Tooling management system based on mes

Abstract

The application relates to the field of automatic control technology, and particularly discloses a tooling use management system based on MES, which receives and stores pre-action declarations of a distributed control unit by establishing a dynamic intention coordination library in the MES; when an abnormal signal contradictory to a tooling logic state is perceived, conflict identification is performed based on the pre-action declaration, and a logic consistency measurement value is calculated; when the measurement value exceeds a dynamic decision threshold, a virtual resource occupation identifier is created to block a physical scheduling instruction, and a proactive perception verification process is started; finally, state synchronization is performed according to a verification result; if it is a false alarm, the virtual identifier is converted into a formal resource allocation; and if it is true, the identifier is removed and the pre-action declaration is re-coordinated; through the intention coordination and virtual resource resolution mechanism, the application solves the positive feedback oscillation problem caused by local faults of a distributed control system, and improves system stability and resource utilization efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automation control, and particularly relates to a tooling usage management system based on MES. BACKGROUND

[0002] In the intelligent manufacturing environment, the manufacturing execution system (MES) as the core connecting the planning layer and the control layer needs to effectively manage the tooling resources. The existing tooling management system based on MES usually adopts a distributed control architecture, and each control unit (such as a robot controller, an AGV scheduling system, etc.) autonomously decides according to local state information, acquires the tooling state through a perception network and responds. Although this architecture improves the flexibility of the system, its control logic is mainly based on the "state-reaction" mode, that is, the system directly triggers the corresponding scheduling instruction according to the real-time state fed back by the sensor. In the actual running process, the system realizes the allocation and scheduling of the tooling through basic functions such as tooling identification, state monitoring and instruction issuing, and forms a relatively solid tooling management process.

[0003] The existing technology has the following deficiencies:

[0004] When the system scale expands and the number of control units increases, due to the lack of global coordination mechanism of each distributed control unit, the instantaneous failure of a single perception unit may cause "positive feedback oscillation" between multiple autonomous decision-making subsystems. Specifically, the false alarm signal of a certain sensor (such as an RFID reader) will be transmitted between the MES tooling management, AGV scheduling and robot control subsystems, and the responses made by each system based on local decision logic will excite each other and continuously amplify. This system-level oscillation problem has the characteristics of strong concealment, fast propagation speed and wide influence range, and since it is caused by the defect of the system architecture rather than a simple component failure, it is difficult to effectively solve by traditional methods. SUMMARY

[0005] The purpose of the present application is to provide a tooling usage management system based on MES to solve the problems in the background.

[0006] The purpose of the present application can be achieved by the following technical solutions:

[0007] The tooling usage management system based on MES comprises:

[0008] An intention coordination module is configured to establish and maintain a dynamic intention coordination library in the MES, and receive and store pre-action declarations from multiple distributed control units, wherein each pre-action declaration contains an identifier of a target tooling, a type of planned operation and a timestamp.

[0009] The intelligent perception and conflict identification module, when the MES receives an abnormal signal contradicting the current logical state of the target tooling through the perception network, takes the pre-action declaration stored in the dynamic intention coordination library as a comparison benchmark, triggers the conflict identification mechanism, and calculates a generated logical consistency metric value;

[0010] The dynamic decision and resource resolution module takes the generated logical consistency metric value as a judgment basis, and when the logical consistency metric value exceeds a preset dynamic decision threshold, the following processing is performed:

[0011] Based on the conflict identification result, a virtual resource occupation identifier is created for the relevant distributed control unit in the resource management layer of the MES, and based on this identifier, the target tooling re-deployment instruction issued to the physical execution unit is blocked;

[0012] Based on the established virtual resource occupation identifier, an active perception verification process is started, and a diagnostic device is dispatched to immediately confirm the real physical state of the target tooling;

[0013] The state convergence and system synchronization module takes the confirmation result returned by the active perception verification process as input and performs the following operations:

[0014] When the abnormal signal is a false alarm, the virtual resource occupation identifier is converted into a formal resource allocation record;

[0015] When the abnormal signal is true, the virtual resource occupation identifier is removed and the pre-action declaration is re-coordinated according to the actual resource state.

[0016] As a further scheme of the present application, the dynamic intention coordination library is established and maintained, specifically including:

[0017] A tooling use intention graph based on a time sequence relationship is constructed, the tooling use intention graph takes a target tooling identifier as a node and a time sequence of a pre-action declaration as an edge, forming a networked structure with a time-dependent relationship;

[0018] The pre-action declarations from different distributed control units are spatio-temporally aligned and conflict-resolved to generate a unified intention description framework;

[0019] The validity state of the pre-action declaration is dynamically maintained based on the timestamp attribute of the pre-action declaration, expired pre-action declarations are automatically subjected to archiving and invalidation processing, and real-time consistency maintenance is performed on the dynamic intention coordination library.

[0020] As a further scheme of the present application, the calculation of the generated logical consistency metric value specifically includes:

[0021] A multi-dimensional evidence chain is established to time-space align all pre-action declarations of the corresponding target tool in the dynamic intention coordination library with the received abnormal signal, and to extract features in three dimensions of time overlap, operation mutual exclusivity and declaration confidence level;

[0022] A confidence accumulation calculation is performed within a time window, a degree of compliance of each pre-action declaration is weighted and accumulated based on a time sequence relationship between a timestamp of the pre-action declaration and a time of occurrence of the abnormal signal, and an initial consistency score is generated;

[0023] A conflict propagation depth analysis is implemented, a range of chain reactions possibly triggered by the abnormal signal is evaluated by tracking a dependency relationship network between the pre-action declarations, and the initial consistency score is dynamically revised based on a propagation depth;

[0024] A final logical consistency measurement value is output.

[0025] As a further scheme of the application, the virtual resource occupation identifier is created for the relevant distributed control unit in the resource management layer of the MES, and specifically comprises:

[0026] The virtual resource occupation identifier with a time limit attribute is created, and the virtual resource occupation identifier contains a target tool identifier, an occupation start time, a predicted duration and a credibility score;

[0027] A decision tree of instruction blocking is established, and different levels of physical scheduling instructions of complete blocking, delayed execution or conditional release are dynamically selected according to the credibility score and the time limit attribute of the virtual resource occupation identifier;

[0028] A resource state mirror synchronization is implemented, and a logical resource view parallel to a physical tool state is established in the MES resource management layer while maintaining the virtual resource occupation;

[0029] When the predicted duration is reached or a verification result is received, a clearing process of the virtual resource occupation identifier is automatically triggered.

[0030] As a further scheme of the application, the decision tree of instruction blocking specifically comprises:

[0031] A decision condition evaluation layer is constructed, the virtual resource occupation identifier is divided into three confidence levels of high, medium and low based on the credibility score, and the occupation duration is divided into three time intervals of short-term, medium-term and long-term according to the time limit attribute;

[0032] Differentiated blocking strategies are configured for combinations of different confidence levels and time intervals, including completely blocking all physical scheduling instructions, delaying the execution of non-critical instructions, and releasing instructions with emergency priority;

[0033] According to the system real-time load state and production task emergency degree, the judgment threshold and the blocking strength of each condition in the decision tree are dynamically adjusted;

[0034] By monitoring the system state change after the instruction blocking, the strategy parameters in the decision tree are automatically corrected, and adaptive optimization of the blocking strategy is realized.

[0035] As a further scheme of the application, the starting active sensing verification process specifically includes:

[0036] Based on the logical consistency measurement value and the key level of the target tooling, the emergency degree of the verification task is calculated and the execution order is dynamically adjusted;

[0037] When multiple verification tasks are concurrent, according to the real-time working load and the moving path of the diagnostic equipment, the appropriate diagnostic equipment is intelligently matched to execute verification;

[0038] The visual recognition equipment, the radio frequency identification reader and the Internet of Things sensor jointly participate in the verification process, and the reliability of the confirmation result is improved through multi-source data cross verification;

[0039] The verification process real-time monitoring is designed, the execution progress of the verification task is tracked, and when the verification is not completed within the preset time, a backup verification scheme is started.

[0040] As a further scheme of the application, when the abnormal signal is confirmed to be a false alarm, the virtual resource occupation identifier is converted into a formal resource allocation record, specifically including:

[0041] The time attribute and the control unit association information in the virtual resource occupation identifier are reserved, and only the state mark is updated from virtual occupation to formal allocation;

[0042] In the logical resource view of the MES resource management layer, the state of the target tooling is adjusted from to-be-verified to allocated, and the control unit information to which it belongs is updated;

[0043] According to the production task priority, the blocked physical scheduling instruction is recovered in batches, and the tooling allocation instruction of the high-priority task is preferentially recovered;

[0044] The consistency of the verification resource allocation record and the actual resource usage state of each distributed control unit is verified, and the system is ensured to smoothly transit to the normal operation state.

[0045] As a further scheme of the application, when the abnormal signal is confirmed to be true, the virtual resource occupation identifier is removed and the pre-action declaration is re-coordinated according to the actual resource state, specifically including:

[0046] The virtual occupation state of the target tooling is immediately removed, and the related system resource lock is released;

[0047] Based on the abnormality confirmation result, the historical declaration credibility of the related distributed control unit is attenuated;

[0048] All pre-action declarations affected by the target tool state change are identified, and a dependency graph is established between the declarations.

[0049] The time urgency, task criticality and declaration credibility of each pre-action declaration are comprehensively considered to generate a new tool allocation scheme.

[0050] As a further scheme of the application, the construction process of the dependency graph is:

[0051] The target tool identifier, operation type and timestamp triple information contained in each pre-action declaration are extracted to establish a standardized expression of the declaration elements.

[0052] Based on the sequence constraint of tool use, resource exclusivity rule and process connection requirement, the time sequence dependency, resource mutual exclusion and logical association relationship between different declarations are identified.

[0053] The pre-action declaration is taken as a node, and the dependency relationship is taken as an edge to construct a dependency graph containing time sequence link, resource competition and logical constraint.

[0054] The beneficial effects of the application are:

[0055] (1) The application introduces a "dynamic intention coordination library" and a "logical consistency measurement" mechanism, which improves the response mode of the system from passive coping to active prediction and collaborative decision-making. Specifically, after receiving an abnormal signal, the system does not immediately trigger the scheduling action of the physical world, but first performs multidimensional consistency analysis of the signal and the pre-action declaration of each control unit at the logical level. When a high degree of contradiction is identified between the signal and the overall intention, a "virtual resource occupation identifier" buffer mechanism is created to "freeze" the resource state at the logical level before confirming the physical fact, effectively blocking the propagation path of the false signal between cross-domain control systems. This method converts potential "physical oscillation" into "logical verification", enabling the system to maintain stable operation in complex environments and avoiding production stagnation and equipment idling caused by false reports.

[0056] (2) This invention cleverly resolves the contradiction between security and efficiency through a parallel processing mechanism of "virtual resource occupancy identifier" and "active perception verification". The creation of the virtual identifier ensures that high-risk erroneous scheduling instructions are reliably blocked during the verification period, thus guaranteeing system security. The system does not completely stop, but synchronously schedules diagnostic equipment to efficiently confirm the true state of the target tooling, and through the "state convergence and system synchronization" mechanism, it can quickly re-coordinate based on the verification results. If it is a false alarm, the virtual identifier can be instantly converted into a formal allocation, production can be resumed, and the interruption time can be shortened; if it is true, a new and feasible tooling allocation scheme can be quickly generated based on the updated resource state and dependency graph. This not only reduces production waste caused by anomaly investigation, but also improves the turnover rate of tooling resources and the continuity of the entire production process through precise resource state management and rapid system recovery capabilities. Attached Figure Description

[0057] The invention will now be further described with reference to the accompanying drawings.

[0058] Figure 1 This is a flowchart of the system of the present invention. Detailed Implementation

[0059] 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.

[0060] Please see Figure 1 As shown, this invention is a tooling usage management system based on MES, comprising:

[0061] The intent coordination module is used to establish and maintain a dynamic intent coordination library within the MES, and to receive and store pre-action declarations from multiple distributed control units. Each pre-action declaration contains the identifier of the target tool, the type of operation to be performed, and a timestamp.

[0062] The intelligent sensing and conflict identification module, when the MES receives an abnormal signal that contradicts the current logical state of the target tooling through the sensing network, uses the pre-action declarations stored in the dynamic intent coordination library as a comparison benchmark to trigger the conflict identification mechanism and calculate and generate a logical consistency metric value.

[0063] The dynamic decision-making and resource resolution module uses the generated logical consistency metric as the basis for judgment. When the logical consistency metric exceeds the preset dynamic decision threshold, the following processing is performed:

[0064] Based on the conflict recognition result, a virtual resource occupation identifier is created for the relevant distributed control unit in the resource management layer of the MES, and based on this identifier, the target tooling re-deployment instruction issued to the physical execution unit is blocked;

[0065] Based on the established virtual resource occupation identifier, an active sensing verification process is started, and a diagnostic device is dispatched to immediately confirm the real physical state of the target tooling;

[0066] The state convergence and system synchronization module takes the confirmation result returned by the active sensing verification process as input and performs the following operations:

[0067] When the confirmation exception signal is a false alarm, the virtual resource occupation identifier is converted into a formal resource allocation record;

[0068] When the confirmation exception signal is true, the virtual resource occupation identifier is removed and the pre-action declaration is re-coordinated according to the actual resource state.

[0069] In the intent coordination module, the process of establishing a dynamic intent coordination library within the manufacturing execution system (MES) first needs to build a tooling use intent graph based on the time sequence relationship. The graph takes the target tooling identifier as the node, and each node records the basic attributes and current state of the tooling. The time sequence of the pre-action declaration constitutes the edge of the graph, and these edges not only record the time sequence of the declaration, but also contain the dependency relationship and constraint conditions between the declarations. In the construction process, the system will analyze the timestamp information of each pre-action declaration to determine the order of the declarations and establish the time dependency relationship. For declarations with sequential constraints, the system will establish directed edges between the corresponding nodes to indicate that these declarations must be executed in a specific order. For declarations that can be executed in parallel, the system will establish undirected edges to indicate that there is no strict time sequence relationship between these declarations. The network structure formed in this way can reveal the time dependency relationship between tooling use intents, providing a basis for subsequent conflict detection and resource coordination.

[0070] After receiving the pre-action declarations from different distributed control units, the system needs to perform a spatio-temporal alignment process. This process first converts the local timestamps used by each control unit into the global time reference of the manufacturing execution system. Then, depending on the type of operation planned in the pre-action declaration, the system determines the length of the time window required for each operation. For complex operations involving multiple tooling, the system decomposes them into basic operations and determines the temporal constraints between them. Based on the spatio-temporal alignment, the system detects possible conflicts between the pre-action declarations of different control units. Conflict detection mainly focuses on declarations with overlapping time windows, checking whether they make mutually exclusive operation requests for the same tooling. When conflicts are detected, the system resolves them by adjusting the operation time windows or reallocating tooling resources, taking into account factors such as the priority of the pre-action declaration, the degree of time urgency, and the importance of the operation. Finally, a unified intent description framework is generated, which can fully and conflict-free express the tooling use intent of all control units.

[0071] The maintenance of the dynamic intent coordination library includes the management of the effective state of the pre-action declaration. The system sets the start time and end time of the effective period for each declaration according to the timestamp attribute of the pre-action declaration. Within the declaration effective period, the system marks it as active and participates in the subsequent conflict detection and resource allocation process. When the declaration exceeds its effective period, the system automatically converts its state to expired and removes it from the current active intent set. For expired pre-action declarations, the system decides whether to delete them immediately or archive them based on their historical importance. For expired declarations with important reference value, the system transfers them to the historical record library and establishes corresponding index information for subsequent query and analysis. At the same time, the system regularly cleans up those no longer needed expired declarations, releases storage space, and ensures the running efficiency of the coordination library.

[0072] To ensure the real-time consistency of the dynamic intent coordination library, the system needs to establish a maintenance mechanism. This mechanism first includes the processing of new pre-action declarations. When a new pre-action declaration is received, the system immediately performs spatio-temporal alignment and conflict detection to ensure that there is no conflict between the new declaration and existing declarations. For new declarations with conflicts, the system will handle them according to the preset conflict resolution strategy, which may include adjusting the time window of the new declaration, modifying the time schedule of existing declarations, or rejecting the new declaration. Second, the system monitors the execution of the tooling operation and adjusts the relevant declaration state and time schedule when it finds that the actual operation deviates from the pre-action declaration. In addition, the system also establishes a tracking mechanism for the dependency relationship between declarations. When the state of a declaration changes, it automatically checks and updates other declarations that have a dependency relationship with it. Through these systematic maintenance measures, the dynamic intent coordination library always maintains an accurate and consistent state, providing reliable support for the coordinated control of tooling resources.

[0073] For the timestamp information of pre-action declarations, the system records the submission time, planned start time, and planned completion time, and calculates the relative relationship between these time points. When building the tool usage intent graph, the system analyzes the time intervals between different declarations to determine whether they have overlapping or connecting relationships in time. When performing conflict detection, the system compares the usage time windows of different declarations for the same tool and calculates their time overlap. When the overlap exceeds the set threshold, it is determined that there is a time conflict. For detected time conflicts, the system eliminates conflicts by adjusting the time window of the declaration. The principle of adjustment is to minimize the impact on the original declaration arrangement while ensuring that the adjusted time window meets the technical requirements of each operation. Through these specific data processing and method steps, the dynamic intent coordination library can effectively manage and maintain the intent information of tool usage.

[0074] In the intelligent perception and conflict identification module, when the manufacturing execution system receives abnormal signals that contradict the current logical state of the target tool through the perception network, the system starts the conflict identification mechanism. This process first needs to establish a multi-dimensional evidence chain to systematically compare the received abnormal signals with all pre-action declarations for the corresponding target tool in the dynamic intent coordination library. In the spatio-temporal alignment process, the system extracts the time overlap feature by comparing the time of the abnormal signal and the time window of the planned operation in the pre-action declaration to calculate the degree of coincidence in the time dimension. The extraction of the operation mutual exclusivity feature requires analyzing whether the operation state indicated by the abnormal signal and the operation type planned to be executed in the pre-action declaration have a logical conflict relationship. The determination of the declaration confidence level feature is based on factors such as the reliability of the source of the pre-action declaration, the historical accuracy rate, and the age of the declaration. These three dimensions of feature data collectively constitute the evidence basis for evaluating the credibility of the abnormal signal.

[0075] When performing the confidence accumulation calculation within the time window, the system first determines the relative position relationship between the time of the abnormal signal occurrence and each pre-action declaration timestamp. For each pre-action declaration, the system analyzes the time difference between the planned execution time and the time of the abnormal signal occurrence, and divides the declaration into different time proximity levels according to the size of the time difference. Declarations with a time difference within 1 minute belong to the highest proximity level, 1 to 5 minutes belong to the medium proximity level, and more than 5 minutes belong to the lower proximity level. The system sets a corresponding base weight for each proximity level, with the highest proximity level being 0.9, the medium proximity level being 0.6, and the lower proximity level being 0.3. Then the system checks whether the operation type of the pre-action declaration is consistent with the state indicated by the abnormal signal. If it is consistent, it adds a consistency bonus of 0.2 to the base weight of the declaration, and if it is not consistent, it deducts a consistency penalty of 0.2. Finally, the weighted scores of all pre-action declarations are accumulated to obtain the initial consistency score.

[0076] When performing the conflict propagation depth analysis, the system traces the dependency relationship network between the pre-action declarations. This process first starts from the pre-action declaration directly related to the abnormal signal, performs a breadth-first search along the dependency edges, and records the number of affected declaration nodes. The number of declaration nodes in the first layer of direct dependency is denoted as N1, the number of declaration nodes in the second layer of indirect dependency is denoted as N2, and the number of declaration nodes in the third layer of further dependency is denoted as N3. The system calculates the conflict propagation coefficient based on these data, and the calculation process is: 1 plus N1 multiplied by 0.3, plus N2 multiplied by 0.2, plus N3 multiplied by 0.1. Then the system analyzes the criticality of each declaration on the dependency path. For declarations on the critical path, their propagation weight is increased by 1.5 times, and for declarations not on the critical path, their propagation weight remains 1. Finally, the system divides the initial consistency score by the conflict propagation coefficient, and then multiplies it by the critical path adjustment factor to obtain the modified consistency score.

[0077] When outputting the final logical consistency metric value, the system maps the modified consistency score to the value range of 0 to 100. The mapping process uses a piecewise linear conversion method. When the modified score is less than or equal to 0, the logical consistency metric value is 0; when the modified score is between 0 and 50, the logical consistency metric value is equal to the modified score multiplied by 2; when the modified score is between 50 and 100, the logical consistency metric value is equal to 100 minus the modified score multiplied by 0.5; when the modified score is greater than or equal to 100, the logical consistency metric value remains 100. This final logical consistency metric value reflects the overall deviation degree of the abnormal signal from the intent of tool usage, with a lower value indicating a higher likelihood of false positives and a higher value indicating a higher likelihood of reflecting the true state. The system compares this metric value with the pre-set dynamic decision threshold as the basis for subsequent resource coordination decisions.

[0078] In the dynamic decision and resource resolution module, when creating a virtual resource occupation identifier in the resource management layer of the manufacturing execution system, the specific attribute composition of the identifier needs to be determined first. The virtual resource occupation identifier contains four basic elements: target tooling identifier for uniquely determining the tooling equipment involved, occupation start time record identifier for the creation time, expected duration setting identifier for the validity period, and credibility score for reflecting the reliability of the identifier. In the creation process, the system will calculate the credibility score according to the logical consistency measure value in the conflict identification result. The calculation method is to subtract the logical consistency measure value from 100, and then divide the result by 100 to convert it to a decimal value between 0 and 1. The determination of the expected duration needs to consider multiple factors, including the severity of abnormal signals, the average time required by the system to handle similar situations in the past, and the configuration of the current available diagnostic resources. The base duration is set to 10 minutes, and then adjusted according to the credibility score. For every 0.1 increase in the credibility score, the expected duration increases by 2 minutes.

[0079] When establishing the decision tree for instruction blocking, first build the decision condition evaluation layer. The system divides the credibility score of the virtual resource occupation identifier into three confidence levels: scores between 0.7 and 1.0 are high confidence, scores between 0.4 and 0.7 are medium confidence, and scores between 0 and 0.4 are low confidence. At the same time, the time interval is divided into three levels according to the occupation duration: 1 hour or less is short-term, 1 to 4 hours is medium-term, and 4 hours or more is long-term. This division is based on statistical analysis of tool usage patterns and can better reflect the characteristics of resource occupation under different circumstances. In the division process, the system will consider factors such as the usage frequency of the tool, its importance in the current production task, and the availability of alternative tools. For high-usage-frequency key tools, the system will use stricter time interval division standards: the short-term interval is shortened to 30 minutes, the medium-term interval is 30 minutes to 2 hours, and the long-term interval is 2 hours or more.

[0080] The system configures differentiated blocking strategies for different combinations of confidence levels and time intervals. For high confidence levels combined with short-term time intervals, a complete blocking strategy is adopted, stopping all physical scheduling instructions related to the tool. For high confidence levels combined with medium-term time intervals, a delayed execution strategy is implemented, postponing the execution of non-critical task tool scheduling instructions for 30 minutes, but allowing emergency production task instructions to pass. For medium confidence levels combined with short-term time intervals, a conditional release strategy is adopted, allowing instruction execution only when there is a backup tool available. For low confidence levels, regardless of the time interval, only monitoring and recording are performed without substantive blocking. The implementation of these strategies requires modification of the tool state identification in the manufacturing execution system. When a tool is marked as a virtual occupied state, the corresponding scheduling instructions will be redirected to the pending queue or directly rejected according to the preset strategy.

[0081] The system dynamically adjusts the decision threshold and blocking strength in the decision tree according to real-time load state and production task urgency. The evaluation of real-time load state is based on the number of production tasks being executed in the manufacturing execution system, device utilization rate indicators and the length of the pending work order queue. When the system load exceeds 85%, the threshold of all blocking strategies will automatically increase by 0.1 to reduce the impact on production progress. When the system load is less than 30%, the threshold is correspondingly reduced by 0.1 to enhance the safety protection of the system. The judgment of production task urgency is based on factors such as task priority, delivery deadline urgency and customer level. For urgent tasks, the system temporarily adjusts the blocking strategy to allow necessary tool scheduling instructions to continue execution under medium confidence levels. This dynamic adjustment mechanism updates the adjustment parameters every 5 minutes by monitoring the running state data of the production system.

[0082] By monitoring the system state changes after instruction blocking, the system automatically corrects the strategy parameters in the decision tree. The monitoring content includes the number of blocked instructions, the importance distribution of blocked instructions, the impact of blocking duration on production progress, etc. The system records the processing results of each instruction blocking, including whether the final confirmed abnormal signal is real, the production delay caused during the blocking process, and whether it triggers other chain problems. Based on these historical data, the system optimizes the strategy parameters once a month, adjusting the confidence level threshold and the length of the time interval. The optimization process uses the gradient descent method, with the goal of minimizing the production loss caused by false blocking as the objective function, gradually adjusting the parameters in the decision tree. At the same time, the system will establish a strategy effect evaluation mechanism to evaluate the decision of each instruction blocking after the event, and feedback the evaluation results to the decision parameter adjustment process, forming a closed-loop optimization.

[0083] In the state convergence and system synchronization module, when the abnormal signal is confirmed to be a false alarm, the system initiates a conversion process from the virtual resource occupation identifier to the formal resource allocation record. This conversion first preserves all the time attributes recorded in the virtual resource occupation identifier, including the start time of occupation, the estimated duration, and the actual occupation duration. At the same time, the control unit association information is completely preserved, ensuring that the affiliation between the tooling and the corresponding control unit does not change. The system updates the identifier's state marker from "virtual occupation" indicating temporary occupation to "formal allocation" indicating formal allocation; this state change is recorded in the manufacturing execution system's operation log with the timestamp and the reason for the change. During the state marker update process, the system checks the integrity of the virtual resource occupation identifier to ensure that all necessary information is properly preserved, providing a complete data foundation for subsequent resource management and scheduling.

[0084] In the logical resource view of the manufacturing execution system resource management layer, the system performs a target tooling state synchronization operation. This operation adjusts the target tooling's state in the logical resource view from "to be verified" indicating the need for verification to "allocated" indicating that it has been allocated. During the state adjustment process, the system updates the tooling's control unit information, establishing a formal binding relationship between the tooling and the control unit that made the pre-action declaration. At the same time, the system recalculates the tooling's estimated available time, adding the operation execution time based on the operation type and estimated time in the pre-action declaration to the original time. For composite operations that require the use of multiple toolings in succession, the system synchronously updates the states of related toolings to ensure that the logical resource view accurately reflects the overall usage state of the tooling group.

[0085] When restoring the blocked physical scheduling instructions in batches according to production task priority, the system first prioritizes all blocked instructions. The sorting criteria include the production urgency of the task, the customer level, the delivery deadline urgency, and the critical path position of the task in the entire production plan. The system divides the tasks into three priority levels: priority 1 for critical tasks that directly affect the main production plan, priority 2 for secondary tasks that affect local production progress, and priority 3 for auxiliary tasks that can be flexibly adjusted. The restoration process starts with priority 1 tasks, with the number of restored instructions controlled within 20% of the total blocked instructions, and the interval between adjacent batches kept at least 30 seconds to observe the system's restoration state. After each batch of instructions is restored and executed, the system monitors the load of the tooling scheduling system. If the load exceeds the safe threshold of 80%, it suspends the subsequent instruction restoration and continues after the load decreases.

[0086] When verifying the consistency between the resource allocation records and the actual resource usage status of each distributed control unit, the system adopts a two-way verification mechanism. First, it sends a status query request from the manufacturing execution system to each distributed control unit, collecting the tool usage status data recorded by each unit. Then it compares the tool status in the resource allocation records with the actual status feedback by the control units, checking whether they are consistent. For the status records with differences, the system will start detailed cause analysis to determine whether the cause of the difference is data synchronization delay, network transmission error or system logic error. According to the analysis results of the difference, the system performs corresponding data correction operation to ensure that the resource allocation records of the manufacturing execution system are completely consistent with the actual status of the distributed control units. This verification process continues until all resource status differences are eliminated, and the system confirms that it has reached a stable running state.

[0087] When confirming that the abnormal signal is true, the system immediately releases the virtual occupation status of the target tool. This operation includes clearing the virtual occupation identifier of the tool in the manufacturing execution system, removing the temporary status marker in the resource management layer, and releasing various system resource locks set due to virtual occupation. System resource locks include tool usage permission lock, scheduling instruction processing lock, and access lock of related control units, etc. During the process of releasing the virtual occupation status, the system will record the timestamp, the execution personnel or system identifier, and the reason for releasing, which will be an important basis for system operation audit. At the same time, the system will send a status update notification to all related distributed control units to ensure that each unit can timely learn about the change of tool status.

[0088] Based on the results of this exception confirmation, the system performs decay processing on the historical declaration credibility of the related distributed control units. The credibility decay calculation adopts an exponential decay method, multiplying the historical declaration credibility of the control unit by a decay coefficient of 0.9. For the control units that have experienced similar abnormal situations within the last 24 hours, an additional penalty decay of 0.05 is applied. If the decayed credibility value is less than 0.3, the system will mark the control unit as an object that needs to be closely monitored, and will conduct more strict review on its pre-action declaration in subsequent processing. The credibility decay data will be updated to the capability evaluation file of the control unit, serving as a reference basis for future allocation of tool resources. The system will also record the specific reasons and related data of the decay processing to ensure the transparency and traceability of the processing process.

[0089] When identifying all pre-action statements affected by the target tool state change, the system employs a dependency-based propagation analysis. This analysis first determines the pre-action statements directly related to the target tool, including statements that plan to use the tool, statements that have a time precedence relationship with the tool usage, and statements that have a resource contention relationship with the tool usage. Then the system recursively searches along the dependency chains between statements to find all indirectly affected statements. During the search process, the system records the degree of association between each affected statement and the target tool state change, which is calculated based on the length of the dependency path and the strength of the dependency relationship. The degree of association decays by 50% for each layer of path length, and a statement with a strong dependency relationship has twice the degree of association of a statement with a weak dependency relationship. Finally, the system establishes a complete list of affected statements, providing a data foundation for subsequent re-coordination.

[0090] When constructing the dependency graph between statements, the system first extracts the triple information of target tool identification, operation type, and timestamp contained in each pre-action statement. During the extraction process, each statement element is standardized. The target tool identification is uniformly converted to the standard encoding format in the manufacturing execution system, the operation type is mapped to the corresponding item in the standard operation type dictionary, and the timestamp is uniformly converted to a relative timestamp in seconds. The standardized statement elements are organized in a fixed format to ensure data consistency in subsequent processing. The system generates a unique statement identifier for each pre-action statement, which consists of three parts: control unit number, generation timestamp, and sequence number, used to uniquely identify the statement in the dependency graph.

[0091] Based on the sequence constraints of tool usage, resource exclusivity rules, and process connection requirements, the system identifies the dependencies between different statements. For multiple statements involving the same tool, the system checks whether their time windows overlap. Statements with overlapping time windows exceeding 5 minutes are marked as having a resource contention relationship. For statements with process sequence, the system determines their time sequence dependency relationship based on the process flowchart. The statement of the previous process must be executed before the statement of the subsequent process. For statements using tools with mutual exclusivity, even if their time windows do not overlap, they are also marked as having a logical constraint relationship. The strength of the dependency relationship varies depending on the type of constraint. The strength value of the resource contention relationship is 0.8, the strength value of the time sequence dependency relationship is 1.0, and the strength value of the logical constraint relationship is 0.6.

[0092] The system constructs a dependency graph containing time sequence link, resource competition and logical constraint, taking pre-action declaration as node and dependency relationship as edge. In the process of graph construction, the size of node is determined according to the importance degree of declaration, and the importance degree is calculated according to the production task priority, tooling value and operation complexity involved in the declaration. The thickness of edge is drawn according to the strength of dependency relationship, and the higher the strength value is, the thicker the edge is. The graph adopts hierarchical layout, the nodes of the same time sequence level are arranged on the same horizontal line, the nodes with resource competition relationship are connected by dashed line, and the nodes with time sequence dependency relationship are connected by solid arrow. The system sets attribute records for each node and edge in the graph, including declaration details of node, dependency type and strength value of edge, which supports subsequent graph analysis and query operation.

[0093] The working principle of the present application is as follows: a dynamic intention coordination library is constructed to receive and store pre-action declarations from the distributed control unit, and a tooling use intention graph is established; when an abnormal signal is perceived, multi-dimensional conflict identification is carried out based on the pre-action declaration, and a logical consistency measure value is calculated; when the measure value exceeds a threshold value, a virtual resource occupation identifier is created and physical scheduling instructions are blocked, and a proactive perception verification process is started; finally, state synchronization is performed according to the verification result, if it is a false alarm, the virtual identifier is converted into a formal resource allocation record, if it is true, the virtual identifier is removed and the pre-action declaration is re-coordinated, so as to avoid system oscillation while ensuring production continuity.

[0094] The above describes one embodiment of the present application in detail, but the content described is only the preferred embodiment of the present application, and cannot be considered as limiting the scope of the present application. Any equivalent changes and improvements made within the scope of the present application should still belong to the scope of the present application.

Claims

1. A tooling usage management system based on MES, characterized in that, include: The intent coordination module is used to establish and maintain a dynamic intent coordination library within the MES, and to receive and store pre-action declarations from multiple distributed control units. Each pre-action declaration contains the identifier of the target tool, the type of operation to be performed, and a timestamp. The intelligent sensing and conflict identification module, when the MES receives an abnormal signal that contradicts the current logical state of the target tooling through the sensing network, uses the pre-action declarations stored in the dynamic intent coordination library as a comparison benchmark to trigger the conflict identification mechanism and calculate and generate a logical consistency metric value. The calculation of the generated logical consistency metric specifically includes: Establish a multi-dimensional evidence chain, and align the received abnormal signals with all pre-action declarations of the corresponding target tooling in the dynamic intent collaboration library in time and space, and extract features in three dimensions: time overlap, operation mutual exclusion and declaration confidence level. The confidence level is accumulated within the time window. Based on the temporal relationship between the timestamp of the pre-action declaration and the time of occurrence of the abnormal signal, the compliance degree of each pre-action declaration is weighted and accumulated to generate an initial consistency score. Conduct in-depth conflict propagation analysis, assess the scope of chain reactions that anomalous signals may trigger by tracing the dependency network between pre-action statements, and dynamically revise the initial consistency score based on the propagation depth; Output the final logical consistency metric; The dynamic decision-making and resource resolution module uses the generated logical consistency metric as the basis for judgment. When the logical consistency metric exceeds the preset dynamic decision threshold, the following processing is performed: Based on the conflict identification results, a virtual resource occupancy identifier is created for the relevant distributed control unit in the resource management layer of the MES, and the target tooling reallocation instruction issued to the physical execution unit is blocked based on this identifier. Based on the established virtual resource occupancy identifier, the active perception and verification process is initiated, and diagnostic equipment is dispatched to instantly confirm the actual physical state of the target tooling. The state convergence and system synchronization module takes the confirmation result returned by the active sensing verification process as input and performs the following operations: When the abnormal signal is confirmed to be a false alarm, the virtual resource occupancy identifier is converted into a formal resource allocation record; Once the abnormal signal is confirmed to be real-time, the virtual resource occupancy flag is removed, and the pre-action declaration is re-coordinated based on the actual resource status.

2. The tooling usage management system based on MES according to claim 1, characterized in that, The establishment and maintenance of the dynamic intent collaboration library specifically includes: Construct a tooling usage intent graph based on temporal relationships. The tooling usage intent graph uses the target tooling identifier as nodes and the time sequence of pre-action declarations as edges to form a network structure with temporal dependencies. Spatiotemporal alignment and conflict resolution are performed on pre-action declarations from different distributed control units to generate a unified intent description framework; The validity status of pre-action declarations is dynamically maintained based on the timestamp attribute of the declarations. Expired pre-action declarations are automatically archived and invalidated, and the dynamic intent coordination library is maintained in real time for consistency.

3. The tooling usage management system based on MES according to claim 1, characterized in that, The step of creating virtual resource occupancy identifiers for relevant distributed control units in the resource management layer of MES specifically includes: Create a virtual resource occupancy identifier with a time-sensitive attribute. The virtual resource occupancy identifier includes the target tooling identifier, occupancy start time, expected duration, and credibility score. Establish a decision tree for command blocking, and dynamically select different levels of physical scheduling commands, such as complete blocking, delayed execution, or conditional release, based on the credibility score and timeliness attribute of the virtual resource occupancy identifier. Implement resource status mirroring synchronization, and while maintaining virtual resource occupation, establish a logical resource view in the MES resource management layer that runs parallel to the physical tooling status; When the expected duration is reached or a verification result is received, the process of clearing the virtual resource occupancy identifier is automatically triggered.

4. The tooling usage management system based on MES according to claim 3, characterized in that, The establishment of the decision tree for instruction blocking specifically includes: A decision-making condition assessment layer is constructed, which divides virtual resource occupancy identifiers into three confidence levels: high, medium, and low, based on their credibility scores. At the same time, the occupancy duration is divided into three time intervals: short-term, medium-term, and long-term, based on the timeliness attribute. Differentiated blocking strategies are configured for combinations of different confidence levels and time intervals, including completely blocking all physical scheduling instructions, delaying the execution of non-critical instructions, and allowing instructions with urgent priority. Based on the real-time system load status and the urgency of production tasks, dynamically adjust the judgment thresholds and blocking strengths of each condition in the decision tree; By monitoring the changes in system state after the command is blocked, the strategy parameters in the decision tree are automatically corrected to achieve adaptive optimization of the blocking strategy.

5. The tooling usage management system based on MES according to claim 1, characterized in that, The initiation of the active perception verification process specifically includes: Based on the logical consistency metric and the criticality level of the target tooling, the urgency of the verification tasks is calculated and the execution order is dynamically adjusted. When multiple verification tasks are performed concurrently, the appropriate diagnostic device is intelligently allocated to perform the verification based on the real-time workload and movement path of the diagnostic device. Coordinate visual recognition devices, RFID readers, and IoT sensors to participate in the verification process, and improve the reliability of the confirmation results through cross-verification of multi-source data; The design verification process is monitored in real time, and the execution progress of the verification task is tracked. If the verification is not completed within the preset time, the backup verification plan is activated.

6. The tooling usage management system based on MES according to claim 1, characterized in that, When the abnormal signal is confirmed to be a false alarm, the virtual resource occupancy identifier is converted into a formal resource allocation record, specifically including: The time attribute and control unit association information in the virtual resource occupancy identifier are retained, and only its status flag is updated from virtual occupancy to formal allocation; In the logical resource view of the MES resource management layer, the status of the target tooling is changed from pending verification to assigned, and its control unit information is updated. The blocked physical scheduling instructions will be restored in batches according to the priority of production tasks, with priority given to restoring tooling allocation instructions for high-priority tasks. Verify the consistency between resource allocation records and the actual resource usage status of each distributed control unit to ensure a smooth transition of the system to normal operation.

7. The tooling usage management system based on MES according to claim 1, characterized in that, When the abnormal signal is confirmed to be real-time, the virtual resource occupancy flag is removed and the pre-action declaration is re-coordinated based on the actual resource status. Specifically, this includes: Immediately release the virtual occupancy status of the target tool and release the relevant system resource locks; Based on the results of this anomaly confirmation, the credibility of historical claims of the relevant distributed control units will be reduced. Identify all pre-action declarations affected by changes in the target tooling state and establish a dependency graph between the declarations; Taking into account the time urgency, task criticality, and credibility of each pre-action declaration, a new tooling allocation scheme is generated.

8. The tooling usage management system based on MES according to claim 7, characterized in that, The process of constructing the dependency graph is as follows: Extract the target tooling identifier, operation type, and timestamp triplet information contained in each pre-action declaration, and establish a standardized expression for the declaration elements; Based on the constraints of the order of tooling use, the rules of resource exclusivity, and the requirements of process connection, identify the temporal dependencies, resource mutual exclusions, and logical relationships between different declarations; Using pre-action declarations as nodes and dependencies as edges, a dependency graph containing temporal links, resource contention, and logical constraints is constructed.

Images