A production equipment state monitoring and early warning method and system based on cloud edge-end cooperation

Abstract

The application provides a production equipment state monitoring and early warning method and system based on cloud edge end cooperation, multi-source time sequence operation data of a target production equipment is collected, and data filtering is performed on the multi-source time sequence operation data to obtain an edge feature data set; a fault correlation feature sequence is generated based on a historical fault case library and the edge feature data set; a state evolution track of the target production equipment is constructed according to the deviation degree of the fault correlation feature sequence and a device operation standard template, and a preliminary early warning instruction and an adapted early warning response strategy are generated according to the matching result of the state evolution track and a preset early warning threshold; the edge end combines the multi-source time sequence operation data collected in real time to locally verify the preliminary early warning instruction, and generates a monitoring and early warning signal of the target production equipment according to the early warning verification result and the early warning response strategy. The technical scheme provided by the application can realize state early warning of the target production equipment under the difference of cloud edge environment and information asymmetry.

Description

Technical Field

[0001] This application relates to the field of equipment monitoring and early warning technology, and more specifically, to a method and system for monitoring and early warning of production equipment status based on cloud-edge-device collaboration. Background Technology

[0002] With the rapid development of industrial internet and IoT technologies, equipment monitoring and early warning has become a core component of industrial intelligence. By deploying various sensors, the system can collect multi-dimensional operational data of equipment in real time and transmit it to cloud platforms or edge computing nodes via the network. The platform uses big data analysis and machine learning models to process and recognize the data in real time, accurately assessing the health status of the equipment. Once abnormal features or performance degradation trends are detected, the system automatically triggers tiered early warnings, thereby effectively ensuring production safety, improving operational efficiency, and reducing maintenance costs.

[0003] In existing equipment monitoring and early warning systems, key features reflecting equipment operating status are extracted from real-time sensor data. These real-time features are then compared and matched with preset thresholds, static rule bases, or dynamic model benchmarks trained on historical data. When feature values ​​exceed safe ranges, conform to specific fault modes, or exhibit their changing trends, performance degradation is predicted, and the system automatically triggers corresponding level early warning signals according to preset rules, thereby achieving early identification of abnormal equipment conditions. However, in cloud-edge-device collaborative production equipment status monitoring and early warning systems, there are issues of cloud-edge environment differences and information asymmetry. Early warning decisions rely entirely on data uploaded to the cloud and static models trained on historical global data. In actual industrial production, the specific operating conditions, environmental interference, and instantaneous states of edge devices are highly dynamic and specific, resulting in differences between the current real-time operating data distribution of the equipment and the historical data distribution used for cloud model training. Furthermore, the degradation mode of the equipment may deviate from the preset template due to changes in operating conditions. This mismatch can lead to false alarms in early warning commands issued by the cloud. Therefore, how to achieve status early warning for target production equipment under cloud-edge environment differences and information asymmetry has become a challenge for the industry. Summary of the Invention

[0004] This application provides a method and system for monitoring and early warning of the status of production equipment based on cloud-edge-device collaboration, which can realize the status warning of target production equipment under the conditions of cloud-edge environment differences and information asymmetry.

[0005] Firstly, this application provides a method for monitoring and early warning of production equipment status based on cloud-edge-device collaboration, comprising the following steps: Multi-source time-series operation data of the target production equipment are collected at the edge, and the multi-source time-series operation data is filtered based on the equipment operation characteristics to obtain an edge feature dataset; The edge feature dataset is associated with the device identification information of the target generation device and uploaded to the cloud. Fault association analysis is performed on the edge feature dataset in the cloud based on the historical fault case library, thereby generating a fault association feature sequence with time stamp. Based on the degree of deviation between the fault-related feature sequence and the equipment operation standard template, the state evolution trajectory of the target production equipment is constructed in the cloud, and based on the matching result of the state evolution trajectory and the preset warning threshold, preliminary warning instructions and adapted warning response strategies are generated. The preliminary warning command and warning response strategy are sent to the corresponding edge device. The edge device performs local verification of the preliminary warning command by combining the real-time collected multi-source time-series operation data, and generates a monitoring warning signal for the target production equipment based on the warning verification result and the warning response strategy.

[0006] In some embodiments, performing data filtering based on device operation characteristics on the multi-source time-series operational data to obtain an edge feature dataset specifically includes: Extract all operational feature dimensions of the target production equipment; Invalid data is removed from the multi-source time-series runtime data based on all runtime feature dimensions to obtain a preprocessed time-series dataset. The preprocessed time series dataset is used to extract time series feature parameters that match the operational feature dimensions of the target production equipment to form an edge feature dataset.

[0007] In some embodiments, associating the edge feature dataset with the device identification information of the target generating device and uploading it to the cloud specifically includes: Obtain the equipment identification information of the target production equipment; The device identification information is used to perform label embedding processing on the edge feature dataset to obtain a feature dataset with identification labels; Perform cloud-transmission-adaptive serialization encoding on the labeled feature dataset to obtain the encoded transmission dataset; The encoded transmission dataset is uploaded to a designated cloud storage node based on a preset cloud communication protocol.

[0008] In some embodiments, performing fault correlation analysis on the edge feature dataset based on a historical fault case library in the cloud, and then generating a fault correlation feature sequence with time stamps, specifically includes: Obtain a historical fault case library of the target production equipment, and construct a standardized fault feature index for the historical fault case library to generate a fault feature index library; A feature similarity matching algorithm is used to compare and associate the edge feature dataset with the fault feature index to obtain preliminary fault association results; Extract the time-series timestamps from the edge feature dataset and perform time dimension calibration on the preliminary fault association results to obtain the time-series association feature set; The time-series associated feature set is serialized and encapsulated to generate a fault-associated feature sequence with time stamps.

[0009] In some embodiments, constructing the state evolution trajectory of the target production equipment in the cloud based on the degree of deviation between the fault association feature sequence and the equipment operation standard template specifically includes: Obtain the equipment operation standard template corresponding to the target production equipment, and perform feature standardization processing on the equipment operation standard template to generate a standard feature benchmark set; The deviation between the fault-related feature sequence and the standard feature benchmark set is calculated on a time-series basis to obtain the time-series deviation dataset; Extract the time stamp information and corresponding deviation values ​​from the time series deviation dataset to generate trajectory data sources; Based on the trajectory, a data source is constructed to fit the state evolution path of the target production equipment, generating the state evolution trajectory of the target production equipment.

[0010] In some embodiments, the preliminary warning instruction and warning response strategy are sent to the corresponding edge device, and the edge device performs local verification of the preliminary warning instruction in conjunction with real-time collected multi-source time-series operational data. Specifically, this includes: Obtain the device identification information at the target edge to generate an edge communication index; Based on the edge communication index, the preliminary early warning command and early warning response strategy are encapsulated and transmitted to obtain the data packet to be sent. The data packets are transmitted to the corresponding target edge terminal via the cloud-edge communication link. The target edge terminal receives the data packets and parses and restores them to obtain preliminary warning instructions and warning response strategies. The target edge collects multi-source time-series operation data of the target production equipment in real time and generates a real-time verification dataset; Based on the verification rules in the early warning response strategy, the real-time verification dataset is matched and verified with the feature conditions of the preliminary early warning instruction to obtain the early warning verification result.

[0011] In some embodiments, generating a monitoring and early warning signal for the target production equipment based on the early warning verification result and the early warning response strategy specifically includes: The validity of the early warning verification results is determined, and a verification validity determination result is generated. Based on the verification validity determination result, the signal generation rules in the early warning response strategy are matched to obtain the signal generation rule set; Based on the signal generation rule set, the verification validity determination result is mapped to the warning signal parameters to generate warning signal parameters; The early warning signal parameters are standardized and encoded to generate monitoring and early warning signals for the target production equipment.

[0012] Secondly, this application provides a production equipment status monitoring and early warning system based on cloud-edge-device collaboration, used to execute a production equipment status monitoring and early warning method based on cloud-edge-device collaboration. The system includes: The acquisition module is used to acquire multi-source time-series operation data of the target production equipment at the edge, and to perform data filtering based on equipment operation characteristics on the multi-source time-series operation data to obtain an edge feature dataset; The processing module is used to associate the edge feature dataset with the device identification information of the target generation device and upload it to the cloud. On the cloud, the edge feature dataset is analyzed for fault association based on the historical fault case library, thereby generating a fault association feature sequence with time stamp. The processing module is also used to construct the state evolution trajectory of the target production equipment in the cloud according to the degree of deviation between the fault association feature sequence and the equipment operation standard template, and generate preliminary warning instructions and adapted warning response strategies based on the matching results of the state evolution trajectory and the preset warning threshold. The execution module is used to send the preliminary warning command and warning response strategy to the corresponding edge terminal. The edge terminal combines the real-time collected multi-source time-series operation data to perform local verification of the preliminary warning command, and generates a monitoring warning signal for the target production equipment based on the warning verification result and the warning response strategy.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described cloud-edge-device collaborative production equipment status monitoring and early warning method.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for monitoring and warning the status of production equipment based on cloud-edge-device collaboration.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The cloud-edge-device collaborative production equipment status monitoring and early warning method and system provided in this application firstly collects multi-source time-series operational data of the target production equipment at the edge, and performs data filtering based on equipment operation characteristics to obtain an edge feature dataset. Secondly, the edge feature dataset is associated with the equipment identification information of the target production equipment and uploaded to the cloud. In the cloud, fault correlation analysis is performed on the edge feature dataset based on a historical fault case library to generate a fault correlation feature sequence with time stamps. Then, based on the degree of deviation between the fault correlation feature sequence and the equipment operation standard template, the state evolution trajectory of the target production equipment is constructed in the cloud, and a preliminary early warning instruction and an adapted early warning response strategy are generated based on the matching result of the state evolution trajectory and a preset early warning threshold. Finally, the preliminary early warning instruction and early warning response strategy are sent to the corresponding edge, and the edge verifies the preliminary early warning instruction locally in conjunction with the real-time collected multi-source time-series operational data, and generates a monitoring and early warning signal for the target production equipment based on the early warning verification result and the early warning response strategy.

[0016] Therefore, this application can achieve status early warning of target production equipment under the conditions of cloud-edge environment differences and information asymmetry. First, the edge end filters multi-source time-series operation data based on equipment operation characteristics and generates an edge feature dataset, which can reduce the data transmission bandwidth occupation of the cloud-edge end and the cloud computing load, and improve the efficiency and targeting of data preprocessing. Second, after associating the edge feature dataset with equipment identification information, it is uploaded to the cloud and fault correlation analysis is carried out to generate fault correlation feature sequences with time stamps. This enables accurate traceability of equipment operation data. Combined with matching analysis of historical fault case database, it improves the accuracy of fault feature identification. The time stamp provides data support for subsequent time-series judgment of equipment status evolution. Then, based on the degree of deviation between the fault correlation feature sequence and the equipment operation standard template, a system is constructed. The system tracks the state evolution trajectory and generates preliminary warning commands and adaptation strategies based on the matching results between the trajectory and the warning threshold. This effectively identifies the continuous evolution trend of equipment from normal to abnormal, avoiding false alarms and missed alarms caused by the degradation mode of the target production equipment deviating from the preset template due to changes in operating conditions. Furthermore, the adapted warning response strategy makes the warning handling measures more targeted. Finally, the preliminary warning commands and strategies are sent to the edge for local verification and generate monitoring warning signals. This allows for the verification of cloud commands based on real-time operational data collected at the edge, avoiding false alarms caused by differences in cloud and edge environments and information asymmetry, thereby improving the accuracy and reliability of the warning results. In summary, the technical solution provided in this application can achieve state warnings for target production equipment under conditions of differences in cloud and edge environments and information asymmetry. Attached Figure Description

[0017] Figure 1This is an exemplary flowchart of a cloud-edge-device collaborative production equipment status monitoring and early warning method according to some embodiments of this application; Figure 2 This is an exemplary flowchart illustrating the determination of fault-related feature sequences according to some embodiments of this application; Figure 3 This is a schematic diagram of the structure of a cloud-edge-device collaborative production equipment status monitoring and early warning system according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a computer device that implements a cloud-edge-device collaborative production equipment status monitoring and early warning method according to some embodiments of this application. Detailed Implementation

[0018] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] refer to Figure 1 The figure is an exemplary flowchart of a cloud-edge-device collaborative production equipment status monitoring and early warning method according to some embodiments of this application. The figure mainly includes the following steps: In step S101, multi-source time-series operation data of the target production equipment are collected at the edge, and the multi-source time-series operation data are filtered based on the equipment operation characteristics to obtain an edge feature dataset.

[0020] It should be noted that, in this application, the edge terminal refers to the edge computing and processing node in the cloud-edge-device collaborative production equipment status monitoring and early warning system. It is usually deployed on-site at the target production equipment and consists of hardware components such as an embedded processor, sensor data acquisition interface, local storage module, and communication module. It is the hub connecting the cloud platform and the production equipment terminal. The edge terminal can not only collect multi-source time-series operating data such as vibration frequency, operating temperature, voltage, current, and speed of the target production equipment in real time, but also receive preliminary early warning commands and adapted early warning response strategies issued by the cloud. Finally, it outputs the status monitoring and early warning signal of the target production equipment, thereby reducing the data processing pressure and network bandwidth transmission cost of the cloud while ensuring the real-time performance and reliability of equipment status monitoring and early warning.

[0021] In specific implementation, multi-source time-series operation data of the target production equipment is collected through the sensor data acquisition interface in the edge terminal. The multi-source time-series operation data includes, but is not limited to, the vibration frequency, operating temperature, voltage, current, and speed of the target production equipment. During the acquisition, the sensor data acquisition interface is connected to the vibration sensor, temperature sensor, current and voltage sensor, and speed sensor of the target production equipment, which will not be elaborated here.

[0022] It should be noted that the multi-source time-series operational data in this application refers to an ordered set of operational parameters with time stamps collected by the edge terminal for the target production equipment in different operational dimensions. The multi-source time-series operational data reflects the real-time operational status of the target production equipment at different times, providing a data foundation for subsequent edge terminal feature screening and cloud status analysis.

[0023] In some embodiments, the process of filtering the multi-source time-series operational data based on device operational characteristics to obtain an edge feature dataset is achieved through the following steps: Extract all operational feature dimensions of the target production equipment; Invalid data is removed from the multi-source time-series runtime data based on all runtime feature dimensions to obtain a preprocessed time-series dataset. The preprocessed time series dataset is used to extract time series feature parameters that match the operational feature dimensions of the target production equipment to form an edge feature dataset.

[0024] In specific implementation, firstly, based on the design technical manual of the target production equipment, a feature enumeration method is used to filter parameters that can characterize the equipment's operating status, thereby extracting all operating feature dimensions of the target production equipment. These operating feature dimensions include the vibration frequency, operating temperature, voltage, current, and rotational speed of the target production equipment. The equipment operating feature dimensions refer to a set of various physical quantities and performance indicators that can reflect the operating conditions of the target production equipment. Then, time-series operating data corresponding to each operating feature dimension is extracted from the multi-source time-series operating data, and invalid data is removed from the time-series operating data corresponding to each operating feature dimension. Missing values ​​are filled using linear interpolation, and outliers deviating from the normal operating range are... The 3σ principle is used to identify and eliminate data. For mixed noisy data, a moving average filtering method is used for smoothing, thus obtaining a preprocessed time series dataset. The preprocessed time series dataset refers to the set of valid time series operation data that only contains the dimensions of equipment operation characteristics after invalid data processing. Finally, for the time series operation data corresponding to each type of operation characteristic dimension in the preprocessed time series dataset, time domain feature extraction methods (i.e., calculating mean, variance, and peak value) and frequency domain feature extraction methods (i.e., obtaining the main frequency and frequency band energy through fast Fourier transform) are used to calculate feature parameters, extracting time series feature parameters that match the operation characteristic dimensions of the target production equipment, and then combining the extracted feature parameters into an edge feature dataset.

[0025] It should be noted that, in this application, the edge feature dataset refers to the set of time-series feature parameters extracted from the preprocessed time-series dataset that correspond one-to-one with the dimensions of device operation features. By determining the edge feature dataset, the preprocessing results of multi-source time-series operation data at the edge can be accepted. By retaining the time-domain and frequency-domain feature parameters that match the dimensions of device operation features and removing invalid, noisy, and redundant data, the amount of data transmitted from the edge to the cloud can be effectively reduced, saving communication bandwidth resources and improving data transmission efficiency.

[0026] In step S102, the edge feature dataset is associated with the device identification information of the target generating device and uploaded to the cloud. The edge feature dataset is then analyzed for fault association based on the historical fault case library in the cloud, thereby generating a fault association feature sequence with time stamps.

[0027] It should be noted that the cloud in this application refers to a distributed computing and storage integrated platform deployed in a remote industrial data center to support cloud-edge-device collaborative production equipment status monitoring and early warning. Its hardware consists of a multi-node high-performance server cluster, a highly available distributed storage array, and a dedicated communication gateway for industrial IoT, which can provide a remote integrated platform for full-process data processing and decision support for production equipment status monitoring and early warning.

[0028] In some embodiments, the process of associating the edge feature dataset with the device identification information of the target generation device and uploading it to the cloud is achieved through the following steps: Obtain the equipment identification information of the target production equipment; The device identification information is used to perform label embedding processing on the edge feature dataset to obtain a feature dataset with identification labels; Perform cloud-transmission-adaptive serialization encoding on the labeled feature dataset to obtain the encoded transmission dataset; The encoded transmission dataset is uploaded to a designated cloud storage node based on a preset cloud communication protocol.

[0029] In specific implementation, firstly, the hardware serial number, factory unique code, and device identity code assigned by the Industrial Internet of Things (IIoT) identification system pre-stored in the embedded control module of the target production equipment are read to obtain the device identification information of the target production equipment. This device identification information refers to structured data containing the unique identity information of the target production equipment. Secondly, a key-value pair tag embedding mechanism is used to associate and map the device identification information as tag keys with the edge feature dataset. By adding the device identity code corresponding to the device identification information to the header data area of ​​the edge feature dataset, the binding of the device identification information and the edge feature dataset is achieved, resulting in a feature dataset with tags. This feature dataset with tags refers to a composite dataset that integrates the unique identity tag of the target device and the device's operational feature data. Then, the feature dataset with tags is subjected to cloud transmission adaptation serialization encoding, using an industrial protocol. The Buffers serialization protocol first defines a corresponding serialization data description file based on the data structure of the tagged feature dataset, clarifying the type, length, and arrangement order of the feature data fields of the device identity code and feature dataset. Then, based on the data description file, the tagged feature dataset is structurally converted and compressed to obtain the encoded transmission dataset. The encoded transmission dataset refers to a compact data carrier that has undergone serialization encoding and compression and is suitable for long-distance transmission between cloud and edge. Finally, based on a preset industrial IoT communication protocol, such as the MQTT protocol for device communication, a stable data link is established with the cloud through the edge communication module according to the pre-configured cloud communication parameters (i.e., including cloud server address, communication port, device authentication information, and data transmission topic), and the encoded transmission dataset is sent to the designated distributed storage node in the cloud.

[0030] It should be noted that, in this embodiment, the specified storage node refers to a dedicated storage unit pre-allocated in the cloud for storing specific device operating characteristic data, which can support the data retrieval needs of subsequent cloud-based fault correlation analysis and other operations.

[0031] In some embodiments, reference Figure 2 As shown in the figure, this is an exemplary flowchart of determining a fault association feature sequence according to some embodiments of this application. In this embodiment, fault association analysis is performed on the edge feature dataset based on a historical fault case library in the cloud to generate a fault association feature sequence with time stamps. This can be achieved by the following steps: In step S1021, a historical fault case library of the target production equipment is obtained, and a standardized fault feature index is constructed on the historical fault case library to generate a fault feature index library. In step S1022, a feature similarity matching algorithm is used to compare and associate the edge feature dataset with the fault feature index library to obtain preliminary fault association results; In step S1023, the time-series timestamps in the edge feature dataset are extracted, and the preliminary fault association results are calibrated in terms of time dimension to obtain the time-series association feature set; In step S1024, the time-series associated feature set is serialized and encapsulated to generate a fault-associated feature sequence with time stamps.

[0032] In practice, firstly, a historical fault case library of the same model and operating conditions as the target production equipment is retrieved from the cloud-based equipment lifecycle management database. This historical fault case library includes, but is not limited to, historical fault repair records, fault diagnosis reports, and fault characteristic monitoring data. Subsequently, fault tree analysis is used to decompose each type of fault event in the historical fault case library into fault modes, extracting the corresponding fault characteristic parameters for each type of fault (e.g., bearing wear corresponds to the vibration acceleration time-domain peak value, characteristic frequency amplitude, and bearing temperature rise rate; poor gear meshing corresponds to the vibration meshing frequency harmonic amplitude and sideband energy ratio; in the power system, motor overload...). Corresponding to the effective value of stator current, winding temperature, and the system oil pressure fluctuation amplitude and flow attenuation rate corresponding to hydraulic pump failure, the fault occurrence threshold and fault evolution law are analyzed. Then, through inverted index construction technology, fault types are associated and mapped with corresponding fault feature parameters to establish a standardized retrieval structure with feature parameters as index keys and fault information as index values, generating a fault feature index library. This fault feature index library refers to a structured database built based on historical fault cases that enables the association and retrieval of feature parameters and fault information. Next, a cosine similarity matching algorithm is used to perform an association comparison between the edge feature dataset and the fault feature index library, first comparing the edge features... Fault feature parameters in the edge feature dataset and fault feature index are converted into feature vectors of equal dimension. The cosine similarity between the two sets of feature vectors is then calculated to quantify the feature matching degree. Fault feature parameters with cosine similarity values ​​greater than or equal to a similarity threshold are selected to obtain preliminary fault association results. These preliminary fault association results refer to the set of fault information potentially associated with the edge feature dataset, selected through feature similarity matching. Then, pre-embedded time-series timestamp information is extracted from the edge feature dataset, and the fault feature parameters in the preliminary fault association results are sorted according to chronological order. Fault association information is bound to the corresponding timestamp to achieve alignment and calibration of fault association results in the time dimension, resulting in a time-series association feature set. The time-series association feature set refers to the set of fault association feature parameters with added timestamp information. Finally, using structured data encapsulation technology, the feature parameters, timestamps, and corresponding fault association information in the time-series association feature set are systematically integrated according to a preset time-series data format (the preset time-series data format is a structured key-value pair nested format commonly used in the Industrial Internet of Things). Data verification fields and format identification fields are added to generate a fault association feature sequence with time stamps, which will not be elaborated here.

[0033] It should be noted that the fault association feature sequence with time stamps in this application refers to a serialized data set containing fault association information and corresponding timestamps organized in chronological order. Since the preliminary fault association results without time dimension calibration are only the matching information of discrete fault features and edge features, they cannot reflect the evolution of equipment operation features over time. However, the occurrence and development of equipment faults have obvious temporal characteristics. Therefore, by determining the fault association feature sequence with time stamps, the change process of fault association features within the equipment operation cycle can be effectively restored, providing a unified temporal benchmark for subsequent time-series deviation calculation with the equipment operation standard template and the construction of a continuous equipment state evolution trajectory.

[0034] In step S103, based on the degree of deviation between the fault-related feature sequence and the equipment operation standard template, a state evolution trajectory of the target production equipment is constructed in the cloud, and based on the matching result of the state evolution trajectory and the preset warning threshold, a preliminary warning instruction and an adapted warning response strategy are generated.

[0035] In some embodiments, the following steps are used to construct the state evolution trajectory of the target production equipment in the cloud based on the degree of deviation between the fault association feature sequence and the equipment operation standard template: Obtain the equipment operation standard template corresponding to the target production equipment, and perform feature standardization processing on the equipment operation standard template to generate a standard feature benchmark set; The deviation between the fault-related feature sequence and the standard feature benchmark set is calculated on a time-series basis to obtain the time-series deviation dataset; Extract the time stamp information and corresponding deviation values ​​from the time series deviation dataset to generate trajectory data sources; Based on the trajectory, a data source is constructed to fit the state evolution path of the target production equipment, generating the state evolution trajectory of the target production equipment.

[0036] In specific implementation, firstly, the equipment operation standard template corresponding to the target production equipment is retrieved from the cloud-based equipment file management library. This template represents the optimal operating characteristic dataset of the target production equipment under the same operating conditions. Then, a feature normalization method (such as min-max normalization) is used to map various feature parameters in the equipment operation standard template to the same dimension and dimensional space as the fault-related feature sequence, eliminating dimensional differences between different parameters and generating a standard feature benchmark set that corresponds one-to-one with the fault-related feature sequence. This standard feature benchmark set refers to the set of standard operating feature parameters of the equipment that has been normalized and has the same dimension as the fault-related feature sequence. Secondly, a relative deviation rate algorithm is used to calculate the time-series deviation between the fault-related feature sequence and the standard feature benchmark set. Feature parameter values ​​corresponding to each time marker in the fault-related feature sequence are extracted according to a preset time step and compared with the standard feature parameter values ​​of the same dimension and time sequence in the standard feature benchmark set. The relative deviation rate is calculated between the two. The deviation rate quantifies the degree of feature deviation of a single time-series node. The deviation calculation results of all time-series nodes are integrated in chronological order to obtain a time-series deviation dataset. The time-series deviation dataset refers to a structured dataset containing the feature deviation values ​​of all time-series nodes and sorted by time markers. Then, pre-bound time marker information and corresponding deviation calculation results are extracted from the time-series deviation dataset to generate a trajectory construction data source with time markers as the horizontal axis and deviation values ​​as the vertical axis. The trajectory construction data source refers to a data set composed of valid time markers and corresponding deviation values ​​used for fitting the equipment state evolution path. Finally, a cubic spline interpolation algorithm is used to fit the state evolution path of the target production equipment based on the trajectory construction data source and perform smoothing processing. With time markers as independent variables and deviation values ​​as dependent variables, a smooth cubic spline function is constructed to interpolate the discrete trajectory construction data source to obtain a continuous and smooth equipment state change curve, thereby generating the state evolution trajectory of the target production equipment.

[0037] It should be noted that the state evolution trajectory of the target production equipment in this application refers to a visualized curve reflecting the continuous change process of the equipment's operating characteristics deviating from the standard state. Since the discrete time-series deviation dataset can only reflect the degree of deviation of the equipment's operating characteristics from the standard state at a single time point, it cannot intuitively reflect the continuous evolution trend of the equipment's state from normal to abnormal. Moreover, most production equipment failures are dynamic processes that develop gradually. The deviation data of a single node is insufficient to support a comprehensive assessment of the equipment's health status. Therefore, the discrete time-series deviation values ​​are fitted into a continuous and smooth state change curve through path fitting, which intuitively presents the evolution law of the equipment's state over time, so as to locate the start time, development rate and severity of the abnormal state. This provides a visualized and quantifiable basis for subsequent dynamic matching with the preset warning threshold, effectively avoiding the problem of false alarms or missed alarms caused by a single threshold judgment.

[0038] In some embodiments, the generation of preliminary warning instructions and adapted warning response strategies based on the matching results between the state evolution trajectory and the preset warning threshold is achieved through the following steps: Obtain the preset multi-level early warning thresholds of the target production equipment, and standardize the multi-level early warning thresholds to generate a standardized early warning threshold set; A threshold matching algorithm is used to compare the state evolution trajectory with a standardized early warning threshold set point by point to obtain the threshold matching result; Based on the threshold matching result, a preset warning level classification rule is invoked to generate a warning level determination result; Based on the warning level determination result, a preset response strategy library is associated to generate an appropriate warning response strategy; The warning level determination results and warning response strategies are integrated and encapsulated into instructions to generate preliminary warning instructions.

[0039] In specific implementation, firstly, the preset multi-level early warning thresholds of the target production equipment are obtained (the specific thresholds can be set according to actual needs or expert knowledge, and are not limited here). Then, a feature normalization method (such as min-max normalization) is used to map the multi-level early warning thresholds to a dimension space with the same dimensionality as the deviation value of the state evolution trajectory, eliminating the interference of dimensional differences on the matching results, and generating a standardized early warning threshold set consistent with the determination dimension of the state evolution trajectory. The standardized early warning threshold set refers to a set of multi-level threshold intervals divided according to early warning levels after normalization processing. Secondly, the state evolution trajectory is compared point-by-point with the standardized early warning thresholds in the standardized early warning threshold set, i.e., according to... The deviation value of each node on the state evolution trajectory is extracted in chronological order. The standardized warning threshold interval to which this deviation value belongs is determined. A threshold matching label corresponding to each time node is recorded. The threshold matching labels of all time nodes are combined in chronological order to obtain the threshold matching result. The threshold matching result refers to the set of point-by-point matching determination records between the deviation values ​​of each node on the state evolution trajectory and the standardized warning threshold interval. Further, based on the threshold matching result, a preset warning level classification rule is invoked (i.e., the warning level classification rule is based on the four-level threshold intervals corresponding to the standardized warning threshold set, including low, medium, and high-level threshold intervals, combined with the threshold intervals of the state evolution trajectory nodes in the threshold matching result). The number of consecutive occurrences of matching tags is defined. Specifically, a low warning level is determined when a single time node matches a low-level threshold range and subsequent time nodes return to the normal range; a medium warning level is determined when multiple time nodes within a set time window match a medium-level threshold range; and a high warning level is determined when multiple time nodes within a set time window match a high-level threshold range. (The time window can be set according to actual needs and is not limited here.) This warning level classification rule predefines the warning level determination logic corresponding to different threshold matching tag combinations. By executing this warning level classification rule, the threshold matching results are judged, generating a unique warning level determination result. The warning level determination result refers to the classification and judgment conclusion of the risk level of the current operating status of the equipment. Then, based on the warning level determination result, a preset response strategy library is retrieved. This response strategy library pre-stores the operation and maintenance measures, response time requirements, and execution entities corresponding to different warning levels (i.e., low warning level matches operation and maintenance measures such as cloud-based remote continuous monitoring and parameter trend tracking, with a response time requirement of completing data verification within 24 hours, and the execution entity being the cloud-based operation and maintenance management system; medium warning level matches measures such as increasing the frequency of equipment inspections and investigating potential abnormal causes by on-site operation and maintenance personnel, with a response time requirement of conducting on-site inspections within 12 hours, and the execution entity being the on-site operation and maintenance team).The high-alert-level matching equipment involves suspending operation and conducting specialized inspections and maintenance on core fault-related components. The response time requirement is to initiate a shutdown inspection process within 4 hours (executed by the equipment technical maintenance team). The appropriate alert response strategy is matched from the response strategy library based on the alert level. This appropriate alert response strategy refers to a set of specific operation and maintenance solutions matching the alert level. Finally, structured instruction encapsulation technology is used to integrate the alert level determination result with the alert response strategy, adding an instruction generation timestamp and a unique device identifier. The instructions are encoded according to a preset instruction data format to generate a preliminary alert instruction that can be sent to the edge. Further details are omitted here.

[0040] It should be noted that the preliminary warning instruction in this application refers to structured instruction data that includes the warning level and response strategy. By determining the preliminary warning instruction, the warning level judgment result and the appropriate warning response strategy can be integrated. At the same time, it carries the unique device identifier and the instruction generation timestamp to accurately deliver it to the target edge. This provides a clear judgment benchmark and execution basis for the edge to conduct local verification by combining real-time multi-source time-series operation data, effectively avoiding false alarms and missed alarms caused by single cloud-based analysis.

[0041] In step S104, the preliminary warning instruction and warning response strategy are sent to the corresponding edge device. The edge device performs local verification of the preliminary warning instruction by combining the real-time collected multi-source time-series operation data, and generates a monitoring warning signal for the target production equipment based on the warning verification result and the warning response strategy.

[0042] In some embodiments, the preliminary warning instruction and warning response strategy are sent to the corresponding edge device, and the edge device performs local verification of the preliminary warning instruction by combining real-time collected multi-source time-series operational data. This is achieved through the following steps: Obtain the device identification information at the target edge to generate an edge communication index; Based on the edge communication index, the preliminary early warning command and early warning response strategy are encapsulated and transmitted to obtain the data packet to be sent. The data packets are transmitted to the corresponding target edge terminal via the cloud-edge communication link. The target edge terminal receives the data packets and parses and restores them to obtain preliminary warning instructions and warning response strategies. The target edge collects multi-source time-series operation data of the target production equipment in real time and generates a real-time verification dataset; Based on the verification rules in the early warning response strategy, the real-time verification dataset is matched and verified with the feature conditions of the preliminary early warning instruction to obtain the early warning verification result.

[0043] In specific implementation, firstly, the device identification information of the target edge terminal is obtained, and the device identification information is used as the edge terminal communication index. The edge terminal communication index refers to structured data that can uniquely identify the communication identity of the target edge terminal and communicate accurately with the edge terminal. Secondly, based on the edge terminal communication index, the message encapsulation specification of the MQTT industrial communication protocol is adopted to associate and bind the preliminary warning command, the warning response strategy and the edge terminal communication index. Data transmission header verification field, command priority identifier field and transmission timeout retransmission identifier field are added to obtain the data packet to be sent. The data packet to be sent refers to a structured data carrier that can be stably transmitted in the cloud-edge communication link after being encapsulated by the communication protocol. Furthermore, through a pre-set highly reliable cloud-edge communication link, a point-to-point communication connection is established between the cloud and the target edge terminal based on the 5G industrial gateway. Data packets are transmitted to the corresponding target edge terminal at a preset transmission rate. The target edge terminal restores the content of the data packets through its built-in communication protocol parsing module to obtain preliminary warning instructions and warning response strategies. Then, multi-source time-series operation data of the target production equipment is collected through the sensor data acquisition interface in the target edge terminal, and the multi-source time-series operation data is used as a real-time verification dataset to generate a real-time verification dataset. The real-time verification dataset refers to the set of real-time operation feature data of the equipment used to match and verify the preliminary warning instructions.Finally, two types of core verification rules are extracted from the early warning response strategy: feature parameter threshold interval comparison rules and time-series trend consistency verification rules. Simultaneously, feature parameter dimensions and time-series timestamp information that perfectly correspond to the feature conditions of the initial early warning instruction are extracted from the real-time verification dataset to achieve precise mapping and alignment of feature dimensions. Then, according to the feature parameter threshold interval comparison rules, the real-time value of each feature parameter in the real-time verification dataset is compared dimension-by-dimensionally with the threshold interval (including the lower and upper limits of the early warning threshold) of the corresponding feature parameter in the initial early warning instruction. The matching status (match / mismatch) of a single-dimensional feature parameter is recorded, generating a single-dimensional threshold matching result set. Finally, based on the time-series trend consistency verification rules, a linear regression algorithm is used to calculate the trend slope of each feature parameter in the real-time verification dataset over time. The direction of change of the rate is consistent with the trend change requirements of the preset characteristic parameters in the initial warning instruction, generating a time-series trend matching result set. The direction of change includes rising, falling, and stable. Then, according to the preset weight allocation in the warning response strategy, quantitative weights are assigned to the single-dimensional threshold matching result set and the time-series trend matching result set respectively (the specific weights can be set according to actual needs, and are not limited here), and normalized weighted calculations are performed to obtain the comprehensive matching degree. The comprehensive matching degree is then compared with the preset verification validity threshold in the warning response strategy. If the comprehensive matching degree is greater than or equal to the verification validity threshold, the warning instruction is determined to be valid; if the comprehensive matching degree is less than the verification validity threshold, the warning instruction is determined to be invalid. Finally, the single-dimensional threshold matching result, the time-series trend matching result, the comprehensive matching degree value, and the warning judgment result are combined into the warning verification result.

[0044] It should be noted that the early warning verification result in this application refers to the conclusion data used to determine whether the preliminary early warning command is effective. As the output of the local verification process at the edge, the early warning verification result can verify whether the preliminary early warning command issued by the cloud is consistent with the actual operating status of the device based on the device operation data collected in real time at the edge. It can effectively filter out false alarms or missed alarms caused by factors such as historical data deviations and data transmission delays in the cloud, thereby improving the accuracy and reliability of the entire monitoring and early warning system.

[0045] In some embodiments, generating a monitoring and early warning signal for the target production equipment based on the early warning verification result and the early warning response strategy is achieved through the following steps: The validity of the early warning verification results is determined, and a verification validity determination result is generated. Based on the verification validity determination result, the signal generation rules in the early warning response strategy are matched to obtain the signal generation rule set; Based on the signal generation rule set, the verification validity determination result is mapped to the warning signal parameters to generate warning signal parameters; The early warning signal parameters are standardized and encoded to generate monitoring and early warning signals for the target production equipment.

[0046] In specific implementation, firstly, the warning judgment results in the warning verification results are distinguished as either valid or invalid warning commands, generating a verification validity judgment result. This verification validity judgment result refers to the conclusion that the warning verification results meet the conditions for triggering the generation of a warning signal. Secondly, based on the verification validity judgment result, a pre-stored signal generation rule library in the warning response strategy is retrieved, and signal generation logic corresponding to different judgment conclusions is matched. For example, "valid warning command" matches the corresponding level of warning signal generation rule, and "invalid warning command" matches the warning termination rule, resulting in a signal generation rule set. This signal generation rule set refers to the rules matched with the verification validity judgment result, used to guide the generation of warning signal parameters. A set of rules is generated; then, based on the signal generation rule set, the validity judgment results are mapped item by item to the warning signal parameters, extracting the unique device identifier, warning level, signal trigger timestamp, and response measure summary information to generate warning signal parameters containing signal type, level identifier, content description, and execution requirements. The warning signal parameters refer to the data units that constitute the monitoring warning signal; finally, the warning signal parameters are standardized and encapsulated using a signal encoding protocol commonly used in the industrial Internet of Things field, adding signal format identifier fields, data integrity verification fields, and receiver parsing guidance fields to ensure that the generated signal can be parsed by the device monitoring terminal, thereby generating the monitoring warning signal for the target production equipment.

[0047] It should be noted that the monitoring and early warning signal in this application refers to the signal that characterizes the early warning status of the target production equipment and guides the operation and maintenance. The monitoring and early warning signal can be directly pushed to relevant terminals and platforms to support operation and maintenance decisions.

[0048] Furthermore, in another aspect of this application, in some embodiments, this application provides a production equipment status monitoring and early warning system based on cloud-edge-device collaboration, referencing... Figure 3 The figure is a schematic diagram of the structure of a cloud-edge-device collaborative production equipment status monitoring and early warning system according to some embodiments of this application. The cloud-edge-device collaborative production equipment status monitoring and early warning system includes: a data acquisition module 201, a processing module 202, and an execution module 203, which are described below: The acquisition module 201 in this application is mainly used to acquire multi-source time-series operation data of the target production equipment at the edge, and to perform data filtering based on equipment operation characteristics on the multi-source time-series operation data to obtain an edge feature dataset; Processing module 202, in this application, is mainly used to associate the edge feature dataset with the device identification information of the target generating device and upload it to the cloud. On the cloud, it performs fault association analysis on the edge feature dataset based on the historical fault case library, and then generates a fault association feature sequence with time stamp. The processing module 202 is further configured to construct the state evolution trajectory of the target production equipment in the cloud based on the degree of deviation between the fault association feature sequence and the equipment operation standard template, and generate preliminary warning instructions and adapted warning response strategies based on the matching results of the state evolution trajectory and the preset warning threshold. The execution module 203 in this application is mainly used to send the preliminary warning instruction and warning response strategy to the corresponding edge terminal. The edge terminal performs local verification of the preliminary warning instruction in combination with the real-time collected multi-source time-series operation data, and generates a monitoring warning signal for the target production equipment based on the warning verification result and the warning response strategy.

[0049] In addition, this application also provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute the above-described cloud-edge-device collaborative production equipment status monitoring and early warning method.

[0050] In some embodiments, reference Figure 4 The figure is a schematic diagram of the structure of a computer device implementing a cloud-edge-device collaborative production equipment status monitoring and early warning method according to some embodiments of this application. The cloud-edge-device collaborative production equipment status monitoring and early warning method in the above embodiments can... Figure 4 The computer device shown is used to implement this, and the computer device includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

[0051] The processor 301 can be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more devices used to control the execution of the cloud-edge-device collaborative production equipment status monitoring and early warning method described in this application.

[0052] The communication bus 302 can be used to transmit information between the aforementioned components.

[0053] The memory 303 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via the communication bus 302. The memory 303 may also be integrated with the processor 301.

[0054] The memory 303 stores program code for executing the scheme of this application, and its execution is controlled by the processor 301. The processor 301 executes the program code stored in the memory 303. The program code may include one or more software modules. In the above embodiments, the determination of the production equipment status monitoring and early warning method based on cloud-edge-device collaboration can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0055] Communication interface 304 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0056] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0057] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0058] In addition, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for monitoring and warning the status of production equipment based on cloud-edge-device collaboration.

[0059] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0060] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for monitoring and early warning of production equipment status based on cloud-edge-device collaboration, characterized in that, Includes the following steps: Multi-source time-series operation data of the target production equipment are collected at the edge, and the multi-source time-series operation data is filtered based on the equipment operation characteristics to obtain an edge feature dataset; The edge feature dataset is associated with the device identification information of the target generation device and uploaded to the cloud. Fault association analysis is performed on the edge feature dataset in the cloud based on the historical fault case library, thereby generating a fault association feature sequence with time stamp. Based on the degree of deviation between the fault-related feature sequence and the equipment operation standard template, the state evolution trajectory of the target production equipment is constructed in the cloud, and based on the matching result of the state evolution trajectory and the preset warning threshold, preliminary warning instructions and adapted warning response strategies are generated. The preliminary warning command and warning response strategy are sent to the corresponding edge device. The edge device performs local verification of the preliminary warning command by combining the real-time collected multi-source time-series operation data, and generates a monitoring warning signal for the target production equipment based on the warning verification result and the warning response strategy.

2. The method as described in claim 1, characterized in that, The edge feature dataset obtained by performing data filtering based on device operation characteristics on the multi-source time-series operational data specifically includes: Extract all operational feature dimensions of the target production equipment; Invalid data is removed from the multi-source time-series runtime data based on all runtime feature dimensions to obtain a preprocessed time-series dataset. The preprocessed time series dataset is used to extract time series feature parameters that match the operational feature dimensions of the target production equipment to form an edge feature dataset.

3. The method as described in claim 1, characterized in that, The process of associating the edge feature dataset with the device identification information of the target generating device and uploading it to the cloud specifically includes: Obtain the equipment identification information of the target production equipment; The device identification information is used to perform label embedding processing on the edge feature dataset to obtain a feature dataset with identification labels; Perform cloud-transmission-adaptive serialization encoding on the labeled feature dataset to obtain the encoded transmission dataset; The encoded transmission dataset is uploaded to a designated cloud storage node based on a preset cloud communication protocol.

4. The method as described in claim 1, characterized in that, The process of performing fault correlation analysis on the edge feature dataset based on a historical fault case library in the cloud, and then generating a fault correlation feature sequence with time stamps, specifically includes: Obtain a historical fault case library of the target production equipment, and construct a standardized fault feature index for the historical fault case library to generate a fault feature index library; A feature similarity matching algorithm is used to compare and associate the edge feature dataset with the fault feature index to obtain preliminary fault association results; Extract the time-series timestamps from the edge feature dataset and perform time dimension calibration on the preliminary fault association results to obtain the time-series association feature set; The time-series associated feature set is serialized and encapsulated to generate a fault-associated feature sequence with time stamps.

5. The method as described in claim 1, characterized in that, Based on the degree of deviation between the fault association feature sequence and the equipment operation standard template, the state evolution trajectory of the target production equipment is constructed in the cloud, specifically including: Obtain the equipment operation standard template corresponding to the target production equipment, and perform feature standardization processing on the equipment operation standard template to generate a standard feature benchmark set; The deviation between the fault-related feature sequence and the standard feature benchmark set is calculated on a time-series basis to obtain the time-series deviation dataset; Extract the time stamp information and corresponding deviation values ​​from the time series deviation dataset to generate trajectory data sources; Based on the trajectory, a data source is constructed to fit the state evolution path of the target production equipment, generating the state evolution trajectory of the target production equipment.

6. The method as described in claim 1, characterized in that, The preliminary warning command and warning response strategy are sent to the corresponding edge device. The edge device performs local verification of the preliminary warning command by combining real-time collected multi-source time-series operational data. Specifically, this includes: Obtain the device identification information at the target edge to generate an edge communication index; Based on the edge communication index, the preliminary early warning command and early warning response strategy are encapsulated and transmitted to obtain the data packet to be sent. The data packets are transmitted to the corresponding target edge terminal via the cloud-edge communication link. The target edge terminal receives the data packets and parses and restores them to obtain preliminary warning instructions and warning response strategies. The target edge collects multi-source time-series operational data of the target production equipment in real time and generates a real-time verification dataset. Based on the verification rules in the early warning response strategy, the real-time verification dataset is matched and verified with the feature conditions of the preliminary early warning instruction to obtain the early warning verification result.

7. The method as described in claim 1, characterized in that, Based on the early warning verification results and early warning response strategy, the monitoring early warning signals generated for the target production equipment specifically include: The validity of the early warning verification results is determined, and a verification validity determination result is generated. Based on the verification validity determination result, the signal generation rules in the early warning response strategy are matched to obtain the signal generation rule set; Based on the signal generation rule set, the verification validity determination result is mapped to the warning signal parameters to generate warning signal parameters; The early warning signal parameters are standardized and encoded to generate monitoring and early warning signals for the target production equipment.

8. A production equipment status monitoring and early warning system based on cloud-edge-device collaboration, used to execute the production equipment status monitoring and early warning method based on cloud-edge-device collaboration as described in any one of claims 1 to 7, characterized in that, The system includes: The acquisition module is used to acquire multi-source time-series operation data of the target production equipment at the edge, and to perform data filtering based on equipment operation characteristics on the multi-source time-series operation data to obtain an edge feature dataset; The processing module is used to associate the edge feature dataset with the device identification information of the target generation device and upload it to the cloud. On the cloud, the edge feature dataset is analyzed for fault association based on the historical fault case library, thereby generating a fault association feature sequence with time stamp. The processing module is also used to construct the state evolution trajectory of the target production equipment in the cloud according to the degree of deviation between the fault association feature sequence and the equipment operation standard template, and generate preliminary warning instructions and adapted warning response strategies based on the matching results of the state evolution trajectory and the preset warning threshold. The execution module is used to send the preliminary warning command and warning response strategy to the corresponding edge terminal. The edge terminal combines the real-time collected multi-source time-series operation data to perform local verification of the preliminary warning command, and generates a monitoring warning signal for the target production equipment based on the warning verification result and the warning response strategy.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the cloud-edge-device collaborative production equipment status monitoring and early warning method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the cloud-edge-device collaborative production equipment status monitoring and early warning method as described in any one of claims 1 to 7.

Images