spring assembly according to the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0015] Example 1, Figure 1 The present invention provides an optimization method for aperture monitoring and control in sheet spring assembly, comprising the following steps: S1. Collect the insertion depth and propulsion force sequence of the guide pin through multiple penetrations. After aligning the depths, select the penetration process with the smallest propulsion force fluctuation and take its upper quantile value as the maximum allowable propulsion force. Perform envelope analysis on the maximum radial offset of the pin trajectory to obtain the guide sleeve constraint range and determine the standard guide gap, forming a standard aperture monitoring and control set.

[0016] In this embodiment, the sequence of insertion depth and propulsion force of multiple penetrations of the guide pin is collected. After aligning the depths, the penetration process with the smallest propulsion force fluctuation is selected, and its upper quantile value is taken as the maximum allowable propulsion force. Envelope analysis is performed on the maximum radial offset of the pin trajectory to obtain the guide sleeve constraint range and determine the standard guide clearance, forming a standard aperture monitoring and control set, specifically: In the free assembly state of the sheet spring laminations, the insertion depth time series and the corresponding axial thrust time series of the guide pin are collected during multiple penetration processes. Each penetration process is resampled at equal intervals according to the insertion depth, so that all penetration processes form a one-to-one corresponding set of thrust sequence at the same depth sampling point. For the propulsion force sequence set, the discrete amplitude difference of the propulsion force value is calculated at each depth sampling point, and the discrete amplitude difference of each penetration process is accumulated over the entire depth range. The penetration process with the smallest accumulated value is determined as the target penetration process. The thrust force values ​​corresponding to all depth sampling points during the target penetration process are summarized into a unified thrust force sample sequence. After sorting the sample sequence by value, the thrust force value located at the preset cumulative percentage position is selected as the maximum allowable thrust force. Spatial coordinate reconstruction is performed on the pin trajectory data during the target penetration process. The collected two-dimensional position coordinate sequence is synchronously mapped with the insertion depth sequence to establish a three-dimensional spatial trajectory curve with insertion depth as the independent variable and radial offset as the dependent variable. The radial offset value is calculated point by point for the spatial trajectory curve within the full insertion depth range, and the spatial coordinate points where the radial offset changes locally are selected as boundary characterization points. All boundary characterization points are projected onto the same cross-sectional plane, and the minimum envelope circle that can completely cover all projection points is calculated. The radius of the minimum envelope circle is taken as the guide constraint radius. The difference between the guide constraint radius and the outer diameter of the pin is used as the standard guide clearance. The standard guide clearance is combined with the maximum allowable thrust to form a standard aperture monitoring and control set that includes a single guide clearance value and a single maximum allowable thrust value.

[0017] In this embodiment, the insertion depth time series is a continuous record of the axial position of the guide pin changing over time during the penetration process. It is acquired in real time by a displacement sensor and the current position of the pin is recorded at fixed time intervals to reflect the displacement trajectory of the pin from entering the hole to complete penetration.

[0018] In this embodiment, the axial thrust time series is a continuous record of the axial resistance experienced by the guide pin during the penetration process as a function of time. It is synchronously acquired by a force sensor and corresponds one-to-one with the insertion depth time series. It is used to characterize the resistance response of the hole structure at different positions to the pin.

[0019] In this embodiment, equal-interval resampling involves re-dividing the original acquired insertion depth sequence into sampling points according to a uniform depth interval, and interpolating or matching the original thrust data to ensure that all penetration processes have corresponding thrust data at the same depth node, thereby achieving comparability between different penetration processes.

[0020] In this embodiment, the discrete amplitude difference of the propulsion force is the difference between the maximum and minimum values ​​of the propulsion force corresponding to different penetration processes at the same depth sampling point.

[0021] In this embodiment, the cumulative discrete amplitude difference across the entire depth range is the sum of the discrete amplitude differences at all depth sampling points for each penetration process. The smaller the value, the more stable the propulsion fluctuation of the penetration process is along the entire path.

[0022] In this embodiment, the penetration process with the smallest cumulative discrete amplitude difference across the entire depth range is defined as the target penetration process.

[0023] In this embodiment, the unified thrust sample sequence is a data sequence formed by summarizing the thrust values ​​corresponding to all depth sampling points during the target penetration process in depth order.

[0024] In this embodiment, the preset cumulative percentage position is a cumulative percentage position selected after sorting the thrust sample sequence by numerical value. It is used to determine the representative thrust level in the higher range of the sample, so as to reflect the larger drag level but avoid the influence of extreme outliers.

[0025] In this embodiment, the upper quantile thrust value is the thrust value located at the preset cumulative percentage position in the sorted thrust sample sequence, and this value is defined as the maximum allowable thrust.

[0026] In this embodiment, spatial coordinate reconstruction involves synchronously fusing the original acquired two-dimensional planar position data with the corresponding insertion depth data to construct three-dimensional trajectory data containing axial position and radial offset information.

[0027] In this embodiment, the three-dimensional spatial trajectory curve is a continuous trajectory curve established with the insertion depth as the main axis direction and the radial offset as the lateral change amount, which is used to describe the degree of radial deviation of the pin at each depth position during the penetration process.

[0028] In this embodiment, the radial offset value is the lateral deviation distance of the pin relative to the guide axis at a certain depth position, which is calculated by position sensing data.

[0029] In this embodiment, the boundary characterization points are spatial coordinate points where local extreme changes in radial offset are selected after calculating the radial offset value of the spatial trajectory curve point by point within the entire insertion depth range. Specifically, the radial offset values ​​of adjacent positions are compared point by point along the insertion depth direction. When the radial offset value at a certain point increases or decreases simultaneously with its adjacent points before and after it, that point is determined to be a local extreme point, and all local extreme points are used as boundary characterization points.

[0030] Projecting all boundary characterization points onto the same cross-sectional plane yields a set of projected points. The coordinates of any projected point in this set are denoted as: in, Let the coordinates of the projection point be the coordinates of the point in the first direction within the cross-sectional plane. Let be the coordinates of the projection point in the second direction within the cross-sectional plane. The subscript 'i' indicates the index of the projection point in its set, and the value of 'i' ranges from 1 to the total number of projection points contained in the set. The minimum envelope circle is the circle with the smallest radius that can completely cover all projection points on the aforementioned cross-sectional plane. Let the coordinates of the center of this minimum envelope circle be: ,in Let be the coordinate value of the center of the circle in the first direction within the cross-sectional plane. Let be the coordinates of the center of the circle in the second direction within the cross-sectional plane; let r be the radius of the minimum envelope circle. For any candidate center position c, its corresponding coverage radius is the maximum distance between the candidate center and all projection points, i.e., first calculate the distance from the candidate center to each projection point. Euclidean distance: The maximum value among these distances is then taken as the coverage radius corresponding to the candidate circle center. The minimum envelope circle is the circle that minimizes the aforementioned coverage radius among all possible candidate circle center positions, and its radius r is the coverage radius corresponding to this minimum value, i.e. , where P represents the set of all projected points. The radius r mentioned above is taken as the guiding constraint radius; In this embodiment, the standard guide clearance is the difference between the guide constraint radius and the outer diameter of the pin, which represents the minimum radial relaxation required by the guide structure to ensure that the pin can cover the maximum offset path.

[0031] In this embodiment, the standard aperture monitoring control set is a parameter combination consisting of the standard guide clearance and the maximum permissible thrust, including a single guide clearance value and a single maximum permissible thrust value, which are used as the reference control parameters for the aperture monitoring process.

[0032] It should be noted that this embodiment extracts a representative thrust level based on the penetration process with minimal thrust fluctuation, and determines the guide clearance by combining trajectory envelope analysis. This achieves the simultaneous determination of guide constraint strength and propulsion capability, enabling the aperture monitoring process to reflect the problem of insufficient true aperture while avoiding misjudgment caused by excessive constraint.

[0033] S2, in the two-dimensional parameter space of the standard aperture monitoring and control set, several combinations of guide clearance and maximum permissible propulsion force are generated by Latin hypercube sampling to obtain the second aperture monitoring and control set.

[0034] In this embodiment, the second aperture monitoring and control set is obtained by generating several combinations of guide clearances and maximum permissible propulsion forces using Latin hypercube sampling within the two-dimensional parameter space of the standard aperture monitoring and control set, specifically as follows: In the two-dimensional parameter space consisting of the guide clearance and the maximum permissible thrust, the upper and lower bounds of the guide clearance are determined by expanding outwards from the guide clearance value in the standard aperture monitoring and control set, and the upper and lower bounds of the maximum permissible thrust are determined by expanding outwards from the maximum permissible thrust value in the standard aperture monitoring and control set. The upper and lower bounds of the above two dimensions together form the sampling boundary range. The guide clearance is divided into several equal-width intervals between its upper and lower bounds, and the maximum permissible thrust is divided into the same number of equal-width intervals between its upper and lower bounds, so that the number of intervals in the two dimensions is equal. Each interval of the guide gap is randomly paired with each interval of the maximum permissible thrust, so that each interval pair contains one guide gap interval and one maximum permissible thrust interval, and any interval participates in pairing only once in the entire pairing process; In each interval pair, a specific value is extracted from the guide gap interval and the maximum allowable thrust interval according to the same relative position, and the two specific values ​​are combined into a set of sampling points. The sampling points extracted from all interval pairs are summarized to form a set of two-dimensional coordinate points covering the entire sampling boundary range; Remove sampling points from the set of two-dimensional coordinate points that are located outside the edge of the sampling boundary range; Define all remaining sampling points as the set of optimization control parameters.

[0035] In this embodiment, the two-dimensional parameter space is a planar range constructed with guide clearance and maximum permissible propulsion force as the two coordinate axes, used to represent the distribution area of ​​all possible combinations of control parameters.

[0036] In this embodiment, the upper and lower bounds of the values ​​are the minimum and maximum values ​​of each coordinate axis obtained by expanding the guide gap value and the maximum allowable propulsion force value in the standard aperture monitoring and control set to both sides by a preset amplitude. The sampling boundary range is jointly enclosed by the upper and lower bounds of the values ​​of each coordinate axis.

[0037] In this embodiment, the interval grid is a number of equal-width sub-intervals that divide the guide gap between its upper and lower bounds, and the same number of equal-width sub-intervals that divide the maximum permissible thrust between its upper and lower bounds. All the sub-intervals in the above two dimensions form a rectangular grid on the two-dimensional plane.

[0038] In this embodiment, pairing and reassembly involves randomly matching each sub-interval on the guide clearance axis with each sub-interval on the maximum permissible thrust axis, so that each guide clearance sub-interval is paired with only one maximum permissible thrust sub-interval, and each sub-interval participates in pairing only once in the entire pairing process.

[0039] In this embodiment, sampling points are extracted by selecting a specific value in each interval pair after pairing, within the guide gap sub-interval and the maximum allowable propulsion sub-interval, respectively, according to the same relative position ratio, and the two specific values ​​constitute a set of sampling points.

[0040] In this embodiment, the two-dimensional coordinate point set is a point set formed by summarizing the sampling points extracted from all interval pairs. This point set represents a specific aperture monitoring and control sampling combination in the two-dimensional parameter space.

[0041] In this embodiment, the boundary constraint range is the effective area enclosed by the guide gap value and the maximum allowable propulsion force value in the standard aperture monitoring and control set, which are extended to both sides by a preset amount.

[0042] In this embodiment, the set of optimized control parameters is the set of control parameter combinations remaining after filtering the set of two-dimensional coordinate points and removing sampling points located outside the boundary constraint range. This set is used to cover different combinations of path constraint strength and propulsion capability.

[0043] It should be noted that this section uses uniform coverage sampling within the two-dimensional space formed by the guide gap and the maximum permissible propulsion force to ensure that different combinations of control parameters can participate in the subsequent detection and evaluation process.

[0044] S3, based on the radial constraint range collected and analyzed by the inner wall trajectory of the sleeve, converts the guide gap into the sleeve adjustment displacement, and determines the output torque based on the relationship between the driving current and the propulsion force response, thereby converting the aperture monitoring and control set into the output torque and the sleeve adjustment displacement.

[0045] In this embodiment, the radial constraint range is analyzed based on the trajectory acquisition of the inner wall of the sleeve, the guide clearance is converted into the sleeve adjustment displacement, and the output torque is determined based on the relationship between the driving current and the propulsion force response. This transforms the aperture monitoring and control set into output torque and sleeve adjustment displacement. Specifically: During the penetration of the guide pin, the radial displacement time series of the inner wall of the sleeve is obtained by synchronous sampling through the displacement sensing unit of the inner wall of the sleeve. The radial displacement time series is resampled and aligned according to the insertion depth. The displacement values ​​corresponding to the maximum radial convergence position and the maximum radial expansion position of the inner wall of the sleeve within the full penetration depth range are extracted respectively. The difference between the two is calculated as the radial variation range, and the radial variation range is used as the actual radial constraint range of the sleeve. The guide gap value corresponding to each sampling point in the optimization control parameter set is compared with the actual radial constraint range of the sleeve. The deviation between the guide gap value and the median value of the radial constraint range is calculated. The deviation is used as the radial adjustment compensation amount of the sleeve, and the compensation amount is converted into the displacement control command of the sleeve radial drive mechanism. The current input sequence and the corresponding propulsion output sequence of the pin drive system during the penetration process are collected, aligned in a time synchronization manner, and a discrete corresponding data set of current value and propulsion value is constructed. The discrete corresponding data group is segmented according to the current value interval. The arithmetic mean of the propulsion values ​​contained in each current interval is calculated as the average propulsion output value corresponding to that current interval, and the correspondence between each current interval and the average propulsion output value is established. In the correspondence, find the minimum current range in which the average thrust output reaches or exceeds the maximum allowable thrust, take the lower boundary of the current range as the current input boundary value, and convert the current input boundary value into a torque output limit command for the drive system. The displacement control command and the torque output limiting command are combined as the actual execution command for each set of control parameters in the set of optimization control parameters, and are used to execute the guide pin feedback aperture monitoring process.

[0046] In this embodiment, the sleeve inner wall displacement sensing unit is a displacement detection device arranged on the inner wall of the guide sleeve or its supporting structure, used to acquire the displacement change of the inner diameter of the sleeve in the radial direction in real time during the pin penetration process.

[0047] In this embodiment, the radial displacement time series is a data set recording the displacement change of the inner wall of the sleeve along the radial direction over time during the pin penetration process.

[0048] In this embodiment, resampling alignment involves remapping the radial displacement time series according to the insertion depth series, so that each radial displacement data point corresponds to a unique insertion depth position.

[0049] In this embodiment, the maximum radial convergence position is the depth at which the inner wall of the sleeve contracts most significantly inward within the full penetration depth range, and the displacement value corresponding to this position is the maximum negative value.

[0050] In this embodiment, the position of maximum radial expansion is the depth at which the inner wall of the sleeve expands outward most significantly within the full penetration depth range, and the displacement value corresponding to this position is the maximum positive value.

[0051] In this embodiment, the radial variation range is the difference between the displacement value corresponding to the maximum radial convergence position and the displacement value corresponding to the maximum radial expansion position.

[0052] In this embodiment, the actual radial constraint range of the sleeve is determined by the radial variation interval, that is, the continuous interval range from the maximum radial convergence displacement value to the maximum radial expansion displacement value.

[0053] In this embodiment, the median value of the radial constraint range is the arithmetic mean of the maximum radial convergence displacement value and the maximum radial expansion displacement value.

[0054] Let R_mid be the midpoint value of the radial constraint range, and let C be the guide gap value corresponding to a certain sampling point in the optimization control parameter set. Then, the radial adjustment compensation amount Δ is calculated by the following formula: Δ=C-R_mid; when Δ is positive, it means that the sleeve needs to expand outward, and when Δ is negative, it means that the sleeve needs to contract inward. The radial adjustment compensation amount is converted into a displacement control command for the sleeve radial drive mechanism.

[0055] In this embodiment, the current input sequence is a data set of current signals received by the actuator that drives the pin movement during the penetration process, which varies with time.

[0056] In this embodiment, the thrust output sequence is a set of axial thrust data received by the pin at the same time point, which corresponds one-to-one with the current input sequence.

[0057] In this embodiment, the discrete corresponding data group is a set of data pairs formed by aligning the current input sequence and the propulsion output sequence in time. Each data group contains a current value and its corresponding propulsion value.

[0058] In this embodiment, the current interval is a number of continuous sub-intervals divided according to the value range of the current input sequence.

[0059] In this embodiment, the average thrust output is the arithmetic mean of all thrust values ​​within each current range.

[0060] In this embodiment, the mapping relationship is a correspondence established between each current range and its corresponding average propulsion output.

[0061] Let the current intervals be arranged in ascending order of current value as I_1, I_2, ..., I_n, and the corresponding average thrust output values ​​be F_1, F_2, ..., F_n. Let the maximum allowable thrust be denoted as F_max. Starting from the first interval, compare the average thrust output value of each interval with F_max in sequence. Find the first interval I_k that satisfies F_k≥F_max, and use the lower boundary current value of this interval I_k as the current input boundary value.

[0062] In this embodiment, the torque output limiting command is a control signal that converts the current input boundary value into a limit on the output capability of the drive system, used to limit the maximum force during the pin advance process.

[0063] In this embodiment, the actual control target set is a combination of control parameters consisting of the sleeve radial adjustment displacement command and the drive system torque output limit command, which directly acts on the guide pin feedback aperture monitoring process.

[0064] It should be noted that this section transforms the guide clearance and maximum permissible thrust from abstract control parameters into specific executable displacement and torque control commands, thereby achieving direct control of path constraint strength and thrust capability. This allows the aperture monitoring process to maintain sensitivity to aperture axis offset while avoiding forced penetration to cover defects.

[0065] S4, transform each second aperture monitoring and control set and apply it to the process of guide pin feedback aperture monitoring to obtain the actual assembly diameter capacity and hole shaft deviation status of the sheet spring lamination hole group.

[0066] In this embodiment, the process of converting and applying each second aperture monitoring and control set to the guide pin feedback aperture monitoring to obtain the actual assembly bore capacity and bore shaft deviation status of the sheet spring lamination hole group is as follows: Each set of control parameters in the optimization control parameter set is applied to the guide pin feedback aperture monitoring process, so that each set of control parameters corresponds to an independent through-detection cycle. Within each detection cycle, the pin penetration process is executed according to the sleeve adjustment displacement and torque output limits corresponding to the set of control parameters, and the pin insertion depth sequence, axial thrust sequence, and sleeve radial displacement sequence are collected simultaneously during the penetration process. The three types of sequences collected within the same detection period are time-synchronized and aligned based on the insertion depth, and then divided into several continuous segments according to the insertion depth. Within each depth segment, the difference between the maximum and minimum thrust is extracted as the thrust variation amplitude, and the difference between the sleeve radial displacement at the start and end positions of the segment is extracted as the radial convergence variation. The thrust variation amplitude and radial convergence variation within the same segment are paired to establish the relationship between thrust and radial constraint variation segment by segment. In all depth segments, segments whose thrust variation exceeds the average thrust variation of the entire depth are selected, and the insertion depth position corresponding to the segment is mapped to the cumulative thickness position of the sheet spring lamination, as candidate lamination positions where insufficient aperture constraint or hole axis offset occurs. The penetration process of each detection cycle is judged as a whole: if the pin reaches the preset termination depth, it is judged as fully penetrated; if the pin stops advancing before reaching the termination depth, it is judged as not fully penetrated; if the pin completes penetration but the advancing force is continuously higher than the average advancing force of the detection cycle in multiple consecutive depth segments, the consecutive segment is recorded as the blocking segment. The results of the penetration judgment of all detection cycles and the candidate layer positions are summarized to form the characteristics of the hole capacity of the sheet spring lamination hole group under the action of each set of control parameters and the hole axis deviation distribution results.

[0067] In this embodiment, the output torque is the maximum driving torque that the pin drive system is allowed to apply during the penetration process. It is obtained by reverse calculation through the mapping relationship between current and propulsion force, and is used to limit the maximum squeezing capacity of the pin when it encounters the aperture constraint.

[0068] In this embodiment, the sleeve adjustment displacement is the actual adjustment amount of the inner wall of the guide sleeve in the radial direction, which is calculated from the guide clearance and is used to control the allowable radial offset space of the pin during the penetration process.

[0069] In this embodiment, the detection cycle is the complete process from the pin entering the stacked hole group to the completion of penetration or termination due to obstruction under the action of a set of control parameters in the optimization control parameter set. All data collected within this cycle belong to the same control condition.

[0070] In this embodiment, within each detection cycle, the pin penetration process is executed according to the sleeve adjustment displacement and torque output limit corresponding to the set of control parameters, and the pin insertion depth sequence, axial thrust sequence, and sleeve radial displacement sequence are collected simultaneously.

[0071] In this embodiment, the three types of sequences collected within the same detection period are time-synchronized and aligned based on the insertion depth, and then divided into several continuous segments at equal intervals.

[0072] In this embodiment, the thrust variation range is the difference between the maximum and minimum thrust values ​​within each depth segment.

[0073] In this embodiment, the radial convergence change of the sleeve is the difference between the radial displacement of the sleeve at the end position and the beginning position within each depth segment.

[0074] In this embodiment, the propulsion force variation amplitude within the same depth segment is paired with the sleeve radial convergence variation to form segment-by-segment propulsion force and radial constraint variation correlation data.

[0075] Suppose there are N depth segments. Let F_j be the magnitude of the thrust variation in the j-th segment, and let F_j be the average magnitude of the thrust variation across all segments throughout the entire depth range. If the j-th segment satisfies F_j>F_mean, then the segment is marked as a candidate abnormal segment; the insertion depth interval corresponding to the candidate abnormal segment is mapped to the cumulative thickness position of the sheet spring lamination to obtain the candidate lamination position.

[0076] In this embodiment, the penetration process of each detection cycle is judged as a whole: if the pin reaches the preset termination depth, it is judged as fully penetrated; if the pin stops advancing before reaching the termination depth, it is judged as not fully penetrated; if the advancing force of the pin is higher than the average advancing force of the detection cycle in three or more consecutive depth segments, the above consecutive segments are recorded as blocking segments.

[0077] In this embodiment, the path capability is determined by a combination of the penetration completion status and the distribution of obstruction sections. Penetration is completed and there are no obstruction sections, which is considered fully penetrable. Penetration is completed but there are obstruction sections, which is considered partially penetrable. Penetration is not completed, which is considered impenetrable.

[0078] In this embodiment, the penetration determination results of all detection cycles are summarized to obtain the penetration state distribution under different combinations of control parameters, which serves as the characterization result of the actual assembly through diameter capability of the sheet spring lamination hole group.

[0079] In this embodiment, the marked candidate layer positions in all detection cycles are statistically analyzed to obtain the frequency of occurrence of each layer position being marked, which is used as the result of the hole axis deviation distribution.

[0080] It should be noted that this section transforms the penetration behavior under different control combinations into penetration state determination and candidate abnormal layer positions, realizing a direct mapping from raw process data to assembly bore diameter capability and hole-shaft deviation state, so that the test results simultaneously meet the needs of assembly judgment and quality diagnosis.

[0081] S5 extracts features from each monitoring process, trains each second aperture monitoring and control set based on the Q-learning algorithm, and outputs the corresponding assembly reachability excitation value.

[0082] In this embodiment, the feature extraction for each monitoring process, the training of each second aperture monitoring and control set based on the Q-learning algorithm, and the output of the corresponding assembly reachability incentive value are specifically as follows: Based on the monitoring results of the penetration process corresponding to each set of optimization control parameters, the insertion depth sequence, propulsion force sequence and sleeve radial displacement sequence are aligned with the insertion depth as the reference according to the detection cycle, and divided into several fixed depth window units according to equal intervals, with each window unit serving as a local penetration state node. Within each local penetration state node, the ratio of the change in thrust between adjacent sampling points to the corresponding depth change is calculated point by point along the depth direction, forming a thrust change gradient sequence; The average value of the thrust difference between all adjacent sampling points within the window is calculated as the mean difference within the window. The number of sampling points where the thrust difference between adjacent sampling points exceeds the mean difference within the window is counted as the number of thrust abrupt changes. Calculate the difference between the end position and the beginning position of the sleeve radial displacement within the window, and use it as the sleeve radial convergence change. The length of the depth interval in which the thrust is continuously higher than the average thrust within the window is counted and taken as the penetration and stagnation depth length. The propulsion force change gradient sequence, the number of propulsion force abrupt changes, the sleeve radial convergence change, and the penetration resistance depth length are combined in sequence to form a node state feature vector. State transition relationships are established between adjacent depth windows. The difference between the node state feature vector of the current window and the node state feature vector of the next adjacent window constitutes a state transition event. The state transition events of all detection cycles are summarized to form a state transition set. The number of two types of transition events is counted from the state transition set as transition features: the first type is the transition event where the number of thrust change events increases and the change in sleeve radial convergence decreases; the second type is the transition event where the absolute value of the thrust change gradient is less than a preset proportional threshold of the average thrust change gradient within the window and the length of the penetration blocking depth increases. The number of the two types of transition features is concatenated to the end of the corresponding node state feature vector to form the global state feature sequence for each detection period. Using the combination of control parameters in the optimal control parameter set as the action identifier, the global state feature sequence of the corresponding detection cycle as the state input, and the degree of penetration completion, the distribution range of hole shaft deviation position, and the length of the blocking depth of the detection cycle as feedback signals, a cumulative update process for the state transition corresponding to each action is constructed. During each update, the feedback signal corresponding to the current action is successively added to the cumulative evaluation value of the action. The cumulative evaluation value obtained after all detection cycles are completed is used as the assembly reachability incentive value corresponding to the action. Repeat the above training process for all combinations of control parameters in the set of optimal control parameters to generate the assembly reachability incentive value sequence corresponding to each set of control parameters.

[0083] In this embodiment, the detection cycle is a complete through-process completed under the action of a set of control parameters in the optimization control parameter set, and all data collected within this cycle belong to the same control condition.

[0084] In this embodiment, the propulsion force sequence and the sleeve radial displacement sequence are synchronized in time based on the insertion depth, so that the three types of sequences have a one-to-one correspondence at the same depth position.

[0085] In this embodiment, the fixed depth window unit divides the entire length into several continuous equal depth intervals according to the insertion depth, with each interval covering a depth range of the same length.

[0086] In this embodiment, the local penetration state node is a feature description unit corresponding to each depth window unit, which includes the propulsion force change, radial constraint change and stagnation behavior features within the window.

[0087] In this embodiment, the thrust change gradient sequence is a sequence formed by calculating the ratio of the difference between adjacent thrust sampling points to the depth difference between those two points along the depth direction within each window. Let the depth values ​​of adjacent depth sampling points within a certain window be d_j and d_{j+1}, and the corresponding thrust values ​​be F_j and F_{j+1}, then the thrust change gradient at that location is v_j = (F_{j+1} - F_j) / (d_{j+1} - d_j). All v_j values ​​are arranged in depth order to form the thrust change gradient sequence.

[0088] In this embodiment, the number of thrust abrupt changes is the number of sampling points within the window where the difference between adjacent thrust sampling points exceeds the average thrust difference of all adjacent sampling points within the window.

[0089] In this embodiment, the radial convergence change of the sleeve is the difference between the radial displacement of the sleeve within the window and the starting position of the window.

[0090] In this embodiment, the penetration blocking depth length is the length of the depth interval within the window where the propulsion force is continuously higher than the average propulsion force of the window.

[0091] In this embodiment, the node state feature vector is a multi-dimensional data vector formed by sequentially splicing together the propulsion change gradient sequence, the number of propulsion abrupt changes, the sleeve radial convergence change, and the penetration resistance depth length.

[0092] In this embodiment, the state transition relationship is the change relationship of the node state feature vector between adjacent depth windows, which is formed by the difference between the node state feature vector of the previous window and the node state feature vector of the next window.

[0093] In this embodiment, the state transition set is a data set obtained by summarizing the state transition relationships of all adjacent windows within all detection cycles.

[0094] In this embodiment, the first type of transition feature is the number of transition events in the state transition set where the number of propulsion mutations increases and the change in sleeve radial convergence decreases.

[0095] In this embodiment, the second type of transfer feature is the number of transfer events in the state transfer set where the absolute value of the thrust change gradient is less than a preset proportional threshold of the average thrust change gradient within the window and the length of the penetration blockage depth increases.

[0096] In this embodiment, the global state feature sequence is an overall feature sequence formed by concatenating each node state feature vector with two types of transition features in sequence, wherein the two types of transition features are concatenated at the end of the node state feature vector.

[0097] In this embodiment, the action index maps each set of control parameters in the optimization control parameter set to a unique identifier number.

[0098] In this embodiment, the state input is a global state feature sequence corresponding to a certain action index.

[0099] In this embodiment, the feedback signal is formed by a combination of three components: the degree of penetration completion, the distribution range of hole shaft deviation position, and the length of the obstruction depth.

[0100] The quantification method for the degree of penetration completion is as follows: if the pin reaches the preset termination depth position, the value is 1; if the pin does not reach the preset termination depth position, the value is 0.

[0101] The quantification method for the distribution range of hole axis deviation is as follows: statistically analyze the lamination positions corresponding to all candidate abnormal segments marked in step four, and calculate their span ratio within the total lamination thickness range. Specifically, divide the difference between the largest and smallest abnormal lamination positions by the total lamination thickness to obtain a normalized value between 0 and 1.

[0102] The quantification method for the retardation depth length is as follows: sum up the depth lengths of all retardation segments marked in step four, and then divide by the total thickness of the laminate to obtain a normalized value between 0 and 1.

[0103] Let A be the quantization value of the degree of penetration completion, B be the quantization value of the distribution range of the hole shaft deviation position, and C be the quantization value of the length of the hindrance depth. Then the feedback signal r = α·A + β·B - γ·C, where α, β, and γ are weighting coefficients that satisfy α + β + γ = 1.

[0104] In this embodiment, the state transition update process involves cumulatively correcting the historical evaluations of an action index after each detection cycle based on the state input and feedback signals corresponding to the current action index. Let r_k be the feedback signal for the k-th detection cycle corresponding to a certain action index, and let r_k be the cumulative evaluation value after the n-th update. .

[0105] In this embodiment, the assembly accessibility excitation value is the cumulative evaluation value obtained after all action indexes have completed all detection cycles. Its value is determined by the degree of penetration completion, the length of the obstruction depth, and the distribution range of the hole axis deviation position.

[0106] In this embodiment, the assembled reachability incentive value sequence is the set of evaluation values ​​obtained after training all combinations of control parameters in the set of optimization control parameters.

[0107] It should be noted that this section transforms the behavior under different control combinations into a unified feature sequence and performs cumulative evaluation, thereby realizing a quantitative mapping from detection data to the merits of control strategies. This enables subsequent screening of the optimal control combination that achieves a balance between assembly accessibility and defect identification capability based on the excitation value.

[0108] S6. Construct a three-dimensional response surface based on the second aperture monitoring and control set and the corresponding assembly reachability excitation value, and analyze and select the optimal aperture monitoring and control set for application to guide pin feedback aperture monitoring.

[0109] In this embodiment, the construction of a three-dimensional response surface based on the second aperture monitoring and control set and the corresponding assembly reachability excitation value, and the analysis and selection of the optimal aperture monitoring and control set for application to guide pin feedback aperture monitoring, specifically involves: Based on each set of control parameters in each optimization control parameter set and its corresponding assembly reachability excitation value, a three-dimensional discrete data point set is constructed with the guide clearance as the first coordinate axis, the maximum allowable propulsion force as the second coordinate axis, and the assembly reachability excitation value as the third coordinate axis. The three-dimensional discrete data points are arranged in order of their values ​​in the two directions of guide clearance and maximum permissible propulsion force to form a three-dimensional data matrix; Interpolation is performed on the unsampled regions between adjacent data points in the three-dimensional data matrix to obtain a continuously distributed set of three-dimensional response surface data. In the three-dimensional response surface data set, for each data point, its adjacent data points in the guide gap direction and its adjacent data points in the maximum allowable propulsion direction are extracted respectively. The excitation value difference between the current data point and each adjacent data point is calculated one by one to obtain the local gradient change sequence of the data point. In the three-dimensional response surface data set, data points whose excitation values ​​are greater than all adjacent data points in both the guide gap direction and the maximum allowable propulsion force direction are selected as the set of local peak points; In the set of local peak points, the local gradient change sequence corresponding to each point is extracted and the sum of the absolute values ​​of all differences in each sequence is calculated as the total gradient change. The data point with the smallest total gradient change is selected as the optimal candidate point. Within the neighborhood of the optimal candidate point, the dispersion of the excitation value of each data point is statistically analyzed, and the data point with the most stable distribution of excitation value in the neighborhood and the largest excitation value is selected as the optimal aperture monitoring and control set. The guide clearance and maximum allowable thrust corresponding to the optimal aperture monitoring and control set are converted into sleeve adjustment displacement and output torque control commands, which are then applied to the guide pin feedback aperture monitoring process.

[0110] In this embodiment, the set of optimal control parameters is a set of discrete control combinations generated in the two-dimensional parameter space of guide clearance and maximum permissible thrust through Latin hypercube sampling. Each control combination has completed penetration detection and reinforcement learning training and obtained the corresponding assembly reachability excitation value.

[0111] In this embodiment, the assembly reachability excitation value is a scalar result obtained by accumulating feedback on each control combination in terms of penetration completion degree, obstruction distribution and hole-axis deviation identification results based on the Q-learning algorithm. It is used to comprehensively measure the effectiveness of the control combination in assembly inspection.

[0112] In this embodiment, the three-dimensional data point set is a spatial point set formed by mapping the guide clearance value of each set of control parameters in the optimization control parameter set as the first dimension coordinate, the maximum allowable propulsion force as the second dimension coordinate, and the corresponding assembly reachability excitation value as the third dimension coordinate, according to a one-to-one correspondence.

[0113] In this embodiment, the three-dimensional data matrix is ​​a two-dimensional array structure formed by sorting the data points in the three-dimensional data point set according to the guide gap value from smallest to largest, and then arranging the data points with the same guide gap value according to the maximum allowable propulsion force from smallest to largest, and filling the matrix row by row in the sorting order. The rows of the matrix correspond to the guide gap sorting position, the columns of the matrix correspond to the maximum allowable propulsion force sorting position, and each element in the matrix is ​​the assembly reachability excitation value of the corresponding data point. A blank position indicates that there is no sampling point at the intersection of the grid.

[0114] In this embodiment, the three-dimensional response surface dataset is a continuous dataset formed by interpolating and filling in the missing positions in the three-dimensional data matrix. For the missing positions between adjacent sampled data points, a linear interpolation method is used to distribute the excitation value increment according to the distance ratio between two known data points along the guide gap direction or the maximum allowable propulsion force direction, so that the completed data shows a continuous changing trend in the global range.

[0115] In this embodiment, the local gradient change sequence is formed by extracting neighboring data points in the guide gap direction and the maximum allowable propulsion direction for each data point in the three-dimensional response surface dataset, calculating the difference between the excitation value of the current data point and the excitation values ​​of each neighboring data point in that direction, and combining the differences in each direction in sequence. Let m be the number of neighboring points of a certain data point in the guide gap direction and n be the number of neighboring points in the maximum allowable propulsion direction, and let Δ1, Δ2, ..., Δ_{m+n} be the difference between the excitation values ​​of the point and each neighboring point. Then the local gradient change sequence of the point is [Δ1, Δ2, ..., Δ_{m+n}].

[0116] In this embodiment, the local peak point set is a set of data points selected from the three-dimensional response surface data set that meets the following conditions: the excitation value of the data point in the guide gap direction is greater than the excitation value of all adjacent points in that direction, and the excitation value in the maximum allowable thrust direction is greater than the excitation value of all adjacent points in that direction.

[0117] In this embodiment, the total gradient change is the sum of the absolute values ​​of the differences in the local gradient change sequence of a data point, used to measure the drastic change in the excitation value at the location of the data point. Let the local gradient change sequence of a data point be [Δ1, Δ2, ..., Δ_{m+n}], then the total gradient change of that data point is G = |Δ1| + |Δ2| + ... + |Δ_{m+n}|.

[0118] In this embodiment, the optimal candidate point is the data point with the smallest total gradient change selected from the set of local peak points.

[0119] In this embodiment, the neighborhood range is a local spatial region centered on the optimal candidate point and extending along the guide gap direction and the maximum allowable propulsion direction by one adjacent interval. This region includes the optimal candidate point itself and its adjacent data points.

[0120] In this embodiment, the variance of the excitation value is the result of statistically analyzing the dispersion of the assembly reachability excitation values ​​of all data points within the neighborhood of the optimal candidate point. Let there be L data points in the neighborhood, with excitation values ​​E1, E2, ..., E_L. First, the average value of the excitation values ​​within the neighborhood is calculated. Then calculate the mean of the squares of the deviations from the average at each point to obtain the variance of the incentive value. The smaller the variance, the more concentrated the distribution of excitation values ​​in the neighborhood, and the higher the stability.

[0121] In this embodiment, the optimal aperture monitoring control set is the combination of control parameters corresponding to the data point with the smallest excitation value variance and the largest self-excitation value selected within the neighborhood of the optimal candidate point.

[0122] In this embodiment, the sleeve adjustment displacement and output torque control command are the actual execution control parameters obtained by reverse conversion of the guide clearance and the maximum allowable thrust in the optimal aperture monitoring and control set through the aforementioned mapping relationship. The guide clearance corresponds to the displacement control amount of the sleeve radial drive mechanism, and the maximum allowable thrust corresponds to the upper limit of the torque output of the drive system.

[0123] It should be noted that this part achieves global modeling of the coupling relationship between guide clearance and maximum permissible propulsion force by uniformly mapping discrete control parameters and reinforcement learning output results to a continuous three-dimensional response space. The optimal control combination is determined by jointly selecting local peak points and neighborhood variance constraints, thereby improving the ability to distinguish between insufficient aperture and aperture offset while ensuring assembly accessibility.

[0124] Example 2: The present invention also includes an aperture monitoring and control optimization system for sheet spring assembly, comprising the following steps.

[0125] The present invention also includes an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform an aperture monitoring and control optimization method for sheet spring assembly.

[0126] The present invention also includes a computer-readable storage medium storing a computer program that, when executed by a processor, implements a method for optimizing aperture monitoring and control in sheet spring assembly.

[0127] In the embodiments provided by this invention, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0128] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0129] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0130] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0131] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0132] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0133] In the embodiments provided in this disclosure, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0134] It should be noted that, in this disclosure, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element limited by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0135] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for monitoring and controlling the aperture of a sheet spring assembly, characterized in that, Includes the following steps: The insertion depth and propulsion force sequence of multiple penetrations of the guide pin are collected. After aligning the depths, the penetration process with the smallest propulsion force fluctuation is selected. The upper quantile value of its propulsion force is taken as the maximum allowable propulsion force. Envelope analysis is performed on the maximum radial offset of the pin trajectory to obtain the guide sleeve constraint range and determine the standard guide clearance, thus forming a standard aperture monitoring and control set. In the two-dimensional parameter space of the standard aperture monitoring and control set, several combinations of guide clearance and maximum allowable propulsion force are generated by Latin hypercube sampling to obtain the second aperture monitoring and control set; Based on the radial constraint range of the inner wall trajectory acquisition and analysis, the guide clearance is converted into the sleeve adjustment displacement, and the output torque is determined based on the relationship between the driving current and the propulsion force response, thereby converting the aperture monitoring and control set into output torque and sleeve adjustment displacement. Each second aperture monitoring and control set is transformed and applied to the process of guide pin feedback aperture monitoring to obtain the actual assembly diameter capacity and hole axis deviation status of the sheet spring lamination hole group. Feature extraction is performed on each monitoring process, and the Q-learning algorithm is used to train each second aperture monitoring and control set to output the corresponding assembly reachability incentive value. A three-dimensional response surface is constructed based on the second aperture monitoring and control set and the corresponding assembly reachability excitation value, and the optimal aperture monitoring and control set is analyzed and screened for application to guide pin feedback aperture monitoring.

2. The aperture monitoring and control optimization method for sheet spring assembly according to claim 1, characterized in that, The sequence of insertion depth and propulsion force of the guide pin through multiple penetrations is collected. After aligning the depths, the penetration process with the smallest propulsion force fluctuation is selected, and its upper quantile value is taken as the maximum allowable propulsion force. Envelope analysis is performed on the maximum radial offset of the pin trajectory to obtain the guide sleeve constraint range and determine the standard guide clearance, forming a standard aperture monitoring and control set, specifically: In the free assembly state of the sheet spring laminations, the insertion depth time series and the corresponding axial thrust time series of the guide pin are collected during multiple penetration processes. Each penetration process is resampled at equal intervals according to the insertion depth, so that all penetration processes form a one-to-one corresponding set of thrust sequence at the same depth sampling point. For the propulsion force sequence set, the discrete amplitude difference of the propulsion force value is calculated at each depth sampling point, and the discrete amplitude difference of each penetration process is accumulated over the entire depth range. The penetration process with the smallest accumulated value is determined as the target penetration process. The thrust force values ​​corresponding to all depth sampling points during the target penetration process are summarized into a unified thrust force sample sequence. After sorting the sample sequence by value, the thrust force value located at the preset cumulative percentage position is selected as the maximum allowable thrust force. Spatial coordinate reconstruction is performed on the pin trajectory data during the target penetration process. The collected two-dimensional position coordinate sequence is synchronously mapped with the insertion depth sequence to establish a three-dimensional spatial trajectory curve with insertion depth as the independent variable and radial offset as the dependent variable. The radial offset value is calculated point by point for the spatial trajectory curve within the full insertion depth range, and the spatial coordinate points where the radial offset changes locally are selected as boundary characterization points. All boundary characterization points are projected onto the same cross-sectional plane, and the minimum envelope circle that can completely cover all projection points is calculated. The radius of the minimum envelope circle is taken as the guide constraint radius. The difference between the guide constraint radius and the outer diameter of the pin is used as the standard guide clearance. The standard guide clearance is combined with the maximum allowable thrust to form a standard aperture monitoring and control set that includes a single guide clearance value and a single maximum allowable thrust value.

3. The aperture monitoring and control optimization method for sheet spring assembly according to claim 2, characterized in that, The second aperture monitoring and control set is obtained by generating several combinations of guide clearance and maximum permissible propulsion force using Latin hypercube sampling within the two-dimensional parameter space of the standard aperture monitoring and control set. Specifically: In the two-dimensional parameter space consisting of the guide clearance and the maximum permissible thrust, the upper and lower bounds of the guide clearance are determined by expanding outwards from the guide clearance value in the standard aperture monitoring and control set to both sides. The upper and lower bounds of the maximum permissible thrust are determined by expanding outwards from the maximum permissible thrust value in the standard aperture monitoring and control set to both sides. The upper and lower bounds together form the sampling boundary range. The guide clearance is divided into several equal-width intervals between its upper and lower bounds, and the maximum permissible thrust is divided into the same number of equal-width intervals between its upper and lower bounds, so that the number of intervals in the two dimensions is equal. Each interval of the guide gap is randomly paired with each interval of the maximum permissible thrust, so that each interval pair contains one guide gap interval and one maximum permissible thrust interval, and any interval participates in pairing only once in the entire pairing process; In each interval pair, a specific value is extracted from the guide gap interval and the maximum allowable thrust interval according to the same relative position, and the two specific values ​​are combined into a set of sampling points. The sampling points extracted from all interval pairs are summarized to form a set of two-dimensional coordinate points covering the entire sampling boundary range; Remove sampling points from the set of two-dimensional coordinate points that are located outside the edge of the sampling boundary range; Define all remaining sampling points as the set of optimization control parameters.

4. The aperture monitoring and control optimization method for sheet spring assembly according to claim 3, characterized in that, The radial constraint range based on the acquisition and analysis of the inner wall trajectory of the sleeve is used to convert the guide clearance into the sleeve adjustment displacement. The output torque is determined based on the relationship between the driving current and the propulsion force response. Thus, the aperture monitoring and control set is transformed into output torque and sleeve adjustment displacement. Specifically: During the penetration of the guide pin, the radial displacement time series of the inner wall of the sleeve is obtained by synchronous sampling through the displacement sensing unit of the inner wall of the sleeve. The radial displacement time series is resampled and aligned according to the insertion depth. The displacement values ​​corresponding to the maximum radial convergence position and the maximum radial expansion position of the inner wall of the sleeve within the full penetration depth range are extracted respectively. The difference between the two is calculated as the radial variation range, and the radial variation range is used as the actual radial constraint range of the sleeve. The guide gap value corresponding to each sampling point in the optimization control parameter set is compared with the actual radial constraint range of the sleeve. The deviation between the guide gap value and the median value of the radial constraint range is calculated. The deviation is used as the radial adjustment compensation amount of the sleeve, and the compensation amount is converted into the displacement control command of the sleeve radial drive mechanism. The current input sequence and the corresponding propulsion output sequence of the pin drive system during the penetration process are collected, aligned in a time synchronization manner, and a discrete corresponding data set of current value and propulsion value is constructed. The discrete corresponding data group is segmented according to the current value interval. The arithmetic mean of the propulsion values ​​contained in each current interval is calculated as the average propulsion output value corresponding to that current interval, and the correspondence between each current interval and the average propulsion output value is established. In the correspondence between each current range and the average thrust output, find the minimum current range where the average thrust output reaches or exceeds the maximum allowable thrust. Use the lower boundary of this current range as the current input boundary value and convert the current input boundary value into a torque output limit command for the drive system. The displacement control command and the torque output limiting command are combined as the actual execution command for each set of control parameters in the set of optimization control parameters, and are used to execute the guide pin feedback aperture monitoring process.

5. The aperture monitoring and control optimization method for sheet spring assembly according to claim 4, characterized in that, The process of converting and applying each second aperture monitoring and control set to the guide pin feedback aperture monitoring, to obtain the actual assembly bore capacity and bore shaft deviation status of the sheet spring lamination hole group, specifically involves: Each set of control parameters in the optimization control parameter set is applied to the guide pin feedback aperture monitoring process, so that each set of control parameters corresponds to an independent through-detection cycle. Within each detection cycle, the pin penetration process is executed according to the sleeve adjustment displacement and torque output limits corresponding to the set of control parameters, and the pin insertion depth sequence, axial thrust sequence, and sleeve radial displacement sequence are collected simultaneously during the penetration process. The three types of sequences collected within the same detection period are time-synchronized and aligned based on the insertion depth, and then divided into several continuous segments according to the insertion depth. Within each depth segment, the difference between the maximum and minimum thrust is extracted as the thrust variation amplitude, and the difference between the sleeve radial displacement at the start and end positions of the segment is extracted as the radial convergence variation. The thrust variation amplitude and radial convergence variation within the same segment are paired to establish the relationship between thrust and radial constraint variation segment by segment. In all depth segments, segments whose thrust variation exceeds the average thrust variation of the entire depth are selected, and the insertion depth position corresponding to the segment is mapped to the cumulative thickness position of the sheet spring lamination, as candidate lamination positions where insufficient aperture constraint or hole axis offset occurs. The penetration process of each detection cycle is judged as a whole: if the pin reaches the preset termination depth, it is judged as fully penetrated; if the pin stops advancing before reaching the termination depth, it is judged as not fully penetrated; if the pin completes penetration but the advancing force is continuously higher than the average advancing force of the detection cycle in multiple consecutive depth segments, the segment in which the advancing force is continuously higher than the average advancing force of the detection cycle is recorded as the blocking segment. The results of the penetration judgment of all detection cycles and the candidate layer positions are summarized to form the characteristics of the hole capacity of the sheet spring lamination hole group under the action of each set of control parameters and the hole axis deviation distribution results.

6. The aperture monitoring and control optimization method for sheet spring assembly according to claim 5, characterized in that, The process involves feature extraction for each monitoring process, training each second aperture monitoring and control set based on the Q-learning algorithm, and outputting the corresponding assembly reachability incentive value. Specifically: Based on the monitoring results of the penetration process corresponding to each set of optimization control parameters, the insertion depth sequence, propulsion force sequence and sleeve radial displacement sequence are aligned with the insertion depth as the reference according to the detection cycle, and divided into several fixed depth window units according to equal intervals, with each window unit serving as a local penetration state node. Within each local penetration state node, the ratio of the change in thrust between adjacent sampling points to the corresponding depth change is calculated point by point along the depth direction, forming a thrust change gradient sequence; The average value of the thrust difference between all adjacent sampling points within the window is calculated as the mean difference within the window. The number of sampling points where the thrust difference between adjacent sampling points exceeds the mean difference within the window is counted as the number of thrust abrupt changes. Calculate the difference between the end position and the beginning position of the sleeve radial displacement within the window, and use it as the sleeve radial convergence change. The length of the depth interval in which the thrust is continuously higher than the average thrust within the window is counted and taken as the penetration and stagnation depth length. The propulsion force change gradient sequence, the number of propulsion force abrupt changes, the sleeve radial convergence change, and the penetration resistance depth length are combined in sequence to form a node state feature vector. State transition relationships are established between adjacent depth windows. The difference between the node state feature vector of the current window and the node state feature vector of the next adjacent window constitutes a state transition event. The state transition events of all detection cycles are summarized to form a state transition set. The number of two types of transition events is counted from the state transition set as transition features: the first type is the transition event where the number of thrust change events increases and the change in sleeve radial convergence decreases; the second type is the transition event where the absolute value of the thrust change gradient is less than a preset proportional threshold of the average thrust change gradient within the window and the length of the penetration blocking depth increases. The number of the two types of transition features is concatenated to the end of the corresponding node state feature vector to form the global state feature sequence for each detection period. Using the combination of control parameters in the optimal control parameter set as the action identifier, the global state feature sequence of the corresponding detection cycle as the state input, and the degree of penetration completion, the distribution range of hole shaft deviation position, and the length of the blocking depth of the detection cycle as feedback signals, a cumulative update process for the state transition corresponding to each action is constructed. During each update, the feedback signal corresponding to the current action is successively added to the cumulative evaluation value of the action. The cumulative evaluation value obtained after all detection cycles are completed is used as the assembly reachability incentive value corresponding to the action. Repeat the above training process for all combinations of control parameters in the set of optimal control parameters to generate the assembly reachability incentive value sequence corresponding to each set of control parameters.

7. The aperture monitoring and control optimization method for sheet spring assembly according to claim 6, characterized in that, The process involves constructing a three-dimensional response surface based on the second aperture monitoring and control set and the corresponding assembly reachability excitation value, and analyzing and selecting the optimal aperture monitoring and control set for application to guide pin feedback aperture monitoring. Specifically: Based on each set of control parameters in each optimization control parameter set and its corresponding assembly reachability excitation value, a three-dimensional discrete data point set is constructed with the guide clearance as the first coordinate axis, the maximum allowable propulsion force as the second coordinate axis, and the assembly reachability excitation value as the third coordinate axis. The three-dimensional discrete data points are arranged in order of their values ​​in the two directions of guide clearance and maximum permissible propulsion force to form a three-dimensional data matrix; Interpolation is performed on the unsampled regions between adjacent data points in the three-dimensional data matrix to obtain a continuously distributed set of three-dimensional response surface data. In the three-dimensional response surface data set, for each data point, its adjacent data points in the guide gap direction and its adjacent data points in the maximum allowable propulsion direction are extracted respectively. The excitation value difference between the current data point and each adjacent data point is calculated one by one to obtain the local gradient change sequence of the data point. In the three-dimensional response surface data set, data points whose excitation values ​​are greater than all adjacent data points in both the guide gap direction and the maximum allowable propulsion force direction are selected as the set of local peak points; In the set of local peak points, the local gradient change sequence corresponding to each point is extracted and the sum of the absolute values ​​of all differences in each sequence is calculated as the total gradient change. The data point with the smallest total gradient change is selected as the optimal candidate point. Within the neighborhood of the optimal candidate point, the dispersion of the excitation value of each data point is statistically analyzed, and the data point with the most stable distribution of excitation value in the neighborhood and the largest excitation value is selected as the optimal aperture monitoring and control set. The guide clearance and maximum allowable thrust corresponding to the optimal aperture monitoring and control set are converted into sleeve adjustment displacement and output torque control commands, which are then applied to the guide pin feedback aperture monitoring process.

8. A system for optimizing aperture monitoring and control using a sheet spring assembly as described in any one of claims 1-7, characterized in that, It includes a through-modeling module, a sampling expansion module, a parameter mapping module, a path detection module, a reinforcement learning module, and an optimal control module; The through-modeling module is used to collect the insertion depth and propulsion force sequence of multiple penetrations of the guide pin. After aligning the depths, the penetration process with the smallest propulsion force fluctuation is selected, and its upper quantile value is taken as the maximum allowable propulsion force. Envelope analysis is performed on the maximum radial offset of the pin trajectory to obtain the guide sleeve constraint range and determine the standard guide clearance, forming a standard aperture monitoring and control set. The sampling extension module is used to generate several combinations of guide clearance and maximum allowable propulsion force in the two-dimensional parameter space of the standard aperture monitoring and control set using Latin hypercube sampling, so as to obtain the second aperture monitoring and control set. The parameter mapping module is used to collect and analyze the radial constraint range based on the trajectory of the inner wall of the sleeve, convert the guide clearance into the sleeve adjustment displacement, and determine the output torque based on the relationship between the driving current and the propulsion force response, thereby converting the aperture monitoring and control set into output torque and sleeve adjustment displacement. The bore detection module is used to convert and apply each second bore monitoring and control set to the process of guide pin feedback bore monitoring to obtain the actual assembly bore capacity and bore shaft deviation status of the sheet spring lamination bore group. The reinforcement learning module is used to extract features from each monitoring process, train each second aperture monitoring and control set based on the Q-learning algorithm, and output the corresponding assembly reachability incentive value. The optimal control module is used to construct a three-dimensional response surface based on the second aperture monitoring control set and the corresponding assembly reachability excitation value, and to analyze and screen the optimal aperture monitoring control set for application to guide pin feedback aperture monitoring.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the aperture monitoring and control optimization method for sheet spring assembly as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the aperture monitoring and control optimization method for sheet spring assembly as described in any one of claims 1 to 7.

Images