A space application type scheduling method and system for a multi-channel parallel machining machine tool

Abstract

The application provides a space application type scheduling method and system of a multi-channel parallel machining tool, relates to the technical field of tool machining, and comprises the following steps: obtaining part machining data and generating machining tasks of each channel, and dividing the machining tasks into a plurality of machining task units; when any channel executes a machining task unit, determining a machining application space and performing overlap auditing with a preset occupied space set, and if there is overlap, the unit is added to a waiting sequence; after responding to a channel release signal, updating the occupied space set, and based on the updated occupied space set, performing space overlap auditing on the waiting sequence one by one, and performing scheduling response processing according to the auditing result, so as to realize multi-channel parallel machining. The method binds the machining task unit with space occupation, performs space overlap auditing before execution, makes conflicting tasks enter the waiting sequence, and automatically re-audits the waiting sequence after space release to realize admission, so that the overall efficiency and safety of multi-channel parallel machining are improved.

Description

Technical Field

[0001] This invention belongs to the field of machine tool processing technology, specifically relating to a space-based scheduling method for multi-channel parallel machining tools. Background Technology

[0002] In multi-channel parallel machining centers, multiple machining channels simultaneously execute different CNC program segments to improve overall machining capacity. However, multiple moving parts share a local machining space, leading to a significant increase in the risk of dynamic interference. Existing technologies typically employ channel-level synchronization or online full 3D collision detection for collision protection. The former only achieves time alignment of machining cycles in each channel and cannot characterize the occupancy and mutual exclusion of spatial resources, causing multiple executing parts to potentially enter the same hazardous space simultaneously. The latter relies on high-fidelity geometric models and real-time calculations, resulting in a high computational burden and a tendency to trigger overly conservative shutdowns, reducing parallel machining capacity. Consequently, it is difficult to balance the safety and efficiency of multi-channel parallel machining. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a space-request-based scheduling method for multi-channel parallel machining tools to solve the aforementioned problems. This method binds machining task units to space occupancy, performs space overlap review before execution to put conflicting tasks into a waiting sequence, and automatically re-reviews the waiting sequence after space is released to achieve admission, thereby avoiding dynamic collisions between multiple channels and improving the overall efficiency and safety of multi-channel parallel machining.

[0004] To address the aforementioned technical problems, this invention provides a space-based scheduling method for multi-channel parallel machining tools, comprising the following steps: Acquire part processing data, generate channel processing tasks corresponding to each processing channel in a multi-channel parallel machining center based on the part processing data, and divide the channel processing tasks into several processing task units so that each processing channel processes the part based on the corresponding several processing task units; When any processing channel executes to a processing task unit, a processing application space is determined based on the processing task unit, and a space overlap review is performed on the processing application space based on a preset occupied space set to obtain a space review result; If the spatial review result indicates that there is overlap, the processing task unit is added to a preset waiting sequence; In response to a channel release signal of any processing channel, a processing release space is determined based on the channel release signal, and a preset occupied space set is updated based on the processing release space; Based on the updated preset occupied space set, the processing application space corresponding to each processing task unit in the preset waiting sequence is checked for space overlap one by one to obtain the waiting space check result; Based on the waiting space review results, the processing task units in the preset waiting sequence are scheduled and responded to to achieve multi-channel parallel processing.

[0005] In the above scheme, the channel processing tasks are divided into processing task units, each corresponding to an independent processing application space. When any processing channel executes to a processing task unit, the processing application space is checked against a preset occupied space set for overlap. Processing task units that overlap with the preset occupied space set are identified and blocked before entering the corresponding space, avoiding dynamic collisions between multiple channels. If the space check result indicates overlap, the processing task unit is added to a preset waiting sequence, allowing conflicting tasks to queue in an orderly manner instead of being directly shut down, reducing capacity loss due to conservative shutdowns. In response to the channel release signal, the processing release space is determined and the preset occupied space set is updated. Each processing task unit in the preset waiting sequence is checked for overlap one by one, so that waiting tasks are automatically re-admitted after the space is released. Based on the waiting space check result, the processing task units in the preset waiting sequence are scheduled and responded to, achieving multi-channel parallel processing while ensuring space safety, thus improving overall processing efficiency and safety.

[0006] Furthermore, the step of dividing the channel processing task into several processing task units includes: For each channel processing task, the process attributes are segmented based on the process switching point of the processing task in that channel to obtain a process segmentation task set; The process segmentation task segment set is physically boundary-segmented to obtain a physically boundary-segmented task segment set. The physical boundary segmentation task segment set is spatially stable segmented to obtain a spatially stable segmentation task segment set. Deadlock avoidance segmentation is performed based on a preset occupation time threshold and the spatial stability segmentation task set to obtain a deadlock avoidance segmentation task set. The deadlock avoidance segmentation task segment set is subjected to anomaly recovery boundary segmentation to obtain anomaly recovery boundary segmentation task segment set; The set of abnormal recovery boundary segmentation task segments is divided into regions by application and alignment to obtain the plurality of processing task units.

[0007] In the above scheme, the process attribute segmentation makes different processes correspond to different tool postures and spatial envelopes; the physical boundary segmentation uses natural positions such as the tool drop point and tool lift point as boundaries; the spatial stability segmentation ensures that the spatial occupancy envelope within the same task unit is basically stable; the deadlock avoidance segmentation limits the single occupancy time by setting a preset occupancy time threshold, reducing the risk of loop waiting; the anomaly recovery boundary segmentation provides natural boundaries for breakpoint recovery; and the region application alignment segmentation makes each task unit correspond to an independent region application and release, thereby jointly ensuring that the machining task unit has clear spatial boundaries and manageability, so that subsequent spatial overlap review and waiting sequence management can be based on stable and independent spatial objects, reducing the probability of deadlock and starvation in multi-channel concurrent scheduling.

[0008] Furthermore, the step of dividing the channel processing task into several processing task units, so that each processing channel processes parts based on the corresponding several processing task units, includes: After dividing the channel processing task into several processing task units, spatial modeling is performed on each processing task unit based on a preset machine tool parameter set to obtain the processing application space corresponding to each processing task unit. Based on the preset forced processing dependencies and the preset machine tool parameter set, the scheduling parameters of each processing task unit are constructed to obtain the task scheduling parameter set corresponding to each processing task unit; Based on the processing application space and task scheduling parameter set corresponding to each processing task unit, the scheduling of the plurality of processing task units is optimized to obtain an ordered task unit sequence corresponding to each processing channel, so that each processing channel executes each processing task unit in sequence based on the ordered task unit sequence, thereby processing the parts.

[0009] In the above scheme, after dividing the channel processing task into several processing task units, spatial modeling is performed on each processing task unit based on a preset machine tool parameter set to obtain the processing application space corresponding to each processing task unit, thus directly associating each processing task unit with the processing application space. Scheduling parameters are constructed for each processing task unit based on preset mandatory processing dependencies and the preset machine tool parameter set, resulting in a task scheduling parameter set for each processing task unit. Then, based on the processing application space and task scheduling parameter set corresponding to each processing task unit, the scheduling of several processing task units is optimized to obtain an ordered task unit sequence for each processing channel. Each processing channel then executes each processing task unit sequentially based on the ordered task unit sequence. By performing spatial modeling and optimized scheduling before execution, the overlapping identification of processing application spaces and the temporal coordination of preset mandatory processing dependencies are brought forward to the scheduling stage. This allows the ordered task unit sequence to avoid spatial conflicts and dependency contradictions at the source, thereby reducing the frequency of spatial review blocking and invalid waiting during multi-channel parallel processing and improving overall processing efficiency.

[0010] Furthermore, based on the preset forced machining dependencies and the preset machine tool parameter set, scheduling parameters are constructed for each machining task unit to obtain the task scheduling parameter set corresponding to each machining task unit, including: For each processing task unit, the preceding task set and the following task set of the processing task unit are determined based on the preset forced processing dependency relationship; Based on the preset machine tool parameter set and the machining task unit, the machining time is measured to obtain the basic execution time; The scheduling evaluation time is obtained by superimposing a predicted waiting penalty term on the basic execution time based on a preset conflict penalty coefficient. The task scheduling parameter set corresponding to the processing task unit is obtained based on the pre-task set, post-task set, basic execution time, and scheduling evaluation time of the processing task unit.

[0011] In the above scheme, for each processing task unit, the pre-processing task set and post-processing task set are determined based on the preset mandatory processing dependencies to clarify the mandatory timing constraints between task units. The processing time is measured based on the preset machine tool parameter set and the processing task unit to obtain the basic execution time, providing an execution time benchmark. A predicted waiting penalty is added to the basic execution time based on a preset conflict penalty coefficient to obtain the scheduling evaluation time, incorporating the potential waiting cost caused by spatial overlap into the time evaluation. Finally, the task scheduling parameter set corresponding to the processing task unit is obtained based on the pre-processing task set, post-processing task set, basic execution time, and scheduling evaluation time. Thus, the task scheduling parameter set simultaneously carries the process logic defined by the preset mandatory processing dependencies, the processing time reflected by the basic execution time, and the spatial conflict penalty information contained in the scheduling evaluation time. This provides a comprehensive decision-making basis for subsequent optimized scheduling based on processing application space and the task scheduling parameter set, taking into account both process feasibility and spatial competition avoidance. This allows the resulting ordered task unit sequence to effectively reduce the risk of spatial overlap and invalid waiting during multi-channel parallel processing, improving overall scheduling efficiency.

[0012] Furthermore, the optimized scheduling of the plurality of processing task units based on the processing request space and task scheduling parameter set corresponding to each processing task unit, to obtain an ordered sequence of task units corresponding to each processing channel, includes: Construct a dependency constraint graph and a space conflict matrix based on the processing application space and task scheduling parameter set corresponding to each processing task unit; Based on the preset processing execution time, the dependency constraint graph, and the spatial conflict matrix, the plurality of processing task units are comprehensively sorted to obtain a set of candidate sorting schemes; The waiting time evaluation value corresponding to each candidate sorting scheme in the candidate sorting scheme set is obtained by estimating the duration of each candidate sorting scheme. The optimal solution is searched based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain the target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on the target sorting scheme.

[0013] In the above scheme, the dependency constraint graph ensures that the candidate sorting scheme set satisfies the preset mandatory processing dependencies, the spatial conflict matrix quantifies the competition intensity of processing application space, the duration estimation transforms spatial competition into a comparable waiting time evaluation value, and the optimal solution search selects the target sorting scheme with the minimum waiting cost in the feasible solution space. Thus, by jointly optimizing the processing application space and task scheduling parameter set during the scheduling stage, multi-channel spatial competition and dependency conflicts are resolved in advance, reducing the frequency of space overlap review blocking and invalid waiting during runtime, and improving the execution smoothness of the ordered task unit sequence.

[0014] Further, the step of searching for the optimal solution based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain the target sorting scheme, and generating an ordered task unit sequence corresponding to each processing channel based on the target sorting scheme, includes: The optimal solution is searched based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain the initial sorting scheme; The waiting percentage is determined based on a preset waiting percentage threshold and the initial sorting scheme, and the waiting percentage determination result is obtained. When the waiting percentage determination result is that the duration exceeds the standard, a set of high-conflict task units is determined based on the spatial conflict matrix and several processing task units, and the set of high-conflict task units is further subdivided to obtain several subdivided processing task units. Based on several subdivided processing task units and several other processing task units, spatial modeling, scheduling parameter construction, and scheduling optimization are performed again to obtain an updated target sorting scheme. Then, based on the updated target sorting scheme, the waiting ratio is re-determined until the waiting ratio determination result is that the duration has not exceeded the standard. The updated target sorting scheme is determined as the target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on the target sorting scheme.

[0015] In the above scheme, after obtaining the initial sorting scheme by searching for the optimal scheme based on the waiting time evaluation value corresponding to each candidate sorting scheme, the waiting ratio is determined based on the preset waiting ratio threshold and the initial sorting scheme. When the time exceeds the limit, a set of high-conflict task units is determined based on the spatial conflict matrix and several processing task units, and the set of high-conflict task units is further subdivided into several subdivided processing task units. Based on the subdivided processing task units and the remaining processing task units, spatial modeling, scheduling parameter construction and scheduling optimization are performed again to obtain the updated target sorting scheme and the waiting ratio is determined again until the time does not exceed the limit. The above scheme adaptively subdivides the set of high-conflict task units through a waiting ratio feedback mechanism, and iteratively performs spatial modeling, scheduling parameter construction, and schedule optimization until the preset waiting ratio threshold is met, so that the granularity of task unit division and spatial conflict intensity are dynamically matched. When the waiting ratio determination result is that the duration does not exceed the standard, the updated target sorting scheme is determined as the target sorting scheme and an ordered task unit sequence is generated. Thus, in the scheduling stage, the task unit granularity is adaptively adjusted to reduce the long-term occupation of processing application space by high-conflict task units, reduce the overall waiting time evaluation value, and improve the execution efficiency of multi-channel parallel processing.

[0016] Furthermore, based on a preset set of occupied space, the processing application space is subjected to a space overlap review to obtain a space review result, including: Based on a preset safety margin, the processing application space and the preset occupied space set are respectively expanded outward to obtain the corresponding expanded application space and expanded occupied space set; Based on the preset axis-aligned bounding box algorithm, a rough bounding box judgment is performed on the set of the outward application space and the outward occupied space to obtain a candidate overlapping pair list; The candidate overlapping pair list and the outer space occupied set are accurately overlapped based on the preset geometric intersection determination algorithm to obtain the space review result.

[0017] In the above scheme, the processing application space and the preset occupied space set are expanded outward based on a preset safety margin to obtain the corresponding expanded application space and expanded occupied space set, thus compensating for the spatial uncertainty caused by thermal deformation, clamping errors, and control lag during processing. A preset axis-aligned bounding box algorithm is used to perform a coarse bounding box judgment on the expanded application space and the expanded occupied space set, resulting in a candidate overlapping pair list. This quickly eliminates obviously non-intersecting space pairs, significantly reducing the number of objects requiring further judgment. A preset geometric intersection judgment algorithm is used to perform a precise overlap judgment on the candidate overlapping pair list and the expanded occupied space set, obtaining the space review result. By expanding the safety margin to provide engineering safety margin, reducing the computational scale through coarse bounding box judgment, and performing precise overlap judgment only after the coarse judgment passes, the space review result significantly reduces the computational burden while ensuring judgment accuracy. This enables online real-time execution of space overlap review in multi-channel concurrent scenarios, ensuring that conflict detection between the processing application space and the preset occupied space set is both reliable and efficient.

[0018] Furthermore, the process application space corresponding to each processing task unit in the preset waiting sequence is checked for overlap based on the updated preset occupied space set to obtain the waiting space check result, including: Determine that the preset waiting sequence is not empty, obtain the waiting time of each processing task unit, and traverse the preset waiting sequence based on the preset aging coefficient, the preset basic process priority and the waiting time of each processing task unit to determine the effective priority of each processing task unit. The preset waiting sequence is reordered based on the effective priority of each processing task unit to obtain the re-examination waiting sequence; Based on the updated preset occupied space set, the processing application space corresponding to each processing task unit in the re-examination waiting sequence is reviewed one by one for space overlap to obtain the waiting space review result.

[0019] In the above scheme, when the preset waiting sequence is not empty, the waiting time of each processing task unit is obtained. Based on the preset aging coefficient, the preset basic process priority, and the waiting time of each processing task unit, the preset waiting sequence is traversed to determine the effective priority of each processing task unit. The waiting time is then converted into a dynamic priority gain through the preset aging coefficient, which, together with the preset basic process priority, determines the re-review order. The preset waiting sequence is reordered based on each effective priority to obtain the re-review waiting sequence. Based on the updated preset occupied space set, the processing application space corresponding to each processing task unit in the re-review waiting sequence is reviewed one by one for space overlap to obtain the waiting space review result. Through the dynamic calculation of effective priorities, processing task units that have been waiting for a long time are given higher re-review priority, avoiding some channels from being unable to obtain processing application space for a long time due to continuous space conflicts. Each review is conducted and the preset occupied space set is updated immediately after each approval to ensure that subsequent reviews in the same release event are based on the latest occupied state, eliminating the approval of conflicting processing application spaces. This ensures the fairness of the waiting queue while guaranteeing space mutual exclusion, and improves the overall responsiveness of multi-channel parallel processing.

[0020] Furthermore, the space-request scheduling method for multi-channel parallel machining tools further includes: During multi-channel parallel machining, the actual axis position feedback information of each machining channel is obtained in real time; The actual axial distance between corresponding axes in adjacent machining channels is determined based on the actual axis position feedback information of each machining channel. When the actual axial distance is less than the preset safety distance, a shutdown protection command is generated to stop the multi-channel parallel machining tool from running. When the actual axial distance is less than the preset warning distance, a speed reduction operation command is generated to cause the multi-channel parallel machining tool to reduce its speed. The current machine tool processing status is saved based on the shutdown protection command or the speed reduction operation command; In response to the machine tool recovery signal, and based on the machine tool recovery signal and the current machine tool processing state, the state of the multi-channel parallel machining machine tool is restored so that each processing channel can continue to process parts.

[0021] In the above scheme, real-time acquisition of actual axis position feedback information and determination of actual axial distance provide real-time execution layer protection for multi-channel parallel machining tools. When the actual axial distance is lower than the preset safety distance, a stop protection command is directly generated to completely block the actual collision risk. When the actual axial distance is lower than the preset warning distance, a deceleration operation command is generated to reduce the movement speed in advance to buffer the approaching trend of anomalies. At the same time, the current machining state of the machine tool is saved based on the stop protection command or deceleration operation command to ensure that there is a state rollback benchmark after an anomaly occurs. In response to the machine tool recovery signal, the state is restored based on the saved current machine tool machining state, so that each machining channel can continue part processing from the saved point, avoiding secondary anomalies caused by state loss or misalignment during the recovery stage. Thus, while ensuring the safety of multi-channel parallel machining, the efficiency of anomaly recovery and the continuity of operation are improved.

[0022] The present invention also provides a space-request scheduling system for multi-channel parallel machining tools, comprising: The task unit division module is used to acquire part processing data, generate channel processing tasks corresponding to each processing channel in a multi-channel parallel machining center based on the part processing data, and divide the channel processing tasks into several processing task units so that each processing channel can perform part processing based on the corresponding several processing task units. The spatial overlap review module is used to determine the processing application space based on the processing task unit when any processing channel executes to a processing task unit, and to perform spatial overlap review on the processing application space based on a preset occupied space set to obtain the spatial review result; The waiting sequence management module is used to determine that the space review result indicates overlap, and then add the processing task unit to a preset waiting sequence; A space release response module is used to respond to a channel release signal of any processing channel, determine the processing release space based on the channel release signal, and update the preset occupied space set based on the processing release space; The waiting sequence re-examination module is used to perform space overlap review on the processing application space corresponding to each processing task unit in the preset waiting sequence based on the updated preset occupied space set, and obtain the waiting space review result. The scheduling response module is used to perform scheduling response processing on each processing task unit in the preset waiting sequence based on the waiting space review result, so as to realize multi-channel parallel processing.

[0023] In the above scheme, the task unit division module acquires part processing data and generates channel processing tasks corresponding to each processing channel, dividing the channel processing tasks into several processing task units, providing independent and manageable objects for space management; the space overlap review module determines the processing application space based on the processing task unit when the processing channel executes to the processing task unit, and performs space overlap review with the preset occupied space set to prevent conflicting tasks from entering the shared space; the waiting sequence management module adds the processing task unit to the preset waiting sequence when the space review result shows overlap, so that conflicting tasks are queued in an orderly manner instead of being directly stopped; the space release response module responds to the channel release signal, determines the processing release space and updates the preset occupied space set; the waiting sequence re-review module performs space overlap review on each processing task unit in the preset waiting sequence based on the updated preset occupied space set; the scheduling response module performs scheduling response processing on each processing task unit in the preset waiting sequence based on the waiting space review result. The modules work together to form a closed loop of conflict detection, queuing, release re-review and scheduling response, which reduces unplanned downtime and invalid waiting caused by space overlap while ensuring the security of the preset occupied space set, and improves the execution efficiency and safety of multi-channel parallel processing. Attached Figure Description

[0024] Figure 1 This is a schematic flowchart of a space-request scheduling method for a multi-channel parallel machining center, provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the timing interaction in a spatial request scheduling method for a multi-channel parallel machining center provided in an embodiment of the present invention; Figure 3 This is a top view of the machine tool spatial layout of a multi-channel parallel machining center in a spatial request scheduling method for a multi-channel parallel machining center provided in an embodiment of the present invention; Figure 4 This is a step-by-step schematic diagram of generating machining request space in a space request-based scheduling method for a multi-channel parallel machining tool according to an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the hierarchical protection principle of adjacent channel corresponding axis distance monitoring in a spatial request scheduling method for a multi-channel parallel machining center according to an embodiment of the present invention. Figure 6 This is a schematic diagram of a space-request scheduling system architecture for a multi-channel parallel machining tool, provided as an embodiment of the present invention. Detailed Implementation

[0025] 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 This embodiment provides a space-requesting scheduling method for multi-channel parallel machining tools, including the following steps: Step S1: Obtain part processing data, generate channel processing tasks corresponding to each processing channel in the multi-channel parallel machining center based on the part processing data, and divide the channel processing tasks into several processing task units so that each processing channel processes the part based on the corresponding several processing task units; Step S2: When any processing channel executes to a processing task unit, a processing application space is determined based on the processing task unit, and a space overlap review is performed on the processing application space based on a preset occupied space set to obtain a space review result; Step S3: Determine that the space review result indicates overlap, and add the processing task unit to the preset waiting sequence; Step S4: In response to a channel release signal of any processing channel, determine the processing release space based on the channel release signal, and update the preset occupied space set based on the processing release space; Step S5: Based on the updated preset occupied space set, perform space overlap review on the processing application space corresponding to each processing task unit in the preset waiting sequence one by one to obtain the waiting space review result; Step S6: Based on the waiting space review results, schedule and respond to each processing task unit in the preset waiting sequence to achieve multi-channel parallel processing.

[0027] Please see Figure 2 and Figure 3 In one embodiment, an anti-collision management module is introduced to perform unified spatial scheduling management of multi-channel machining tools. Each machining channel is equipped with at least one private waiting area, which is a pre-defined set of safe retreat poses. Under a unified coordinate system, it does not overlap with all shared danger zones, and the private waiting areas of different channels do not interfere with each other. When the current task unit is completed and the subsequent task unit has not yet been approved, the channel first retreats to its private waiting area to wait.

[0028] In practice, the part machining data is first acquired, and machining tasks for each machining channel are generated and divided into multiple machining task units. Each machining task unit is associated with the geometric space information occupied by the task. When loading the machining program G-code, each machining channel registers the geometric space information corresponding to each machining task unit to the anti-collision management module according to the task unit identifier. The anti-collision management module performs 3D modeling according to the channel number and task unit identifier to form the machining application space for each task unit.

[0029] Each channel executes in G-code order; before any channel reaches a certain task unit, it executes auxiliary code M100Pn, where n is the task unit identifier, submits a space usage request for that unit to the module, and then enters a waiting state and starts a waiting timer.

[0030] Upon receiving an application, the module locates the corresponding 3D spatial model (i.e., the processing space) based on the channel number and task unit identifier. It then reviews any overlaps based on a preset set of occupied spaces, which is the collection of occupied spaces from all currently approved and in-use task units. If the applied space does not overlap with any space in the occupied space set, the space is marked as used and added to the occupied space set. The application is approved, the channel's timer is stopped and reset, and subsequent G-code execution continues. If overlaps exist, the space is marked as unused and placed in a preset waiting sequence for orderly waiting.

[0031] During the waiting period, if the maximum waiting time threshold is reached... If approval is obtained within the specified time, the timer stops and is reset to zero; if the timeout occurs, the channel enters a timeout alarm state, an alarm is issued first and subsequent scheduling of the channel is suspended, the channel is kept in a private waiting area, and the timeout task identifier, waiting start time and current waiting sequence status are recorded for manual intervention or resumption of scheduling.

[0032] After the current task unit completes its execution, the channel returns to the private waiting area, executes auxiliary code M101Pn, and submits a space release request to the module. Upon module acceptance, the space occupied by the task unit is removed from the preset space set, and the space set is updated. Subsequently, the processing space requests of each task unit in the waiting sequence are re-examined for overlap. Units that no longer overlap are approved and scheduled for execution, while those that still overlap continue to wait for the next release and review. By timely updating the space set and automatically re-examining, waiting tasks that meet the conditions can be automatically admitted after space release, maximizing multi-channel parallel processing efficiency while ensuring safety.

[0033] It should be noted that, Figure 2 This is a temporal interaction diagram of a space-request scheduling method for multi-channel parallel machining tools. Figure 3 This is a top view of the machine tool space layout for a multi-channel parallel machining center. Figure 2 and Figure 3 A unified coordinate system is used, which is the machine tool's global XY top-view coordinate system. Figure 2 The horizontal rectangles in the diagram represent Channel 1, Channel 2, the anti-collision management module, and the region status, respectively; the vertical arrows represent the direction of time progression; solid arrows are the instruction signals sent by the processing channel to the anti-collision management module, and dashed arrows are the feedback signals returned by the module to the processing channel; dashed text boxes are used to annotate the atomic execution rules for re-examination after release; the M100 instruction is used to submit a processing space usage request, and the M101 instruction is used to submit a processing space release request; TU1 and TU2 refer to different processing task units. Figure 3 The red dashed rectangles in the diagram represent shared danger zones; the blue dashed rectangles represent private waiting zones; the blank rectangles represent processing / execution positions; the blue dashed arrows represent return paths; the two sets of processing / execution positions on the left and right correspond to Channel 1 and Channel 2 respectively, and each set of processing / execution positions is equipped with its own private waiting zone. During the channel waiting process, the user returns to the corresponding private waiting zone along the return path. Both processing / execution positions are adjacent to shared danger zones.

[0034] Furthermore, in one embodiment, the channel processing task is divided into several processing task units, including the following multi-level segmentation process. The task unit refers to the smallest processing segment within the same processing channel that can be independently requested, executed, and released as an independent area. Each task unit corresponds to at least a defined processing instruction segment, a defined space occupancy area, an estimated execution duration, and a complete application, review, processing, and release cycle, fully compatible with the scheduling mechanism constructed by the subsequent anti-collision management module based on space application, waiting for arbitration, and space release.

[0035] First, process attribute segmentation is performed. For each channel's machining task, segmentation is based on the process switching points of that channel's machining task. When the machining process spans different process types, it is preferentially divided into different task units. The different process types include, but are not limited to: end face turning, external turning, internal turning, grooving, drilling, milling, tapping, and the switching between roughing and finishing. Different processes typically correspond to different tool postures, machining objects, and spatial occupancy envelopes. By segmenting through process attributes, each of the divided task units has a relatively single and clear process attribute, avoiding a decrease in the accuracy of subsequent spatial overlap verification due to excessive differences in spatial envelopes.

[0036] Then, physical boundary segmentation is performed. The task segmentation set obtained by segmenting the process attributes is segmented using natural physical locations such as the tool drop point, tool lift point, step switching point, tool switching point, spindle or turret switching point, channel synchronization waiting point, position point before entering the shared area, and exit point after completing the processing in the shared area as boundaries. Preferably, the tool drop point is used as the starting point of the task unit and the position of lifting the tool to exit the current processing area is used as the ending point of the task unit, thus obtaining the physical boundary segmentation task segmentation set.

[0037] Next, spatial stability segmentation is performed. The task segment set is divided according to the physical boundary, and further segmented based on the spatial occupancy stability of the machining trajectory. For continuous trajectories within the same process, if the spatial occupancy envelope of the tool, turret, spindle head, or oscillating head is basically stable throughout the entire trajectory segment and does not need to exit the current shared area midway, then the trajectory segment is maintained within the same task unit. If any of the following occurs: a significant change in the spatial occupancy envelope, the tool entering or exiting a new shared danger zone, the attitude change of the executing component exceeding a preset threshold, the machining object switching from an outer surface to an inner cavity or inner hole, or subsequent machining requiring the re-application of new area resources, then a new task unit is defined at the corresponding location. Spatial stability segmentation ensures that the spatial occupancy envelope corresponding to the same task unit is basically stable, thereby determining a machining application space with clear geometric boundaries and stable shape for each task unit, facilitating accurate spatial overlap review by the anti-collision management module.

[0038] Deadlock avoidance partitioning is then performed. A pre-set threshold for the duration of partition occupation is used. , This indicates the maximum time a single task unit is allowed to continuously occupy the shared area. The purpose is to reduce the risk of circular waiting and deadlocks caused by prolonged occupation in multi-channel concurrent scheduling by limiting the duration of a single space occupation. This is based on a preset occupation time threshold. Deadlock avoidance segmentation is performed on the spatial stability segmentation task set. For continuous long trajectories that are expected to occupy the shared area for a long time and are prone to competition and conflict with adjacent channels, they are further subdivided into multiple execution times that do not exceed the occupation time threshold. Shorter task units are used to shorten the duration of a single occupation, reduce the possibility of loop waiting, and provide a basis for adaptive subdivision and adjustment based on the waiting ratio in subsequent scheduling.

[0039] Then, anomaly recovery boundary segmentation is performed. For the task segment set obtained after deadlock avoidance segmentation, suitable natural boundary locations for abnormal stop recovery, manual intervention recovery, or breakpoint continuation are identified, such as the location where a processing segment successfully ends and the tool can be safely retracted. These locations are prioritized as task unit boundaries. Anomaly recovery boundary segmentation provides natural boundaries for breakpoint recovery. When an abnormal stop occurs, the system can use these boundaries to re-establish consistency between task units, region occupancy states, and channel states, thereby improving the system's robustness and recoverability.

[0040] Finally, the region application alignment and segmentation are performed. The anomaly recovery boundary segmentation task set is segmented to ensure that each processing task unit corresponds exactly to one space region application submitted to the collision avoidance management module and one corresponding space release. That is, when a processing process needs to reapply for a new space occupation area from the collision avoidance management module, a new task unit is divided at the application point.

[0041] For ease of understanding, this embodiment uses the single-channel turning of a disc-shaped part in a dual-spindle multi-turret milling and turning center as an example. Assume that one channel needs to sequentially perform end-face turning, external turning, and internal turning, according to the above-mentioned segmentation rules: First, process attribute segmentation is performed at the process switching points from end face turning to external cylindrical turning and from external cylindrical turning to internal hole turning. Then, physical boundary segmentation is performed based on the physical boundaries such as the tool entry point and tool exit point of each process. Next, it is confirmed that the spatial occupancy envelope of each trajectory segment is stable, and no further spatial stability segmentation is required. At the same time, the expected duration of each process segment occupying the shared area does not exceed the preset occupancy duration threshold, and the deadlock avoidance segmentation maintains the original task segment unchanged. The position after each segment finishes machining and the tool is lifted naturally constitutes the abnormal recovery boundary. Finally, the segmentation is aligned according to the area application, requiring each task segment to correspond to an independent area application and release.

[0042] Based on this, three machining task units are ultimately obtained: The first task unit is the end-face turning task unit, which covers the machining process from the entry of the end-face turning tool into the end-face cutting position to the lifting and exiting of the tool after the end-face turning is completed. Its spatial occupancy is jointly determined by the workpiece end-face area and the local structures of the tool turret, tool, and front end of the spindle; The second task unit is the external turning task unit, which starts when the external turning tool enters the external diameter cutting starting point and ends when the tool is lifted and exited after the external diameter turning is completed. Its spatial occupancy surrounds the outer diameter contour of the workpiece; The third task unit is the internal turning task unit, which starts when the internal turning tool enters the hole opening to start cutting and ends when the internal hole machining is completed and exits the hole opening. Its spatial occupancy is determined by the envelope formed by the internal turning tool shank extending into the hole and the contour of the workpiece cavity. The above three task units correspond to three independent area application, review, processing, and release cycles. That is, before executing each task unit, the channel first applies for the corresponding spatial area, and then releases the area after completion. If the applied area overlaps with the currently occupied area of ​​other channels, the task unit enters the waiting sequence and is re-reviewed after the relevant area is released. The task unit division rules and the application, waiting, and release mechanisms of the collision avoidance management module form a direct correspondence, jointly ensuring the safety and efficiency of multi-channel parallel processing.

[0043] Further, please see Figure 4 In one embodiment, the channel processing task is divided into several processing task units, so that each processing channel processes parts based on the corresponding processing task unit, including: Based on a preset machine tool parameter set, spatial modeling is performed on each machining task unit to obtain the machining request space corresponding to each machining task unit. Scheduling parameters are constructed for each machining task unit based on preset mandatory machining dependencies and the preset machine tool parameter set to obtain the task scheduling parameter set corresponding to each machining task unit. The machining task units are then optimized and scheduled based on their corresponding machining request spaces and task scheduling parameter sets to obtain an ordered task unit sequence for each machining channel. Each machining channel then executes each machining task unit sequentially based on this ordered task unit sequence, thereby machining the part. By prioritizing spatial overlap identification and temporal coordination of mandatory dependencies in the scheduling stage, the ordered task unit sequence avoids spatial conflicts and dependency contradictions at the source, reducing the frequency of spatial audit blocking and invalid waiting during runtime, and improving overall machining efficiency.

[0044] For spatial modeling, the preset machine tool parameter set includes at least the actual geometric dimensions of the execution components such as the spindle head, cutting tool, turret, or oscillating head. A geometric space occupancy area is generated for each task unit as a machining application space. This machining application space serves as the spatial representation object when the task unit submits its application, the judgment object when the anti-collision management module reviews overlapping execution areas, and the area identifier object when the space is released and reclaimed after the task unit completes. The basic geometric calculation of the space occupancy area can employ existing methods in the art, such as swept volume, bounding box, convex hull, or 2D projected closed contour extrusion. This embodiment preferably uses a combination of local occupancy model, toolpath sweeping, and safety margin expansion to generate the machining application space.

[0045] The process of generating machining application space is as follows: Based on the start and end boundaries of the task unit, the corresponding toolpath trajectory segments are extracted from the NC program or CAM toolpath, and the curve trajectory is discretized into a sequence of tool point positions according to a preset step size. Based on the actual dimensions of the tool body, tool holder, tool shank, and the local structures of the spindle head, turret, or tilting head near the workpiece in the preset machine tool parameter set, a local occupancy model is established to characterize the instantaneous occupancy range of the execution component at a single tool point position on the surrounding space. This local occupancy model is mapped to each discrete tool point position in a unified coordinate system according to the posture information of the toolpath trajectory, obtaining the instantaneous space occupancy volume corresponding to each tool point position. The union of all instantaneous space occupancy volumes is taken to form the original space occupancy volume, i.e., the swept envelope formed by the local occupancy model moving along the toolpath trajectory. To compensate for thermal deformation, clamping errors, control lag, execution errors, and braking margins during processing, the original space occupancy is expanded and a safety margin is added to obtain the final space occupancy area. This safety margin includes a fixed safety boundary, process error compensation, control lag compensation, and thermal deformation compensation to ensure that no actual collision occurs even with the aforementioned disturbances. Finally, the final space occupancy area is simplified to generate a coarse bounding box and a fine geometry. The task unit identifier, its associated channel number, region geometric parameters, and estimated execution time are registered in the anti-collision management module for subsequent application, review, and release management.

[0046] Simultaneously, scheduling parameters are constructed based on preset mandatory machining dependencies and preset machine tool parameter sets. The preset mandatory machining dependencies refer to unchangeable machining sequence constraints determined by the part's process, such as the order of end face turning and external diameter turning within the same channel, or the synchronization requirements of certain processes between different channels. The constructed task scheduling parameter set includes at least the expected execution time of the task unit, a list of other task units with which it has mandatory sequential dependencies, and optional priority information.

[0047] After obtaining the processing request space and task scheduling parameter set for each task unit, optimized scheduling is performed. Optimized scheduling aims to minimize spatial conflicts and invalid waiting times in multi-channel parallel processing. It comprehensively utilizes the geometric overlap between the processing request spaces of each task unit and the mandatory dependencies in the task scheduling parameter set to uniformly sort the task units of each processing channel, forming an ordered sequence of task units. By prioritizing spatial overlap identification and temporal coordination of mandatory dependencies during the scheduling stage before processing execution, the ordered sequence of task units avoids spatial conflicts and dependency contradictions at the source, reducing the frequency of spatial review blocking and invalid waiting time during operation, and improving overall processing efficiency.

[0048] Taking the machining of a disc-shaped part by a dual-spindle multi-turret milling and turning center as an example, assume that the machining task of the channel is divided into three task units: end face turning, external turning, and internal turning. Based on a preset machine tool parameter set, a local occupancy model is established for each task unit using the corresponding tool and the local structure of the turret or spindle front end, and a machining application space is generated by sweeping along the corresponding cutting trajectory. Simultaneously, based on the mandatory machining dependency relationship of end face turning, external turning, and internal turning required by the machining process, a task scheduling parameter set is constructed for each task unit, clarifying its sequence constraints. Optimized scheduling, based on these dependencies and the potential overlap between the three machining application spaces, arranges an ordered sequence of task units for the channel, enabling each task unit to achieve maximum parallelism and conflict-free operation with task units in other channels in both space and time, while satisfying the process sequence. Each task unit has clear spatial boundaries and scheduling attributes, laying the foundation for the real-time application, review, and scheduling response of the subsequent collision avoidance management module.

[0049] It should be noted that, Figure 4 This is a step-by-step schematic diagram of generating the processing application space in this embodiment. Five independent boxes are arranged horizontally in the diagram. Each box is accompanied by a text description. The solid blue arrows between adjacent boxes represent the sequence of processes. The three-dimensional model in each box is the spatial geometric entity generated in the corresponding step. Dashed lines are used to mark the boundaries and allowance areas.

[0050] The first set of boxes corresponds to the local occupancy model generation step. The internal 3D model is a local occupancy model composed of the tool, tool holder, and spindle head local structures. This model represents the instantaneous space occupied by the execution component at a single tool position point. The second set of boxes corresponds to the toolpath sweeping step, which contains multiple sets of local occupancy models. The blue dashed curve represents the discretized toolpath trajectory, illustrating the mapping and arrangement of the local occupancy models along the toolpath trajectory at each tool position point. The third set of boxes corresponds to the original space occupancy volume generation step. The semi-transparent tubular entity with a central dashed line inside is formed by taking the union of all instantaneous space occupancy volumes. The original space occupancy body is the swept envelope geometry; the fourth set of boxes corresponds to the superimposed safety margin step, with the blue dashed outline on the outside of the entity representing the outward expansion of the safety margin boundary, indicating that the safety margin is superimposed on the original space occupancy body to obtain the expanded occupancy body; the fifth set of boxes corresponds to the final space occupancy area step, with the complete semi-transparent entity inside representing the final space occupancy area after the margin expansion is completed. The dashed box text in the lower right corner of the box indicates the three uses of this geometry: registering the area with the anti-collision management module, performing space overlap review, and releasing and recycling the area after the task is completed.

[0051] Furthermore, in one embodiment, scheduling parameters are constructed for each machining task unit based on a preset forced machining dependency relationship and a preset machine tool parameter set, resulting in a task scheduling parameter set corresponding to each machining task unit, including: For each processing task unit, the preceding task set and the following task set of the processing task unit are determined based on the preset forced processing dependency relationship; the processing time is measured based on the preset machine tool parameter set and the processing task unit to obtain the basic execution time; the predicted waiting penalty term is added to the basic execution time based on the preset conflict penalty coefficient to obtain the scheduling evaluation time; the task scheduling parameter set corresponding to the processing task unit is obtained based on the preceding task set, the following task set, the basic execution time, and the scheduling evaluation time of the processing task unit.

[0052] After obtaining the processing request space and task scheduling parameter set corresponding to each processing task unit, optimized scheduling is performed to generate an ordered sequence of task units corresponding to each processing channel. First, a dependency constraint graph is constructed based on the preceding and following task sets in each task scheduling parameter set. Then, based on each processing request space, geometric interference checks are used to predict whether any two task units have spatial overlap, and a spatial conflict matrix is ​​constructed. The dependency constraint graph ensures that the sequencing scheme satisfies the process logic constraints, and the spatial conflict matrix quantifies the competition intensity between different task units due to shared space resources.

[0053] The time parameters for each task unit are determined for sorting. The base execution time refers to the estimated execution time determined solely by the toolpath and auxiliary actions of that task unit, without considering conflicts or waiting with other channel task units. The toolpath corresponding to the task unit is extracted from the NC program or CAM toolpath and decomposed according to motion type into rapid traverse segments, linear cutting segments, circular arc segments, and fixed cycle segments. For linear cutting segments, the time is obtained by dividing the segment length by the effective feed rate; for circular cutting segments, the arc length is calculated based on the arc radius and central angle, and then divided by the effective feed rate to obtain the time; for rapid traverse positioning segments, the rapid traverse time is calculated by dividing the rapid traverse length by the effective rapid traverse speed; if multi-axis linkage segments exist, the linear axis motion time and rotary axis motion time are calculated separately, and the larger value is taken as the execution time of that segment. Simultaneously, auxiliary action times, including tool change time, turret indexing time, spindle start / stop time, and clamping / unlocking time, are obtained from the machine tool parameter table or process database and summed to obtain the total auxiliary action time. In addition, a trajectory switching correction time has been added. This is to compensate for dynamic effects such as acceleration / deceleration, corner deceleration, and interpolation transitions. Therefore, the basic execution time of the task unit... Represented as: in The sum of the fast-shift times, This represents the total cutting time. The total time for a fixed cycle.

[0054] The scheduling evaluation time is formed by adding a predicted waiting penalty term on top of the basic execution time. Predicting the waiting penalty item The determination is based on the number of conflicting task units involved in the task unit, the average occupation time of the conflicting task units, and the preset waiting penalty coefficient λ, and is expressed as follows: Where λ is the waiting penalty coefficient. As an indicator of conflict intensity, This represents the average occupation time of relevant conflicting task units. Scheduling evaluation time. The potential spatial conflict cost is quantified into a time increment, which is used to evaluate the urgency of candidate task units during the ranking process.

[0055] Subsequently, based on the dependency constraint graph, spatial conflict matrix, and the aforementioned duration parameters, all processing task units are comprehensively sorted to generate a candidate sorting scheme set. When generating candidate sorting schemes, a topological sort is first performed according to the dependency constraint graph to ensure that the sorting position of any task unit does not violate the order specified by its preceding task set. For parallel candidate task units without mandatory dependencies, multiple possible execution sequences are generated through permutations, combinations, or order swaps.

[0056] For each candidate sorting scheme, the earliest start time and waiting time of each task unit are calculated via simulation along the time axis. The basic ready time of task unit u is defined. Given the maximum value of the idle time of its channel and the completion times of all its preceding tasks; and considering spatial conflict constraints, the earliest start time of task unit u. Pick Release time with conflict zone The larger value in; the waiting time of task unit u. That is and The difference, and the completion time of the task unit. The waiting time evaluation value of the candidate sorting scheme is obtained by summing the waiting times of all task units in each channel of the entire sorting scheme.

[0057] Finally, based on the principle of minimizing the waiting time evaluation value, the target sorting scheme is searched from the candidate sorting scheme set. The search method can employ known combinatorial optimization or heuristic search methods in the field, such as branch and bound with pruning, genetic algorithms, or simulated annealing. (Schedule evaluation time) It can be used to guide heuristic selection or local sorting during the search process, but the final quality of the solution is determined by the cumulative waiting time calculated based on the basic execution time and spatial conflict simulation. After obtaining the target sorting solution, an ordered sequence of task units corresponding to each processing channel is generated accordingly. Each channel will execute the processing task units in sequence according to this sequence, thereby significantly reducing the blocking frequency and invalid waiting time caused by spatial overlap in multi-channel parallel processing at the source.

[0058] Taking the machining of disc-shaped parts in a dual-spindle multi-turret milling and turning center as an example, the channel machining tasks are divided into end-face turning task unit TU1, external turning task unit TU2, and internal turning task unit TU3. Based on process baseline requirements, TU1 is a mandatory prerequisite task for TU2 and TU3, and TU1 points to TU2 and TU3 when constructing the dependency constraint diagram. The spatial conflict matrix reflects the overlapping relationships between the machining application spaces of each task unit and their spaces with other channel task units. The basic execution time of TU1 is calculated as follows: the total length of the end-face turning toolpath divided by the end-face feed rate plus the approach and retraction rapid traverse time; the basic execution time of TU2 is calculated as the total external turning cutting length divided by the external feed rate plus the approach and exit time; and the basic execution time of TU3 is calculated as the total length of the internal turning toolpath divided by the internal feed rate plus the entry and exit time of the hole and the necessary deceleration correction time. If the machining application space of TU2 conflicts significantly with adjacent channels, a higher prediction waiting penalty term is superimposed on it, which increases its scheduling evaluation time accordingly. When generating candidate sorting schemes, since there is no mandatory dependency between TU2 and TU3, multiple candidate sequences such as TU1-TU2-TU3 or TU1-TU3-TU2 can be formed. By simulating and calculating the total waiting time of each candidate sequence for the channel and adjacent channels, the scheme with the minimum total waiting time is selected as the target sorting scheme, ultimately generating an ordered sequence of task units for that channel. The scheduling stage fully considers process sequence constraints and spatial competition relationships, making the actual processing execution smoother and more efficient.

[0059] Furthermore, in one embodiment, the plurality of processing task units are optimized and scheduled based on the processing application space and task scheduling parameter set corresponding to each processing task unit to obtain an ordered sequence of task units corresponding to each processing channel, including: A dependency constraint graph and a spatial conflict matrix are constructed based on the processing application space and task scheduling parameter set corresponding to each processing task unit. The processing task units are then comprehensively sorted based on the preset processing execution time, the dependency constraint graph, and the spatial conflict matrix to obtain a candidate sorting scheme set. The duration of each candidate sorting scheme in the candidate sorting scheme set is estimated to obtain a waiting time evaluation value corresponding to each candidate sorting scheme. An optimal scheme search is performed based on the waiting time evaluation values ​​corresponding to each candidate sorting scheme to obtain a target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on the target sorting scheme.

[0060] First, based on the pre-task set and post-task set in the task scheduling parameter set corresponding to each processing task unit, a dependency constraint graph is constructed to describe the mandatory order between task units. Simultaneously, based on the processing application space corresponding to each processing task unit, geometric interference checks are used to predict whether there is spatial overlap between any two task units, and a spatial conflict matrix is ​​constructed. The dependency constraint graph ensures that the sequencing scheme meets the process logic constraints defined by the preset mandatory processing dependencies, and the spatial conflict matrix quantifies the competition intensity between different task units due to shared processing space. Together, they constitute the hard constraints and optimization basis for scheduling.

[0061] To quantitatively compare the time performance of different sorting schemes, time parameters for sorting are determined for each task unit, and the duration of the task unit is divided into basic execution time and scheduling evaluation time. The basic execution time refers to the estimated execution time determined solely by the toolpath trajectory and auxiliary actions of the task unit itself, without considering conflicts or waiting with other channel task units. The toolpath trajectory corresponding to the task unit is extracted from the NC program or CAM toolpath and decomposed according to motion type into rapid traverse segment, linear cutting segment, circular arc segment, fixed cycle segment, etc. For the linear cutting segment, the machining time is calculated by dividing the segment length by the effective feed rate; for the circular arc cutting segment, the arc length is calculated based on the arc radius and central angle, and then the arc length is divided by the effective feed rate to obtain the machining time; for the rapid traverse positioning segment, the rapid traverse time is calculated by dividing the rapid traverse length by the effective rapid traverse speed; for the multi-axis linkage segment, the linear axis motion time and the rotary axis motion time are calculated separately, and the larger value is taken as the execution time of the segment. Simultaneously, auxiliary action times, including tool change time, turret indexing time, spindle start / stop time, and clamping / unclamping time, are obtained from the machine tool parameter table or process database, and summed to obtain the total auxiliary action time. In addition, a trajectory switching correction time is added to compensate for dynamic effects such as acceleration / deceleration, corner deceleration, and interpolation transition.

[0062] On top of the base execution time, a predicted waiting penalty is superimposed to form the scheduling evaluation time. This predicted waiting penalty is an estimated waiting cost introduced based on the anticipated spatial conflict between the processing space requested by this task unit and the processing space requested by other channel task units. It transforms potential spatial conflicts into a penalty in the time dimension, enabling the sequencing process to proactively avoid highly conflicting tasks. The predicted waiting penalty is determined based on the number of conflicting task units involved in this task unit, the average occupation time of the conflicting task units, and a preset waiting penalty coefficient.

[0063] Subsequently, based on the dependency constraint graph, spatial conflict matrix, and the aforementioned duration parameters, all processing task units are comprehensively sorted to generate a candidate sorting scheme set. When generating candidate sorting schemes, a topological sort is first performed according to the dependency constraint graph to ensure that the sorting position of any task unit does not violate the order specified by its preceding task set. For parallel candidate task units without mandatory dependencies, multiple possible execution sequences are generated through permutations, combinations, or order swaps. For each candidate sorting scheme, the earliest start time and waiting time of each task unit are simulated and calculated along the time axis.

[0064] Finally, based on the minimum waiting time evaluation value, the target sorting scheme is searched from the candidate sorting scheme set. The search method can employ known combinatorial optimization or heuristic search methods in the field, such as branch and bound with pruning, genetic algorithms, or simulated annealing algorithms. After obtaining the target sorting scheme, an ordered task unit sequence corresponding to each processing channel is generated. Each channel executes the processing task units sequentially according to this sequence. This pre-emptively resolves spatial contention and dependency conflicts among multiple channels during the scheduling stage, reducing the blocking frequency and invalid waiting time caused by spatial overlap during runtime, and improving the execution smoothness of the ordered task unit sequence.

[0065] Taking the machining of disc-shaped parts in a dual-spindle multi-turret milling and turning center as an example, the channel machining tasks are divided into end-face turning task unit TU1, external turning task unit TU2, and internal turning task unit TU3. Based on process baseline requirements, TU1 is a mandatory prerequisite task for TU2 and TU3. When constructing the dependency constraint diagram, TU1 points to TU2 and TU3 respectively. The spatial conflict matrix reflects the overlapping relationships between the machining application spaces of each task unit and their spaces with other channel task units. The basic execution time of TU1 is calculated as follows: the total length of the end-face turning toolpath divided by the end-face feed rate plus the approach and retraction rapid traverse time; the basic execution time of TU2 is calculated as the total external turning cutting length divided by the external feed rate plus the approach and exit time; the basic execution time of TU3 is calculated as the total length of the internal turning toolpath divided by the internal feed rate plus the entry and exit time of the hole and the necessary deceleration correction time. If the machining application space of TU2 conflicts significantly with adjacent channels, a higher prediction waiting penalty term is superimposed on it, which increases its scheduling evaluation time accordingly. When generating candidate sorting schemes, since there is no mandatory dependency between TU2 and TU3, multiple candidate sequences can be formed, such as TU1 preceding TU2 preceding TU3 or TU1 preceding TU3 preceding TU2. By simulating and calculating the total waiting time of each candidate sequence for the channel and adjacent channels, the scheme with the minimum total waiting time is selected as the target sorting scheme, ultimately generating the ordered task unit sequence for that channel. The scheduling stage fully incorporates process sequence constraints and spatial competition relationships to ensure the high efficiency and safety of actual processing execution.

[0066] Further, in one embodiment, an optimal solution search is performed based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain a target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on the target sorting scheme, including: Optimal scheme search is performed based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain an initial sorting scheme; waiting ratio determination is performed based on a preset waiting ratio threshold and the initial sorting scheme to obtain a waiting ratio determination result; when the waiting ratio determination result indicates that the duration exceeds the standard, a high-conflict task unit set is determined based on the spatial conflict matrix and several processing task units, and the high-conflict task unit set is further subdivided into several subdivided processing task units; spatial modeling, scheduling parameter construction, and scheduling optimization are performed again based on several subdivided processing task units and the remaining several processing task units to obtain an updated target sorting scheme, and waiting ratio determination is performed again based on the updated target sorting scheme until the waiting ratio determination result indicates that the duration does not exceed the standard; the updated target sorting scheme is determined as the target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on the target sorting scheme.

[0067] First, an optimal solution search is performed based on the estimated waiting time values ​​of each candidate sorting scheme to obtain an initial sorting scheme. This search process aims to minimize the overall waiting time and employs a combined optimization method guided by a dependency constraint graph and a spatial conflict matrix. After dividing the processing tasks of each channel into several task units, the estimated execution time, spatial occupancy area, pre- and post-task dependencies, and priority information of each task unit are read to form a set of task units. A directed dependency constraint graph is established based on the relationship between pre- and post-task tasks, generating feasible candidate sequences that satisfy the dependency constraints. A conflict relationship matrix is ​​established based on the spatial occupancy area corresponding to each task unit; if the spatial occupancy areas of two task units overlap after interference checking, a conflict relationship is recorded.

[0068] Each candidate sequence is evaluated, and its execution process is simulated along the timeline, taking into account the expected execution time of each task unit and the regional conflict relationship. Indicators such as the waiting time of each channel, the start and end times of each task unit, the cumulative waiting time, and the percentage of waiting time are calculated. The objective function value for each candidate sequence is calculated, and the sorting scheme corresponding to the optimal value is selected as the initial sorting scheme. The objective function can take various forms: minimizing the total waiting time is the goal, where the waiting time of the i-th task unit due to regional conflict under sorting scheme S is... Then the objective function J is the waiting time of all task units. The sum: Alternatively, with the goal of minimizing the percentage of waiting time, let the total processing time be... Total waiting time is Then the objective function J is: Alternatively, a weighted objective function can be used, which sums the total waiting time, the percentage of waiting time, and the conflict cost in a weighted manner, with each weight coefficient set according to the actual process requirements.

[0069] After obtaining the initial sorting scheme, the waiting percentage is determined based on a preset waiting percentage threshold and the initial sorting scheme. The preset waiting percentage threshold is the proportion of the maximum allowable waiting time to the total processing time, pre-set by the system according to production cycle requirements, reflecting the tolerable upper limit of processing efficiency loss. The ratio of the total waiting time under the initial sorting scheme to the total processing time is calculated and compared with the preset waiting percentage threshold to obtain the waiting percentage determination result. If the determination result indicates that the time exceeds the limit, it means that the granularity of the current task unit division is not sufficient to effectively resolve spatial competition. Then, a high-conflict task unit set is determined based on the spatial conflict matrix and several current processing task units. The high-conflict task unit set refers to the set of task units that have high spatial conflict intensity with other task units, long occupation time, and significant contribution to the total waiting time. For this set of high-conflict task units, the execution unit further subdivides each high-conflict task unit into several subdivided processing task units with shorter execution time and smaller space occupation range. This shortens the duration of single space occupation and reduces the area occupied by single space occupation, thereby reducing the probability of long-term spatial overlap with other channel task units and providing more granular scheduling objects for subsequent rescheduling.

[0070] After obtaining several subdivided processing task units, these are merged with the remaining unsubdivided processing task units. The entire process of spatial modeling, scheduling parameter construction, and optimized scheduling is then re-executed. This involves rebuilding the task unit set, reconstructing the dependency constraint graph and conflict relationship matrix, and again evaluating candidate sequences and comparing them with the objective function to obtain an updated target sorting scheme. The waiting percentage is then determined again for the updated target sorting scheme. This process is iterated until a waiting percentage determination result shows that the time has not exceeded the limit, i.e., the total waiting time percentage has dropped below the preset waiting percentage threshold, or the preset maximum number of iterations has been reached. At this point, the updated target sorting scheme that finally meets the waiting percentage requirement is determined as the target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on this target sorting scheme.

[0071] In this embodiment, dynamic matching between task unit partitioning granularity and spatial conflict intensity is achieved. When the waiting ratio in the initial scheduling exceeds the standard, this embodiment automatically identifies and refines those high-conflict task units that cause serious blocking, gradually reducing the overall waiting time in the iteration until the efficiency requirements are met; when the waiting ratio does not exceed the standard, an ordered task unit sequence is directly output, effectively controlling the waiting cost of multi-channel parallel processing in the scheduling stage, and ensuring the continuity and efficiency of processing execution.

[0072] Furthermore, in one embodiment, the processing application space is reviewed for overlap based on a preset set of occupied space to obtain a space review result, including: Based on a preset safety margin, the processing application space and the preset occupied space set are respectively expanded outward to obtain the corresponding expanded application space and expanded occupied space set, in order to compensate for the spatial uncertainty caused by thermal deformation, clamping error and control lag during processing; based on a preset axis alignment bounding box algorithm, the expanded application space and the expanded occupied space set are coarsely bounded box judged to obtain a candidate overlapping pair list, and obviously non-intersecting space pairs are quickly screened out; based on a preset geometric intersection judgment algorithm, the candidate overlapping pair list is precisely overlap judged to obtain the space review result.

[0073] After receiving a space usage application, the anti-collision management module employs a layered, progressive space overlap determination mechanism for review. First, based on a preset safety margin, the application space and each used space in the preset occupied space set are expanded outwards to obtain the corresponding expanded application space and expanded occupied space set. The preset safety margin is a space buffer value set to compensate for potential thermal deformation, clamping errors, control lag, model errors, and braking margins during processing. A safety buffer layer is reserved outside the nominal geometric boundary, ensuring that even with the presence of these disturbances, the review results guarantee that no physical collisions will occur during actual processing.

[0074] Subsequently, a preliminary bounding box algorithm is used to coarsely determine the bounding box structure of the extended application space and the extended occupied space set, resulting in a candidate overlapping pair list. In this embodiment, for each used space in the extended application space and the extended occupied space set, a separate axis-aligned bounding box is generated. If the projection intervals of two bounding boxes are completely separated in any of the X, Y, or Z coordinate directions (i.e., the maximum value of one bounding box in that direction is less than the minimum value of the other bounding box), they are directly determined not to overlap. Only when the projection intervals of two bounding boxes intersect in all three coordinate directions are the space pairs included in the candidate overlapping pair list. By quickly filtering out most obviously non-overlapping space pairs through the coarse bounding box determination, the number of objects for subsequent precise determination is significantly reduced, thus lowering the computational burden under multi-channel concurrent conditions.

[0075] For each spatial pair that enters the candidate overlap pair list through coarse judgment, a precise overlap determination is performed based on a preset geometric intersection determination algorithm. It should be noted that the precise geometric intersection determination can employ polyhedral intersection determination based on the separating axis theorem, Boolean intersection determination of swept or stretched bodies, or boundary representation volume intersection determination based on triangular meshes. In a preferred embodiment, if both the processing application space and the used space are expressed using closed contour stretched bodies, the precise judgment process consists of two steps: the first step is planar projection determination, transforming the bottom projection contours of the two stretched bodies to the same reference plane to determine whether the two closed contours intersect or have an inclusion relationship; the second step is stretching height interval determination, determining whether the height intervals of the two stretched bodies in the stretching direction intersect. Only when the bottom projection contours intersect or contain each other and the height intervals in the stretching direction also intersect are the two spatial objects ultimately determined to overlap.

[0076] After comparing the requested space with all currently used spaces one by one, the space review results are generated. If, through precise overlap determination, there is an overlap between the requested space and any used space, the space review result is determined to be overlapping, the application is rejected, the requested space is marked as a space to be used, and the corresponding processing task unit is written into a preset waiting sequence; if there is no overlap between the requested space and any used space, the space review result is determined to be overlapping, the application is approved, the requested space is marked as a space to be used and added to a preset occupied space set, and an approval signal is returned to the corresponding processing channel.

[0077] This embodiment takes a dual-spindle multi-turret milling and turning center as an example, assuming that a task unit in channel one submits a usage application. The anti-collision management module first reads the machining application space corresponding to the task unit from the regional registration library according to the channel number and task unit identifier, and converts it to the global coordinate system along with all currently used spaces. Then, it applies a safety margin expansion to all space objects. Next, it quickly filters out obviously non-overlapping objects through axis alignment bounding box coarse judgment, and only performs precise judgment on candidate overlapping pairs based on the projection of the bottom surface of the stretched body and the height range. Finally, it decides whether to approve the application or add it to the waiting sequence based on the judgment result. Through the layered mechanism of providing engineering safety margin through safety margin expansion, reducing the calculation scale through bounding box coarse judgment, and ensuring judgment accuracy through precise intersection judgment, the space overlap review can meet the reliability requirements while having the ability to be executed online in real time, providing a basic guarantee for the safety and efficiency of multi-channel parallel machining.

[0078] Furthermore, in one embodiment, based on the updated preset occupied space set, the processing application space corresponding to each processing task unit in the preset waiting sequence is checked for space overlap one by one to obtain the waiting space check result, including: The preset waiting sequence is determined to be non-empty. The waiting time of each processing task unit is obtained. Based on the preset aging coefficient, the preset basic process priority, and the waiting time of each processing task unit, the preset waiting sequence is traversed to determine the effective priority of each processing task unit. The preset waiting sequence is reordered based on the effective priority of each processing task unit to obtain the re-examination waiting sequence. Based on the updated preset occupied space set, the processing application space corresponding to each processing task unit in the re-examination waiting sequence is reviewed for space overlap one by one to obtain the waiting space review result.

[0079] After completing the current processing task unit and submitting a release request in any processing channel, the anti-collision management module responds to the release signal of that channel, determines the corresponding processing release space, and updates the preset occupied space set. If the preset waiting sequence is determined to be non-empty, the processing request space corresponding to each processing task unit in the waiting sequence is checked for space overlap one by one based on the updated preset occupied space set to obtain the waiting space check result.

[0080] Specifically, firstly, the waiting time of each processing task unit in the preset waiting sequence is obtained. This waiting time is calculated by subtracting the enqueue time of the task unit when it entered the waiting sequence from the current re-examination time. Then, based on the preset aging coefficient, the preset basic process priority, and the waiting time of each processing task unit, the preset waiting sequence is traversed and the effective priority of each processing task unit is determined.

[0081] It should be noted that the preset aging coefficient is a rate factor used to convert waiting time into priority gain, reflecting the system's sensitivity to the increased urgency of task units due to accumulated waiting time; the preset basic process priority is the inherent priority of task units pre-assigned by factors such as process criticality and cycle time sensitivity. By multiplying the waiting time by the preset aging coefficient as a dynamic gain and adding it to the preset basic process priority, waiting task units that have not received approval for a long time are given a higher effective priority, avoiding indefinite postponement of execution due to continuous spatial conflicts.

[0082] After determining the effective priority of each processing task unit, the preset waiting sequence is reordered according to the effective priority to obtain the re-examination waiting sequence. During sorting, it is preferable to arrange them in descending order of effective priority; when multiple task units have the same effective priority, they are further arranged according to the order in which they entered the waiting sequence to ensure fairness of first-come, first-served under the same priority conditions.

[0083] Subsequently, based on the updated preset occupied space set, the processing application space corresponding to each processing task unit is reviewed for overlap according to the order in the re-review waiting sequence. This review process preferably adopts an atomic execution method of approval one by one and immediate update. That is, for each task unit awaiting review in the re-review waiting sequence, it is sequentially determined whether its processing application space overlaps with the currently used space set. Once it is determined that the processing application space of a task unit does not overlap with any currently used space, the application is immediately approved, its processing application space is marked as used space and immediately added to the preset occupied space set, and then the review of the next task unit in the re-review waiting sequence continues. If the processing application space of a task unit still overlaps with the currently used space, its position in the waiting sequence is retained, awaiting the triggering of the next space release event. By updating the used space set after each approval, it is ensured that subsequent reviews in the same release event are always based on the latest occupied state, fundamentally eliminating the potential risk of simultaneously approving two pending applications with conflicting space in the same re-review process.

[0084] In the waiting queue re-examination stage after space release, the dynamic calculation and sorting of effective priorities give long-waiting task units the opportunity for priority re-examination, maintaining the fairness of multi-channel scheduling; the atomic execution mechanism of reviewing and updating the occupied space set one by one maximizes the use of available space resources released by each space release while ensuring space mutual exclusion security, thereby improving the overall responsiveness and processing efficiency of the multi-channel parallel processing system.

[0085] Further, please see Figure 5 In one embodiment, the space-requesting scheduling method for a multi-channel parallel machining center further includes: During multi-channel parallel machining, the actual axis position feedback information of each machining channel is acquired in real time. This information is collected and reported in real time by each axis servo driver through the encoder interface, representing the current actual position of each motion axis in a unified coordinate system.

[0086] Based on the actual axis position feedback information of each machining channel, the actual axial distance between corresponding axes in adjacent machining channels on the machine tool structure is determined. Adjacent machining channels refer to pairs of channels in the physical layout of the machine tool whose execution component movement areas are adjacent and may approach or interfere with each other. Corresponding axes refer to axes that have the same motion meaning between these channels and can be mapped to a unified coordinate system for distance comparison, including linear axes and rotary axes. For linear axes, the actual axial distance is defined as the absolute value of the difference between the positions of corresponding axes in two channels; for rotary axes, the actual axial distance is defined as the absolute value of the difference between the angular positions of corresponding axes in two channels, and can be converted into equivalent linear distance using equivalent radii.

[0087] Minimum safety distances and warning distances are pre-set for the corresponding axes of each pair of adjacent channels. The minimum safety distance consists of a static safety clearance and a dynamic additional safety clearance. The static safety clearance compensates for inherent geometric occupancy and errors, while the dynamic additional safety clearance prevents dangerous situations where timely braking is impossible due to relative approach speed and control delay. The static safety clearance is determined comprehensively based on the minimum structural clearance of the mechanical body, the outward expansion of the tool or actuator, the additional clearance between the workpiece and the fixture, measurement and control error margins, thermal deformation compensation, and fixed safety margins. The dynamic additional safety clearance is calculated based on the relative approach speed of the two channels along that axis, the total delay time from detection to control activation, and the equivalent deceleration, including the sum of the distance traveled during the control delay and the relative braking distance. Since the structural shapes and braking conditions of adjacent channels may differ in the forward and reverse directions, it is preferable to set separate minimum safety distances for the forward and reverse directions. The warning distance is based on the minimum safety distance with an added warning buffer, used to trigger deceleration intervention in advance before danger approaches.

[0088] During the machining process, the calculated actual axial distance is compared with a preset distance threshold in real time. When the actual axial distance of any corresponding axis is less than the corresponding preset warning distance, a deceleration command is generated, causing the relevant channel to automatically reduce its movement speed along the approach direction to slow down the approach trend and buy time for subsequent shutdown. When the actual axial distance further decreases to less than the corresponding preset safety distance, a shutdown protection command is generated, causing the multi-channel parallel machining tool to stop immediately, completely blocking the actual collision risk from the bottom layer.

[0089] While generating shutdown protection commands or speed reduction commands, the control system automatically saves the current machine tool processing status. To support consistent recovery after abnormal shutdowns, it is preferable to save a status snapshot at the task unit boundary, which includes at least the channel number, the identifier of the currently executing or pending task unit, the current position of each channel or its private waiting area pose, the set of currently approved occupied space areas, the complete status of the preset waiting sequence, the current value of the waiting timer, and the set of currently successfully completed task units. Setting the status snapshot point at the task unit boundary is because the task unit has a complete application, approval, processing, and release lifecycle, and using this as a recovery benchmark can naturally reconstruct the correspondence between the task unit, the area occupancy status, and the channel status.

[0090] Once the anomaly is resolved, in response to the machine tool recovery signal, the multi-channel parallel machining center is restored to its current state based on the previously saved machining status. During restoration, each channel is restored to its pose at the boundary of its corresponding task unit according to the most recent valid state snapshot. The area occupancy state and waiting queue state are re-established, and the area application and execution process is restarted from the boundary of that task unit. This allows each machining channel to continue machining parts from the saved point, avoiding secondary anomalies caused by state loss or misalignment. This improves anomaly recovery efficiency and operational continuity while ensuring machining safety.

[0091] It should be noted that, Figure 5 This diagram illustrates the principle of graded protection for monitoring the distance between corresponding axes in adjacent channels. The rectangles at the top represent the axes corresponding to channel i and channel j, respectively. These are the corresponding axes of adjacent processing channels. Blue arrows indicate that the two axes are moving towards each other. The vertical dashed lines represent the current position references for the two axes, and the dashed lines between the two references indicate the actual distance D. Blue bidirectional dashed lines indicate the warning distance. The red double-sided dashed line indicates the minimum safe distance. The horizontal axis at the bottom of the diagram points in the direction of increasing distance. The axis divides the system into three sections: green for normal operation, yellow for warning and deceleration, and red for shutdown protection. The dividing points correspond to... The two dashed boxes on the right side of the figure record the distance calculation formula and the definition of each distance parameter, respectively. The dashed box at the bottom indicates the device control triggering rules corresponding to different actual distances.

[0092] It should be noted that the global parameters involved in this embodiment, including attitude change threshold, toolpath offset length, envelope error threshold, maximum waiting time threshold, aging coefficient, early warning buffer, area safety expansion, and objective function weight coefficient, are all determined by the machine tool parameter library, process parameter library, offline simulation results, or measured calibration data. When not given in advance in specific implementation, they can be calibrated and set according to the machine tool structure, control cycle, and safety requirements to ensure the adaptability and reliability of the entire scheduling and protection system under different machine tools and working conditions.

[0093] Please see Figure 6 This embodiment also provides a space-request scheduling system for multi-channel parallel machining tools, including: The task unit division module is used to acquire part processing data, generate channel processing tasks corresponding to each processing channel in a multi-channel parallel machining center based on the part processing data, and divide the channel processing tasks into several processing task units so that each processing channel can perform part processing based on the corresponding several processing task units. The spatial overlap review module is used to determine the processing application space based on the processing task unit when any processing channel executes to a processing task unit, and to perform spatial overlap review on the processing application space based on a preset occupied space set to obtain the spatial review result; The waiting sequence management module is used to determine that the space review result indicates overlap, and then add the processing task unit to a preset waiting sequence; A space release response module is used to respond to a channel release signal of any processing channel, determine the processing release space based on the channel release signal, and update the preset occupied space set based on the processing release space; The waiting sequence re-examination module is used to perform space overlap review on the processing application space corresponding to each processing task unit in the preset waiting sequence based on the updated preset occupied space set, and obtain the waiting space review result. The scheduling response module is used to perform scheduling response processing on each processing task unit in the preset waiting sequence based on the waiting space review result, so as to realize multi-channel parallel processing.

[0094] In this embodiment, the task unit division module acquires part processing data and generates channel processing tasks corresponding to each processing channel, dividing the channel processing tasks into several processing task units, providing independent and manageable objects for space management; the space overlap review module determines the processing application space based on the processing task unit when the processing channel executes to the processing task unit, and performs space overlap review with the preset occupied space set to prevent conflicting tasks from entering the shared space; the waiting sequence management module adds the processing task unit to the preset waiting sequence when the space review result shows overlap, so that conflicting tasks are queued in an orderly manner instead of being directly stopped; the space release response module responds to the channel release signal, determines the processing release space and updates the preset occupied space set; the waiting sequence re-review module performs space overlap review on each processing task unit in the preset waiting sequence one by one based on the updated preset occupied space set; the scheduling response module performs scheduling response processing on each processing task unit in the preset waiting sequence based on the waiting space review result. The modules work together to form a closed loop of conflict detection, queuing, release re-review and scheduling response, which reduces unplanned downtime and invalid waiting caused by space overlap while ensuring the security of the preset occupied space set, thereby improving the execution efficiency and safety of multi-channel parallel processing.

[0095] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A space-request-based scheduling method for multi-channel parallel machining centers, characterized in that, Includes the following steps: Acquire part processing data, generate channel processing tasks corresponding to each processing channel in a multi-channel parallel machining center based on the part processing data, and divide the channel processing tasks into several processing task units so that each processing channel processes the part based on the corresponding several processing task units; When any processing channel executes to a processing task unit, a processing application space is determined based on the processing task unit, and a space overlap review is performed on the processing application space based on a preset occupied space set to obtain a space review result; If the spatial review result indicates that there is overlap, the processing task unit is added to a preset waiting sequence; In response to a channel release signal of any processing channel, a processing release space is determined based on the channel release signal, and a preset occupied space set is updated based on the processing release space; Based on the updated preset occupied space set, the processing application space corresponding to each processing task unit in the preset waiting sequence is checked for space overlap one by one to obtain the waiting space check result; Based on the waiting space review results, the processing task units in the preset waiting sequence are scheduled and responded to to achieve multi-channel parallel processing.

2. The spatial request-based scheduling method for a multi-channel parallel machining center according to claim 1, characterized in that, The process of dividing the channel processing task into several processing task units includes: For each channel processing task, the process attributes are segmented based on the process switching point of the processing task in that channel to obtain a process segmentation task set; The process segmentation task segment set is physically boundary-segmented to obtain a physically boundary-segmented task segment set. The physical boundary segmentation task segment set is spatially stable segmented to obtain a spatially stable segmentation task segment set. Deadlock avoidance segmentation is performed based on a preset occupation time threshold and the spatial stability segmentation task set to obtain a deadlock avoidance segmentation task set. The deadlock avoidance segmentation task segment set is subjected to anomaly recovery boundary segmentation to obtain anomaly recovery boundary segmentation task segment set; The set of abnormal recovery boundary segmentation task segments is divided into regions by application and alignment to obtain the plurality of processing task units.

3. The space-request scheduling method for a multi-channel parallel machining center according to claim 1, characterized in that, The step of dividing the channel processing task into several processing task units, so that each processing channel processes parts based on the corresponding several processing task units, includes: After dividing the channel processing task into several processing task units, spatial modeling is performed on each processing task unit based on a preset machine tool parameter set to obtain the processing application space corresponding to each processing task unit. Based on the preset forced processing dependencies and the preset machine tool parameter set, the scheduling parameters of each processing task unit are constructed to obtain the task scheduling parameter set corresponding to each processing task unit; Based on the processing application space and task scheduling parameter set corresponding to each processing task unit, the scheduling of the plurality of processing task units is optimized to obtain an ordered task unit sequence corresponding to each processing channel, so that each processing channel executes each processing task unit in sequence based on the ordered task unit sequence, thereby processing the parts.

4. The space-request scheduling method for a multi-channel parallel machining center according to claim 3, characterized in that, Based on preset forced machining dependencies and preset machine tool parameter sets, scheduling parameters are constructed for each machining task unit to obtain the task scheduling parameter set corresponding to each machining task unit, including: For each processing task unit, the preceding task set and the following task set of the processing task unit are determined based on the preset forced processing dependency relationship; Based on the preset machine tool parameter set and the machining task unit, the machining time is measured to obtain the basic execution time; The scheduling evaluation time is obtained by superimposing a predicted waiting penalty term on the basic execution time based on a preset conflict penalty coefficient. The task scheduling parameter set corresponding to the processing task unit is obtained based on the pre-task set, post-task set, basic execution time, and scheduling evaluation time of the processing task unit.

5. The spatial request-based scheduling method for a multi-channel parallel machining center according to claim 3, characterized in that, The optimization scheduling of the plurality of processing task units based on the processing request space and task scheduling parameter set corresponding to each processing task unit yields an ordered sequence of task units corresponding to each processing channel, including: Construct a dependency constraint graph and a space conflict matrix based on the processing application space and task scheduling parameter set corresponding to each processing task unit; Based on the preset processing execution time, the dependency constraint graph, and the spatial conflict matrix, the plurality of processing task units are comprehensively sorted to obtain a set of candidate sorting schemes; The waiting time evaluation value corresponding to each candidate sorting scheme in the candidate sorting scheme set is obtained by estimating the duration of each candidate sorting scheme. The optimal solution is searched based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain the target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on the target sorting scheme.

6. The space-request scheduling method for a multi-channel parallel machining center according to claim 5, characterized in that, The process of searching for the optimal solution based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain the target sorting scheme, and generating an ordered task unit sequence corresponding to each processing channel based on the target sorting scheme, includes: The optimal solution is searched based on the waiting time evaluation value corresponding to each candidate sorting scheme to obtain the initial sorting scheme; The waiting percentage is determined based on a preset waiting percentage threshold and the initial sorting scheme, and the waiting percentage determination result is obtained. When the waiting percentage determination result is that the duration exceeds the standard, a set of high-conflict task units is determined based on the spatial conflict matrix and several processing task units, and the set of high-conflict task units is further subdivided to obtain several subdivided processing task units. Based on several subdivided processing task units and several other processing task units, spatial modeling, scheduling parameter construction, and scheduling optimization are performed again to obtain an updated target sorting scheme. Then, based on the updated target sorting scheme, the waiting ratio is re-determined until the waiting ratio determination result is that the duration has not exceeded the standard. The updated target sorting scheme is determined as the target sorting scheme, and an ordered task unit sequence corresponding to each processing channel is generated based on the target sorting scheme.

7. The space-request scheduling method for a multi-channel parallel machining center according to claim 1, characterized in that, Based on a preset set of occupied space, the processing application space is subjected to space overlap review to obtain space review results, including: Based on a preset safety margin, the processing application space and the preset occupied space set are respectively expanded outward to obtain the corresponding expanded application space and expanded occupied space set; Based on the preset axis-aligned bounding box algorithm, a rough bounding box judgment is performed on the set of the outward application space and the outward occupied space to obtain a candidate overlapping pair list; The candidate overlapping pair list and the outer space occupied set are accurately overlapped based on the preset geometric intersection determination algorithm to obtain the space review result.

8. The space-request scheduling method for a multi-channel parallel machining center according to claim 1, characterized in that, The process application space corresponding to each processing task unit in the preset waiting sequence is checked for overlap one by one based on the updated preset occupied space set to obtain the waiting space check result, including: Determine that the preset waiting sequence is not empty, obtain the waiting time of each processing task unit, and traverse the preset waiting sequence based on the preset aging coefficient, the preset basic process priority and the waiting time of each processing task unit to determine the effective priority of each processing task unit. The preset waiting sequence is reordered based on the effective priority of each processing task unit to obtain the re-examination waiting sequence; Based on the updated preset occupied space set, the processing application space corresponding to each processing task unit in the re-examination waiting sequence is reviewed one by one for space overlap to obtain the waiting space review result.

9. The space-request scheduling method for a multi-channel parallel machining center according to claim 1, characterized in that, Also includes: During multi-channel parallel machining, the actual axis position feedback information of each machining channel is obtained in real time; The actual axial distance between corresponding axes in adjacent machining channels is determined based on the actual axis position feedback information of each machining channel. When the actual axial distance is less than the preset safety distance, a shutdown protection command is generated to stop the multi-channel parallel machining tool from running. When the actual axial distance is less than the preset warning distance, a speed reduction operation command is generated to cause the multi-channel parallel machining tool to reduce its speed. The current machine tool processing status is saved based on the shutdown protection command or the speed reduction operation command; In response to the machine tool recovery signal, and based on the machine tool recovery signal and the current machine tool processing state, the state of the multi-channel parallel machining machine tool is restored so that each processing channel can continue to process parts.

10. A space-request scheduling system for multi-channel parallel machining tools, characterized in that, include: The task unit division module is used to acquire part processing data, generate channel processing tasks corresponding to each processing channel in a multi-channel parallel machining center based on the part processing data, and divide the channel processing tasks into several processing task units so that each processing channel can perform part processing based on the corresponding several processing task units. The spatial overlap review module is used to determine the processing application space based on the processing task unit when any processing channel executes to a processing task unit, and to perform spatial overlap review on the processing application space based on a preset occupied space set to obtain the spatial review result; The waiting sequence management module is used to determine that the space review result indicates overlap, and then add the processing task unit to a preset waiting sequence; A space release response module is used to respond to a channel release signal of any processing channel, determine the processing release space based on the channel release signal, and update the preset occupied space set based on the processing release space; The waiting sequence re-examination module is used to perform space overlap review on the processing application space corresponding to each processing task unit in the preset waiting sequence based on the updated preset occupied space set, and obtain the waiting space review result. The scheduling response module is used to perform scheduling response processing on each processing task unit in the preset waiting sequence based on the waiting space review result, so as to realize multi-channel parallel processing.

Images