A cloud-edge collaboration-based electroplating equipment management method and system

Abstract

The present application relates to the technical field of industrial internet equipment management, in particular to a kind of electroplating equipment management method and system based on cloud edge cooperation, current density is collected and change trajectory is constructed, state jump frequency and direction feature are extracted by dividing interval, accordingly, the conductive state is marked as fluctuation or stable category, load current is synchronously collected and upper load interval is determined, when the equipment is in fluctuation category and load continuously remains in upper load interval, equipment number and time mark are extracted and electroplating equipment management result is generated.The present application extracts jump feature according to the interval expression and migration relationship of current density change trajectory, identifies the stability degree of conductive state in combination with the frequency and direction alternation law, correlates state feature and high load continuous distribution area, locks the key period of conductive fluctuation and load concentration synchronization, improves risk positioning accuracy, enhances the depth of equipment operation behavior analysis and the pertinence of management basis.

Description

Technical Field

[0001] This invention relates to the field of industrial internet equipment management technology, and in particular to a cloud-edge collaborative electroplating equipment management method and system. Background Technology

[0002] The field of industrial internet equipment management technology encompasses a comprehensive system for monitoring the operation of electroplating production equipment and its supporting process systems, recording production processes, and controlling equipment scheduling. It primarily targets rectifier power supplies, electroplating tanks, circulating jet systems, transmission mechanisms, and automatic loading / unloading devices in electroplating production lines. By deploying current acquisition units, jet flow detection devices, temperature acquisition devices, and equipment operation status acquisition interfaces, it continuously collects data on current density, linear velocity, tank solution circulation status, and sheet material conveying status during the electroplating process. Data transmission between the field control unit and the upper-level system is achieved through industrial communication methods. The system records equipment operation, stores process parameters, tracks production batches, and manages maintenance cycles. Simultaneously, it integrates field control logic to adjust the frequency conversion parameters of the jet pump, the speed of the transmission steel belt, and the positions of the anode and cathode shields, forming a complete technical system covering equipment access, data acquisition, communication transmission, operation recording, and control execution. This system has a clear application foundation in the standardized design, modular structure, and long-term storage of production records for electroplating equipment.

[0003] Among them, the cloud-edge collaborative electroplating equipment management method and system refers to the operation management of electroplating tanks, jet pumps, transmission steel belts, rectifier power supplies, anode and cathode shields, and automatic loading and unloading systems in vertical continuous electroplating equipment. This involves collecting current density values, jet flow parameters, tank temperature data, sheet material transmission speed, and equipment start / stop signals at the equipment side; and determining the current density range, jet flow interval, temperature upper and lower limits, and linear speed matching relationship at the edge side according to preset rules, generating record data for abnormal states, and sending the processed operation data to the cloud according to a communication protocol. In the cloud, the data is categorized, stored, and retrieved based on equipment number, production batch information, electroplating time parameters, and historical operation records. Maintenance tasks are scheduled for the rectifier power supply, jet pumps, and transmission mechanisms according to a preset maintenance schedule. Remote commands are used to adjust the jet pump frequency conversion parameters, transmission speed settings, and anode and cathode shield positions. The technical aspects involved include electroplating current acquisition, jet flow detection, temperature acquisition, transmission protocol setting, equipment number management, production record archiving, and remote parameter configuration processes.

[0004] Existing technologies focus on processing collected parameters and equipment operation information, emphasizing the parameter values ​​themselves and their correlation with production information. They pay less attention to the changing trajectory of the same parameter during continuous operation. When parameter values ​​remain within the normal range for a long period but exhibit continuous and recurring changes, it is difficult to effectively distinguish the relevant change processes. For example, when the conductivity fluctuates frequently and the load remains concentrated at a high level, the recorded results are still mainly based on discrete parameters, lacking a joint characterization of the continuity of change and the state of load concentration. This leads to a lag in the identification of abnormal signs, making it difficult to promptly detect operating phases with cumulative risk characteristics, thus affecting the precision of equipment management decisions. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a method and system for managing electroplating equipment based on cloud-edge collaboration. The technical solution is as follows: A cloud-edge collaborative electroplating equipment management method includes the following steps: S1: Collect the current density sampling value of the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, generate sampling time markers, bind the equipment number to form sampling records, collect them according to the equipment number and sort them according to the sampling time, associate adjacent sampling values ​​to form a continuous current density change trajectory, and generate equipment current time series dataset. S2: Based on the device current time series dataset, read the continuous change trajectory of current density and the rated parameters of the device, divide the current density intervals and number them, match the sampled values ​​in sequence to form an interval number sequence, identify the number change time and change direction, and form a set of conduction state jump feature; S3: Extract the number change time markers and directions from the set of conductive state jump features, sort them by device number to form a number change time sequence, count the change frequency and read the state judgment parameters. When the direction alternates and the frequency is greater than the threshold, it is marked as fluctuating; otherwise, it is marked as stable, thus forming an edge conductive state identifier set. S4: Read the device number and conductivity status mark in the edge conductivity status identifier set, obtain the load sampling value and time mark with the device number, form a load change trajectory, read the rated load parameter, divide the load interval and match the sampling value, form a load interval number sequence, and generate a device load distribution record set; S5: Based on the device load distribution record set, call the conductivity status marker, establish a corresponding relationship, determine the upper load interval, identify continuous distribution segments, and when the fluctuation category and number fall into the upper part, extract the device number and load sampling time marker to generate electroplating equipment management results.

[0006] A cloud-edge collaborative electroplating equipment management system, the system comprising: The sampling and collection module collects the current density sampling value of the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, generates a sampling time marker, binds the equipment number to form a sampling record, collects the data according to the equipment number and sorts it according to the sampling time, associates adjacent sampling values ​​to form a continuous current density change trajectory, and generates a device current time series dataset. The interval jump module, based on the device current time series dataset, reads the continuous change trajectory of current density and the device rated parameters, divides the current density intervals and numbers them, sequentially matches the sampled values ​​to form an interval number sequence, identifies the number change time and change direction, and forms a set of conduction state jump feature sets. The state discrimination module extracts the number change time markers and directions from the set of conductive state transition features, sorts them by device number to form a number change time sequence, counts the change frequency and reads the state judgment parameters. When the direction alternates and the frequency is greater than the threshold, it is marked as fluctuating; otherwise, it is marked as stable, forming an edge conductive state identifier set. The load mapping module reads the device number and conductivity status mark from the edge conductivity status identifier set, obtains the load sampling value and time mark with the device number, forms the load change trajectory, reads the rated load parameter, divides the load interval and matches the sampling value, forms the load interval number sequence, and generates a device load distribution record set. The high-load screening module, based on the equipment load distribution record set, calls the conductivity status marker, establishes a corresponding relationship, determines the upper load interval, identifies continuous distribution segments, and when the fluctuation category and number fall into the upper part, extracts the equipment number and load sampling time marker, and generates electroplating equipment management results.

[0007] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this invention, the current density change trajectory formed by continuous sampling is expressed in intervals, and the conductivity state transition features are extracted based on the interval number migration relationship. A time-series feature system reflecting the rhythm and trend of state changes is constructed. The stability of the conductivity state is identified by combining the frequency and direction of number changes, thereby improving the ability to identify hidden fluctuation behavior. At the same time, high-load continuous distribution areas are screened based on the load interval arrangement position, and the conductivity state features are correlated with the high-load continuous distribution features to form an identification result that takes into account both state fluctuation features and load aggregation features. This helps to pinpoint the key time period when conductivity fluctuations and load concentration occur simultaneously, improves the accuracy of risk positioning, and enhances the depth of equipment operation behavior analysis and the pertinence of management basis. Attached Figure Description

[0008] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a flowchart illustrating the acquisition process of S1 in this invention; Figure 3This is a flowchart illustrating the acquisition process of S2 in this invention; Figure 4 This is a flowchart illustrating the acquisition process of S3 in this invention; Figure 5 This is a flowchart illustrating the acquisition process of S4 in this invention; Figure 6 This is a flowchart of the acquisition process for S5 of the present invention. Detailed Implementation

[0009] The following detailed explanation illustrates the specific implementation methods: The basic implementation examples are as follows: Figure 1 As shown: A cloud-edge collaborative electroplating equipment management method includes the following steps: S1: Collect the current density sampling value at the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, and generate a sampling time stamp simultaneously. Bind the current density sampling value, sampling time stamp and device number obtained in the same sampling to form a sampling record. Collect the sampling records according to the device number, and arrange them in the order of sampling time under the same device number to form a current density sampling value sequence. Associate the current density sampling values ​​corresponding to adjacent sampling times in the same current density sampling value sequence to form a continuous current density change trajectory. Combine the corresponding device number and sampling time stamp to generate a device current time series dataset. The equipment current time series dataset includes equipment operation profiles, current change characteristics, conductivity characterization, and time series correlation information.

[0010] Please see Figure 2 Specifically, S1 includes: S11: Collect the current density sampling value at the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, and simultaneously generate a sampling time stamp. The edge acquisition node reads the device number and cloud-edge collaborative communication frame stamp, writes the current density sampling value, sampling time stamp, and device number obtained in the same sampling into the same sampling entry, verifies whether the three items are bound, and generates a sampling binding identifier; as detailed below: The current density sampling value at the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank is collected. The edge processing node obtains the analog voltage signal through the FL-2 type manganese copper shunt physical device connected in series on the cathode bus of the electroplating tank. The built-in ADS1256 type 24-bit high-precision analog-to-digital converter core chip is used as the existing acquisition technology. The weak voltage at both ends of the shunt is captured through the SPI serial bus at a hardware sampling frequency of 1000Hz. The captured 0 to 75mV analog voltage is processed in the firmware layer using a specific scaling factor formula. Converted to real-time current intensity, where The current conversion ratio factor representing amperes per millivolt. This represents the upper limit of the rated output current of the rectifier power supply, and is set to 2000A. The rated voltage drop of the shunt is 75mV. If the analog voltage measured by the current core chip is 37.5mV, the real-time current is accurately calculated to be 1000A through multiplication. The static value of the effective contact area of ​​the electrodes, 40dm², is read from the non-volatile memory EEPROM. 2 Divide the 1000A real-time current by 40dm 2 The current density sample value is 25.0 A / dm². 2 While completing the numerical conversion, the system calls the externally mounted DS3231 hardware real-time clock crystal oscillator to obtain the current Unix millisecond-level timestamp, synchronously generating a sampling time stamp in the format of 1717200000500. It retrieves a predefined cross-bit fixed string device number, such as RECT-001-CN, from the configuration area, and extracts the cloud-edge collaborative communication frame marker within the current lifecycle from the underlying receive buffer of the network communication protocol stack. This marker consists of a 16-bit version number, a 32-bit random sequence number, and an appended eight-bit cyclic redundancy check code, such as V1-A5B6C7D8. The system then writes the current density sample value 25.0, the sampling time stamp 1717200000500, and the device number RECT-001-CN into a continuous... A single sampling entry is formed in the memory address block, and a field integrity verification action is performed. By calling the memory pointer, the offset addresses corresponding to the four fields are checked in turn to see if there is a non-empty data stream. The verification weight factor is set to 0.25. If the current density sampling value is not empty, 0.25 is added. If the time stamp conforms to the decimal positive integer format, another 0.25 is added. If the device number conforms to the preset 10-character length, another 0.25 is added. If the CRC check code of the communication frame mark matches, the final 0.25 is added. When the sum of the verification weights is exactly equal to the logical baseline value of 1.00, it is determined that the three key contents and auxiliary contents in the sampling entry are logically bound. A one-byte hexadecimal identifier 0xFF is inserted into the header of the data packet as a sampling binding identifier.

[0011] S12: Based on the sampling entry corresponding to the sampling binding identifier, the device number, sampling time stamp, and cloud-edge collaborative communication frame stamp are called. The sampling entries are aggregated according to the device number, and the sampling entries under the same device number are arranged in order according to the sampling time stamp. An order stamp is written for the arrangement position to obtain the sampling sorting identifier; as detailed below: Based on the sampling entries corresponding to the sampling binding identifier, the edge processing unit uses a direct memory access (DMA) controller to read data packets with the 0xFF identifier in batches from the FIFO circular queue of the memory buffer. Using a preset packet bit offset parsing protocol, it extracts the device number RECT-001-CN with a field offset of 12 to 22 bytes, and simultaneously extracts the sampling timestamp with a field offset of 2 to 10 bytes, as well as the cloud-edge collaborative communication frame marker at the end of the packet. Using the device number as a hash index key, it physically categorizes the large number of mixed sampling entries, grouping all sampling entries belonging to the RECT-001-CN number into a temporary linear linked list storage space. It then calls in-situ quicksort logic to adjust the memory position of the entries in the linked list based on their time dimension. The timestamps of the first entry (1717200000500) and the second entry (1717200001500) are compared. If 1717200001500 is greater than 1717200000500, both entries are kept in the storage space. The physical order remains unchanged. If a backflow anomaly is found during the investigation where the subsequent sampling time is less than the previous sampling time, a memory pointer swap operation is performed to exchange the physical addresses of the two. If a logical conflict is encountered where two adjacent timestamps are exactly the same, it indicates that the clock is jittering or there is an inconsistency due to double triggering. In this case, an inconsistency correction operation is performed to forcibly increment the subsequent timestamp by 1 microsecond to avoid logical collisions. All 500 sampling entries in the linked list are traversed until the timestamps are arranged in a strictly monotonically increasing order. A 16-bit unsigned integer value incrementing from 00001 to 00500 is added to the metadata area of ​​each entry as a sequence marker. If the sampling timestamp of an entry is the 10th position in the sequence, its sequence marker is written as 0x000A. The continuity of the sorting is verified by checking whether the difference between two adjacent sequence markers in the sequence is always equal to 1. When all the collected entries are assigned a unique sequence index and the logical order is completely consistent with the time stamp growth direction, the status code 0xAA is written to the control field of the data frame, thus obtaining the sampling sorting identifier containing the time topology.

[0012] S13: Based on the sampling sorting identifier, retrieve the current density sample values ​​and sampling time markers under the same device number, sequentially associate adjacent current density sample values ​​according to the sampling time order, and synchronize the association relationship, device number, and sampling time marker to the cloud collaborative node to generate a device current time series set; specifically as follows: Based on the sampling sequence identifier, the edge communication gateway uses its built-in W5500 hardware network protocol stack chip to sequentially retrieve the current density sample value and its corresponding sampling time marker belonging to device number RECT-001-CN from the sorted linear linked list via the RJ45 Ethernet physical interface. It then reads the current density value of 25.0 A / dm², marked with sequence number 00001.2 The current density value of 25.2 A / dm² is associated with the sequence number 00002. 2 Establish a logical pointer association between the two; The time interval between these two is calculated as The milliseconds are used to encapsulate this association with the device number RECT-001-CN and the start sampling time marker to construct a publish message payload conforming to the MQTT transmission protocol. This encapsulated data frame is then sent to the static IP address 192.168.10.101 specified by the remote cloud collaborative node via the network physical layer interface. Upon receiving the network data packet, the cloud collaborative node stores it in a distributed time-series database using the device number as the primary key and the sampling time as the sorting key. It compares the sequence markers of two adjacent received packets. If the current received packet is marked 00005 and the previous packet is marked 00004, the transmission link is considered logically continuous. If an inconsistency is found where the sequence marker jumps directly from 00004 to 00006, it indicates a data packet loss at the network physical link layer. In this case, an inconsistency handling mechanism is immediately triggered, sending a dedicated retransmission request frame to the edge gateway. In cases of continuous data, the received current density sample values ​​are processed point-to-point according to the time axis. If occasional data loss occurs, a linear interpolation formula is applied. Complete the missing points, among which This represents the estimated current density value to be interpolated. and These represent the adjacent effective current density sample values, respectively. The target timestamp representing the lost point is used to locate the location and the arithmetic mean of the valid sampled timestamps before and after it is taken. and The timestamps representing the valid points before and after are used to record this dynamic vector trajectory. The above actions are repeated until all 500 sampling points in this batch have completed topological reorganization in the cloud database, and finally a device current timing set is generated in the cloud storage layer.

[0013] S2: Based on the continuous change trajectory of current density in the device current time series dataset, the current density sampled values ​​are read sequentially according to the sampling time marker. The device rated parameter record established before the acquisition is read according to the device number to obtain the lower limit, upper limit and number of intervals of rated current density. The numerical range between the lower limit and upper limit of rated current density is divided into multiple current density intervals at equal intervals according to the number of intervals. The intervals are assigned interval numbers in ascending order of current density. Each current density sampled value is matched to the corresponding current density interval in sequence to obtain the corresponding interval number and form an interval number sequence. Adjacent interval numbers are compared. When the interval number corresponding to the later sampling time is inconsistent with the interval number corresponding to the previous sampling time, the later sampling time marker is used as the number change time marker. The number change direction is determined according to the size relationship of the interval numbers before and after the change, forming a set of conduction state jump feature. The set of conduction state transition characteristics includes state transition characteristics, fluctuation trend characteristics, change activity characteristics, and state transition characterization.

[0014] Please see Figure 3 Specifically, S2 includes: S21: Based on the continuous current density change trajectory in the device current time-series dataset, the edge collaborative nodes sequentially read the current density sampled values ​​according to the sampling time marker, and read the device rated parameter records established before the data acquisition according to the device number. They then determine whether the device number corresponds to the lower limit of rated current density, the upper limit of rated current density, and the number of intervals, and generate a rated interval benchmark; specifically as follows: Based on the continuous current density variation trajectory in the device current timing dataset, the edge processing node retrieves a real-time data frame marked RECT-001-CN from the local asynchronous queue by executing the Ethernet interface driver. Using the data offset pointer, it precisely locates bytes 16 to 24 of the message and extracts the double-precision floating-point current density sample value of 25.0 A / dm². 2 The remote configuration synchronization protocol is started simultaneously. The high-speed key-value retrieval technology of the in-memory database Redis is used to access the device parameter mirror table in the distributed storage layer. The primary key RECT-001-CN is input, and the associated production batch process package is retrieved. The lower limit of the rated current density is parsed from the JSON format string of the process package. 0.0A / dm 2 ,in The minimum safe current density boundary value required for the electroplating process is represented by the upper limit value of the rated current density. 50.0 A / dm 2 ,in This represents the maximum allowable current density limit for this electroplating process, and reads the number of integer variable ranges. The set value is 20, of which This represents the total number of spatial intervals for discretizing the total flow velocity space. Data item integrity verification logic is executed. A memory comparison function is called to check whether the currently active device number RECT-001-CN is in the same logical link block as the three parameters retrieved and returned. The verification mask is set to 0x07. If the lower limit exists, a bitwise AND operation is performed on the first bit of the mask; if the upper limit exists, a bitwise AND operation is performed on the second bit; if the number of intervals is a positive integer greater than zero, a bitwise AND operation is performed on the third bit. The final mask result is compared with the baseline value 0x07 for numerical equality. When they are completely equal, it is determined that the three key operational constraint parameters under that device number have completed static binding and dynamic correspondence. The logic controller is called to encapsulate the above four parameters into a fixed-length 64-byte interval attribute control block, and a globally unique memory handle is allocated to this control block as a reference for subsequent calculations, generating the rated interval baseline.

[0015] S22: The edge collaboration node retrieves the lower limit of rated current density, the upper limit of rated current density, and the number of intervals from the rated interval reference. It then divides the range between the lower and upper limits of rated current density into equal intervals based on the number of intervals, assigns interval numbers in ascending order of current density, and sequentially matches the sampled current density values ​​to the corresponding current density intervals, generating an interval number sequence. The details are as follows: Use the lower limit of rated current density of 0.0 A / dm from the rated range reference. 2 The upper limit of rated current density is 50.0 A / dm³. 2 The number of intervals is 20. The edge processing nodes perform subtraction operations through a floating-point arithmetic unit and call the formula. Calculate the unit interval step width, where The uniform span boundary value represents each subdivided current density interval. The upper limit is 50.0. The lower limit is 0.0. Given 20 intervals, the unit interval span is calculated using the formula. Based on this calibration value, a contiguous array space is allocated in memory starting from address 0.0. A loop-stepping logic is used to calculate the numerical boundaries of each interval and establish a mapping table. [0.0, 2.5) is assigned interval number 1, [2.5, 5.0) is assigned interval number 2, and so on, mapping [22.5, 25.0) to interval number 10, and [25.0, 27.5) to interval number 11. To avoid ambiguity in the assignment of critical points, a left-closed, right-open interval discrimination criterion is uniformly adopted across the entire network, and this applies only to intervals equal to the upper limit value. Extreme sampled values ​​are uniformly and forcibly assigned to interval number 20. Sampled values ​​are read sequentially according to the sampling time marker 1717200000500, and the current density sampled value is set to 25.0 A / dm. 2 Substituting the boundary comparison logic, we perform greater than or equal to and less than comparisons. We determine that 25.0 exactly satisfies the logical condition that the value lies between the left boundary 25.0 and the right boundary 27.5 of interval 11. We extract the integer number 11 and read the sampled value of 25.2 A / dm² at the next time step 1717200001500. 2 The numerical boundary comparison determines that it still falls within the range and continues to match number 11. If the sampled value at the third time step is 27.6 A / dm², 2 If it exceeds the right boundary of interval 11, then the upward incremental traversal logic is executed to determine that it falls within the interval [27.5, 30.0) of number 12. The corresponding number 12 is extracted, and this series of numbers 11, 11, and 12 generated in the order of timestamps are encapsulated into a data packet payload with a timestamp index to generate an interval number sequence.

[0016] S23: Based on the interval numbering sequence, the cloud-based collaborative node reads the adjacent interval numbers according to the sampling time marker, compares the interval number corresponding to the later sampling time with the interval number corresponding to the previous sampling time, and when the two interval numbers are inconsistent, registers the later sampling time marker as the number change time marker. The direction of the number change is determined based on the relationship between the interval numbers before and after the change, forming a set of conductivity state transition features; specifically as follows: Based on the interval numbering sequence, the cloud-based collaborative node obtains the interval change data packets pushed from the edge side by listening to a specific Topic under the MQTT protocol. According to the sampling time markers 1717200000500 and 1717200001500, it extracts adjacent interval number values ​​11 and 12 from the data packet payload, performs a logical XOR operation to determine if they are consistent. If the XOR result is 0, it is determined that the number has not changed, and continues to extract the numbers 11 and 12 corresponding to the next pair of adjacent time points 1717200001500 and 1717200002500, performing a numerical subtraction formula. Calculate the interval difference, where Represents the range of change. This represents the interval number of the next sampling time and has a value of 12. The interval number representing the previous sampling time is 11. The calculated result is 1. The logic comparator determines that the result is not equal to 0, thus identifying an anomaly where the number of the next sampling time is inconsistent with the previous time. If, during the verification process, a significant disordered misalignment is found between adjacent numbers, such as a jump from 11 to 16, with a span greater than the preset step size of 3, it indicates that the transformer has experienced a sudden pulse surge interference. In this case, the inconsistency anomaly protection mechanism is triggered, marking this point as a 0x99 pulse fault code and suspending the direction determination. If the change is normal, the time stamp 1717200002500 is immediately written into the node of the sudden event record linked list as the number change time stamp. Simultaneously, the direction determination logic is started, calling the next number 12 and the previous number 11 to compare their magnitudes and calculate their difference. Since the difference is positive, the Boolean value "True" is written in the direction of change field to indicate that the conductivity level is on an upward trend. If the difference is negative, "False" is written to indicate a downward trend. The four key attributes, including device ID, number change time stamp 1717200002500, change direction True, and the difference in interval number before and after the change, are encapsulated into an edge telemetry and telematics message to form a conductivity state jump feature set.

[0017] S3: Extract the number change time stamp and number change direction from the conductivity state transition feature set. Collect the number change records corresponding to the same device according to the device number and arrange them according to the order of the number change time stamp to form the number change time sequence. Statistically count the number change frequency based on the number change time sequence. Extract the arrangement relationship of the number change direction in the number change time sequence. Read the stable frequency range and direction alternation judgment conditions from the state judgment parameters established before data collection. When the number change direction alternates between rising and falling within the state judgment time window, and the number change frequency is greater than the stable frequency threshold, mark the conductivity state of the corresponding device as the fluctuation category. When the number change frequency is less than or equal to the stable frequency threshold, or the number change direction does not alternate between rising and falling, mark the conductivity state of the corresponding device as the stable category. Output the conductivity state mark according to the device number to form an edge conductivity state identifier set. The edge conductivity status identifier set includes status category labels, stability levels, fluctuation risk identifiers, and device status indexes.

[0018] Please see Figure 4 Specifically, S3 includes: S31: Extract the time stamp and direction of number change from the conductivity state transition feature set. The edge collaboration node reads the device number and the cloud-edge collaboration transmission mark, gathers the corresponding number change records of the same device according to the device number, compares the order of the number change time stamps, and arranges the number change records in chronological order to generate the number change time sequence; as detailed below: The edge processing logic layer extracts the number change timestamps and number change directions from the conductivity state transition feature set. It then retrieves the feature packet containing device number RECT-001-CN by reading the hardware circular queue cache space of the dual-port RAM. Using the protocol parsing engine, it reads the cloud-edge collaborative transmission flag V1-A5B6C7D8 at offset address 0x04. All transition event records under the same device number RECT-001-CN are pulled from the memory mount point to a temporary sorting buffer. Finally, it calls the double-precision timestamp comparison logic to retrieve the change timestamp 1717200002500 of the first record and the timestamp of the second record. A numerical subtraction operation is performed on 1717200005500. If the difference is greater than 0, the subsequent record is determined to be later in the timeline. If network congestion causes out-of-order arrival of data packets at the transport layer, and the difference between adjacent timestamps is less than 0, a reordering mechanism is immediately triggered. This mechanism uses pointer interaction to quickly swap the physical storage locations of the two records to restore the correct time topology. The direction flag (True) for the jump from 11 to 12 is associated with the direction flag (False) for the jump from 12 back to 11. By traversing the 100 jump records in the entire buffer, the time interval between each pair of adjacent jump records is calculated using the formula. ,in This represents the time difference between two consecutive conduction transition events. This represents the latest change time stamp with a value of 1717200005500. The previous change time stamp is set to 1717200002500, and the calculated time interval is 3000 milliseconds. All jump records that have passed the time monotonicity check are chained together in an absolutely increasing sequence. Each linked list node contains a unique device identifier, the change time point, and a Boolean direction bit representing the rise or fall. A 32-bit cyclic redundancy check code is appended to the end of the data packet to verify the integrity of the sorted data stream, and a numbered change time sequence is generated.

[0019] S32: Based on the sequence of number changes, the system calls the number change time stamp, number change direction, and status determination time window for the same device number. The edge collaboration node counts the number change records within the window and determines whether the number change direction alternates between rising and falling according to the order of the number change time stamps to obtain the jump determination identifier; as follows: Based on the sequence of number changes, the number change time stamp, change direction, and status determination time window under the same device number RECT-001-CN are called. The edge processing node reads the status determination time window parameter setting of 60 seconds (60,000 milliseconds) from the configuration register. Taking the latest sampling timestamp 1717200060000 as the right boundary, it traces backward to the left boundary 1717200000000. The total number of number change records falling within this 60,000 millisecond interval is counted by the logic accumulator. The number change direction bit sequence within this time span is traversed, such as [True, False, True, False, True]. The XOR operation of adjacent direction bits is performed. The first True bit is XORed with the second False bit. Since the directions are inconsistent, a logic 1 is output. The second False bit is XORed with the third True bit to get a logic 1. If all eleven consecutive XOR results in the sequence are logic 1, the direction sequence is determined to meet the strict alternation rule of rising and falling, and the frequency calculation formula is called. Calculate the transition rate, where This represents the frequency of state transitions per minute. The total number of number changes captured within the window, with a value of 12. The state determination time window is converted to a floating-point value in minutes, with a value of 1.0 minute, and then substituted into the formula to calculate... The frequency is set to 12 times per minute. A memory comparison instruction is invoked to match the real-time frequency value of 12 with a preset stability threshold. If the total number of records is found to be 12 during the check, the error is considered. If there are 0 instances of inconsistency, it means that the rectifier power supply output is absolutely still and there is no external disturbance. In this case, the alternation mask is set to 0 and the high-frequency judgment is skipped. At the same time, the alternation feature of the direction sequence is encoded into a binary word mask. If it is completely alternating, the mask bit is set to 1. If there are two consecutive True and True increases, the mask bit is cleared to zero. The frequency calculation results, the direction alternation mask, and the window boundary timestamp set are encapsulated into a multi-dimensional feature vector to obtain the jump judgment identifier.

[0020] S33: Based on the jump judgment identifier, the cloud collaborative node reads the stable frequency threshold and direction alternation judgment condition established before data collection. When the number change frequency is greater than the stable frequency threshold and the number change direction alternates between rising and falling, it is marked as a fluctuation category; otherwise, it is marked as a stable category. Conductivity status markers are output according to the device number, forming an edge conductivity status identifier set; as detailed below: Based on the transition detection identifier, the cloud-based collaborative node extracts the stable frequency threshold by reading a pre-set XML configuration file. Alternating determination conditions for direction; Using historical big data statistical benchmark formula Calculate the decision boundary, where This represents the critical frequency threshold that distinguishes between a steady state and a fluctuating state. This represents the historical average frequency of background fluctuations during production line operation under no-load conditions, with a value of 3 times per minute. The standard deviation of the frequency under historical stable operating conditions is represented by a value of 1 time per minute; The stable frequency threshold is obtained by substituting the values ​​into the calculation example. The system retrieves the real-time frequency value of 12 times / minute from the transition judgment flag, performs a value comparison, and determines that 12 is greater than 5. Simultaneously, it retrieves the direction alternation mask bit. If the mask bit is 1, the direction alternation judgment condition is met. A logical AND operation is performed to check if both conditions are true simultaneously. That is, when (12>5) and (direction alternation == True), the corresponding status bit of RECT-001-CN is written to the device status register as hexadecimal code 0x03, marking it as a fluctuation category. If the frequency statistics value is only 2 times / minute, then 2 is determined to be less than or equal to 5. If the direction sequence is [True, True, False] causing inconsistent alternation, the status bit is written to hexadecimal code 0x01 to mark it as a stable category. The generated conductivity status tag is concatenated with the device number RECT-001-CN to generate a JSON data format message conforming to the IoT transmission specification. The status field contains a numerical fluctuation level. This result set is pushed to the global status monitoring tablespace, summarized according to the device physical topology, and distributed to each network monitoring terminal to form an edge conductivity status identifier set.

[0021] S4: Read the device number and corresponding conductivity status mark in the edge conductivity status identifier set. Use the device number as an index to obtain the load current sampling value and load sampling time mark of the corresponding electroplating rectifier power supply. Form the load current change trajectory according to the load sampling time sequence. Read the device rated load parameter record established before the acquisition according to the device number to obtain the lower limit value, upper limit value and number of load intervals of the rated load current. Divide the value range between the lower limit value and the upper limit value of the rated load current into multiple load intervals at equal intervals according to the number of load intervals. Assign load interval numbers in order from low to high. Match the load current sampling value to the corresponding load interval to obtain the load interval number corresponding to the load sampling time mark. Form the load interval number sequence according to the load sampling time sequence to generate the device load distribution record set. The equipment load distribution record set includes load level distribution, load concentration degree, load duration characteristics, and load range coverage characteristics.

[0022] Please see Figure 5 Specifically, S4 includes: S41: Read the device number and corresponding conductivity status marker from the edge conductivity status identifier set. The edge collaborative node writes the device number into the load acquisition index, obtains the corresponding electroplating rectifier power supply load current sampling value and load sampling time marker, compares the order of the load sampling time markers, and associates the load current sampling values ​​in time sequence to generate a load current change trajectory; specifically as follows: The edge processing node reads the device number and corresponding conductivity status flag from the edge conductivity status identifier set. It then extracts the device identifier string RECT-001-CN by accessing a specific address space in its internal static random access memory and obtains the hexadecimal status bit 0x03 (fluctuation category flag) determined in step S33. This device number is used as a hash index key and passed to the real-time data routing table. The underlying hardware driver is then called to access the DHK-3000 open-loop Hall current sensor physical device, which is clipped onto the positive DC output cable of the rectifier power supply, as an existing load data acquisition technique. The device outputs a corresponding 0 to 5V analog voltage signal and sends it to the hexadecimal successive approximation ADC channel. The sensor's rated transformation ratio is set to 600. If the current measured voltage value is 2.6V, it is accurately converted into a 1560.0A load current sampling value through multiplication. The load sampling time mark 1717200000600 generated by the same DS3231 hardware real-time clock crystal oscillator via external hardware SPI timing logic is read synchronously to ensure the consistency of the acquisition link clock source. This time mark is synchronized and aligned with the current density time reference in the previous S11 step. Perform the subtraction calculation formula ,in This represents the clock time offset between the two sampling chains. This represents the current load sampling time stamp with a value of 1717200000600. The current density time base is represented by the value 1717200000500. The calculated offset is 100 milliseconds. The maximum allowable tolerance for synchronization is set to 150 milliseconds. If 100 milliseconds is less than 150 milliseconds, the load data is considered valid. If the calculated offset is greater than 150 milliseconds, it indicates a synchronization failure between two independent hardware clocks. This triggers a synchronization exception interrupt, discards the load data, and simultaneously broadcasts a clock calibration to the network. The system then continuously retrieves sampling packets from the local circular buffer to obtain the load current value for the next time step. The values ​​565.0A and their timestamps 1717200001600 are compared using a logic comparator. The current timestamp is extracted and subtracted from the previous timestamp. If the result is positive, the time axis is determined to be monotonically increasing. The values ​​1560.0A, 1565.0A, etc., are written into a dynamically allocated memory buffer in chronological order. A pointer variable pointing to the physical address of the next sampling point is inserted into the data unit of each sample value. By traversing the memory pointer address chain and performing data connectivity verification, the load current change trajectory is generated.

[0023] S42: Based on the load current change trajectory, the device number and cloud-edge collaborative transmission marker are retrieved. The device's rated load parameter record established before data acquisition is read according to the device number. The lower limit of the rated load current, the upper limit of the rated load current, and the number of load intervals are obtained. The correspondence between the device number and the rated load content is determined, and a load interval identifier is generated. Details are as follows: Based on the load current change trajectory, the device ID RECT-001-CN and the cloud-edge collaborative transmission marker V1-A5B6C7D8 are invoked. The edge processing node executes a network security transmission protocol to initiate an HTTP GET request to the high-performance Apache CouchDB non-relational database cluster deployed in the cloud. The device's rated parameter document in XML format is retrieved via a TLSv1.3 encrypted network channel. A string parser is used to locate the specific load parameter node and extract the lower limit of the rated load current from the fixed configuration items. It is 0.0A, where The minimum base no-load current value represents the design operating current of the rectifier power supply, from which the upper limit value of the rated load current is extracted. It is 3000.0A, of which This represents the maximum load current limit that the transformer is allowed to operate safely for a long period of time, and the number of intervals read. There are 12, of which This represents the total number of segmented spaces used to segment and label the total operating power space. The parameter integrity comparison logic is executed, retrieving the currently active device ID (RECT-001-CN) and performing a character-by-character XOR operation with the primary key field in the document. If the result is zero, it is determined that the device ID and parameter record have a logical correspondence. Simultaneously, the data verification process is initiated, summing the binary raw data of the three parameters to obtain a checksum value. This checksum value is then compared with the checksum field at the end of the document for equality. If the two are equal, it is confirmed that the parameters did not undergo bit flipping during network transmission, and the extracted parameters are validated. The system performs a filtering process to determine whether the upper limit value of 3000.0 is greater than the lower limit value of 0.0, and whether the number of intervals (12) is within the preset reasonable integer range of [1, 100]. If both results are true, the system calls the interval division function to segment the full range. The lower limit value, upper limit value, number of intervals, and corresponding device number are encapsulated into a parameter description frame with a length of 128 bytes. A hexadecimal control word 0x5A is inserted at the frame header as a guide identifier. A unique global resource descriptor is assigned to this set of parameters. The parameter description frame is then sent to the load interval calculation buffer via the internal bus to generate a load interval identifier.

[0024] S43: Based on the load interval identifier, retrieve the load current sample value and load sampling time marker from the load current change trajectory, match the load current sample value to the corresponding load interval, obtain the load interval number corresponding to the load sampling time marker, arrange them in order of load sampling time, and synchronize them to the cloud collaborative node to generate a device load distribution record set; specifically as follows: Based on the load interval identifier, the load current sample value of 1560.0A and the load sampling time marker 1717200000600 from the load current change trajectory are retrieved. The edge processing node executes the equidistant division logic through the floating-point arithmetic unit and calls the interval calculation formula. The interval span of the load sub-interval is calculated, where This represents the precise current step value for each load sub-interval. The upper limit of the rated load current is 3000.0A. Given the lower limit of the rated load current of 0.0A and the number of intervals of 12, substituting these values ​​into a specific calculation example yields the following result. A. Based on this width, construct a range mapping table, mapping [0.0, 250.0) to load range 1, [250.0, 500.0) to load range 2, and so on, mapping [1500.0, 1750.0) to load range 7. To prevent classification confusion caused by overlapping critical values, the left-closed, right-open assignment principle is also applied globally, and it is additionally stipulated that when the load current sample value is exactly equal to the upper limit value... If the current load is forced to be assigned to load interval 12, the current load current sample value of 1560.0A is retrieved sequentially. An interval range matching operation is performed to determine if 1560.0A is greater than or equal to 1500.0A and less than 1750.0A. If the result is true, the corresponding load interval number 7 is extracted. Then, the sample value of 1565.0A corresponding to the next time interval 1717200001600 is extracted. After comparison, it is determined that it still belongs to interval number 7. The obtained interval numbers are arranged in ascending order according to the load sampling time marker, constructing a structure such as [171720000]. The key-value pair sequence [0600:7,1717200001600:7] is encapsulated into the data field of a TCP packet using the IoT transmission protocol. The static IP address of the cloud collaborative node, 192.168.10.101, is filled into the packet header. The data stream is pushed to the cloud collaborative node through the physical network card. After receiving the packet, the cloud collaborative node performs an unpacking action and inserts the data items in the sequence into the Device_Load table of the distributed load record database. After the data of all time nodes is entered into the database, a device load distribution record set is generated.

[0025] S5: Based on the load sampling time stamp and load interval number sequence in the equipment load distribution record set, read the corresponding conductivity status mark in the edge conductivity status mark set according to the equipment number, establish the correspondence between the load interval number sequence and conductivity status mark of the same equipment, determine the load interval in the second half of the arrangement as the upper load interval according to the order of load interval number from low to high, identify the continuous distribution segment of load interval number falling into the upper load interval along the load sampling time sequence, when the conductivity status mark of the corresponding equipment is of the fluctuation type, and the load interval number remains in the upper load interval within the continuous distribution segment, extract the corresponding equipment number and the load sampling time stamp within the continuous distribution segment, and generate the electroplating equipment management result; The management results for electroplating equipment include key monitoring targets, periods of high load fluctuation, operational risk indicators, and management priority information.

[0026] Please see Figure 6 Specifically, in S5, it includes: S51: Based on the load sampling time stamp and load interval number sequence in the device load distribution record set, the edge collaboration node reads the corresponding conductivity status mark in the edge conductivity status identifier set according to the device number, verifies the correspondence between the load interval number sequence under the same device number and the conductivity status mark, and establishes a load status correspondence table; as detailed below. Based on the load sampling timestamps and load interval number sequences in the device load distribution record set, the edge processing node executes the network data exchange protocol and uses high-performance gRPC remote procedure call technology based on the underlying HTTP / 2 protocol as the existing data communication extraction method. It remotely retrieves the time-series data stream from the Device_Load table in the distributed database, inputs the primary key RECT-001-CN, and extracts the hexadecimal conductive status flag 0x03 written in step S33 using the offset pointer. A buffer with address alignment is allocated in the local memory mapping area. The load sampling timestamp sequence containing timestamps such as 1717200000600 is associated bit-by-bit with the corresponding load interval number 7. A data synchronization verification action is performed. The Hamming distance between the device number string in the load data packet and the device number string in the status identifier set is calculated by calling the memory comparison operator. If the calculated Hamming distance is greater than 0, it indicates a serious cross-slot error in network routing. The data inconsistency rejection logic is immediately triggered, the current load entry is discarded, and a log is written to the central log service layer. Error status code 0x404 is entered. If the distance is 0, the number is considered to be completely consistent. The time alignment logic is started synchronously. The load sampling time stamp 1717200000600 is retrieved and the corresponding current density timestamp 1717200000500 in the previous sequence is subtracted. The clock synchronization offset difference on the physical time axis is calculated and the nearest time node is found in the cache index. The alignment tolerance threshold is set to 500 milliseconds. If the measured difference of 100 milliseconds is within the range [0, 500], the time base is considered to be successfully matched. Each group [ The load interval number: status flag] is encapsulated into a new specific format data unit through logical addition. For each device number, 500 sampling points are traversed in a loop to check if there is a coexisting item with status flag 0x03 and number 7 in the sequence. A 16-bit composite index identifier is assigned to each pair of successfully matched data items, and a logical association matrix is ​​established in the data table space with device number as the first-order index and timestamp as the second-order index. The current memory operation chain is closed by writing the end character 0xED, and a load status correspondence table is established.

[0027] S52: Retrieve the load interval number sequence, conductivity status flag, and cloud-edge collaborative transmission flag from the load status correspondence table. Determine the second half of the arrangement positions according to the load interval numbers in ascending order. Filter the load interval numbers falling into the second half of the arrangement positions. Merge consecutive number positions along the load sampling time sequence to obtain the upper load continuous segment; details are as follows: The edge processing node retrieves the load interval number sequence, conductivity status flag, and cloud-edge collaborative transmission flag V1-A5B6C7D8 from the load status mapping table, and reads the total number of load intervals from the configuration register. This integer value is used to perform interval division boundary calculation. Calling the dangerous load determination formula Determine the starting position of the upper high-risk load zone, where Integer access boundary index number representing the high-load operating range. The total number of intervals is 12. This represents the floor function in mathematics, calculated by substituting 12 into the formula. If the item with the interval number in the range [7,12] is determined to belong to the high-risk, high-load segment in the second half of the arrangement, then the numerical filtering operator is activated to extract each number from the load interval number sequence [7,7,8,6,7]. Perform logical comparison operations The matching action under the option category determines that numbers 7, 7, and 8 meet the conditions, while number 6 does not. If, during scanning, inconsistent illegal data exceeding the rated range of 12, such as number 13, is found, indicating a memory overflow error, the interception mechanism is triggered to directly block the pointer and reset the current pointer offset. The numbers that meet the conditions and their corresponding timestamps 1717200000600 and 1717200001600 are stored in a temporary stack. Time continuity merging processing is performed, retrieving two adjacent timestamps that meet the conditions and performing a subtraction operation. If the difference is equal to the basic sampling period of 1000 milliseconds, it is determined to be physically continuous, and 1717200000600 and 1717200001600 are merged into a continuous time segment descriptor. The break tolerance coefficient is set to 1.5. If the difference between adjacent times is greater than 1.5, the time segment descriptor is not allowed to break. If the time interval is milliseconds, the merging of the current segment is terminated and a new segment record is started. The entire stack space is traversed, and the scattered number positions are converted into length information with a start timestamp of 1717200000600 and an end timestamp of 1717200010600. A hexadecimal identifier 0x7F is inserted at the beginning of each merged segment to represent high load characteristics. The total number of sampling points contained in the continuous segment is counted and its weight ratio in the total period is calculated. When the ratio is greater than the preset baseline value of 0.1, it is marked as an active segment, and the upper load continuous segment is obtained.

[0028] S53: Based on the continuous upper load segment, the cloud-based collaborative node reads the conductivity status marker corresponding to the same device number, determines whether the conductivity status marker is a fluctuation category, and verifies whether the load interval number within the continuous distribution segment remains within the upper load interval. It then extracts the device number and the load sampling time marker within the continuous distribution segment to generate the electroplating equipment management result; specifically as follows: Based on the upper load continuum, the cloud-based collaborative node retrieves the data load containing Device_ID RECT-001-CN from the edge side by executing a remote procedure call instruction. It parses the conductivity status flag field in the data load and extracts its value 0x03. A logical judgment instruction is invoked to compare the extracted value 0x03 with the preset fluctuation category benchmark value 0x03 for numerical equality. When the output logic is true, the second-level verification protocol is executed. The timestamp interval [1717200000600, 1717200010600] within the upper load continuum is retrieved. The distributed load database is accessed to retrieve the load interval number corresponding to each millisecond within this interval. A full matching operation is performed to determine whether all load interval numbers within the interval are consistently greater than or equal to 7. If a sudden drop in the number to 5 is detected at time point 1717200005600, the node will be flagged. Anomalies, indicating a deliberate reduction in load during production or a sudden circuit break in the plating solution, immediately trigger an inconsistent termination mechanism. This forcibly truncates the effective topological boundary of the current continuous segment and removes the timing items following the abrupt change point. Simultaneously, a 0xCC load reduction status code is written to the database. If the numbers corresponding to all 50 sampling points within the interval are consistently within the range of [7,12], the device is determined to be in a high-risk state of conductivity fluctuation under high load. The device number RECT-001-CN, along with the corresponding start time 1717200000600 and end time 1717200010600, are extracted. This information is then encapsulated according to the IoT management message format. Descriptive codes such as "Category:Fluctuation,Load_Level:High" are written into the message body, and the cumulative integral formula for power consumption efficiency is invoked. Calculate the total operating energy consumption during this abnormally high-risk period, of which Represents the total electrical charge consumed within a high-risk continuous segment, expressed in ampere-seconds. This is the instantaneous load current sample value, currently input as 1560.0A. The discrete sampling step time interval is currently set to 1.0 seconds. If the integral value exceeds the preset safety boundary threshold, an early warning level code 0xEE will be added to the management result. The result will be distributed to the management terminal display unit through a network publish / subscribe mode to generate the electroplating equipment management result.

[0029] A cloud-edge collaborative electroplating equipment management system, the system comprising: The sampling and collection module collects the current density sampling value of the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, generates a sampling time marker, binds the equipment number to form a sampling record, collects the data according to the equipment number and sorts it according to the sampling time, associates adjacent sampling values ​​to form a continuous current density change trajectory, and generates a device current time series dataset. The interval jump module, based on the equipment current time series dataset, reads the continuous change trajectory of current density and the equipment rated parameters, divides the current density intervals and numbers them, sequentially matches the sampled values ​​to form an interval number sequence, identifies the time and direction of number change, and forms a set of conduction state jump feature; The state discrimination module extracts the number change time markers and directions from the set of conductive state jump features, sorts them by device number to form a number change time sequence, counts the change frequency and reads the state judgment parameters. When the direction alternates and the frequency is greater than the threshold, it is marked as fluctuating; otherwise, it is marked as stable, forming an edge conductive state identifier set. The load mapping module reads the device number and conductivity status mark from the edge conductivity status identifier set, obtains the load sampling value and time mark with the device number, forms the load change trajectory, reads the rated load parameters, divides the load interval and matches the sampling value, forms the load interval number sequence, and generates a device load distribution record set. The high-load screening module, based on the equipment load distribution record set, calls the conductivity status marker, establishes a corresponding relationship, determines the upper load range, identifies continuous distribution segments, and when the fluctuation category and number fall into the upper range, it extracts the equipment number and load sampling time marker to generate electroplating equipment management results.

[0030] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A cloud-edge collaborative electroplating equipment management method, characterized in that, Includes the following steps: S1: Collect the current density sampling value of the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, generate sampling time markers, bind the equipment number to form sampling records, collect them according to the equipment number and sort them according to the sampling time, associate adjacent sampling values ​​to form a continuous current density change trajectory, and generate equipment current time series dataset. S2: Based on the device current time series dataset, read the continuous change trajectory of current density and the rated parameters of the device, divide the current density intervals and number them, match the sampled values ​​in sequence to form an interval number sequence, identify the number change time and change direction, and form a set of conduction state jump feature; S3: Extract the number change time markers and directions from the set of conductive state jump features, sort them by device number to form a number change time sequence, count the change frequency and read the state judgment parameters. When the direction alternates and the frequency is greater than the threshold, it is marked as fluctuating; otherwise, it is marked as stable, thus forming an edge conductive state identifier set. S4: Read the device number and conductivity status mark in the edge conductivity status identifier set, obtain the load sampling value and time mark with the device number, form the load change trajectory, read the rated load parameters, divide the load interval and match the sampling value to form the load interval number sequence, and generate the device load distribution record set.

2. The electroplating equipment management method based on cloud-edge collaboration according to claim 1, characterized in that: The device current time series dataset includes device operation profile, current change characteristics, conductivity intensity characterization, and time series correlation information. The conductivity state jump feature set includes state transition characteristics, fluctuation trend characteristics, change activity characteristics, and state transition characterization. The edge conductivity state identifier set includes state category labels, stability level, fluctuation risk identifier, and device state index. The device load distribution record set includes load level distribution, load concentration degree, load continuity characteristics, and load range coverage characteristics.

3. The electroplating equipment management method based on cloud-edge collaboration according to claim 1, characterized in that: The steps for obtaining S1 are as follows: S11: Collect the current density sampling value of the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, and generate a sampling time stamp simultaneously. The edge acquisition node reads the device number and cloud-edge collaborative communication frame mark, writes the current density sampling value, sampling time stamp, and device number obtained in the same sampling into the same sampling entry, checks whether the three items are bound, and generates a sampling binding identifier. S12: Based on the sampling entry corresponding to the sampling binding identifier, call the device number, sampling time marker and cloud-edge collaborative communication frame marker, collect the sampling entries according to the device number, arrange the sampling entries under the same device number according to the order of the sampling time marker, and write the order marker for the arrangement position to obtain the sampling sorting identifier; S13: Based on the sampling sorting identifier, call the current density sampling value and sampling time marker under the same device number, associate adjacent current density sampling values ​​in sequence according to the sampling time, and synchronize the association relationship, device number and sampling time marker to the cloud collaborative node to generate a device current time sequence set.

4. The electroplating equipment management method based on cloud-edge collaboration according to claim 1, characterized in that: The steps for obtaining S2 are as follows: S21: Based on the continuous current density change trajectory in the device current time series dataset, the edge collaborative node reads the current density sampling value sequentially according to the sampling time marker, reads the device rated parameter record established before the acquisition according to the device number, and determines whether the device number corresponds to the lower limit of rated current density, the upper limit of rated current density, and the number of intervals, and generates the rated interval benchmark. S22: Call the lower limit of rated current density, the upper limit of rated current density, and the number of intervals in the rated interval reference. The edge collaboration node divides the range between the lower limit of rated current density and the upper limit of rated current density equally according to the number of intervals. It assigns interval numbers in order of current density intervals from low to high. It then matches the current density sampled values ​​to the corresponding current density intervals in sequence to generate an interval number sequence. S23: According to the interval numbering sequence, the cloud collaborative node reads the adjacent interval numbers according to the sampling time marker, compares the interval number corresponding to the next sampling time with the interval number corresponding to the previous sampling time, and when the two interval numbers are inconsistent, registers the next sampling time marker as the number change time marker, determines the number change direction according to the size relationship of the interval numbers before and after the change, and forms a set of conductivity state jump feature.

5. The electroplating equipment management method based on cloud-edge collaboration according to claim 1, characterized in that: The steps for obtaining S3 are as follows: S31: Extract the number change time marker and number change direction from the conductivity state transition feature set. The edge collaboration node reads the device number and cloud-edge collaboration transmission marker, collects the number change records corresponding to the same device according to the device number, compares the order of the number change time markers, arranges the number change records in chronological order, and generates the number change time sequence. S32: Based on the number change time sequence, call the number change time marker, number change direction and status judgment time window under the same device number, the edge collaboration node counts the number change record occurrences in the window, and judges the alternating arrangement of number change direction of rising and falling according to the order of number change time markers to obtain the jump judgment identifier; S33: According to the jump determination identifier, the cloud collaborative node reads the stable frequency threshold and direction alternation determination condition established before the acquisition. When the number change frequency is greater than the stable frequency threshold and the number change direction is alternating between rising and falling, it is marked as a fluctuation category; otherwise, it is marked as a stable category. The conductive state mark is output according to the device number to form an edge conductive state identifier set.

6. The electroplating equipment management method based on cloud-edge collaboration according to claim 1, characterized in that: The steps for obtaining S4 are as follows: S41: Read the device number and corresponding conductivity status mark in the edge conductivity status identifier set. The edge collaboration node writes the device number into the load acquisition index, obtains the corresponding electroplating rectifier power supply load current sampling value and load sampling time mark, compares the order of the load sampling time marks, associates the load current sampling values ​​in time order, and generates a load current change trajectory. S42: Based on the load current change trajectory, call the device number and cloud-edge collaborative transmission mark, read the device rated load parameter record established before the data acquisition according to the device number, obtain the lower limit value of rated load current, the upper limit value of rated load current and the number of load intervals, determine the correspondence between the device number and the rated load content, and generate the load interval identifier. S43: Based on the load interval identifier, call the load current sample value and load sampling time marker in the load current change trajectory, match the load current sample value to the corresponding load interval, obtain the load interval number corresponding to the load sampling time marker, arrange them in the order of load sampling time and synchronize them to the cloud collaborative node to generate a device load distribution record set.

7. The electroplating equipment management method based on cloud-edge collaboration according to claim 1, characterized in that: The steps for obtaining the result are as follows: The method further includes: S5: Based on the equipment load distribution record set, call the conductivity status marker, establish a corresponding relationship, determine the upper load interval, identify continuous distribution segments, and when the fluctuation category and number fall into the upper part, extract the equipment number and load sampling time marker to generate electroplating equipment management results. The management results of the electroplating equipment include key monitoring targets, periods of high load fluctuation, operational risk indicators, and management priority information.

8. The electroplating equipment management method based on cloud-edge collaboration according to claim 7, characterized in that: The steps for obtaining S5 are as follows: S51: Based on the load sampling time marker and load interval number sequence in the device load distribution record set, the edge collaboration node reads the corresponding conductivity status marker in the edge conductivity status identifier set according to the device number, verifies the correspondence between the load interval number sequence under the same device number and the conductivity status marker, and establishes a load status correspondence table. S52: Call the load interval number sequence, conductivity status mark and cloud-edge collaborative transmission mark in the load status correspondence table, divide the second half of the arrangement position according to the load interval number from low to high, filter the load interval numbers that fall into the second half of the arrangement position, and merge the consecutive number positions along the load sampling time sequence to obtain the upper load continuous segment. S53: Based on the upper load continuous segment, the cloud collaborative node reads the conductivity status mark corresponding to the same device number, determines whether the conductivity status mark is a fluctuation category, and verifies whether the load interval number in the continuous distribution segment remains in the upper load interval. It then extracts the device number and the load sampling time mark in the continuous distribution segment to generate the electroplating equipment management result.

9. A cloud-edge collaborative electroplating equipment management system, characterized in that, The system is used in the cloud-edge collaborative electroplating equipment management method according to any one of claims 1-8, the system comprising: The sampling and collection module collects the current density sampling value of the output terminal of the electroplating rectifier power supply corresponding to the electroplating tank, generates a sampling time marker, binds the equipment number to form a sampling record, collects the data according to the equipment number and sorts it according to the sampling time, associates adjacent sampling values ​​to form a continuous current density change trajectory, and generates a device current time series dataset. The interval jump module, based on the device current time series dataset, reads the continuous change trajectory of current density and the device rated parameters, divides the current density intervals and numbers them, sequentially matches the sampled values ​​to form an interval number sequence, identifies the number change time and change direction, and forms a set of conduction state jump feature sets. The state discrimination module extracts the number change time markers and directions from the set of conductive state transition features, sorts them by device number to form a number change time sequence, counts the change frequency and reads the state judgment parameters. When the direction alternates and the frequency is greater than the threshold, it is marked as fluctuating; otherwise, it is marked as stable, forming an edge conductive state identifier set. The load mapping module reads the device number and conductivity status mark from the edge conductivity status identifier set, obtains the load sampling value and time mark with the device number, forms the load change trajectory, reads the rated load parameter, divides the load interval and matches the sampling value, forms the load interval number sequence, and generates a device load distribution record set. The high-load screening module, based on the equipment load distribution record set, calls the conductivity status marker, establishes a corresponding relationship, determines the upper load interval, identifies continuous distribution segments, and when the fluctuation category and number fall into the upper part, extracts the equipment number and load sampling time marker, and generates electroplating equipment management results.

Images