Intelligent Control System and Method for Electroplating Production Line Based on Digital Twin

Abstract

This invention relates to the field of intelligent control technology for electroplating production lines, and discloses an intelligent management and control system and method for electroplating production lines based on digital twins. The method includes: acquiring faulty crane information and generating segmented spatiotemporal feasible regions for the crane; generating a timeout risk batch queue; generating a degraded operation scheduling scheme using a greedy search and local exchange algorithm; calculating a dragging amount correction coefficient and an estimated actual dragging amount based on over-immersion time, plating solution viscosity, and surface tension; generating an actual cleaning batch sequence and a cross-contamination risk list; performing a mass conservation deduction on the countercurrent cleaning digital twin model using the correction parameters; generating an emergency isolation rinsing command and performing cross-module bidirectional constraint transmission; issuing control commands and performing parameter correction after fault recovery.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for electroplating production lines, and more specifically, to an intelligent control system and method for electroplating production lines based on digital twins. Background Technology

[0002] In multi-plating-type mixed-line electroplating production lines, the digital twin system uses a degraded scheduling module and a cleaning section contamination control module to map and control the crane operation scheduling and the countercurrent cleaning process in real time. Multiple cranes coordinate to perform the workpiece transfer task between tanks according to the scheduling schedule. The cleaning section tracks the concentration changes of characteristic metal ions in each level of the cleaning tank through a countercurrent cleaning digital twin model, and optimizes the water replenishment flow and assesses the risk of cross-contamination based on the standard workpiece dragging parameters and the scheduled plating sequence.

[0003] When a crane suddenly stops due to mechanical failure, the track position occupied by the faulty crane will physically divide the continuous operating space into multiple isolation sections. The degraded scheduling module will use alternative cranes to urgently rescue workpieces at risk of timeout. The rescue order is determined by the accessibility of the alternative cranes and the urgency of the timeout, rather than by the original scheduling order.

[0004] However, existing digital twin models for the cleaning section still use standard drag volume parameters and the original scheduling of plating sequences for contamination risk assessment, which has the following drawbacks: First, prolonged immersion leads to a significantly higher actual drag volume of the workpiece than the standard parameters, causing the digital twin model for the cleaning section to underestimate the total amount of metal ions entering the cleaning tank by using the standard drag volume. Second, emergency scheduling changes the plating sequence of the workpiece entering the cleaning section, potentially causing highly cross-contamination-sensitive plating combinations to appear adjacently. The digital twin model for the cleaning section, using the original scheduling sequence, cannot identify such new risks. The dual information of increased drag volume and changed plating sequence is not perceived across modules in the digital twin system, leading to an underestimation of the cross-contamination risk in the cleaning section, which may cause surface quality defects in the workpiece. Summary of the Invention

[0005] This invention provides an intelligent control system and method for electroplating production lines based on digital twins, which solves the technical problems in related technologies such as the lack of effective degraded operation scheduling means when the gantry crane of the electroplating production line fails, the inability to accurately assess the changes in the amount of ultra-immersion batches carried, and the difficulty in controlling the risk of cross-contamination in the cleaning section in real time.

[0006] This invention discloses an intelligent control method for an electroplating production line based on digital twins, comprising the following steps: receiving a crane fault alarm signal, obtaining the stop position coordinates and physical dimension parameters of the faulty crane, dividing the production line track space into multiple isolated operating sections, and generating segmented spatiotemporal feasible domains for available cranes in each isolated section;

[0007] Get a snapshot of the soaking progress of all batches in production, calculate the remaining time of each batch before the maximum allowable soaking time limit, and generate a queue of batches with timeout risk by sorting them in ascending order of remaining time.

[0008] With the goal of minimizing the total timeout loss of all in-process batches, a degraded operation scheduling scheme is generated under the constraints of the segmented driving time-space feasible region through a fast scheduling algorithm that combines greedy search and local exchange.

[0009] Based on the superimmersion time of each superimmersion batch and the viscosity and surface tension parameters of the plating solution corresponding to the plating type, nonlinear correction calculation is performed using the superimmersion drag amount correction function to output the drag amount correction coefficient and the estimated value of the actual drag amount for each superimmersion batch.

[0010] Based on the estimated arrival time of each batch entering the cleaning section in the downgraded operation scheduling scheme, the actual cleaning batch sequence is generated, and the cross-contamination sensitivity matrix between plating types is queried to generate a cross-contamination risk list.

[0011] Replace the standard drag volume parameter in the countercurrent cleaning digital twin model with the estimated actual drag volume, perform mass conservation extrapolation of each characteristic metal ion according to the actual cleaning batch sequence, and output the residual concentration distribution of each cleaning tank.

[0012] The residual concentration distribution is compared with the allowable residual concentration threshold. The flushing flow rate and duration of the exceeding section are calculated and an emergency isolation flushing command is generated. The flushing time constraint is sent back to the degraded scheduling module for cross-module constraint transmission.

[0013] Issue downgraded operation scheduling instructions and emergency isolation flushing instructions, and after the fault is restored, correct and update the parameters of the ultra-immersion towing amount correction function based on the measured concentration data.

[0014] Furthermore, the generation of segmented spatiotemporal feasible regions for usable train travel within each isolation segment includes:

[0015] In the spatial dimension, the movement range of each available crane is limited by the two ends of the isolated operating section, one end being the edge of the section occupied by the faulty crane, and the other end being the physical endpoint of the production line track or the edge of the adjacent fault point; in the temporal dimension, the positional distance between any two available cranes in the same isolated section at the same time is not less than half of the sum of their respective physical dimensions.

[0016] Obtain the rated operating speed, acceleration parameters, and current load status of each normal operating vehicle. For each available vehicle, starting from its current position, calculate the farthest and nearest positions that the vehicle can reach at each time point according to the motion model of accelerating to the rated speed, running at a constant speed, and decelerating to a stop. Use the motion envelope as the boundary constraint curve of the vehicle in the segmented driving spatiotemporal feasible domain.

[0017] Furthermore, in the timeout risk batch queue, the remaining time of each batch is the maximum allowable immersion time limit for the corresponding plating type minus the immersion time already taken at the time of the failure. Batches with negative remaining time indicate that they have exceeded the maximum allowable immersion time limit at the time of the failure and are placed at the front of the timeout risk batch queue. The over-immersion time of each batch is initialized to zero at the time of the failure and is continuously updated over time until the batch is removed from the tank.

[0018] Furthermore, the execution process of the fast scheduling algorithm includes:

[0019] According to the order of the timeout risk batch queue, each batch is assigned an alternative execution vehicle and a retrieval time. For the current batch, the vehicle that makes the retrieval time of the batch earliest is selected from its corresponding alternative execution vehicle set. The corresponding time period of the vehicle is marked as occupied and its remaining available time period is updated.

[0020] After completing the initial allocation of all batches, the alternative execution of train allocation and extraction time of two batches with adjacent priorities are swapped in turn. The total timeout loss after the swap is calculated. If the total timeout loss after the swap is less than the total timeout loss before the swap, the swap is accepted and the degraded operation scheduling scheme is updated. Otherwise, the original allocation remains unchanged.

[0021] Repeat the local swap evaluation until no swap is accepted in a complete round of traversal, and output the degraded operation scheduling scheme.

[0022] The total timeout loss is the sum of the product of the plating type loss weight coefficient of each batch and the cumulative over-immersion time of that batch. In the local exchange assessment, the arrival order of each batch entering the cleaning section after the exchange is also estimated. If the exchange leads to an increase in the number of adjacent plating type combinations with high cross-contamination sensitivity, a penalty term is applied to the exchange scheme. The penalty term is the product of the penalty weight coefficient and the number of newly added highly sensitive adjacent combinations.

[0023] Furthermore, in the over-immersion drag amount correction function, the drag amount correction coefficient is equal to the product of the correction amplitude coefficient and the over-immersion time ratio term, viscosity ratio term, and surface tension ratio term; wherein, the over-immersion time ratio term is the over-immersion time nonlinear exponent power of the ratio of the over-immersion time of this batch to the reference over-immersion time, the viscosity ratio term is the viscosity influence exponent power of the ratio of the viscosity of the corresponding plating solution of this batch to the viscosity of the reference plating solution, and the surface tension ratio term is the surface tension influence exponent power of the ratio of the reference surface tension to the surface tension of the corresponding batch; the actual drag amount estimate is equal to the product of the standard drag amount parameter of the corresponding plating type of this batch and the drag amount correction coefficient; the reference over-immersion time, reference plating solution viscosity, reference surface tension, correction amplitude coefficient, over-immersion time nonlinear exponent, viscosity influence exponent, and surface tension influence exponent are obtained by fitting historical process test data.

[0024] Furthermore, when calculating the drag amount correction coefficient, the current plating solution temperature of the corresponding tank for each ultra-immersion batch is also obtained, and the temperature influence factor is multiplied into the ultra-immersion drag amount correction function. The temperature influence factor is the temperature influence exponent of the ratio of the current plating solution temperature to the reference temperature of the corresponding tank for that batch.

[0025] Furthermore, in the mass conservation deduction, after the i-th batch is cleaned, the concentration of characteristic metal ion j in the m-th cleaning tank is equal to the concentration of the ion in the cleaning tank before the batch enters, plus the product of the estimated actual drag amount of the batch and the concentration of the ion in the corresponding plating solution of the batch divided by the effective volume of the cleaning tank, minus the product of the outflow rate of the cleaning tank and the cleaning duration divided by the effective volume and multiplied by the concentration before entering the tank, plus the product of the countercurrent replenishment flow rate of the next cleaning tank and the cleaning duration divided by the effective volume and multiplied by the concentration of the next tank before entering the tank;

[0026] The final cleaning tank is replenished with fresh pure water, and the concentration of the next cleaning tank is zero. The cleaning duration of each batch is taken from the actual residence time of the batch in the corresponding cleaning tank in the downgraded operation scheduling scheme.

[0027] In the process of mass conservation deduction, the secondary drag effect when the batch is transferred between cleaning tanks of each level is also calculated. When the workpiece is lifted from the current cleaning tank and sent to the next cleaning tank, the amount of ions carried by the liquid film attached to the workpiece surface is estimated by multiplying the concentration of the current cleaning tank and the secondary drag volume and included in the concentration update of the next cleaning tank. The secondary drag volume is adjusted using the same drag volume correction coefficient as the initial drag volume.

[0028] Furthermore, the process of determining the flushing flow rate and duration for the excess section includes:

[0029] Using a sequential quadratic programming algorithm, the optimization objective is to minimize the total flushing water volume of each excessive cleaning section. The constraint is that the concentration of target characteristic ions in each excessive cleaning section does not exceed the allowable residual concentration threshold after flushing. The upper bound constraint is the maximum allowable flushing flow rate of each cleaning section. The flushing flow rate and duration of each excessive cleaning section are then solved.

[0030] The concentration of characteristic metal ions after rinsing is calculated based on a first-order decay model of continuous dilution. This concentration is equal to the exponential function value of the quotient obtained by multiplying the predicted residual concentration before rinsing by the product of the negative rinsing flow rate and the duration, and dividing by the effective volume of the cleaning tank at that stage.

[0031] The cross-module constraint transmission includes: assessing whether the flushing duration of each exceeding standard section is completed within the arrival time interval of the corresponding adjacent batch in the degraded operation scheduling scheme; if the flushing duration is greater than the arrival time interval, the constraint information of insufficient time interval is sent back to the degraded scheduling module; an isolation waiting time of not less than the difference between the flushing duration and the arrival time interval is inserted between adjacent rescue batches involving highly sensitive combinations; or the subsequent batches are allocated to the standby cleaning section.

[0032] Furthermore, the cross-module constraint transmission is an iterative feedback process. After the downgraded scheduling module makes local scheduling adjustments, the cleaning section pollution control module re-executes the generation of the actual cleaning batch sequence and cross-contamination risk list, quality conservation deduction, and threshold comparison with the updated downgraded operation scheduling scheme. If there are still sections exceeding the standard, the constraint information is sent back to the downgraded scheduling module for further adjustment. The iterative process sets a maximum iteration limit. The iteration terminates when the maximum iteration limit is reached or all allowable residual concentration threshold constraints are met.

[0033] The correction update includes: acquiring measured ion concentration data of each ultraimmersion batch in each cleaning tank during the actual cleaning process, calculating the prediction deviation between the measured concentration value and the predicted value of the countercurrent cleaning digital twin model, aiming to minimize the sum of squares of prediction deviations of all ultraimmersion batches in each cleaning tank, refitting the correction amplitude coefficient, ultraimmersion time nonlinearity index, viscosity influence index and surface tension influence index in the ultraimmersion drag amount correction function using the least squares method, and correcting the effective volume and mixing efficiency parameters of each cleaning tank; wherein, the relative change between the parameter value after each correction and the parameter value before correction does not exceed the preset upper limit of the rate of change.

[0034] This invention provides an intelligent control system for electroplating production lines based on digital twins, used to execute the intelligent control method for electroplating production lines based on digital twins as described in any one of claims 1 to 9, characterized in that it includes:

[0035] The fault perception and segment division module is used to receive the crane fault alarm signal, obtain the stopping position coordinates and physical size parameters of the faulty crane, divide the production line track space into multiple isolated operating segments, and generate the segmented spatiotemporal feasible domain of the available cranes in each isolated segment.

[0036] The timeout risk assessment module is used to obtain a snapshot of the soaking progress of all batches in production, calculate the remaining time of each batch before the maximum allowable soaking time limit, and generate a timeout risk batch queue by sorting the batches in ascending order of remaining time.

[0037] The degraded scheduling module is used to generate a degraded operation scheduling scheme under the constraints of the segmented driving time and space feasible region by using a fast scheduling algorithm that combines greedy search and local exchange, with the goal of minimizing the total timeout loss of all in-process batches.

[0038] The drag amount correction module is used to perform nonlinear correction calculations based on the over-immersion time of each over-immersion batch and the viscosity and surface tension parameters of the plating solution corresponding to the plating type, using the over-immersion drag amount correction function, and outputs the drag amount correction coefficient and the estimated value of the actual drag amount for each over-immersion batch.

[0039] The cleaning sequence and risk identification module is used to generate the actual cleaning batch sequence based on the expected arrival time of each batch entering the cleaning section in the degraded operation scheduling scheme, and to generate a cross-contamination risk list by querying the cross-contamination sensitivity matrix between plating types.

[0040] The cleaning section pollution control module is used to replace the standard drag volume parameter in the countercurrent cleaning digital twin model with the estimated actual drag volume, perform mass conservation deduction of each characteristic metal ion according to the actual cleaning batch sequence, and output the residual concentration distribution of each cleaning tank.

[0041] The emergency flushing and cross-module constraint transmission module is used to compare the residual concentration distribution with the allowable residual concentration threshold, solve the flushing flow rate and duration for the exceeding section and generate an emergency isolation flushing command, and send the flushing time constraint back to the degraded scheduling module for cross-module constraint transmission.

[0042] The instruction issuance and parameter correction module is used to issue downgrade operation scheduling instructions and the emergency isolation flushing instructions, and to correct and update the parameters of the ultra-immersion towing amount correction function based on the measured concentration data after the fault is recovered.

[0043] This invention provides an intelligent control method for electroplating production lines based on digital twins. It solves the technical problem of underestimated cross-contamination risk caused by deviations in drag volume parameters and missing plating sequence information in emergency scenarios of crane malfunctions in existing electroplating production line digital twin systems. The method achieves the following technical effects: By generating segmented spatiotemporal feasible domains for the crane, it provides feasible spatial constraints based on track isolation constraints for solving degraded operation scheduling schemes; by incorporating ultra-immersion time, plating solution viscosity, and surface tension into the drag volume calculation in a nonlinear form through an ultra-immersion drag volume correction function, it ensures that the estimated total amount of metal ions entering the cleaning tank is consistent with the ultra-immersed workpiece. The actual surface condition is matched; by jointly querying the actual cleaning batch sequence and the cross-contamination sensitivity matrix between plating types, the adjacent combinations of highly sensitive plating types that do not exist in the original schedule due to the change in the rescue order are identified; through the bidirectional constraint transmission between the degraded scheduling module and the cleaning section contamination control module, the rescue scheduling can meet the contamination isolation time requirements of the cleaning section while ensuring the timeliness of overtime protection, reducing the possibility of timing conflicts when the two subsystems make independent decisions; through parameter correction and update after fault recovery, the predicted parameters of the countercurrent cleaning digital twin model can be continuously corrected based on the measured data. Attached Figure Description

[0044] Figure 1 This is a flowchart of the intelligent control method for electroplating production lines based on digital twins provided in an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of the production line track section division and the distribution of each traveling car position provided in an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of the sorting of remaining time for timeout risk batches provided in an embodiment of the present invention;

[0047] Figure 4 This is a schematic diagram comparing the soaking time of each batch with the maximum allowable soaking time provided in the embodiments of the present invention;

[0048] Figure 5 This is a schematic diagram illustrating the contribution of timeout losses for each batch in the degraded operation scheduling scheme provided in this embodiment of the invention;

[0049] Figure 6 This is a schematic diagram of the correction coefficient for the ultra-immersion batch dragging amount and the actual dragging amount provided in the embodiments of the present invention;

[0050] Figure 7 The three-stage countercurrent cleaning tank Cr provided in this embodiment of the invention 6 A schematic diagram comparing the ⁺ concentration projection results with the threshold;

[0051] Figure 8 This is a schematic diagram of a heatmap showing the changes in cross-contamination sensitivity levels provided in an embodiment of the present invention;

[0052] Figure 9 This is a schematic diagram comparing the model parameters before and after fault recovery provided in an embodiment of the present invention. Detailed Implementation

[0053] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be described in detail below with reference to the accompanying drawings. 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.

[0054] In multi-plating-type mixed-line electroplating production lines, the digital twin system uses a degraded scheduling module and a cleaning section contamination control module to map and control the crane operation scheduling and the countercurrent cleaning process in real time. Multiple cranes coordinate to perform the workpiece transfer task between tanks according to the scheduling schedule. The cleaning section tracks the concentration changes of characteristic metal ions in each level of the cleaning tank through a countercurrent cleaning digital twin model, and optimizes the water replenishment flow and assesses the risk of cross-contamination based on the standard workpiece dragging parameters and the scheduled plating sequence.

[0055] When a crane suddenly stops due to mechanical failure, the track position occupied by the faulty crane physically divides the continuous operating space into multiple isolated sections. Workpieces in the faulty area face the risk of over-immersion due to timeout. The degraded scheduling module uses alternative cranes to urgently rescue workpieces at risk of over-immersion. The rescue order is determined by the accessibility of the alternative crane and the urgency of the timeout, rather than by the original scheduled plating sequence. However, the digital twin model of the cleaning section still uses standard drag volume parameters and the original scheduled plating sequence for contamination risk assessment, resulting in the following two information biases: First, over-immersion alters the wettability and coating morphology of the workpiece surface, leading to a significantly higher actual drag volume than the standard parameters. The digital twin model of the cleaning section, using the standard drag volume, underestimates the total amount of metal ions entering the cleaning tank. Second, the rescue scheduling changes the plating sequence of workpieces entering the cleaning section, potentially causing highly cross-contamination-sensitive plating combinations to appear adjacently. The digital twin model of the cleaning section, using the original scheduling sequence, cannot identify such new risks. The dual information of increased drag volume and change in plating sequence was not perceived across modules in the digital twin system, resulting in a serious underestimation of the risk of cross-contamination in the cleaning section, which may cause defects in the surface quality of the workpiece.

[0056] This embodiment provides an intelligent control method for electroplating production lines based on digital twins. It corrects and quantifies the over-immersion dragging amount in emergency scenarios of crane failure, regenerates the actual cleaning batch sequence, and performs bidirectional constraint transmission between the degraded scheduling module and the cleaning section contamination control module, outputting coordinated and consistent degraded operation scheduling instructions and emergency isolation rinsing instructions.

[0057] At least one embodiment of the present invention discloses an intelligent control method for electroplating production lines based on digital twins, such as... Figure 1 As shown, it includes the following steps:

[0058] It should be understood that the execution entity of this embodiment is the processor or server in the digital twin system of the electroplating production line. The digital twin system of the electroplating production line maintains real-time mapping data of the physical space of the production line, including the number, position coordinates, motion status, and physical dimension parameters of each crane, the position of each tank and the current batch information, as well as the concentration monitoring values ​​of multi-component characteristic metal ions in each level of the cleaning tank in the cleaning section. In addition, the digital twin system of the electroplating production line pre-stores standard drag-carrying parameters, plating solution viscosity parameters, plating solution surface tension parameters, the maximum allowable immersion time limit for each batch, and the cross-contamination sensitivity matrix between plating types.

[0059] Step 1: Obtain fault information and generate segmented spatiotemporal feasible regions for driving within the isolation section;

[0060] The system receives a crane malfunction alarm signal and obtains the malfunctioning crane's number, stopping position coordinates, and physical dimensions. Based on the malfunctioning crane's stopping position coordinates and physical dimensions, it calculates the track section occupied by the malfunctioning crane and divides the production line track space into multiple isolated operating sections bounded by the malfunctioning section. It then obtains the current position coordinates and motion status of each normal crane, determines the isolation operating section to which each normal crane belongs, and generates segmented spatiotemporal feasible regions for usable cranes within each isolation section.

[0061] It should be noted that the aforementioned segmented operational spacetime feasible domain refers to the set of constraints on the movement range of available trains in both time and space dimensions within each isolated operational segment. Spatially, the movement range of available trains is limited by the two boundaries of the isolated segment, with one end being the edge of the section occupied by the faulty train, and the other end being the physical endpoint of the production line track or the edge of an adjacent fault point. Temporally, when multiple available trains exist within the same isolated segment, the movement trajectories of each train must not overlap spatially on the time axis; that is, the distance between the positions of any two trains at the same moment must not be less than half the sum of their respective physical dimensions.

[0062] Furthermore, to more accurately describe the motion capability of available vehicles within each isolation section, the rated operating speed, acceleration parameters, and current load state of each normal vehicle are acquired when generating the segmented spatiotemporal feasible domain. The acceleration and deceleration processes of the vehicles are incorporated into the boundary calculation of the segmented spatiotemporal feasible domain. Specifically, for each available vehicle, starting from its current position, the farthest and closest positions that the vehicle can reach at each time point are calculated according to the motion model of accelerating to rated speed, running at a constant speed, and decelerating to a stop. This motion envelope is used as the boundary constraint curve of the vehicle in the segmented spatiotemporal feasible domain.

[0063] Step 2: Generate a timeout risk batch queue;

[0064] Obtain a snapshot of the immersion progress of all batches in production at the moment the fault occurred. This snapshot includes the tank number of each batch, the immersion time already elapsed, and the corresponding plating type. Identify the batches in production within each tank in the area originally managed by the faulty crane, and calculate the remaining time for each batch before the maximum allowable immersion time limit. ,in The calculation method is as follows:

[0065]

[0066] in, For the first The maximum allowable immersion time for each batch of plating type. The time when the fault occurred Soaking time for each batch This refers to the batch number. Each batch will be sorted by remaining time. Arrange the batches in ascending order to generate a timeout risk queue. Simultaneously record the plating type identifier and the elapsed immersion time for each batch. The soaking time has exceeded the limit. It is initialized to zero at the moment of failure and continuously updated over time until the batch is extracted from the tank.

[0067] It should be noted that in the aforementioned batch queue with timeout risk, the remaining time... A negative value indicates that the maximum allowable soaking time has been exceeded at the time of the failure. The absolute value is the soaking time that has been exceeded. Such batches are at the front of the timeout risk batch queue and have the highest rescue priority.

[0068] Step 3: Generate a degraded operation scheduling plan;

[0069] For each batch in the timeout risk batch queue, analyze whether its location in the segmented driving spatiotemporal feasible domain is within the reachable range of a normal driving machine. Specifically, determine whether the position coordinates of the batch's location in the segmented driving spatiotemporal feasible domain fall within the spatial projection range of an available driving machine belonging to the same isolation section as the segmented driving spatiotemporal feasible domain. If so, mark the available driving machine as the alternative driving machine for that batch.

[0070] To minimize the total timeout loss of all in-process batches With the primary objective, a fast scheduling algorithm combining greedy search and local swapping is used to generate a degraded scheduling scheme. Total timeout penalty. The definition of is:

[0071]

[0072] in, This represents the total number of batches in the timeout risk batch queue. For the first The loss weighting coefficient corresponding to the plating type in each batch For the first The cumulative overimmersion time for each batch, calculated from the actual time the replaced vehicle was extracted from the tank. This is the batch number.

[0073] Furthermore, the total amount of the aforementioned timeout losses middle, The value of is directly related to the degraded operation scheduling scheme: for degraded operation scheduling schemes that allocate alternative execution routes and extract time is The One batch, ,in This represents the soaking time of this batch at the time the malfunction occurred. This is to replace the actual time when the alternative execution crane arrives at the tank containing the batch and completes the extraction; different alternative execution crane allocation schemes and extraction time arrangements will result in different... The value is selected, thus affecting The size of the allocation scheme is determined by the fast scheduling algorithm, which compares different allocation schemes. The optimal solution is determined by the size of the equation.

[0074] The execution process of the above fast scheduling algorithm includes the following sub-steps:

[0075] Step 301: According to the order in the timeout risk batch queue, assign alternative execution vehicles and extraction times to each batch in sequence. For the current batch, select the vehicle that makes the extraction time of the batch earliest from its corresponding alternative execution vehicle set, mark the corresponding time period of the vehicle as occupied, and update the remaining available time period of the vehicle in the segmented vehicle spatiotemporal feasible domain.

[0076] Step 302: After completing the initial allocation of all batches, a partial swap evaluation is performed on adjacent priority batch pairs. Specifically, the alternative execution vehicle allocation and retrieval time of two adjacent batches in the timeout risk batch queue are swapped sequentially, and the total timeout loss after the swap is calculated. ,like If the swap is accepted, the downgraded operation scheduling scheme will be updated; otherwise, the original allocation will remain unchanged.

[0077] Step 303: Repeat the local exchange evaluation in step 302 until no exchange is accepted in a complete round of traversal. Output a degraded operation scheduling scheme that includes the alternative execution train number, extraction time and transfer path for each timeout risk batch.

[0078] It should be noted that the above transfer path refers to the movement trajectory of the alternative execution crane after extracting the batch from the current tank to the next target tank. The next target tank is the tank for the subsequent process of the batch in the original schedule. If the tank for the subsequent process is not within the reach of the alternative execution crane, the batch will be transferred to a temporary storage location or a spare tank with the same function within the reach of the alternative execution crane.

[0079] Furthermore, in order to take into account the pollution control requirements of the subsequent cleaning section during the scheduling solution process, in the local exchange evaluation in step 302, in addition to calculating the total timeout loss... In addition, the arrival order of each batch entering the cleaning section after the swap is estimated. If the swap leads to an increase in the number of adjacent plating combinations with high cross-contamination sensitivity, a penalty will be imposed on the swap plan. The evaluation indicators will be adjusted to ,in As a penalty weighting coefficient, This determines the number of newly added highly sensitive adjacent combinations.

[0080] Step 4: Calculate the correction factor for the towing amount of the over-immersion batch and the estimated value of the actual towing amount;

[0081] Based on the soaking time of each soaking batch and the corresponding plating solution viscosity of the plating type and surface tension The parameters are calculated using a nonlinear correction function for the over-immersion drag amount, and the drag amount correction coefficient for each over-immersion batch is output. and estimated actual towing volume .in, This is the batch number.

[0082] The expression form of the over-immersion drag correction function is:

[0083] in, For the first Correction factor for the amount of towed goods in each batch; For the first Estimated actual towing volume for each batch; For the first Standard drag quantity parameters for each batch corresponding to the plating type; For the first The soaking time for each batch has exceeded the limit; For reference overimmersion time; For the first The viscosity of the plating solution for each batch corresponding to the plating type; For reference plating solution viscosity; For the first The surface tension of the corresponding plating type in each batch; For reference surface tension; To correct the amplitude coefficient; The nonlinear exponent of the superimmersion time; The viscosity-affecting index; This represents the surface tension influence index. The above reference values ​​and index parameters were obtained by fitting historical process test data.

[0084] It should be noted that the nonlinearity of the above-mentioned ultra-immersion drag amount correction function is reflected in the following: as the ultra-immersion time increases, the continuous growth of the coating on the workpiece surface leads to an increase in surface roughness. At the same time, the wettability of the plating solution on the workpiece surface is enhanced. The combined effect of these two factors causes the drag amount to increase at a rate greater than linear. Therefore, the ultra-immersion time term adopts a nonlinear exponent greater than 1. Fitting was performed. In addition, the higher the viscosity of the plating solution, the thicker the liquid film adheres, and the lower the surface tension, the more difficult it is for the liquid film to detach. Therefore, the viscosity term and the surface tension term enter the ultra-immersion drag correction function in a positive and negative correlation, respectively.

[0085] Furthermore, to improve the accuracy of towing amount correction, the towing amount correction coefficient is calculated... At the same time, the current plating solution temperature of the corresponding tank for each ultra-immersion batch is also obtained. The temperature influence factor is incorporated into the ultraimmersion drag correction function. The temperature influence factor is... The form is multiplied into the ultra-immersion drag correction function, where For the first The current plating solution temperature of each batch corresponds to the tank. For reference temperature, The temperature effect index reflects the combined impact of increased plating bath temperature leading to decreased viscosity and changes in surface tension on the drag volume.

[0086] Step 5: Generate the actual cleaning batch sequence and cross-contamination risk list;

[0087] Based on the estimated arrival time of each batch of equipment to be replaced by the crane in the cleaning section under the degraded operation scheduling scheme, the actual cleaning batch sequence is generated by sorting them in ascending order of arrival time. Each element in the actual cleaning batch sequence includes a batch number, a plating type identifier, and an estimated arrival time.

[0088] The cross-contamination sensitivity matrix between plating types is queried to obtain the cross-contamination sensitivity level and corresponding allowable residual concentration threshold between each pair of adjacent batches in the actual cleaning batch sequence. The aforementioned cross-contamination sensitivity matrix between plating types is a pre-stored two-dimensional lookup table. The rows and columns of the matrix correspond to the various plating types involved in the production line. The elements of the matrix record the quality impact level and corresponding allowable residual concentration threshold caused by the residual characteristic metal ions of the preceding plating type on the surface of the workpiece in the subsequent plating type. ,in This serves as a type identifier for the preceding plating type. This serves as a type identifier for subsequent plating processes. This is a characteristic metal ion identifier for the preceding plating type.

[0089] The cross-contamination sensitivity levels of adjacent batch pairs in the actual cleaning batch sequence are compared with the sensitivity levels of the same batch pairs in the original schedule. If the sensitivity level of a pair of adjacent batches in the actual cleaning batch sequence is higher than that of the pair in the original schedule, or if the pair of adjacent batches is not adjacent in the original schedule, then the pair of adjacent batches is marked as a new or changed cross-contamination risk item, and a cross-contamination risk list is generated after summarizing.

[0090] It should be noted that the cross-contamination sensitivity matrix between the above plating types is an asymmetric matrix, that is, the plating types... As a prelude, a type of plating As a subsequent sensitivity level, it is related to the plating type. As a prelude, a type of plating The sensitivity levels may differ in subsequent processes because different characteristic metal ions interfere with different plating processes to varying degrees.

[0091] Step 6: Perform mass conservation deduction on the countercurrent cleaning digital twin model based on the corrected parameters;

[0092] In the countercurrent cleaning digital twin model, the independent concentration monitoring values ​​of multi-component characteristic metal ions in each cleaning stage of the current cleaning tank are obtained as the initial state for simulation. The estimated actual drag volume of each ultra-immersion batch is then used. Replace the standard towing volume parameters originally used in the countercurrent cleaning digital twin model. .

[0093] Based on the actual cleaning batch sequence, mass conservation deductions were performed for each characteristic metal ion. For the [number]th batch... The first batch entered the first During the cleaning process in the first-stage cleaning tank, characteristic metal ions The concentration update calculation is as follows:

[0094] in, For the first After the first batch of cleaning is completed, the first... Characteristic metal ions in the cleaning tank The concentration; For the first The first batch entered the first... Characteristic metal ions in the cleaning tank The concentration; For the first Estimated actual towing volume for each batch; For the first Characteristic metal ions in the plating solution corresponding to each batch of plating type The concentration; For the first Effective volume of the cleaning tank; For the first The outflow rate of the cleaning tank; For the first The first cleaning tank to the second The counter-current water supply flow rate of the cleaning tank; For the first The batch in The cleaning duration in the cleaning tank; This refers to the batch number; Identifiers for characteristic metal ions; This refers to the number of stages in the cleaning tank. The value is derived from the actual residence time of this batch in the corresponding cleaning tank in the degraded operation scheduling plan. Different batches will have different values ​​due to different degraded operation scheduling plans. Values, This directly reflects the impact of scheduling timing on the predicted cleaning concentration.

[0095] The above formula is used to progressively extrapolate the cleaning process for each level of the cleaning tank and each batch, outputting the residual concentration distribution of each characteristic metal ion in each level of the cleaning tank after each batch is cleaned. .

[0096] It should be noted that the countercurrent water replenishment direction in the above countercurrent cleaning digital twin model is a step-by-step flow from the final cleaning tank to the first cleaning tank, i.e. This indicates that the overflow water from the subsequent tank enters the preceding tank. For the final cleaning tank, its makeup water source is fresh, pure water. ,in This represents the total number of stages in the cleaning tank.

[0097] Furthermore, to more accurately reflect the impact of over-immersion batches on the concentration in the cleaning tanks, the secondary dragging effect during batch transfer between different cleaning tank levels was also considered in the mass conservation derivation. Specifically, when the workpiece moves from the first... The first cleaning tank is lifted out and sent into the second cleaning tank. During the first cleaning tank, the amount of ions carried by the liquid film adhering to the workpiece surface is based on the current first cleaning tank. Concentration of the cleaning tank and towing volume The product of is estimated, and this ion quantity is included in the first... In the concentration update equation of the stage cleaning tank. Among them, For the workpiece from the first The secondary drag volume when the primary cleaning tank is removed. Use the same towing amount correction factor as the initial towing amount. Adjustments will be made.

[0098] Step 7: Generate emergency isolation flushing instructions and perform cross-module constraint transfer;

[0099] The predicted residual concentration distribution of each cleaning section Permissible residual concentration thresholds for subsequent batches in the cross-contamination risk list A comparison was performed. Specifically, for each risk in the cross-contamination risk list, the characteristic metal ions of the preceding plating type were obtained from each level of the cleaning tank after the previous batch of cleaning was completed. The predicted residual concentration is used to determine whether it exists. If such a situation exists, then the corresponding cleaning stage level will be adjusted. The corresponding time window is marked as the out-of-range segment.

[0100] For each exceeding the standard range, with the hard constraint of reducing the residual concentration of the target characteristic ions to below the allowable residual concentration threshold, a sequential quadratic programming algorithm is used to solve for the flushing flow rate required for enhanced flushing. and duration The sequential quadratic programming algorithm uses the total flushing water volume of each exceeding-standard cleaning section as the optimization objective, the constraint that the concentration of target characteristic ions in each exceeding-standard cleaning section does not exceed the allowable residual concentration threshold after flushing as the constraint, and the upper bound constraint that the maximum allowable flushing flow rate of each cleaning section is used as the upper bound constraint. The input is the set of exceeding-standard sections. The predicted residual concentration and allowable residual concentration threshold for each exceeding cleaning standard section are output as the optimal flushing flow rate for each exceeding cleaning standard section. and duration The mathematical expression of the above optimization problem is:

[0101] in, This is the set of all cleaning stages that exceed the standard. The elements in the table represent the stages of each excessive cleaning segment. ; To enhance the effect after rinsing Characteristic metal ions in the cleaning tank The concentration, according to the mass conservation equation... and Perform calculations for variables; For the first Flushing flow rate of the cleaning section exceeding the standard; For the first The rinsing duration of the cleaning section exceeding the standard; For the first The maximum permissible rinsing flow rate of the graded cleaning tank.

[0102] Specifically, The calculation method is as follows: during the flushing duration Inside, with flushing flow rate Continuing towards the first The first-stage cleaning tank is filled with pure water and an equal volume of tank liquid is simultaneously discharged. Based on a first-order decay model of continuous dilution, the rinsing is completed after the first stage... Characteristic metal ions in the cleaning tank The concentration is:

[0103] in, Before rinsing Characteristic metal ions in the cleaning tank The predicted residual concentration, For the first The effective volume of the cleaning tank is: This represents an exponential function. After solving, it outputs the emergency isolation flushing instructions for each exceeding the standard section, including the flushing flow rate. and duration .

[0104] After generating the emergency isolation flushing command, assess the time required for emergency isolation flushing of each affected section. Does the corresponding arrival time interval for adjacent batches appear in the degraded operation scheduling scheme? It can be completed within [timeframe]. If... If the time interval is insufficient, the constraint information is sent back to the degraded scheduling module, triggering a local scheduling adjustment. This local scheduling adjustment includes the following two processing methods: First, inserting a time interval of not less than [a certain value] between adjacent rescue batches involving highly sensitive combinations. First, the isolation waiting time; second, if inserting the isolation waiting time causes the over-immersion time of subsequent batches to exceed the acceptable range, then the subsequent batches will be reassigned to the standby cleaning section. After the adjustment is completed, the downgraded operation scheduling plan and emergency isolation flushing instructions will be updated.

[0105] It should be noted that the above-mentioned cross-module constraint transmission process is a single-round feedback process, that is, the degraded scheduling module determines the final scheme after making a local adjustment based on the returned time interval constraints. After the scheduling adjustment, the cleaning section pollution control module re-executes the deduction of steps 5 and 6 and the threshold comparison of step 7 with the updated degraded operation scheduling scheme to verify whether the adjusted degraded operation scheduling scheme meets all allowable residual concentration threshold constraints.

[0106] Furthermore, to achieve better global scheduling quality during cross-module constraint transmission, the aforementioned single-round feedback process is extended to an iterative feedback process. Specifically, after the pollution control module in the cleaning section re-executes the simulation and comparison, if there are still sections exceeding the standard, the constraint information is again transmitted back to the degraded scheduling module, which then performs further local scheduling adjustments. The iterative process has a maximum iteration limit; iteration terminates when the maximum iteration limit is reached or all allowable residual concentration threshold constraints are met, and the result of the last iteration is used as the final solution.

[0107] Step 8: Issue control commands and perform parameter correction after fault recovery;

[0108] The finalized degraded operation scheduling instructions are issued to the motion controllers of each alternative execution vehicle. These instructions include the target tank number, extraction time, and transfer path for each alternative execution vehicle. Emergency isolation flushing instructions are issued to the actuators of each excessive cleaning section. These instructions include the flushing flow rate for each excessive cleaning section. Duration and startup time.

[0109] After the fault is recovered, the measured ion concentration data of each ultra-immersion batch in each cleaning tank during the actual cleaning process are obtained, and the measured ion concentration data are compared with the predicted values ​​of the countercurrent cleaning digital twin model in step 6. Compare and calculate the prediction bias. ,in These are the measured concentration values.

[0110] Based on prediction bias The parameters in the ultra-immersion drag correction function , , and A correction and update are performed. Specifically, with the objective of minimizing the sum of squared prediction deviations for all ultra-immersion batches in each cleaning tank, the above parameters are refitted using the least squares method. The inputs to the least squares method are the ultra-immersion time, solution viscosity, surface tension, and corresponding prediction deviations for each ultra-immersion batch, and the output is the updated parameters. , , and Simultaneously, the effective volume of each cleaning tank in the countercurrent cleaning digital twin model was determined. By correcting the mixing efficiency parameters, the prediction accuracy of the countercurrent cleaning digital twin model can be continuously improved in subsequent fault emergency scenarios.

[0111] It should be noted that, in order to avoid excessive parameter shifts caused by data fluctuations from a single fault event during the above parameter correction and update process, constraints are imposed on the parameter update magnitude, namely, the relative change between the parameter value after each correction and the parameter value before correction does not exceed the preset upper limit of the change rate.

[0112] The intelligent control method for electroplating production lines based on digital twins provided in this embodiment overcomes the problem of underestimated cross-contamination risk caused by two factors in the original digital twin system: deviation of drag quantity parameters and lack of plating sequence information. This is achieved through cross-module linkage perception and bidirectional constraint transmission between the degraded scheduling module and the contamination control module of the cleaning section in emergency scenarios of gantry malfunction.

[0113] Specifically, the generation of segmented feasible spatiotemporal domains for the electroplating production line's digital twin system enables the system to quickly determine the available movement range of each normal crane based on track isolation constraints after a crane malfunction, thus providing accurate feasible spatial constraints for solving degraded operation scheduling schemes. The over-immersion drag amount correction function replaces the standard drag amount parameters used in the countercurrent cleaning digital twin model by incorporating over-immersion time, plating solution viscosity, and surface tension in a nonlinear form into the drag amount calculation, ensuring that the estimated total amount of metal ions entering the cleaning tank matches the actual surface state of the over-immersed workpiece. The generation of the actual cleaning batch sequence and the joint query of the cross-contamination sensitivity matrix between plating types identify adjacent combinations of highly sensitive plating types that were not present in the original schedule due to changes in the rescue order. The countercurrent cleaning digital twin model uses the corrected drag amount parameters and the actual cleaning batch sequence as inputs to perform mass conservation deduction, ensuring that the residual concentration prediction results reflect the actual contamination load distribution under fault scenarios. The bidirectional constraint transmission between the generation of emergency isolation rinsing instructions and the degraded operation scheduling scheme ensures that the emergency scheduling meets the contamination isolation time requirements of the cleaning section while guaranteeing the timeliness of overdue protection, avoiding potential timing conflicts that might occur when the two subsystems make independent decisions. Parameter correction and updates after fault recovery allow the parameters of the ultra-immersion dragging amount correction function and the mixed parameters of the cleaning tank to be continuously corrected based on measured data, thereby improving the prediction accuracy of the countercurrent cleaning digital twin model in subsequent similar fault scenarios. Therefore, this implementation method, through the above multi-step collaborative processing, enables the electroplating production line digital twin system to output globally coordinated and consistent degraded operation scheduling instructions and emergency isolation rinsing instructions in the event of a crane malfunction, reducing the possibility of underestimating the risk of cross-contamination in the cleaning section.

[0114] The following is an example of an application of the present invention, such as Figure 2-9 As shown, the implementation process is as follows:

[0115] A multi-plating mixed-line electroplating production line (hereinafter referred to as Line A) experienced a crane malfunction alarm at 09:27:43 on March 14, 20XX. Line A has a total track length of 48.0m and is equipped with 5 cranes (Crane 1 to Crane 5), currently handling mixed-line production of four plating types: nickel, chromium, zinc, and copper. The malfunctioning crane was Crane 3, with its stopping position at track 28.6m. Its physical dimensions (track occupies) are 1.2m, therefore the malfunctioning crane occupies the track section from 28.0m to 29.2m. Using this section as a boundary, the production line track is divided into two isolated operating sections: the left section covers 0 to 28.0m, and the right section covers 29.2m to 48.0m.

[0116] The digital twin system reads the real-time status of each train in normal operation at the moment the fault occurred. Train 1 and Train 2 are located in the left section, and Train 4 and Train 5 are located in the right section. Combining the rated operating speed and acceleration parameters of each train, segmented spatiotemporal feasible regions for train operation are generated for the left section (Train 1 and Train 2) and the right section (Train 4 and Train 5), respectively.

[0117] Table 1. Traffic status and section at the time of the fault.

[0118]

[0119] The faulty crane 3 was originally responsible for the tank area covering the track from 18.0m to 38.0m, involving a total of 6 batches in production (batch 1 to batch 6). The digital twin system read the soaking progress snapshot at 09:27:43, according to the formula... Calculate the remaining time for each batch. Taking batch 2 (chrome plating) as an example... s, s, then The value 's' indicates that the batch had been submerged for 53 seconds when the malfunction occurred, representing the duration of the over-immersion. Initialize to 53 seconds and continuously accumulate. After sorting by remaining time in ascending order, generate a timeout risk batch queue, with batch 2 at the front and having the highest priority.

[0120] Table 2. Snapshots of the soaking progress of the batch at the time of failure and ranking of timeout risks.

[0121]

[0122] The tank positions for each batch in production fall within either the left-side section (0.0–28.0 m) or the right-side section (29.2–48.0 m). The digital twin system determines the alternative operating vehicles for each batch based on the segmented spatiotemporal feasible domain. Tanks 06, 09, 11, and 12 are located in the left-side section and are handled by vehicles 1 and 2; tanks 14 and 17 are located in the right-side section and are handled by vehicles 4 and 5.

[0123] Following the greedy allocation in step 301, batch 2 (already over-immersed, ranked first) is prioritized. The current position of crane 2 is 19.7m, closest to tank 09 (located at 22.3m), with an estimated travel time of approximately 34 seconds. Crane 2 is assigned to retrieve batch 2 at 09:28:17. Subsequently, batch 5 (crane 4, retrieved at 09:28:52), batch 1 (crane 1, retrieved at 09:29:05), and so on, are allocated sequentially. After the initial allocation, step 302, a local exchange evaluation, is performed. After one round of traversal, no exchange is accepted, and the final degraded operation scheduling scheme is directly output. Weighting coefficients for each plating type: Chromium plating... Nickel plating galvanized Copper plating .

[0124] Calculate the timeout penalty using batch 2 as an example: Extracted at 94 seconds after the fault occurred. s, s, but s, the loss contributed by this batch is .

[0125] Table 3 Degraded Operation Scheduling Scheme

[0126]

[0127] Ultra-immersion batch 2 (chrome plating) s), batch 5 (nickel plated, s), Batch 1 (galvanized, s) Drag amount correction is required. The digital twin system reads the physical property parameters of the plating solution for each plating type from the pre-stored process database, and the reference parameters are determined by fitting historical process test data: s, mPa·s, mN·m⁻¹, °C, , , , , .

[0128] Taking batch 2 (chrome plating) as an example, after incorporating the temperature influence factor into the correction function, the correction coefficient is calculated by substituting each parameter:

[0129] Table 4. Calculation results of corrected towing volume for ultra-immersion batches

[0130]

[0131] Based on the estimated arrival time of each batch sent to the cleaning section in the downgraded operation scheduling plan, the actual cleaning batch sequence is generated as follows: Batch 2 (chrome plating) → Batch 5 (nickel plating) → Batch 1 (zinc plating) → Batch 4 (copper plating) → Batch 6 (nickel plating) → Batch 3 (zinc plating). The original schedule was: Batch 1 (zinc plating) → Batch 2 (chrome plating) → Batch 3 (zinc plating) → Batch 4 (copper plating) → Batch 5 (nickel plating) → Batch 6 (nickel plating).

[0132] The sensitivity matrix for cross-contamination between plating types was queried, and the sensitivity level of each pair of adjacent plating type combinations in the actual sequence was retrieved. Taking the "chromium plating → nickel plating" combination as an example, hexavalent chromium ions (… It has a strong inhibitory effect on the nickel layer crystallization process in nickel plating, with a sensitivity level of high (level 3) and an allowable residual concentration threshold. mg·L⁻¹; while the sensitivity level of the "chromium plating → zinc plating" combination in the original schedule was only medium (level 2). This highly sensitive combination was not adjacent in the original schedule and is a newly added risk item, which is recorded in the cross-contamination risk list.

[0133] Table 5 Actual Cleaning Batch Sequence and Cross-Contamination Risk Identification

[0134]

[0135] The cleaning section of production line A adopts a three-stage countercurrent cleaning structure. The effective volume of each cleaning tank is... L, water flow rate L·min⁻¹, the counter-current water supply flow rate is the same. At the time of the failure, in the first-stage cleaning tank... Initial monitoring concentration mg·L⁻¹, Level 2 is mg·L⁻¹, Level 3 is mg·L⁻¹.

[0136] Taking batch 2 (chrome plating) entering the first-stage cleaning tank as an example, substitute the mass conservation formula. Given: L, in the chromium plating solution concentration mg·L⁻¹, L, L·min⁻¹, L·min⁻¹, the residence time of batch 2 in the first-level cleaning tank in the degraded operation scheduling scheme. min:

[0137] The calculated result is far higher than the permissible residual concentration threshold. mg·L⁻¹ triggers subsequent over-limit segment identification and emergency flushing procedures.

[0138] The results of step 6 show that after batch 2 is cleaned, in the first-stage cleaning tank... The residual concentration was 0.1837 mg·L⁻¹, exceeding the threshold of 0.08 mg·L⁻¹, and the first-stage cleaning tank was marked as exceeding the standard. Maximum allowable flushing flow rate for each cleaning tank. L·min⁻¹, effective volume L.

[0139] Taking the first-stage cleaning tank as an example, the required rinsing parameters are calculated using a first-order attenuation model. The maximum rinsing flow rate is used. Substituting L·min⁻¹ into the solution, we find the solution that satisfies... Minimum flushing time required for mg·L⁻¹:

[0140] In the downgraded operation scheduling plan, the time interval between batch 2 (chrome plating) and batch 5 (nickel plating) arriving at the cleaning section s min, much shorter than the required rinsing time At min, cross-module constraint feedback is triggered. Upon receiving the constraint information, the degraded scheduling module inserts an isolation waiting time between batch 2 and batch 5, delaying the actual delivery time of batch 5 to the cleaning section to 09:43:00. This ensures that batch 5 is received only after the first-level cleaning tank completes its emergency flushing, and updates the degraded operation scheduling plan. The cleaning section contamination control module re-executes the simulation with the updated plan, verifying that the concentrations in each level of the cleaning tank meet the threshold constraints after the adjustment, and then the iteration terminates.

[0141] Table 6 Results of Emergency Isolation Flushing Commands and Cross-Module Constraint Transmission

[0142]

[0143] Production line A completed the repair and restoration of the faulty overhead crane 3 at 11:52:00 on March 14, 20XX. The digital twin system collected measured ion concentration data from three ultra-immersion batches in each cleaning tank during this emergency response, compared them with the predicted values ​​from step 6, and calculated the prediction deviation. Taking batch 2 in the first cleaning tank... For example: predicted value mg·L⁻¹, Measured value mg·L⁻¹, prediction bias mg·L⁻¹ indicates that the actual towing volume is slightly higher than the corrected estimate, and the parameters need to be adjusted. Adjust appropriately upwards.

[0144] All deviation data from the three batches in the tertiary cleaning tank were summarized, and the correction function parameters were refitted using the least squares method. The parameter comparison before and after correction is as follows: Updated from 0.38 to 0.41 (relative change of 7.9%, which does not exceed the preset change rate limit of 10%). Updated from 1.25 to 1.27. It remains unchanged from 0.60. The value has been updated from 0.45 to 0.47. Simultaneously, the effective volume of the first-stage cleaning tank... The calibration was changed from 800L to 792L (hybrid dead zone correction). The updated parameters were written into the digital twin system parameter library for use in subsequent fault emergency scenarios.

[0145] Table 7. Model parameter calibration results after fault recovery

[0146]

[0147] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

[0148] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only 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 noted that due to the limitations of textual expression, there are objectively infinite specific structures. For those skilled in the art, several improvements, modifications, or changes can be made without departing from the principles of the present invention, and the above technical features can also be combined in an appropriate manner. These improvements, modifications, changes, or combinations, or the direct application of the inventive concept and technical solution to other situations without modification, should all be considered within the scope of protection of the present invention.

Claims

1. A method for intelligent control of an electroplating production line based on digital twins, characterized in that, Includes the following steps: Receive the crane fault alarm signal, obtain the stopping position coordinates and physical size parameters of the faulty crane, divide the production line track space into multiple isolated operating sections, and generate the segmented spatiotemporal feasible domain of the available cranes in each isolated section; Get a snapshot of the soaking progress of all batches in production, calculate the remaining time of each batch before the maximum allowable soaking time limit, and generate a queue of batches with timeout risk by sorting them in ascending order of remaining time. With the goal of minimizing the total timeout loss of all in-process batches, a degraded operation scheduling scheme is generated under the constraints of the segmented driving time-space feasible region through a fast scheduling algorithm that combines greedy search and local exchange. Based on the superimmersion time of each superimmersion batch and the viscosity and surface tension parameters of the plating solution corresponding to the plating type, nonlinear correction calculation is performed using the superimmersion drag amount correction function to output the drag amount correction coefficient and the estimated value of the actual drag amount for each superimmersion batch. Based on the estimated arrival time of each batch entering the cleaning section in the downgraded operation scheduling scheme, the actual cleaning batch sequence is generated, and the cross-contamination sensitivity matrix between plating types is queried to generate a cross-contamination risk list. Replace the standard drag volume parameter in the countercurrent cleaning digital twin model with the estimated actual drag volume, perform mass conservation extrapolation of each characteristic metal ion according to the actual cleaning batch sequence, and output the residual concentration distribution of each cleaning tank. The residual concentration distribution is compared with the allowable residual concentration threshold. The flushing flow rate and duration of the exceeding section are calculated and an emergency isolation flushing command is generated. The flushing time constraint is sent back to the degraded scheduling module for cross-module constraint transmission. Issue downgraded operation scheduling instructions and emergency isolation flushing instructions, and after the fault is restored, correct and update the parameters of the ultra-immersion towing amount correction function based on the measured concentration data.

2. The intelligent control method for electroplating production lines based on digital twins according to claim 1, characterized in that, The generation of segmented spatiotemporal feasible regions for trains to travel within each isolation segment includes: In the spatial dimension, the movement range of each available crane is limited by the two ends of the isolated operating section, one end being the edge of the section occupied by the faulty crane, and the other end being the physical endpoint of the production line track or the edge of the adjacent fault point; in the temporal dimension, the positional distance between any two available cranes in the same isolated section at the same time is not less than half of the sum of their respective physical dimensions. Obtain the rated operating speed, acceleration parameters, and current load status of each normal operating vehicle. For each available vehicle, starting from its current position, calculate the farthest and nearest positions that the vehicle can reach at each time point according to the motion model of accelerating to the rated speed, running at a constant speed, and decelerating to a stop. Use the motion envelope as the boundary constraint curve of the vehicle in the segmented driving spatiotemporal feasible domain.

3. The intelligent control method for electroplating production lines based on digital twins according to claim 1, characterized in that, In the timeout risk batch queue, the remaining time for each batch is the maximum allowable immersion time limit for the corresponding plating type minus the immersion time already elapsed at the time of the failure. Batches with negative remaining time indicate that the maximum allowable immersion time limit has been exceeded at the time of the failure and are placed at the front of the timeout risk batch queue. The over-immersion time for each batch is initialized to zero at the moment of the fault and is continuously updated over time until the batch is extracted from the tank.

4. The intelligent control method for electroplating production lines based on digital twins according to claim 1, characterized in that, The execution process of the fast scheduling algorithm includes: According to the order of the timeout risk batch queue, each batch is assigned an alternative execution vehicle and a retrieval time. For the current batch, the vehicle that makes the retrieval time of the batch earliest is selected from its corresponding alternative execution vehicle set. The corresponding time period of the vehicle is marked as occupied and its remaining available time period is updated. After completing the initial allocation of all batches, the alternative execution of train allocation and extraction time of two batches with adjacent priorities are swapped in turn. The total timeout loss after the swap is calculated. If the total timeout loss after the swap is less than the total timeout loss before the swap, the swap is accepted and the degraded operation scheduling scheme is updated. Otherwise, the original allocation remains unchanged. Repeat the local swap evaluation until no swap is accepted in a complete round of traversal, and output the degraded operation scheduling scheme. The total timeout loss is the sum of the product of the plating type loss weight coefficient of each batch and the cumulative over-immersion time of that batch. In the local exchange assessment, the arrival order of each batch entering the cleaning section after the exchange is also estimated. If the exchange leads to an increase in the number of adjacent plating type combinations with high cross-contamination sensitivity, a penalty term is applied to the exchange scheme. The penalty term is the product of the penalty weight coefficient and the number of newly added highly sensitive adjacent combinations.

5. The intelligent control method for electroplating production lines based on digital twins according to claim 1, characterized in that, In the overimmersion drag amount correction function, the drag amount correction coefficient is equal to the product of the correction amplitude coefficient and the overimmersion time ratio, viscosity ratio, and surface tension ratio. Wherein, the superimmersion time ratio term is the superimmersion time nonlinear exponent of the ratio of the superimmersion time of this batch to the reference superimmersion time, the viscosity ratio term is the viscosity influence exponent of the ratio of the viscosity of the corresponding plating solution of this batch to the viscosity of the reference plating solution, and the surface tension ratio term is the surface tension influence exponent of the ratio of the reference surface tension to the surface tension of the corresponding batch. The estimated actual drag amount is equal to the product of the standard drag amount parameter of the corresponding plating type for that batch and the drag amount correction coefficient; the reference overimmersion time, reference plating solution viscosity, reference surface tension, correction amplitude coefficient, overimmersion time nonlinearity index, viscosity influence index and surface tension influence index are obtained by fitting historical process test data.

6. The intelligent control method for electroplating production lines based on digital twins according to claim 5, characterized in that, When calculating the drag amount correction coefficient, the current plating solution temperature of the corresponding tank for each ultra-immersion batch is also obtained, and the temperature influence factor is multiplied into the ultra-immersion drag amount correction function. The temperature influence factor is the temperature influence exponent of the ratio of the current plating solution temperature to the reference temperature of the corresponding tank for that batch.

7. The intelligent control method for electroplating production lines based on digital twins according to claim 1, characterized in that, In the mass conservation deduction, after the i-th batch is cleaned, the concentration of characteristic metal ion j in the m-th cleaning tank is equal to the concentration of the ion in the cleaning tank before the batch enters, plus the product of the estimated actual drag amount of the batch and the concentration of the ion in the corresponding plating solution of the batch divided by the effective volume of the cleaning tank, minus the product of the outflow rate of the cleaning tank and the cleaning duration divided by the effective volume and multiplied by the concentration before entering the tank, plus the product of the countercurrent replenishment flow rate of the next cleaning tank and the cleaning duration divided by the effective volume and multiplied by the concentration before entering the next tank; The final cleaning tank is replenished with fresh pure water, and the concentration of the next cleaning tank is zero. The cleaning duration of each batch is taken from the actual residence time of the batch in the corresponding cleaning tank in the downgraded operation scheduling scheme. In the process of mass conservation deduction, the secondary drag effect when the batch is transferred between cleaning tanks of each level is also calculated. When the workpiece is lifted from the current cleaning tank and sent to the next cleaning tank, the amount of ions carried by the liquid film attached to the workpiece surface is estimated by multiplying the concentration of the current cleaning tank and the secondary drag volume and included in the concentration update of the next cleaning tank. The secondary drag volume is adjusted using the same drag volume correction coefficient as the initial drag volume.

8. The intelligent control method for electroplating production lines based on digital twins according to claim 1, characterized in that, The process of determining the flushing flow rate and duration for the excessive section includes: Using a sequential quadratic programming algorithm, the optimization objective is to minimize the total flushing water volume of each excessive cleaning section. The constraint is that the concentration of target characteristic ions in each excessive cleaning section does not exceed the allowable residual concentration threshold after flushing. The upper bound constraint is the maximum allowable flushing flow rate of each cleaning section. The flushing flow rate and duration of each excessive cleaning section are then solved. The concentration of characteristic metal ions after rinsing is calculated based on a first-order decay model of continuous dilution. This concentration is equal to the exponential function value of the quotient obtained by multiplying the predicted residual concentration before rinsing by the product of the negative rinsing flow rate and the duration, and dividing by the effective volume of the cleaning tank at that stage. The cross-module constraint transmission includes: assessing whether the flushing duration of each exceeding standard section is completed within the arrival time interval of the corresponding adjacent batch in the degraded operation scheduling scheme; if the flushing duration is greater than the arrival time interval, the constraint information of insufficient time interval is sent back to the degraded scheduling module; an isolation waiting time of not less than the difference between the flushing duration and the arrival time interval is inserted between adjacent rescue batches involving highly sensitive combinations; or the subsequent batches are allocated to the standby cleaning section.

9. The intelligent control method for electroplating production lines based on digital twins according to claim 1, characterized in that, The cross-module constraint transmission is an iterative feedback process. After the downgraded scheduling module makes local scheduling adjustments, the cleaning section pollution control module re-executes the generation of the actual cleaning batch sequence and cross-contamination risk list, quality conservation deduction and threshold comparison with the updated downgraded operation scheduling scheme. If there are still sections that exceed the standard, the constraint information is sent back to the downgraded scheduling module for further adjustment. The iteration process sets a maximum iteration limit. The iteration terminates when the maximum iteration limit is reached or all allowable residual concentration threshold constraints are met. The correction update includes: acquiring the measured ion concentration data of each ultraimmersion batch in each cleaning tank during the actual cleaning process, calculating the prediction deviation between the measured concentration value and the predicted value of the countercurrent cleaning digital twin model, aiming to minimize the sum of squares of prediction deviations of all ultraimmersion batches in each cleaning tank, refitting the correction amplitude coefficient, ultraimmersion time nonlinear index, viscosity influence index and surface tension influence index in the ultraimmersion drag amount correction function using the least squares method, and correcting the effective volume and mixing efficiency parameters of each cleaning tank; Among them, the relative change between the parameter value after each correction and the parameter value before correction does not exceed the preset upper limit of the rate of change.

10. A digital twin-based intelligent control system for electroplating production lines, used to execute the digital twin-based intelligent control method for electroplating production lines as described in any one of claims 1 to 9, characterized in that, include: The fault perception and segment division module is used to receive the crane fault alarm signal, obtain the stopping position coordinates and physical size parameters of the faulty crane, divide the production line track space into multiple isolated operating segments, and generate the segmented spatiotemporal feasible domain of the available cranes in each isolated segment. The timeout risk assessment module is used to obtain a snapshot of the soaking progress of all batches in production, calculate the remaining time of each batch before the maximum allowable soaking time limit, and generate a timeout risk batch queue by sorting the batches in ascending order of remaining time. The degraded scheduling module is used to generate a degraded operation scheduling scheme under the constraints of the segmented driving time and space feasible region by using a fast scheduling algorithm that combines greedy search and local exchange, with the goal of minimizing the total timeout loss of all in-process batches. The drag amount correction module is used to perform nonlinear correction calculations based on the over-immersion time of each over-immersion batch and the viscosity and surface tension parameters of the plating solution corresponding to the plating type, using the over-immersion drag amount correction function, and outputs the drag amount correction coefficient and the estimated value of the actual drag amount for each over-immersion batch. The cleaning sequence and risk identification module is used to generate the actual cleaning batch sequence based on the expected arrival time of each batch entering the cleaning section in the degraded operation scheduling scheme, and to generate a cross-contamination risk list by querying the cross-contamination sensitivity matrix between plating types. The cleaning section pollution control module is used to replace the standard drag volume parameter in the countercurrent cleaning digital twin model with the estimated actual drag volume, perform mass conservation deduction of each characteristic metal ion according to the actual cleaning batch sequence, and output the residual concentration distribution of each cleaning tank. The emergency flushing and cross-module constraint transmission module is used to compare the residual concentration distribution with the allowable residual concentration threshold, solve the flushing flow rate and duration for the exceeding section and generate an emergency isolation flushing command, and send the flushing time constraint back to the degraded scheduling module for cross-module constraint transmission. The instruction issuance and parameter correction module is used to issue downgrade operation scheduling instructions and the emergency isolation flushing instructions, and to correct and update the parameters of the ultra-immersion towing amount correction function based on the measured concentration data after the fault is recovered.

Images