Internet of things data full-link diagnostic analysis method and system

Abstract

The invention relates to the technical field of data processing, in particular to an Internet of Things data full-link diagnostic analysis method and system, and the method comprises the steps: carrying out the data similarity analysis processing of collected standard full-link terminal received data and real full-link terminal received data, and generating Internet of Things full-link analysis data; and when the Internet of Things full-link analysis data is abnormal, performing data preprocessing on the collected full-link node real-time operation data to generate full-link node real-time operation feature data, and performing full-link fault type identification processing on the full-link node real-time operation feature data and the established full-link fault type text feature data to obtain a full-link fault type text feature data. And full-link real-time fault type text feature data is generated, and finally Internet of Things full-link diagnosis data is constructed and pushed to an Internet of Things full-link diagnosis platform, so that accurate diagnosis and timely feedback of the Internet of Things data full-link are realized.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a full-link diagnosis and analysis method and system for Internet of Things data. Background Art

[0002] With the development of network technology, the network structure of enterprises is becoming increasingly complex. A large number of devices are interconnected through the network and exchange data, causing the amount of data involved in the Internet of Things to increase exponentially. Therefore, accurate full-link diagnostic technology for the Internet of Things becomes particularly important.

[0003] The Chinese invention patent with publication number CN111143160B introduces a system full-link monitoring method and device, which periodically calculates and stores the total log volume of each system, determines the starting data volume and the end data volume of each link based on the total log volume, obtains the processing status data of each link based on the starting data volume and the end data volume, and compares the processing status data with the processing benchmark conditions, generates alarm information based on the comparison results, monitors the operation status of each link in real time, locates abnormal nodes in abnormal links, and plans link capacity based on the operation status of each link. It is impossible to accurately identify the type of link abnormality, and it is difficult to ensure the timeliness of the monitoring results by monitoring the links through large-scale data processing. Summary of the invention

[0004] 1. Technical issues to be resolved In order to solve the deficiencies in the background technology, the present invention provides a method and system for full-link diagnosis and analysis of IoT data, so as to achieve accurate diagnosis and timely feedback of the full-link of IoT data.

[0005] (II) Technical solution A full-link diagnostic analysis method for IoT data includes the following steps: S1. Collect standard full-link terminal reception data and real full-link terminal reception data; S2. Perform data similarity analysis and processing based on the data received by the standard full-link terminal and the data received by the real full-link terminal to generate IoT full-link analysis data. When the data is normal, end the IoT full-link diagnosis operation. S3. When the IoT full-link analysis data is abnormal, collect the real-time operation data of the full-link nodes; S4, performing data preprocessing on the full-link node operation data to generate real-time operation feature data of the full-link node; S5. Establishing text feature data of all-link fault types; S6. Perform full-link fault type identification processing according to the full-link node real-time operation feature data and the full-link fault type text feature data to generate full-link real-time fault type text feature data; S7. Construct the IoT full-link diagnostic data based on the IoT full-link analysis data, the real-time operation feature data of the full-link nodes, and the real-time fault type text feature data of the full-link, and push it to the IoT full-link diagnostic platform.

[0006] The present invention generates Internet of Things full-link analysis data by performing data similarity analysis on the collected standard full-link terminal reception data and the real full-link terminal reception data. When the Internet of Things full-link analysis data is abnormal, data preprocessing is performed on the collected full-link node real-time operation data to generate full-link node real-time operation feature data, and full-link fault type identification processing is performed with the established full-link fault type text feature data to generate full-link real-time fault type text feature data. Finally, the Internet of Things full-link diagnostic data is constructed and pushed to the Internet of Things full-link diagnostic platform to achieve accurate diagnosis and timely feedback of the full-link of Internet of Things data.

[0007] Preferably, the specific steps of collecting the standard full-link terminal reception data and the real full-link terminal reception data are as follows: S11. Collect production data from the IoT data source online through the industrial plant data collection unit, and push it to the industrial plant data terminal receiving unit through the IoT communication network. At the same time, preset the collected production data as standard full-link terminal receiving data. , the production data of the IoT data source includes but is not limited to any one of thermal power data, hydropower data, wind power data, photovoltaic data, natural gas data, and coal data; Receive production data pushed by the IoT communication network through the industrial plant data terminal receiving unit to generate real full-link terminal receiving data .

[0008] Preferably, data similarity analysis is performed based on the standard full-link terminal received data and the real full-link terminal received data to generate IoT full-link analysis data. When it is normal, the specific steps of ending this IoT full-link diagnosis operation are as follows: S21, calculating the standard full-link terminal receiving data by using a cosine similarity algorithm and the real full-link terminal receives data Similarity between ; The cosine similarity algorithm formula is as follows: , in, represents the similarity between the standard full-link terminal receiving data and the real full-link terminal receiving data, and Characteristic vectors respectively representing the standard full-link terminal receiving data and the real full-link terminal receiving data; S22, set the similarity threshold to , the similarity between the standard full-link terminal receiving data and the real full-link terminal receiving data and the similarity threshold Perform numerical comparison and generate IoT full-link analysis data based on the numerical comparison results ; like , it means that the IoT full-link transmission process is normal, and the IoT full-link analysis data is output The result is normal and is pushed to the IoT full-link diagnosis platform through the IoT communication network, ending this IoT full-link diagnosis operation. like , it means that the IoT full-link transmission process is abnormal, and the IoT full-link analysis data is output is abnormal.

[0009] By performing data similarity analysis on the collected standard full-link terminal reception data and the real full-link terminal reception data, a preliminary analysis is made on whether there are any abnormalities in the entire IoT link, avoiding real-time monitoring of each key node in the entire link at all times, and effectively reducing the cost of full-link diagnosis of IoT data.

[0010] Preferably, when the IoT full-link analysis data is abnormal, the specific steps of collecting the real-time operation data of the full-link nodes are as follows: S31, when the IoT full-link analysis data When an abnormality occurs, the real-time operation data of each key node in the entire IoT link is collected online through the IoT link node data management platform to generate a real-time operation data set for the entire link node. ,in, It represents the first data in the entire IoT link collected online through the IoT link node data management platform. Real-time operation data of key nodes, Indicates the total number of real-time operation data of all-link nodes, including but not limited to packet loss rate, network delay, available bandwidth, IP address, CPU usage and network protocol information.

[0011] Preferably, the specific steps of performing data preprocessing on the full-link node operation data to generate the full-link node real-time operation feature data are as follows: S41, using the exponential sliding filtering method to perform data noise reduction processing on each full-link node real-time operation data set in turn, to generate a full-link node real-time operation feature data set ,in, It represents the first data in the entire IoT link collected online through the IoT link node data management platform. The real-time operation data of the key nodes are processed by data noise reduction to obtain the real-time operation characteristic data of the entire link nodes.

[0012] Preferably, the specific steps of establishing the full-link fault type text feature data are as follows: S51. Collect the operation characteristic data of each key node when different faults occur in the entire IoT link through the IoT platform, and generate a sample data set of operation characteristic of the fault node in the entire link ,in, Indicates Full-link fault node operation feature sample data, Indicates the total number of running feature sample data of all-link fault nodes; S52, using an agglomerative hierarchical clustering algorithm to perform data clustering processing on the full-link fault node operation feature sample data in the full-link fault node operation feature sample data set to generate a full-link fault node operation feature clustering sample data set ,in, Represents the first Clustering sample data of running characteristics of nodes with full-link failures. Indicates the total number of running feature clustering sample data of all-link fault nodes; the data clustering process is as follows: S521. Treat each full-link fault node operation feature sample data in the full-link fault node operation feature sample data set as a cluster, calculate the Euclidean distance between clusters, and obtain a distance matrix; the Euclidean distance calculation formula is as follows: , in, Representation vector and vector The Euclidean distance between and Represents the centers of any two clusters dimensional feature vector, Representation vector In the The values ​​in the dimensions, Representation vector In the The value of the dimension; S522, merging the two clusters corresponding to the minimum values ​​in the distance matrix into a new cluster, recalculating the Euclidean distance between the new cluster and other clusters, and updating the distance matrix; S523, setting a Euclidean distance threshold. If the minimum value in the distance matrix is ​​greater than or equal to the Euclidean distance threshold, the data clustering processing operation is terminated to obtain a full-link fault node operation feature clustering sample data set; otherwise, return to S512 until the minimum value in the distance matrix is ​​greater than or equal to the Euclidean distance threshold; S53, storing the full-link fault node operation feature clustering sample data in the full-link fault node operation feature clustering sample data set in the full-link fault type identification unit, and entering corresponding text feature data for each type of full-link fault node operation feature clustering sample data through the full-link fault type text entry interface to generate a full-link fault type text feature data set ,in, Indicates The full-link fault type text feature data corresponding to the full-link fault node operation feature clustering sample data.

[0013] The agglomerative hierarchical clustering algorithm is used to perform data clustering on the collected full-link fault node operation feature sample data, establish a good and reliable full-link fault analysis standard, and provide a tool for full-link fault type identification. At the same time, the bottom-up clustering method can intuitively identify the differences between different data, ensuring the quality and accuracy of the clustering results.

[0014] Preferably, the specific steps of performing full-link fault type identification processing according to the full-link node real-time operation characteristic data and the full-link fault type text characteristic data to generate the full-link real-time fault type text characteristic data are as follows: S61, through the teaching and learning optimization algorithm, the real-time operation feature data set of the full-link node of the Internet of Things and the full-link fault type text feature data in the full-link fault type text feature data set are processed for full-link fault type identification, and a full-link real-time fault type feature data set is generated. ; S611, construct a student population, set the size of the student population to , the current number of iterations is The maximum number of iterations is And the search space dimension of the full-link fault type text feature data is ; Randomly generate in the full link fault type text feature data search space Full-link fault type text feature data, each full-link fault type text feature data corresponds to a student individual in the student population, and the initial position set of the student population is initialized as ,in, Indicates the first The initial position of each student; S612, calculate the fitness function value of each student in the student population, arrange the student individuals in the student population from large to small according to the fitness function value, and select the student individual with the highest fitness function value as the teacher individual; the fitness function formula is as follows: , in, Indicates the first The fitness function value of each student is Indicates the first The Euclidean distance between the full-link fault node operation feature clustering sample data corresponding to the full-link fault type text feature data corresponding to the individual students and the real-time operation feature data of the full-link node of the Internet of Things in the real-time operation feature data set of the full-link node of the Internet of Things, Indicates the correction value; S613, during the teaching phase, each student individual in the student population updates its position in the full-link fault type text feature data search space by learning from the teacher individual; the position update formula is as follows: , in, represents a random number that is uniformly distributed between (0,1). represents the teaching factor, represents the position of individual teachers, Indicates the student population The location of each student after the location update, Indicates the student population The current location of each student, represents the average position of all individual students; S614, in the learning phase, each student individual in the student population updates their position in the full-link fault type text feature data search space through mutual learning; the position update formula is as follows: , in, represents the first student randomly selected from the student population The current location of each student, Represents a random number that obeys a uniform distribution between (0,1); S615, calculating the fitness function value of each student individual in the student population after the position is updated in the full-link fault type text feature data search space, if the fitness function value of the student individual after the position is updated is greater than the fitness function value before the position is updated, then the new position is used to replace the original position; otherwise, the original position is retained; S616, if the fitness function value of the student individual is greater than the fitness function value of the teacher individual, the student individual is used as a new teacher individual; otherwise, directly enter S617; S617, determine the current number of iterations Is it less than the maximum number of iterations? , if the current number of iterations Less than the maximum number of iterations , then the current number of iterations Add 1 and return to S613; otherwise, take the current teacher individual as the global optimal solution, output the full-link fault type text feature data corresponding to the global optimal solution and perform data identification to generate the full-link real-time fault type text feature data .

[0015] Through the teaching and learning optimization algorithm, the real-time operation feature data set of the full-link node and the full-link fault type text feature data are processed for full-link fault type identification, which improves the speed and efficiency of the identification process and realizes the scientific identification of the full-link fault type. At the same time, the teaching and learning optimization algorithm can effectively avoid the recognition result from falling into the local optimal solution, thereby ensuring the accuracy of the recognition result.

[0016] Preferably, the specific steps of constructing the IoT full-link diagnostic data based on the IoT full-link analysis data, the real-time operation feature data of the full-link nodes, and the real-time fault type text feature data of the full-link, and pushing it to the IoT full-link diagnostic platform are as follows: S71, analyzing the IoT full-link data , the full-link node real-time operation feature data set and the full-link real-time fault type characteristic data Combine data to build IoT full-link diagnostic data .

[0017] S72, transmitting the IoT full-link diagnostic data through the IoT communication network Push to the IoT full-link diagnosis platform to end this IoT full-link diagnosis operation.

[0018] The present invention also discloses an Internet of Things data full-link diagnosis and analysis system, including a full-link terminal receiving data acquisition unit, a full-link terminal receiving data analysis unit, a full-link node real-time operation data acquisition unit, a full-link node real-time operation data preprocessing unit, a full-link fault type text feature data construction unit, a full-link fault type identification unit and an Internet of Things full-link diagnosis data construction unit; The full-link terminal receiving data collection unit collects production data from the Internet of Things data source online through the industrial plant data collection unit, and pushes it to the industrial plant data terminal receiving unit, and presets the collected production data as standard full-link terminal receiving data; receives the production data pushed by the Internet of Things communication network through the industrial plant data terminal receiving unit, and generates real full-link terminal receiving data; The full-link terminal reception data analysis unit calculates the similarity between the standard full-link terminal reception data and the real full-link terminal reception data by using a cosine similarity algorithm, and compares the similarity with a set similarity threshold, and generates IoT full-link analysis data according to the numerical comparison result; When the IoT full-link analysis data is abnormal, the full-link node real-time operation data collection unit collects the real-time operation data of each key node in the IoT full-link online through the IoT link node data management platform to generate the full-link node real-time operation data; The full-link node real-time operation data preprocessing unit uses an exponential sliding filter method to perform data noise reduction processing on the full-link node real-time operation data to generate full-link node real-time operation feature data; The full-link fault type text feature data construction unit uses an agglomerative hierarchical clustering algorithm to perform data clustering processing on the collected full-link fault node operation feature sample data to generate full-link fault node operation feature cluster sample data, and enters corresponding text feature data for various types of full-link fault node operation feature cluster sample data through a full-link fault type text entry interface to generate full-link fault type text feature data; The full-link fault type identification unit performs full-link fault type identification processing on the full-link node real-time operation feature data set and the full-link fault type text feature data through a teaching and learning optimization algorithm to generate full-link real-time fault type feature data; The IoT full-link diagnostic data construction unit constructs the IoT full-link diagnostic data by combining the IoT full-link analysis data, the full-link node real-time operation feature data set and the full-link real-time fault type feature data, and pushes it to the IoT full-link diagnostic data through the IoT communication network.

[0019] (III) Beneficial effects 1. The present invention generates IoT full-link analysis data by performing data similarity analysis on the collected standard full-link terminal reception data and the real full-link terminal reception data. When the IoT full-link analysis data is abnormal, data preprocessing is performed on the collected real-time operation data of the full-link nodes to generate real-time operation feature data of the full-link nodes, and full-link fault type identification processing is performed with the established full-link fault type text feature data to generate full-link real-time fault type text feature data. Finally, IoT full-link diagnostic data is constructed and pushed to the IoT full-link diagnostic platform to achieve accurate diagnosis and timely feedback of IoT data full-link. 2. By performing data similarity analysis on the collected standard full-link terminal reception data and the real full-link terminal reception data, a preliminary analysis is made on whether there is any abnormality in the entire IoT link, avoiding real-time monitoring of each key node in the entire link at all times, and effectively reducing the cost of IoT data full-link diagnosis; 3. Use the agglomerative hierarchical clustering algorithm to perform data clustering on the collected full-link fault node operation feature sample data, establish a good and reliable full-link fault analysis standard, and provide a tool for full-link fault type identification. At the same time, the bottom-up clustering method can intuitively identify the differences between different data, ensuring the quality and accuracy of the clustering results; 4. The teaching and learning optimization algorithm is used to perform full-link fault type identification processing on the real-time operation feature data set of the full-link nodes and the full-link fault type text feature data, thereby improving the speed and efficiency of the identification process and realizing scientific identification of full-link fault types. At the same time, the teaching and learning optimization algorithm can effectively avoid the recognition results from falling into the local optimal solution and ensure the accuracy of the recognition results. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0021] Figure 1 A flowchart of a full-link diagnostic analysis method for Internet of Things data provided by the present invention; Figure 2 A module schematic diagram of a full-link diagnostic analysis system for Internet of Things data provided by the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0024] Embodiment 1 is as follows: See also Figure 1 , a full-link diagnosis and analysis method for IoT data, comprising the following steps: S1. Collect standard full-link terminal reception data and real full-link terminal reception data; S11. Collect production data from the IoT data source online through the industrial plant data collection unit, and push it to the industrial plant data terminal receiving unit through the IoT communication network. At the same time, preset the collected production data as standard full-link terminal receiving data. , the production data of the IoT data source includes but is not limited to any one of thermal power data, hydropower data, wind power data, photovoltaic data, natural gas data, and coal data; Receive production data pushed by the IoT communication network through the industrial plant data terminal receiving unit to generate real full-link terminal receiving data .

[0025] S2. Perform data similarity analysis and processing based on the data received by the standard full-link terminal and the data received by the real full-link terminal to generate IoT full-link analysis data. When the data is normal, end the IoT full-link diagnosis operation. S21, calculating the standard full-link terminal receiving data by using a cosine similarity algorithm and the real full-link terminal receives data Similarity between ; The cosine similarity algorithm formula is as follows: , in, represents the similarity between the standard full-link terminal receiving data and the real full-link terminal receiving data, and Characteristic vectors respectively representing the standard full-link terminal receiving data and the real full-link terminal receiving data; S22, set the similarity threshold to , the similarity between the standard full-link terminal receiving data and the real full-link terminal receiving data and the similarity threshold Perform numerical comparison and generate IoT full-link analysis data based on the numerical comparison results ; like , it means that the IoT full-link transmission process is normal, and the IoT full-link analysis data is output The result is normal and is pushed to the IoT full-link diagnosis platform through the IoT communication network, ending this IoT full-link diagnosis operation. like , it means that the IoT full-link transmission process is abnormal, and the IoT full-link analysis data is output is abnormal.

[0026] S3. When the IoT full-link analysis data is abnormal, collect the real-time operation data of the full-link nodes; S31, when the IoT full-link analysis data When an abnormality occurs, the real-time operation data of each key node in the entire IoT link is collected online through the IoT link node data management platform to generate a real-time operation data set for the entire link node. ,in, It represents the first data in the entire IoT link collected online through the IoT link node data management platform. Real-time operation data of key nodes, Indicates the total number of real-time operation data of all-link nodes, including but not limited to packet loss rate, network delay, available bandwidth, IP address, CPU usage and network protocol information.

[0027] S4, performing data preprocessing on the full-link node operation data to generate real-time operation feature data of the full-link node; S41, using the exponential sliding filtering method to perform data noise reduction processing on each full-link node real-time operation data set in turn, to generate a full-link node real-time operation feature data set ,in, It represents the first data in the entire IoT link collected online through the IoT link node data management platform. The real-time operation data of the key nodes are processed by data noise reduction to obtain the real-time operation characteristic data of the entire link nodes.

[0028] S5. Establishing text feature data of all-link fault types; S51. Collect the operation characteristic data of each key node when different faults occur in the entire IoT link through the IoT platform, and generate a sample data set of operation characteristic of the fault node in the entire link ,in, Indicates Full-link fault node operation feature sample data, Indicates the total number of running feature sample data of all-link fault nodes; S52, using an agglomerative hierarchical clustering algorithm to perform data clustering processing on the full-link fault node operation feature sample data in the full-link fault node operation feature sample data set to generate a full-link fault node operation feature clustering sample data set ,in, Represents the first Clustering sample data of running characteristics of nodes with full-link failures. Indicates the total number of running feature clustering sample data of all-link fault nodes; the data clustering process is as follows: S521. Treat each full-link fault node operation feature sample data in the full-link fault node operation feature sample data set as a cluster, calculate the Euclidean distance between clusters, and obtain a distance matrix; the Euclidean distance calculation formula is as follows: , in, Representation vector and vector The Euclidean distance between and Represents the centers of any two clusters dimensional feature vector, Representation vector In the The values ​​in the dimensions, Representation vector In the The value of the dimension; S522, merging the two clusters corresponding to the minimum values ​​in the distance matrix into a new cluster, recalculating the Euclidean distance between the new cluster and other clusters, and updating the distance matrix; S523, setting a Euclidean distance threshold. If the minimum value in the distance matrix is ​​greater than or equal to the Euclidean distance threshold, the data clustering processing operation is terminated to obtain a full-link fault node operation feature clustering sample data set; otherwise, return to S512 until the minimum value in the distance matrix is ​​greater than or equal to the Euclidean distance threshold; S53, storing the full-link fault node operation feature clustering sample data in the full-link fault node operation feature clustering sample data set in the full-link fault type identification unit, and entering corresponding text feature data for each type of full-link fault node operation feature clustering sample data through the full-link fault type text entry interface to generate a full-link fault type text feature data set ,in, Indicates The full-link fault type text feature data corresponding to the full-link fault node operation feature clustering sample data.

[0029] S6. Perform full-link fault type identification processing according to the full-link node real-time operation feature data and the full-link fault type text feature data to generate full-link real-time fault type text feature data; S61, through the teaching and learning optimization algorithm, the real-time operation feature data set of the full-link node of the Internet of Things and the full-link fault type text feature data in the full-link fault type text feature data set are processed for full-link fault type identification, and a full-link real-time fault type feature data set is generated. ; S611, construct a student population, set the size of the student population to , the current number of iterations is The maximum number of iterations is And the search space dimension of the full-link fault type text feature data is ; Randomly generate in the full link fault type text feature data search space Full-link fault type text feature data, each full-link fault type text feature data corresponds to a student individual in the student population, and the initial position set of the student population is initialized as ,in, Indicates the first The initial position of each student; S612, calculate the fitness function value of each student in the student population, arrange the student individuals in the student population from large to small according to the fitness function value, and select the student individual with the highest fitness function value as the teacher individual; the fitness function formula is as follows: , in, Indicates the first The fitness function value of each student is Indicates the first The Euclidean distance between the full-link fault node operation feature clustering sample data corresponding to the full-link fault type text feature data corresponding to the individual students and the real-time operation feature data of the full-link node of the Internet of Things in the real-time operation feature data set of the full-link node of the Internet of Things, Indicates the correction value; S613, during the teaching phase, each student individual in the student population updates its position in the full-link fault type text feature data search space by learning from the teacher individual; the position update formula is as follows: , in, represents a random number that is uniformly distributed between (0,1). represents the teaching factor, represents the position of individual teachers, Indicates the student population The location of each student after the location update, Indicates the student population The current location of each student, represents the average position of all individual students; S614, in the learning phase, each student individual in the student population updates their position in the full-link fault type text feature data search space through mutual learning; the position update formula is as follows: , in, represents the first student randomly selected from the student population The current location of each student, Represents a random number that obeys a uniform distribution between (0,1); S615, calculating the fitness function value of each student individual in the student population after the position is updated in the full-link fault type text feature data search space, if the fitness function value of the student individual after the position is updated is greater than the fitness function value before the position is updated, then the new position is used to replace the original position; otherwise, the original position is retained; S616, if the fitness function value of the student individual is greater than the fitness function value of the teacher individual, the student individual is used as a new teacher individual; otherwise, directly enter S617; S617, determine the current number of iterations Is it less than the maximum number of iterations? , if the current number of iterations Less than the maximum number of iterations , then the current number of iterations Add 1 and return to S613; otherwise, take the current teacher individual as the global optimal solution, output the full-link fault type text feature data corresponding to the global optimal solution and perform data identification to generate the full-link real-time fault type text feature data .

[0030] S7. Constructing IoT full-link diagnostic data based on the IoT full-link analysis data, the real-time operation feature data of the full-link nodes, and the real-time fault type text feature data of the full-link, and pushing it to the IoT full-link diagnostic platform; S71, analyzing the IoT full-link data , the full-link node real-time operation feature data set and the full-link real-time fault type characteristic data Combine data to build IoT full-link diagnostic data .

[0031] S72, transmitting the IoT full-link diagnostic data through the IoT communication network Push to the IoT full-link diagnosis platform to end this IoT full-link diagnosis operation.

[0032] Embodiment 2 is as follows: See also Figure 2 , an Internet of Things data full-link diagnosis and analysis system, including a full-link terminal reception data acquisition unit, a full-link terminal reception data analysis unit, a full-link node real-time operation data acquisition unit, a full-link node real-time operation data preprocessing unit, a full-link fault type text feature data construction unit, a full-link fault type identification unit and an Internet of Things full-link diagnosis data construction unit; The full-link terminal receiving data collection unit collects production data from the Internet of Things data source online through the industrial plant data collection unit, and pushes it to the industrial plant data terminal receiving unit, and presets the collected production data as standard full-link terminal receiving data; receives the production data pushed by the Internet of Things communication network through the industrial plant data terminal receiving unit, and generates real full-link terminal receiving data; The full-link terminal reception data analysis unit calculates the similarity between the standard full-link terminal reception data and the real full-link terminal reception data by using a cosine similarity algorithm, and compares the similarity with a set similarity threshold, and generates IoT full-link analysis data according to the numerical comparison result; When the IoT full-link analysis data is abnormal, the full-link node real-time operation data collection unit collects the real-time operation data of each key node in the IoT full-link online through the IoT link node data management platform to generate the full-link node real-time operation data; The full-link node real-time operation data preprocessing unit uses an exponential sliding filter method to perform data noise reduction processing on the full-link node real-time operation data to generate full-link node real-time operation feature data; The full-link fault type text feature data construction unit uses an agglomerative hierarchical clustering algorithm to perform data clustering processing on the collected full-link fault node operation feature sample data to generate full-link fault node operation feature cluster sample data, and enters corresponding text feature data for various types of full-link fault node operation feature cluster sample data through a full-link fault type text entry interface to generate full-link fault type text feature data; The full-link fault type identification unit performs full-link fault type identification processing on the full-link node real-time operation feature data set and the full-link fault type text feature data through a teaching and learning optimization algorithm to generate full-link real-time fault type feature data; The IoT full-link diagnostic data construction unit constructs IoT full-link diagnostic data by combining the IoT full-link analysis data, the full-link node real-time operation feature data set and the full-link real-time fault type feature data, and pushes it to the IoT full-link diagnostic data through the IoT communication network.

[0033] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0034] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A full-link diagnostic analysis method for Internet of Things data, characterized in that: The steps include: S1. Collect standard full-link terminal reception data and real full-link terminal reception data; S2. Perform data similarity analysis and processing based on the data received by the standard full-link terminal and the data received by the real full-link terminal to generate IoT full-link analysis data. When the data is normal, end the IoT full-link diagnosis operation. S3. When the IoT full-link analysis data is abnormal, collect the real-time operation data of the full-link nodes; S4, performing data preprocessing on the full-link node operation data to generate real-time operation feature data of the full-link node; S5. Establishing text feature data of all-link fault types; S6. Perform full-link fault type identification processing according to the full-link node real-time operation feature data and the full-link fault type text feature data to generate full-link real-time fault type text feature data; S7. Construct the IoT full-link diagnostic data based on the IoT full-link analysis data, the real-time operation feature data of the full-link nodes, and the real-time fault type text feature data of the full-link, and push it to the IoT full-link diagnostic platform.

2. According to claim 1, a method for full-link diagnosis and analysis of Internet of Things data is characterized in that: The S1 comprises the following steps: S11. Collect production data from the IoT data source online through the industrial plant data collection unit, and push it to the industrial plant data terminal receiving unit through the IoT communication network. At the same time, preset the collected production data as standard full-link terminal receiving data. , the production data of the IoT data source includes but is not limited to any one of thermal power data, hydropower data, wind power data, photovoltaic data, natural gas data, and coal data; Receive production data pushed by the IoT communication network through the industrial plant data terminal receiving unit to generate real full-link terminal receiving data .

3. The method for full-link diagnosis and analysis of Internet of Things data according to claim 2 is characterized in that: The S2 comprises the following steps: S21, calculating the standard full-link terminal receiving data by using a cosine similarity algorithm and the real full-link terminal receives data Similarity between ; S22, set the similarity threshold to , the similarity between the standard full-link terminal receiving data and the real full-link terminal receiving data and the similarity threshold Perform numerical comparison and generate IoT full-link analysis data based on the numerical comparison results ; like , it means that the IoT full-link transmission process is normal, and the IoT full-link analysis data is output The result is normal and is pushed to the IoT full-link diagnosis platform through the IoT communication network, ending this IoT full-link diagnosis operation. like , it means that the IoT full-link transmission process is abnormal, and the IoT full-link analysis data is output is abnormal.

4. The method for full-link diagnosis and analysis of Internet of Things data according to claim 3 is characterized in that: The S3 comprises the following steps: S31, when the IoT full-link analysis data When an abnormality occurs, the real-time operation data of each key node in the entire IoT link is collected online through the IoT link node data management platform to generate a real-time operation data set for the entire link node. ,in, It represents the first data in the entire IoT link collected online through the IoT link node data management platform. Real-time operation data of key nodes, Indicates the total number of real-time operation data of all-link nodes, including but not limited to packet loss rate, network delay, available bandwidth, IP address, CPU usage and network protocol information.

5. The method for full-link diagnosis and analysis of Internet of Things data according to claim 4 is characterized in that: The S4 comprises the following steps: S41, using the exponential sliding filtering method to perform data noise reduction processing on each full-link node real-time operation data set in turn, to generate a full-link node real-time operation feature data set ,in, It represents the first data in the entire IoT link collected online through the IoT link node data management platform. The real-time operation data of the key nodes are processed by data noise reduction to obtain the real-time operation characteristic data of the entire link nodes.

6. The method for full-link diagnosis and analysis of Internet of Things data according to claim 5 is characterized in that: The S5 comprises the following steps: S51. Collect the operation characteristic data of each key node when different faults occur in the entire IoT link through the IoT platform, and generate a sample data set of operation characteristic of the fault node in the entire link ,in, Indicates Full-link fault node operation feature sample data, Indicates the total number of running feature sample data of all-link fault nodes; S52, using an agglomerative hierarchical clustering algorithm to perform data clustering processing on the full-link fault node operation feature sample data in the full-link fault node operation feature sample data set to generate a full-link fault node operation feature clustering sample data set ,in, Represents the first Clustering sample data of running characteristics of nodes with full-link failures. Indicates the total number of running feature clustering sample data of all-link fault nodes; the data clustering process is as follows: S521, treating each full-link fault node operation feature sample data in the full-link fault node operation feature sample data set as a cluster, calculating the Euclidean distance between clusters, and obtaining a distance matrix; S522, merging the two clusters corresponding to the minimum values ​​in the distance matrix into a new cluster, recalculating the Euclidean distance between the new cluster and other clusters, and updating the distance matrix; S523, setting a Euclidean distance threshold. If the minimum value in the distance matrix is ​​greater than or equal to the Euclidean distance threshold, the data clustering processing operation is terminated to obtain a full-link fault node operation feature clustering sample data set; otherwise, return to S512 until the minimum value in the distance matrix is ​​greater than or equal to the Euclidean distance threshold; S53, storing the full-link fault node operation feature clustering sample data in the full-link fault node operation feature clustering sample data set in the full-link fault type identification unit, and entering corresponding text feature data for each type of full-link fault node operation feature clustering sample data through the full-link fault type text entry interface to generate a full-link fault type text feature data set ,in, Indicates The full-link fault type text feature data corresponding to the full-link fault node operation feature clustering sample data.

7. The method for full-link diagnosis and analysis of Internet of Things data according to claim 6 is characterized in that: The S6 comprises the following steps: S61, through the teaching and learning optimization algorithm, the real-time operation feature data set of the full-link node of the Internet of Things and the full-link fault type text feature data in the full-link fault type text feature data set are processed for full-link fault type identification, and a full-link real-time fault type feature data set is generated. .

8. The method for full-link diagnosis and analysis of Internet of Things data according to claim 7 is characterized in that: The S61 comprises the following steps: S611, construct a student population, set the size of the student population to , the current number of iterations is The maximum number of iterations is And the search space dimension of the full-link fault type text feature data is ; Randomly generate in the full link fault type text feature data search space full-link fault type text feature data, each full-link fault type text feature data corresponds to a student individual in the student population, and an initial position set of the student population is initialized; S612, calculating the fitness function value of each student individual in the student population, arranging each student individual in the student population from large to small according to the fitness function value, and selecting the student individual with the highest fitness function value as the teacher individual; S613, during the teaching phase, each student individual in the student population updates its position in the full-link fault type text feature data search space by learning from the teacher individual; S614, in the learning stage, each student individual in the student population updates the position in the full-link fault type text feature data search space through mutual learning; S615, calculating the fitness function value of each student individual in the student population after the position is updated in the full-link fault type text feature data search space, if the fitness function value of the student individual after the position is updated is greater than the fitness function value before the position is updated, then the new position is used to replace the original position; otherwise, the original position is retained; S616, if the fitness function value of the student individual is greater than the fitness function value of the teacher individual, the student individual is used as a new teacher individual; otherwise, directly enter S617; S617, determine the current number of iterations Is it less than the maximum number of iterations? , if the current number of iterations Less than the maximum number of iterations , then the current number of iterations Add 1 and return to S613; otherwise, take the current teacher individual as the global optimal solution, output the full-link fault type text feature data corresponding to the global optimal solution and perform data identification to generate the full-link real-time fault type text feature data .

9. The method for full-link diagnosis and analysis of Internet of Things data according to claim 8, characterized in that: The S7 comprises the following steps: S71, analyzing the IoT full-link data , the full-link node real-time operation feature data set and the full-link real-time fault type characteristic data Combine data to build IoT full-link diagnostic data ; S72, transmitting the IoT full-link diagnostic data through the IoT communication network Push to the IoT full-link diagnosis platform to end this IoT full-link diagnosis operation.

10. A system for implementing the full-link diagnostic analysis method for Internet of Things data as described in any one of claims 1-9.

Images