A configurable process engine execution method based on industrial production processes

Abstract

The application discloses an industrial production process-based configurable process engine execution method, and relates to the technical field of industrial software process management. The method comprises the following steps: receiving process model definition data, analyzing and obtaining node configuration information and flow logic, wherein the nodes comprise process nodes, quality inspection nodes and equipment linkage nodes; in response to a process start instruction, creating a process instance and driving the first node to execute; in the node execution process, analyzing a dynamic binding relationship, calling an external business system interface to obtain or write business data according to the binding relationship, and calculating a next execution node according to the flow logic; real-time monitoring of abnormal events, when an abnormal event is monitored, interrupting the current process according to an abnormal processing strategy and jumping to an abnormal processing branch node for execution; collecting execution track data and business data of the process instance, and storing the data to generate traceable process execution history records. The application realizes configurable execution and abnormal closed-loop processing of industrial processes.

Description

Technical Field

[0001] This application relates to the field of industrial software technology, and in particular to a configurable process engine execution method based on industrial production processes. Background Technology

[0002] With the rapid development of Industry 4.0 and intelligent manufacturing, manufacturing enterprises are increasingly demanding production process management. Production process management is a core aspect of manufacturing enterprise operations management, involving multiple aspects such as production planning, process route design, process scheduling and execution, quality inspection and control, and equipment status monitoring.

[0003] Traditional production process management methods mainly take the following forms: First, paper-based document circulation, which uses paper work orders, process cards, inspection sheets, and other documents to transmit production information. This is inefficient, prone to errors, and difficult to trace, and can no longer meet the management needs of modern manufacturing. Second, traditional software systems, which use customized MES, ERP, and other systems to manage production processes. Although this achieves informatization, it suffers from long development cycles, high costs, and poor flexibility. When the enterprise's business processes are adjusted, they need to be recoded and modified. Third, general workflow engines, which use open-source workflow engines such as Activiti and Flowable. These are mainly designed for general approval process scenarios and lack sufficient support for the specific industrial production processes, procedures, and equipment linkage scenarios.

[0004] Existing production process management technologies have the following main shortcomings:

[0005] First, there is a lack of in-depth support for industrial scenarios. Existing workflow engines mainly support simple approval workflows and cannot effectively handle complex processes in industrial production, such as multi-process parallelism, material flow between processes, process parameter control, quality inspection nodes, and equipment linkage triggering—all unique to industrial scenarios. Specifically, general-purpose workflow engines do not define industry-specific node types such as process nodes, quality inspection nodes, and equipment linkage nodes, making it difficult to embed processing parameters, inspection standards, and equipment instructions from the production site into the process execution logic.

[0006] Secondly, there is a disconnect between processes and business data. Existing process engines are often independent of business data management, making it difficult to associate process nodes with business forms, equipment data, quality data, etc., thus hindering data-driven process execution. When a process reaches a certain work step, it cannot automatically load the corresponding process form, automatically obtain the real-time status of the bound equipment, or automatically send processing parameters to the equipment, resulting in data silos between process execution and production operations.

[0007] Third, the anomaly handling mechanism is inadequate. Unexpected situations such as equipment failure, material shortages, and quality abnormalities frequently occur during production. Current technologies lack flexible anomaly identification and handling mechanisms, failing to support scenarios involving abnormal process branches, process rollbacks, and emergency order insertions. More critically, existing technologies struggle to pre-set anomaly capture nodes and anomaly handling branches within the process model. When an anomaly occurs, it cannot automatically identify the anomaly type and redirect to the pre-defined processing path, relying solely on manual intervention or process interruption.

[0008] Fourth, weak process monitoring and traceability capabilities. The lack of real-time monitoring and visualization of the production process makes it impossible to promptly identify bottlenecks and anomalies, and hinders full-process quality traceability. When product quality issues arise, it is impossible to quickly trace back all process nodes the product has undergone, the personnel involved, equipment parameters, and inspection data at each node, making it difficult to define quality responsibility and improve processes. Summary of the Invention

[0009] This application provides a configurable process engine execution method based on industrial production processes, aiming to solve the problems of poor flexibility in industrial process management, data silos, and delayed response to production anomalies caused by the lack of node definitions for specific industrial scenarios, the lack of process and business data association capabilities, and the imperfect anomaly handling mechanism in the prior art.

[0010] Firstly, a configurable process engine execution method based on industrial production processes is provided, including:

[0011] S1, receive process model definition data, parse the process model definition data to obtain node configuration information and flow logic, the node configuration information includes standard process nodes and industrial extended nodes, the industrial extended nodes include process nodes, quality inspection nodes and equipment linkage nodes.

[0012] S2, in response to the process start command, create a process instance and initialize process variables, and drive the execution of the first node of the process instance based on the data defined by the process model;

[0013] S3, during node execution, parse the pre-configured dynamic binding relationship, call the external business system interface to obtain or write business data according to the dynamic binding relationship, and calculate the next execution node according to the flow logic in the data defined by the process model;

[0014] S4 monitors abnormal events in real time during the process flow. When an abnormal event is detected, the current process is interrupted and the execution jumps to the abnormal handling branch node according to the preset abnormal handling strategy until a recovery instruction is received or the preset automatic recovery conditions are met.

[0015] S5. During process execution, collect the execution trajectory data and business data of the process instance, and associate and store the execution trajectory data and business data to generate a traceable process execution history.

[0016] Optionally, in the above scheme, before receiving the process model definition data, step S1 further includes a process modeling step:

[0017] The process involves receiving drag-and-drop operations from the user through a visual designer and generating a flowchart that includes the industrial extension nodes.

[0018] The flowchart is validated for legality, and the validated flowchart is serialized into a JSON format flowchart model definition file;

[0019] The attribute configuration of the industrial extension node includes: processing parameters of the process node, qualification judgment rules of the quality inspection node, and control instructions of the equipment linkage node.

[0020] Optionally, in the above scheme, step S3, parsing the pre-configured dynamic binding relationship, includes:

[0021] Parse the form templates bound to the nodes, establish the mapping relationship between form fields and process variables, and realize the automatic loading of business form data;

[0022] The business rules bound to the nodes are parsed. These business rules include node entry rules, node execution rules, and node departure rules. The flow conditions of the nodes are controlled based on these business rules.

[0023] The device information bound to the node is parsed, device control commands are generated, and the data adaptation layer is used to send them to the production equipment.

[0024] Optionally, in the above scheme, the exception handling strategy in step S4 includes:

[0025] When a device fault code or quality failure flag is detected, an exception token is generated, and task dispatching to downstream nodes is stopped.

[0026] The abnormality type is matched with a preset abnormality handling branch, which includes process rollback to a specified node, triggering an alarm to notify relevant personnel, or transferring to a rework / scrap handling node.

[0027] Receive feedback on the processing results for the exception handling branch, compensate and update the process variables based on the feedback on the processing results, and resume process execution.

[0028] Optionally, in the above scheme, a process version management step may be included after step S5:

[0029] Receive modification instructions for the current process model and generate a new version of the process model;

[0030] Mark the running process instance as a pending migration state, and perform path mapping on the process instance in the pending migration state based on the node definition of the new version of the process model;

[0031] While ensuring the consistency of business data, the process instance in the state to be migrated will be smoothly migrated to the new version of the process model for continued execution.

[0032] Secondly, a configurable process engine system based on industrial production processes is provided, including:

[0033] The instantiation module is used to receive process model definition data, parse the process model definition data to obtain node configuration information and flow logic. The node configuration information includes standard process nodes and industrial extended nodes. The industrial extended nodes include process nodes, quality inspection nodes and equipment linkage nodes. In response to the process start command, the module creates a process instance and initializes process variables, and drives the execution of the first node of the process instance based on the process model definition data.

[0034] The execution scheduling module is used to parse the pre-configured dynamic binding relationship during the node execution process, call the external business system interface to obtain or write business data according to the dynamic binding relationship, and calculate the next execution node according to the flow logic in the data defined by the process model.

[0035] The exception monitoring module is used to monitor exception events in real time during the process flow. When an exception event is detected, the current process is interrupted and the execution jumps to the exception handling branch node according to the preset exception handling strategy until a recovery instruction is received or the preset automatic recovery conditions are met.

[0036] The historical archiving module is used to collect execution trajectory data and business data of process instances during process execution, and to associate and store the execution trajectory data and business data to generate a traceable process execution history.

[0037] Optionally, the above solution further includes: a modeling module, used to receive drag-and-drop operations from the user through a visual designer, generate a flowchart containing the industrial extension nodes; perform legality verification on the flowchart, serialize the verified flowchart into a JSON format process model definition file; and generate process model definition data; wherein the attribute configuration of the industrial extension nodes includes: processing parameters of process nodes, qualification judgment rules of quality inspection nodes, and control instructions of equipment linkage nodes.

[0038] Optionally, in the above scheme, the execution scheduling module includes:

[0039] The rule engine unit is used to parse and execute the business rule expressions configured on the node;

[0040] The data adaptation unit is used to dynamically load external data source drivers via the SPI mechanism to achieve communication with heterogeneous databases and production equipment;

[0041] The asynchronous task unit is used to encapsulate time-consuming external system calls into asynchronous tasks, and decouples them through message queues to avoid blocking the core process engine.

[0042] Thirdly, an electronic device is provided, including a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the steps of the above-described method.

[0043] Fourthly, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0044] Compared with the prior art, this application has at least the following beneficial effects:

[0045] Based on further analysis and research into the problems of existing technologies, this application recognizes that existing technologies suffer from poor flexibility in industrial process management, data silos, and delayed responses to production anomalies due to a lack of node definitions for specific industrial scenarios, a lack of ability to correlate process and business data, and an imperfect anomaly handling mechanism. By receiving and parsing process model definition data containing industrial extended nodes, the process engine natively supports industrial-specific scenarios such as process nodes, quality inspection nodes, and equipment linkage nodes. By parsing dynamic binding relationships and calling external business system interfaces, real-time interaction between process nodes and business data is achieved, breaking down data silos between process execution and production operations. Furthermore, by real-time monitoring of abnormal events and activating preset anomaly handling branch nodes, automatic identification and process-level response to anomalies are achieved. This results in improved scenario adaptability of industrial process management, deep integration of process and business data, shortened production anomaly response time, and enhanced visualization and traceability of process execution. Attached Figure Description

[0046] Figure 1 This is a flowchart illustrating a configurable process engine execution method based on an industrial production process, provided as an embodiment of this application.

[0047] Figure 2 An overall architecture diagram of a process engine system provided in one embodiment of this application;

[0048] Figure 3 A schematic diagram of the interface of a visual process designer provided in one embodiment of this application;

[0049] Figure 4 This is a schematic diagram of the structure of an industrial expansion node type provided in one embodiment of this application;

[0050] Figure 5 A schematic diagram illustrating a dynamic binding mechanism for process nodes provided in one embodiment of this application;

[0051] Figure 6 A schematic diagram illustrating the working principle of a process execution engine provided in one embodiment of this application;

[0052] Figure 7 This is a schematic diagram of the interface of a process monitoring and analysis system provided in one embodiment of this application. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0054] In the description of this application, unless otherwise stated, the terms "including", "comprising", "having", etc., also mean "not limited to" (certain units, components, materials, steps, etc.).

[0055] In one embodiment, reference Figure 1 This provides a configurable process engine execution method based on industrial production processes, including:

[0056] S1, receive process model definition data, parse the process model definition data to obtain node configuration information and flow logic, the node configuration information includes standard process nodes and industrial extended nodes, the industrial extended nodes include process nodes, quality inspection nodes and equipment linkage nodes.

[0057] S2, in response to the process start command, create a process instance and initialize process variables, and drive the execution of the first node of the process instance based on the data defined by the process model;

[0058] S3, during node execution, parse the pre-configured dynamic binding relationship, call the external business system interface to obtain or write business data according to the dynamic binding relationship, and calculate the next execution node according to the flow logic in the data defined by the process model;

[0059] S4 monitors abnormal events in real time during the process flow. When an abnormal event is detected, the current process is interrupted and the execution jumps to the abnormal handling branch node according to the preset abnormal handling strategy until a recovery instruction is received or the preset automatic recovery conditions are met.

[0060] S5. During process execution, collect the execution trajectory data and business data of the process instance, and associate and store the execution trajectory data and business data to generate a traceable process execution history.

[0061] In this embodiment, a configurable process engine execution method based on an industrial production process is provided, including:

[0062] S1, the process engine system first receives process model definition data through the process modeling module. This process model definition data is generated by the visual process designer, which uses HTML5 Canvas / SVG technology to implement the process design canvas, supporting interactive operations such as dragging and dropping to add node components, connecting them with lines, and configuring attributes. The left side of the designer is a component library panel, displaying various available node components; the middle is the canvas area, supporting operations such as dragging, connecting, selecting, and deleting nodes; the right side is an attribute configuration panel, dynamically displaying configurable attributes based on the selected node type.

[0063] Based on the standard BPMN 2.0 nodes, this invention defines extended node types specific to industrial production. These include: process nodes representing production processes, configured with attributes such as process code, name, working hours, and processing parameters; quality inspection nodes representing quality inspection processes, configured with attributes such as inspection items, standards, and pass / fail rules; and equipment linkage nodes enabling linkage with production equipment, configured with attributes such as equipment code, acquisition parameters, and control commands. Furthermore, industrial extended nodes may also include material flow nodes (recording material codes, quantities, and locations), anomaly capture nodes (listening for and capturing specified types of anomaly events), and anomaly handling nodes (executing anomaly handling logic).

[0064] The process model is stored in JSON format, including metadata such as node list, connection relationships, node attributes, and process variables, and supports serialization and deserialization of the process model. The ProcessDefinitionParser is responsible for parsing the JSON-formatted process model and constructing the process definition object in memory.

[0065] S2, upon receiving the process start command, the runtime service component (RuntimeService) creates a process instance, assigns a unique instance identifier, initializes process variables, and generates a token to drive process execution. The process instance maintains runtime information such as the current list of active nodes, historical execution records, and process status. The execution engine component (ExecutionEngine) drives the execution of the process instance, places the token on the starting node, and triggers the executor call of the first node.

[0066] S3, each node, upon execution, first parses the pre-configured dynamic binding relationships. This dynamic binding mechanism includes: form binding, where process nodes can be bound to business form templates, configuring the mapping relationship between form fields and process variables, enabling automatic loading and saving of form data during process execution, and supporting advanced features such as master-child table association, field linkage, and data validation; personnel binding, supporting various personnel allocation strategies, including designated personnel, role allocation, department allocation, and dynamic calculation, with configurable task assignment, countersigning, OR signing, and transfer processing modes; equipment binding, where process nodes can be bound to production equipment, configuring equipment codes, collected parameters, control commands, and other information, enabling linkage between process execution and equipment status, and supporting functions such as equipment status detection, processing parameter issuance, and completion signal reception; and rule binding, supporting the configuration of business rules on nodes, including entry rules (determining whether entry into the node is possible), execution rules (node ​​processing logic), and exit rules (determining whether exit from the node is possible), with rule expressions supporting syntax elements such as variables, operators, and functions.

[0067] The data adaptation layer enables data interaction between the process engine and external systems, including read and write adaptation of form data, equipment data, and quality data. For time-consuming node executions (such as waiting for equipment processing or external system calls), an asynchronous execution mechanism is used to avoid blocking the process engine. After the node is completed, the engine is notified to continue execution via callback or message queue.

[0068] After a node completes its execution, the next execution node is calculated based on the flow logic. For parallel gateways, a multi-token mechanism is supported to enable parallel execution and convergence of branch processes; for conditional gateways, the flow direction is determined based on the node execution result and the result of the rule expression calculation.

[0069] S4 employs an event-driven mechanism for exception handling. Exception events generated during process execution are monitored by an exception catcher. Exception catcher nodes are specifically defined to listen for specified types of exception events, and upon triggering, the process is routed to the corresponding exception handling branch. Configurable exception identification rules include: equipment malfunction exceptions (identified based on equipment status codes or fault signals), quality exceptions (identified based on non-conforming marks in quality inspection results), timeout exceptions (identified based on task processing time exceeding a preset threshold), and material exceptions (identified based on material shortages or incorrect materials). Exception rules support combined conditions and priority settings.

[0070] The exception handling strategy supports multiple methods: suspending the process for manual handling, automatically transferring to the exception handling node, triggering alarms to notify relevant personnel, executing preset compensation operations, and rolling back the process to the previous node or a specified node. For recoverable exceptions, an exception handling process branch is triggered; for severe exceptions, process rollback or termination is executed. All exception information is recorded in the exception log, including exception type, occurrence node, exception time, and processing result. The process supports instance rollback operations, allowing reversal to the previous step or any historical node. Rollback automatically undoes executed operations and data changes, and supports lifecycle operations such as pausing, resuming, and terminating the process.

[0071] S5's HistoryService component records historical data of process execution, including the process instance's start time, end time, execution path, start and end times of each node, personnel involved, processing results, and changes in process variables. The monitoring and analysis module provides real-time status queries for process instances, including currently active nodes, pending tasks, and process variable values, and supports visual replay of the process execution path. The performance monitoring component statistically analyzes performance metrics such as execution time, waiting time, and processing count for each process and node, identifying process bottlenecks and inefficient steps. The alarm notification component supports configuring process alarm rules, triggering alarm notifications for situations such as unprocessed tasks, abnormal process termination, and node execution failures. Alarm methods support multiple channels including in-site messages, SMS, email, and WeChat Work. The data analysis component performs multi-dimensional analysis of process execution data, generating reports such as process efficiency reports, node processing statistics, and personnel workload analysis, providing data support for process optimization.

[0072] The method in this embodiment is implemented based on a configurable process engine system, which adopts a layered architecture design and includes the following core modules:

[0073] (1) Process Modeling Layer: Provides a visual process designer, supports process modeling based on the BPMN2.0 specification, and extends the industry-specific node types to realize the graphical configuration of the process.

[0074] (2) Process Engine Core Layer: Responsible for the creation, execution, scheduling and management of process instances, including core components such as process parser, execution scheduler, rule engine and state manager.

[0075] (3) Node execution layer: encapsulates the execution logic of various process nodes, including the actuator implementation of standard BPMN nodes and industrial extended nodes.

[0076] (4) Data adaptation layer: realizes data interaction between the process engine and external systems, including reading and writing adaptation of form data, equipment data and quality data.

[0077] (5) Monitoring and display layer: Provides visualization monitoring, data analysis and report display functions for process execution.

[0078] In one embodiment, before receiving the process model definition data, step S1 further includes a process modeling step:

[0079] The process involves receiving drag-and-drop operations from the user through a visual designer and generating a flowchart that includes the industrial extension nodes.

[0080] The flowchart is validated for legality, and the validated flowchart is serialized into a JSON format flowchart model definition file;

[0081] The attribute configuration of the industrial extension node includes: processing parameters of the process node, qualification judgment rules of the quality inspection node, and control instructions of the equipment linkage node.

[0082] In one embodiment, before receiving the process model definition data, step S1 further includes a process modeling step:

[0083] The process modeling module provides a visual process designer. The front end of this designer is developed based on the Vue.js or React framework, and AntV X6 or bpmn-js is used as the process diagram drawing engine. The designer uses HTML5 Canvas / SVG technology to implement the process design canvas, and supports interactive operations such as dragging and dropping to add nodes, connecting them with lines, selecting, deleting, and configuring attributes.

[0084] The designer interface is divided into several functional areas: the left side is the component library panel, which displays various available node components, including standard BPMN2.0 nodes (such as start events, end events, user tasks, gateways, etc.) and industrial extended nodes defined in this invention (such as process nodes, quality inspection nodes, equipment linkage nodes, material flow nodes, anomaly capture nodes, anomaly handling nodes, etc.); the middle is the canvas area, where users can add nodes from the component library panel to the canvas by dragging and dropping, and establish flow relationships between nodes by connecting them. It supports operations such as dragging and moving nodes, adding, deleting, and modifying connections, and selecting and deleting nodes; the right side is the attribute configuration panel, which dynamically displays configurable attribute fields according to the node type selected by the user.

[0085] After completing the flowchart in the canvas area, users can save, preview, and publish the flowchart.

[0086] When a user triggers a publish operation, the process designer performs a validity check on the currently drawn flowchart. Validation includes, but is not limited to: the process must have one and only one start node, the process must have at least one end node, all nodes must have reachable flow paths, the integrity of node attribute configurations (e.g., process nodes must be configured with process codes and standard working hours, quality inspection nodes must be configured with inspection items and pass / fail rules, etc.), and the uniqueness of gateway node branch conditions.

[0087] After successful validation, the flowchart is serialized into a JSON-formatted process model definition file. This file includes a list of nodes (each node contains metadata such as node type, node identifier, node name, location coordinates, and attribute configuration), connection relationships (each connection contains source node identifier, target node identifier, and transition condition expression), process variable definitions, and process version information. The JSON format storage method supports the serialization and deserialization of the process model, facilitating parsing, execution, and version management by the process engine.

[0088] Specifically, the attribute configurations for each industrial extension node type are as follows:

[0089] Process nodes are used to represent production processes. Their attribute configurations include: process code (a unique code that identifies the process), process name (the Chinese name of the process), standard working hours (the planned execution time of the process), machining parameters (such as cutting speed, feed rate, machining depth, and other process parameters), bound equipment (the equipment code or equipment type for which the process is executed), dependencies of preceding processes, and relationships of subsequent processes.

[0090] Quality inspection nodes are used to represent quality inspection links. Their attribute configurations include: inspection items (quality characteristics that need to be inspected), inspection standards (the acceptable range or judgment criteria for each inspection item), acceptance judgment rules (such as "all items are qualified, then it is considered qualified" or "key items are qualified and the number of non-key items that are unqualified does not exceed X"), non-conformance handling process (such as the process jump configuration for handling methods such as rework, scrapping, and downgrading), and inspection methods (full inspection, sampling inspection, first piece inspection, etc.).

[0091] The equipment linkage node is used to achieve linkage with production equipment. Its attribute configuration includes: equipment code (unique identifier of the target equipment), control commands (processing commands or operation commands that need to be issued to the equipment, such as start, stop, parameter setting, etc.), collected parameters (status parameters or processing data that need to be collected from the equipment, such as temperature, pressure, speed, processing count, etc.), status detection rules (detection conditions for determining whether the equipment is ready or whether the processing is completed), and timeout threshold (maximum waiting time for the equipment to respond).

[0092] In addition, the attribute configurations for material flow nodes include: material code, flow quantity, source location, and target location; the attribute configurations for anomaly capture nodes include: anomaly type, anomaly identification rules, and priority; and the attribute configurations for anomaly handling nodes include: handling method, handling personnel, notification method, and compensation operation.

[0093] This invention provides a visual process modeling method for industrial production, comprising:

[0094] (1) Graphical process designer: The process design canvas is implemented using HTML5 Canvas / SVG technology, and interactive operations such as drag-and-drop addition of node components, connection of lines, and attribute configuration are supported. The designer provides functional areas such as component library panel, attribute configuration panel, and process preview panel.

[0095] (2) Industrial Extension Node Types: Based on the standard BPMN2.0 node, define node types specific to industrial production, as shown in Table 1.

[0096] ;

[0097] (3) Process model storage structure: The process model definition is stored in JSON format, including metadata such as node list, connection relationship, node attributes, process variables, etc., and supports serialization and deserialization of the process model.

[0098] In one embodiment, step S3, parsing the pre-configured dynamic binding relationship, includes:

[0099] Parse the form templates bound to the nodes, establish the mapping relationship between form fields and process variables, and realize the automatic loading of business form data;

[0100] The business rules bound to the nodes are parsed. These business rules include node entry rules, node execution rules, and node departure rules. The flow conditions of the nodes are controlled based on these business rules.

[0101] The device information bound to the node is parsed, device control commands are generated, and the data adaptation layer is used to send them to the production equipment.

[0102] In one embodiment, step S3, parsing the pre-configured dynamic binding relationship, includes:

[0103] Each process node, quality inspection node, and other industrial extension nodes can be bound to one or more business form templates. When a node is executed, the process engine reads the bound form template definition through the data adaptation layer. This definition includes the form layout, field types, field names, and mapping relationships with process variables.

[0104] Form binding supports multiple mapping modes: one is one-way mapping, where form field values ​​are automatically filled from process variables to form controls; the second is two-way mapping, where data entered by the user in the form is automatically synchronized to the corresponding process variables; and the third is master-child table association mapping, which supports the linkage between master table records and child table details, such as the work order master table being associated with multiple process execution record child tables.

[0105] Field mapping supports configuration of data validation rules, such as required field validation, data type validation, and value range validation. When a node is executed, the process engine automatically loads data from the process variables and populates the form according to the mapping relationship; when the node is executed, the engine automatically writes the form data to the corresponding process variables, realizing seamless flow of business data during process execution.

[0106] In addition, form binding supports dynamic form functionality, which allows for dynamic control of the visibility, editability, or dropdown option data source of form fields based on process variable values. For example, the inspection items in a quality control node can dynamically load the corresponding inspection standard template based on the current process type.

[0107] The node supports configuring three types of business rule expressions. The rule expressions support syntax elements such as variables, operators, and functions. The rule engine component is responsible for parsing and executing these rules.

[0108] Node entry rules are used to determine whether a process instance meets the conditions for entering that node. When a token flows to a node, the execution engine first calculates the entry rule expression. If the result is true, entry to the node is allowed and execution begins; if the result is false, the token remains at the current node or an exception handling branch is triggered. Entry rules can be used to implement precondition checks in production processes, such as "the next process can only proceed if the previous process has passed inspection."

[0109] Node execution rules are used to define the business processing logic within a node. For process nodes, execution rules can define equipment readiness checks at the start of the process, parameter monitoring logic during processing, and completion signal confirmation rules after the process is completed. Node execution rules are parsed and executed by the rule engine during node execution, and the node execution behavior can be dynamically adjusted based on the rule calculation results.

[0110] Node departure rules are used to determine whether to allow execution to proceed to downstream nodes after a node has completed its execution. When a node completes execution, the execution engine calculates the departure rule expression. If the result is true, a token is generated and the node proceeds to the calculated next node; if the result is false, the node remains in a completed state but does not proceed immediately, waiting for conditions to be met before automatic or manual intervention. Departure rules can be used to implement conditional branch control in production processes, such as "if quality inspection is qualified, proceed to the next process; if it is unqualified, proceed to the rework node."

[0111] The rule engine supports compound conditional expressions, such as "${qualityResult == 'PASS' &&defectCount == 0}" or "${temperature >= 180 && temperature <= 220}", and supports dynamically extending custom functions through the SPI mechanism to meet the complex judgment needs in industrial production.

[0112] Process nodes and equipment linkage nodes support binding to specific production equipment information, including equipment code, equipment type, equipment interface protocol, collection parameter list, control command template, etc. During node execution, the process engine communicates and interacts with the production equipment through a data adaptation layer.

[0113] The device binding resolution process includes: First, obtaining the device's connection information (such as IP address, port number, and communication protocol type) from the device master data based on the device code configured in the node; second, parsing the control instruction template configured in the node, filling the process parameters in the process variables into the instruction template, and generating specific device control instructions, such as "send the machining parameter G code to the CNC machine tool" or "send a start signal to the PLC"; then, calling the device communication driver through the data adaptation layer to send the control instructions to the target device; finally, listening to the status signal or completion signal returned by the device and updating the device status to the process variables.

[0114] The data adaptation layer supports multiple industrial communication protocols, including OPC UA, Modbus, MQTT, and HTTP API. It dynamically loads data source drivers for different protocols via the SPI mechanism, enabling interconnection with heterogeneous production equipment. Device binding also supports configuration of status detection rules, such as "start timing after the device status changes to RUNNING" or "mark the node as complete upon receiving a completion signal," and supports setting timeout thresholds, triggering an exception handling branch upon timeout.

[0115] For the collection of equipment data, the equipment binding can be configured to collect a list of parameters (such as temperature, pressure, speed, processing count, etc.). During the execution of nodes, the process engine collects equipment data at the configured frequency and stores the collected data in the process execution history, so as to realize the synchronous tracking of equipment status and process execution.

[0116] This invention designs a dynamic binding mechanism between process nodes and business elements:

[0117] (1) Form binding: Work process nodes can be bound to business form templates, and the mapping relationship between form fields and process variables can be configured to realize the automatic loading and saving of form data when the process is executed. It supports advanced features such as master-child table association, field linkage, and data validation.

[0118] (2) Personnel Binding: Supports multiple personnel allocation strategies, including designated personnel, role allocation, department allocation, dynamic calculation, etc. Configurable processing modes include task assignment, countersigning, OR signing, and transfer.

[0119] (3) Equipment Binding: Process nodes can be bound to production equipment, and equipment codes, collected parameters, control commands and other information can be configured to realize the linkage between process execution and equipment status. It supports functions such as equipment status detection, processing parameter distribution and completion signal reception.

[0120] (4) Rule binding: Supports configuring business rules on nodes, including entry rules (determining whether to enter the node), execution rules (node ​​processing logic), and exit rules (determining whether to leave the node). Rule expressions support syntax elements such as variables, operators, and functions.

[0121] In one embodiment, the exception handling strategy in step S4 includes:

[0122] When a device fault code or quality failure flag is detected, an exception token is generated, and task dispatching to downstream nodes is stopped.

[0123] The abnormality type is matched with a preset abnormality handling branch, which includes process rollback to a specified node, triggering an alarm to notify relevant personnel, or transferring to a rework / scrap handling node.

[0124] Receive feedback on the processing results for the exception handling branch, compensate and update the process variables based on the feedback on the processing results, and resume process execution.

[0125] In one embodiment, the exception handling strategy in step S4 includes:

[0126] Anomaly handling employs an event-driven mechanism, with an anomaly catcher monitoring various anomaly events in real time during process execution. Anomaly identification rules support multiple configuration methods: equipment fault anomalies are identified based on equipment status codes or fault signals; anomalies are triggered when the equipment status code collected by the linked node matches a preset fault code; quality anomalies are determined based on the inspection results of quality control nodes; anomalies are triggered when non-conforming marks or inspection values ​​in the quality control data exceed the range of the conformity judgment rules; timeout anomalies are identified based on task processing time exceeding a preset threshold; and material anomalies are identified based on material shortages or material errors. Anomaly rules support combined conditions and priority settings; for example, it can be configured to "trigger the highest priority anomaly when the equipment fault code is E001 and the current process is a critical process."

[0127] When the exception catcher identifies a matching exception event, the process engine generates an exception token. An exception token is a special type of token, distinct from normal process tokens, and it identifies the exception type, the node where it occurred, the time of the exception, and exception details. Upon generating an exception token, the execution engine immediately stops dispatching normal tasks from the current node to downstream nodes and suspends the execution of the main process to prevent erroneous flow under abnormal conditions.

[0128] The process model definition pre-configures exception capture nodes and corresponding exception handling nodes. Exception capture nodes are specifically designed to listen for specified types of exception events and define the mapping relationship between triggering conditions and exception handling branches. When an exception event is captured, the process engine matches the pre-defined exception handling branch based on the exception type.

[0129] The exception handling branch supports multiple handling strategies:

[0130] Process rollback to a specified node is suitable for anomalies that can be traced and corrected. The process engine supports instance rollback operations, allowing reversal to the previous step or any historical node. During rollback, executed operations and data changes are automatically undone, restoring the process state to its pre-execution state at the target node for easy re-execution. For example, when quality inspection finds a non-conformity, rollback can be performed to the previous process node for rework.

[0131] Alarm notifications are applicable to abnormal scenarios requiring manual intervention. Alarm notifications support multiple channels, including in-app messages, SMS, email, and WeChat Work. Notification content can be dynamically populated with abnormal information such as abnormal type, occurrence time, and abnormal details. Alarm rules can be configured with parameters such as notified personnel, notification method, and repeat notification interval.

[0132] The transfer to the rework / scrap handling node is applicable to quality anomaly scenarios. Based on the configuration of the anomaly handling branch, the process engine transfers the token to the preset rework node or scrap node to execute the preset rework procedure or scrap approval process. The rework node can be configured with rework procedures, rework inspection requirements, etc.; the scrap node can be configured with scrap approval processes, material scrap registration, etc.

[0133] For severe anomalies, the anomaly handling branch can also be configured with strategies such as suspending the process for manual handling or executing preset compensation operations.

[0134] Once the exception handling branch has completed execution, the workflow engine receives feedback on the processing result. This feedback can come from manual operation (e.g., operator confirmation that the fault has been resolved) or automatic system judgment (e.g., equipment status restored to ready, quality re-inspection passed, etc.).

[0135] Based on the feedback of the processing results, the process engine performs compensation update operations. The compensation operations include: modifying the changed process variables according to the processing results (such as updating equipment status flags, resetting timeout counters, recording rework counts, etc.); revoking the lock status of the exception token on the process; clearing unresolved flags in the exception log; and updating business data according to the compensation rules (such as adjusting production plans, updating material inventory, etc.).

[0136] After the compensation update is completed, the process engine resumes process execution. Recovery methods include: converting the exception token to a normal token and re-executing from the node where the exception occurred; transferring the token to the recovery node specified in the exception handling branch to continue execution; or re-executing the rolled-back node sequence according to the rollback configuration. The process supports lifecycle operations such as pause, resume, and terminate, ensuring a smooth continuation of process execution after exception handling.

[0137] All anomaly information is recorded in the anomaly log, including anomaly type, occurrence point, anomaly time, anomaly details, handling strategy, handling result, handling personnel, and handling time, providing data support for process optimization and problem tracing.

[0138] This invention designs a comprehensive process exception handling mechanism:

[0139] (1) Exception Node Types: Exception capture nodes and exception handling nodes are specifically defined, and exception handling branches can be preset in the process. Exception capture nodes listen for exception events of a specified type, and when triggered, the process flows to the corresponding exception handling branch.

[0140] (2) Anomaly identification rules: Supports configuration of anomaly identification conditions, including various anomaly triggering conditions such as equipment fault codes, quality non-conformity marks, and unprocessed timeouts. Anomaly rules support combination of conditions and priority settings.

[0141] (3) Exception handling strategy: Supports multiple exception handling methods, including: process suspension waiting for manual handling, automatic flow to exception handling node, triggering alarm to notify relevant personnel, executing preset compensation operation, process rollback to the previous node or specified node, etc.

[0142] (4) Process rollback and recovery: Supports rollback operations for process instances, allowing reversal to the previous step or any historical node. During rollback, executed operations and data changes are automatically undone. Supports lifecycle operations such as process pause, recovery, and termination.

[0143] In one embodiment, step S5 is followed by a process version management step:

[0144] Receive modification instructions for the current process model and generate a new version of the process model;

[0145] Mark the running process instance as a pending migration state, and perform path mapping on the process instance in the pending migration state based on the node definition of the new version of the process model;

[0146] While ensuring the consistency of business data, the process instance in the state to be migrated will be smoothly migrated to the new version of the process model for continued execution.

[0147] In one embodiment, step S5 is followed by a process version management step:

[0148] The process modeling module supports version management of published process models. When a user modifies an existing process model through the visual designer (such as adding nodes, modifying node attributes, adjusting flow connections, changing exception handling branches, etc.) and triggers a release operation, the process engine receives the modification instruction and performs a validity check on the process model. After the check passes, the system saves the modified process model as the new version.

[0149] The version management mechanism maintains an independent version sequence for each process model. Version information includes metadata such as version number (using semantic versioning specifications, such as major version number, minor version number, and revision number), publisher, release time, and change description. Historical versions are fully preserved, supporting version rollback and version comparison functions. Users can view any historical version of the process definition in the process designer, or create new branch versions based on historical versions.

[0150] The release of process models supports multiple strategies: First, a new version can be released directly, and newly launched process instances will use the latest version by default; second, a canary release can be implemented, where the new version of the process can be specified to only be effective for some instances (such as canary screening by production workshop, product type or work order number), and then fully switched after verification, so as to reduce the production risks brought about by process changes.

[0151] When it is necessary to migrate running process instances from an older version to a newer version, the system first identifies and filters the running process instances. Administrators can specify the migration scope, such as migrating all running instances, migrating only instances on specific nodes, or migrating only instances started within a specified time range.

[0152] The selected process instance is marked as pending migration. Before the migration is executed, the system saves a snapshot of the instance's current state, including runtime information such as the current active node, process variable values, pending task list, token distribution, and exception records.

[0153] Path mapping is the core of version migration. The system maps and transforms nodes of old version process instances based on the node definitions of the new version's workflow model. Mapping rules include: node identifier mapping (establishing a correspondence between old and new versions of nodes based on node codes or names); node attribute mapping (converting attribute values ​​of old version nodes according to the attribute definitions of the new version, such as setting default values ​​for newly added required attributes and ignoring deleted attributes); workflow path mapping (recalculating reachable downstream nodes for the current active node based on the workflow logic of the old and new versions); and task mapping (migrating pending tasks from nodes in the old version to the task list of the corresponding nodes in the new version).

[0154] For nodes added in the new version, the system supports configuring a default initial state (such as pending execution, automatic skip, etc.); for nodes deleted in the new version, the system performs a compatibility check during migration to ensure that no process instance is currently stuck on the deleted node, otherwise the node must be completed or rolled back before the migration is performed.

[0155] The migration process employs a transaction mechanism to ensure data consistency. The system incorporates the version migration operations of process instances into unified transaction management, including updating process definition references, resetting process instance states, relocating activity nodes, reallocating task lists, and converting process variables. During the migration process, if any step fails, the system automatically performs a transaction rollback, restoring the process instance to its state before the migration and recording the migration failure in a log.

[0156] After migration, the process instance continues to execute under the new version of the process model. The execution engine uses the new version's node definitions, flow logic, dynamic binding configurations, and exception handling strategies in subsequent workflows. For issues such as missing process variables or format incompatibility caused by version differences during the migration process, the system provides a compensation mechanism, such as filling in with preset default values ​​or performing data conversion through custom scripts.

[0157] After the migration is complete, the system records a migration log, including metadata such as migration time, operator, list of migrated process instances, version change information (from old version to new version), and migration result (success / failure and reason for failure). The migrated process instances can be viewed normally in the monitoring and analysis module, and historical execution records retain execution data from the old version, ensuring traceability throughout the entire process.

[0158] In addition, the system supports a canary migration strategy for process instances, which means migrating a small number of test instances first to verify the stability and compatibility of the new version, and then migrating the remaining instances in batches after confirming that there are no problems, thereby further reducing the impact of version changes on the production process.

[0159] In one embodiment, the process execution engine of the present invention further includes the following technical method:

[0160] (1) Process instance management: When a process starts, a process instance is created, a unique instance identifier is assigned, and process variables are initialized. The process instance maintains runtime information such as the current list of active nodes, historical execution records, and process status.

[0161] (2) Token-driven mechanism: The process execution is driven by a token, which is transferred between nodes to trigger node execution. It supports a multi-token mechanism generated by parallel gateways to achieve parallel execution and convergence of branch processes.

[0162] (3) Asynchronous execution optimization: For node executions that take a long time (such as waiting for equipment processing or external system calls), an asynchronous execution mechanism is adopted to avoid blocking the process engine. After the node is completed, the engine is notified to continue execution through callback or message queue.

[0163] (4) Transaction Management: Process status changes and business data operations are incorporated into unified transaction management to ensure data consistency. Support for transaction commit, rollback, and compensation mechanisms.

[0164] This invention provides comprehensive process monitoring and analysis capabilities:

[0165] (1) Real-time process tracking: Provides real-time status query of process instances, including information such as current active nodes, pending tasks, and process variable values. Supports visual replay of process execution paths.

[0166] (2) Process performance monitoring: Statistical analysis of performance indicators such as execution time, waiting time, and number of processing times for each process and node to identify process bottlenecks and inefficient links.

[0167] (3) Alarms and Notifications: Supports configuring process alarm rules, such as triggering alarm notifications in cases of unprocessed task timeouts, abnormal process termination, and node execution failures. Alarm methods support multiple channels such as in-site messages, SMS, email, and WeChat Work.

[0168] (4) Process data analysis: Perform multi-dimensional analysis on process execution data to generate reports such as process efficiency reports, node processing statistics, and personnel workload analysis, providing data support for process optimization.

[0169] This invention supports version management and dynamic updates of the process:

[0170] (1) Version control: The process model supports version management. Each release generates a new version and retains historical version information. Version information includes metadata such as version number, publisher, release time, and change description.

[0171] (2) Gray release: Supports gray release strategy for processes. New versions of processes can be specified to be effective only for some instances. After verification, they can be fully switched to reduce the risk of process changes.

[0172] (3) Instance migration: Supports version migration of process instances, which can migrate running instances from the old version to the new version. The migration process ensures the integrity of instance status and business data.

[0173] Compared with the prior art, the present invention has the following beneficial effects:

[0174] (1) Significantly reduce the threshold for process configuration: By configuring the production process through a visual "drag, drop, and drag" method, business personnel can design and adjust the process without programming knowledge, and the response time for process changes is shortened from days to hours.

[0175] (2) Deeply adapted to industrial scenarios: The specially designed industrial extension node type effectively supports industrial-specific scenarios such as process management, quality inspection, and equipment linkage, filling the application gap of general workflow engines in the industrial field.

[0176] (3) Achieve deep integration of process and business: Through dynamic binding mechanism, realize flexible association between process nodes and forms, personnel, equipment and rules, and build a data-driven process execution system.

[0177] (4) Enhanced anomaly handling capabilities: A comprehensive anomaly identification and handling mechanism supports a variety of anomaly handling strategies, effectively responding to various emergencies in the production process and ensuring production continuity.

[0178] (5) Enhance process visualization and control: Real-time process monitoring and analysis capabilities enable managers to fully grasp the production progress, promptly identify and resolve problems, and improve the level of production management.

[0179] (6) Support continuous process optimization: Version management and canary release mechanism support safe iterative updates of the process, and process data analysis provides quantitative basis for process optimization, promoting continuous improvement of production process.

[0180] The following describes the implementation architecture of the method in this application.

[0181] System Deployment Architecture: The process engine system of this invention can be deployed using a microservice architecture. Core services include process design services, process execution services, task scheduling services, and monitoring and analysis services. Each service can be deployed independently and scaled horizontally, supporting containerized deployment and Kubernetes orchestration. The process engine supports persistence of process data in various databases (MySQL, PostgreSQL, Oracle, etc.), supports Redis caching to improve performance, and supports RocketMQ / Kafka message queues for asynchronous processing.

[0182] Workflow Designer Implementation: The workflow designer's front-end is developed based on the Vue.js / React framework, using AntV X6 or bpmn-js as the flowchart drawing engine. The left side of the designer is a component library panel displaying various available node components; the middle is the canvas area, supporting operations such as dragging, connecting, selecting, and deleting nodes; the right side is a property configuration panel, dynamically displaying configurable properties based on the selected node type. The designer supports saving, previewing, and publishing the workflow, performing validity checks on the workflow model during publishing.

[0183] Workflow Engine Core Implementation: The core workflow engine is implemented in Java and built using the Spring Boot framework. Core components include:

[0184] ProcessDefinitionParser: A process definition parser responsible for parsing JSON-formatted process models and constructing process definition objects in memory.

[0185] RuntimeService: A runtime service that provides interfaces for starting, querying, and terminating process instances.

[0186] TaskService: The task service is responsible for tasks such as creation, assignment, completion, and transfer.

[0187] ExecutionEngine: The execution engine drives the execution of process instances, manages the flow of tokens, and schedules nodes.

[0188] HistoryService: This service records historical data of process execution, supporting querying and tracing.

[0189] Node Executor Implementation: Various node executors implement the NodeExecutor interface, defining the execute() method to encapsulate the node execution logic. Process node executors are responsible for loading and binding forms, assigning processing personnel, and waiting for task completion; device node executors are responsible for issuing control commands, monitoring device status, and receiving completion signals; gateway node executors are responsible for calculating flow conditions and controlling branch convergence. Executors support dynamic expansion via the SPI mechanism, facilitating the addition of custom node types.

[0190] Exception handling implementation: Exception handling adopts an event-driven mechanism. Exception events generated during process execution are monitored by an exception catcher. The exception catcher determines the exception handling strategy based on the exception type and configuration rules. For recoverable exceptions, an exception handling process branch is triggered; for severe exceptions, process rollback or termination is executed. All exception information is recorded in the exception log, including exception type, occurrence point, exception time, and handling result.

[0191] Taking the implementation of the MES system in the machining workshop of an automotive parts manufacturing company as an example, the specific application of this invention is illustrated:

[0192] Scenario Description: This company produces automotive engine parts, involving multiple processes such as turning, milling, drilling, heat treatment, grinding, and quality inspection. It needs to manage complex production processes such as process routes, equipment scheduling, and quality traceability.

[0193] Process Design: Process engineers use a process designer to drag and drop process nodes, configuring processing parameters, equipment requirements, and time quotas for each process. Quality inspection nodes are added after critical processes, configuring inspection items and acceptance criteria. Anomaly handling branches are set up to handle abnormal situations such as equipment failures and quality defects.

[0194] Process Execution: The production planner issues production work orders, the process engine creates process instances, and drives the execution of each process sequentially. Process nodes automatically send machining instructions to CNC machine tools, collecting equipment operating data and machining parameters. Quality inspection nodes record inspection data; if data is found to be non-conforming, a rework or scrap process is triggered.

[0195] Process monitoring: Workshop managers can view the production progress, equipment status, and quality of each work order in real time through a large monitoring screen. The system automatically calculates indicators such as processing time, equipment utilization rate, and first-pass yield for each process, providing data support for production optimization.

[0196] In one embodiment, a configurable process engine system based on an industrial production process is provided, comprising:

[0197] The instantiation module is used to receive process model definition data, parse the process model definition data to obtain node configuration information and flow logic. The node configuration information includes standard process nodes and industrial extended nodes. The industrial extended nodes include process nodes, quality inspection nodes and equipment linkage nodes. In response to the process start command, the module creates a process instance and initializes process variables, and drives the execution of the first node of the process instance based on the process model definition data.

[0198] The execution scheduling module is used to parse the pre-configured dynamic binding relationship during the node execution process, call the external business system interface to obtain or write business data according to the dynamic binding relationship, and calculate the next execution node according to the flow logic in the data defined by the process model.

[0199] The exception monitoring module is used to monitor exception events in real time during the process flow. When an exception event is detected, the current process is interrupted and the execution jumps to the exception handling branch node according to the preset exception handling strategy until a recovery instruction is received or the preset automatic recovery conditions are met.

[0200] The historical archiving module is used to collect execution trajectory data and business data of process instances during process execution, and to associate and store the execution trajectory data and business data to generate a traceable process execution history.

[0201] In one embodiment, the system further includes: a modeling module, configured to receive drag-and-drop operations from a user via a visual designer, generate a flowchart containing the industrial extension nodes; perform a validity check on the flowchart, serialize the valid flowchart into a JSON-formatted process model definition file; and generate process model definition data; wherein the attribute configuration of the industrial extension nodes includes: processing parameters of process nodes, qualification judgment rules of quality inspection nodes, and control instructions of equipment linkage nodes.

[0202] In one embodiment, the execution scheduling module includes:

[0203] The rule engine unit is used to parse and execute the business rule expressions configured on the node;

[0204] The data adaptation unit is used to dynamically load external data source drivers via the SPI mechanism to achieve communication with heterogeneous databases and production equipment;

[0205] The asynchronous task unit is used to encapsulate time-consuming external system calls into asynchronous tasks, and decouples them through message queues to avoid blocking the core process engine.

[0206] In one embodiment, an electronic device is provided, including a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the steps of the method described in the embodiments above.

[0207] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the embodiments above.

[0208] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A configurable process engine execution method based on industrial production processes, characterized in that, include: S1, receive process model definition data, parse the process model definition data to obtain node configuration information and flow logic, the node configuration information includes standard process nodes and industrial extended nodes, the industrial extended nodes include process nodes, quality inspection nodes and equipment linkage nodes. S2, in response to the process start command, create a process instance and initialize process variables, and drive the execution of the first node of the process instance based on the data defined by the process model; S3, during node execution, parse the pre-configured dynamic binding relationship, call the external business system interface to obtain or write business data according to the dynamic binding relationship, and calculate the next execution node according to the flow logic in the data defined by the process model; S4 monitors abnormal events in real time during the process flow. When an abnormal event is detected, the current process is interrupted and the execution jumps to the abnormal handling branch node according to the preset abnormal handling strategy until a recovery instruction is received or the preset automatic recovery conditions are met. S5. During process execution, collect the execution trajectory data and business data of the process instance, and associate and store the execution trajectory data and business data to generate a traceable process execution history.

2. The method according to claim 1, characterized in that, In step S1, before receiving the process model definition data, a process modeling step is also included: The process involves receiving drag-and-drop operations from the user through a visual designer and generating a flowchart that includes the industrial extension nodes. The flowchart is validated for legality, and the validated flowchart is serialized into a JSON format flowchart model definition file; The attribute configuration of the industrial extension node includes: processing parameters of the process node, qualification judgment rules of the quality inspection node, and control instructions of the equipment linkage node.

3. The method according to claim 1, characterized in that, In step S3, parsing the pre-configured dynamic binding relationship includes: Parse the form templates bound to the nodes, establish the mapping relationship between form fields and process variables, and realize the automatic loading of business form data; The business rules bound to the nodes are parsed. These business rules include node entry rules, node execution rules, and node departure rules. The flow conditions of the nodes are controlled based on these business rules. The device information bound to the node is parsed, device control commands are generated, and the data adaptation layer is used to send them to the production equipment.

4. The method according to claim 1, characterized in that, The exception handling strategy in step S4 includes: When a device fault code or quality failure flag is detected, an exception token is generated, and task dispatching to downstream nodes is stopped. The abnormality type is matched with a preset abnormality handling branch, which includes process rollback to a specified node, triggering an alarm to notify relevant personnel, or transferring to a rework / scrap handling node. Receive feedback on the processing results for the exception handling branch, compensate and update the process variables based on the feedback on the processing results, and resume process execution.

5. The method according to claim 1, characterized in that, Step S5 is followed by a process version management step: Receive modification instructions for the current process model and generate a new version of the process model; Mark the running process instance as a pending migration state, and perform path mapping on the process instance in the pending migration state based on the node definition of the new version of the process model; While ensuring the consistency of business data, the process instance in the state to be migrated will be smoothly migrated to the new version of the process model for continued execution.

6. A configurable process engine system based on industrial production processes, characterized in that, include: The instantiation module is used to receive process model definition data, parse the process model definition data to obtain node configuration information and flow logic. The node configuration information includes standard process nodes and industrial extended nodes. The industrial extended nodes include process nodes, quality inspection nodes and equipment linkage nodes. In response to the process start command, the module creates a process instance and initializes process variables, and drives the execution of the first node of the process instance based on the process model definition data. The execution scheduling module is used to parse the pre-configured dynamic binding relationship during the node execution process, call the external business system interface to obtain or write business data according to the dynamic binding relationship, and calculate the next execution node according to the flow logic in the data defined by the process model. The exception monitoring module is used to monitor exception events in real time during the process flow. When an exception event is detected, the current process is interrupted and the execution jumps to the exception handling branch node according to the preset exception handling strategy until a recovery instruction is received or the preset automatic recovery conditions are met. The historical archiving module is used to collect execution trajectory data and business data of process instances during process execution, and to associate and store the execution trajectory data and business data to generate a traceable process execution history.

7. The system according to claim 6, characterized in that, Also includes: The modeling module is used to receive drag-and-drop operations from the user through a visual designer and generate a flowchart that includes the industrial extension nodes. The flowchart is validated for legality, and the validated flowchart is serialized into a JSON format flowchart model definition file; Generate process model definition data; wherein, the attribute configuration of the industrial extension node includes: processing parameters of the process node, qualification judgment rules of the quality inspection node, and control instructions of the equipment linkage node.

8. The system according to claim 6, characterized in that, The execution scheduling module includes: The rule engine unit is used to parse and execute the business rule expressions configured on the node; The data adaptation unit is used to dynamically load external data source drivers via the SPI mechanism to achieve communication with heterogeneous databases and production equipment; The asynchronous task unit is used to encapsulate time-consuming external system calls into asynchronous tasks, and decouples them through message queues to avoid blocking the core process engine.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the steps of the method as described in any one of claims 1 to 5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images