Digital quantity acquisition and communication transmission method

Abstract

The invention discloses a digital quantity acquisition and communication transmission method, which relates to the technical field of big data acquisition and comprises a digital quantity acquisition step and a digital quantity communication transmission step. Wherein the digital quantity acquisition method comprises the following steps: acquiring data information through a data extraction module, performing data information dimension reduction processing on the extracted data information, preprocessing the acquired abnormal data information, and realizing loss calculation of the data information through line loss fluctuation; the digital quantity communication transmission method comprises the steps that multi-channel information transmission is conducted on data information, in the information transmission process, sensing of the data information is achieved through a network sensing module, and fault diagnosis of data information communication abnormity is achieved through an improved GAN algorithm module; therefore, integration of digital quantity acquisition and communication is realized. The data acquisition capability is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of big data acquisition, and more particularly to a digital quantity acquisition and communication transmission method. Background Art

[0002] In industrial automation control, we often encounter concepts such as switch quantities, digital quantities, analog quantities, and pulse quantities. Digital quantity acquisition modules are used to acquire digital quantities. Generally, digital quantities are signals composed of 0s and 1s, often encoded and arranged in a regular pattern. However, for switch quantities, a closed contact is considered a 1, and an open contact is a 0. A switch input module converts switch quantities into digital signals for acquisition. Analog quantities can be quantized using threshold values, with values ​​below that value representing 0 and values ​​greater than or equal to that value representing 1. The DM-TC-DI switch input isolator is suitable for contact switches, NAMUR proximity switches, OC signals, or level inputs. After isolation, the signals are transmitted to the control system or other instrumentation via relay contact outputs (or OC outputs or level outputs).

[0003] In the data communication process, existing technologies mostly use data information collection and data communication through data communication transmission modules. This separate operation method often results in equipment waste, extremely low work efficiency, serious waste of materials, and cannot achieve the integration of digital quantity collection and communication. Summary of the Invention

[0004] In view of the shortcomings of the above technologies, the present invention discloses a digital quantity acquisition and communication transmission method, which greatly improves the data acquisition capability.

[0005] In order to achieve the above technical effects, the present invention adopts the following technical solutions: A digital quantity acquisition and communication transmission method, including the following methods: It includes digital quantity acquisition steps and digital quantity communication transmission steps; wherein: The digital quantity acquisition method is: The data extraction module obtains data information, performs dimensionality reduction on the extracted data information, pre-processes the collected abnormal data information, and calculates the loss of data information through line loss fluctuations; The digital communication transmission method is: Multi-channel information transmission is performed on data information. During the information transmission process, data information is perceived through the network perception module, and fault diagnosis of data information communication anomalies is achieved through the improved GAN algorithm module; thereby realizing the integration of digital quantity acquisition and communication.

[0006] As a further technical solution of the present invention, the data extraction module realizes the control of data information extraction through the CC2530 chip.

[0007] As a further technical solution of the present invention, the transmission of data information is achieved through a ZigBee communication module, wherein the ZigBee communication module is provided with an MSP430F249 single-chip microcomputer control unit.

[0008] As a further technical solution of the present invention, the method for analyzing abnormal data information is: The collected network data communication data sequence is divided into multiple segments of equal length, and the average value of each segment is calculated. The original communication data sequence is recorded as , the PAA algorithm is used to reduce the dimension of communication data information. The sequence after dimension reduction is expressed as: (1) In formula (1), Indicates the length of the collected original data information sequence, represents the length of the communication data information sequence after dimensionality reduction, Represents the average value of the collected original data information sequence, It represents the fragment of the original data sequence collected. After being processed by the PAA algorithm, the length of the original data communication is reduced from down to ; Preprocess the abnormal network large data set and use the interpolation method to deal with the missing values ​​in the network large data set. The missing value function is expressed as: (2) In formula (2), The serial number representing the missing value, represents the ordinal number of the non-missing value, Represents the interpolation cardinality. For the periodic analysis of the transmission of abnormal large data sets on the network, network communication has large fluctuations. The continuous fluctuation value of the communication protocol in the data channel has continuously decreased to a stable state. The network channel communication function is expressed as: (3) In formula (3), Indicates the statistical period of data communication, Indicates the year-on-year fluctuation value of the network communication protocol. Indicates the average year-on-year fluctuation value within the data communication statistical period. Indicates the month-on-month fluctuation value of the data communication environment. Indicates the average month-on-month fluctuation value within the data communication statistical period; The continuous fluctuation value function of communication line loss is expressed as: (4) In formula (4), is the continuous fluctuation value of line loss, is the line loss rate of the previous communication line statistical cycle, Line loss rate for the next communication line statistical cycle; A classification and regression tree is used to establish an abnormal data communication identification model, and the feature space is divided into two categories to distinguish normal network data communication from abnormal network data communication. The Gini coefficient function of the probability distribution is expressed as: (5) In formula (5), Indicates the number of communication protocol classes in the data communication process. Indicates that the communication protocol sample point belongs to The probability of the class is obtained by calculating the Gini index of the current data communication sample set, using data communication to summarize all the values ​​of each abnormal data information feature, calculating the Gini coefficient, and then finding the corresponding optimal segmentation features and values ​​to generate a decision tree for abnormal information identification.

[0009] As a further technical solution of the present invention, the classification and regression tree includes a classifier.

[0010] As a further technical solution of the present invention, the network perception module adopts the STM32F103ZET6 microcontroller unit to realize the perception calculation of communication data information.

[0011] As a further technical solution of the present invention, the network perception module is further provided with an isolation circuit.

[0012] As a further technical solution of the present invention, the isolation circuit is a photoelectric isolator of the ADUM1250 chip.

[0013] As a further technical solution of the present invention, the improved GAN algorithm module includes a communication feature extraction model, a communication data generation model and Communication data discrimination model , wherein the output end of the communication feature extraction model is connected to the input end of the communication data generation model, and the output end of the communication data generation model is connected to the input end of the communication data discrimination model.

[0014] As a further technical solution of the present invention, the working method of the communication feature extraction model is: The network data information features are extracted through a multi-layer perceptron. The output function of the communication feature extraction model is expressed as: (6) In formula (6), Represents the discriminant model in the GAN network, Represents the generation model in the GAN network, represents the true distribution of the input communication network parameters, represents the distribution of input noise data, represents the data sampling process of the model, represents the noise data in the communication network parameter data, Represents the data communication node in the GAN network; The communication data generation model realizes the generation of communication data through matrix functions. The method for the communication data generation model to realize data information generation is: The input vectors for different network states are , network communication data is represented as ,When a fault occurs in the power communication network, the network state information generation function during the fault period is: (7) In formula (7), Indicates Time The value of the network parameters, Indicates the fault duration, The overall network communication data information, The entirety of the input vector representing the state of the network, Indicates the number of input vectors; The communication data discrimination model realizes the discrimination of the network data information communication process through the loss function, where the loss function is: (8) In formula (8), Represents the loss value output by the communication data discrimination model, Indicates the leaf node during communication diagnosis, Indicates the number of leaf nodes during communication diagnosis, represents the regularization parameter for communication diagnosis, The communication matrix function value representing different communication data information losses during communication diagnosis, Represents the structural parameters of the model.

[0015] The beneficial positive effects of the present invention are: The invention comprises a digital quantity acquisition step and a digital quantity communication transmission step; wherein: the digital quantity acquisition method comprises: obtaining data information through a data extraction module, performing dimensionality reduction processing on the extracted data information, pre-processing the collected abnormal data information, and realizing loss calculation of the data information through line loss fluctuation; the digital quantity communication transmission method comprises: performing multi-channel information transmission on the data information, realizing data information perception through a network perception module during the information transmission process, and realizing fault diagnosis of data information communication abnormality through an improved GAN algorithm module; thereby realizing digital quantity acquisition and communication as one. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive work, among which: Figure 1 This is a schematic diagram of the abnormal data communication identification process in the present invention; Figure 2 Schematic diagram of the hardware circuit structure of the concentrator in the present invention; Figure 3 Schematic diagram of the isolation circuit in the present invention. DETAILED DESCRIPTION

[0017] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the embodiments described herein are only used to illustrate and explain the present invention and are not used to limit the present invention.

[0018] like Figure 1-Figure 3 As shown, A digital quantity acquisition and communication transmission method includes the following methods: It includes digital quantity acquisition steps and digital quantity communication transmission steps; wherein: The digital quantity acquisition method is: The data extraction module obtains data information, performs dimensionality reduction on the extracted data information, pre-processes the collected abnormal data information, and calculates the loss of data information through line loss fluctuations; The digital communication transmission method is: Multi-channel information transmission is performed on data information. During the information transmission process, data information is perceived through the network perception module, and fault diagnosis of data information communication anomalies is achieved through the improved GAN algorithm module; thereby realizing the integration of digital quantity acquisition and communication.

[0019] In a specific embodiment, the present invention integrates network data information collection and communication, realizes synchronous calculation of data information, and greatly improves network data information perception and information processing capabilities.

[0020] In a specific embodiment, the data extraction module implements the control of data information extraction through the CC2530 chip.

[0021] In a specific embodiment, data information extraction control is achieved by using the CC2530 chip, which has 8KB of memory. This approach can provide power for a variety of data information, facilitating the configuration of information collection terminals. The CC2530 chip can also simultaneously receive control commands for data information collection, ensuring the effectiveness of the system's control over the terminal. The wireless communication module uses the H685 TD-LTE device to meet the communication needs of various application scenarios.

[0022] In a specific embodiment, the transmission of data information is achieved through a ZigBee communication module, wherein the ZigBee communication module is provided with an MSP430F249 single-chip microcomputer control unit.

[0023] In a specific embodiment, data communication is achieved through an MSP430F249 single-chip microcomputer control unit. A concentrator can also be used, using the MSP430F249 single-chip microcomputer as the main control module. The MSP single-chip microcomputer has a variety of addressing modes, is equipped with registers and on-chip RAM storage, and has a rich set of control instructions, enabling users to debug programs on the microcontroller in real time. The concentrator hardware structure also includes peripheral circuits such as a data storage module, a power module, and a clock module. A voltage regulator chip converts the output voltage and provides it to the communication module and the main control module. The voltage regulator chip uses an RH5RE33 to convert the output voltage to 3.3V.

[0024] The method for abnormal data information analysis is: In a network system, the network communication platform provides support for the storage, processing, and analysis of large amounts of data communications. The present invention performs dimensionality reduction on data with large volumes and low information density, facilitating further data mining, reducing the space used for data storage, and improving the efficiency of identifying abnormal data communications. The network data collected by the acquisition layer has high real-time performance, and the dimensions of the original data communication are relatively large. The present invention uses the Piecewise Aggregate Approximation (PAA) method to reduce the data dimensionality of the collected information and extract data sequence features by averaging the network data sequence in segments. In specific embodiments, other data dimensionality reduction methods can also be used to implement abnormal data information analysis and processing.

[0025] The collected network data communication data sequence is divided into multiple segments of equal length, and the average value of each segment is calculated. The original communication data sequence is recorded as , the PAA algorithm is used to reduce the dimension of communication data information. The sequence after dimension reduction is expressed as: (1) In formula (1), Indicates the length of the collected original data information sequence, represents the length of the communication data information sequence after dimensionality reduction, Represents the average value of the collected original data information sequence, It represents the fragment of the original data sequence collected. After being processed by the PAA algorithm, the length of the original data communication is reduced from down to During the dimensionality reduction process, each segment is replaced by its average value, reflecting the overall situation of each segment. The reduced sequence reflects the changing trend of the entire original sequence. This method has a good dimensionality reduction effect on various time series big data generated by power equipment such as electricity meters, collectors, and concentrators. After dimensionality reduction of data communication big data, abnormal data communication is identified to promptly detect possible network data anomalies.

[0026] Preprocess the abnormal network large data set and use the interpolation method to deal with the missing values ​​in the network large data set. The missing value function is expressed as: (2) In formula (2), The serial number representing the missing value, represents the ordinal number of the non-missing value, Represents the interpolation cardinality. For the periodic analysis of the transmission of abnormal large data sets on the network, network communication has large fluctuations. The continuous fluctuation value of the communication protocol in the data channel has continuously decreased to a stable state. The network channel communication function is expressed as: (3) In formula (3), Indicates the statistical period of data communication, Indicates the year-on-year fluctuation value of the network communication protocol. Indicates the average year-on-year fluctuation value within the data communication statistical period. Indicates the month-on-month fluctuation value of the data communication environment. Indicates the average month-on-month fluctuation value within the data communication statistical period; During data communication, the line loss rate during network communication will increase. To prevent other factors from causing circuit communication loss and leading to an increase in line loss rate, a fixed number of working days is set as the statistical window. The continuous fluctuation value function of communication line loss is expressed as: (4) In formula (4), is the continuous fluctuation value of line loss, is the line loss rate of the previous communication line statistical cycle, Line loss rate for the next communication line statistical cycle; A classification and regression tree is used to establish an abnormal data communication identification model, and the feature space is divided into two categories to distinguish normal network data communication from abnormal network data communication. The Gini coefficient function of the probability distribution is expressed as: (5) In formula (5), Indicates the number of communication protocol classes in the data communication process. Indicates that the communication protocol sample point belongs to The probability of the class is obtained by calculating the Gini index of the current data communication sample set, using data communication to summarize all the values ​​of each abnormal data information feature, calculating the Gini coefficient, and then finding the corresponding optimal segmentation features and values ​​to generate a decision tree for abnormal information identification.

[0027] Preprocessing, calculation and data information processing are performed through abnormal network large data sets to improve the communication capability of abnormal data information.

[0028] In the above embodiment, the classification and regression tree includes a classifier.

[0029] In a specific embodiment, the classifier is a classifier with decision-making capabilities, enabling the communication of network data information according to different data communication protocols, communication nodes, and communication types. In the above-mentioned embodiment, interpolation is an important method for approximating discrete functions. It can be used to estimate the approximate value of a function at other points based on the value of the function at a finite number of points. Interpolation is used to fill the gaps between pixels during image transformation.

[0030] In specific applications, the interpolation problem is formulated as follows: Assuming the interval Real-valued function on In this interval different points The value at is , requiring an estimate exist A certain point The basic idea is to find a function ,exist On the node with The function values ​​are the same (sometimes, even the first-order derivative values ​​are the same), use The value of as a function This method uses a pre-selected set of simple functions to parameters Function class Find the conditions that meet the Function , and As Here is called the interpolated function, is called the interpolation node. is called the interpolation function class, and the above equation is called the interpolation condition. The function that satisfies the above formula is called the interpolation function. It is called the interpolation remainder. When the minimum closed interval is , the corresponding interpolation is called interpolation; otherwise, it is called extrapolation. To avoid the large fluctuations that may occur with high-order interpolation, piecewise low-order interpolation is often used in practical applications to improve the degree of approximation. For example, piecewise linear interpolation or piecewise cubic Hermite interpolation can be used to approximate a known function, but their overall smoothness is poor. In other embodiments, a global piecewise interpolation method, cubic spline interpolation, can also be used to improve data communication sampling or computing power.

[0031] In a specific embodiment, the network perception module uses an STM32F103ZET6 microcontroller unit to implement perception calculation of communication data information.

[0032] The RF transceiver inside the sensing module uses the SX1278 chip, which supports multiple modulation methods such as LoRa, GFSK, FSK, and MSK. The output power is +20dBm when the voltage fluctuates within 100mW.

[0033] In a specific embodiment, the network perception module uses the Hi2115 chip in the radio frequency communication circuit, supports the CoAP protocol and UDP / TCP protocol, and realizes bidirectional data transmission from the serial port to the network through the Modbus protocol. The 8-channel data acquisition interface and RS485 bus interface reserved on the module hardware board are compatible with a variety of network sensors, greatly improving the sensor driving capability of the module. In order to reduce the impact of noise signals in the power communication network on the module's perception accuracy, anti-interference design is taken into account in the hardware design, and an isolation circuit is added to achieve bidirectional isolation and photoelectric isolation. In a specific embodiment, the network awareness module is further provided with an isolation circuit.

[0034] In a specific embodiment, the isolation circuit is a photoelectric isolator of the ADUM1250 chip.

[0035] In a specific embodiment, the optoelectronic isolator uses the LTV-816S-TA1-C chip, and the bidirectional isolation chip selects the ADUM1250. Optocouplers and isolated power supplies are used at both ends of the network perception module to achieve electrical isolation, while also eliminating the impact of certain acquisition channel noise on the module. The bidirectional isolation chip isolates the external and internal signals in the module, ensuring that the external signal is not directly connected to the main control unit, thereby ensuring a more stable processor and ensuring that the accuracy of the perception module's acquisition network parameters is not affected. This gives the module strong anti-interference capabilities and prevents interference between modules.

[0036] In a specific embodiment, the improved GAN algorithm module includes a communication feature extraction model, a communication data generation model and Communication data discrimination model , wherein the output end of the communication feature extraction model is connected to the input end of the communication data generation model, and the output end of the communication data generation model is connected to the input end of the communication data discrimination model.

[0037] In a specific embodiment, the working method of the communication feature extraction model is: The network data information features are extracted through a multi-layer perceptron. The output function of the communication feature extraction model is expressed as: (6) In formula (6), Represents the discriminant model in the GAN network, Represents the generation model in the GAN network, represents the true distribution of the input communication network parameters, represents the distribution of input noise data, represents the data sampling process of the model, represents the noise data in the communication network parameter data, Represents the data communication node in the GAN network.

[0038] Formula (1) is used to complete the game of generative adversarial networks. The generative model samples from real data, and the discriminative model learns based on the distribution pattern of real data. In a specific embodiment, the concept of generative adversarial networks (GANs) is introduced into the field of network fault detection and diagnosis. A fault diagnosis model is constructed based on GANs. A large number of data sets are obtained based on a small number of labeled data sets for training the fault diagnosis model. The generator and discriminator are used to generate and classify network communication parameter samples, respectively.

[0039] In a specific embodiment, the communication data generation model realizes the generation of communication data through a matrix function, and the method for the communication data generation model to realize data information generation is: The network fault diagnosis model first determines the characteristics corresponding to different communication network states, and then identifies the fault. The input vectors of different network states are , network communication data is represented as ,When a fault occurs in the power communication network, the network state information generation function during the fault period is: (7) In formula (7), Indicates Time The value of the network parameters, Indicates the fault duration, The overall network communication data information, The entirety of the input vector representing the state of the network, Indicates the number of input vectors; In a specific embodiment, a communication data generation model provides coded data during application, wherein the data information is divided into data equal to or less than the maximum data transfer unit (MTU) specified on the communication path, and the data is transmitted and the receiving device recovers and reproduces the data through this structure. During network data communication, the transmitting device generates data storage by dividing the coded data used as the transmission target into multiple packets of data equal to or less than the maximum data transfer unit (MTU) specified on the communication path and additional information used as arrangement information of the divided data, and sequentially transmits the packets, such as NAL unit fragments obtained by further dividing the NAL unit, which are stored in each packet, and transmits the packets. The receiving device arranges the NAL unit fragments stored in the packet in a divided manner with reference to the additional information of the packet and reconstructs and decodes the NAL unit.

[0040] In a specific embodiment, when generating data information, the data information being transmitted is encoded or arranged in a matrix to achieve data information generation, so as to improve the data information generation capability.

[0041] In a specific embodiment, the communication data discrimination model realizes the discrimination of the network data information communication process through a loss function, wherein the loss function is: (8) In formula (8), Represents the loss value output by the communication data discrimination model, Indicates the leaf node during communication diagnosis, Indicates the number of leaf nodes during communication diagnosis, represents the regularization parameter for communication diagnosis, The communication matrix function value representing different communication data information losses during communication diagnosis, Represents the structural parameters of the model.

[0042] The final loss value is calculated using formula (6). When implementing communication data judgment, discriminant analysis is also called linear discriminant analysis. It is a statistical method that uses samples of known categories to establish a discriminant model to discriminate samples of unknown categories. The purpose of discriminant analysis is to establish classification rules composed of numerical indicators for data of known categories, and then apply such rules to classify samples of unknown categories. The characteristics of discriminant analysis are to summarize the regularity of objective object classification based on the data information of several samples of each category that have been mastered in history, and to establish discriminant formulas and discriminant criteria. When performing discriminant analysis, it is usually necessary to divide the data into two parts. One part is the training model data, and the other part is the verification model data. In one embodiment, a model is first fitted by training the training set data. Then, the other part is used to verify the model effect. If it also performs well on the test set data, it means that the fitting model is very good. In discriminant analysis, according to the nature of the data, it is divided into discriminant analysis of qualitative data and discriminant analysis of quantitative data; different discriminant criteria are used, and there are discriminant methods such as Fisher, Bayesian, and distance. The principle of distance discrimination is to determine the distance between each sample and its parent population. This involves establishing a distance discrimination function for each parent population based on the data. Each sample's data is then substituted into the function to determine the distance between each sample and its parent population. The sample is then assigned to the parent population with the smallest distance. This data information function enables the diagnosis of abnormal information during data communication.

[0043] During the experiment, the CPU used was an AMD 2600 with a 3.40 GHz clock speed and 16 GB of RAM, the graphics card was a GTX 1070 Ti with 8 GB of video memory, the operating system was Windows 10, and the software version was TensorFlow 1.14. The system type was a three-cell experimental environment with a cell radius of 500 meters, a system bandwidth of 5 MHz, soft frequency reuse, a base station transmit power of 43 dBm, a number of users ranging from 50 to 100, and a simulation time of 60,000 TTLS. The data used in the experiment was flow statistical feature data for power communication network services, which includes a feature set of all available statistical features. In the experimental dataset, 230 features represent a single service data flow. The data types in the dataset include various service types, such as web browsing, email, file transfer, network status, database, and multimedia. The network communication platform was built on Linux, and the configuration parameters are shown in Table 1.

[0044] Table 1 Network communication platform configuration parameters Serial number project Configuration 1 CPU main frequency 3.9GHz 2 Memory 8GB 3 operating system Cent0S6 4 JDK JDK 1.8 5 Hadoop Hadoop 2.4.1 6 Hbase HBase 0.96.2 7 Virtual Machine VMware Workstation 6.5

[0045] The experimental data samples are shown in Table 2.

[0046] Table 2 Experimental data samples Serial number Data Name Data Representation Data Type Data volume 1 Real-time load collection SSCC Varchar 4572 2 Electric energy data EEN Varchar 6523 3 Statistics TTJ Varchar 4127 4 Power consumption status data YHJ Varchar 8536 5 Abnormal event information YCHQ Varchar 7465 6 Creation time CTT Datatime 4108

[0047] During data communication and data collection testing, the control center issues control commands, and the power terminal collection nodes respond to these commands, collect power data, and transmit the data back to the terminal nodes. To verify the system's long-term data collection stability, multiple nodes were used in the experimental environment for data collection, with a collection frequency of 100 times per hour and a test duration of 100 hours. Scheme 1 (CAN bus data communication) and Scheme 2 (Internet network data communication) were used for data communication. Based on the aforementioned data information, a comparative analysis was conducted, and 10,000 data collection attempts were performed. No system errors occurred before 40 hours of testing. After 40 hours, the number of errors gradually increased across all systems. Scheme 1 experienced errors around 50 hours, while Scheme 2 experienced an increasing number of errors around 44 hours. By the 100-hour mark, the number of errors for Scheme 1 and Scheme 2 was 16 and 13, respectively. This suggests that the system may be subject to interference from factors such as signal noise during data transmission, which may reduce signal strength at the collection nodes and cause packet loss, leading to an increase in the number of errors.

[0048] The system of the present invention only experiences erroneous collection after about 65 hours, with the highest number of erroneous collections being 6 times. The target error rate of the system is 0.006%. The data communication collection function of the system has good reliability and can maintain good data transmission effect in complex communication environments.

[0049] The system has high real-time requirements for business data. Data processing needs to meet the real-time requirements of the system. The experimental data in Table 2 are used for testing to compare the data processing time of Solution 1 and Solution 2. The system of the present invention takes the shortest time to process various types of data communication data, distributes data processing tasks to other nodes in the network communication platform, and finally returns and summarizes the processing results, which speeds up the data processing time and meets the real-time requirements of the system's business functions. The shortest processing time for data sample No. 6 is 7.85ms, and the longest processing time for data sample No. 4 is 48.8ms.

[0050] The processing time for each set of sample data using Scheme 1 was greater than 20ms, with a maximum processing time of 50.09ms and a minimum of 20.05ms. The minimum processing time for Scheme 2 was 24.43ms, and the processing time for data sample 5 was as high as 45.06ms. While the average processing time for Schemes 1 and 2 was higher than that of the system of the present invention, they still met the real-time requirements of the system. However, they required more system resources to process large amounts of data communications.

[0051] Although specific embodiments of the present invention have been described above, those skilled in the art will appreciate that these specific embodiments are merely illustrative, and that those skilled in the art may omit, substitute, and modify the above methods and details in various ways without departing from the principles and spirit of the present invention. For example, combining the above method steps to perform substantially the same functions in substantially the same manner to achieve substantially the same results falls within the scope of the present invention. Therefore, the scope of the present invention is limited solely by the appended claims.

Claims

1. A digital quantity acquisition and communication transmission method, characterized in that: Includes the following methods: It includes digital quantity acquisition steps and digital quantity communication transmission steps; wherein: The digital quantity acquisition method is: The data extraction module obtains data information, performs dimensionality reduction on the extracted data information, pre-processes the collected abnormal data information, and calculates the loss of data information through line loss fluctuations; The digital communication transmission method is: Multi-channel information transmission is performed on data information. During the information transmission process, data information is perceived through the network perception module, and fault diagnosis of data information communication anomalies is achieved through the improved GAN algorithm module; thereby realizing the integration of digital quantity acquisition and communication.

2. A digital quantity acquisition and communication transmission method according to claim 1, characterized in that: The data extraction module realizes the control of data information extraction through the CC2530 chip.

3. The digital quantity acquisition and communication transmission method according to claim 1, characterized in that: The data information is transmitted through the ZigBee communication module, wherein the ZigBee communication module is provided with an MSP430F249 single-chip microcomputer control unit.

4. The digital quantity acquisition and communication transmission method according to claim 1, characterized in that: The method for abnormal data information analysis is: The collected network data communication data sequence is divided into multiple segments of equal length, and the average value of each segment is calculated. The original communication data sequence is recorded as , the PAA algorithm is used to reduce the dimension of communication data information. The sequence after dimension reduction is expressed as: （1） In formula (1), Indicates the length of the collected original data information sequence, represents the length of the communication data information sequence after dimensionality reduction, Represents the average value of the collected original data information sequence, It represents the fragment of the original data sequence collected. After being processed by the PAA algorithm, the length of the original data communication is reduced from down to ; Preprocess the abnormal network large data set and use the interpolation method to deal with the missing values ​​in the network large data set. The missing value function is expressed as: （2） In formula (2), The serial number representing the missing value, represents the ordinal number of the non-missing value, Represents the interpolation cardinality. For the periodic analysis of the transmission of abnormal large data sets on the network, network communication has large fluctuations. The continuous fluctuation value of the communication protocol in the data channel has continuously decreased to a stable state. The network channel communication function is expressed as: （3） In formula (3), Indicates the statistical period of data communication, Indicates the year-on-year fluctuation value of the network communication protocol. Indicates the average year-on-year fluctuation value within the data communication statistical period. Indicates the month-on-month fluctuation value of the data communication environment. Indicates the average month-on-month fluctuation value within the data communication statistical period; The continuous fluctuation value function of communication line loss is expressed as: （4） In formula (4), is the continuous fluctuation value of line loss, is the line loss rate of the previous communication line statistical cycle, Line loss rate for the next communication line statistical cycle; A classification and regression tree is used to establish an abnormal data communication identification model, and the feature space is divided into two categories to distinguish normal network data communication from abnormal network data communication. The Gini coefficient function of the probability distribution is expressed as: （5） In formula (5), Indicates the number of communication protocol classes in the data communication process. Indicates that the communication protocol sample point belongs to The probability of the class is obtained by calculating the Gini index of the current data communication sample set, using data communication to summarize all the values ​​of each abnormal data information feature, calculating the Gini coefficient, and then finding the corresponding optimal segmentation features and values ​​to generate a decision tree for abnormal information identification.

5. A digital quantity acquisition and communication transmission method according to claim 4, characterized in that: Classification and regression trees include classifiers.

6. The digital quantity acquisition and communication transmission method according to claim 1, characterized in that: The network perception module uses the STM32F103ZET6 microcontroller unit to realize the perception and calculation of communication data information.

7. A digital quantity acquisition and communication transmission method according to claim 6, characterized in that: The network perception module is also provided with an isolation circuit.

8. The digital quantity acquisition and communication transmission method according to claim 7, characterized in that: The isolation circuit is a photoelectric isolator of the ADUM1250 chip.

9. The digital quantity acquisition and communication transmission method according to claim 1, characterized in that: The improved GAN algorithm module includes a communication feature extraction model, a communication data generation model and a communication data discrimination model, wherein the output end of the communication feature extraction model is connected to the input end of the communication data generation model, and the output end of the communication data generation model is connected to the input end of the communication data discrimination model.

10. The digital quantity acquisition and communication transmission method according to claim 9, characterized in that: The communication feature extraction model works as follows: The network data information features are extracted through a multi-layer perceptron. The output function of the communication feature extraction model is expressed as: （6） In formula (6), Represents the discriminant model in the GAN network, Represents the generation model in the GAN network, represents the true distribution of the input communication network parameters, represents the distribution of input noise data, represents the data sampling process of the model, represents the noise data in the communication network parameter data, Represents the data communication node in the GAN network; The communication data generation model realizes the generation of communication data through matrix functions. The method for the communication data generation model to realize data information generation is: The input vectors for different network states are , network communication data is represented as ,When a fault occurs in the power communication network, the network state information generation function during the fault period is: （7） In formula (7), Indicates Time The values ​​of the network parameters, Indicates the fault duration, The overall network communication data information, The entirety of the input vector representing the state of the network, Indicates the number of input vectors; The communication data discrimination model realizes the discrimination of the network data information communication process through the loss function, where the loss function is: （8） In formula (8), Represents the loss value output by the communication data discrimination model, Indicates the leaf node during communication diagnosis, Indicates the number of leaf nodes during communication diagnosis, represents the regularization parameter for communication diagnosis, The communication matrix function value representing different communication data information losses during communication diagnosis, Represents the structural parameters of the model.

Images