End-to-end packet loss recovery method, electronic equipment and medium for wet copper production process

Abstract

The present invention relates to the technical field of end-to-end packet loss recovery, and in particular to an end-to-end packet loss recovery method, electronic device, and medium for a wet copper production process, comprising: quantifying the degree of physical influence of each production process parameter on production stability and establishing an influence coefficient matrix; establishing a real-time level matrix for data transmission of each production process parameter based on the maximum allowable delay requirement of the control closed loop of each production process parameter; generating dynamic priority weights for each production process parameter based on the current production status, the influence coefficient matrix, and the real-time level matrix; allocating communication resources based on the dynamic priority weights, and establishing a multi-level collaborative recovery strategy for the sensor layer, the transmission layer, and the control layer. The present invention breaks through the limitations of traditional single-layer recovery through dual evaluation of the influence coefficient matrix and the real-time level matrix of multiple parameters, achieves end-to-end optimization from the sensor to the control center, and is suitable for complex industrial scenarios such as wet copper production.

Description

Technical Field

[0001] The present invention relates to the technical field of end-to-end packet loss recovery, and in particular to an end-to-end packet loss recovery method, electronic equipment, and medium for a wet copper production process. Background Art

[0002] As a core technology for metal extraction, the wet copper production process relies heavily on real-time monitoring and dynamic control of key parameters such as solution pH, temperature, flow rate, metal ion concentration, and current density for production stability. In industrial field environments, sensor networks need to continuously transmit collected process data to the control center to support precise adjustment of the closed-loop control system. However, the wet metallurgical production environment is characterized by strong electromagnetic interference, frequent equipment vibration, and complex network topology. This makes the data transmission link from the sensor layer to the control layer susceptible to interference, and frequently causes problems such as data loss and transmission delays. Such data anomalies not only cause deviations in process parameters and increased energy consumption, but can also cause equipment overload or process loss, seriously threatening production safety and efficiency.

[0003] Existing methods for recovering industrial data packet loss often focus on single-layer optimization and lack end-to-end, global collaborative design. For example, general packet loss recovery solutions typically employ fixed priority or static redundancy strategies, failing to consider the dynamic nature of multi-parameter coupled control in wet copper production processes. This leads to rigid resource allocation and inefficient recovery. Furthermore, existing technologies fail to effectively coordinate recovery actions across the sensor, transport, and control layers. Summary of the Invention

[0004] In view of at least one of the above technical problems, the present invention provides an end-to-end packet loss recovery method and system for a wet copper production process, which can effectively solve the problems in the background technology.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] End-to-end packet loss recovery method for wet copper production process, including:

[0007] Quantify the physical impact of each production process parameter on production stability and establish an influence coefficient matrix;

[0008] Establishing a real-time level matrix for data transmission of each of the production process parameters based on the maximum allowable delay requirement of the closed-loop control of each of the production process parameters;

[0009] Generating a dynamic priority weight of each of the production process parameters based on the current production status, the influence coefficient matrix, and the real-time level matrix;

[0010] Allocate communication resources according to the dynamic priority weights and establish a multi-level collaborative recovery strategy for the sensor layer, the transport layer, and the control layer;

[0011] The multi-level collaborative recovery strategy includes:

[0012] In the sensor layer, a redundant sensor group is set, and data of the redundant sensor group is triggered to be preferentially replaced based on the dynamic priority weight;

[0013] At the transport layer, allocating adaptive retransmission times and channel preemption priorities according to the dynamic priority weights;

[0014] At the control layer, lost parameters are compensated based on historical data and model predictions.

[0015] Furthermore, generating a dynamic priority weight of each of the production process parameters includes:

[0016] Collect sensor data in real time through edge computing nodes;

[0017] Based on the process deviation requirements and the parameter coupling relationship, dynamically generate a first weight matrix corresponding to the influence coefficient matrix;

[0018] Based on the real-time status of the transmission channel, dynamically generating a second weight matrix corresponding to the real-time level matrix;

[0019] The dynamic priority weight of each of the production process parameters is obtained according to the first weight matrix and the second weight matrix.

[0020] Furthermore, obtaining the dynamic priority weight of each of the production process parameters includes:

[0021] The first weight matrix and the second weight matrix are weightedly fused to obtain the dynamic priority weight of each production process parameter.

[0022] Furthermore, when collecting sensor data in real time through edge computing nodes, it also includes:

[0023] The sensor data is subjected to sliding window filtering and normalization processing.

[0024] Furthermore, performing communication resource allocation according to the dynamic priority weight includes:

[0025] Classifying the communication resources into high priority data, medium priority data, and low priority data according to the dynamic priority weights;

[0026] Specifically, dedicated time slots are reserved for the high-priority data, dynamic bandwidth is allocated for the medium-priority data, and fragmented parallel transmission is enabled for the low-priority data.

[0027] Furthermore, the protocol for the fragmented parallel transmission is a TCP / IP fragmentation strategy, and the number of fragments is dynamically adjusted according to the current network bandwidth.

[0028] Furthermore, triggering preferential replacement of data of the redundant sensor group based on the dynamic priority weight includes:

[0029] Setting a primary sensor in the redundant sensor group;

[0030] Calculating and comparing the real-time confidence scores of each sensor in the redundant sensor group;

[0031] When the primary sensor data is lost, the sensor data corresponding to the highest real-time confidence score is selected as the backup sensor data.

[0032] Furthermore, the loss parameters are compensated based on historical data and model prediction, including:

[0033] Establish a parameter prediction model based on LSTM neural network. The input parameters include historical process data series and current values ​​of related parameters, and the output parameter prediction value.

[0034] When the duration of data packet loss exceeds the maximum allowable delay corresponding to the production process parameter, the parameter prediction value is input into the control closed loop as compensation data.

[0035] An electronic device, comprising:

[0036] a memory for storing computer program instructions;

[0037] A processor is configured to execute the computer program instructions to implement the end-to-end packet loss recovery method for the wet copper production process as described above.

[0038] A computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the end-to-end packet loss recovery method for the wet copper production process as described above.

[0039] The technical solution of the present invention can achieve the following technical effects:

[0040] The present invention overcomes the limitations of traditional single-layer recovery through dual evaluation of a multi-parameter influence coefficient matrix and a real-time level matrix, achieving end-to-end optimization from sensors to the control center. Compared with general packet loss recovery methods, this solution significantly improves data transmission reliability, control stability, and resource utilization through end-to-end dynamic optimization and layered collaboration, making it suitable for complex industrial scenarios such as wet copper production. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 A flow chart of an end-to-end packet loss recovery method for a wet copper production process;

[0043] Figure 2 A schematic diagram of a process for generating dynamic priority weights for each production process parameter;

[0044] Figure 3 A flowchart for triggering preferential replacement of data from redundant sensor groups based on dynamic priority weights. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0046] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly attached to the other element or there may be an intermediate element. When an element is referred to as being "connected to" another element, it may be directly connected to the other element or there may be an intermediate element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only implementation methods.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0048] Example 1

[0049] like Figure 1 As shown in Figure 1, the end-to-end packet loss recovery method for the wet copper production process includes:

[0050] S1: Quantify the physical impact of each production process parameter on production stability and establish an influence coefficient matrix;

[0051] S2: Based on the maximum allowable delay requirements of the control closed loop of each production process parameter, a real-time level matrix for data transmission of each production process parameter is established;

[0052] S3: Generate dynamic priority weights for each production process parameter based on the current production status, impact coefficient matrix, and real-time level matrix;

[0053] In this embodiment, since the static priority cannot adapt to the problem of parameter coupling change and working condition change, this step sets a dynamic priority, and the weight is adaptively adjusted with the production status, resource allocation is more accurate, and the system robustness is improved. The current production status, the influence coefficient matrix and the real-time level matrix are selected as the core basis to generate the dynamic priority weight; the current production status is an adaptive response to the dynamic evolution of the working condition, which will change the importance of the parameters, and can analyze the sensor data stream in real time to detect abnormal events; since the influence of various production process parameters on production stability in wet copper production is nonlinear and highly coupled, this step establishes a wet copper production process. The influence coefficient matrix of production parameters can be used to assign sensitivity coefficients to each parameter through process mechanism analysis or historical data modeling, quantify its physical correlation, avoid applying equal force to all parameters, provide a physical basis for subsequent dynamic priority, and ensure that key parameters are prioritized; the control loops of different parameters have orders of magnitude differences in tolerance to delays (for example, current density requires millisecond-level response, and temperature can tolerate second-level delays). This step sets the real-time level matrix for data transmission. According to the requirements of the control algorithm, the maximum allowable delay level can be set for each parameter. The generated matrix avoids low-timeliness parameters occupying high-real-time resources and optimizes network utilization.

[0054] S4: Allocate communication resources based on dynamic priority weights and establish a multi-level collaborative recovery strategy for the sensor layer, transport layer, and control layer;

[0055] Specifically, the purpose of communication resource allocation is to establish deterministic transmission paths and reserve operational space for recovery strategies to ensure their smooth execution. Communication resource allocation is performed first, and global resource pre-optimization lays the foundation for layered recovery actions, thereby avoiding local decision conflicts and improving overall efficiency. Multi-layer coordinated recovery is then implemented, converting dynamic priority weights into coordinated recovery actions at the sensor, transport, and control layers. End-to-end assurance is achieved through layered governance and cross-layer linkage.

[0056] The multi-level collaborative recovery strategy includes:

[0057] At the sensor layer, a redundant sensor group is set up, and the data of the redundant sensor group is preferentially replaced based on the dynamic priority weight. The industrial field sensors of hydrometallurgy are susceptible to strong acid corrosion and electromagnetic interference, which may cause single-point data anomalies or loss, and simple master-slave switching may switch to the same faulty equipment. Therefore, this step sets up a redundant sensor group at the sensor layer to avoid single-point failures and systematic errors caused by principle defects, ensure the reliability of data after switching, and at the same time, preferential replacement ensures that the switching is triggered only for high-weight parameters to avoid instability caused by frequent switching of low-value sensors.

[0058] At the transport layer, adaptive retransmission times and channel preemption priorities are allocated according to dynamic priority weights. Network conditions fluctuate violently in the wet copper production environment. In order to solve the problem that fixed retransmissions are prone to data recovery failures with high dynamic priority weights and bandwidth waste with data with low dynamic priority weights, this embodiment sets an adaptive retransmission number and a mechanism to dynamically adjust the maximum retransmission number according to the parameter weights to avoid low-value parameters from crowding out network resources. The channel preemption priority implements cross-level preemption based on the weight difference to ensure breakthrough timeliness of key parameters, and interrupted data packets need to be retransmitted from the beginning. During implementation, strict priority scheduling may cause systemic risks. In order to solve the natural defects of strict priority scheduling, an anti-starvation mechanism can be introduced at the transport layer. The same data packet can be preempted at most twice, or timeout to trigger priority increase, to prevent long-term backlog of data with low dynamic priority weights from causing system abnormalities.

[0059] At the control layer, lost parameters are compensated based on historical data and model predictions.

[0060] The present invention overcomes the limitations of traditional single-layer recovery through dual evaluation of a multi-parameter influence coefficient matrix and a real-time level matrix, achieving end-to-end optimization from sensors to the control center. Compared with general packet loss recovery methods, this solution significantly improves data transmission reliability, control stability, and resource utilization through end-to-end dynamic optimization and layered collaboration, making it suitable for complex industrial scenarios such as wet copper production.

[0061] Based on the above embodiments, Figure 2 As shown, the dynamic priority weights of each production process parameter are generated, including:

[0062] S31: Real-time collection of sensor data through edge computing nodes;

[0063] Specifically, since the wet production site has strong acid corrosion and electromagnetic interference, in order to reduce the interference of the specific environment, this step selects edge computing nodes to collect sensor data, shorten the signal transmission distance, complete the analog-to-digital conversion on-site, and avoid interference with microvolt-level analog signals. At the same time, the use of edge nodes to collect data allows the sealing node to be directly installed on the trough bracket, eliminating the need for large-area cable trough laying.

[0064] S32: Based on the process deviation requirements and the parameter coupling relationship, dynamically generate a first weight matrix corresponding to the influence coefficient matrix;

[0065] S33: Based on the real-time status of the transmission channel, dynamically generate a second weight matrix corresponding to the real-time level matrix;

[0066] S34: Obtain the dynamic priority weight of each production process parameter according to the first weight matrix and the second weight matrix.

[0067] This embodiment obtains process deviation requirements, parameter coupling relationships, and real-time status of transmission channels from the real-time data collected by the edge node in step S31, and converts the edge real-time data stream into a dual-weight matrix by combining the influence coefficient matrix and the real-time level matrix. This quantifies the physical impact of the parameters on production stability, reflects real-time working condition changes, and adjusts the transmission priority of the parameters according to the real-time status of the network to resolve resource competition in congested scenarios, coordinate physical requirements and network status, and achieve global optimal decision-making.

[0068] Based on the above embodiment, the dynamic priority weight of each production process parameter is obtained, including:

[0069] The first weight matrix and the second weight matrix are weightedly fused to obtain the dynamic priority weight of each production process parameter.

[0070] To improve the quality of collected sensor data, real-time collection of sensor data through edge computing nodes also includes:

[0071] Perform sliding window filtering and normalization on sensor data.

[0072] In this embodiment, the wet copper production site is affected by instantaneous pulses caused by strong electromagnetic interference, periodic noise caused by equipment vibration, and thermal noise of the sensor itself. To ensure the quality of the collected sensor data, this step preprocesses the sensor data. To unify the dimension of the sensor data, this step provides normalization processing. The purpose of sliding window filtering is to reduce the impact of industrial site noise and improve the authenticity of data in a strong interference environment. The window size is set by a window width adaptive algorithm, and median filtering is selected for sliding window filtering. Median filtering is used to eliminate instantaneous interference pulses of the electrode, and exponential smoothing is used to predict trends to improve the accuracy of sensor data.

[0073] As a preferred embodiment of this invention, communication resource allocation is performed according to dynamic priority weights, including:

[0074] Classify communication resources into high priority data, medium priority data and low priority data according to dynamic priority weights;

[0075] Among them, dedicated time slots are reserved for high-priority data, dynamic bandwidth is allocated for medium-priority data, and fragmented parallel transmission is enabled for low-priority data.

[0076] Specifically, in order to avoid communication network congestion and maximize the utilization of network communication resources, this embodiment divides communication resources into three priority layers, hard isolates high-priority data, reserves dedicated time slots for high-priority data, ensures absolute priority of control instructions, uses dynamic bandwidth sharing for medium-priority data, ensures that parameters such as temperature are delivered within an acceptable delay, and breaks low-priority data into fragments for transmission, utilizing the idle gaps of high-latency channels for transmission, thus avoiding network congestion while completing transmission.

[0077] To ensure highly reliable transmission of non-critical data, the protocol for sharded parallel transmission is the TCP / IP sharding strategy, and the number of shards is dynamically adjusted according to the current network bandwidth.

[0078] Specifically, low-priority data does not need to be real-time, but its ultimate integrity must be guaranteed. This embodiment uses a TCP / IP fragmentation strategy to split data packets that exceed the network's carrying capacity into multiple small fragments for transmission, dispersing the risk of large data to independent fragments and reassembling them at the receiving end, thereby improving the success rate of low-priority data transmission without affecting high-priority channels. In the wet copper production scenario, the network fluctuates violently, and a fixed number of fragments will waste resources. This strategy improves network bandwidth utilization by dynamically adjusting the number of fragments. Setting a fragment cache pool at the receiving end can avoid the abandonment of the entire packet due to the loss of a single fragment, significantly improving the transmission reliability and efficiency in the high packet loss environment of the industrial network.

[0079] Based on the above embodiments, Figure 3 As shown, triggering the preferential replacement of data from redundant sensor groups based on dynamic priority weights includes:

[0080] Set the primary sensor in the redundant sensor group;

[0081] Calculate the real-time confidence score of each sensor in the redundant sensor group and compare them;

[0082] When the primary sensor data is lost, the sensor data corresponding to the highest real-time confidence score is selected as the backup sensor data.

[0083] In this embodiment, a main sensor is set. Designating the main sensor can simplify the decision-making process. As long as the set main sensor is normal, the system can output a constant data source to avoid data jumps. Due to the harsh environment of wet copper production, strong acid and high current density can lead to a high sensor failure rate. Therefore, this step sets a data preferential replacement mechanism in which the system can quickly switch to other sensors as backup sensors when the main sensor fails. To avoid inferior switching, it is not enough to simply set the main and backup sensors. It must be combined with a confidence scoring mechanism to calculate the confidence score of each sensor in the redundant sensor group in real time. When the main sensor loses a packet, it automatically switches to the backup sensor data with the highest score. This is more adaptable to changing industrial environments than fixed backup sensors. It should be noted that the data preferential replacement mechanism can be designed with a safety boundary. When the confidence scores of each sensor in the redundant sensor group do not meet the standard, the system cannot switch at will, and predictive compensation or safe shutdown should be triggered.

[0084] As a preferred embodiment of this invention, the loss parameter compensation is predicted based on historical data and a model, including:

[0085] Establish a parameter prediction model based on LSTM neural network. The input parameters include historical process data series and current values ​​of related parameters, and the output parameter prediction value.

[0086] Specifically, this embodiment selects historical process data sequences and current values ​​of associated parameters as inputs to the LSTM prediction model based on the physical and chemical nature of hydrometallurgy and the real-time requirements of the control system. The historical sequences can capture the dynamic process of the electrolysis reaction, such as the decay period of the current density. The associated parameters will directly affect the control target parameters through physical coupling or control loops. For example, there is a strong coupling between the cell voltage and temperature, and a sudden voltage change will immediately affect the current density. At the same time, the current values ​​of the associated parameters provide the instantaneous state of the system, avoiding the cumulative error of pure extrapolation. When data is lost, the historical process data sequences and the current values ​​of the associated parameters become the true observation of the process state.

[0087] When the duration of data packet loss exceeds the maximum allowable delay of the corresponding production process parameter, the parameter prediction value is input into the control loop as compensation data.

[0088] Example 2

[0089] An electronic device, comprising:

[0090] a memory for storing computer program instructions;

[0091] A processor is configured to execute computer program instructions to implement the end-to-end packet loss recovery method for the wet copper production process as described in the first embodiment.

[0092] Example 3

[0093] A computer-readable storage medium is characterized in that it stores computer program instructions, and when the instructions are executed by a processor, it implements the end-to-end packet loss recovery method for the wet copper production process as described in Example 1.

[0094] Those skilled in the art will appreciate that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An end-to-end packet loss recovery method for a wet copper production process, characterized in that: include: Quantify the physical impact of each production process parameter on production stability and establish an influence coefficient matrix; Establishing a real-time level matrix for data transmission of each of the production process parameters based on the maximum allowable delay requirement of the closed-loop control of each of the production process parameters; Generating a dynamic priority weight of each of the production process parameters based on the current production status, the influence coefficient matrix, and the real-time level matrix; Allocate communication resources according to the dynamic priority weights and establish a multi-level collaborative recovery strategy for the sensor layer, the transport layer, and the control layer; The multi-level collaborative recovery strategy includes: In the sensor layer, a redundant sensor group is set, and data of the redundant sensor group is triggered to be preferentially replaced based on the dynamic priority weight; At the transport layer, allocating adaptive retransmission times and channel preemption priorities according to the dynamic priority weights; At the control layer, lost parameters are compensated based on historical data and model predictions.

2. The end-to-end packet loss recovery method for the wet copper production process according to claim 1, characterized in that: Generating dynamic priority weights for each of the production process parameters, including: Collect sensor data in real time through edge computing nodes; Based on the process deviation requirements and the parameter coupling relationship, dynamically generate a first weight matrix corresponding to the influence coefficient matrix; Based on the real-time status of the transmission channel, dynamically generating a second weight matrix corresponding to the real-time level matrix; The dynamic priority weight of each of the production process parameters is obtained according to the first weight matrix and the second weight matrix.

3. The end-to-end packet loss recovery method for the wet copper production process according to claim 2, characterized in that: Obtaining the dynamic priority weight of each of the production process parameters includes: The first weight matrix and the second weight matrix are weightedly fused to obtain the dynamic priority weight of each production process parameter.

4. The end-to-end packet loss recovery method for the wet copper production process according to claim 2, characterized in that: When collecting sensor data in real time through edge computing nodes, it also includes: The sensor data is subjected to sliding window filtering and normalization processing.

5. The end-to-end packet loss recovery method for the wet copper production process according to claim 1, characterized in that: Allocating communication resources according to the dynamic priority weights includes: Classifying the communication resources into high priority data, medium priority data, and low priority data according to the dynamic priority weights; Specifically, dedicated time slots are reserved for the high-priority data, dynamic bandwidth is allocated for the medium-priority data, and fragmented parallel transmission is enabled for the low-priority data.

6. The end-to-end packet loss recovery method for the wet copper production process according to claim 5, characterized in that: The protocol for the fragmented parallel transmission is the TCP / IP fragmentation strategy, and the number of fragments is dynamically adjusted according to the current network bandwidth.

7. The end-to-end packet loss recovery method for a wet copper production process according to claim 1, characterized in that: Triggering preferential replacement of data of the redundant sensor group based on the dynamic priority weight includes: Setting a primary sensor in the redundant sensor group; Calculating and comparing the real-time confidence scores of each sensor in the redundant sensor group; When the primary sensor data is lost, the sensor data corresponding to the highest real-time confidence score is selected as the backup sensor data.

8. The end-to-end packet loss recovery method for a wet copper production process according to claim 1, characterized in that: Compensate for lost parameters based on historical data and model predictions, including: Establish a parameter prediction model based on LSTM neural network. The input parameters include historical process data series and current values ​​of related parameters, and the output parameter prediction value. When the duration of data packet loss exceeds the maximum allowable delay corresponding to the production process parameter, the parameter prediction value is input into the control closed loop as compensation data.

9. An electronic device, characterized in that: include: a memory for storing computer program instructions; A processor, configured to execute the computer program instructions to implement the end-to-end packet loss recovery method for the wet copper production process according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that Computer program instructions are stored, and when the instructions are executed by a processor, the end-to-end packet loss recovery method of the wet copper production process according to any one of claims 1 to 8 is implemented.

Images