Crankshaft conveying and framing beat control method based on multi-station cooperation

Abstract

The invention discloses a crankshaft conveying and framing beat control method based on multi-station cooperation, and relates to the technical field of mechanical manufacturing automation control, and the method comprises the following steps: collecting the action time of each station, the signal arrival time and the stay time, building a unified time axis, marking the peak value and the valley value of the action curve of each station, and generating a beat phase band; the time rhythm is used for representing whole-line time rhythm; and aligning the starting point and the ending point of the action of each station one by one according to the beat phase belt, identifying the abutting area and the neutral gear area between the adjacent stations, and generating a delay fingerprint list for recording the time sequence deviation characteristics between the stations. According to the method, multi-station time sequence coordination is achieved through a unified time axis and a beat phase belt, the station rhythm is dynamically adjusted through delayed fingerprints, phase dislocation indexes and a beat slip list, beat conflicts and oscillation are eliminated through double-track instruction issuing and an inverse-phase breathing window, the beat error is reduced by 60%, the synchronization precision reaches the millisecond level, and the production stability is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical manufacturing automation control, in particular to a crankshaft conveying and framing beat control method based on multi-station cooperation. BACKGROUND

[0002] The crankshaft conveying and framing beat control based on multi-station cooperation refers to the beat control of the operation rhythm of multiple stations such as material moving, angle correction, pneumatic measurement, visual detection, classification framing, and empty frame circulating feeding in the crankshaft automatic detection and boxing equipment, so as to realize the beat coordination control of the whole line operation. The method takes the beat control system as the core, matches the action time, detection period, and conveying rhythm of each station in time sequence, so that the crankshaft keeps continuous circulation without waiting and stagnation from conveying to framing. The system dynamically adjusts the conveying rhythm and framing instruction by real-time acquisition of the state signals of each station, so as to ensure the accurate synchronization of NG and OK crankshafts, material frame switching, and stacking actions, so as to realize the cooperation and balance between high-precision detection and high-efficiency framing, and significantly improve the automation degree and beat stability of the production line.

[0003] The prior art has the following disadvantages: In the prior art, the multi-station cooperative conveying and framing system generally adopts fixed beat and linear feedback mechanism for control. When the action of any station is delayed or the detection signal is offset, the system usually recovers the beat synchronization through adaptive path re-planning. However, under the condition of high-speed beat operation, if the transmission delay of the feedback signal exists in microseconds, and the delay time is just offset with the beat adjustment period, the control logic will appear phase offset superposition, resulting in continuous misplacement of the beat correction command and the actual state feedback. This misplacement is repeatedly accumulated in the control loop, which is easy to form closed loop oscillation, so that the program enters the continuous correction state and cannot converge, finally causing the action conflict of the station, the disorder of the conveying path, and even the whole line shutdown.

[0004] The above information disclosed in the background section is only used to strengthen the understanding of the background of the present disclosure, so it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The purpose of the present application is to provide a crankshaft conveying and framing beat control method based on multi-station cooperation to solve the problems in the background.

[0006] In order to achieve the above purpose, the present application provides the following technical scheme: a crankshaft conveying and framing beat control method based on multi-station cooperation, comprising the following steps: Collect the action time, signal arrival time and dwell time of each station, establish a unified time axis, mark the peak and valley values ​​of the action curve of each station, and generate a beat phase band to characterize the time rhythm of the entire line. Align the start and end points of each workstation's actions one by one according to the beat phase band, identify the overlapping and gap areas between adjacent workstations, and generate a delay fingerprint list to record the timing deviation characteristics between each workstation. Based on the delayed fingerprint list, an avoidance window is set for each workstation, and a fixed micro gap is inserted in front of the conflicting area to form a phase misalignment index list, which is used to establish the anti-collision distribution of action time; The sequence of workstation actions is rearranged based on the phase misalignment index list. A forward shift is applied to workstations with dense actions, and a backward shift is applied to workstations with lagging actions. A cycle time slip list is generated to balance the cycle time of the workstations. A signal distribution gate is established based on the rhythm sliding list. Fast response commands are directly output, and response delay commands are input into the buffer zone to generate a dual-track command list for distinguishing signal transmission paths. Based on the dual-track command list, a reverse breathing window is established at the end of each beat cycle. A short tightening operation is performed first, followed by a slow relaxation operation to eliminate beat conflicts and oscillations. The correction results are then written back to the beat phase band.

[0007] Preferably, the steps for generating the beat phase band are as follows: Collect the action time, signal arrival time and action dwell time of each workstation, and establish a unified time axis with the initialization of the entire line as the zero point; Draw the motion line segment and signal arrival position for each workstation on a unified timeline, and label the motion duration and dwell time; The peak and valley values ​​of the action segments of each workstation on the time axis are calibrated, and time overlap and gap intervals are distinguished by color. The beat phase band is constructed based on the peak-valley distribution results and the time overlap ratio, and is used for beat rearrangement, phase avoidance and beat slip in subsequent steps.

[0008] Preferably, the delayed fingerprint list generation steps are as follows: Extract the start and end times of the action of each workstation in the beat phase band on a unified time axis; Compare the action times of the immediate preceding and following workstations in each beat cycle to identify time overlap areas and time gap areas. Generate a list of delayed fingerprints based on the overlap length or gap length, and record the time overlap length, time gap length, action start offset value, and action end offset value; The data in the delayed fingerprint list is categorized and archived, and the corresponding and gap areas are divided according to the length range.

[0009] Preferably, the process for forming the misphase index list is as follows: Based on the delayed fingerprint list, identify workstation pairs with overlapping times and set an avoidance window to postpone the start time of the action of the later workstation; Insert a micro-gap with a length greater than the time overlap length into the avoidance window, and record the adjusted start time and end time of the action; Record the original start time, micro-gap length, adjusted time, and cycle number as a misphase index list; The list of misaligned indexes is validated to ensure that the time sequence is reasonable, no new overlapping conflicts are caused, and the cycle period limit is not exceeded.

[0010] Preferably, the process for generating the beat slip list is as follows: Extract the adjusted start time and end time of each workstation from the misphase index list, and calculate the workstation motion density within each cycle. Apply a forward movement to workstations in the motion-intensive section and readjust the motion time segment; Apply a backward shift to workstations in sparse motion zones and assess the remaining space in the cycle time. All the movement adjustment information after moving forward and backward is compiled into a beat sliding list, and the sliding time and corresponding beat cycle number are recorded.

[0011] Preferably, the forward and backward shifts are set based on the motion density level. The forward shift is controlled to prevent time overlap in the motion-dense section, and the backward shift is controlled within the remaining time of the beat cycle. After each adjustment, a response buffer time is reserved to absorb instruction transmission delay.

[0012] Preferably, the dual-track command list generation process is as follows: Extract the start time of the action for each station in the cycle slide list, and calculate the average response time and maximum offset time of the control commands; Based on response time, control commands are divided into fast-response commands and delayed-response commands. Set a direct output path for fast response commands and a path that includes a buffer for delayed response commands; Set the buffer dwell time and calculate the advance issuance time based on the target time of the instruction; All control information is aggregated to generate a dual-track command list, and structural consistency and timing conflict checks are performed.

[0013] Preferably, based on the dual-track command list, a reverse breathing window is established at the end of each beat cycle, and a short tightening operation and a slow loosening operation are performed. The steps for writing the correction results back to the beat phase band are as follows: Extract the action end time of each workstation in the dual-track command list within each cycle, and perform difference analysis with the cycle end time to identify the adjustable time segment at the end of the cycle. Within the identified time interval, set up a reverse breathing window and divide it into a short tightening zone and a slow relaxation zone; Within the short, tight zone, move the workstations with later completion times forward and recalculate the motion rhythm; Within the gradually relaxed zone, the end time of some workstations will be extended according to the fluctuations in movement, while maintaining a time-cleaning zone; The time adjustment data for tightening and loosening is written back to the beat phase band and convergence and stability are verified.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention establishes a unified timeline and cycle phase band to perform time-series quantitative management of the action data of each workstation, which were originally operating independently. Combined with a delay fingerprint list and a phase misalignment index list, it accurately identifies and dynamically avoids time overlaps, gaps, and delays between multiple workstations, achieving overall coordination and real-time correction of the rhythm of multiple workstation actions. By using a cycle sliding list to adjust the sequence of workstation actions forward and backward, the overall rhythm tends to be balanced, avoiding localized cycle congestion. Simultaneously, a dual-track command list manages rapid response commands and delayed commands separately, eliminating execution deviations caused by signal interference and command stacking. Finally, a reverse-phase breathing window is set at the end of each cycle, utilizing a short tightening and slow widening self-adjustment process to actively eliminate cycle misalignment and closed-loop oscillations, achieving dynamic convergence and continuous stable operation of the entire line's cycle. Compared to existing control methods that rely on fixed cycle time and linear feedback mechanisms, this invention reduces cycle time error by more than 60%, improves multi-station synchronization accuracy to the millisecond level, significantly enhances the automation level, cycle time stability, and equipment uptime of the production line, and effectively solves the problem of instability in multi-station collaborative systems under high-speed cycle time conditions. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the crankshaft conveyor frame assembly cycle control method based on multi-station collaboration of the present invention. Detailed Implementation

[0017] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0018] This invention provides, for example Figure 1 The crankshaft conveyor frame assembly cycle control method shown includes the following steps: Collect the action time, signal arrival time and dwell time of each station, establish a unified time axis, mark the peak and valley values ​​of the action curve of each station, and generate a beat phase band to characterize the time rhythm of the entire line. To achieve rhythmic control of multi-station collaborative operation during crankshaft conveying and framing, a unified timeline for the entire production line needs to be established, and the motion patterns and signal behaviors of each station need to be clearly defined. Therefore, it is essential to accurately collect the station action time, signal arrival time, and action dwell time, and process them sequentially to form a rhythmic phase band. The detailed implementation steps are as follows: Data acquisition was performed on each workstation involved in cycle time coordination along the entire conveyor and framing production line. Taking a production line with six key workstations as an example: material transfer station, angle correction station, pneumatic measurement station, vision inspection station, sorting and framing station, and empty frame circulating loading station. Signal triggers were installed at the main execution points of each workstation to detect the start time of the action; position sensors were installed at the end points of each action to calibrate the completion time; and time recording nodes were set in the signal path to accurately record the time taken for the signal to travel from the trigger source to the next level response mechanism. For example, at the angle correction station, the recording showed that the total time from start to completion was 0.76 seconds, with a signal delay of 0.06 seconds and a dwell time of 0.41 seconds. At the vision inspection station, the total action time was 0.93 seconds, with a signal transmission time of 0.03 seconds and a dwell time of 0.58 seconds. The above time data is organized sequentially to ensure that three key time types are collected for each workstation: action start time, action completion time, and signal transmission delay, forming a basic time dataset.

[0019] The aforementioned time data is mapped to a common timeline using a unified reference point. This timeline is initialized with the entire production line as its zero point, and the complete running time of each production cycle is set as a cycle boundary, for example, 3.5 seconds. Using this timeline as the horizontal axis, the action periods of each workstation are plotted as continuous line segments, with signal arrival times marked as vertical pointer lines. The action duration and dwell time are respectively marked above and below the timeline. For example, at the start of the timeline at 2.1 seconds, the material handling station begins operation, lasting 0.42 seconds, with the signal arrival point at 2.13 seconds; at 2.2 seconds, the vision inspection station signal is triggered, lasting 0.93 seconds, with signal transmission of 0.03 seconds, and the signal response point at 2.23 seconds. By distributing and plotting these specific data, it is possible not only to observe whether there is time overlap between the actions of each workstation, but also to further infer whether the signal response has sufficient buffer time, providing an accurate basis for subsequent timing judgments.

[0020] After establishing a unified timeline, curve analysis was performed on the motion segments of each workstation on the timeline to clearly identify peak and trough values. A peak value is defined as the highest duration of an action in the historical records of that workstation, and a trough value is the shortest interval between two actions. Taking the angle correction workstation as an example, its average action time is 0.52 seconds, while within a certain cycle, the collected action time reaches 0.83 seconds, which is its peak value. In the empty frame cyclic loading workstation, the interval between two adjacent actions is only 0.13 seconds; this time period is marked as the trough area. To ensure the accuracy of data analysis, five sets of continuous cycle data were used for comparative sampling to exclude occasional anomalies. The peak and trough values ​​of each workstation are marked with red and blue dots on the unified timeline to facilitate the identification of high-conflict areas and densely overlapping time segments.

[0021] Based on the peak-valley distribution results, an overlap analysis of the time coverage between adjacent workstations is performed on a unified time axis. According to the time overlap ratio and cumulative duration distribution, multiple cycle phase intervals are divided. These are specifically color-coded: areas with a time overlap exceeding 0.25 seconds are marked in red, indicating potential phase conflict zones; areas with a time gap exceeding 0.5 seconds are marked in green, indicating cycle idle zones; and continuous non-overlapping segments are marked in yellow, indicating cycle safety buffer zones. For example, if a 0.33-second time overlap is found between the inspection station and the framing station in the sixth cycle, this segment is marked as a red phase conflict zone; and a 0.62-second gap is observed between the loading station and the transfer station, which is marked as a green cycle idle zone. Finally, these cycle phase intervals are integrated to form cycle phase bands, with each time interval including its start point, end point, duration, and corresponding workstation number, stored in tabular form. This provides accurate timing data for subsequent cycle rearrangement, phase avoidance, and cycle sliding steps.

[0022] Align the start and end points of each workstation's actions one by one according to the beat phase band, identify the overlapping and gap areas between adjacent workstations, and generate a delay fingerprint list to record the timing deviation characteristics between each workstation. To accurately identify time conflicts and downtime periods between multiple stations during crankshaft conveying and framing, based on the construction of cycle phase bands, action sequence alignment analysis is performed, and conflicting and downtime regions are extracted pair by pair, ultimately forming a list of delay fingerprints for quantifying control error characteristics. The specific operation includes the following steps: Extract the action time segment of each workstation on a unified time axis within the beat phase band, and unfold its start and end times one by one, matching them sequentially with the corresponding beat cycle according to workstation number. Within each cycle, use the workstation's start time as a time reference point, and compare it with the end time of the immediately preceding workstation within the same beat, calculating the time difference between the two. Taking the first cycle as an example, the angle correction workstation's action starts at 1.26 seconds, the vision inspection workstation's action starts at 1.92 seconds, and the angle correction workstation's end time is 2.04 seconds. Therefore, the vision inspection workstation's start time is 0.12 seconds earlier than the completion time of the preceding workstation's action, constituting an explicit time overlap. This process requires a full analysis of the temporal dependencies between all workstations within each beat cycle to ensure coverage of all potentially conflicting workstation pair combinations.

[0023] Based on the alignment analysis results, the time relationships between all workstation pairs are divided into two categories: regions with overlapping time periods and regions with time gaps. For overlapping time periods, a negative value is obtained by subtracting the ending time of the preceding workstation from the start time of the subsequent workstation's action; the overlap length is the absolute value of this negative value. For time gaps, a positive time difference is considered a time gap region, and its length is the positive value. Taking the second cycle as an example, the pneumatic measurement workstation ends at 1.50 seconds, and the visual inspection workstation begins at 1.68 seconds, with a 0.18-second interval, forming a gap region. In practical operation, the timing judgment of each workstation pair should be combined with observations from multiple consecutive cycle periods to avoid misjudging the time trend due to a single fluctuation. It is generally recommended to use the time series of five consecutive cycle periods as the smallest unit of analysis for summary statistics.

[0024] After completing the identification of overlaps and gaps in all workstation pairs, a delay fingerprint list is established. The list is constructed following a cycle time arrangement with the cycle period on the horizontal axis and the workstation pair number on the vertical axis. Detailed records are kept of the time overlap length, time gap length, action start offset, and action end offset within each cycle. Taking the analysis data of the visual inspection workstation and the framing workstation over five cycle periods as an example, if there are more than three overlaps exceeding 0.15 seconds, they are classified as moderately conflicting workstation pairs and recorded as having continuous offsetting characteristics. Simultaneously, the frequency, maximum, minimum, and average values ​​of various conflicts or gaps should be recorded in the list to assess the overall stability of the workstation pair's temporal fluctuations. This list can be created manually or exported as a data table from the temporal identification process. The key is the accuracy, continuity, and detail of the data collection to ensure sufficient temporal resolution to capture minute delays.

[0025] Each data point in the delay fingerprint list is categorized and archived, and classified into levels according to the length range of overlapping or gap regions. Workstation pairs with significant time conflict characteristics are specially marked. The classification criteria can be set as follows: overlap length less than 0.05 seconds is a mild conflict region, 0.05 to 0.15 seconds is a moderate conflict region, and more than 0.15 seconds is a severe conflict region; gap length less than 0.10 seconds is a short gap region, and more than 0.10 seconds is a long gap region. After classification, a forced avoidance strategy can be configured for severe conflict regions, and a buffer correction action can be embedded for long gap regions, serving as the parameter input source for subsequent phase misalignment scheduling and slippage processing.

[0026] Based on the delayed fingerprint list, an avoidance window is set for each workstation, and a fixed micro gap is inserted in front of the conflicting area to form a phase misalignment index list, which is used to establish the anti-collision distribution of action time; To avoid execution conflicts caused by time overlap between multi-station actions, based on the identification of conflicting areas in the delay fingerprint list, avoidance time is introduced for subsequent workstations, and a misalignment index list is constructed by inserting specific micro-gaps to achieve time decoupling of action sequence. The specific detailed steps include the following: For workstation pairs marked with an overlap length exceeding 0.08 seconds in the delayed fingerprint list, their start and end times are read one by one, and they are grouped and processed according to each cycle. Taking the second cycle as an example, the angle correction workstation's action starts at 1.32 seconds and ends at 2.05 seconds; the vision inspection workstation's action starts at 1.98 seconds in this cycle, resulting in a 0.07-second overlap. Although the overlap length is slightly below the warning value, combined with the subsequent framing workstation's start time of 2.84 seconds, the interval is insufficient to form a stable buffer. Therefore, this overlap is still included in the avoidance adjustment scope. Identified workstation pairs should not be merged by average, but should be analyzed independently cycle by cycle to ensure that all occasional conflicts are handled.

[0027] For each pair of workstations with overlapping time intervals, an avoidance window is set. This window is inserted as a micro-gap before the start time of the subsequent workstation's action, and its length must be greater than the detected overlap duration. Taking the material handling station and the vision inspection station as an example, if an overlap length of 0.11 seconds is found in the fourth cycle, to avoid phase shift caused by error accumulation, the micro-gap is set to 0.15 seconds, that is, the original start time of the vision inspection station is delayed from 2.31 seconds to 2.46 seconds. This micro-gap length must not be less than 0.12 seconds or more than 0.18 seconds to ensure scheduling flexibility and cycle rhythm consistency. At the same time, a time of 0.02 seconds should be reserved as a signal response hysteresis margin to avoid compensation failure caused by fluctuations in command transmission time.

[0028] After inserting the avoidance window, the start and end times of the subsequent workstation's actions are recalculated, and the adjusted time information is recorded to form a phase misalignment index list. The list should include the original action start time, micro-gap length, adjusted action start time, adjusted action end time, workstation number, cycle number, and source of conflict. For example, for the sorting and framing workstation, if its original action time period was 2.84 to 3.52 seconds, and the start time changes to 2.94 seconds after the avoidance window adjustment, its action end time is delayed to 3.62 seconds, with a delay recorded as 0.10 seconds. All records are summarized and organized into a phase misalignment index list according to cycle number, and then sorted by workstation number from smallest to largest. This index table will provide a high-precision reference for subsequent action rearrangement and sliding adjustments.

[0029] The phase misalignment index list is verified for completeness and feasibility to ensure that the start and end times of the adjusted actions do not exceed the cycle limit and do not create new overlap risks with other non-conflicting workstations. If new conflicts arise in the third cycle due to forward or backward misalignment, such as a less than 0.06-second interval between the adjusted visual inspection workstation start time of 2.46 seconds and the new start time of the framing workstation of 2.94 seconds, the micro-gap needs to be reallocated, or a reserved buffer segment needs to be introduced at the upstream workstation. Simultaneously, each adjusted action should be checked to ensure that its end time exceeds the cycle endpoint; for example, when the cycle limit is 3.60 seconds, no action should extend beyond that time point. Through the above verification methods, the phase misalignment index list can be ensured to have three basic characteristics: clear action time sequence, reasonable micro-gap insertion, and complete elimination of overlap, significantly improving the timing stability of multi-workstation action coordination.

[0030] The sequence of workstation actions is rearranged based on the phase misalignment index list. A forward shift is applied to workstations with dense actions, and a backward shift is applied to workstations with lagging actions. A cycle time slip list is generated to balance the cycle time of the workstations. After completing the phase misalignment index list and inserting micro-gap adjustments, it is necessary to optimize the time cycle distribution of the entire line to avoid excessively tight local cycles or excessively long gaps. To achieve a balanced and coordinated operating state, based on the start and end times of the actions recorded in the phase misalignment index list, the regions with dense and sparse actions within the cycle should be identified. Then, by shifting the start time of the actions forward or backward, the overall sequence of actions at each workstation should be rearranged to finally generate a cycle slip list. This process includes the following steps: Extract the adjusted start and end times of all workstations from the misalignment index list, and calculate the number of overlapping workstation actions per unit time in each cycle. Taking the fourth cycle as an example, if the actions of three workstations—material handling, pneumatic measurement, and visual inspection—overlap simultaneously within the time interval of 2.40 seconds to 2.90 seconds, and the total number of workstations covered is more than 3 out of 6, then this segment can be identified as an action-dense area. Conversely, if only one workstation (such as a sorting and framing workstation) is active within the time interval of 3.00 seconds to 3.50 seconds, and there is no overlap of actions from other workstations before and after it, then this segment is identified as an action-sparse area. In the determination process, sampling or mean estimation methods are not used; instead, point-by-point accumulation statistics are performed every 0.1-second time interval within each cycle to ensure the completeness and continuity of density identification.

[0031] Based on the density analysis results, the actions of workstations in the dense action zone are moved forward. The forward movement should adhere to three principles: no overlap in time, no reversal of sequence, and no conflict in response. For example, if the original start time of the material handling station in the fourth cycle is 2.43 seconds and the end time is 2.84 seconds, and it overlaps with the vision inspection station by 0.31 seconds within that time period, then the start time of the station's action is moved forward to 2.28 seconds, and the end time is simultaneously moved forward to 2.69 seconds, thus removing it from the overlapping area with other workstation actions. The magnitude of the forward movement is set according to the density level, generally controlled between 0.10 and 0.20 seconds to avoid new conflicts or signal response mismatches caused by excessive forward movement. Simultaneously, a response buffer of at least 0.04 seconds should be maintained to absorb any delays that may occur between command issuance and machine start-up.

[0032] For workstation actions in sparsely populated areas, a shift is applied to fully utilize the remaining cycle time and achieve a balanced distribution of rhythm across the entire line. Taking the sorting and framing workstation as an example, if its original start time is 3.01 seconds and end time is 3.55 seconds, and the total cycle time is 3.60 seconds, this workstation has a 0.05-second usable tail segment. Furthermore, since its initial action does not overlap with other workstations, its entire action is shifted back by 0.10 seconds, changing the start time to 3.11 seconds and the end time to 3.65 seconds. This shift operation precisely expands the utilization efficiency of sparse areas, not only avoiding wasted cycle time resources but also freeing up more safe operating space for other workstations to move forward. In addition, the shift operation also needs to assess whether the remaining space in the current cycle time is sufficient. If insufficient, some of the shifting actions need to be split and processed in the initial segment of the next cycle.

[0033] All workstation actions that have undergone forward and backward repositioning are rearranged and compiled into a rhythmic sliding list. The list should include the workstation number, original action start time, adjusted action start time, adjustment direction (forward or backward), sliding time, new action end time, and corresponding rhythmic cycle number. For example, in the fifth rhythmic cycle, the original start time of the visual inspection workstation was 1.92 seconds, which was adjusted to 1.78 seconds after forward repositioning, with a sliding time of 0.14 seconds, resulting in a new end time of 2.67 seconds. This record should be clearly marked as a forward repositioning adjustment. This list, once completed, will serve as the timing input for signal scheduling and rhythmic breathing strategies, ensuring that each workstation operates as closely as possible to the rhythmic beat without conflict.

[0034] A signal distribution gate is established based on the rhythm sliding list. Fast response commands are directly output, and response delay commands are input into the buffer zone to generate a dual-track command list for distinguishing signal transmission paths. After adjusting the workstation's cycle time, to ensure that the timing of control command output perfectly matches the actual rhythm of the workstation's movements, a signal distribution structure tailored to different response characteristics needs to be established. Control commands are categorized into fast output and buffered output based on their response speed differences, and processed separately via dual-track paths to generate a dual-track command list, achieving reasonable signal path distribution and synchronous control. The specific steps are as follows: The start time of actions at all stations in the cycle time sliding list is extracted, and combined with the historical response time of each control command, the average response time and maximum offset time required between the command being issued from the master control terminal and the start of the action at the target station are calculated. During the statistical process, specific actions at each station are independently evaluated using actual operating data. For example, the average time from issuing the rotation command to the start of the action at the angle correction station is 0.022 seconds, with a maximum fluctuation of 0.004 seconds; while the response time for the push command at the empty frame loading station is 0.066 seconds, with a maximum fluctuation of 0.014 seconds. Based on this evaluation standard, control commands with a response time not exceeding 0.030 seconds and a fluctuation not exceeding 0.005 seconds are classified as fast-response commands, while commands with a response time exceeding 0.030 seconds or a fluctuation exceeding 0.005 seconds are classified as delayed-response commands. This classification is based on actual execution characteristics, avoiding the problem of traditional control systems that only make rough classifications based on station type.

[0035] Based on the command response classification results, different issuance paths are set for fast-response commands and delayed-response commands. Fast-response commands use a direct output method; at the action start time marked in the cycle time list, the control issuing end directly outputs the signal, which arrives at the workstation immediately without delay. For example, in the sixth cycle, the target start time for the image acquisition command at the vision inspection station is 2.34 seconds, and its control signal will be issued at the exact 2.34-second mark, quickly triggering the action start. Delayed-response commands, on the other hand, have a buffer inserted in the transmission path. Their control signal will be issued earlier and remain in the buffer for a period before reaching the execution end. For example, in the fifth cycle, the target start time for the material transfer station action is 1.87 seconds, and the average response time of the control command is 0.062 seconds. The control end will issue the signal at 1.808 seconds in advance, and the signal will remain in the buffer for 0.062 seconds before arriving at the execution end at 1.87 seconds, ensuring precise alignment of the action timing.

[0036] For all delayed response commands, the dwell time of the buffer band is set individually based on their response characteristics. The dwell time consists of the average response time plus the redundancy protection time, which is set according to the actual fluctuation range, generally not less than 0.008 seconds. For example, in an empty frame supply station, the average response time of the transmission command is 0.071 seconds, and the maximum deviation is 0.012 seconds. To prevent commands from failing due to early or delayed execution, the dwell time of the buffer band is set to 0.083 seconds. The control signal will be injected into the buffer band according to this dwell time and will be transmitted exactly before the target action time. During implementation, it must be ensured that each buffer path signal corresponds to only one control command to prevent overlapping dwell times that could lead to triggering errors. The entry time of the buffer band is precisely determined by the cycle slip list, and the exit time must be exactly consistent with the start time of the station action, with an error not exceeding 0.005 seconds, to ensure execution accuracy.

[0037] Based on all the control commands for the established paths, compile and output a complete dual-track command list. This list should include the following: workstation number, cycle number, control command number, target action name, target action start time, control command issuance time, response time, whether a buffer zone is present, buffer zone dwell time, and command transmission path type. For example, in the seventh cycle, the control command number for the angle correction workstation is CMD013, the target action is spindle rotation start, the target start time is 2.15 seconds, the control signal is directly output at 2.15 seconds, and the path type is fast channel; while in the same cycle, the target action for the sorting and framing workstation control command CMD028 is gripping and descending, the target start time is 3.02 seconds, the command issuance time is 2.938 seconds, the buffer zone dwell time is 0.082 seconds, and the path type is buffer channel. This list provides a precise basis for subsequent cycle closed-loop control, signal phase correction, and dynamic adjustment.

[0038] The generated dual-track command list undergoes structural consistency verification and timing conflict analysis. First, it is ensured that all control commands are included in the scheduling scope without omission, and that each command is correctly classified and assigned a transmission path. Second, the buffer band setting time is verified to match the control response time, confirming that the alignment accuracy between the advance command issuance time and the actual action time of the workstation is within the allowable range. Third, the concentrated issuance of all rapid commands within the same cycle time is checked to avoid command interference caused by an excessive number of signals in the same time period; if necessary, the issuance time of some rapid commands is adjusted to 0.003 seconds before the target action time. Finally, timing simulation is performed on the command distribution of all workstations on the entire production line over 10 consecutive cycle times to ensure that there are no abnormalities such as cross-triggered events, delayed responses, or signal failures.

[0039] According to the dual-track command list, a reverse breathing window is established at the end of each beat cycle. A short tightening operation is performed first, followed by a slow relaxation operation to eliminate beat conflicts and oscillations. The correction results are then written back to the beat phase band. After completing the arrangement of the dual-track command list and the allocation of execution paths, in order to avoid the accumulation and loss of control of beat errors at the cycle boundaries, a time adjustment structure with compression and release capabilities must be constructed at the end of each beat cycle. By setting a reverse-phase breathing window, the orderly compression and slow release of the action rhythm can be achieved, thereby eliminating local beat conflicts and synchronously writing back to the beat phase band. The specific steps include: Based on the generated dual-track command list, the end time of each workstation's action within each cycle is extracted and compared with the theoretical end time of the current cycle to identify the remaining adjustable time segment at the end of the cycle. Taking the eighth cycle as an example, the theoretical end time of this cycle is 3.60 seconds. The end time of the material transfer station in this cycle is 3.43 seconds, the angle correction station is 3.51 seconds, and the framing station is 3.58 seconds. There is a significant difference between these three at the end of the cycle. After comparison, it was found that the framing station has only 0.02 seconds left before the end of the cycle, indicating a risk of critical action backlog. Therefore, it is necessary to set a dynamic adjustment segment of the cycle boundary, i.e., a reverse breathing window, based on the latest end time of all workstations, and reserve a boundary compressible time of not less than 0.06 seconds.

[0040] Two functional time periods are divided within the identified cycle end window. The first functional time period is designated as a short, tightening zone, used to compress workstations with dense overlapping actions and a high risk of delay. Specifically, workstations with later action end times are moved forward. The compression amount is dynamically calculated based on the distance between the workstation and the cycle end, generally controlled between 0.01 and 0.03 seconds. For example, in the ninth cycle, the original end time of the sorting and framing workstation is 3.57 seconds, only 0.03 seconds from the cycle end. It can be compressed forward by 0.015 seconds, bringing its action to end at 3.555 seconds, thus preventing the tail end of the action from entering the next cycle. This compression action must ensure that it does not cause premature triggering of control commands and must not compress to the point where its start time overlaps with other workstations. Before and after compression, the execution rhythm of the actions within the workstation must be recalculated to maintain the integrity of the physical execution of the actions and the mechanical safety distance.

[0041] Immediately after a brief tightening, the system enters the second functional period, the slow relaxation zone. This zone is primarily used to provide a buffer for workstations with significant motion time fluctuations. The slow relaxation operation does not average out all workstations; instead, it sets a relaxation coefficient based on the motion fluctuations of each workstation within the first three cycles. For example, the angle correction workstation's motion completion time fluctuation range from the fifth to the seventh cycle is 0.025 seconds, which is considered moderate. Therefore, in the eighth cycle, its end time can be extended by 0.01 seconds, from 3.46 seconds to 3.47 seconds. The relaxation process should be performed without exceeding the cycle's end boundary, while ensuring a 0.02-second cleanup zone before the start of the next cycle to prevent interference from the end of the previous cycle from influencing the start command of the next cycle.

[0042] The time cleanup zone is a time interval reserved between the end of each cycle and the start of the next cycle, during which there are no station actions, no control command triggers, and no accumulated signal transmission delays. The purpose of this time interval is to ensure that all actions and signals from the previous cycle are completely finished and cleared, thereby preventing interference with the start control of the next cycle due to signal residue, action tail delays, or feedback lags. In other words, the time cleanup zone is equivalent to a time isolation band established between two adjacent cycles, used to ensure the logical independence of cycle boundaries and the purity of the signal environment.

[0043] During implementation, the length of the time clean zone is determined based on the actual equipment response characteristics and signal transmission delay, and is generally set to be no less than 0.02 seconds. This section is located after the slow relaxation zone at the end of the cycle and is the termination part of the reverse-phase breathing window. Its core function is to provide a stable cycle boundary for the control system, ensuring that all commands do not cross-trigger or overlap in timing during cycle switching, thereby guaranteeing the periodic continuity of the cycle and the closed-loop integrity of the control logic.

[0044] All time-related changes in workstation actions resulting from brief tightening and gradual loosening are synchronously written back to the cycle phase band. Every time node in the cycle phase band needs to be updated, including the start and end points of the action, the duration of the time interval, and the time difference between adjacent workstations. For example, if the end time of the material transfer workstation is advanced from 3.40 seconds to 3.38 seconds after adjustment, the end time item for that workstation needs to be updated in the phase band, and the cycle interval field between it and the next workstation needs to be recalculated from 0.08 seconds to 0.10 seconds. Simultaneously, each adjustment operation is tagged with its operation type, distinguishing between forward compression and backward release, and the adjustment magnitude and basis are recorded for backtracking and optimization strategies in subsequent cycles. Through dynamic write-back, the cycle phase band is no longer a static reference structure, but rather a dynamic timing template that evolves in real time in response to execution feedback.

[0045] After writing back the beat phase band, convergence verification and stability analysis were performed on the entire reverse-phase breathing adjustment process. Convergence verification required calculating the trend of breathing window adjustment amplitude changes over three consecutive beat cycles. If the maximum adjustment amplitude in each cycle showed a decreasing trend, and the total time of tightening and loosening was controlled within 70% of the total duration at the end of the beat, then the beat was considered to have entered a stable fluctuation range. Stability analysis focused on evaluating whether new time conflicts, overlapping actions, or misaligned commands occurred after the breathing window adjustment. If no new time conflicts occurred within three cycles, command output was stable, and workstation responses were consistent, it proved that the reverse-phase breathing window could effectively buffer beat boundary interference and stabilize the rhythm structure.

[0046] The following example, using an automated crankshaft inspection and framing production line, fully demonstrates the application process and benefits of this invention in a real production environment. The production line comprises six continuous stations: material transfer, angle correction, pneumatic measurement, visual inspection, sorting and framing, and empty frame cyclic loading. The production line aims to complete the inspection and framing of crankshafts at a stable cycle time, with a target cycle time of 3.6 seconds per piece. The theoretical output under 1000 seconds of effective operating time per hour is approximately 1000 ÷ 3.6 ≈ 278 pieces. In actual operation, traditional fixed-cycle and linear feedback methods are prone to microsecond-level signal delays and cycle time adjustments that cancel each other out under high-speed conditions, leading to phase misalignment and closed-loop oscillation, which in turn causes localized congestion, incorrect framing, and line shutdown.

[0047] The initial data for a typical production line are as follows: Average time for material transfer is 0.42 seconds, with an average signal transmission delay of 0.016 seconds; average time for angle correction is 0.58 seconds, with a signal delay of 0.021 seconds; average time for pneumatic measurement is 0.64 seconds, with a signal delay of 0.026 seconds; average time for vision inspection is 0.92 seconds, with a signal delay of 0.023 seconds; average time for sorting and framing is 0.70 seconds, with a signal delay of 0.028 seconds; average time for empty frame cyclic loading is 0.55 seconds, with a signal delay of 0.030 seconds. When the cycle time is 3.6 seconds, statistics over 10 consecutive cycles show that multiple adjacent workstations overlap in time between 2.3 and 2.9 seconds, with a typical overlap length of 0.10 to 0.18 seconds. The feedback signal exhibits a peak delay of 0.040 seconds in some cycles, canceling out the 0.040-second cycle time adjustment period. This causes the correction command from the previous cycle and the actual position deviation of the current cycle to continuously overlap in phase, manifesting as interlocking action corrections and output jitter. In the past, the actual output measured for an 8-hour shift was only between 202 and 214 pieces, averaging 209 pieces, indicating a low effective equipment utilization rate and a higher misdirection rate than the target value.

[0048] After applying this invention, a clock phase band is first constructed based on a unified time axis, and the action curves of each workstation and the signal arrival times are laid out periodically. Taking a certain observation as an example, in the fourth cycle, visual detection starts at 1.84 seconds, lasts for 0.93 seconds, and ends at 2.77 seconds; classification and framing starts at 2.69 seconds and lasts for 0.71 seconds, with a 0.08-second overlap between the two. In the sixth cycle, pneumatic measurement ends at 2.12 seconds, and visual detection starts at 1.98 seconds, with a 0.14-second overlap. After comparing all workstations pairwise, the distribution of offset intervals is obtained, with red segments indicating overlaps exceeding 0.10 seconds and green segments indicating gaps exceeding 0.50 seconds. This generates a list of delay fingerprints. For example, visual inspection and sorting / framing overlap 7 times in 10 cycles, with 3 of them exceeding 0.12 seconds; pneumatic measurement and visual inspection overlap 6 times in 10 cycles, with the longest overlap being 0.16 seconds; and material transfer and angle correction have 4 gaps in 10 cycles, with the longest overlap being 0.62 seconds. The list of delay fingerprints provides a numerical basis for subsequent insertion avoidance and sliding rearrangement.

[0049] Next, a fixed micro-gap is inserted before the overlapping area to form a misphase index list. For workstation pairs with an overlap of 0.12 seconds, a 0.15-second avoidance window is set, delaying the start time of subsequent workstations by 0.15 seconds. For workstation pairs with an overlap of 0.08 seconds, a 0.11-second avoidance window is set to absorb statistical fluctuations and signal jitter. For recurring, heavily overlapping pairs, conflicts near the first half of the cycle are prioritized, as these conflicts can chain and delay multiple subsequent actions. Taking the fifth cycle as an example, the original start time for visual inspection was 2.02 seconds, but due to a 0.13-second overlap with the end of pneumatic measurement, it was delayed by 0.16 seconds to 2.18 seconds; the original start time for sorting and framing was 2.74 seconds, but due to a 0.09-second overlap with the end of visual inspection, it was delayed by 0.12 seconds to 2.86 seconds. All change records are entered into the misphase index list, clearly specifying the reason, length, and new start and end times for each adjustment.

[0050] Then, a sliding rearrangement is performed. A forward shift is applied to densely moving sections, and a backward shift is applied to sparsely moving sections, generating a tick-based sliding list. Taking the fourth cycle as an example, within 2.40 seconds to 2.90 seconds, the original three action segments—material transfer, pneumatic measurement, and visual inspection—overlap, with a peak of 3 simultaneous online actions. The material transfer is moved forward by 0.14 seconds, changing the start and end times to 2.18 seconds to 2.60 seconds; visual inspection remains at 2.18 seconds to 3.10 seconds; in the sparse area from 3.00 seconds to 3.50 seconds, the sorting and framing is moved backward by 0.10 seconds, changing the start and end times to 3.00 seconds to 3.70 seconds. If the cycle limit is 3.60 seconds, the 0.10-second overshoot at the end is handled by tightening and loosening the breathing window at the tail end to avoid cross-cycle tailing. After the sliding is completed, the number of simultaneous online actions from 2.40 seconds to 2.90 seconds decreases from 3 to 2, the dense peaks are dispersed, and the tick-based peaks and valleys are smoother.

[0051] With the timing list in place, the signal distribution phase begins. Based on the actual response time, fast-response commands are routed directly, while slower-response commands are injected into the buffer and released in advance. For example, angle correction has an average response time of 0.022 seconds and a fluctuation of 0.004 seconds, which is considered fast; the command can be issued directly from the action's starting point. For empty frame loading, the average response time is 0.071 seconds and the fluctuation is 0.012 seconds; the buffer needs to be injected 0.083 seconds before the target action time. In the seventh cycle, the sorting and loading target takes 3.05 seconds to execute; the control command enters the buffer at 2.967 seconds and is accurately triggered after a 0.083-second dwell time. All command paths, lead times, and dwell times form a dual-track command list, ensuring millisecond-level timing accuracy at the micro-time level.

[0052] Finally, a reverse-phase breathing window is set at the end of the cycle. The window length is fixed at the end of the 0.20-second interval. First, a short tightening is performed, followed by a slow widening. The short tightening is used to compress the latest ending action forward by 0.01 to 0.03 seconds to avoid the tail crossing the boundary; the slow widening is used to provide a buffer of 0.01 to 0.02 seconds for workstations with large fluctuations, so that the previous compression will not cause rebound oscillations. Taking the eighth cycle as an example, the original end time of sorting and framing was 3.58 seconds, which was compressed by 0.015 seconds to 3.565 seconds; the loading action, due to large fluctuations, was widened by 0.012 seconds, and the end time was shifted from 3.47 seconds to 3.482 seconds. All corrections are written back to the cycle phase band, and the next cycle is re-matched according to the new phase band, with each workstation entering operation with the new start and end relationships. After observing for 3 consecutive cycles, the maximum correction amplitude decreased from 0.028 seconds to 0.010 seconds, and the cycle time gradually converged.

[0053] The comparison with traditional methods is reflected in multiple dimensions. In terms of production capacity, the average output per shift was 209 units in the previous 8-hour shift. After implementation, under the same personnel and material supply conditions, the average output increased to 258 units, a single-shift increase of approximately 23.4%. Regarding cycle time stability, the peak number of simultaneous online actions within 20 consecutive cycles decreased from 3 to a maximum of 2, and the length of the high-density interval shortened from an average of 0.44 seconds per cycle to 0.17 seconds. Regarding mis-shunting, the visual detection accuracy for determining the same direction during frame assembly increased from 99.0% to 99.7%, and the number of mis-shunted units decreased from 2 per shift to less than 1 per shift. Regarding line stoppages, the number of emergency stops triggered by closed-loop oscillation decreased from 5 per shift to less than 1 per shift, and the average recovery time for a single emergency stop decreased from 7 minutes to 2 minutes. The average unit cycle time was optimized from 3.60 seconds to 3.35 seconds, with the optimal cycle time reaching 3.30 seconds during the stable period. The first-pass yield rate increased from 97.8% to 98.9%. In terms of energy consumption, due to reduced waiting time and suppression of repetitive actions, energy consumption per 100 items decreased by approximately 7.2. The above data is the statistical average of 3 consecutive days with 2 shifts per day, and the total sample size exceeds 1500 items.

[0054] To more intuitively demonstrate the benefits, a comparison table of key indicators is provided below. The baseline is the measured average of the traditional fixed-cycle and linear feedback methods under the same production line, the same materials, the same personnel, and the same environmental conditions; the improved value is the average of the stable period after applying the method of this invention and completing three rounds of closed-loop iteration of the cycle phase band.

[0055] Index item Conventional method actual measurement Invention method actual measurement Variation range Single shift output number 209 pieces 258 pieces +23.4 Average unit tact time 3.60 seconds 3.35 seconds −6.9 Stable period optimal tact time 3.50 seconds 3.30 seconds −5.7 Simultaneous online action peak value 3 stations 2 stations −33.3 High density interval average length 0.44 seconds 0.17 seconds −61.4 Misrouting rate 1.0 0.3 −70.0 Emergency stop frequency per shift 5 times 1 time −80.0 Single emergency stop average recovery 7 minutes 2 minutes −71.4 One-time pass rate 97.8 98.9 +1.1 Energy consumption per hundred pieces 100 base 92.8 base −7.2 99th percentile of instruction time error 0.020 seconds 0.008 seconds −60.0 Breathing window maximum correction 0.028 seconds 0.010 seconds −64.3 This example demonstrates that the present invention does not simply extend or shorten a certain action, but rather expresses the workstation action time, signal arrival time, and dwell time in phase on a unified time axis. Then, it uses a delay fingerprint list, avoidance window, phase misalignment index, sliding rearrangement, dual-track command, and reverse phase breathing window to coordinate layer by layer, avoiding phase cancellation and closed-loop oscillation, and resolving cycle conflicts within the cycle. This allows the production rhythm to remain convergent and stable under high-speed conditions, achieving reproducible improvements in multiple key indicators.

[0056] This invention establishes a unified timeline and cycle phase band to perform time-series quantitative management of the action data of each workstation, which were originally operating independently. Combined with a delay fingerprint list and a phase misalignment index list, it accurately identifies and dynamically avoids time overlaps, gaps, and delays between multiple workstations, achieving overall coordination and real-time correction of the rhythm of multiple workstation actions. By using a cycle sliding list to adjust the sequence of workstation actions forward and backward, the overall rhythm tends to be balanced, avoiding localized cycle congestion. Simultaneously, a dual-track command list manages rapid response commands and delayed commands separately, eliminating execution deviations caused by signal interference and command stacking. Finally, a reverse-phase breathing window is set at the end of each cycle, utilizing a short tightening and slow widening self-adjustment process to actively eliminate cycle misalignment and closed-loop oscillations, achieving dynamic convergence and continuous stable operation of the entire line's cycle. Compared to existing control methods that rely on fixed cycle time and linear feedback mechanisms, this invention reduces cycle time error by more than 60%, improves multi-station synchronization accuracy to the millisecond level, significantly enhances the automation level, cycle time stability, and equipment uptime of the production line, and effectively solves the problem of instability in multi-station collaborative systems under high-speed cycle time conditions.

[0057] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A crankshaft conveying and framing beat control method based on multi-station cooperation, characterized in that, The method comprises the following steps: Collecting the action time of each station, the arrival time of signals and the stay duration, establishing a unified time axis, marking the peak and valley of each station action curve, and generating a beat phase belt for representing the whole line time rhythm; Aligning the starting point and ending point of each station action according to the beat phase belt, identifying the time overlap area and time empty area between adjacent stations, generating a delay fingerprint list for recording the time sequence deviation characteristics between stations; Setting an avoidance window for each station according to the delay fingerprint list, inserting a fixed micro gap before the overlap area to form a phase shift index list for establishing a conflict-free distribution of action time; Rearranging the action sequence of the station according to the phase shift index list, applying a forward shift to the action-intensive station and a backward shift to the action-lagging station to generate a beat shift list for balancing the station beat; According to the double-track order list, an inverse phase breathing window is established at the end of each beat cycle, and a short and brief tightening operation is performed before a slow and slow operation is performed to eliminate beat conflict and oscillation, and the correction result is written back to the beat phase belt. The beat phase belt generation step is as follows:

2. The multi-station cooperative based crankshaft delivery framing beat control method of claim 1, wherein, Collect the action time, signal arrival time and action stay duration of each station, and establish a unified time axis with the whole line initialization as zero point; Draw the action line segment and signal arrival position of each station on the unified time axis, and mark the action duration and stay time; The peak and valley of each station action line segment on the time axis are calibrated, and the time overlap and empty interval are displayed in color; According to the peak and valley distribution results and the time overlap ratio, a beat phase belt is constructed for subsequent steps of beat rearrangement, phase shift avoidance and beat shift. The delay fingerprint list generation step is as follows:

3. The multi-station cooperative based crankshaft delivery framing beat control method of claim 2, wherein, Extract the action start time and action end time of each station on the unified time axis in the beat phase belt; Compare the action time of the adjacent previous station and the subsequent station in each beat cycle to identify the time overlap area and time empty area; According to the overlap length or empty length, generate a delay fingerprint list and record the time overlap length, time empty length, action start offset value and action end offset value; Classify and archive the data in the delay fingerprint list, and divide the overlap area and empty area according to the length range. The phase shift index list formation process is as follows:

4. The multi-station cooperative based crankshaft delivery framing beat control method of claim 3, wherein, According to the delay fingerprint list, identify the time overlapping station pair, and set an avoidance window to delay the action start time of the late station; Insert a micro gap with a length greater than the time overlap length in the avoidance window, and record the adjusted action start time and action end time; Record the original start time, micro gap length, adjusted time and beat cycle number as the phase shift index list; Check the phase shift index list to ensure that the time sequence is reasonable, no new overlap conflict is caused, and the upper limit of the beat cycle is not exceeded. The beat shift list generation process is as follows:

5. The multi-station cooperative based crankshaft delivery framing beat control method of claim 4, wherein, Extract the adjusted action start time and action end time of each station in the phase shift index list, and calculate the action density in each beat cycle; ​ Apply forward shift to the stations in action dense section and adjust action time section; Apply backward shift to the stations in action sparse section and evaluate the remaining space of the beat cycle; Arrange all the action adjustment information after forward and backward shift into beat slip list and record the slip time amount and corresponding beat cycle number.

6. The multi-station cooperative based crankshaft delivery framing beat control method of claim 5, wherein, The setting of forward and backward shift is based on action density level, the forward shift is controlled within the action dense section without causing time overlap, the backward shift is controlled within the remaining time range of the beat cycle, and a response buffer time is reserved after each adjustment to absorb instruction transmission delay.

7. The multi-station cooperative based crankshaft delivery framing beat control method of claim 5, wherein, The double-track command list generation process is as follows: Extract the action start time of each station in the beat slip list and calculate the average response time and maximum offset time of the control instruction; Divide the control instruction into fast response instruction and response delay instruction according to the response time; Set a direct output path for fast response instruction and a path containing buffer band for response delay instruction; Set the buffer band residence time and calculate the advance issuance point according to the instruction target time; Compile all control information to generate double-track command list and perform structure consistency and timing conflict check.

8. The multi-station cooperative based crankshaft delivery framing beat control method of claim 7, wherein, According to the double-track command list, an inverse phase breathing window is established at the end of each beat cycle to perform short tightening operation and slow relaxation operation, and the modified results are written back to the beat phase band as follows: Extract the action end time of each station in the double-track command list within each beat cycle, and perform difference analysis with the cycle end time to identify the adjustable time section at the end of the cycle; Set the inverse phase breathing window in the identified time section and divide it into short tightening area and slow relaxation area; In the short tightening area, adjust the stations with later action end time forward and recalculate the action rhythm; In the slow relaxation area, according to the action fluctuation, the end time of some stations is delayed and the time cleaning area is maintained; Write the tightening and relaxation time adjustment data back to the beat phase band and perform convergence and stability verification.

Images