An electric heating table production process control method and system based on the internet of things

Abstract

The present application relates to the technical field of industrial automation control and intelligent manufacturing, and particularly relates to an electric heating table production process control method and system based on the Internet of Things. The method comprises the following steps: obtaining an ERP order and extracting a process characteristic vector; performing hierarchical matching and station reorganization to generate an equipment parameter formula; performing interface access and message conversion to obtain unified working condition data flow; performing dual-core abnormality judgment and stop-line lock material alarm three-level linkage based on the unified working condition data flow to obtain a linkage state set; performing thermodynamic load estimation and feedforward proportional integral duty cycle adjustment based on the linkage state set and an aging temperature curve to obtain a heating control amount; performing primary key registration and batch association to obtain a traceability data set; and performing loss function construction and iterative updating based on the traceability data set to obtain an updated formula. The present application realizes intelligent closed-loop control of the electric heating table production process, and significantly improves production quality and energy utilization efficiency.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control and intelligent manufacturing technology, and in particular to a method and system for controlling the production process of an electric heating table based on the Internet of Things. Background Technology

[0002] In the field of industrial automation control and intelligent manufacturing technology, existing solutions typically build production management and data acquisition links around enterprise resource planning systems, shop floor execution systems, testing equipment, aging equipment, and field acquisition devices. These solutions have limitations such as the disconnect between equipment parameter formulas and real-time workstation operating data, the separation of anomaly judgment and production stoppage material locking control, and the failure to integrate the number of online units and temperature status into a unified aging control link.

[0003] Existing methods often rely on a processing path of order placement, workstation data collection, manual verification, and post-event traceability. In scenarios involving mixed-flow production of electric heating tables, frequent workstation switching, and parallel operation of testing and aging, quality anomalies are prone to spread and heating regulation mismatch at aging workstations can occur. It is difficult to achieve stable implementation of anomaly judgment and line stoppage material locking control based on equipment parameter formulas for a unified operating condition data stream, and to generate heating control quantities by combining the number of online units and temperature status.

[0004] Existing technologies generally suffer from common shortcomings in the joint processing of equipment parameter formulas, unified operating condition data streams, line stop material locking control, online unit count, temperature status, heating control quantities, and traceability datasets. These shortcomings include fragmented workstation data acquisition links, reliance on single-point alarms for quality linkage processing, disconnection between aging linkage processing and previous linkage status, and separation between traceability collection processing and formula update processing. As a result, it is difficult to form a consistent process of data acquisition, judgment, control, collection, and updating in the production process control scenario of electric heating tables, making it difficult to maintain a continuous correspondence between production cycle time, quality linkage, aging control, and traceability records.

[0005] Electric heating table production generally employs a parallel production model of discrete and mixed-flow processes. Multiple models need to be switched within a single shift on the same production line, including frequent switching between different rated power levels and between square and round tabletops, resulting in significant changeover time. Existing manufacturing execution systems primarily rely on barcode scanners to upload station records, with data granularity at one record per unit, failing to reflect real-time parameters at the workstation level, including soldering iron temperature, electric screwdriver torque, test waveforms, and aging chamber temperature and humidity. Key equipment such as automatic screwdrivers, automatic testing equipment, and aging chambers mostly use standalone serial communication, unable to receive reverse control commands from the cloud. Energy management is relatively crude; aging chambers typically maintain a fixed heating temperature, regardless of the number of units online, resulting in significant energy waste per shift. Quality traceability is time-consuming; when market complaints arise, engineers must manually link data from multiple independent spreadsheets, which is inefficient and prone to errors. Summary of the Invention

[0006] To address the above problems, this invention provides a method and system for controlling the production process of electric heating tables based on the Internet of Things (IoT). This system solves the problem of how to perform anomaly detection and production stoppage control on a unified operating condition data stream according to equipment parameter formulas, and how to generate heating control quantities by combining the number of online units and temperature status.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] This invention provides a method for controlling the production process of an electric heating table based on the Internet of Things, comprising:

[0009] S100: Obtain ERP orders, perform feature extraction and vector generation processing to obtain process feature vectors;

[0010] The ERP order specifically includes an order number, model number, planned quantity, delivery batch, power specification field, and color code field. The power specification field is used to perform cross-validation with the model number to generate a rated power identifier. The color code field is used to perform a corresponding match with the color coding table in the material master data to generate a color identifier. The model number serves as a mapping index for subsequent recipe calls to retrieve the corresponding workstation parameters.

[0011] S200. Based on the process feature vector, perform hierarchical matching and workstation number reorganization to obtain the equipment parameter formula;

[0012] S300. Based on the device parameter formula, perform interface access and message conversion processing to obtain a unified operating condition data stream;

[0013] S400. Based on the unified working condition data stream, perform dual-core anomaly judgment and three-level linkage processing of line stop and material locking alarm to obtain a linkage status set.

[0014] S500. Based on the linkage state set and aging temperature curve parameters, perform thermodynamic load estimation and feedforward proportional-integral dynamic duty cycle adjustment to obtain the heating control quantity.

[0015] S600. Based on the heating control quantity and the barcode scanning record, perform primary key registration and batch association processing to obtain a traceability dataset;

[0016] S700. Based on the traceability dataset, a loss function is constructed and iteratively updated to obtain an updated formula.

[0017] Furthermore, the process of feature extraction and vector generation includes:

[0018] Perform field integrity checks, version number checks, and timestamp sorting on ERP orders, remove duplicate messages, and retain the current valid order version;

[0019] Read the model number, power specification field, and color code field from the ERP order, and identify the desktop shape identifier from the first workstation image;

[0020] The identification of the desktop shape identifier includes: acquiring a top view image of the desktop by the first workstation image acquisition unit of the production line, performing contour segmentation, edge closure and template matching, and identifying the desktop outer contour with a preset shape template library;

[0021] The power specification field is cross-verified with the model number to generate a rated power identifier;

[0022] The color code field is matched with the color coding table in the material master data to generate a color identifier;

[0023] Once the desktop shape identifier, rated power identifier, and color identifier have all passed the corresponding verification with the model mapping table, the order number, model number, desktop shape identifier, rated power identifier, color identifier, order version number, and generation timestamp are encapsulated in a fixed field order to generate a process feature vector.

[0024] Simultaneously, a difference comparison is performed on the process feature vectors of adjacent orders. When the desktop shape identifier changes, a fixture replacement trigger mark is recorded. When the rated power identifier changes, a test threshold replacement trigger mark is recorded. When the color identifier changes, an exterior part replacement trigger mark is recorded. All of these trigger marks are written into the process feature vector.

[0025] Furthermore, the process of performing hierarchical matching and workstation number reorganization includes:

[0026] The hierarchical matching process includes:

[0027] Perform tooling layer matching on the desktop shape identifier, and call the fixture parameter segment and assembly torque parameter segment corresponding to the square desktop or round desktop;

[0028] Perform electrical layer matching on the rated power label, and call the test upper and lower limit parameter segments and aging temperature curve parameter segments corresponding to the power level;

[0029] Perform appearance layer matching on the color identifier and call the appearance component switching mark and packaging switching mark corresponding to the color code;

[0030] The workstation number reorganization process includes: generating workstation numbers according to raw material kitting workstation, SMT placement workstation, ICT testing workstation, functional testing workstation, aging workstation, assembly workstation and packaging workstation, and then binding each parameter segment in the parameter template to the corresponding workstation number to form a workstation-level recipe record.

[0031] Generate a recipe version number, an effective status flag, a call source flag, and a distribution order flag for the workstation-level recipe record;

[0032] When the trigger marker in the process feature vector includes a fixture change trigger marker, the assembly station and fixture preparation station will be placed in the preceding order of issuance; when it includes a test threshold change trigger marker, the functional test station and aging station will be placed in the preceding order of issuance; when it includes an appearance part change trigger marker, the packaging station and appearance part material preparation station will be placed in the preceding order of issuance.

[0033] The packaged device parameter formula includes the workstation number, solder paste thickness parameter, screw torque parameter, test upper and lower limit parameter, aging temperature curve parameter, formula version number, effective status flag, call source flag, and distribution sequence flag.

[0034] Furthermore, the process of interface access processing includes:

[0035] Read the corresponding parameter segment in the equipment parameter recipe according to the workstation number, and establish the access sequence table for the current batch according to the issuance sequence mark;

[0036] Print parameter labels and equipment status labels corresponding to the solder paste thickness parameters of the SMT placement equipment;

[0037] Automatic screw-driving equipment should be equipped with labels for the torque values ​​corresponding to the screw torque parameters, the nozzle number label, and the tightening completion label.

[0038] Load the test equipment with the reset pulse width data label, electrical parameter label and test conclusion label corresponding to the upper and lower limit parameters of the test;

[0039] Load the aging temperature curve parameters into the aging chamber PLC with labels for the aging chamber's internal temperature, external temperature, number of online units, and heating status.

[0040] Furthermore, the message conversion process includes:

[0041] The serial port acquisition frame is split into the device number field, workstation number field, parameter identifier field, parameter value field, acquisition timestamp field, and recipe version number field, and the serial port acquisition frame is converted into an MQTT message;

[0042] The MQTT message is sent via a transparent serial bridge, carrying TLS 1.3 encryption configuration and watchdog reset configuration;

[0043] The converted messages are continuously written to the edge gateway buffer in chronological order, generating a unified operating condition data stream containing fields such as device number, workstation number, parameter identifier, parameter value, collection timestamp, recipe version number, interface type, message subject, and data status.

[0044] Furthermore, the process of handling dual-core anomaly detection includes:

[0045] The dual-core A7 processor receives the unified operating condition data stream and performs message proxying and formula comparison, while the MCU coprocessor receives the anomaly judgment result and performs line stop material locking control and safety relay driving.

[0046] Data frames corresponding to the test station, data frames corresponding to the assembly station, and data frames with normal data status fields are filtered out from the unified working condition data stream to form a message queue to be judged.

[0047] Read the workstation number field and recipe version number field from the message queue to be judged, and retrieve the parameter segment of the same workstation and version from the equipment parameter recipe;

[0048] A range comparison is performed on the reset pulse width data. When the reset pulse width data falls outside the range of 160ms to 220ms, a reset anomaly flag is generated.

[0049] A threshold comparison is performed on the actual torque value corresponding to the screw torque parameter, and a torque anomaly flag is generated when it exceeds the allowable range.

[0050] A test exception flag is generated when the test device returns a failure status and the parameter identifier field points to a functional test item.

[0051] Furthermore, the three-level linkage process for line stoppage and material locking alarms includes:

[0052] After receiving a reset error flag, torque error flag, or test error flag, the MCU coprocessor executes the actions in the order of line stop control, yellow light drive, and feeder locking.

[0053] The line stop control outputs a line stop signal through a safety relay, which acts on the operating circuit of the production line where the current workstation is located;

[0054] The locking control writes a locking control frame to the feeder, triggering the feeder to stop the next package of components from entering the current batch.

[0055] The alarm driver writes a yellow light drive signal to the alarm light;

[0056] Wait for the safety relay to return an execution confirmation signal and the feeder to return a material locking confirmation signal. If they do not return within the preset period, write an action unconfirmed flag and maintain the line stop state.

[0057] The equipment number field, workstation number field, exception type field, line stop status field, yellow light status field, material lock status field, data collection timestamp field, and recipe version number field are encapsulated into a linkage status set.

[0058] Furthermore, the process of estimating the thermodynamic load includes:

[0059] Read the shutdown status field, exception type field, and recipe version number field from the linkage status set;

[0060] When the outage status field indicates that the current batch is in an outage state, the current batch is written to the aging freeze area and a freeze mark is generated;

[0061] When the stop status field indicates that it is in the release status, the transmission cycle signal, the in-room count signal and the out-of-room count signal are collected from the aging room PLC, and the online station digital segment is generated by combining the product serial number mapping relationship.

[0062] Read the temperature inside and outside the aging chamber and generate a temperature status field;

[0063] It receives the online station digital segment, the aging chamber internal temperature field, the aging chamber external temperature field, and the aging temperature curve parameters. It identifies the heat capacity range based on the current batch's preheating stage, constant temperature stage, or cooling stage, and then identifies the heat dissipation status based on the internal and external temperature difference, outputting the basic duty cycle value.

[0064] Furthermore, the process of feedforward proportional-integral dynamic duty cycle adjustment includes:

[0065] The feedforward control unit receives the basic duty cycle value and temperature state change results, and performs pilot correction on the heating tube drive quantity;

[0066] The proportional-integral control unit receives the feedforward correction result and the current temperature field in the aging chamber, performs continuous tracking correction, and generates the heating tube PWM duty cycle field.

[0067] When the batch is in a frozen state, switch the base duty cycle value to the frozen duty cycle value and stop the amplification calculation for the number of new units.

[0068] The workstation number field, recipe version number field, online station number field, aging chamber internal temperature field, aging chamber external temperature field, heating tube PWM duty cycle field, heating stage field, freeze status field, and control timestamp field are encapsulated into heating control quantities and written to the edge gateway buffer and aging chamber PLC execution register area.

[0069] Furthermore, an IoT-based electric heating table production process control system, used to execute the method described in any of the above embodiments, includes:

[0070] The order parsing module extracts the desktop shape, rated power, and color to generate a process feature vector;

[0071] The recipe calling module calls the solder paste thickness parameters, screw torque parameters, test upper and lower limit parameters, and aging temperature curve parameters corresponding to the workstation number to generate the equipment parameter recipe;

[0072] The workstation access module collects real-time workstation status data and converts serial port interface data into message queue telemetry transmission protocol messages to generate a unified work status data stream.

[0073] The quality linkage module performs anomaly detection, line stop control, and material locking control, and generates a linkage status set.

[0074] The aging linkage module calculates the pulse width modulation duty cycle of the heating tube by combining the number of online units and the temperature status, and generates the heating control quantity.

[0075] The traceability and aggregation module associates product serial numbers with operating condition data, personnel identification, equipment identification, material batches, process versions, and environmental data to generate traceability datasets.

[0076] The formula update module combines data on changeover duration, number of quality anomalies, and energy consumption to perform iterative updates and generate updated formulas.

[0077] The key innovations of this invention include:

[0078] (1) Construct a heterogeneous processing architecture consisting of a dual-core A7 processor and an MCU coprocessor. The dual-core A7 processor receives a unified working condition data stream and performs message proxy and recipe comparison. The MCU coprocessor receives the abnormal judgment result and performs line stop and material lock control and safety relay drive. On this basis, threshold comparison and range comparison are performed on the reset pulse width data, torque value and solder paste thickness value. Consistency comparison is performed on the work station number and recipe version number. When any comparison is abnormal, a reset abnormal mark, torque abnormal mark or test abnormal mark is generated, and the three-level linkage action of line stop control, material lock control and alarm light drive is triggered. At the same time, it waits for the execution confirmation signal and maintains the line stop state when no confirmation is received, and outputs the linkage state set.

[0079] (2) Based on the stop status field of the linkage status set, when it is in the release state, the transmission cycle signal, the entry count signal and the exit count signal are collected from the aging room PLC to generate the online unit digital segment, and the temperature inside the aging room and the temperature outside the aging room are read; according to the aging period, the number of online units and the temperature difference between inside and outside the current batch, the output basic duty cycle value is estimated by the thermodynamic model, and then the heating tube PWM duty cycle field is generated by feedforward control and proportional integral control; when the stop status field indicates the stop, the current batch is written into the aging freeze zone, the duty cycle value is switched to the freeze value and the amplification calculation of the newly added units is stopped, and finally the heating control quantity is output.

[0080] (3) Establish a closed-loop process from process feature vector generation, equipment parameter formula distribution, unified operating condition data stream acquisition, linkage status set and heating control quantity output, to traceability dataset registration and loss function-driven formula iteration update; wherein the formula update processing constructs evaluation items around the changeover time, number of quality anomalies and energy consumption data, performs local correction on patch parameter segment, assembly parameter segment, test parameter segment and aging parameter segment according to the source of anomaly, and writes back the approved new formula version number to the formula call processing.

[0081] The following are its main beneficial effects:

[0082] (1) By using a dual-core heterogeneous architecture and a three-level linkage action, the problems of separation between anomaly detection and stop-line material locking control and easy failure of single-point alarms in the existing solution are solved. Real-time linkage of the entire link from anomaly detection to physical execution is realized, ensuring that batches with quality defects are stopped, locked and alarms are isolated in time, effectively preventing defective products from spreading to downstream workstations and improving the reliability and safety of the production process.

[0083] (2) By dynamically identifying the number of online units, estimating using a thermodynamic model, and implementing feedforward proportional-integral control, combined with freezing mode switching, the defect of the traditional aging chamber's fixed temperature setpoint being disconnected from actual load changes is overcome. This achieves automatic adaptation of the aging temperature to the number of online units and the ambient temperature difference, precisely adjusting heating in the released state, and maintaining heat preservation and avoiding energy waste in the stopped state, thereby significantly improving the energy utilization efficiency and temperature control stability of the aging process.

[0084] (3) By using a closed-loop data system across the entire supply chain and loss function-driven iterative updates of the formula, the disconnect between traceability collection and formula optimization in existing technologies has been changed. This enables targeted correction of equipment parameter formulas based on real production data (changeover time, number of quality anomalies, energy consumption), allowing subsequent batch orders to directly call the optimized formula, thereby continuously improving changeover efficiency, product quality consistency, and energy economy in mixed-flow production.

[0085] (4) By extracting order features and recording mixed-flow changeover trigger markers, trigger markers for changing fixtures, test thresholds, or appearance parts are automatically generated when adjacent batches change in tabletop shape, power, or color, and these markers are written into the process feature vector. This solves the problems of easy errors and large cycle time losses when manually switching parameters in multi-variety mixed-flow production, realizes adaptive pre-configuration of station parameters, and shortens changeover time.

[0086] (5) By using the execution confirmation flag in the heating control quantity and the abnormal status field in the traceability dataset, combined with the writing to the time-series database, the entire process from aging control command to execution result is traceable. This solves the problem of separating aging process control records from product quality records in the existing technology, and provides complete data link support for abnormal batch review and complaint follow-up. Attached Figure Description

[0087] Figure 1 A flowchart illustrating a production process control method for an IoT-based electric heating table, provided as an embodiment of this application;

[0088] Figure 2 This is a structural block diagram of an IoT-based electric heating table production process control system provided in an embodiment of this application. Detailed Implementation

[0089] To enable those skilled in the art to better understand the technical solution, the present invention will be described in detail below with reference to embodiments. The description in this part is only exemplary and explanatory, and should not be used to limit the scope of protection of the present invention in any way.

[0090] Example 1: Refer to Figure 1 This is a flowchart illustrating a method for controlling the production process of an electric heating table based on the Internet of Things, according to an embodiment of the present invention. The process may include at least steps S100-S700:

[0091] S100: Obtain ERP orders, perform feature extraction and vector generation processing to obtain process feature vectors;

[0092] S200. Based on the process feature vector, perform hierarchical matching and workstation number reorganization to obtain the equipment parameter formula;

[0093] S300. Based on the device parameter formula, perform interface access and message conversion processing to obtain a unified operating condition data stream;

[0094] S400. Based on the unified working condition data stream, perform dual-core anomaly judgment and three-level linkage processing of line stop and material locking alarm to obtain a linkage status set.

[0095] S500. Based on the linkage state set and aging temperature curve parameters, perform thermodynamic load estimation and feedforward proportional-integral dynamic duty cycle adjustment to obtain the heating control quantity.

[0096] S600. Based on the heating control quantity and the barcode scanning record, perform primary key registration and batch association processing to obtain a traceability dataset;

[0097] S700. Based on the traceability dataset, a loss function is constructed and iteratively updated to obtain an updated formula.

[0098] S100: Obtain ERP orders, perform feature extraction and vector generation processing to obtain process feature vectors;

[0099] The ERP order specifically includes an order number, model number, planned quantity, delivery batch, power specification field, and color code field. The power specification field is used to perform cross-validation with the model number to generate a rated power identifier. The color code field is used to perform corresponding matching with the color coding table in the material master data to generate a color identifier. The model number serves as a mapping index for subsequent recipe calling and processing to retrieve the corresponding workstation parameters.

[0100] Specifically, S100 is executed by the order parsing module, which is deployed in the cloud service layer and maintains bidirectional data communication with the Enterprise Resource Planning (ERP) system. The order parsing module initiates an order capture process when a new order, inserted order, canceled order, modified order, or urgent order is marked. It receives the order number, model number, planned quantity, delivery batch, power specification, and color code fields from the ERP order. Subsequently, it performs field integrity verification, version number verification, and timestamp sorting on the ERP order, eliminating duplicate messages and retaining the currently valid order version. When the same order number corresponds to multiple revision records, the order parsing module selects the current record based on the latest timestamp and transfers the old version to the revision record area, thereby forming the basic order dataset. This basic order dataset constitutes the direct input source for S100. Considering the technical scenarios of electric heating tables, including square and round tabletops and multi-power mixed-current switching, the order parsing module does not directly output production scheduling results. Instead, it first organizes the basic order dataset into feature extraction objects for the control link, which are then used for subsequent feature recognition and field mapping.

[0101] Furthermore, the extraction of the desktop shape is completed by the first workstation image acquisition unit on the production line. Specifically, the first workstation image is a top-down view image of the desktop captured by the industrial camera in the first workstation image acquisition unit when the first electric heating table desktop arrives at the first workstation on the production line. After the order parsing module performs contour segmentation, edge closure, and template matching processing on this image, the outer contour of the desktop is matched with a preset shape template library to identify the corresponding shape, thereby outputting a desktop shape identifier. The desktop shape identifier includes a square desktop identifier or a circular desktop identifier.

[0102] Specifically, the image acquisition unit at the first workstation of the production line includes an industrial camera, a supplementary lighting component, and a position trigger switch. When the first electric heating table arrives at the first workstation, the position trigger switch sends an acquisition trigger signal, the industrial camera acquires a top-down view image of the table, and the order parsing module receives the top-down view image and performs contour segmentation, edge closure, and template matching processing. It then identifies the outer contour of the table against a preset shape template library and outputs a table shape identifier. The power extraction is performed by the order parsing module through cross-validation of the power specification field and model number in the ERP order. Specifically, the power specification field is read first, and then the model mapping table is called to confirm the rated power level, generating a rated power identifier. The color extraction is performed by the order parsing module reading the color code field in the ERP order and matching it against the color encoding table in the material master data, generating a color identifier. Understandably, the minimum set of parameters constituting the process feature vector consists of the table shape identifier, rated power identifier, and color identifier. The model number is stored as a mapping index in the same record for subsequent recipe call processing to retrieve the corresponding workstation parameters. If the desktop shape identifier is inconsistent with the preset shape of the model number, or the rated power identifier is inconsistent with the preset power of the model number, the order parsing module will write the current order into the abnormal record area and trigger a manual review mark, and the current order will be suspended from entering the subsequent main link.

[0103] Furthermore, after extracting the three core features of desktop shape, power, and color, the order parsing module performs field encapsulation and vector generation processing on the order basic dataset. Specifically, it writes the order number, model number, desktop shape identifier, rated power identifier, color identifier, order version number, and generation timestamp in a fixed field order to generate a process feature vector record. The process feature vector record is registered in the database as a "process feature vector field set", where the desktop shape identifier represents the fixture switching direction, the rated power identifier represents the control board test threshold and aging gear direction, the color identifier represents the switching direction of appearance parts and packaging parts, the model number represents the mapping index relationship, the order version number represents the version to which the current record belongs, and the generation timestamp represents the timing position of the current record entering the main link. To adapt to mixed-flow changeover scenarios, the order parsing module also performs difference comparison processing on the process feature vector field sets of adjacent orders. When the desktop shape identifier changes, it records the fixture change trigger mark; when the rated power identifier changes, it records the test threshold change trigger mark; when the color identifier changes, it records the appearance component change trigger mark. The trigger marks are written into the process feature vector field set, thereby enabling the S100 output product to have both feature description function and subsequent scheduling index function.

[0104] In one engineering embodiment, after the cloud service layer receives an order for a batch of electric heating tables, the order parsing module first reads the model number, power specification field, and color code field from the ERP order. Then, the first-station image acquisition unit on the production line acquires a top-down view image of the tabletop when the first piece arrives, identifies the tabletop shape identifier corresponding to the circular tabletop, and confirms the rated power identifier from the power specification field and the color identifier from the color code field. When the tabletop shape identifier, the rated power identifier, and the color identifier all pass the corresponding verification with the model mapping table, the order parsing module generates a process feature vector field set containing the order number, model number, tabletop shape identifier, rated power identifier, color identifier, order version number, generation timestamp, and trigger marker. The process feature vector field set is written into the order parsing cache area for direct access in S200 by "inputting the process feature vector into the recipe for processing". If the current batch of orders is subsequently revised, the order parsing module re-fetches the revised ERP order, generates a new order version number and overwrites the valid identifier of the old version in the main link, while retaining the old version record for audit tracking. In other words, the output product of S100 is a process feature vector. The output field names include order number, model number, desktop shape identifier, rated power identifier, color identifier, order version number, generation timestamp, and trigger flag. The process feature vector enters the recipe call processing of S200 in the next main step.

[0105] The technical effect of this step can be summarized as follows: S100 converts the ERP order into a process feature vector for the production control chain, forming a unified entry point between the front-end order information and the back-end workstation parameters. The desktop shape identifier is generated by the image acquisition unit of the first workstation on the production line and enters the same field set along with the rated power identifier and the color identifier, making the feature recognition chain clearer in mixed-flow changeover scenarios. The process feature vector carries the order version number, generation timestamp, and trigger marker, providing a continuous connection and audit record basis for subsequent input received by S200.

[0106] S200. Based on the process feature vector, perform hierarchical matching and workstation number reorganization to obtain the equipment parameter formula;

[0107] Specifically, S200 is executed by the recipe calling module, which is located in the cloud service layer and connected to the order parsing module, recipe storage unit, workstation mapping unit, and version management unit. The input source for the recipe calling process is the process feature vector output by S100. The process feature vector includes at least the order number, model number, desktop shape identifier, rated power identifier, color identifier, order version number, generation timestamp, and trigger flag. The recipe storage unit is a parameter library organized by workstation number, which records solder paste thickness parameters, screw torque parameters, test upper and lower limit parameters, and aging temperature curve parameters corresponding to different model numbers. The workstation mapping unit is a rule table that establishes a correspondence between model numbers and parameter templates of each production workstation. The version management unit is a management unit that registers, freezes, replaces, and backtracks different order version numbers, recipe version numbers, and effective status flags. After receiving the process feature vector, the formula retrieval module first performs order version number verification and model number verification, then reads the desktop shape identifier, rated power identifier, and color identifier to form the formula retrieval conditions for the current order. These conditions are then sent to the formula storage unit to query the corresponding parameter template. When the current order is in a scenario of insertion, cancellation, or modification, the version management unit synchronously records the current retrieval batch and replacement batch. When the current order is in a regular production scheduling scenario, the version management unit records the current retrieval batch and the effective start time. This step corresponds to the cloud-based Advanced Planning and Scheduling (APS) receiving the ERP order and organizing the subsequent formula retrieval technical chain around the three-dimensional features of desktop shape, power, and color. The formula object covers key parameters of the patch, assembly, testing, and aging processes.

[0108] Furthermore, the recipe calling module performs layered matching processing on the process feature vector. Tooling layer matching is performed on the desktop shape identifier, calling the fixture parameter segment and assembly torque parameter segment corresponding to the square or round desktop; electrical layer matching is performed on the rated power identifier, calling the test upper and lower limit parameter segment and aging temperature curve parameter segment corresponding to the power level; and appearance layer matching is performed on the color identifier, calling the appearance component switching mark and packaging switching mark corresponding to the color code. Understandably, the minimum set of core parameters constituting the equipment parameter recipe consists of solder paste thickness parameters, screw torque parameters, test upper and lower limit parameters, and aging temperature curve parameters. The solder paste thickness parameter corresponds to the printing thickness setting of the SMT placement station, the screw torque parameter corresponds to the torque setting of the automatic screwdriver or servo screwdriver, the test upper and lower limit parameters correspond to the judgment thresholds of the ICT testing station and the functional testing station, and the aging temperature curve parameter corresponds to the time-segmented temperature setting of the aging chamber. Color-related fields, fixture switching marks, and packaging switching marks are supplementary fields of the recipe and are written together when the current order involves appearance component switching or packaging component switching. If a complete parameter template directly corresponding to the model number exists in the recipe storage unit, the recipe calling module directly reads the complete parameter template. If the complete parameter template corresponding to the model number has not yet been registered, the workstation mapping unit segments and combines them according to the desktop shape identifier, rated power identifier, and color identifier to generate a temporary parameter template. The temporary parameter template is then submitted to the version management unit and registered as pending review. Temporary parameter templates in the pending review state do not enter the distribution chain but remain in the manual review area. After this processing, the recipe calling process does not simply call a single model number segment, but instead splits the process feature vector into three categories of search conditions: shape, power, and color, and then synthesizes them into the workstation-level parameter set corresponding to the current order.

[0109] Furthermore, after obtaining the parameter template, the recipe calling module performs workstation number reorganization and version encapsulation processing on the parameter template. Specifically, the workstation mapping unit first generates workstation numbers according to the raw material kitting workstation, SMT placement workstation, ICT testing workstation, functional testing workstation, aging workstation, assembly workstation, and packaging workstation, and then binds each parameter segment in the parameter template to the corresponding workstation number to form a workstation-level recipe record; the version management unit then generates a recipe version number, an effective status flag, a call source flag, and a distribution order flag for the current workstation-level recipe record, where the recipe version number corresponds to the effective version of the current search batch, the effective status flag is used to distinguish between currently effective records and historically frozen records, the call source flag is used to indicate whether the current record comes from direct retrieval or segment combination, and the distribution order flag is used to constrain the order in which each workstation receives parameters. If the trigger marker in the process feature vector indicates the presence of a fixture change trigger marker, the recipe calling module will prioritize the assembly station and fixture preparation station in the order of issuance. If the trigger marker indicates the presence of a test threshold change trigger marker, the functional test station and aging station will be prioritized in the order of issuance. If the trigger marker indicates the presence of an appearance part change trigger marker, the packaging station and appearance part material preparation station will be prioritized in the order of issuance. Thus, the equipment parameter recipe is not only a set of parameters but also includes station numbers, effective status markers, and issuance order markers. In version replacement scenarios, the recipe calling module first freezes the old recipe version number, then writes the new recipe version number, and records the replacement action in the audit log of the version management unit for subsequent traceability, aggregation, and processing.

[0110] In one engineering embodiment, after the order parsing module outputs the process feature vector of a certain batch of electric heating tables, the recipe calling module reads the model number, tabletop shape identifier, rated power identifier, and color identifier. First, it retrieves the fixture parameter segment and assembly torque parameter segment corresponding to the round tabletop from the recipe storage unit, and then retrieves the test upper and lower limit parameter segments and aging temperature curve parameter segments corresponding to the power level. After that, it adds appearance component switching marks according to the color identifier to form the workstation-level parameter set of the current order. When the batch belongs to the inserted batch, the version management unit freezes the original recipe version number and generates a new recipe version number and effective status mark for the current workstation-level parameter set. Then, it is handed over to the workstation mapping unit to sort by workstation number and output the equipment parameter recipe of the current order. At this time, the output field names of the equipment parameter formula include the workstation number, solder paste thickness parameter, screw torque parameter, test upper and lower limit parameter, aging temperature curve parameter, formula version number, effective status flag, call source flag and distribution sequence flag. The output field names in "Input the equipment parameter formula into the workstation access processing" in S300 are directly read and used as the parameter benchmark before the real-time working condition data of the workstation is collected and converted.

[0111] The technical effect of this step can be summarized as follows: S200 transforms the process feature vector into equipment parameter recipes organized by workstation number, ensuring seamless integration between order-side features and workstation-side parameters within the same workflow. The equipment parameter recipes include core parameters for placement, assembly, testing, and aging processes, and are inscribed with recipe version numbers, effective status markers, and distribution sequence markers, providing clear parameter basis for subsequent input received by S300. The recipe call processing also incorporates version replacement records for order insertion, cancellation, and modification scenarios, making the entire recipe flow chain more complete.

[0112] S300. Based on the device parameter formula, perform interface access and message conversion processing to obtain a unified operating condition data stream;

[0113] Specifically, step S300 is executed by the workstation access module, which is deployed at the edge control layer and connected to the recipe calling module, field device interfaces, and edge gateway. It receives the device parameter recipe output by step S200 as input for this step. The device parameter recipe is a set of parameters organized by workstation number, including workstation number, solder paste thickness parameter, screw torque parameter, test upper and lower limit parameters, aging temperature curve parameter, recipe version number, effective status flag, calling source flag, and distribution sequence flag. The workstation access module consists of an interface access unit, a message conversion unit, a transparent serial bridge, an encrypted transmission unit, and a watchdog reset unit. The interface access unit is responsible for accessing the field interfaces of SMT placement equipment, automatic screw-driving equipment, testing equipment, and the aging chamber programmable logic controller (PLC). The message conversion unit is responsible for organizing different interface formats into a unified message structure. The transparent serial bridge is responsible for serial port forwarding. The encrypted transmission unit is responsible for link encryption of the unified message. The watchdog reset unit is responsible for resetting and reconnecting when the acquisition link is interrupted. The workstation access processing is initiated in scenarios such as new batch launch, order insertion switching, recipe version switching, and equipment reconnection. First, the corresponding parameter segment is read according to the workstation number, and then the access sequence table for the current batch is established according to the issuance order, for subsequent workstation collection actions to call.

[0114] Furthermore, after receiving the equipment parameter formula, the interface access unit first performs a corresponding registration of the workstation number and equipment number, and includes the SMT placement equipment, automatic screw-driving equipment, testing equipment, and aging chamber PLC into the current acquisition queue; then, it distinguishes RS-232 interface data and RS-485 interface data according to the interface type, and loads the corresponding parameter tags according to the workstation number. For the SMT placement equipment, it loads the printing parameter tag and equipment status tag corresponding to the solder paste thickness parameter; for the automatic screw-driving equipment, it loads the torque value tag, gun head number tag, and tightening completion tag corresponding to the screw torque parameter; for the testing equipment, it loads the reset pulse width data tag, electrical parameter tag, and test conclusion tag corresponding to the test upper and lower limit parameters; for the aging chamber PLC, it loads the aging temperature curve parameter corresponding to the aging chamber internal temperature tag, aging chamber external temperature tag, online unit count tag, and heating status tag. After receiving the aforementioned acquisition tag, the message conversion unit splits the serial port acquisition frame into a device number field, a workstation number field, a parameter identifier field, a parameter value field, an acquisition timestamp field, and a recipe version number field. It then converts the serial port acquisition frame into a Message Queuing Telemetry Transport (MQTT) message. The MQTT message is sent via the transparent serial bridge, with the sending link carrying Transport Layer Security (TLS) 1.3 encryption configuration and watchdog reset configuration. If the current device does not return a complete acquisition frame within the polling cycle, the interface access unit registers the disconnection status, retains the previous round's valid acquisition frame, and writes an exception record to the edge gateway. If the current device returns an acquisition frame with missing fields, the message conversion unit writes a missing field flag and stops the current frame from being enqueued.

[0115] Furthermore, after the message conversion unit completes the MQTT message generation, it performs unified field processing and time sequence processing on the current message to form the unified operating condition data stream. The unified operating condition data stream is a collection of workstation operation records written continuously in time sequence. The minimum set of core fields includes the device number field, workstation number field, parameter identifier field, parameter value field, collection timestamp field, recipe version number field, interface type field, message subject field, and data status field. Among them, the device number field represents the device from which the current data comes, the workstation number field represents the workstation to which the current data belongs, the parameter identifier field represents whether the current parameter belongs to solder paste thickness related data, screw torque related data, test related data, or aging related data, the parameter value field represents the current collection result, the collection timestamp field represents the time position when the current collection result enters the main link, the recipe version number field represents the parameter benchmark corresponding to the current collection result, the interface type field represents the RS-232 interface or RS-485 interface, the message subject field represents the edge gateway subscription path, and the data status field represents the normal status, disconnected status, or missing field status. After generating the unified working condition data stream, the workstation access module writes the unified working condition data stream of the current batch into the edge gateway cache, and sorts it by workstation number and collection timestamp for use in "inputting the unified working condition data stream into quality linkage processing" in S400; in other words, the output product of S300 is the unified working condition data stream, and the output fields are named as device number field, workstation number field, parameter identifier field, parameter value field, collection timestamp field, recipe version number field, interface type field, message subject field, and data status field.

[0116] In one engineering embodiment, after the recipe calling module outputs the equipment parameter recipe for a batch of electric heating tables with a specific round tabletop, rated power, and color, the workstation access module first establishes a collection queue for the current batch according to the workstation number, and then connects the SMT placement equipment, automatic screw-driving equipment, functional testing workstation, and aging chamber PLC to the edge gateway. Subsequently, the interface access unit reads the torque value from the automatic screw-driving equipment, the reset pulse width data from the testing equipment, and the number of online units and temperature status from the aging chamber PLC. The message conversion unit sequentially splits and converts the data into MQTT messages. After the transparent serial bridge completes serial port forwarding, the MQTT messages enter the edge gateway buffer and are assembled into the unified operating condition data stream in chronological order. If a recipe version switch occurs midway through the current batch, the workstation access module stops the subsequent collection frames corresponding to the old recipe version number from being queued, reloads the new recipe version number into the collection tags of each workstation, and then continues to collect and generate a new unified operating condition data stream. This ensures that when the S400 receives input, the unified operating condition data stream maintains a correspondence with the currently effective equipment parameter recipe.

[0117] The technical effects of this step can be summarized as follows: S300 establishes a direct association between the equipment parameter recipe and the field interfaces of each workstation, with parameter benchmarking, interface access, and message conversion completed in the same link. The unified operating condition data stream carries the workstation number, parameter identifier, and recipe version number, providing S400 with a clear basis for comparison when receiving input. After the transparent serial bridge, TLS1.3 encryption configuration, and watchdog reset configuration are included in the main link of this step, the field acquisition link record is more complete.

[0118] S400. Based on the unified working condition data stream, perform dual-core anomaly judgment and three-level linkage processing of line stop and material locking alarm to obtain a linkage status set.

[0119] Specifically, S400 is executed by a quality linkage module, which is located at the edge control layer and connected to the workstation access module, recipe call module, safety relay, feeder, and alarm light. The quality linkage module includes a dual-core A7 processor, an MCU coprocessor, a recipe comparison unit, a message proxy unit, an anomaly determination unit, a line stop control unit, a material locking control unit, and a status write-back unit. The dual-core A7 processor is responsible for receiving the unified operating condition data stream, performing message proxying and recipe comparison, while the MCU coprocessor is responsible for receiving anomaly determination results, performing line stop material locking control, and driving the safety relay. The input sources for S400 are the unified operating condition data stream generated by S300 and the device parameter recipe generated by S200 and in an effective state. The unified operating condition data stream includes a device number field, a workstation number field, a parameter identifier field, a parameter value field, a collection timestamp field, a recipe version number field, an interface type field, a message subject field, and a data status field. The device parameter recipe includes test upper and lower limit parameters, screw torque parameters, solder paste thickness parameters, aging temperature curve parameters, workstation number, effective status flag, and recipe version number. The quality linkage processing is initiated when the test device's acquisition frame arrives at the edge gateway buffer, and a quality comparison context for the current batch is established according to the workstation number and recipe version number, so that data frames in the same batch and the current effective parameter benchmark enter the same judgment link.

[0120] Furthermore, the message proxy unit first filters out the data frames corresponding to the test station, the data frames corresponding to the assembly station, and the data frames whose data status field is in a normal state from the unified working condition data stream, and then organizes them according to the execution order of the collection timestamp field to form a message queue to be judged; the formula comparison unit reads the station number field and the formula version number field in the message queue to be judged, retrieves the parameter segments of the same station and version from the equipment parameter formula, and matches the parameter identification field with the test upper and lower limit parameters, screw torque parameters, or solder paste thickness parameters one by one. The anomaly determination unit is a determination unit that performs threshold comparison, range comparison, and state consistency comparison. Threshold comparison is for the test result field and torque result field, range comparison is for the reset pulse width data, torque value, and thickness value, and consistency comparison is for the correspondence between the workstation number, equipment number, and recipe version number. When there is a workstation number and recipe version number that are inconsistent in the current message queue to be determined, the parameter identifier field cannot be mapped to the current parameter segment, or the data status field is in a disconnected state or missing field state, the anomaly determination unit writes the current data frame into the anomaly buffer and registers the current batch as a pending linkage state. The reset pulse width data is output by the Automatic Test Equipment (ATE). The dual-core A7 processor extracts the reset pulse width data from the unified operating condition data stream and performs a comparison according to the reset pulse width range in the test upper and lower limit parameters. When the reset pulse width data falls outside the range of 160ms to 220ms, the anomaly determination unit generates a reset anomaly flag and sends it to the MCU coprocessor. When the actual torque value corresponding to the screw torque parameter exceeds the allowable range of the current workstation, the anomaly determination unit generates a torque anomaly flag. When the test equipment returns a failure status and the corresponding parameter identifier field points to a functional test item, the anomaly determination unit generates a test anomaly flag. The reset anomaly flag, torque anomaly flag, and test anomaly flag are merged within the current batch to form a linked trigger condition.

[0121] Furthermore, after receiving the linkage trigger condition, the MCU coprocessor activates the stop control unit, the material locking control unit, and the alarm drive unit. The stop control unit outputs a stop signal through a safety relay. This stop signal acts on the operating loop of the production line where the current workstation is located, causing the testing workstation, assembly workstation, and upstream conveying unit to stop releasing materials. The material locking control unit writes a material locking control frame to the feeder, which triggers the feeder to stop the next package of control boards or components from entering the current batch. The alarm drive unit writes a yellow light drive signal to the alarm light, separating abnormal batches from normal batches on the field display. Before executing an action, the MCU coprocessor reads the workstation number, equipment number, and data acquisition timestamp fields of the current batch, and then generates an action sequence record. The action sequence is: line stop control, yellow light drive, feeder locking, and status write-back. After the current safety relay returns an execution confirmation signal, the line stop control unit writes a line stop completion flag. After the current feeder returns a locking confirmation signal, the locking control unit writes a locking completion flag. If any confirmation signal does not return within a preset waiting period, the status write-back unit writes an action unconfirmed flag and maintains the current line stop state. Understandably, the core fields constituting the minimum action set for line stop and locking control are the line stop signal, yellow light drive signal, and feeder locking signal. The safety relay drive record, action sequence record, and action unconfirmed flag belong to the same link record fields and follow the current batch into subsequent steps. During the execution of the MCU coprocessor's actions, the dual-core A7 processor continues to receive subsequent data frames, but writes a freeze flag to newly entering data frames of the current batch, preventing them from entering the normal release link.

[0122] In one engineering embodiment, after the current batch of electric heating tables completes power-on testing at the functional testing station, the testing equipment uploads reset pulse width data via the unified operating condition data stream. The dual-core A7 processor extracts the station number field, parameter identifier field, parameter value field, acquisition timestamp field, and formula version number field from the current data frame, and then retrieves the upper and lower limit parameters of the same version from the equipment parameter formula. If the current reset pulse width data is an abnormal value exceeding the range of 160ms to 220ms, the abnormality judgment unit generates a reset abnormality mark. The MCU coprocessor then drives the safety relay to output a stop signal and writes a yellow light drive signal to the yellow light, while simultaneously writing a feeder locking signal to the feeder. The status write-back unit then encapsulates the equipment number field, station number field, abnormality type field, stop status field, yellow light status field, locking status field, acquisition timestamp field, and formula version number field corresponding to the current action to generate the linkage status set, and sends the linkage status set to the "input the linkage status set into aging linkage processing" call in S500. In other words, the output field names of the linkage status set include the equipment number field, workstation number field, abnormality type field, production stop status field, yellow light status field, material lock status field, data collection timestamp field, and recipe version number field. The abnormality type field indicates a reset abnormality, torque abnormality, or test abnormality. The production stop status field indicates whether the current production line is in a production stop state. The yellow light status field indicates the current alarm light status. The material lock status field indicates the current feeder status. The linkage status set is used as one of the inputs for aging linkage processing in subsequent steps.

[0123] The technical effect of this step can be summarized as follows: S400 performs batch-wise comparison between the unified operating condition data stream and the equipment parameter formula at the edge control layer, and anomaly detection, line stop control, and material locking control enter the same linkage link. The dual-core A7 processor and the MCU coprocessor have clearly defined roles, and the reset pulse width data, line stop signal, yellow light drive signal, and feeder material locking signal form a closed-loop record within the same batch. The linkage status set includes anomaly type field and action status field, providing a clear status basis for subsequent input reception by S500.

[0124] S500. Based on the linkage state set and aging temperature curve parameters, perform thermodynamic load estimation and feedforward proportional-integral dynamic duty cycle adjustment to obtain the heating control quantity.

[0125] Specifically, S500 is executed by the aging linkage module, which is located at the edge control layer and connected to the quality linkage module, the aging chamber programmable logic controller, the temperature acquisition unit, and the heating tube drive unit. The input source of S500 is the linkage status set generated by S400. The linkage status set includes the equipment number field, workstation number field, abnormality type field, line stop status field, yellow light status field, material lock status field, acquisition timestamp field, and formula version number field. After receiving the linkage status set, the aging linkage module first reads the line stop status field, the abnormality type field, and the formula version number field, and then calls the aging temperature curve parameters corresponding to the current batch in S200 to establish the control context of the current batch aging workstation. The online number of units in the aging linkage process refers to the number of electric heating tables currently in the aging chamber and within the control range of the current formula version number. The temperature status refers to the temperature acquisition result composed of the temperature inside and outside the aging chamber. The PWM duty cycle of the heating tube is Pulse Width Modulation, used to characterize the on / off ratio of the heating tube within the control cycle. The heating control quantity is a set of aging execution instructions calculated by combining the online number of units in the current batch, the temperature status, the aging temperature curve parameters, and the linkage status. The aging linkage module starts the recalculation process when a new batch enters the aging stage, the formula version is switched, the line stop status changes, or the online number of units changes, so that the current aging stage continuously receives the previous quality linkage results within the same control link and synchronously adjusts the heat treatment status of the current batch.

[0126] Furthermore, the aging linkage module includes a status filtering unit, a unit count identification unit, a temperature processing unit, a thermodynamic model unit, a feedforward control unit, a proportional-integral control unit, and a control encapsulation unit. The status filtering unit first performs batch filtering and station filtering on the linkage status set, extracting the station number field corresponding to the aging station and its associated stop status field. When the stop status field indicates that the current batch is in a stop status, the status filtering unit writes the current batch into the aging freeze zone and sends the freeze mark to the subsequent control link. When the stop status field indicates that the current batch is in a release status, the status filtering unit sends the current batch to the unit count identification unit. The unit count identification unit collects the transmission cycle signal, entry count signal, and exit count signal from the aging chamber's programmable logic controller, and combines this with the product serial number mapping relationship of the current batch to count and process the electric heating tables in the aging chamber, generating an online unit number segment. The temperature processing unit reads the temperature inside the aging chamber from the temperature acquisition point inside the aging chamber and the temperature outside the aging chamber from the temperature acquisition point outside the aging chamber, and performs same-period merging according to the acquisition timestamp to generate a temperature status field. Understandably, the minimum set of fields constituting the core input of the aging linkage processing is the stop status field, the recipe version number field, the online station number field, the aging chamber internal temperature field, and the aging chamber external temperature field. The stop status field is used to determine whether the current batch is allowed to continue heating, the recipe version number field is used to retrieve the current aging temperature curve parameters, the online station number field is used to characterize the current heat load level, and the aging chamber internal temperature field and the aging chamber external temperature field are used to characterize the current temperature status. The transmission cycle signal, the chamber entry count signal, and the chamber exit count signal are auxiliary fields for unit number identification and are recorded along with the current batch.

[0127] Further, the thermodynamic model unit receives the online station digital segment, the aging chamber internal temperature field, the aging chamber external temperature field, and the aging temperature curve parameters, and performs heat load estimation according to the aging period of the current batch. The aging temperature curve parameters are a set of curve parameters called from the equipment parameter recipe by S200 and sent to the aging station, including the target temperature range and corresponding time period of the current batch in the preheating, isothermal, and cooling stages. The thermodynamic model unit first identifies the current heat capacity range based on the online station digital segment, then identifies the current heat dissipation state based on the temperature difference between the aging chamber internal and external temperatures, and then generates a basic duty cycle value based on the target temperature range to which the current aging period belongs. The feedforward control unit receives the basic duty cycle value and the temperature state change result of the current batch, and performs pilot correction on the heating tube drive quantity; the proportional-integral control unit receives the correction result of the feedforward control unit and the current aging chamber internal temperature field, performs continuous tracking correction on the heating tube drive quantity, and generates the heating tube PWM duty cycle. When the current batch is in the release state and the online unit number segment changes, the aging linkage module immediately triggers recalculation; when the current batch is in the frozen state, the feedforward control unit switches the base duty cycle value to the frozen duty cycle value, and the proportional-integral control unit maintains the temperature tracking link for the current cycle and stops amplifying the calculation for newly added units. Therefore, the aging linkage processing is not driven by a fixed temperature, but rather sends the status information from the linkage status set, the aging temperature curve parameters in the equipment parameter formula, and the online unit count and temperature status in the aging chamber to the same control link to complete the duty cycle calculation.

[0128] Furthermore, after obtaining the PWM duty cycle of the heating tube, the control encapsulation unit encapsulates the aging execution information within the current control cycle into the heating control quantity. The heating control quantity includes a workstation number field, a recipe version number field, an online workbench digital segment, an aging chamber indoor temperature field, an aging chamber outdoor temperature field, a heating tube PWM duty cycle field, a heating stage field, a freeze state field, and a control timestamp field. The workstation number field represents the aging workstation to which the current control action belongs; the recipe version number field represents the recipe version corresponding to the current heating control quantity; the online workbench digital segment represents the current heat load quantity; the aging chamber indoor temperature field and the aging chamber outdoor temperature field represent the current temperature status; the heating tube PWM duty cycle field represents the on / off ratio of the heating tube within the current control cycle; the heating stage field represents whether the current batch is in the preheating stage, constant temperature stage, or cooling stage; the freeze state field represents whether the current batch is in a frozen state or a released state; and the control timestamp field represents the time position at which the current heating control quantity enters the main link. The control encapsulation unit writes the heating control quantity into the edge gateway cache and simultaneously writes it into the execution register of the aging chamber programmable logic controller. In subsequent steps, the heating control quantity enters the "input traceability and aggregation processing of the heating control quantity" in S600, serving as one of the direct sources of product serial number-associated operating condition data and process version. When the current aging chamber programmable logic controller returns to the execution confirmation state, the control encapsulation unit adds an execution confirmation flag to the heating control quantity; when the current execution confirmation state is missing, the control encapsulation unit adds an unconfirmed flag and retains the control record for the current cycle.

[0129] In one engineering embodiment, after a batch of electric heating tables is processed by S400, the quality linkage module outputs the linkage state set containing the release status and sends it to the aging linkage module. The aging linkage module first reads the formula version number field and the stop status field from the linkage state set, and then collects the entry count signal and exit count signal corresponding to the current batch from the aging chamber programmable logic controller, and sorts them to obtain the online digital segment. Subsequently, the temperature sorting unit reads the temperature field inside the aging chamber and the temperature field outside the aging chamber, and the thermodynamic model unit generates the basic duty cycle value by combining the aging temperature curve parameters in the equipment parameter formula. The feedforward control unit and the proportional-integral control unit continue to correct the current control cycle to obtain the current heating tube PWM duty cycle field. When a subsequent abnormal test batch causes a stop status change and enters the linkage state set, the state filtering unit writes the current batch into the aging freeze zone, the control encapsulation unit synchronously rewrites the freeze status field and the heating tube PWM duty cycle field, and then generates a new heating control quantity and sends it to S600 for calling. In other words, the output product of S500 is the heating control quantity. The output field names include the workstation number field, the recipe version number field, the online station number field, the aging chamber temperature field, the aging chamber outside temperature field, the heating tube PWM duty cycle field, the heating stage field, the freezing status field, and the control timestamp field. The heating control quantity is used as the input field set for traceability and aggregation processing in subsequent steps.

[0130] The technical effect of this step can be summarized as follows: S500 connects the linkage state set with the on-site state of the aging process within the same link, and the results of the preceding quality linkage directly enter the aging control process. The online unit count, temperature status, and aging temperature curve parameters jointly participate in the calculation of the heating tube PWM duty cycle, making the control basis of the aging process more complete. The heating control quantity includes a freeze status field and a recipe version number field, providing a clear process record basis for S600 to receive inputs subsequently.

[0131] S600. Based on the heating control quantity and the barcode scanning record, perform primary key registration and batch association processing to obtain a traceability dataset;

[0132] Specifically, S600 is executed by the traceability and aggregation module, which is located in the cloud service layer and connected to the workstation access module, quality linkage module, aging linkage module, time series database, and barcode acquisition terminal. The input source of S600 is the heating control quantity generated by S500, which includes the workstation number field, recipe version number field, online station digital segment, aging chamber indoor temperature field, aging chamber outdoor temperature field, heating tube PWM duty cycle field, heating stage field, freezing status field, and control timestamp field. The product serial number is a unique identifier for each electric heating table, generated at the online scanning station, and is used throughout the entire process of patching, testing, aging, assembly, and packaging. Operating condition data refers to the running records corresponding to the current product serial number in the unified operating condition data stream, the linkage status set, and the heating control quantity. Personnel identification is the identity record of the current station operator generated in the login terminal, card swiping terminal, or shift schedule. Equipment identification is the equipment number in the on-site equipment ledger. Material batches are the scanned batch records at the raw material assembly station, control plate loading station, and appearance component loading station. Process version is the version record corresponding to the current batch during formula call processing and on-site distribution. Environmental data consists of temperature and humidity records generated by temperature and humidity acquisition nodes, temperature acquisition points inside the aging chamber, and temperature acquisition points outside the aging chamber. After receiving the heating control quantity, the traceability and aggregation module first reads the station number field, formula version number field, control timestamp field, and freeze status field, and then enters the product serial number association link.

[0133] Furthermore, the traceability and aggregation module includes a primary key registration unit, a data retrieval unit, a batch association unit, a timing writing unit, and a traceability encapsulation unit. The primary key registration unit first reads the product serial number of the current batch from the online scanning record area, the aging chamber entry record area, and the exit record area, and establishes a correspondence between the product serial number and the current aging batch according to the workstation number field and the control timestamp field, forming a primary key registration table. The data retrieval unit then, based on the primary key registration table, extracts the equipment number field, workstation number field, parameter identifier field, parameter value field, acquisition timestamp field, and formula version number field from the unified operating condition data stream from the workstation access module cache area; extracts the abnormal type field, line stop status field, yellow light status field, material lock status field, and acquisition timestamp field from the linkage status set from the quality linkage module cache area; and extracts the online station digital segment, aging chamber internal temperature field, aging chamber external temperature field, heating tube PWM duty cycle field, heating stage field, freezing status field, and control timestamp field from the heating control quantity from the aging linkage module cache area, thereby forming the operating condition data set corresponding to the current product serial number. Understandably, the core minimum set of fields in the operating condition data set consists of the product serial number field, workstation number field, parameter identifier field, parameter value field, anomaly type field, line stop status field, material locking status field, heating tube PWM duty cycle field, and control timestamp field. The remaining fields are archived along with the current batch for subsequent querying. When a scan record for the current product serial number is missing at any workstation, the primary key registration unit writes the current record to the missing number buffer and extracts adjacent records from the time-sequential records of the workstations before and after the same batch to complete the replacement registration. The replacement result, along with the replacement time, is written to the traceability encapsulation unit.

[0134] Furthermore, after obtaining the set of operating condition data, the batch association unit continues to retrieve personnel identifiers, equipment identifiers, material batches, process versions, and environmental data. Specifically, personnel identifiers are generated by combining workstation login records, shift scheduling records, and card terminal records, with the data retrieval order arranged according to the workstation numbers traversed by the current product serial number; equipment identifiers are generated by corresponding the equipment number in the equipment ledger with the workstation number field; material batches are generated by combining incoming material scanning records, control board loading records, screw batch records, and packaging barcode scanning records; process versions are generated by corresponding the formula version number field with the process release record table; and environmental data consists of temperature and humidity acquisition node records, aging chamber indoor temperature field, aging chamber outdoor temperature field, and temperature and humidity curves generated in chronological order. The batch association unit performs same-batch alignment and same-product serial number merging on the aforementioned records to obtain product-level associated records, and sends the product-level associated records to the traceability packaging unit. The traceability encapsulation unit encapsulates traceability records according to the product serial number field, forming a traceability dataset. The output field names of the traceability dataset include the product serial number field, operating condition data field, personnel identification field, equipment identification field, material batch field, process version field, environmental data field, temperature and humidity curve field, traceability timestamp field, and abnormal status field. The operating condition data field corresponds to the parameter records of the current product at each workstation; the personnel identification field corresponds to the operator records passing through the workstation; the equipment identification field corresponds to the equipment records passing through the workstation; the material batch field corresponds to the batch records of control boards, screws, appearance parts, and packaging parts; the process version field corresponds to the current formula version and the released process version; the environmental data field corresponds to the aging process and workshop environment records; the temperature and humidity curve field corresponds to the environmental change records unfolded in chronological order; and the abnormal status field corresponds to the combined records of linkage status and frozen status. The time-series writing unit then writes the traceability dataset into the time-series database according to the traceability timestamp field and records the write position back to the primary key registry table after the writing is completed.

[0135] In one engineering embodiment, after a batch of electric heating tables completes aging control, the aging linkage module outputs the heating control quantity and sends it to the traceability collection module. The primary key registration unit first reads the product serial number of the current batch from the aging room entry barcode scanning record, and then pulls the unified operating condition data stream, linkage status set, and heating control quantity within the same time period according to the control timestamp field, forming the operating condition data set corresponding to the current product serial number. Subsequently, the batch association unit reads the personnel identification of the current assembly station and testing station from the workstation login terminal, reads the test equipment number, screw-driving equipment number, and aging room equipment number from the equipment ledger, reads the control board batch and appearance part batch from the incoming material barcode scanning record, reads the process version from the process release record table, and then reads the temperature and humidity curve from the temperature and humidity acquisition node, finally forming the traceability dataset and writing it into the time series database. When subsequent abnormal batch reviews or complaint batch re-inspections occur, the traceability collection module reads the traceability dataset according to the product serial number field and returns the parameter trajectory and association records of the current product at each workstation. In other words, the traceability dataset is then entered into the "inputting the traceability dataset into the recipe update process" in S700 in the subsequent steps, serving as a direct input source for organizing changeover time, number of quality anomalies, and energy consumption data.

[0136] The technical effect of this step can be summarized as follows: S600 merges the heating control quantity with previous workstation records, personnel records, equipment records, material records, process records, and environmental records under the same product serial number. The traceability dataset includes operating condition data fields, process version fields, and temperature and humidity curve fields, and the previous operating status is continuously recorded within this step. After the traceability dataset enters S700, the formula update processing has a complete data source.

[0137] S700. Based on the traceability dataset, a loss function is constructed and iteratively updated to obtain the updated formula;

[0138] Specifically, S700 is executed by the formula update module, which is located in the cloud service layer and connected to the traceability collection module, formula call module, version management unit, and scheduling record area. The input source for S700 is the traceability dataset generated by S600. This traceability dataset includes fields for product serial number, operating condition data, personnel identification, equipment identification, material batch, process version, environmental data, temperature and humidity curve, traceability timestamp, and abnormal status. The changeover time in the formula update process refers to the time record from the last piece of the previous batch leaving the critical workstation to the first piece of the next batch entering a stable production state after changes in the tabletop shape, rated power, or color identification of two adjacent batches of electric heating tables. The number of quality anomalies refers to the cumulative number of times the same batch triggers anomaly judgment and enters the linkage state set at the testing station, assembly station, and aging station. The energy consumption data refers to the energy consumption record formed by the heating tube PWM duty cycle field, online digital segment, and power meter acquisition results corresponding to the current batch's operation at the aging station. The formula update module initiates the update process upon batch completion, type change completion, abnormal batch closed-loop archiving, or shift end. It first establishes the update task for the current batch, then reads the traceability dataset and historical formula version records, and enters the indicator processing link.

[0139] Specifically, the formula update module includes an indicator processing unit, a loss function construction unit, an update solution unit, a version review unit, and a release write-back unit. The indicator processing unit first extracts the operating condition data field, abnormal status field, process version field, and traceability timestamp field corresponding to the same product serial number from the traceability dataset, and then performs merging and processing according to batch number, workstation number, and formula version number to form a batch update sample. The changeover time in the batch update sample is calculated by the first workstation image acquisition record, fixture switching record, torque parameter switching record, and test threshold switching record. The number of quality anomalies is accumulated by the anomaly type field, line stop status field, material locking status field, and frozen status field. The energy consumption data is obtained by merging the heating tube PWM duty cycle field, online platform digital segment, aging chamber internal temperature field, aging chamber external temperature field, and power meter record. During the data processing, the indicator processing unit also simultaneously reads the personnel identification field, equipment identification field, and material batch field to identify whether a high number of anomalies in the current batch corresponds to a specific workstation, specific equipment, or specific incoming material batch. When the same process version field corresponds to multiple batch update samples, the indicator processing unit arranges them in order according to the traceability timestamp field, removes incomplete records in the missing number buffer, and retains valid samples that have completed the closed loop. After this processing, the batch update sample contains at least the process version field, changeover duration field, number of quality anomalies field, energy consumption data field, desktop shape identification field, rated power identification field, color identification field, workstation number field, and formula version number field, which can be directly called by subsequent update solution links.

[0140] Specifically, after receiving the batch update samples, the loss function construction unit establishes evaluation items around the changeover duration field, the number of quality anomalies field, and the energy consumption data field, and generates a deviation record table corresponding to the current formula version number. The update solution unit then reads the deviation record table, the currently effective formula record, and the historical version record, and performs iterative updates on the solder paste thickness parameter, screw torque parameter, test upper and lower limit parameters, aging temperature curve parameter, and scheduling weight. The iterative update does not directly replace all parameters, but first splits the chip mounting parameter segment, assembly parameter segment, test parameter segment, and aging parameter segment according to the workstation number, and then performs local corrections according to the source of the anomaly: when the changeover duration field is too large, the scheduling weights related to fixture switching and the workstation switching order are corrected first; when the number of quality anomalies field is too large, the test upper and lower limit parameters and torque parameters are corrected first; when the energy consumption data field is too large, the aging temperature curve parameters and the aging workstation scheduling weights are corrected first. The update solving unit generates candidate formula records after each round of correction and compares these records with historical preferred records of similar products. If the current candidate formula record is superior to the currently effective formula record in the fields of quality anomaly frequency, replacement time, and energy consumption data, it enters the version review unit. If any key field is missing or the batch fluctuation is too large, the current candidate formula record is written to the pending review area and not included in the release process. The version review unit generates a new formula version number, a review timestamp field, and a version status field for candidate formula records entering the review process. The version status field includes candidate status, effective status, and rollback status. Before the new formula version number is officially issued, the currently effective formula record remains unchanged.

[0141] In one engineering embodiment, after a batch of round tabletop electric heaters is traced and collected, the formula update module extracts the time records, anomaly records, and power records of the batch at the first workstation, test workstation, and aging workstation from the traceability dataset, and organizes them to obtain the current batch's model changeover duration field, quality anomaly count field, and energy consumption data field. Subsequently, the loss function construction unit generates a deviation record table corresponding to the current process version field, and the update solution unit performs multiple rounds of correction for the aging parameter segment and test parameter segment to obtain new aging temperature curve parameters and new test upper and lower limit parameters. Then, the version review unit generates a new formula version number. After receiving the approved new formula version number, the release and write-back unit writes the updated formula to the formula storage unit and the corresponding update schedule table to the schedule record area. At the same time, the release and write-back unit writes the new formula version number, effective time field, rollback pointer field, and applicable model number field back to the formula call processing in S200, so that subsequent orders with the same tabletop shape identifier, the same rated power identifier, and the same color identifier can directly call the updated formula when entering S200. In other words, the output product of S700 is an updated formula. The output field names include the formula version number field, the applicable model number field, the work station number field, the solder paste thickness parameter field, the screw torque parameter field, the test upper and lower limit parameter field, the aging temperature curve parameter field, the effective time field, the version status field, and the rollback pointer field. The updated formula is fed back into the formula call processing of S200 to form a closed loop across the main steps.

[0142] The technical effect of this step can be summarized as follows: S700 incorporates the changeover time, number of quality anomalies, and energy consumption data from the traceability dataset into the same update chain, forming a continuous correction process oriented towards the recipe version. The updated recipe is not generated in isolation, but is synchronously associated with the workstation number, process version, and historical records, and has a clear version source when subsequently called by S200. After the updated recipe flows back to the recipe call processing, order parsing, recipe call, quality linkage, aging linkage, and traceability aggregation form a complete closed loop.

[0143] Furthermore, in another preferred embodiment, the present invention implements adaptive scheduling for mixed-flow changeover scenarios: after the desktop shape recognition camera at the first workstation of the production line captures an image of a square or round table, the edge gateway automatically calls the changeover robot to change the fixture and notifies the SMT placement machine in advance to prepare a new stencil, significantly reducing changeover loss time; in terms of quality in-process closed-loop, when the reset pulse width data output by the ATE tester exceeds the preset range, the edge gateway pulls the safety relay coil low within a millisecond response time through the MCU coprocessor, causing the current production line to stop immediately and illuminate a yellow light, while simultaneously issuing a material locking command to the feeder, prohibiting the next batch of components from being put on the line, thus locking the quality risk into this batch; in terms of energy consumption closed-loop, the old The PLC in the aging chamber uploads the number of online units in real time, and the edge gateway runs a simplified thermodynamic model. Based on the number of online units and the internal and external temperature difference, it dynamically adjusts the PWM duty cycle of the heating tubes to stabilize the temperature in the aging chamber within the target range, significantly reducing energy consumption per shift. For rapid traceability, all workstation data is written to a time-series database using the product serial number as the primary key, enabling minute-level traceability and quickly locating the five key elements: personnel, machine, material, method, and environment. For self-learning optimization, the cloud uses changeover time, number of quality anomalies, and energy consumption data as loss functions to continuously iterate and update parameters such as solder paste thickness, screw torque, test upper and lower limits, and aging temperature curve parameters in the equipment parameter formula, continuously improving system performance.

[0144] Example 2: Figure 2 This diagram illustrates a structural block diagram of a production process control system for an electric heating table based on the Internet of Things (IoT) according to an embodiment of the present invention. Figure 2 As shown, the structure may include:

[0145] The order parsing module 01 receives order data from the enterprise resource planning (ERP) system, extracts desktop shape, rated power, and color, and generates a process feature vector. Specifically, the order parsing module is connected to the ERP system, the production line first-station image acquisition unit, and the order cache. It receives the order number, model number, planned quantity, delivery batch, rated power field, and color code field, and simultaneously receives desktop image records output by the production line first-station image acquisition unit. The order parsing module first performs integrity checks and duplicate record removal on the order number, model number, and time record, then performs contour recognition and template correspondence processing on the desktop image record to obtain the desktop shape identifier. At the same time, it performs field mapping on the rated power field and color code field to obtain the rated power identifier and color identifier. The order parsing module encapsulates the order number, model number, desktop shape identifier, rated power identifier, color identifier, order version number, and generation time record into a process feature vector, writes the process feature vector into the order cache, and provides the process feature vector and order version number to the recipe calling module for the recipe calling module to call.

[0146] The recipe calling module 02, connected to the order parsing module, receives the process feature vector and calls the solder paste thickness parameters, screw torque parameters, test upper and lower limit parameters, and aging temperature curve parameters corresponding to the workstation number to generate an equipment parameter recipe. Specifically, the recipe calling module is connected to the order parsing module, the recipe storage unit, the workstation mapping unit, and the version management unit, and receives the model number, desktop shape identifier, rated power identifier, color identifier, and order version number from the process feature vector. The recipe calling module first calls the fixture-related parameter segments based on the model number and desktop shape identifier, then calls the test upper and lower limit parameters and aging temperature curve parameters based on the rated power identifier, and then calls the appearance component switching record based on the color identifier. The workstation mapping unit completes the correspondence between the workstation number and the parameter segments. The recipe calling module encapsulates the solder paste thickness parameters, screw torque parameters, test upper and lower limit parameters, aging temperature curve parameters, workstation number, recipe version number, and issuance order record into an equipment parameter recipe, and writes the equipment parameter recipe into the version management unit and the recipe cache area, providing the equipment parameter recipe to the workstation access module for the workstation access module to receive.

[0147] The workstation access module 03, connected to the recipe calling module, is used to receive the equipment parameter recipe, collect real-time workstation operating data, and convert serial interface data into message queue telemetry transmission protocol messages to generate a unified operating data stream. Specifically, the workstation access module is connected to the recipe calling module, the field device interface, the transparent serial bridge, and the edge gateway, and receives the workstation number, recipe version number, solder paste thickness parameter, screw torque parameter, test upper and lower limit parameter, and aging temperature curve parameter from the equipment parameter recipe. The workstation access module establishes a collection queue based on the workstation number for the placement workstation, assembly workstation, and testing workstation. The serial port interfaces of the workstations and old workstations perform polling to collect parameter identifiers, parameter values, device numbers, and collection time records, and perform status marking on records with missing fields and disconnections. The workstation access module organizes the serial port interface data into message queue telemetry transmission protocol messages via a transparent serial bridge, and then encapsulates the device number, workstation number, parameter identifier, parameter value, collection time record, recipe version number, and data status into a unified working condition data stream. The unified working condition data stream is written into the edge gateway buffer and provided to the quality linkage module, while retaining the collection status records for the traceability and aggregation module to read.

[0148] The quality linkage module 04, connected to the workstation access module and the recipe calling module, is used to receive the unified operating condition data stream and the equipment parameter recipe, perform anomaly judgment, line stop control, and material locking control, and generate a linkage status set. Specifically, the quality linkage module is connected to the workstation access module, the recipe calling module, the safety relay, the feeder, and the alarm light. It receives the equipment number, workstation number, parameter identifier, parameter value, acquisition time record, and recipe version number from the unified operating condition data stream, and receives the test upper and lower limit parameters and the parameter segment corresponding to the workstation number from the equipment parameter recipe. The quality linkage module first performs workstation screening and... Based on the version correspondence, and according to the test upper and lower limit parameters, screw torque parameters, and solder paste thickness parameters, range comparison and status comparison are completed, and anomaly judgment results are formed for the reset pulse width record, torque record, and test record; when the anomaly judgment result meets the linkage triggering condition, the quality linkage module outputs a line stop control record to the safety relay, a material locking control record to the feeder, and a yellow light drive record to the alarm light, and encapsulates the equipment number, workstation number, anomaly type, line stop status, material locking status, yellow light status, collection time record, and formula version number into a linkage status set; the linkage status set is transmitted to the aging linkage module, and simultaneously written into the linkage buffer for the traceability collection module to call.

[0149] The aging linkage module 05, connected to the quality linkage module, receives the linkage status set and calculates the heating tube pulse width modulation duty cycle based on the online unit count and temperature status to generate heating control quantities. Specifically, the aging linkage module is connected to the quality linkage module, the aging chamber programmable logic controller, the temperature acquisition unit, and the heating tube drive unit, and receives the workstation number, shutdown status, abnormality type, acquisition time record, and recipe version number from the linkage status set. The aging linkage module first extracts the entry count record, exit count record, and operating status record from the aging chamber programmable logic controller to form the online unit count, and then extracts the data from the temperature acquisition unit. Temperature records inside and outside the aging chamber are used to establish a temperature status, and the aging temperature curve parameters corresponding to the formula version number are retrieved. The aging linkage module performs heat load sorting, control period identification, and duty cycle calculation based on the shutdown status, number of online units, temperature status, and aging temperature curve parameters to obtain the heating tube pulse width modulation duty cycle. Then, the station number, formula version number, number of online units, temperature inside and outside the aging chamber, heating tube pulse width modulation duty cycle, heating stage, freezing status, and control time records are encapsulated into heating control quantities. The heating control quantities are transmitted to the traceability and collection module and simultaneously written into the aging station execution buffer.

[0150] The traceability and data collection module 06, connected to the workstation access module, the quality linkage module, and the aging linkage module, receives the heating control quantity and associates it with operating condition data, personnel identification, equipment identification, material batch, process version, and environmental data according to the product serial number to generate a traceability dataset. Specifically, the traceability and data collection module is connected to the workstation access module, the quality linkage module, the aging linkage module, the time-series database, and the barcode acquisition terminal. It receives the workstation number, formula version number, number of online units, aging chamber temperature, aging chamber outside temperature, heating tube pulse width modulation duty cycle, and control time record from the heating control quantity, and extracts the product serial number from the barcode acquisition terminal. The traceability and data collection module then retrieves the product serial number from the workstation access module based on the product serial number. The input module retrieves operating condition data, anomaly types and linkage statuses from the quality linkage module, personnel identifiers from the personnel terminal record area, equipment identifiers from the equipment ledger, material batches from the incoming material scanning record area, process versions from the version management unit, and environmental data and temperature and humidity curves from the environmental acquisition record area. It then performs batch merging and product serial number encapsulation on the aforementioned records. The traceability collection module encapsulates product serial numbers, operating condition data, personnel identifiers, equipment identifiers, material batches, process versions, environmental data, temperature and humidity curves, traceability time records, and anomaly statuses into a traceability dataset. This traceability dataset is written into the time-series database and provided to the formula update module for reception.

[0151] The formula update module 07 is connected to the traceability collection module and the formula calling module. It is used to receive the traceability dataset, combine the changeover time, the number of quality anomalies and energy consumption data to perform iterative updates and generate updated formulas. Specifically, the formula update module is connected to the traceability collection module, the formula call module, the version management unit, and the scheduling record area. It receives product serial numbers, operating condition data, process versions, traceability time records, and abnormal states from the traceability dataset. The formula update module first extracts the first workstation switching record, test abnormality record, aging workstation power record, and temperature record from the traceability dataset, and organizes them into changeover duration, number of quality abnormalities, and energy consumption data. Then, it performs sample merging according to workstation number and process version. The formula update module performs iterative updates on solder paste thickness parameters, screw torque parameters, test upper and lower limit parameters, aging temperature curve parameters, and scheduling weights based on the changeover duration, number of quality abnormalities, and energy consumption data to form an updated formula. The updated formula is then written to the version management unit and the scheduling record area. The updated formula is sent back to the formula call module for subsequent orders to receive via the corresponding equipment parameter formula call link. At the same time, it retains version status records and rollback pointer records for subsequent traceability calls.

[0152] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the present invention. These examples are merely for the purpose of helping to understand the method and core ideas of the present invention. The above descriptions are only preferred embodiments of the present invention. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or variations without departing from the principles of the present invention, and can also combine the above technical features in an appropriate manner. These improvements, modifications, variations, or combinations, or the direct application of the concept and technical solution of the present invention to other situations without modification, should all be considered within the scope of protection of the present invention.

Claims

1. A method for controlling the production process of an electric heating table based on the Internet of Things, characterized in that, include: S100: Obtain ERP orders, perform feature extraction and vector generation processing to obtain process feature vectors; The ERP order specifically includes an order number, model number, planned quantity, delivery batch, power specification field, and color code field. The power specification field is used to perform cross-validation with the model number to generate a rated power identifier. The color code field is used to perform a corresponding match with the color coding table in the material master data to generate a color identifier. The model number serves as a mapping index for subsequent recipe calls to retrieve the corresponding workstation parameters. S200. Based on the process feature vector, perform hierarchical matching and workstation number reorganization to obtain the equipment parameter formula; S300. Based on the device parameter formula, perform interface access and message conversion processing to obtain a unified operating condition data stream; S400. Based on the unified working condition data stream, perform dual-core anomaly judgment and three-level linkage processing of line stop and material locking alarm to obtain a linkage status set. S500. Based on the linkage state set and aging temperature curve parameters, perform thermodynamic load estimation and feedforward proportional-integral dynamic duty cycle adjustment to obtain the heating control quantity. S600. Based on the heating control quantity and the barcode scanning record, perform primary key registration and batch association processing to obtain a traceability dataset; S700. Based on the traceability dataset, a loss function is constructed and iteratively updated to obtain an updated formula.

2. The method according to claim 1, characterized in that, The process of feature extraction and vector generation includes: Perform field integrity checks, version number checks, and timestamp sorting on ERP orders, remove duplicate messages, and retain the current valid order version; Read the model number, power specification field, and color code field from the ERP order, and identify the desktop shape identifier from the first workstation image; The identification of the desktop shape identifier includes: acquiring a top view image of the desktop by the first workstation image acquisition unit of the production line, performing contour segmentation, edge closure and template matching, and identifying the desktop outer contour with a preset shape template library; The power specification field is cross-verified with the model number to generate a rated power identifier; The color code field is matched with the color coding table in the material master data to generate a color identifier; Once the desktop shape identifier, rated power identifier, and color identifier have all passed the corresponding verification with the model mapping table, the order number, model number, desktop shape identifier, rated power identifier, color identifier, order version number, and generation timestamp are encapsulated in a fixed field order to generate a process feature vector. Simultaneously, a difference comparison is performed on the process feature vectors of adjacent orders. When the desktop shape identifier changes, a fixture replacement trigger mark is recorded. When the rated power identifier changes, a test threshold replacement trigger mark is recorded. When the color identifier changes, an exterior part replacement trigger mark is recorded. All of these trigger marks are written into the process feature vector.

3. The method according to claim 2, characterized in that, The process of performing hierarchical matching and workstation number reorganization includes: The hierarchical matching process includes: Perform tooling layer matching on the desktop shape identifier, and call the fixture parameter segment and assembly torque parameter segment corresponding to the square desktop or round desktop; Perform electrical layer matching on the rated power label, and call the test upper and lower limit parameter segments and aging temperature curve parameter segments corresponding to the power level; Perform appearance layer matching on the color identifier and call the appearance component switching mark and packaging switching mark corresponding to the color code; The workstation number reorganization process includes: generating workstation numbers according to raw material kitting workstation, SMT placement workstation, ICT testing workstation, functional testing workstation, aging workstation, assembly workstation and packaging workstation, and then binding each parameter segment in the parameter template to the corresponding workstation number to form a workstation-level recipe record. Generate a recipe version number, an effective status flag, a call source flag, and a distribution order flag for the workstation-level recipe record; When the trigger marker in the process feature vector includes a fixture change trigger marker, the assembly station and fixture preparation station will be placed in the preceding order of issuance; when it includes a test threshold change trigger marker, the functional test station and aging station will be placed in the preceding order of issuance; when it includes an appearance part change trigger marker, the packaging station and appearance part material preparation station will be placed in the preceding order of issuance. The packaged device parameter formula includes the workstation number, solder paste thickness parameter, screw torque parameter, test upper and lower limit parameter, aging temperature curve parameter, formula version number, effective status flag, call source flag, and distribution sequence flag.

4. The method according to claim 3, characterized in that, The process of handling interface access includes: Read the corresponding parameter segment in the equipment parameter recipe according to the workstation number, and establish the access sequence table for the current batch according to the issuance sequence mark; Print parameter labels and equipment status labels corresponding to the solder paste thickness parameters of the SMT placement equipment; Automatic screw-driving equipment should be equipped with labels for the torque values ​​corresponding to the screw torque parameters, the nozzle number label, and the tightening completion label. Load the test equipment with the reset pulse width data label, electrical parameter label and test conclusion label corresponding to the upper and lower limit parameters of the test; Load the aging temperature curve parameters into the aging chamber PLC with labels for the aging chamber's internal temperature, external temperature, number of online units, and heating status.

5. The method according to claim 4, characterized in that, The message conversion process includes: The serial port acquisition frame is split into the device number field, workstation number field, parameter identifier field, parameter value field, acquisition timestamp field, and recipe version number field, and the serial port acquisition frame is converted into an MQTT message; The MQTT message is sent via a transparent serial bridge, carrying TLS 1.3 encryption configuration and watchdog reset configuration; The converted messages are continuously written to the edge gateway buffer in chronological order, generating a unified operating condition data stream containing fields such as device number, workstation number, parameter identifier, parameter value, collection timestamp, recipe version number, interface type, message subject, and data status.

6. The method according to claim 5, characterized in that, The process of handling dual-core anomalies includes: The dual-core A7 processor receives the unified operating condition data stream and performs message proxying and formula comparison, while the MCU coprocessor receives the anomaly judgment result and performs line stop material locking control and safety relay driving. Data frames corresponding to the test station, data frames corresponding to the assembly station, and data frames with normal data status fields are filtered out from the unified working condition data stream to form a message queue to be judged. Read the workstation number field and recipe version number field from the message queue to be judged, and retrieve the parameter segment of the same workstation and version from the equipment parameter recipe; A range comparison is performed on the reset pulse width data. When the reset pulse width data falls outside the range of 160ms to 220ms, a reset anomaly flag is generated. A threshold comparison is performed on the actual torque value corresponding to the screw torque parameter, and a torque anomaly flag is generated when it exceeds the allowable range. A test exception flag is generated when the test device returns a failure status and the parameter identifier field points to a functional test item.

7. The method according to claim 6, characterized in that, The process of three-level linkage handling for line stoppage and material locking alarms includes: After receiving a reset error flag, torque error flag, or test error flag, the MCU coprocessor executes the actions in the order of line stop control, yellow light drive, and feeder locking. The line stop control outputs a line stop signal through a safety relay, which acts on the operating circuit of the production line where the current workstation is located; The locking control writes a locking control frame to the feeder, triggering the feeder to stop the next package of components from entering the current batch. The alarm driver writes a yellow light drive signal to the alarm light; Wait for the safety relay to return an execution confirmation signal and the feeder to return a material locking confirmation signal. If they do not return within the preset period, write an action unconfirmed flag and maintain the line stop state. The equipment number field, workstation number field, exception type field, line stop status field, yellow light status field, material lock status field, data collection timestamp field, and recipe version number field are encapsulated into a linkage status set.

8. The method according to claim 7, characterized in that, The process of estimating thermodynamic load includes: Read the shutdown status field, exception type field, and recipe version number field from the linkage status set; When the outage status field indicates that the current batch is in an outage state, the current batch is written to the aging freeze area and a freeze mark is generated; When the stop status field indicates that it is in the release status, the transmission cycle signal, the in-room count signal and the out-of-room count signal are collected from the aging room PLC, and the online station digital segment is generated by combining the product serial number mapping relationship. Read the temperature inside and outside the aging chamber and generate a temperature status field; It receives the online station digital segment, the aging chamber internal temperature field, the aging chamber external temperature field, and the aging temperature curve parameters. It identifies the heat capacity range based on the current batch's preheating stage, constant temperature stage, or cooling stage, and then identifies the heat dissipation status based on the internal and external temperature difference, outputting the basic duty cycle value.

9. The method according to claim 8, characterized in that, The process of feedforward proportional-integral dynamic duty cycle adjustment includes: The feedforward control unit receives the basic duty cycle value and temperature state change results, and performs pilot correction on the heating tube drive quantity; The proportional-integral control unit receives the feedforward correction result and the current temperature field in the aging chamber, performs continuous tracking correction, and generates the heating tube PWM duty cycle field. When the batch is in a frozen state, switch the base duty cycle value to the frozen duty cycle value and stop the amplification calculation for the number of new units. The workstation number field, recipe version number field, online station number field, aging chamber internal temperature field, aging chamber external temperature field, heating tube PWM duty cycle field, heating stage field, freeze status field, and control timestamp field are encapsulated into heating control quantities and written to the edge gateway buffer and aging chamber PLC execution register area.

10. A production process control system for an electric heating table based on the Internet of Things, used to execute the method according to any one of claims 1-9, characterized in that, include: The order parsing module extracts the desktop shape, rated power, and color to generate a process feature vector; The recipe calling module calls the solder paste thickness parameters, screw torque parameters, test upper and lower limit parameters, and aging temperature curve parameters corresponding to the workstation number to generate the equipment parameter recipe; The workstation access module collects real-time workstation status data and converts serial port interface data into message queue telemetry transmission protocol messages to generate a unified work status data stream. The quality linkage module performs anomaly detection, line stop control, and material locking control, and generates a linkage status set. The aging linkage module calculates the pulse width modulation duty cycle of the heating tube by combining the number of online units and the temperature status, and generates the heating control quantity. The traceability and aggregation module associates product serial numbers with operating condition data, personnel identification, equipment identification, material batches, process versions, and environmental data to generate traceability datasets. The formula update module combines data on changeover duration, number of quality anomalies, and energy consumption to perform iterative updates and generate updated formulas.

Images