Industrial linkage control method and system based on cloud database

Abstract

This invention discloses an industrial linkage control method and system based on a cloud database. It tracks and analyzes the status data of target objects on an industrial operation line to identify suspicious processes. Based on the operation location tags of the suspicious processes, it searches the cloud database for status data related to those processes. Based on the status data, it identifies potential abnormal events on the industrial operation line. It reproduces and identifies these potential abnormal events, generating fault attribute codes for the industrial operation line. Based on the fault attribute codes and the cloud database, it determines the faulty hardware operation records of the industrial operation line. By comparing the faulty hardware operation records with the suspicious processes, it implements linkage adjustments to the target range of the industrial operation line. By tracking target objects on the industrial operation line to investigate all processes, identifying the fault attributes corresponding to potential abnormal events, and combining this with hardware fault investigation and calibration using the cloud database, it ensures the accuracy and reliability of the overall fault handling of the operation line.

Description

Technical Field

[0001] This invention relates to the field of cloud data processing, and in particular to an industrial linkage control method and system based on a cloud database. Background Technology

[0002] In smart industrial scenarios, production lines and / or robots form industrial operation lines that integrate numerous production processes into a single time and space, enabling continuous mass production and comprehensive control of hardware equipment. Given the large number and diverse types of hardware equipment in these scenarios, malfunctions are inevitable during operation. These malfunctions manifest not only in mechanical wear and tear but also in the operation of hardware control programs, thus affecting the continuity, stability, and reliability of industrial production across the entire scenario. Current technologies typically employ partial shutdowns for maintenance and repair when malfunctions occur in industrial scenarios. However, this approach comes at the cost of global / partial production stoppages and is mostly implemented manually, impacting the continuity and efficiency of the entire production scenario. It also fails to comprehensively detect faults or make coordinated adjustments to hardware equipment for known faults, reducing the effectiveness and accuracy of troubleshooting and repair. Summary of the Invention

[0003] Considering that existing industrial operation lines rely on partial shutdowns for maintenance when hardware malfunctions, impacting the entire line's operation and lacking the ability to adjust related hardware in response to known faults, the effectiveness and accuracy of fault diagnosis and repair, as well as the continuity and stability of the industrial operation line, cannot be guaranteed. In view of these problems, this invention proposes an industrial linkage control method based on a cloud database, including:

[0004] Track and analyze the status data of target objects on the industrial operation line to identify suspicious processes on the industrial operation line; based on the operation location tags of the suspicious processes, search for status data related to the suspicious processes from the cloud database;

[0005] Based on the status data, potential abnormal events of the industrial operation line are identified; the potential abnormal events are reproduced and identified to generate fault attribute codes for the industrial operation line.

[0006] Based on the fault attribute code and the cloud database, determine the faulty hardware operation record of the industrial operation line; compare the faulty hardware operation record with the suspected process, and implement coordinated adjustment of the target range of the industrial operation line.

[0007] Optionally, the status data of target objects on the industrial operation line are tracked and analyzed to identify suspicious processes on the industrial operation line; based on the operation location tags of the suspicious processes, status data related to the suspicious processes are retrieved from a cloud database, including:

[0008] Track the entire process image of the target object on the industrial operation line, divide and identify the process intervals of the entire process image, and obtain the structural features and manipulated features of the target object for each process.

[0009] A benchmark comparison is performed between the constructed features and the operated features to determine whether the process is a suspicious process.

[0010] Based on the spatial attributes and hardware configuration attributes of the suspected process on the industrial operation line, an operation location tag for the suspected process is formed; the operation location tag is compared with the cloud database to find status data related to the suspected process; wherein, the status data includes hardware operation records related to the suspected process.

[0011] Optionally, based on the status data, potential abnormal events of the industrial operation line are identified; the potential abnormal events are reproduced and identified to generate a fault attribute code for the industrial operation line, including:

[0012] Hardware action information and hardware operating electrical parameters are extracted from the hardware operation records related to the suspected process contained in the status data to determine potential abnormal events of the hardware related to the suspected process; wherein, the potential abnormal events include abnormal action events and / or abnormal operating electrical signal events of the hardware related to the suspected process.

[0013] The potential abnormal events are constructed as feature elements, and recurrence identification is performed based on the feature elements to determine the recurrence time-domain attribute of the potential abnormal events. Based on the recurrence time-domain attribute, the attribute code of the actual fault occurring on the industrial operation line is identified. The attribute code contains a characterization label of the type of the actual fault.

[0014] Optionally, based on the fault attribute code and the cloud database, the faulty hardware operation record of the industrial operation line is determined; by comparing the faulty hardware operation record with the suspected process, a coordinated adjustment is implemented for the target range of the industrial operation line, including:

[0015] By comparing the fault attribute code with the cloud database, the hardware where the actual fault occurred is extracted from the cloud database; based on the correlation between the operating input and output of the hardware and the execution of the suspected process, the faulty hardware working record of the industrial operation line is determined.

[0016] By comparing the faulty hardware operation record with the suspected process, the faulty hardware operation path of the suspected process is determined, and the faulty hardware operation path is adjusted in a coordinated manner within the target area covered by the industrial operation line; wherein, the coordinated adjustment refers to the coordinated adjustment of the operation mode of different faulty hardware on the faulty hardware operation path.

[0017] As one aspect of the present invention, embodiments of the present invention also provide an industrial linkage control system based on a cloud database, including:

[0018] The process calibration module is used to track and analyze the status data of target objects on the industrial operation line, thereby calibrating the suspicious processes on the industrial operation line.

[0019] The process status lookup module is used to search for status data related to the suspicious process from the cloud database based on the operation location tag of the suspicious process;

[0020] A potential anomaly determination module is used to determine potential abnormal events of the industrial operation line based on the status data.

[0021] The fault attribute determination module is used to reproduce and identify the potential abnormal events and generate the fault attribute code of the industrial operation line.

[0022] The work record determination module is used to determine the fault hardware work record of the industrial operation line based on the fault attribute code and the cloud database.

[0023] The linkage adjustment module is used to compare the faulty hardware operation record with the suspicious process and implement linkage adjustment on the target range of the industrial operation line.

[0024] Optionally, the process calibration module is used to track and analyze the status data of target objects on the industrial operation line, thereby calibrating suspicious processes on the industrial operation line, including:

[0025] Track the entire process image of the target object on the industrial operation line, divide and identify the process intervals of the entire process image, and obtain the structural features and manipulated features of the target object for each process.

[0026] A benchmark comparison is performed between the constructed features and the operated features to determine whether the process is a suspicious process.

[0027] The process status lookup module is used to search for status data related to the suspicious process from the cloud database based on the operation location tag of the suspicious process, including:

[0028] Based on the spatial attributes and hardware configuration attributes of the suspected process on the industrial operation line, an operation location tag for the suspected process is formed; the operation location tag is compared with the cloud database to find status data related to the suspected process; wherein, the status data includes hardware operation records related to the suspected process.

[0029] Optionally, the potential anomaly determination module is used to determine potential abnormal events of the industrial operation line based on the status data, including:

[0030] Hardware action information and hardware operating electrical parameters are extracted from the hardware operation records related to the suspected process contained in the status data to determine potential abnormal events of the hardware related to the suspected process; wherein, the potential abnormal events include abnormal action events and / or abnormal operating electrical signal events of the hardware related to the suspected process.

[0031] The fault attribute determination module is used to reproduce and identify the potential abnormal events, and generate fault attribute codes for the industrial operation line, including:

[0032] The potential abnormal events are constructed as feature elements, and recurrence identification is performed based on the feature elements to determine the recurrence time-domain attribute of the potential abnormal events. Based on the recurrence time-domain attribute, the attribute code of the actual fault occurring on the industrial operation line is identified. The attribute code contains a characterization label of the type of the actual fault.

[0033] Optionally, the work record determination module is used to determine the fault hardware work record of the industrial operation line based on the fault attribute code and the cloud database, including:

[0034] By comparing the fault attribute code with the cloud database, the hardware where the actual fault occurred is extracted from the cloud database; based on the correlation between the operating input and output of the hardware and the execution of the suspected process, the faulty hardware working record of the industrial operation line is determined.

[0035] The linkage adjustment module is used to compare the faulty hardware operation record with the suspicious process, and to implement linkage adjustment on the target range of the industrial operation line, including:

[0036] By comparing the faulty hardware operation record with the suspected process, the faulty hardware operation path of the suspected process is determined, and the faulty hardware operation path is adjusted in a coordinated manner within the target area covered by the industrial operation line; wherein, the coordinated adjustment refers to the coordinated adjustment of the operation mode of different faulty hardware on the faulty hardware operation path.

[0037] The beneficial effects of the above-mentioned technical solutions provided in the embodiments of the present invention include at least the following:

[0038] This invention provides an industrial linkage control method and system based on a cloud database. The method tracks and analyzes the status data of target objects on an industrial operation line to identify suspicious processes. Based on the operation location tags of the suspicious processes, it searches the cloud database for status data related to those processes. Based on the status data, it identifies potential abnormal events on the industrial operation line. It reproduces and identifies these potential abnormal events, generating fault attribute codes for the industrial operation line. Based on the fault attribute codes and the cloud database, it determines the faulty hardware operation records of the industrial operation line. By comparing the faulty hardware operation records with the suspicious processes, it implements linkage adjustments to the target range of the industrial operation line. By tracking target objects on the industrial operation line to investigate all processes, identifying the fault attributes corresponding to potential abnormal events, and combining this with hardware fault investigation and calibration using the cloud database, the method ensures the accuracy and reliability of overall fault handling on the operation line.

[0039] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0040] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0041] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0042] Figure 1 This is a flowchart illustrating the industrial linkage control method based on a cloud database provided in an embodiment of the present invention.

[0043] Figure 2 This is a schematic diagram of the structure of an industrial linkage control system based on a cloud database provided in an embodiment of the present invention. Detailed Implementation

[0044] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0045] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," "outer," "far," "near," "front," and "rear," etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the accompanying drawings and are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0046] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0047] Please see Figure 1 As shown, an embodiment of this application provides an industrial linkage control method based on a cloud database. This cloud database-based industrial linkage control method includes:

[0048] Track and analyze the status data of target objects on the industrial operation line to identify suspicious processes on the industrial operation line; based on the operation location tags of suspicious processes, search for status data related to the suspicious processes from the cloud database;

[0049] Based on the status data, identify potential abnormal events on the industrial operation line; reproduce and identify potential abnormal events to generate fault attribute codes for the industrial operation line.

[0050] Based on the fault attribute code and the cloud database, identify the faulty hardware operation records of the industrial operation line; compare the faulty hardware operation records with the suspicious processes, and implement coordinated adjustments to the target range of the industrial operation line.

[0051] The beneficial effects of the above embodiments are that the industrial linkage control method based on cloud database checks all processes by tracking target objects on the industrial operation line, identifies the fault attributes corresponding to potential abnormal events, and combines the cloud database to perform hardware fault diagnosis and calibration, thereby ensuring the accuracy and reliability of fault handling of the entire operation line.

[0052] In another embodiment, the status data of targets on the industrial operation line is tracked and analyzed to identify suspicious processes on the industrial operation line; based on the operation location tags of the suspicious processes, status data related to the suspicious processes is retrieved from a cloud database, including:

[0053] Track the entire process image of the target object on the industrial operation line, divide and identify the process intervals of the entire process image, and obtain the structural features and manipulated features of the target object for each process.

[0054] By comparing the structural features and the manipulated features against a benchmark, it can be determined whether the process is a suspicious process.

[0055] Based on the spatial attributes and hardware configuration attributes of the suspicious process on the industrial operation line, an operation location tag for the suspicious process is generated; the operation location tag is compared with the cloud database to find the status data related to the suspicious process; among which, the status data includes the hardware operation records related to the suspicious process.

[0056] The beneficial effects of the above embodiments are that the industrial operation line may include, but is not limited to, production lines and hardware equipment such as robots; wherein, the production line may include, but is not limited to, several hardware devices corresponding one-to-one with several production processes, and each hardware device may include, but is not limited to, performing different types of operation processes such as cutting, splicing, embedding, and packaging; robots may be set up between different production processes, for example, to transport the corresponding semi-finished products formed in the current production process to the next production process, and / or to perform machine vision recognition on the semi-finished products processed in the current production process to screen out unqualified products, etc. By combining production lines and robots and other hardware equipment, the industrial operation line can maximize the satisfaction of the centralized batch production needs of different products. Understandably, in an industrial production line, a material / product (i.e., the target object) needs to traverse all processes on the industrial production line (i.e., all production lines and all hardware equipment such as robots). By tracking and analyzing the full-process images of the target object on the industrial production line, it is possible to comprehensively and accurately record the operation and processing of the target object by each process (i.e., each production line and each robot) as well as the structural changes of the target object after each process, providing a basis for subsequent judgment on whether each process is normal or not.

[0057] Specifically, the entire process of capturing images of a target object on an industrial operation line can be tracked and filmed. Based on the spatial distribution of all processes within the industrial operation line and preset process switching markers (such as sensor trigger signals and robotic arm movement start points), the entire process image is divided into process intervals to obtain several sub-images, each corresponding to a process under the industrial operation line. Each sub-image is then identified and analyzed. Image segmentation algorithms are used to extract the contour, volume, and surface texture of the target object before and after the process as structural features. Target tracking algorithms are used to extract the displacement trajectory, rotation angle, and deformation of the target object during the process as manipulated features. This yields the structural features (such as changes in the target object's own structural shape) and manipulated features (such as the processing amount and / or spatial transportation and transfer operations applied to the target object by hardware equipment during the corresponding process) of the target object during each process. The aforementioned structural features and manipulated features are then compared with the structural features and manipulated features of the benchmark target object for the corresponding process to obtain structural deviations and manipulated deviations. The benchmark features are derived from the statistical average of the same process during historical normal production periods or the standard values ​​in the design specifications. Deviations are calculated using Euclidean distance or Mahalanobis distance. If the aforementioned structural deviation or manipulated deviation exceeds the corresponding deviation threshold, the corresponding process is judged to be a suspicious process; otherwise, the corresponding process is judged not to be a suspicious process, thereby comprehensively and accurately calibrating processes that may cause anomalies on the industrial operation line.

[0058] Furthermore, each process on the industrial production line is equipped with unique and distinctive hardware devices, and the operating mode of the hardware devices in each process is also fixed. If an anomaly does occur in a certain process, the hardware devices configured for that process will inevitably exhibit corresponding abnormalities during operation. These abnormalities may include, but are not limited to, abnormal external actions of the hardware devices and abnormal internal electrical signal control. Specifically, the spatial attributes of the suspected process on the industrial production line (such as the physical spatial distribution and process sequence distribution of the suspected process in the entire industrial production line process) and hardware configuration attributes (such as the types and models of all hardware devices included in the suspected process) are obtained. These spatial attributes and hardware configuration attributes are then encoded and combined, for example, by concatenating spatial coordinates, process number, equipment model, and controller ID into a string or performing a hash operation to generate a unique operation location tag for the suspected process. Then, using the aforementioned operation location tags as indexes, we search for hardware operation records related to the suspicious process in the cloud database. The cloud database is used to record all hardware and software operation records of the industrial operation line throughout its entire lifecycle, as well as communication and interaction records between the industrial operation line and the outside world. The aforementioned hardware operation records refer to the interaction operation records of the external actions and internal electrical signals (such as instructions) of all hardware equipment under the suspicious process, specifically including equipment logs, sensor timing data, control instruction sequences and their timestamps.

[0059] In another embodiment, based on condition data, potential abnormal events of the industrial operation line are identified; the potential abnormal events are reproduced and identified to generate fault attribute codes for the industrial operation line, including:

[0060] Hardware action information and hardware operating electrical parameters are extracted from the hardware operation records related to the suspicious processes contained in the status data to determine potential abnormal events of the hardware related to the suspicious processes; among which, potential abnormal events include abnormal action events and / or abnormal operating electrical signal events of the hardware related to the suspicious processes.

[0061] Construct feature characteristics of potential abnormal events, identify recurrence based on feature characteristics, and determine the recurrence time domain attribute of potential abnormal events; based on the recurrence time domain attribute, identify the attribute code of the actual fault occurring on the industrial operation line; wherein, the attribute code contains a characterization label of the type of actual fault.

[0062] As can be seen from the preceding content, the beneficial effects of the above embodiments include the status data obtained from the suspicious process, which includes the interaction records of the external actions and internal electrical signals (such as instructions) of all hardware devices under the suspicious process. The status data characterizes the working status of the hardware devices from the perspective of external actions and internal signal control. Specifically, hardware action information (such as hardware action trajectory and / or action location) and hardware operation electrical parameters (such as electrical signal parameters such as instructions and data streams during hardware operation) are extracted from hardware operation records related to suspicious processes contained in the status data. Dynamic time warping (DTW) is performed on the extracted hardware action information and the standard action template of the corresponding hardware to obtain the action drift amplitude. Wavelet transform is performed on the hardware operation electrical parameters to denoise them, and their signal-to-noise ratio is calculated as the electrical parameter interference noise intensity. If the action drift amplitude exceeds a preset amplitude threshold (this threshold is set according to the statistical distribution of action drift in historical normal data, for example, the 95th percentile), and / or the electrical parameter interference noise intensity is lower than a preset intensity threshold (for example, the signal-to-noise ratio is lower than 20dB), it is determined that the corresponding hardware has an abnormal action event and / or an abnormal operation electrical signal event; otherwise, it is determined that the corresponding hardware does not have an abnormal action event and / or an abnormal operation electrical signal event.

[0063] Furthermore, based on the hardware equipment type, process location, duration, magnitude, and operating parameters (such as load and temperature) at the time of the abnormal event and / or abnormal electrical signal event, the suspicious process is predicted for future time periods to determine the recurrence time-domain attributes of the aforementioned potential abnormal events (such as recurrence frequency and / or minimum recurrence interval). Specifically, the above-mentioned feature elements are input into a pre-trained time series prediction model (such as an LSTM or Transformer model). This model is trained based on historical fault data and outputs a probability time series of the same or similar abnormal events occurring in the future. From this, the recurrence frequency (the number of times expected to occur per unit time) and the minimum recurrence interval (the minimum time difference between two adjacent abnormalities) are extracted. If the recurrence frequency exceeds a preset frequency threshold (e.g., more than 0.5 times per hour) and / or the shortest recurrence interval is lower than a preset interval threshold (e.g., less than 10 minutes), then the aforementioned potential abnormal events are determined to be actual faults occurring on the industrial operation line; otherwise, the aforementioned potential abnormal events are determined not to be actual faults occurring on the industrial operation line. Then, the event type of the potential abnormal events that are actual faults occurring on the industrial operation line is obtained (e.g., the specific location attributes of the event occurrence at the hardware or software level and the duration of the event occurrence, etc.). The aforementioned event types are encoded, and attribute codes containing information such as fault category, subcategory, and severity level are generated according to the preset fault classification system. For example, a three-segment code of "category-subcategory-level" is used to provide a basis for subsequent accurate determination of hardware equipment fault conditions and linkage adjustments.

[0064] In another embodiment, the step of extracting hardware action information and hardware operating electrical parameters from the hardware operation records related to the suspicious process contained in the status data, thereby determining potential abnormal events of the hardware related to the suspicious process, can also be implemented as follows:

[0065] For each piece of hardware j related to the suspected procedure, execute:

[0066] Step A1: Acquire the electrical parameters of hardware j at a fixed sampling period. The electrical parameters corresponding to the t-th sampling point of hardware j... When missing, the multi-source fusion estimate of the electrical parameters of hardware j at the t-th sampling point is determined based on the observed electrical parameters of hardware j at neighboring times, the mean electrical parameters of hardware j under normal operating conditions during the same historical period, the personalized time weight of hardware j, and the historical pattern confidence coefficient of hardware j. .

[0067] Specifically, it can be determined according to the following formula. :

[0068]

[0069] in, The sampling point number is a positive integer. Let be the set of neighborhood sampling point indices of the t-th sampling point, containing m sampling points before and after t, where m is a preset positive integer;

[0070] Personalized time weights for hardware j, used to quantize neighborhood sampling points. For the current sampling point The extent of their contribution. Let j be the time decay constant of the hardware. The decay rate is due to Control. When At that time, the weights exactly decayed to their initial values. ,Right now The physical meaning is that the weight decays to The required time step. The larger the value, the slower the weight decay, indicating a smoother change in hardware status and a longer duration of historical data reference value.

[0071] The value is determined based on the standard duration of the process in which hardware j is located, specifically by multiplying the number of sampling points included in the standard duration of that process by an adaptive coefficient. ,Right now ,in This refers to the number of sampling points included in the standard process duration. This is the dynamic adjustment coefficient for hardware j, preset based on the degree of change in the hardware during the process (the smaller value is used for those with more drastic changes); if the process duration is not fixed, the number of sampling points contained in the median of multiple historical cycle durations is taken as the coefficient. .

[0072] Let be the historical pattern confidence coefficient of hardware j. Its physical meaning is the relative importance of historical data during interpolation. It is a positive real number and is determined based on the coefficient of variation of the electrical parameters of hardware j during its historical normal operation. Specifically, its value is the reciprocal of the coefficient of variation. ,in , Let J be the standard deviation of the electrical parameters of hardware j under normal operating conditions during the same historical period. This is the corresponding mean; The average electrical parameters of hardware j under normal operating conditions during the same historical period are obtained by averaging the sampled values ​​at the same time during the historical fault-free period.

[0073] like If there are no missing parts, then let .

[0074] In this invention, the concept of historical contemporaneous period can be defined based on the data acquisition method of this invention—fixed sampling period. Let the sampling point number be... ( (The historical period refers to the same sampling point sequence number in multiple past production cycles.) All historical data points. In industrial production line scenarios, because the production process is usually repeated according to a fixed sequence of operations and cycle time, the same sampling point number... This corresponds to the same process stage (e.g., always corresponding to "2 seconds after the welding process is completed"). Therefore, "historical data from the same period" essentially reflects the normal operating mode of the hardware equipment under the same process stage and operating conditions. This is achieved by statistically averaging these historical data from the same period. The embodiments of the present invention can construct a dynamic benchmark with the same "business context" as the current moment, providing a reliable reference for subsequent data repair and anomaly detection.

[0075] Step A1 Explanation: In industrial data acquisition, data loss often occurs due to sensor malfunctions or communication anomalies. Directly discarding missing data or using simple interpolation would disrupt the temporal continuity and periodicity of the data. Therefore, Step A1 proposes a data repair method based on a weighted fusion of temporal proximity and historical patterns. This step first obtains the electrical parameter observations of hardware j at neighboring times, and simultaneously obtains the average electrical parameter values ​​of the hardware under normal operating conditions during the same historical period. Then, personalized time weights are applied... To control the influence of neighboring observations, the historical model confidence coefficient was used. Adjusting the reliability of historical data from the same period, the two are weighted and fused to obtain a multi-source fusion estimate. Among them, the time decay constant Determined based on process duration, reflecting the urgency of hardware status changes; Trust coefficient. Determined based on hardware operational stability, this step reflects the reliability of historical data from the same period. It adapts to the dynamic characteristics and stability differences of various hardware, generating accurate and reliable interpolated values ​​when data is missing. This avoids information loss associated with traditional simple interpolation methods, providing a complete and continuous data foundation for subsequent anomaly detection and improving data quality in harsh industrial operation environments.

[0076] Step A2: Based on the electrical parameters of hardware j at the t-th sampling point and multi-source fusion estimation value Determine the residual of hardware j at the t-th sampling point. Based on the residual of hardware j at the t-th sampling point Dynamic deviation feature of the previous sampling point And the personalized smoothing coefficient of hardware j Determine the dynamic deviation characteristic of hardware j at the t-th sampling point. .

[0077] Specifically, it can be determined according to the following formula. :

[0078]

[0079] in, is the personalized smoothing coefficient for hardware j. Its physical meaning is the weight of the current residual in the dynamic deviation characteristic. The value range is between (0,1) and can be set according to the noise level of hardware j.

[0080] Among them, initialization ;

[0081] like If so, it is determined that hardware j has abnormal electrical parameter sampling at the t-th sampling point, and an abnormality marker is recorded. Otherwise, record. Regardless of the presence or absence of anomaly markers, the dynamic deviation feature of all sampling points All proceed to step A3 to participate in the comprehensive anomaly score calculation.

[0082] in, The personalized control limit coefficient for hardware j is a multiple threshold of the normal fluctuation range. The value range is a positive real number. It is preset according to the importance level of hardware j in the industrial operation line. The preferred values ​​are 2.5 for key hardware, 3.0 for important hardware, and 3.5 for general hardware. Let be the standard deviation of the residuals of hardware j during the historical fault-free period, obtained by statistical analysis of the residual sequence during the historical fault-free period. .

[0083] Step A2 Explanation: The repaired data may still contain abnormal spikes caused by transient disturbances. Directly using single-point residuals is susceptible to noise and can lead to misjudgments. Therefore, Step A2 uses an exponentially weighted moving average method to construct a dynamic deviation characteristic. This step first calculates the residual between the current observation and the repair baseline. This reflects the degree of real-time deviation; then, it is smoothed using an exponential smoothing formula. The residual sequence is smoothed, giving higher weights to recent residuals while retaining historical cumulative information, to obtain the dynamic deviation feature. Personalized smoothing coefficient This setting can be adjusted based on the hardware noise level; a smaller value should be used when the noise level is high to enhance the smoothing effect. Standard deviation of residuals under normal operating conditions When the deviation exceeds the personalized control limit coefficient, a comparison is made. When the sample size exceeds one-time standard deviation, anomalies are identified and marked as abnormal. This method effectively captures persistent abnormal trends while suppressing random noise interference, achieving a balance between sensitivity and stability.

[0084] Step A3: Based on the dynamic deviation feature of hardware j at the t-th sampling point The standard deviation of the residuals of hardware j during its historical fault-free period. Action information of hardware j at the t-th sampling point The average value of hardware j's action information during historical fault-free periods. and standard deviation and the anomaly markers generated in step A2 Determine the comprehensive anomaly score of hardware j at the t-th sampling point. .

[0085] Specifically, calculate according to the following formula :

[0086]

[0087] in, The action information of hardware j at the t-th sampling point refers to the motion state parameters of the hardware during operation, including at least one of displacement, velocity, acceleration, angle or torque, which is obtained through periodic sampling. and These are the mean and standard deviation of the action information of hardware j during the historical fault-free period, respectively, which are obtained by statistical analysis of the action information sequence during the historical fault-free period. The first preset prevention and elimination item and the second preset release and elimination item are respectively represented by units of 1 and 2. , Since the units are the same, all are taken as the smallest positive values. This can be preset based on experience to prevent the denominator from being zero and thus making the calculation meaningless. For example, the value can be taken as... ; , , respectively, represent the importance weights of the electrical parameters and motion information of hardware j, and respectively, indicating the relative importance of electrical parameters and motion information in the comprehensive anomaly score. , The range of values ​​for all values ​​is positive real numbers and satisfies The specific value can be determined through empirical setting or regression analysis based on historical failure samples; The enhancement coefficient for anomaly labeling refers to the weighting of electrical parameters for points marked as anomalous samples. The multiplier added to the base value is a positive real number, typically between 0.2 and 0.5. The specific value can be preset based on the hardware's sensitivity requirements for anomaly sampling; higher sensitivity requirements result in a higher multiplier. The larger the value; If step A2 determines that the sampling is abnormal, then... ,otherwise .

[0088] like Greater than the preset personalized threshold If , then it is determined that hardware j has a potential abnormal event at the t-th sampling point; where The threshold for determining the anomaly of hardware j is the critical value of the comprehensive anomaly score. It can be set according to the score distribution of historical normal data, such as the 95th percentile of the score sequence, or the mean plus three times the standard deviation, to ensure that the false alarm rate is controlled within the preset range.

[0089] The historical fault-free period refers to the time interval in which the hardware j has been confirmed to be fault-free in its past operation records. It is usually a continuous time data during the stable operation of the equipment, which is used to statistically analyze the fluctuation characteristics under normal conditions.

[0090] Step A3 Explanation: Judging anomalies solely based on electrical parameters may lead to misjudgments due to electromagnetic interference, while motion information reflects the mechanical execution state; the two types of information are complementary. Therefore, Step A3 proposes a multimodal fusion-based comprehensive anomaly scoring method. This step integrates electrical anomaly scores... and abnormality of movement After normalization, the results are weighted and summed, and the anomaly markers generated in step A2 are introduced. The weights of the electrical components are dynamically enhanced through personalized weight fusion. , Add weights to anomaly markers Adjusting the relative importance of the two types of information, a comprehensive anomaly score is finally obtained. The score will be compared with a preset threshold. By comparing data, it is possible to determine whether there are potential abnormal events in the hardware, thereby identifying complex faults that cannot be detected by a single dimension, and improving the accuracy and reliability of anomaly detection. This method achieves multi-source information fusion, improving the accuracy and comprehensiveness of anomaly detection. At the same time, the introduction of anomaly marking provides a dynamic enhancement mechanism for the marked points, automatically increasing the weight of electrical parameters when there are obvious anomalies, and relying more on action information when the electrical parameters are normal, thus realizing adaptive adjustment of the fusion strategy.

[0091] Steps A1-A3 form an organic whole, utilizing a progressive data processing architecture to create a complete potential anomaly event identification chain, from data repair to single-modal anomaly detection and then to multi-modal fusion judgment. This chain has the following advantages: First, all key parameters are individually designed, adapting to the dynamic characteristics, noise levels, and stability differences of different hardware, avoiding the inadequacy of traditional one-size-fits-all methods in diverse industrial scenarios; second, it has low computational complexity, with all operations being linear or simple statistical, allowing for efficient embedding into the real-time data processing flow of industrial control systems, meeting the monitoring requirements of industrial scenarios for low latency and high throughput; third, it directly outputs clear potential anomaly judgment results through comprehensive anomaly scoring, facilitating subsequent system linkage adjustments or alarms, achieving a technological leap from passive maintenance to proactive early warning, and improving the operational continuity and reliability of industrial operation lines.

[0092] In another embodiment, faulty hardware operation records of the industrial operation line are determined based on fault attribute codes and a cloud database; by comparing the faulty hardware operation records with suspected processes, coordinated adjustments are implemented for the target range of the industrial operation line, including:

[0093] By comparing the fault attribute codes with the cloud database, the actual hardware where the fault occurred is extracted from the cloud database; based on the correlation between the operating inputs and outputs of the hardware and the execution of the suspected process, the faulty hardware work records of the industrial operation line are determined.

[0094] By comparing the work records of faulty hardware with the suspected processes, the operating path of the faulty hardware in the suspected process is determined. Based on this, the operating path of the faulty hardware is adjusted in a coordinated manner within the target area covered by the industrial operation line. The coordinated adjustment refers to the coordinated adjustment of the operating modes of different faulty hardware on the operating path of the faulty hardware.

[0095] The above technical solution will be explained below:

[0096] First, the fault attribute code is used as an index to perform a matching query in a preset fault-hardware mapping table in the cloud database to determine the hardware where the actual fault occurred. This fault-hardware mapping table is pre-established based on historical fault data, recording the correspondence between fault types and potentially involved hardware identifiers. Then, the instruction inputs and response outputs during the operation of the affected hardware are cross-correlation analyzed or dynamically time-warped (DTW) with the theoretical input-output baseline of the hardware during the execution of the suspected process. The similarity or distance between the two is calculated as the degree of association; the theoretical baseline is derived from the equipment design specifications or a statistical model of historical normal operation data. If the degree of association exceeds a preset threshold, it is determined that the affected hardware directly caused the fault. In this case, detailed work logs, sensor data, and control instruction sequences of the affected hardware within one time window before and after the fault are extracted from the cloud database as the fault hardware's work record. The preset threshold is set based on the association distribution of similar hardware and processes in historical normal data. If the degree of association does not exceed the preset threshold, it is determined that the affected hardware did not directly cause the fault, and its work record is not extracted from the cloud database.

[0097] Next, by comparing the faulty hardware's work record with the suspected process, and based on the physical connection topology of the industrial operation line and the process flow sequence, combined with the hardware's action timestamps and signal flow directions in the work record, the data interaction links and physical action links between the relevant hardware at the time of the fault occurrence are identified, thus obtaining the faulty hardware's operating path for the suspected process. The faulty hardware's operating path refers to the causally related operational links formed by all faulty hardware within the suspected process during operation, characterizing the correlation between all faulty hardware within the suspected process in the target object's operation processing and / or electrical signals.

[0098] Finally, coordinated adjustments are implemented for the faulty hardware's operating path within the target area covered by the industrial operation line. Coordinated adjustment refers to the collaborative adjustment of the operating modes of different faulty hardware devices along the faulty hardware's operating path. This can be achieved through two methods: rule-based and optimization model-driven approaches.

[0099] (1) Rule-based driven mode: A linkage control rule base is pre-established. Each rule includes triggering conditions (such as fault type and severity), adjustment objects, adjustment actions, and the method for determining the adjustment amount. Adjustment actions include reducing running speed, limiting output torque, delaying command sending time, and modifying motion trajectory. The adjustment amount can be calculated according to the fault severity using empirical formulas or determined by looking up a table. When a fault occurs, the faulty hardware set is traversed, the rules in the rule base are matched, and the adjustment amount of each hardware is calculated. If there are rule conflicts, they are handled according to priority or transferred to the optimization model.

[0100] (2) Optimization Model-Driven Mode: For complex fault scenarios that cannot be covered by the rule base, an optimization model is built in real time with the goal of minimizing the impact on the production process and the adjustment cost. The model considers the impact of each hardware adjustment on the process output, the adjustment cost, hardware safety constraints and collaborative constraints, and solves for the optimal adjustment amount through numerical optimization algorithms.

[0101] The calculated adjustments are converted into specific control commands, which are then sent to each hardware controller via industrial bus or industrial IoT. Command types include speed settings, acceleration settings, torque limits, command delays, trajectory corrections, and mode switching. Each hardware component adjusts its operating parameters according to the commands, achieving coordinated degraded operation. This ensures that the industrial production line can continue operating after a failure or buys time for maintenance, avoiding a complete shutdown and improving the accuracy and reliability of fault handling.

[0102] The above embodiments achieve the following beneficial effects:

[0103] This approach solves the problem of inefficient, manual troubleshooting. By using fault attribute codes as indexes, it matches the actual hardware where the fault occurred in a cloud database, replacing the traditional method of manual, step-by-step troubleshooting. Furthermore, by quantifying the correlation between hardware operating instructions and the execution of suspicious procedures, faulty hardware is only identified when the correlation exceeds a preset threshold. This avoids misjudgments and omissions caused by insufficient human experience or incomplete information, significantly improving the efficiency and accuracy of fault location.

[0104] This approach addresses the issue of fault handling focusing only on a single hardware component and failing to consider related devices. By comparing the faulty hardware's operational records and suspected procedures, it identifies the data interaction and physical interaction links between relevant hardware components at the moment of the fault, constructing a causal operational path for the faulty hardware. This path reveals the propagation relationships and mutual influences of the fault among the hardware components, providing a global perspective for subsequent coordinated adjustments and avoiding localized handling methods.

[0105] This solution addresses the problem of maintenance requiring downtime and disrupting production continuity. By implementing coordinated adjustments to the operating modes of different hardware components along the faulty hardware's path, the industrial production line can continue operating in a degraded mode after a fault occurs, or buy valuable time for maintenance, thus avoiding production interruptions and economic losses caused by a complete line shutdown.

[0106] This solves the problem of the inability to coordinate hardware adjustments for known faults. By employing both rule-based and optimization model-driven approaches, it enables coordinated adjustments to multiple faulty hardware components along the path (such as adjusting command interaction timing, hardware action amplitude, and attitude). This represents a technological leap from single-device adjustment to multi-device coordinated control, ensuring the overall accuracy and reliability of fault handling on the operation line.

[0107] Please see Figure 2 As shown, an embodiment of this application provides an industrial linkage control system based on a cloud database. This cloud-based industrial linkage control system includes:

[0108] The process calibration module is used to track and analyze the status data of target objects on the industrial operation line, thereby calibrating suspicious processes on the industrial operation line;

[0109] The process status lookup module is used to search for status data related to suspicious processes from the cloud database based on the operation location tags of suspicious processes;

[0110] The potential anomaly identification module is used to identify potential abnormal events on the industrial operation line based on condition data.

[0111] The fault attribute determination module is used to reproduce and identify potential abnormal events and generate fault attribute codes for industrial operation lines.

[0112] The work record determination module is used to determine the fault hardware work records of the industrial operation line based on the fault attribute code and the cloud database.

[0113] The linkage adjustment module is used to compare the working records of faulty hardware with suspicious processes and implement linkage adjustments to the target range of the industrial operation line.

[0114] The beneficial effects of the above embodiments are that the cloud-based industrial linkage control system tracks the target objects on the industrial operation line to check all processes, identifies the fault attributes corresponding to potential abnormal events, and combines the cloud database to perform hardware fault diagnosis and calibration, ensuring the accuracy and reliability of the overall fault handling of the operation line.

[0115] In another embodiment, the process calibration module is used to track and analyze the status data of targets on the industrial operation line, thereby calibrating suspicious processes on the industrial operation line, including:

[0116] Track the entire process image of the target object on the industrial operation line, divide and identify the process intervals of the entire process image, and obtain the structural features and manipulated features of the target object for each process.

[0117] By comparing the structural features and the manipulated features against a benchmark, it can be determined whether the process is a suspicious process.

[0118] The process status lookup module is used to search for status data related to suspicious processes from the cloud database based on the operation location tags of the suspicious processes, including:

[0119] Based on the spatial attributes and hardware configuration attributes of the suspicious process on the industrial operation line, an operation location tag for the suspicious process is generated; the operation location tag is compared with the cloud database to find the status data related to the suspicious process; the status data includes the hardware operation records related to the suspicious process.

[0120] In another embodiment, the potential anomaly determination module is used to determine potential abnormal events on the industrial operation line based on condition data, including:

[0121] Hardware action information and hardware operating electrical parameters are extracted from the hardware operation records related to the suspicious processes contained in the status data to determine potential abnormal events of the hardware related to the suspicious processes; among which, potential abnormal events include abnormal action events and / or abnormal operating electrical signal events of the hardware related to the suspicious processes.

[0122] The fault attribute determination module is used to reproduce and identify potential abnormal events, and generate fault attribute codes for industrial operation lines, including:

[0123] Construct feature characteristics of potential abnormal events, identify recurrence based on feature characteristics, and determine the recurrence time domain attribute of potential abnormal events; based on the recurrence time domain attribute, identify the attribute code of the actual fault occurring on the industrial operation line; wherein, the attribute code contains a characterization label of the type of actual fault.

[0124] In another embodiment, the work record determination module is used to determine the faulty hardware work records of the industrial operation line based on the fault attribute code and the cloud database, including:

[0125] By comparing the fault attribute codes with the cloud database, the actual hardware where the fault occurred is extracted from the cloud database; based on the correlation between the operating inputs and outputs of the hardware and the execution of the suspected process, the faulty hardware work records of the industrial operation line are determined.

[0126] The linkage adjustment module is used to compare faulty hardware operation records and suspicious processes to implement linkage adjustments to target ranges of industrial operation lines, including:

[0127] By comparing the work records of faulty hardware with the suspected processes, the operating path of the faulty hardware in the suspected process is determined. Based on this, the operating path of the faulty hardware is adjusted in a coordinated manner within the target area covered by the industrial operation line. The coordinated adjustment refers to the coordinated adjustment of the operating modes of different faulty hardware on the operating path of the faulty hardware.

[0128] The operation and effect of the cloud-based industrial linkage control system of the present invention are consistent with the above-mentioned cloud-based industrial linkage control method, and the description of the cloud-based industrial linkage control system will not be repeated here.

[0129] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. This disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims. Thus, if these modifications and variations of the invention fall within the scope of the claims of the invention and their equivalents, the invention is also intended to include these modifications and variations.

Claims

1. An industrial linkage control method based on a cloud database, characterized in that, include: Track and analyze the status data of target objects on the industrial operation line to identify suspicious processes on the industrial operation line; Based on the operation location tags of the suspected process, retrieve status data related to the suspected process from the cloud database; Based on the condition data, potential abnormal events of the industrial operation line are identified; The potential abnormal events are reproduced and identified to generate fault attribute codes for the industrial operation line. Based on the fault attribute code and the cloud database, determine the faulty hardware operation record of the industrial operation line; compare the faulty hardware operation record with the suspected process, and implement coordinated adjustment of the target range of the industrial operation line.

2. The industrial linkage control method based on a cloud database as described in claim 1, characterized in that: Track and analyze the status data of target objects on the industrial operation line to identify suspicious processes on the industrial operation line; Based on the operation location tag of the suspected process, retrieve status data related to the suspected process from the cloud database, including: Track the entire process image of the target object on the industrial operation line, divide and identify the process intervals of the entire process image, and obtain the structural features and manipulated features of the target object for each process. A benchmark comparison is performed between the constructed features and the operated features to determine whether the process is a suspicious process. Based on the spatial attributes and hardware configuration attributes of the suspected process on the industrial operation line, an operation location tag for the suspected process is formed; the operation location tag is compared with the cloud database to find status data related to the suspected process; wherein, the status data includes hardware operation records related to the suspected process.

3. The industrial linkage control method based on a cloud database as described in claim 1, characterized in that: Based on the condition data, potential abnormal events of the industrial operation line are identified; The potential abnormal events are reproduced and identified to generate fault attribute codes for the industrial operation line, including: Hardware action information and hardware operating electrical parameters are extracted from the hardware operation records related to the suspected process contained in the status data to determine potential abnormal events of the hardware related to the suspected process; wherein, the potential abnormal events include abnormal action events and / or abnormal operating electrical signal events of the hardware related to the suspected process. The potential abnormal events are constructed as feature elements, and recurrence identification is performed based on the feature elements to determine the recurrence time-domain attribute of the potential abnormal events. Based on the recurrence time-domain attribute, the attribute code of the actual fault occurring on the industrial operation line is identified. The attribute code contains a characterization label of the type of the actual fault.

4. The industrial linkage control method based on a cloud database as described in claim 1, characterized in that: Based on the fault attribute code and the cloud database, determine the faulty hardware operation record of the industrial operation line; By comparing the faulty hardware operation record with the suspected process, a coordinated adjustment is implemented for the target range of the industrial operation line, including: By comparing the fault attribute code with the cloud database, the hardware where the actual fault occurred is extracted from the cloud database; based on the correlation between the operating input and output of the hardware and the execution of the suspected process, the faulty hardware working record of the industrial operation line is determined. By comparing the faulty hardware operation record with the suspected process, the faulty hardware operation path of the suspected process is determined, and the faulty hardware operation path is adjusted in a coordinated manner within the target area covered by the industrial operation line; wherein, the coordinated adjustment refers to the coordinated adjustment of the operation mode of different faulty hardware on the faulty hardware operation path.

5. An industrial linkage control system based on a cloud database, characterized in that, include: The process calibration module is used to track and analyze the status data of target objects on the industrial operation line, thereby calibrating the suspicious processes on the industrial operation line. The process status lookup module is used to search for status data related to the suspicious process from the cloud database based on the operation location tag of the suspicious process; A potential anomaly determination module is used to determine potential abnormal events of the industrial operation line based on the status data. The fault attribute determination module is used to reproduce and identify the potential abnormal events and generate the fault attribute code of the industrial operation line. The work record determination module is used to determine the fault hardware work record of the industrial operation line based on the fault attribute code and the cloud database. The linkage adjustment module is used to compare the faulty hardware operation record with the suspicious process and implement linkage adjustment on the target range of the industrial operation line.

6. The industrial linkage control system based on a cloud database as described in claim 5, characterized in that: The process calibration module is used to track and analyze the status data of target objects on the industrial operation line, thereby calibrating suspicious processes on the industrial operation line, including: Track the entire process image of the target object on the industrial operation line, divide and identify the process intervals of the entire process image, and obtain the structural features and manipulated features of the target object for each process. A benchmark comparison is performed between the constructed features and the operated features to determine whether the process is a suspicious process. The process status lookup module is used to search for status data related to the suspicious process from the cloud database based on the operation location tag of the suspicious process, including: Based on the spatial attributes and hardware configuration attributes of the suspected process on the industrial operation line, an operation location tag for the suspected process is formed; the operation location tag is compared with the cloud database to find status data related to the suspected process; wherein, the status data includes hardware operation records related to the suspected process.

7. The industrial linkage control system based on a cloud database as described in claim 5, characterized in that: The potential anomaly determination module is used to determine potential abnormal events of the industrial operation line based on the status data, including: Hardware action information and hardware operating electrical parameters are extracted from the hardware operation records related to the suspected process contained in the status data to determine potential abnormal events of the hardware related to the suspected process; wherein, the potential abnormal events include abnormal action events and / or abnormal operating electrical signal events of the hardware related to the suspected process. The fault attribute determination module is used to reproduce and identify the potential abnormal events, and generate fault attribute codes for the industrial operation line, including: The potential abnormal events are constructed as feature elements, and recurrence identification is performed based on the feature elements to determine the recurrence time-domain attribute of the potential abnormal events. Based on the recurrence time-domain attribute, the attribute code of the actual fault occurring on the industrial operation line is identified. The attribute code contains a characterization label of the type of the actual fault.

8. The industrial linkage control system based on a cloud database as described in claim 5, characterized in that: The work record determination module is used to determine the fault hardware work record of the industrial operation line based on the fault attribute code and the cloud database, including: By comparing the fault attribute code with the cloud database, the hardware where the actual fault occurred is extracted from the cloud database; based on the correlation between the operating input and output of the hardware and the execution of the suspected process, the faulty hardware working record of the industrial operation line is determined. The linkage adjustment module is used to compare the faulty hardware operation record with the suspicious process, and to implement linkage adjustment on the target range of the industrial operation line, including: By comparing the faulty hardware operation record with the suspected process, the faulty hardware operation path of the suspected process is determined, and the faulty hardware operation path is adjusted in a coordinated manner within the target area covered by the industrial operation line; wherein, the coordinated adjustment refers to the coordinated adjustment of the operation mode of different faulty hardware on the faulty hardware operation path.

Images