Industrial one-stop multi-source data acquisition and monitoring system and equipment

Abstract

The invention relates to the technical field of data processing, and discloses an industrial one-stop multi-source data acquisition and monitoring system and equipment, and the system comprises a data acquisition module which is used for obtaining multi-source industrial monitoring data and generating a data stream input node; the data preprocessing module is used for processing the input data in the data flow model to generate a standardized data flow; the data conversion module is used for dynamically converting data nodes in the standardized data stream according to different data protocols to generate a compatible data stream; the task scheduling module is used for optimizing the data flow and distributing the data flow to a target node according to the compatible data flow in combination with a node priority and a dynamic scheduling strategy in the data flow model; and the exception detection module is used for carrying out exception state analysis on the optimized data stream and judging whether an exception exists or not according to a preset rule. According to the technical scheme, the compatibility and reliability of the monitoring system can be improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and more particularly to an industrial one-stop multi-source data acquisition and monitoring system and equipment. Background Art

[0002] Against the backdrop of rapid global economic development, industrial automation and intelligent manufacturing have become important trends driving the upgrading and transformation of various industries. As IoT technology continues to mature, enterprises' demand for multi-source data collection, real-time analysis, and intelligent monitoring is growing rapidly. Especially in the smart fields of smart water, smart cities, smart energy, and smart construction, enterprises expect to improve production efficiency, optimize resource allocation, and further reduce operating costs to enhance market competitiveness through more efficient data management methods.

[0003] In the existing technology, industrial data acquisition and monitoring systems already have basic environmental parameter monitoring functions, such as the collection and upload of temperature, humidity, pressure and other data, and support the integration of multiple communication protocols. However, these systems have great limitations in compatibility and functional scalability in complex scenarios, and are difficult to adapt to the dynamic processing requirements of multi-source data. In addition, existing systems often rely on static rule configuration in task scheduling and data conversion, resulting in insufficient flexibility in practical applications, especially in scenarios of multi-protocol data interaction and complex task distribution.

[0004] To sum up, the existing industrial monitoring systems have significant deficiencies in compatibility, reliability and intelligent data processing, and it is difficult to meet the actual needs for efficient and diversified data management in intelligent business scenarios. Summary of the invention

[0005] The present application provides an industrial one-stop multi-source data acquisition and monitoring system and industrial one-stop multi-source data acquisition and monitoring equipment, aiming to solve the technical problems of the existing industrial monitoring systems in terms of compatibility, reliability and intelligence.

[0006] Additional aspects and advantages of the disclosure will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the disclosure.

[0007] According to a first aspect of the present disclosure, there is provided an industrial one-stop multi-source data acquisition and monitoring system, which includes: a data acquisition module, which is used to access external devices through a variety of data interfaces, obtain multi-source industrial monitoring data and generate data flow input nodes to build a data flow model based on a directed acyclic graph; a data preprocessing module, which is used to format, detect outliers and complete the input data in the data flow model to generate a normalized data flow; a data conversion module, which is used to dynamically convert the data nodes in the normalized data flow according to different data protocols to generate a compatible data flow; a task scheduling module, which is used to optimize the data flow and distribute it to the target node according to the compatible data flow, combined with the node priority and dynamic scheduling strategy in the data flow model; an anomaly detection module, which is used to analyze the abnormal state of the optimized data flow, determine whether there is an anomaly according to preset rules, and generate an alarm signal and trigger an alarm device in an abnormal situation.

[0008] According to a second aspect of the present disclosure, there is provided an industrial one-stop multi-source data acquisition and monitoring device, which is configured with an industrial one-stop multi-source data acquisition and monitoring system as described in the above embodiment, and the device includes: a power supply; a housing, including multiple external device connection interfaces; a main control unit, used to provide the computing and software service deployment functions required for system operation; a communication component, connected to the main control unit, and used to establish various types of communication connections with external data sources through various communication methods; a sensor component, including an environmental monitoring sensor and an equipment status sensor, used to collect temperature and humidity, air pressure, gas concentration and equipment operation parameters in real time; a display and interaction component, including a touch screen and a front-end interface, used to display monitoring data and system status in real time, and provide a user operation interface for parameter configuration and task management.

[0009] It can be seen from the above technical solution that the present disclosure has at least one of the following advantages and positive effects: The industrial one-stop multi-source data acquisition and monitoring system disclosed in the present invention is composed of a data acquisition module, a data preprocessing module, a data conversion module, a task scheduling module and an anomaly detection module. Among them: The data acquisition module is connected to external devices through a variety of data interfaces, supports the acquisition of multi-source industrial monitoring data, and builds a data flow model based on a directed acyclic graph, thereby improving the system's adaptability to multiple data sources and providing a standardized data foundation for subsequent processing.

[0010] The data preprocessing module formats, detects outliers, and completes the input data in the data flow model, eliminating the impact of data redundancy and missing data and making the data flow more standardized. This not only improves the validity and consistency of the data, but also provides high-quality input for subsequent data conversion and task scheduling.

[0011] The data conversion module uses a dynamic conversion mechanism to adapt the data nodes in the standardized data stream according to different data protocols, thereby generating a compatible data stream. This module improves the flexibility and scalability of the system in a multi-protocol communication environment and can adapt to the needs of different industrial application scenarios.

[0012] By combining the node priority and dynamic scheduling strategy in the data flow model, the task scheduling module optimizes the compatible data flow and distributes it to the target node, effectively improving the efficiency of task execution and the ability to allocate resources reasonably, ensuring the orderly transmission and efficient use of data flow in complex environments.

[0013] Finally, the anomaly detection module analyzes the state of the optimized data stream, determines anomalies based on preset rules, and generates alarm signals. This module can respond to abnormal conditions in the system in real time, ensuring the reliability of the data monitoring process and the safety of system operation. The above technical solution significantly enhances the compatibility, reliability, and intelligent data processing capabilities of the system on the basis of achieving collaborative work among various modules, meeting the diverse needs in industrial monitoring scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0015] Figure 1 It is a structural block diagram of an industrial one-stop multi-source data acquisition and monitoring system in one embodiment of the present disclosure; Figure 2 It is a structural block diagram of an industrial one-stop multi-source data acquisition and monitoring system in another embodiment of the present disclosure; Figure 3 A schematic diagram of the workflow of an industrial one-stop multi-source data acquisition and monitoring system in one embodiment of the present disclosure; Figure 4 A schematic diagram of the software service architecture of an industrial one-stop multi-source data acquisition and monitoring system in one embodiment of the present disclosure; Figure 5 This is a user operation page diagram of an industrial one-stop multi-source data acquisition and monitoring system in one embodiment of the present disclosure; Figure 6 It is a schematic diagram of the structure of an industrial one-stop multi-source data acquisition and monitoring device in one embodiment of the present disclosure.

[0016] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts. DETAILED DESCRIPTION

[0017] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with this specification. Instead, they are merely examples of devices and methods consistent with some aspects of this specification as detailed in the appended claims.

[0018] The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit this specification. The singular forms "a", "the" and "the" used in this specification and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0019] It should be understood that although the terms first, second, third, etc. may be used in this specification to describe various information, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this specification, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0020] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more comprehensive and complete and will fully convey the concept of the example embodiments to those skilled in the art.

[0021] In addition, the described features, structures or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the present disclosure.

[0022] Furthermore, the drawings are only schematic illustrations and are not necessarily drawn to scale. The block diagrams shown in the drawings are only functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0023] In an exemplary embodiment of the present disclosure, firstly, an industrial one-stop multi-source data acquisition and monitoring system is provided. Figure 1 The structure block diagram of an industrial one-stop multi-source data acquisition and monitoring system in one embodiment of the present disclosure is schematically shown. Figure 1 As shown, the industrial one-stop multi-source data acquisition and monitoring system includes a data acquisition module 1, a data preprocessing module 2, a data conversion module 3, a task scheduling module 4 and an anomaly detection module 5. Among them, the data acquisition module 1 can be used to access external devices through a variety of data interfaces, obtain multi-source industrial monitoring data and generate data flow input nodes to build a data flow model based on a directed acyclic graph; the data preprocessing module 2 can be used to format, detect outliers and complete the input data in the data flow model to generate a normalized data flow; the data conversion module 3 can be used to dynamically convert the data nodes in the normalized data flow according to different data protocols to generate a compatible data flow; the task scheduling module 4 can be used to optimize the data flow according to the compatible data flow, combined with the node priority and dynamic scheduling strategy in the data flow model, and distribute it to the target node; the anomaly detection module 5 can be used to analyze the abnormal state of the optimized data flow, determine whether there is an abnormality according to preset rules, and generate an alarm signal and trigger the alarm device in an abnormal situation.

[0024] In the one-stop multi-source data acquisition and monitoring system of this embodiment, during operation, the data acquisition module 1 first accesses external devices through a variety of data interfaces, obtains multi-source industrial monitoring data in real time and generates data stream input nodes, thereby constructing a data flow model based on a directed acyclic graph, and providing an organized data basis for subsequent processing. Subsequently, the input data in the data flow model is passed to the data preprocessing module 2 for processing, and the redundant, inconsistent or missing data is corrected into a normalized data flow by formatting, detecting outliers and completing the data, ensuring the integrity and consistency of the data, and laying the foundation for further conversion of the data. Then, the data conversion module 3 dynamically converts the data nodes in the normalized data flow according to the preset multiple communication protocol rules, generates a compatible data flow that can adapt to different communication protocols, thereby meeting the communication requirements in a multi-protocol environment and realizing flexible data transmission. On this basis, the task scheduling module 4 optimizes the data flow according to the compatible data flow, combined with the node priority information and dynamic scheduling strategy in the data flow model, and distributes it to the target node to achieve efficient deployment of tasks and rational use of resources. Finally, the optimized data stream is input into the anomaly detection module 5 for abnormal state analysis. The module determines whether there is an abnormality in the data stream through preset rules, generates an alarm signal when an abnormality is detected, and triggers the alarm device to achieve timely response, thereby effectively ensuring the stable operation of the system and the reliability of data monitoring. The whole process completes the monitoring and management of the whole process from data collection to anomaly detection through the orderly coordinated operation of the modules.

[0025] Next, we will refer to Figure 2 In other embodiments of the present disclosure, the functional principles of the data acquisition module 1, the data preprocessing module 2, the data conversion module 3, the task scheduling module 4 and the anomaly detection module 5 in the industrial one-stop multi-source data acquisition and monitoring system are described in detail.

[0026] The data acquisition module 1 can be used to access external devices through various data interfaces, obtain multi-source industrial monitoring data and generate data flow input nodes to build a data flow model based on a directed acyclic graph. Among them, the external device can represent industrial equipment that can be connected to the data acquisition module 1 through various data interfaces, including but not limited to sensors, controllers, measuring instruments and other industrial equipment that supports data output, and of course, third-party systems and third-party databases. Industrial monitoring data can represent a data set generated by an external device and received by the data acquisition module 1 to reflect the operating status or environmental parameters of the industrial system, including but not limited to the measurement values ​​of temperature, humidity, pressure, flow, current, voltage and other physical or chemical parameters. The data flow model based on the directed acyclic graph can represent the data organization model generated by the data acquisition module 1. The model represents the logical structure of the data flow in the form of a directed acyclic graph (DAG), wherein each node in the graph corresponds to a single task or data unit of data acquisition, and the directed edges in the graph represent the data dependency relationship between the nodes. The model has an acyclic characteristic, ensuring that the data processing flow is executed in a predefined order, while supporting parallel processing of data flows between multiple nodes, thereby improving the processing efficiency of the system and the flexibility of task allocation.

[0027] refer to Figure 2 As shown, in some embodiments, the data acquisition module includes a multi-channel interface unit 11, a data synchronization unit 12, a data mapping unit 13 and a node generation unit 14. Among them: The multi-channel interface unit 11 can be used to access external devices through the target data interface, collect multi-source industrial monitoring data, and generate corresponding data stream input nodes The target data interface may include one or more of a serial port, an Ethernet interface, and an I2C bus.

[0028] The data synchronization unit 12 can be used to receive the original data stream in the data stream input node, and generate a directed acyclic graph by aligning the timestamps and correcting the sequence. The synchronous data flow of the data flow model is as follows: V represents a set of nodes in the data flow model, E Represents a set of directed edges in the data flow model. Specifically, timestamp alignment is performed based on the following formula: in, Representation Node With Node The time deviation, and Respectively represent nodes and timestamp, is the allowed time deviation threshold. When the time deviation meets the above conditions, the data nodes are considered to be aligned.

[0029] Sequence correction adjusts the order of data streams by defining sequence consistency constraints, and its formula is as follows: in, Representation Node With Node The nodes in the data stream are arranged in a time sequence. The values ​​of are sorted to ensure that the time dependencies of all edges in the data flow model are in order. The above timestamp alignment and order correction formulas ensure the logical consistency and timing accuracy of data transmission.

[0030] The data mapping unit 13 may be used to map the predecessor node set of each node in the synchronous data stream according to the predecessor node set of each node in the synchronous data stream. , computing node data input status , and through the processing function Generates the output data of the node , the calculation relationship is: , in, Represents the predecessor node To the input node The weight of Indicates the current node ID. Indicates the predecessor node identifier. Represents the activation function, which can be ReLU or Sigmoid function, etc., to introduce nonlinear mapping to enhance the model's ability to represent data. Representation Node The weight matrix of the processing function, Representation Node This enables dependency mapping between data nodes and ensures the logical integrity and accuracy of the data flow in the model.

[0031] The node generation unit 14 may be used to map the node data stream output by the data mapping unit into a standardized input node set and form an initial path in a directed acyclic graph structure.

[0032] The data preprocessing module 2 can be used to format, detect outliers and complete the input data in the data flow model in sequence to generate a normalized data flow. Specifically, the data preprocessing module 2 formats the input data to unify the data from different sources into a standard format; identifies and processes abnormal data that deviates from the normal range through an outlier detection method to eliminate or mark interfering data; and fills and repairs the missing information in the data through a missing value completion method, thereby generating a normalized data flow.

[0033] In some embodiments, reference Figure 2 As shown, the data preprocessing module includes a field mapping unit 21, an outlier detection unit 22, and a missing value filling unit 23. Among them: The field mapping unit 21 can be used to map the field according to the field mapping rule. , the source data field in the input data Mapping to target format fields , to perform unified conversion and formatting of data fields.

[0034] The outlier detection unit 22 can be used to perform outlier detection on the result of the formatting process by using an outlier detection method based on a standard score, and remove abnormal data. The calculation process of the outlier detection method can be expressed as: Among them, x represents the current value, represents the sample mean, represents the sample standard deviation, Z represents the standardized deviation value of the data point, which is used to measure the degree of deviation between the current value x and the sample mean μ, and is measured in units of sample standard deviation σ. When |Z| exceeds the preset threshold, the data point is determined to be an outlier.

[0035] The missing value filling unit 23 can be used to fill the data after the abnormal value detection by a linear interpolation algorithm to generate the normalized data stream. The calculation process of the linear interpolation algorithm can be expressed as: in, Represents the currently missing data point, whose value is obtained by interpolation calculation. Represents missing values The data value at the previous time point, Represents missing values The data value at the next time point.

[0036] The data conversion module 3 can be used to dynamically convert the data nodes in the normalized data stream according to different data protocols to generate a compatible data stream. Among them, the data protocol can represent a standardized communication specification for defining the format, semantics and rules of data during transmission, including but not limited to Modbus TCP protocol, OPC-UA protocol, HTTP protocol, MQTT protocol, Socket protocol and Web Service protocol and other data communication protocols suitable for industrial environments. A compatible data stream can represent a data stream generated after being processed by the data conversion module 3, and each data node of the data stream is converted into a format and structure that meets the requirements of the target data protocol. The compatible data stream can adapt to the communication requirements of different protocols, ensure the seamless transmission of data between heterogeneous devices or systems, and support data interaction in a multi-protocol environment, thereby improving the adaptability and scalability of the system.

[0037] In some embodiments, dynamically converting the data nodes in the normalized data stream according to different data protocols to generate a compatible data stream specifically includes the following technical steps: First, protocol matching is performed on the data nodes in the normalized data stream, and the target protocol type corresponding to each data node is determined by referring to the preset protocol mapping rule table. The target protocol type includes but is not limited to MQTT, Socket and HTTP. Specifically, the data conversion module determines the target protocol type that matches the data node according to the preset protocol mapping rule table by reading the metadata attributes of each data node. The protocol mapping rule table stores the mapping relationship between the source data type and the target protocol type. In the process of protocol matching, the most suitable data communication protocol is dynamically selected according to the attributes of the node (such as data format, communication interface type), providing a clear rule basis for the subsequent parsing and conversion steps.

[0038] Then, according to the target protocol type, the content of the data node is parsed, and the parsed fields are remapped to the standard field set of the target protocol. Specifically, according to the target protocol type, the content of the data node is structured and parsed. The parsing process includes reading the field structure of the data node, extracting key field information (such as timestamp, device ID, data value, etc.), and remapping the parsed fields according to the standard field set of the target protocol. The standard field set is defined according to the format requirements of the target protocol, including but not limited to protocol header fields, data payload fields, and checksum fields. For example, when mapping a data node to the HTTP protocol, the data value is mapped to the payload field, and the timestamp is mapped to the URL parameter to ensure that the generated data node conforms to the semantics and format specifications of the target protocol.

[0039] Finally, the mapping fields are reorganized according to the encapsulation rules corresponding to the target protocol type, and the compatible data stream matching the target protocol type is generated by adding protocol header information and updating verification information. Specifically, the mapped fields are re-encapsulated according to the encapsulation rules corresponding to the target protocol type. The encapsulation step includes reorganizing the field content according to the standard format of the target protocol, adding protocol header information to identify the data source and transmission path, and updating the verification field to ensure data integrity. The update of the verification field is calculated according to the verification rules of the target protocol, for example, a new verification value is generated based on a CRC check or an MD5 hash value. The data nodes finally encapsulated are reorganized into a compatible data stream that conforms to the target protocol format, which can be adapted to different communication protocols to achieve data transmission, thereby meeting the communication needs in a multi-protocol environment.

[0040] The task scheduling module can be used to optimize the data flow and distribute it to the target node according to the compatible data flow, combined with the node priority and dynamic scheduling strategy in the data flow model. Among them, the node priority can represent the importance ranking basis of the task scheduling module for the execution order and resource allocation of each data node in the data flow model. The dynamic scheduling strategy can represent the set of rules for the task scheduling module to dynamically adjust the processing order and resource allocation of the data flow according to the node priority and real-time operating environment in the data flow model. The dynamic scheduling strategy includes but is not limited to the following rules: give priority to scheduling nodes with lower priority values ​​to maximize resource utilization; in the case of insufficient resources, wait for resource release through the task queuing mechanism; combine the dependencies of the tasks and distribute the tasks on the premise of ensuring the completion of the predecessor nodes.

[0041] In some embodiments, reference Figure 2 As shown, the task scheduling module may include a priority calculation unit 41, a dynamic sorting unit 42 and a data flow allocation unit 43. Among them: The priority calculation unit 41 can be used to calculate the node information in the compatible data stream and the resource consumption of the node. and delay time , according to the priority calculation formula: Prioritize nodes ,in, and They are the weight factors for resource consumption and latency requirements, respectively, used to balance the impact of the two on priority, and the value range meets Resource consumption weight factor It can be used to measure the occupation of system resources by tasks, including but not limited to the utilization rate of computing units such as CPU and GPU during task execution, bandwidth occupation during data transmission, etc. Delay demand weight factor It can be used to measure the timeliness of task execution, including but not limited to task response time and data transmission delay. This unit extracts resource consumption parameters from the data flow model by analyzing the task requirements of each node. and delay requirement parameters Resource consumption parameters Indicates the computing resources required by the node to perform tasks, such as CPU utilization, memory occupancy, etc.; delay requirement parameters Indicates the time sensitivity of node task execution, such as data transmission delay or task completion deadline.

[0042] The dynamic sorting unit 42 can be used to dynamically sort the nodes according to the result of the priority calculation unit, postpone the scheduling of low-priority nodes, and give priority to processing high-priority nodes. The unit first extracts the priority values ​​of all nodes and sorts the nodes in order from low to high. Nodes with low priority are regarded as priority processing nodes, and nodes with high priority are postponed for scheduling. During the sorting process, the unit also combines the node dependencies in the data flow model to ensure that the predecessor node completes the task before the subsequent node to maintain the logical consistency of the data flow. In the case of resource allocation conflicts, the dynamic sorting unit will dynamically adjust the scheduling order of the nodes according to the current system resource status to achieve efficient arrangement of tasks and rational use of resources.

[0043] The data stream allocation unit 43 can be used to distribute the optimized data stream to the corresponding target node according to the scheduling result of the dynamic sorting unit, and adjust the allocation strategy to optimize the data stream transmission efficiency when resources are insufficient. Specifically, the unit first extracts the node sequence after dynamic sorting, and associates it with the target node in the data stream to determine the distribution path of the data stream. During the distribution process, the data stream allocation unit adjusts the allocation strategy according to the resource status of the target node, including allocating bandwidth, computing resources and storage resources. In addition, when resources are insufficient, the unit adopts a dynamic queue mechanism to temporarily store the tasks to be distributed, and redistributes the data stream in priority order after the resources are released to ensure the continuity and transmission efficiency of task distribution. Through the above process, the data stream allocation unit can achieve efficient transmission of data streams and load balancing of target nodes, thereby optimizing the overall operation performance of the system.

[0044] The anomaly detection module 5 is used to analyze the abnormal state of the optimized data stream, determine whether there is an abnormality according to the preset rules, and generate an alarm signal and trigger the alarm device in the case of an abnormality. Figure 2 As shown, the abnormality detection module may include an alarm unit 51 and a data trend analysis unit 52. Among them: The alarm unit 51 can be used to monitor the target parameters in the data stream in real time, and determine whether the data exceeds the normal range through the early warning threshold, so as to generate a corresponding alarm signal and trigger the alarm device.

[0045] The data trend analysis unit 52 can be used to analyze the historical monitoring data stream and the real-time monitoring data stream using a long short-term memory network model, and mark abnormal nodes according to the analysis results. The long short-term memory network model is a deep learning model based on a recursive neural network, which can be specifically used to process and analyze time series data.

[0046] In some embodiments, the historical monitoring data stream and the real-time monitoring data stream are analyzed using a long short-term memory network model, and abnormal nodes are marked according to the analysis results, specifically including the following steps: The historical monitoring data stream and the real-time monitoring data stream are normalized, and a time series matrix is ​​constructed based on the result of the normalization process: in, Indicates Data nodes at time The monitoring value of Indicates the total number of data nodes. Indicates the length of the time series.

[0047] The time series matrix The long short-term memory network model is input, and the state vector of each time step is calculated through the network hidden layer. The calculation formula of the hidden layer state is: in, represents the hidden layer state at the current time step, and represents the network weight matrix, represents the bias vector, represents the activation function, Represents the time step The corresponding input data vector, that is, the historical monitoring data stream and the real-time monitoring data stream at time The observed value at time, Represents the previous time step The hidden layer state.

[0048] The abnormal score of each data node is calculated based on the hidden layer state vector. The score calculation formula is: in, Indicates The anomaly score of each node, Represents the mean of the hidden layer state. When the preset threshold is exceeded, the corresponding data node will be marked as an abnormal node.

[0049] In some embodiments, reference Figure 2 As shown, the industrial one-stop multi-source data acquisition and monitoring system also includes a fault detection module 6, which can be used to monitor the system operation status in real time, perform fault detection on the operation status data based on the support vector machine and isolation forest algorithm, and switch to the backup system through the fault isolation mechanism when a fault is detected.

[0050] Specifically, the fault detection module 6 monitors the system operation status in real time, and collects the system's multi-dimensional operation status data set , get the time The module samples these parameters at fixed time intervals through the data acquisition unit to form a time series matrix. ,in Indicates The sequence value of the operating status parameter, is a The matrix of represents the number of state parameters, Represents the number of time steps. Through this matrix, the real-time and integrity of the system operation status data is ensured, providing high-quality input for subsequent fault detection.

[0051] The fault detection module first uses the support vector machine (SVM) to classify the operating status data and build a classification model to distinguish between normal and abnormal states. The optimization goal of the SVM model is: in, Represents the normal vector of the classification hyperplane, defining the direction of the classification boundary; The bias term representing the classification boundary is used to adjust the boundary position; Indicates The slack variable of samples is used to allow for classification error; is a regularization parameter used to balance the trade-off between model complexity and classification error. In this optimization model, the kernel function Implement nonlinear mapping, where Represents the feature mapping function, which maps the input data to a high-dimensional feature space. The Gaussian kernel function is in the form of: in, and There are two data points, is the kernel width parameter, which controls the nonlinearity of the mapping. The module generates a classification model through the training set and inputs the data in real time during the detection phase. Input to the model, output classification results ,in Indicates whether the current status is normal or abnormal.

[0052] Based on SVM classification detection, the module further uses the isolation forest algorithm (iForest) to mine abnormal data. Isolation forest constructs an isolation tree by randomly splitting the sample feature space and measures the isolation depth of the sample. The specific process of isolation forest processing is as follows: First, from the running status dataset Randomly select a subset from ,in Represents the number of isolated trees.

[0053] Then, randomly select features and segmentation threshold , construct the binary splitting rule: Enter the left subtree, otherwise enter the right subtree, where Indicates that the sample has the feature The value on is a randomly selected threshold.

[0054] Finally, the average isolation depth for each sample is calculated , and calculate the anomaly score according to the following formula: in, For sample The abnormal score of is the average isolation depth, Indicates sample size The standard average isolation depth of is used to normalize the score. When approaching 1, the sample are considered as outliers.

[0055] When the module determines that the system operation status is abnormal through the above detection algorithm, the fault isolation mechanism is immediately activated to switch to the backup system. The module first evaluates the resource status of the backup system, including the availability of CPU, memory and bandwidth, and confirms that the backup system can support the current task requirements; then it performs data synchronization and copies the key parameters of the current task to the backup system through real-time mirroring to ensure that the task can be seamlessly connected after switching; finally, it switches the task to the backup system and stops the faulty operation of the main system. The fault isolation mechanism achieves high reliability and stability of system operation through rapid response, avoiding task interruption and data loss.

[0056] In some embodiments, the workflow in the industrial one-stop multi-source data acquisition and monitoring system is as follows: Figure 3 As shown in the figure, the system first performs an initialization process when it starts. By starting the monitoring serial port sensor, the system accesses external devices through multiple data interfaces to collect data from various monitoring devices. In this process, after the monitoring serial port sensor receives the transmitted data, it obtains the sensor data through polling reading and transmits it to the data processing module. The module first cleans the data to ensure that the data meets the format requirements and completes the abnormal value detection. The processed data is normalized according to the preset standards, and then enters the alarm mechanism judgment link to determine whether the data meets the warning conditions.

[0057] When the data reaches the preset warning conditions, the system will immediately trigger an alarm, the alarm signal will be sent through the speaker alarm device, and the real-time alarm information will be displayed on the LCD screen. At the same time, the system turns on the video monitoring control panel, obtains the real-time video stream and displays it on the monitoring interface. If an intrusion is detected on the monitoring screen, the system immediately starts intrusion detection and determines whether there is any illegal intrusion. If the system confirms that an intrusion has occurred, it automatically triggers the alarm mechanism and makes a corresponding security response. During this process, the system also starts the front-end web interface configuration and dynamically adjusts the parameter configuration of the data acquisition service according to the task requirements. The system sets the priority and task scheduling of real-time data collection according to the current task to ensure the efficiency and accuracy of data processing. After the data is processed, it will be sent to the target node as needed to complete the final data processing task.

[0058] In some embodiments, the software service architecture of the industrial one-stop multi-source data acquisition and monitoring system is as follows: Figure 4 As shown. Specifically, the system first accesses data from a variety of data sources, including third-party databases, sensors, and third-party systems. The system can interact with external devices through a variety of access protocols, such as Oracle, SQL Server, MySQL, REDIS, PostgreSQL and other database types, as well as HTTP, MQTT, Socket, Web Service, ModbusTCP and other communication protocols, so as to achieve comprehensive data collection and aggregation. After data collection, the system will format, clean, convert and process the data through the data processing unit. In this process, the data will be integrated through the connector module, and after data cleaning and formatting, it will enter the data storage stage. At this stage, the system ensures that the data is effectively stored through mass storage and metadata management technology, and can be efficiently queried and analyzed. All processed data will be organized into standardized data streams to ensure its reliability and efficiency in subsequent processing.

[0059] The task configuration module is used to flexibly configure various tasks within the system. Users can define different tasks according to their needs and adjust their priorities. Task scheduling depends on the system's dynamic scheduling strategy. The scheduling algorithm optimizes the execution order of tasks to achieve optimal resource utilization efficiency and data flow rate. During data transmission, the system adjusts the data flow allocation and processing strategy in real time based on the priority of the task and the allocation of resources. The data security module provides protection for the operation of the system and ensures the safe transmission and storage of data. The system uses encryption technology to encrypt sensitive data and protects personal privacy data through desensitization technology. At the same time, through strict permission management, unauthorized users are prevented from accessing data to ensure data integrity and security.

[0060] The system also includes a real-time monitoring and management function that can monitor important indicators such as task execution, system status, and data flow. The monitoring module tracks the flow of data in real time and tracks and analyzes each stage of task execution to ensure the stability of the system during operation. Finally, the data sharing and synchronization module supports multiple synchronization methods, including real-time synchronization, incremental synchronization, and two-way synchronization to ensure the consistency of data inside and outside the system. The system can select the appropriate synchronization method according to needs to achieve synchronous updates of each data node to ensure the timeliness and reliability of data.

[0061] In some embodiments, the user operation page in the industrial one-stop multi-source data acquisition and monitoring system is as follows: Figure 5 As shown in the figure. Specifically, the user interface provides an integrated data processing platform, in which users can select different data input and output methods through the left panel. Users can select data output methods such as XML.Output, JSONOutput, Excel output, etc., and connect them to various devices or data sources through drag-and-drop operations to facilitate data collection, processing and export. The overall process of data flow includes data collection, data processing and final output. Data is first collected by the connected production equipment and Siemens equipment and enters the data flow processing unit. After the data flow is processed by filters, character encoding conversion, etc., a cleaned data flow is generated, which will be used for subsequent processing links, including the removal of outliers. Users can export the data flow through the Microsoft Excel output node to display the results of data processing in real time. In addition, the interface can also display each node and connection path related to the data flow, allowing users to conveniently manage and control the path and configuration of the data flow through simple graphical operations. The operation interface facilitates users to flexibly configure and monitor in real time throughout the data collection and processing process to ensure effective data collection, conversion and accurate output.

[0062] In some embodiments, the software service architecture uses Java to write the backend and Vue to build the front-end interface. The overall architecture design aims to achieve efficient data integration and configuration management. Specifically, the back-end service is developed based on the Spring Boot framework, which is responsible for data processing, task scheduling and business logic implementation. The back-end service has efficient data interaction capabilities with a variety of databases, big data platforms, Internet of Things protocols and text files, ensuring that the system has flexible expansion capabilities and efficiency when processing massive amounts of data. Through plug-in design, the back-end system follows the open and closed principle and can be easily expanded without modifying existing core modules.

[0063] The front-end interface is built with Vue.js, which provides a user-friendly operation interface, allowing users to quickly configure data sources, target databases, and processing flows through simple drag-and-drop operations without writing code, greatly reducing the user's usage threshold. In terms of data storage, the system uses PostgreSQL as the main database solution, which can handle complex SQL queries and advanced data operations to ensure data security, reliability, and efficiency. At the same time, the system uses MinIO as an object storage service, which ensures high availability and scalability of massive data management by supporting multi-threaded and multi-node parallel processing, data redundancy, and automatic fault recovery. In terms of communication, the system uses a combination of HTTP and WebSocket, where HTTP is used for ordinary requests and responses, and WebSocket is used to achieve two-way real-time data transmission between components, improving the efficiency of user interaction and system performance. In order to ensure the high availability and load balancing of the system, the front-end and back-end deployment and request management are reverse-proxyed through Nginx, and at the same time, combined with an efficient caching mechanism and a dynamic and static separation configuration method, the system's performance and stability are further improved.

[0064] The system is deployed using Docker containerization technology. All components are encapsulated through Docker and orchestrated and managed using Docker-Compose. In this way, users can quickly start or stop services through simple Docker commands, achieve flexible environment configuration and resource isolation, and facilitate efficient management of system components and their resources. In the design of the data processing engine, the concepts of streaming computing and task scheduling are borrowed, and the configuration of complex data integration tasks is simplified through a graphical drag-and-drop operation interface. The core modules include data flow engine, task scheduler, and data cleaning processing module, which can support the entire process from multi-source data collection to target database storage, and ensure seamless integration of different data sources and protocols. Through these designs, the system can efficiently process various data types and sources, while ensuring smooth transmission and processing of data flows, providing users with convenient and reliable data management services.

[0065] In some embodiments, reference Figure 2 As shown, the industrial one-stop multi-source data collection and monitoring system also includes a dynamic threshold adjustment module 7, which can be used to dynamically update the warning threshold of each monitoring parameter using a Bayesian optimization algorithm based on the multi-source industrial monitoring data collected in real time and historical data statistics.

[0066] Specifically, the dynamic threshold adjustment module 7 first performs statistical analysis on the multi-source industrial monitoring data and historical data collected in real time. Suppose the monitoring parameter set is ,in Indicates The module collects the current monitoring data in real time through the data acquisition unit. , and combined with historical monitoring data Compute statistical features, including means , Standard Deviation And the distribution density function The calculation formula for statistical analysis can be expressed as: in, Indicates The historical mean of the monitoring parameters reflects the central trend of the data; Indicates The standard deviation of each monitoring parameter reflects the degree of dispersion of the data; is the number of historical data points; Indicates The monitoring parameter is the first Values; It represents the probability density function of the monitoring parameters and describes the distribution characteristics of the data.

[0067] The dynamic threshold adjustment module builds a Bayesian optimization model based on statistical analysis results and real-time data to dynamically adjust the warning threshold of the monitoring parameters. The threshold value is , the goal of Bayesian optimization is to maximize the monitoring accuracy of the system and the sensitivity of anomaly detection. The objective function of the optimization problem is: in, Indicates the threshold The corresponding monitoring effect, Represents the loss function, which is used to measure the data points Exceeding the threshold The punishment of time, Represents data points The probability density function describes the distribution of data. Indicates the candidate value of the current monitoring threshold.

[0068] The loss function is defined as: in, Indicates monitoring data points; Indicates the warning threshold; Indicates the deviation value when a data point exceeds the threshold.

[0069] The Bayesian optimization model estimates the objective function based on Gaussian process regression (GPR) The posterior distribution of : in, Represents sampled data sets, including historical and real-time data; Represents the mean of the posterior distribution, used to estimate effect; Represents the variance of the posterior distribution, which is used to measure the uncertainty of the estimate.

[0070] The module dynamically updates the warning threshold of each monitoring parameter, using the Expected Improvement (EI) criterion: in, Represents the expected improvement value, which is used to measure the selection threshold Potential optimization benefits of Represents the posterior mean of the current threshold; The objective function value representing the current optimal threshold; represents the uncertainty of the current threshold; Indicates the standardized score, reflecting Exceed possibility; represents the cumulative distribution function of the standard normal distribution; Represents the probability density function of the standard normal distribution.

[0071] By maximizing , the module determines the new optimal threshold , and update the warning thresholds of monitoring parameters in real time, dynamically adapting to changes in data, thereby improving the accuracy of anomaly detection and the monitoring efficiency of the system.

[0072] In some embodiments, the workflow of the one-stop multi-source data acquisition and monitoring system for industrial use may include: first, the system is initialized by starting the main control unit, loading the operating system and necessary drivers to ensure the normal operation of the device. Next, the network connection is configured and the 4G module is initialized to ensure that the device can access the network and ensure the smooth subsequent data transmission. At this time, the touch screen will automatically obtain and display the IP address of the terminal, which is convenient for accessing the Web page through the IP. In addition, the environment detection sensor, speaker and camera of the terminal will be automatically initialized and ready for environmental monitoring and data collection. Next, the system deploys data integration services, including installing docker and java environments on the main control system. The database, MinIO and Nginx services are started through the docker containerization technology, and the front-end page is proxy-based through Nginx. At the same time, the Java back-end service is started. The back-end service is responsible for receiving the user's configuration request and managing the data flow. The front-end Vue application interacts with the back-end service through the Web interface, providing a user configuration interface for users to configure the data flow. In order to ensure that the system can start automatically after booting, the systemd service is added, and the software service is set to start automatically when it is powered on. In the user configuration task stage, the user enters the address through the browser to access the Web interface of the data integration service, and enters the task configuration interface after login authentication. Users can select the required data source, target database, and data processing flow through an intuitive drag-and-drop interface. The system supports multiple data source types, including databases, big data platforms, IoT devices, and text files, ensuring flexibility in responding to different business needs. During the data collection process, the backend service collects data from the required devices or databases based on the tasks configured by the user. The data will first be stored locally as an original backup to ensure the reliability and integrity of the data, and then the data will be passed to the next data processing step in the form of a data stream and enter the processing flow.

[0073] During the data processing and push phase, the collected data will be cleaned through the data processing plug-in configured on the Web page to remove outliers and duplicate data to ensure data quality. After the data is standardized, data from different sources will be converted into a unified format, and the output fields and field type parameters will be clearly defined. The processed data will be pushed to the configured target database. It also supports pushing data to the specified target system or external API according to the configuration, realizing multi-platform data sharing and meeting different business needs. In the real-time monitoring and alarm phase, the system will monitor data changes in real time. When an abnormal situation is detected, the system will trigger the alarm mechanism to respond. In addition, the system will monitor the camera screen. When the screen changes, the USB camera will record the abnormal event and issue a sound alarm through the speaker to alert the staff. The data integration service will update the Web interface in real time to display the current monitoring data, task status and system log, so that users can track the operation of the system. Users can view historical data, manage tasks and adjust configurations on the Web interface to ensure that the system continues to run stably and meet business needs. In the task scheduling and management phase, the backend service regularly schedules the configured data collection tasks to ensure the automation and efficiency of the system operation. The system also supports pausing, starting, and deleting tasks. Users can adjust task configurations at any time to optimize data integration workflows based on actual business needs.

[0074] The industrial one-stop multi-source data acquisition and monitoring system in the above embodiment effectively improves the accuracy and system reliability of data acquisition, monitoring and fault diagnosis through multi-level data processing and optimization mechanisms. The data acquisition module accesses external devices through multiple interfaces, obtains and generates data stream input nodes, thereby constructing a data flow model based on a directed acyclic graph to ensure the flexibility and expansibility of data processing. In the data preprocessing stage, the system can format, detect outliers and complete the input data, and then generate a standardized data flow to provide standardized input for subsequent data processing. In the data conversion process, the system can convert nodes of different data protocols according to the target protocol type through dynamic protocol conversion to ensure the compatibility of data flows. This mechanism not only improves the system's adaptability to multiple devices and protocols, but also optimizes the processing efficiency of data flows. In the task scheduling process, the system optimizes the data flow according to the node priority and dynamic scheduling strategy in the data flow model, and distributes it to the target node, further improving the efficiency and accuracy of data flow transmission. Through the anomaly detection module, the system can perform real-time abnormal state analysis on the optimized data stream, generate alarm signals and trigger alarm devices in time when an anomaly occurs, effectively ensuring the stability and security of the system. The technology of using long short-term memory network model for data trend analysis and node anomaly marking further enhances the system's ability to recognize complex data patterns and ensures accurate detection of potential anomalies. In addition, the dynamic threshold adjustment module uses the Bayesian optimization algorithm to dynamically adjust the warning threshold of the monitoring parameters based on the real-time collected data and historical statistical results, ensuring that the system can automatically optimize the warning mechanism according to data changes, thereby improving the intelligence level of the warning. The fault detection module combines the support vector machine with the isolation forest algorithm to monitor the system operation status in real time and quickly switch to the backup system through the fault isolation mechanism when a fault occurs, thereby ensuring the high availability and disaster recovery capability of the system. Through the combination of the above technical means, the efficient integration of data acquisition, processing, storage and monitoring is realized, which significantly improves the automation level, reliability and flexibility of the system, meets the growing demand for multi-source data fusion and real-time processing in industrial monitoring, and has important application value.

[0075] In addition, the embodiments of the present disclosure also provide an industrial one-stop multi-source data acquisition and monitoring device, which is configured with the industrial one-stop multi-source data acquisition and monitoring system in the above embodiments, and the device includes: a power supply; a housing, including multiple external device connection interfaces; a main control unit, used to provide the computing and software service deployment functions required for system operation; a communication component, connected to the main control unit, used to establish various types of communication connections with external data sources through various communication methods; a sensor component, including an environmental monitoring sensor and an equipment status sensor, used to collect temperature and humidity, air pressure, gas concentration and equipment operation parameters in real time; a display and interaction component, including a touch screen and a front-end interface, used to display monitoring data and system status in real time, and provide a user operation interface for parameter configuration and task management.

[0076] Among them, the main control unit uses the Raspberry Pi 5 motherboard, 64-bit ARM architecture, installed with the Raspberry Pi OS system, and the CPU is a quad-core Arm Cortex-A76 @ 2.4GHz, which is responsible for the overall control and data processing of the system, and the software service deployment function. The communication module includes 802.11ac dual-band Wi-Fi, and in addition, the EC200A-CN 4G module is expanded. It uses the USB standard interface to connect to the main control unit and supports data transmission in a variety of network environments. The sensor module includes temperature and humidity sensors, gas sensors, etc., using the STM32F030 motherboard, and the chip uses the SHT20 temperature and humidity chip. The sensor is connected to the main control unit through the USB port, and the sensor communication uses serial communication to monitor environmental data in real time. The display screen uses the MHS3528 3.5-inch touch screen with a resolution of 340*480, provides SPI input for touch information processing, and is inserted and connected with the Pin pin of the main control unit. The screen is located on the front panel of the device, which is convenient for user operation and information display. The remote monitoring and alarm module uses a USB driver-free camera and speaker, which are connected to the main control unit through a standard USB interface. The camera collects video information in real time, and after obtaining the video information, it passes the video stream to the intrusion detection service. When a difference in the picture is detected, an alarm is issued through the speaker. The power supply supports 5V / 5A DC power input and is compatible with the Type-C protocol. The Type-C interface provides a stable power supply to the main control unit.

[0077] Further, refer to Figure 6 As shown, the industrial one-stop multi-source data acquisition and monitoring equipment also includes: software deployment (supporting platform services, touch screen programs, motion detection programs) 61, temperature and humidity sensor interface 62, USB expansion board reserved interface 63, alarm speaker interface 64, camera interface 65, Gigabit Ethernet interface 66, 5V / 5A Type-C power interface 67, 2*micro HDMI dual-channel 4Kp60 interface 68 and 3.5-inch touch screen 69.

[0078] In addition, the industrial one-stop multi-source data acquisition and monitoring equipment in the disclosed embodiment ensures the rapid deployment and flexible application of the equipment through integrated modular design, so that the equipment can easily connect different types of sensors, databases and other data sources to meet the diverse data acquisition needs. At the same time, the low power consumption characteristics and optimized power management strategy of Raspberry Pi 5 are adopted, which not only effectively reduces the overall power consumption of the system and reduces the operating costs, but also improves the energy efficiency of the equipment, ensuring that low energy consumption is maintained during long-term operation. The device supports a variety of data sources and protocols, and there is no need to frequently switch different tools or platforms for data acquisition and processing, which improves the efficiency of data integration and simplifies the workflow. This feature allows users to complete data processing tasks efficiently without complex configuration during use. In addition, the device has multi-functional applications such as intrusion detection alarm and remote real-time monitoring, which can ensure the comprehensive performance and reliability of the equipment under the flexible coordination of hardware and software. In terms of compatibility, the device especially supports mainstream domestic databases and operating systems, ensuring that the system can run stably under different technical environments, providing a safer and more reliable choice, and promoting the development of an autonomous and controllable technology ecosystem. At the same time, the design of the device takes into account the user's operating experience. The front-end application interface built with Vue allows users to configure tasks through drag-and-drop operations without writing code, which greatly reduces the operating threshold and enhances the ease of use of the system. In addition, the lightweight design of the device ensures its flexible deployment in small areas and edge computing applications, meeting the application needs of different environments.

[0079] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0080] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiment of the present disclosure.

[0081] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0082] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. An industrial one-stop multi-source data acquisition and monitoring system, characterized in that: include: The data acquisition module is used to access external devices through various data interfaces, obtain multi-source industrial monitoring data and generate data flow input nodes to build a data flow model based on a directed acyclic graph; A data preprocessing module, used for formatting, detecting outliers and completing the input data in the data flow model to generate a normalized data flow; A data conversion module, used for dynamically converting data nodes in the normalized data stream according to different data protocols to generate a compatible data stream; A task scheduling module, used to optimize the data flow and distribute it to the target node according to the compatible data flow, combined with the node priority and dynamic scheduling strategy in the data flow model; The anomaly detection module is used to analyze the abnormal status of the optimized data stream, determine whether there is an anomaly based on preset rules, and generate an alarm signal and trigger the alarm device under abnormal circumstances.

2. The industrial one-stop multi-source data acquisition and monitoring system according to claim 1 is characterized in that: The data acquisition module comprises: Multi-channel interface unit, used to access external devices through the target data interface, collect multi-source industrial monitoring data, and generate corresponding data stream input nodes , the target data interface includes one or more of a serial port, an Ethernet interface, and an I2C bus; A data synchronization unit is used to receive the original data stream in the data stream input node and generate a directed acyclic graph by aligning the timestamp and correcting the sequence. The synchronous data flow of the data flow model is as follows: V represents a set of nodes in the data flow model, E represents a set of directed edges in the data flow model; A data mapping unit is used to map the predecessor nodes of each node in the synchronous data stream according to the predecessor node set of each node in the synchronous data stream. , computing node data input status , and through the processing function Generates the output data of the node , the calculation relationship between the input state and the output data is: ， in, Represents the predecessor node To the input node The weight of Indicates the current node ID. Indicates the predecessor node identifier. represents the activation function, Representation Node The weight matrix of the processing function, Representation Node The bias vector of The node generation unit is used to map the node data stream output by the data mapping unit into a standardized input node set and form an initial path in a directed acyclic graph structure.

3. The industrial one-stop multi-source data acquisition and monitoring system according to claim 1, characterized in that: The data preprocessing module comprises: Field mapping unit, used to map fields according to the rules , the source data field in the input data Mapping to target format fields , to perform unified conversion and formatting of data fields; An outlier detection unit, used to perform outlier detection on the result of the formatting process by using an outlier detection method based on a standard score, and to remove abnormal data; The missing value filling unit is used to fill the data after the outlier detection by a linear interpolation algorithm to generate the normalized data stream.

4. The industrial one-stop multi-source data acquisition and monitoring system according to claim 1, characterized in that: The step of dynamically converting the data nodes in the normalized data stream according to different data protocols to generate a compatible data stream includes: Perform protocol matching on the data nodes in the normalized data stream, and determine the target protocol type corresponding to each data node by referring to a preset protocol mapping rule table, wherein the target protocol type includes one or more of MQTT, Socket, and HTTP; Parsing the content of the data node according to the target protocol type, and remapping the parsed fields to a standard field set of the target protocol; The mapping fields are reorganized according to the encapsulation rules corresponding to the target protocol type, and the compatible data stream matching the target protocol type is generated by adding protocol header information and updating verification information.

5. The industrial one-stop multi-source data acquisition and monitoring system according to claim 1, characterized in that: The task scheduling module includes: A priority calculation unit is used to calculate the node information in the compatible data stream and the resource consumption of the node. and delay time , according to the priority calculation formula: Prioritize nodes ,in, and represents the weight factor; A dynamic sorting unit, used to dynamically sort the nodes according to the result of the priority calculation unit, and preferentially process the nodes with high priority; The data stream allocation unit is used to distribute the optimized data stream to the corresponding target node according to the scheduling result of the dynamic sorting unit, and adjust the allocation strategy to optimize the data stream transmission efficiency when resources are insufficient.

6. The industrial one-stop multi-source data acquisition and monitoring system according to claim 1, characterized in that: The anomaly detection module comprises: An alarm unit is used to monitor the target parameters in the data stream in real time and determine whether the data exceeds the normal range through the early warning threshold, so as to generate a corresponding alarm signal and trigger the alarm device; The data trend analysis unit is used to analyze the historical monitoring data stream and the real-time monitoring data stream using a long short-term memory network model, and mark abnormal nodes according to the analysis results.

7. The industrial one-stop multi-source data acquisition and monitoring system according to claim 6, characterized in that: The method of analyzing the historical monitoring data stream and the real-time monitoring data stream by using the long short-term memory network model and marking abnormal nodes according to the analysis results includes: The historical monitoring data stream and the real-time monitoring data stream are normalized, and a time series matrix is ​​constructed based on the result of the normalization process: in, Indicates Data nodes at time The monitoring value of Indicates the total number of data nodes. Indicates the length of the time series; The time series matrix The long short-term memory network model is input, and the state vector of each time step is calculated through the network hidden layer. The calculation formula of the hidden layer state is: in, represents the hidden layer state at the current time step, and represents the network weight matrix, represents the bias vector, represents the activation function, Represents the time step The corresponding input data vector, Represents the previous time step The hidden layer state of The abnormal score of each data node is calculated based on the hidden layer state vector. The score calculation formula is: in, Indicates The anomaly score of each node, Represents the mean of the hidden layer state. When the preset threshold is exceeded, the corresponding data node will be marked as an abnormal node.

8. The industrial one-stop multi-source data acquisition and monitoring system according to claim 1, characterized in that: Also includes: The fault detection module is used to monitor the system operation status in real time, perform fault detection on the operation status data based on the support vector machine and isolation forest algorithm, and switch to the backup system through the fault isolation mechanism when a fault is detected.

9. The industrial one-stop multi-source data acquisition and monitoring system according to claim 1, characterized in that: Also includes: The dynamic threshold adjustment module is used to dynamically update the warning threshold of each monitoring parameter using a Bayesian optimization algorithm based on the multi-source industrial monitoring data collected in real time and the statistical results of historical data.

10. An industrial one-stop multi-source data acquisition and monitoring device, equipped with an industrial one-stop multi-source data acquisition and monitoring system as described in any one of claims 1 to 9, characterized in that: The device comprises: power supply; A housing including a plurality of external device connection interfaces; The main control unit is used to provide the computing and software service deployment functions required for system operation; A communication component, connected to the main control unit, and used to establish various types of communication connections with external data sources through various communication modes; Sensor components, including environmental monitoring sensors and equipment status sensors, are used to collect temperature and humidity, air pressure, gas concentration and equipment operating parameters in real time; The display and interaction components, including the touch screen and front-end interface, are used to display monitoring data and system status in real time, and provide a user operation interface for parameter configuration and task management.

Images