Water source industrial water flow data acquisition system in network security mode

Abstract

The invention discloses a water source industrial water flow data acquisition system in a network security mode, and belongs to the technical field of data processing, and the system comprises a data acquisition module which collects water source industrial water flow data in real time and carries out the preprocessing; the data analysis module is used for analyzing the preprocessed industrial water flow data and evaluating and judging whether the real-time flow of the industrial water in the water source area is abnormal or not based on a pre-constructed flow evaluation model; the abnormity analysis module is used for acquiring real-time flow data of the abnormal industrial water of the water source, performing accumulative analysis on the real-time flow data, further determining an abnormal influence factor and an abnormal type, and sending abnormal type data to the encryption transmission module; and the encryption transmission module is used for receiving the abnormal type data, automatically generating an encryption channel and encryption data, and uploading the encryption data to the SIS system. The problem that the redundant load of a system is too large during data transmission due to the fact that existing industrial water flow data acquisition equipment generally adopts remote transmission of water flow data and only has a data transmission function is solved.

Description

Technical Field

[0001] The invention belongs to the technical field of data processing, and in particular relates to an industrial water flow data acquisition system for a water source in a network security mode. Background Art

[0002] SIS, or power plant-level monitoring information system, is based on a real-time database, aims at economical operation and improving the overall benefits of power generation enterprises, and uses advanced computer technology and appropriate professional calculation methods to achieve information integration and sharing within the entire power plant. Specifically, the main functions of SIS in power plants include production process monitoring, performance calculation, equipment life management, vibration monitoring and fault diagnosis, plant-wide load distribution, and operation guidance.

[0003] In the process of thermal power generation / heating, a large amount of industrial water needs to be collected from water sources. The water sources of thermal power plants usually include groundwater, surface water, etc. The water quality of these water sources directly affects the normal operation and environmental protection of thermal power plants. In order to ensure the normal operation of power generation facilities / pump rooms, real-time monitoring of water flow at water sources is required.

[0004] At present, industrial water flow data acquisition equipment is generally used to upload the real-time collected data to the power plant SIS system in real time, so as to facilitate the SIS system to analyze the water flow data. However, the existing industrial water flow data acquisition equipment generally adopts remote transmission of water flow data and only has data transmission function. It is unable to analyze and pre-process the water flow data, resulting in excessive system redundant load during data transmission, affecting the data transmission efficiency.

[0005] Therefore, an industrial water flow data acquisition system for water sources under network security mode is proposed. Summary of the invention

[0006] The purpose of the present invention is to address the deficiencies of the prior art and provide an industrial water flow data acquisition system for a water source under a network security mode, which solves the problem that the existing industrial water flow data acquisition equipment generally adopts remote transmission of water flow data, only has data transmission function, and cannot analyze and preprocess the water flow data, resulting in excessive system redundant load during data transmission. The existing industrial water flow data acquisition equipment generally adopts remote transmission of water flow data, only has data transmission function, and cannot analyze and preprocess the water flow data, resulting in excessive system redundant load during data transmission, affecting data transmission efficiency. In order to solve the above problems, we propose an industrial water flow data acquisition system for a water source under a network security mode. In short, the system includes a data acquisition module, a data analysis module, an abnormal analysis module and an encrypted transmission module. The encrypted transmission module is electrically connected to the data analysis module and the abnormal analysis module respectively, and the data analysis module and the abnormal analysis module are both electrically connected to the data acquisition module.

[0007] The present invention provides a water source industrial water flow data acquisition system in a network security mode, comprising:

[0008] Data acquisition module: used to collect industrial water flow data from water sources in real time and pre-process multi-source heterogeneous industrial water flow data;

[0009] Data analysis module: used to analyze the pre-processed industrial water flow data, and to determine whether there is any abnormality in the real-time flow of industrial water at the water source based on the pre-built flow assessment model;

[0010] Abnormal analysis module: used to obtain the real-time flow data of industrial water in the water source with abnormalities and perform cumulative analysis, determine the abnormal influencing factors based on the cumulative analysis results, and then determine the abnormal type, and send the abnormal type data to the encrypted transmission module;

[0011] Encrypted transmission module: used to receive abnormal type data, self-generate encrypted channels and encrypted data, and upload the encrypted data to the SIS system.

[0012] The present invention provides a water source industrial water flow data acquisition system in a network security mode, and a data acquisition module, comprising:

[0013] Data collection unit: collects time-series multi-source heterogeneous industrial water flow data based on a preset collection method, wherein the industrial water flow data includes water flow information, water quality index information, microbial index information, and organic pollutant index information;

[0014] Preprocessing unit: normalize and preprocess the time-series multi-source heterogeneous industrial water flow data;

[0015] Signal processor: obtains pre-processed industrial water flow data, and performs filtering and amplification processing on the industrial water flow data.

[0016] The present invention provides a water source industrial water flow data acquisition system in a network security mode, and a method for normalizing and preprocessing time-series multi-source heterogeneous industrial water flow data, specifically comprising:

[0017] Loading time-series multi-source heterogeneous industrial water flow data;

[0018] Analyze the meaning of industrial water flow data, determine the variable type of industrial water flow data, and process missing values ​​of industrial water flow data; the variable type of industrial water flow data includes quantitative variables and qualitative variables;

[0019] Obtain the industrial water flow data after missing value processing, and normalize the industrial water flow data. The normalization method is to introduce dummy variables to factorize qualitative variables, and normalize and standardize quantitative variables.

[0020] Load the normalized industrial water flow data, determine the data source types of the multiple groups of industrial water flow data, generate data source index nodes based on the industrial water flow data of the same data source type, generate operation logic indexes based on the data source index nodes, integrate the multiple groups of data source index nodes, and generate index node trees based on the multiple groups of data source index nodes.

[0021] The present invention provides a water source industrial water flow data acquisition system in a network security mode, and the data analysis module includes:

[0022] Data modeling unit: obtain the normalized industrial water flow data stored in the index node tree, and build and train a flow evaluation model based on the industrial water flow data;

[0023] Model processing unit: Analyze the pre-processed industrial water flow data, and determine whether there is any abnormality in the industrial water flow at the water source based on the pre-built flow assessment model.

[0024] The present invention provides a water source industrial water flow data acquisition system in a network security mode, and a method for constructing a flow evaluation model, comprising:

[0025] Obtain the normalized industrial water flow data stored in the index node tree, and divide the industrial water flow data into a training set and a test set, where the ratio of the training set to the test set is 7:3;

[0026] Load the training set, perform dimensionality reduction processing on the training set, and obtain the feature matrix after dimensionality reduction, wherein the principal component analysis method is used to select features with a cumulative contribution of not less than 80% when performing dimensionality reduction processing on the training set;

[0027] Performing clustering processing on the feature matrix after dimensionality reduction to determine multiple groups of combined evaluation predictors, wherein the combined evaluation predictors include at least one group of LSTM prediction models, GBRT prediction models, and SVR prediction models;

[0028] Training the plurality of groups of evaluation predictors respectively, and adjusting the hyperparameters of the evaluation predictors during the training process, wherein the hyperparameters are adjusted by trial and error or Bayesian optimization to output the optimal evaluation predictor;

[0029] Initialize the combined evaluation predictor again using the hyperparameters of the optimal evaluation predictor, calibrate the hyperparameters of the combined evaluation predictor based on the test set, and output the optimal evaluation predictor;

[0030] Based on the optimal evaluation predictor, the industrial water flow of the water source is predicted, the prediction results are denormalized, the prediction values ​​of the test period are obtained, and it is judged whether the prediction values ​​reach the ideal prediction interval. If so, the flow evaluation model is output.

[0031] The present invention provides a water source industrial water flow data acquisition system in a network security mode, and a method for evaluating and judging whether there is an abnormality in the water source industrial water flow based on a pre-built flow evaluation model, specifically comprising:

[0032] The model processing unit establishes a connection with the index node tree and obtains the normalized industrial water flow data from the index node tree;

[0033] The industrial water flow data is processed by dimension reduction to obtain the feature vector after dimension reduction. The industrial water flow data is (a 1 ,a 2 ,a 3 ,...,a n-1 ,a n );

[0034] Among them, H(a n ) represents the eigenvector of industrial water flow data after dimensionality reduction, K is the data type of industrial water flow data, For industrial water flow data conversion information, a n Enter information for industrial water flow data,

[0035] Among them, J(a n ) is the feature recognition value of the feature recognition function:

[0036] Among them, J(a n ) is the feature recognition value of the feature recognition function, b is the feature weight, a n Enter information for industrial water flow data,

[0037] Among them, f(a) is the feature mapping function:

[0038] Among them, ω is the characteristic decomposition coefficient, C is the characteristic mapping constant, and D is the number of external influencing characteristic factors when converting industrial water flow data;

[0039] The predicted value of industrial water flow at the water source is determined based on the combined prediction function and the eigenvector after dimensionality reduction of industrial water flow data:

[0040] Among them, λ is the transformation coefficient of the combined prediction function, and G is the data gradient converted by the traffic assessment model;

[0041] The predicted value of the industrial water flow at the water source is compared with the preset flow range. If the predicted value of the industrial water flow at the water source is not within the preset flow range, it is determined that the real-time flow of the industrial water at the water source is abnormal.

[0042] The present invention provides a water source industrial water flow data acquisition system in a network security mode, and an abnormality analysis module, comprising:

[0043] Data acquisition unit: acquires real-time flow-related data of industrial water at the water source within a preset time period when an abnormality exists;

[0044] Frequency determination unit: grouping the real-time flow-related data of industrial water at the water source within a preset time period according to preset time intervals of different lengths, and determining the long-term frequency and short-term frequency of abnormal data of the real-time flow-related data of industrial water:

[0045] Among them, f long is the long-term frequency of abnormal data in real-time traffic-related data, n is the total number of groups of real-time traffic-related data, m is the number of groups of real-time traffic-related data considered in the short term, and N i,j is the number of times the j-th type of anomaly occurs in the i-th group of time periods, w j is the preset weight of the j-th type of anomaly, α i is the duration weight corresponding to the i-th group of time periods, α g is the duration weight corresponding to the i-th group of time periods, k is the total number of abnormal types, E(t) is the environmental impact function of the abnormal frequency, and f short is the short-term frequency of abnormal data in real-time traffic-related data, N g,j is the number of times the j-th type of anomaly occurs in the g-th group of time periods, g = n-m + 1 is the m-th group of data selected from the n-th group of real-time traffic-related data, T i is the length of the i-th time period, T g is the length of the g-th group of time periods;

[0046] Factor determination unit: Determine the abnormal impact factor based on the long-term abnormal frequency and short-term abnormal frequency of the real-time flow-related data of industrial water:

[0047] Wherein, I is the abnormal impact factor determined based on the real-time flow-related data of industrial water;

[0048] Type determination unit: compares the abnormal impact factor with the corresponding preset range, and then determines the abnormal type based on the preset range-abnormal type data table, and sends the abnormal type and related real-time traffic data to the encrypted transmission module.

[0049] The present invention provides a water source industrial water flow data acquisition system in a network security mode, which receives abnormal type data, self-generates an encrypted channel and encrypted data, and uploads the encrypted data to a SIS system, specifically including:

[0050] The encryption transmission module receives the abnormal type and related real-time flow data, and the encryption transmission module establishes multiple groups of encrypted communication channels with the SIS system, wherein the encrypted communication channels include a group of signaling channels and data encryption channels;

[0051] Encrypt and encapsulate the abnormal type and related real-time traffic data through an encrypted data packet to obtain an encrypted encapsulation package; obtain the SIS system key and the encryption index table, perform similarity detection on the feature value of the encrypted encapsulation package in the encryption index table to obtain an encrypted hash, and use the SIS system key to encrypt the encrypted hash of the encrypted encapsulation package to generate an encrypted transmission ciphertext;

[0052] Load the encrypted transmission ciphertext, insert the encrypted transmission ciphertext into the encrypted encapsulation package, and upload it to the SIS system through the encrypted communication channel.

[0053] The water source industrial water flow data acquisition system under network security mode provided by the present invention solves the problem that the existing industrial water flow data acquisition equipment generally adopts remote transmission of water flow data, has only data transmission function, cannot analyze and pre-process the water flow data, and causes excessive system redundant load during data transmission by analyzing and processing the industrial water flow data of the water source. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the present invention or 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0055] Figure 1 It is a structural schematic diagram of a water source industrial water flow data acquisition system in a network security mode provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0057] Example 1

[0058] like Figure 1 As shown, the water source industrial water flow data acquisition system in network security mode provided by the embodiment of the present invention includes:

[0059] Data acquisition module: used to collect industrial water flow data from water sources in real time and pre-process multi-source heterogeneous industrial water flow data;

[0060] Data analysis module: used to analyze the pre-processed industrial water flow data, and to determine whether there is any abnormality in the real-time flow of industrial water at the water source based on the pre-built flow assessment model;

[0061] Abnormal analysis module: used to obtain the real-time flow data of industrial water in the water source with abnormalities and perform cumulative analysis, determine the abnormal influencing factors based on the cumulative analysis results, and then determine the abnormal type, and send the abnormal type data to the encrypted transmission module;

[0062] Encrypted transmission module: used to receive abnormal type data, self-generate encrypted channels and encrypted data, and upload the encrypted data to the SIS system.

[0063] In this embodiment, the data acquisition module includes at least one group of sampling sensors, and the sampling sensors are used to periodically collect time-series multi-source heterogeneous industrial water flow data, wherein the industrial water flow data includes water flow information, water quality index information, microbial index information, and organic pollutant index information;

[0064] In this embodiment, the water flow information includes but is not limited to the instantaneous flow and cumulative flow of the water flow in the water source, and the cumulative flow is obtained in cycles of days, weeks, months, quarters and years, and the water quality index information includes but is not limited to total phosphorus, chemical oxygen demand and permanganate, pH value, dissolved oxygen, ammonia nitrogen, and heavy metal content in the water source, and the sampling sensor can be a pH sensor, an oxidation-reduction potential sensor, a dissolved oxygen sensor, an ammonia nitrogen sensor, a turbidity sensor, and a conductivity sensor, and each group of the sampling sensors is configured with a unique identification code, and the unique identification code corresponds to the number, location, and encrypted channel information of the sampling sensor;

[0065] In this embodiment, there is a data analysis module and an encryption transmission module. The data analysis module can analyze the pre-processed industrial water flow data, and judge whether there is any abnormality in the industrial water flow at the water source based on the pre-built flow assessment model. The data analysis module can also construct a flow assessment model with a combination of multiple neural networks, and analyze the predicted value of the industrial water flow at the water source based on the industrial water flow data, so as to provide prediction support for the pump room scheduling of the power plant generator set. At the same time, the encryption transmission module can encrypt and transmit abnormal data, and the collection results can be fed back to the SIS system without transmitting a large amount of data, thereby ensuring the security of data transmission and reducing the load and redundancy of the SIS system.

[0066] In this embodiment, the abnormal influencing factors refer to various reasons or factors that cause the abnormal industrial water flow data of the water source to appear. By analyzing the abnormal water flow data, it is possible to identify which factors cause the abnormality;

[0067] In this embodiment, the abnormal type is to classify and identify the detected abnormal phenomenon. According to the characteristics and influencing factors of the abnormality, the abnormal phenomenon can be classified into different types so as to take targeted measures.

[0068] The beneficial effect of the above technical solution is: by analyzing and processing the industrial water flow data of the water source, it solves the problem that the existing industrial water flow data acquisition equipment generally adopts remote transmission of water flow data, has only data transmission function, and cannot analyze and pre-process the water flow data, resulting in excessive system redundant load during data transmission.

[0069] Example 2

[0070] The data acquisition module of the water source industrial water flow data acquisition system in a network security mode provided by the embodiment of the present invention includes:

[0071] Data collection unit: collects time-series multi-source heterogeneous industrial water flow data based on a preset collection method, wherein the industrial water flow data includes water flow information, water quality index information, microbial index information, and organic pollutant index information;

[0072] Preprocessing unit: normalize and preprocess the time-series multi-source heterogeneous industrial water flow data;

[0073] Signal processor: obtains pre-processed industrial water flow data, and performs filtering and amplification processing on the industrial water flow data.

[0074] In this embodiment, the preprocessing unit performs normalization preprocessing on the time-series multi-source heterogeneous industrial water flow data;

[0075] In this embodiment, the preset collection method is periodic collection;

[0076] In this embodiment, the signal processor is used to obtain the pre-processed industrial water flow data, and filter and amplify the industrial water flow data;

[0077] In this embodiment, the sampling sensor and the preprocessing unit are connected by Bluetooth or DTU communication to realize interactive data transmission, and the signal processor is electrically connected to the preprocessing unit, and the signal processor filters and amplifies the industrial water flow data and transmits the data to the data analysis module;

[0078] In this embodiment, the preprocessing unit can perform normalization preprocessing on time-series multi-source heterogeneous industrial water flow data, and the preprocessing unit processes the data by means of meaning analysis, normalization processing, and generating an index node tree, which can facilitate modeling and data analysis of the data analysis module, and the generated index node tree can store the collected data and ensure that the collected data does not need to be transmitted in full, thereby achieving the purpose of reducing the system load.

[0079] The beneficial effect of the above technical solution is that the data acquisition unit can obtain multi-source heterogeneous industrial water flow data in real time, including water flow, water quality indicators, microbial indicators and organic pollutant indicator information. The preprocessing unit normalizes these data, while the signal processor filters and amplifies them to ensure data quality and accuracy. These functions together ensure that the system can efficiently and accurately collect and process industrial water flow data at the water source, and improve water resource management and environmental protection effects.

[0080] Example 3

[0081] The method for normalizing and preprocessing the time-series multi-source heterogeneous industrial water flow data in the water source industrial water flow data collection system in a network security mode provided by the embodiment of the present invention specifically includes:

[0082] Loading time-series multi-source heterogeneous industrial water flow data;

[0083] Analyze the meaning of industrial water flow data, determine the variable type of industrial water flow data, and process missing values ​​of industrial water flow data; the variable type of industrial water flow data includes quantitative variables and qualitative variables;

[0084] Obtain the industrial water flow data after missing value processing, and normalize the industrial water flow data. The normalization method is to introduce dummy variables to factorize qualitative variables, and normalize and standardize quantitative variables.

[0085] Load the normalized industrial water flow data, determine the data source types of the multiple groups of industrial water flow data, generate data source index nodes based on the industrial water flow data of the same data source type, generate operation logic indexes based on the data source index nodes, integrate the multiple groups of data source index nodes, and generate index node trees based on the multiple groups of data source index nodes.

[0086] In this embodiment, the missing value processing of industrial water flow data can be carried out by taking the mode or random forest regression.

[0087] In this embodiment, the index node tree can be understood as a knowledge graph or dynamic database containing multiple groups of index nodes, and the index node tree is stored in the preprocessing unit 120, so that the collected raw data is preprocessed and distributedly stored in the water source industrial water flow data collection system, which ensures the security of the data and the smoothness of data processing;

[0088] The beneficial effects of the above technical solution are: by normalizing and preprocessing the time-series multi-source heterogeneous industrial water flow data, the consistency and comparability of the data are improved. The missing values ​​are processed and the quantitative variables and qualitative variables are normalized, standardized and factorized respectively, making the subsequent analysis more accurate and reliable. By generating data source index nodes and index node trees and integrating multiple groups of data sources, the data management and retrieval efficiency is improved, and the performance and stability of the system in network security mode are enhanced.

[0089] Example 4

[0090] The data analysis module of the water source industrial water flow data acquisition system in a network security mode provided by the embodiment of the present invention includes:

[0091] Data modeling unit: obtain the normalized industrial water flow data stored in the index node tree, and build and train a flow evaluation model based on the industrial water flow data;

[0092] Model processing unit: Analyze the pre-processed industrial water flow data, and determine whether there is any abnormality in the industrial water flow at the water source based on the pre-built flow assessment model.

[0093] The pre-built flow assessment model is a mathematical or statistical model used to analyze and predict industrial water flow behavior. It processes and learns from a large amount of historical water flow data to understand the regularity and abnormal characteristics of these data. It contains the following parts: Input features: The input features include normalized industrial water flow data extracted from the index node tree. The data may cover multiple dimensions, such as time, location, flow value, etc. Data preprocessing: Before building the model, the data is cleaned and preprocessed, including handling missing values, normalization, standardization, etc. This step ensures the consistency of the data and the effectiveness of model training. Model type: Statistical models: Time series analysis models (such as ARIMA, SARIMA) are used to capture trends and periodicity in flow data. Machine learning models: Regression models, support vector machines, neural networks, etc. are used to learn complex patterns in data. Deep learning models: Long short-term memory networks (LSTM), convolutional neural networks (CNN), etc. are particularly suitable for processing time series data and detecting abnormal behavior. Training process: The model is trained using historical data to enable it to recognize normal and abnormal flow patterns. The training process includes the adjustment of model parameters to minimize the prediction error. Model evaluation: After training is completed, the accuracy and robustness of the model are evaluated using a test data set. Common evaluation metrics include mean square error (MSE), mean absolute error (MAE), etc. Anomaly detection: When the model is running in real time, it uses current water flow data as input to predict and evaluate whether there are anomalies. Anomalies may manifest as drastic changes in flow, fluctuations that do not conform to historical patterns, etc.

[0094] The beneficial effect of the above technical solution is: the industrial water flow data is acquired and processed by the data modeling unit, and the flow evaluation model is constructed and trained. The model processing unit uses the pre-built model evaluation to determine whether there is an abnormality in the industrial water flow at the water source. This method improves the accuracy and efficiency of anomaly detection and improves the safety and reliability of industrial water flow management at the water source.

[0095] Example 5

[0096] The embodiment of the present invention provides a water source industrial water flow data acquisition system in a network security mode, and a method for constructing a flow evaluation model, including:

[0097] Obtain the normalized industrial water flow data stored in the index node tree, and divide the industrial water flow data into a training set and a test set, where the ratio of the training set to the test set is 7:3;

[0098] Load the training set, perform dimensionality reduction processing on the training set, and obtain the feature matrix after dimensionality reduction, wherein the principal component analysis method is used to select features with a cumulative contribution of not less than 80% when performing dimensionality reduction processing on the training set;

[0099] Performing clustering processing on the feature matrix after dimensionality reduction to determine multiple groups of combined evaluation predictors, wherein the combined evaluation predictors include at least one group of LSTM prediction models, GBRT prediction models, and SVR prediction models;

[0100] Training the plurality of groups of evaluation predictors respectively, and adjusting the hyperparameters of the evaluation predictors during the training process, wherein the hyperparameters are adjusted by trial and error or Bayesian optimization to output the optimal evaluation predictor;

[0101] Initialize the combined evaluation predictor again using the hyperparameters of the optimal evaluation predictor, calibrate the hyperparameters of the combined evaluation predictor based on the test set, and output the optimal evaluation predictor;

[0102] Based on the optimal evaluation predictor, the industrial water flow of the water source is predicted, the prediction results are denormalized, the prediction values ​​of the test period are obtained, and it is judged whether the prediction values ​​reach the ideal prediction interval. If so, the flow evaluation model is output.

[0103] In this embodiment, when verifying the generalization ability of the model, the training set and the test set are shuffled, and the model is verified based on cross-validation training, so that the error of the trained model is smaller, thereby improving the evaluation accuracy of the model.

[0104] In this embodiment, the combined evaluation predictor includes three groups of combination layers, one group of attention mechanism layers, two groups of fully connected layers and one group of downsampling layers, and the combination layer includes an LSTM layer, a GBRT layer, and an SVR layer. The LSTM layer, the GBRT layer, and the SVR layer contain 32, 64, and 128 LSTM units, GBRT units, and SVR units, respectively.

[0105] In this embodiment, the ideal prediction interval is 0.85-0.9. If it is reached, the flow evaluation model is output.

[0106] The beneficial effect of the above technical solution is: clustering the feature matrix after dimensionality reduction, and then determining multiple groups of combined evaluation predictors, and the combined evaluation predictor is composed of LSTM prediction model, GBRT prediction model, and SVR prediction model, so that the constructed model can significantly reduce boundary effects and overfitting phenomena, and at the same time improve the generalization ability of the model, so that the flow evaluation model can not only be used to evaluate water flow, but also can be used to judge and extract abnormal influencing factors of industrial water flow data, improve the adaptability of the model, and provide auxiliary decision-making support for water intake and use in power plants.

[0107] Example 6

[0108] The method of the water source industrial water flow data acquisition system in a network security mode provided by the embodiment of the present invention, which evaluates and determines whether the water source industrial water flow is abnormal based on a pre-built flow evaluation model, specifically includes:

[0109] The model processing unit establishes a connection with the index node tree and obtains the normalized industrial water flow data from the index node tree;

[0110] The industrial water flow data is processed by dimension reduction to obtain the feature vector after dimension reduction. The industrial water flow data is (a 1 ,a 2 ,a 3 ,...,a n-1 ,a n );

[0111] Among them, H(a n ) represents the eigenvector of industrial water flow data after dimensionality reduction, K is the data type of industrial water flow data, For industrial water flow data conversion information, a n Enter information for industrial water flow data,

[0112] Among them, J(a n ) is the feature recognition value of the feature recognition function:

[0113] Among them, J(a n ) is the feature recognition value of the feature recognition function, b is the feature weight, a n Enter information for industrial water flow data,

[0114] Among them, f(a) is the feature mapping function:

[0115] Among them, ω is the characteristic decomposition coefficient, C is the characteristic mapping constant, and D is the number of external influencing characteristic factors when converting industrial water flow data;

[0116] The predicted value of industrial water flow at the water source is determined based on the combined prediction function and the eigenvector after dimensionality reduction of industrial water flow data:

[0117] Among them, λ is the transformation coefficient of the combined prediction function, and G is the data gradient converted by the traffic assessment model;

[0118] The predicted value of the industrial water flow at the water source is compared with the preset flow range. If the predicted value of the industrial water flow at the water source is not within the preset flow range, it is determined that the real-time flow of the industrial water at the water source is abnormal.

[0119] In this embodiment, when the model processing unit establishes a connection with the index node tree, the model processing unit first configures the index node tree database, downloads and installs the index node tree installation package, and configures the installation port number, installation directory, and data storage directory. After the installation is complete, JDBC is used to connect to the database, and JDBC-API is used to write code to establish a connection, send SQL statements, and process results.

[0120] In this embodiment, the number of external influence characteristic factors is 10-30.

[0121] In this embodiment, the industrial water flow data conversion information refers to the intermediate data used for feature recognition and mapping obtained after data preprocessing and dimensionality reduction. These data play a key role in the feature recognition and mapping process;

[0122] In this embodiment, the industrial water flow data input information refers to the original industrial water flow data directly obtained from the data acquisition module, including the original data without any preprocessing. These data are the basis for subsequent processing and analysis, including: Water flow information: water flow data directly read from the sensor, such as the amount of water flowing per minute. Water quality index information: real-time monitored water quality parameter data, such as pH value, turbidity, dissolved oxygen content, etc. Microbial index information: data on the content of microorganisms in water, such as the number of E. coli. Organic pollutant index information: concentration data of organic pollutants in water, such as the content of harmful substances such as benzene and toluene.

[0123] The beneficial effects of the above technical solution are: by establishing a connection with the index node tree, acquiring and reducing the dimension of industrial water flow data, obtaining a feature vector, and then using feature recognition and mapping functions to determine the feature weight and decomposition coefficient. Based on the combined prediction function and the reduced dimension feature vector, the predicted value of the industrial water flow at the water source is calculated, and compared with the preset range to detect abnormal situations in a timely manner. This method improves the accuracy and timeliness of anomaly detection and enhances the safety and reliability of industrial water flow management at the water source.

[0124] Example 7

[0125] The water source industrial water flow data acquisition system in a network security mode provided by an embodiment of the present invention includes an abnormality analysis module, which includes:

[0126] Data acquisition unit: acquires real-time flow-related data of industrial water at the water source within a preset time period when an abnormality exists;

[0127] Frequency determination unit: grouping the real-time flow-related data of industrial water at the water source within a preset time period according to preset time intervals of different lengths, and determining the long-term frequency and short-term frequency of abnormal data of the real-time flow-related data of industrial water:

[0128] Among them, f long is the long-term frequency of abnormal data in real-time traffic-related data, n is the total number of groups of real-time traffic-related data, m is the number of groups of real-time traffic-related data considered in the short term, and N i,j is the number of times the j-th type of anomaly occurs in the i-th group of time periods, w j is the preset weight of the j-th type of anomaly, α i is the duration weight corresponding to the i-th group of time periods, α g is the duration weight corresponding to the i-th group of time periods, k is the total number of abnormal types, E(t) is the environmental impact function of the abnormal frequency, and f shirt is the short-term frequency of abnormal data in real-time traffic-related data, N g,j is the number of times the j-th type of anomaly occurs in the g-th group of time periods, g = n-m + 1 is the m-th group of data selected from the n-th group of real-time traffic-related data, T i is the length of the i-th time period, T g is the length of the g-th group of time periods;

[0129] Factor determination unit: Determine the abnormal impact factor based on the long-term abnormal frequency and short-term abnormal frequency of the real-time flow-related data of industrial water:

[0130] Wherein, I is the abnormal impact factor determined based on the real-time flow-related data of industrial water;

[0131] Type determination unit: compares the abnormal impact factor with the corresponding preset range, and then determines the abnormal type based on the preset range-abnormal type data table, and sends the abnormal type and related real-time traffic data to the encrypted transmission module.

[0132] In this embodiment, the real-time traffic-related data includes the time, location, and specific traffic value of the abnormality.

[0133] In this embodiment, the preset different lengths are divided into long durations and short durations.

[0134] In this embodiment, g=n-m+1 is to select m groups of data in the most recent time from n groups of real-time traffic-related data. For example, if the total time period is 24 hours, there are 24 groups per hour, and m is the number of groups in the short term considered. For example, if the data of the most recent 6 hours is considered, then m=6;

[0135] In this embodiment, the sum in the formula This means that the data from group n-m+1 to group n are summed, that is, the sum of the abnormal frequencies of the most recent m groups is calculated, and then divided by m to obtain the average abnormal frequency during this period;

[0136] In this embodiment, the long-term frequency of abnormal data in the real-time traffic-related data is the average abnormal frequency of multiple time groups, and the short-term frequency may be the average abnormal frequency in a recent period of time.

[0137] In this embodiment, the preset range-abnormal type data table includes the following columns: abnormal impact factor range: represents different numerical intervals of abnormal impact factors, abnormal type: represents the description of the abnormal type corresponding to a specific impact factor range.

[0138] The beneficial effect of the above technical solution is: through the steps of data acquisition, frequency determination, factor determination and type determination, the industrial water flow data of the water source with abnormalities is analyzed. By calculating the long-term and short-term abnormal frequencies, the abnormal influencing factors are determined, and the abnormal types are determined by comparing with the preset range. This method improves the accuracy and efficiency of abnormal analysis, can timely discover and identify the abnormal types of industrial water flow in water sources, and provides valuable reference information.

[0139] Example 8

[0140] The water source industrial water flow data acquisition system in a network security mode provided by the embodiment of the present invention receives abnormal type data, generates an encrypted channel and encrypted data by itself, and uploads the encrypted data to the SIS system, specifically including:

[0141] The encryption transmission module receives the abnormal type and related real-time flow data, and the encryption transmission module establishes multiple groups of encrypted communication channels with the SIS system, wherein the encrypted communication channels include a group of signaling channels and data encryption channels;

[0142] Encrypt and encapsulate the abnormal type and related real-time traffic data through an encrypted data packet to obtain an encrypted encapsulation package; obtain the SIS system key and the encryption index table, perform similarity detection on the feature value of the encrypted encapsulation package in the encryption index table to obtain an encrypted hash, and use the SIS system key to encrypt the encrypted hash of the encrypted encapsulation package to generate an encrypted transmission ciphertext;

[0143] Load the encrypted transmission ciphertext, insert the encrypted transmission ciphertext into the encrypted encapsulation package, and upload it to the SIS system through the encrypted communication channel.

[0144] In this embodiment, the encrypted communication channel refers to a secure communication channel established between two systems in network security mode to ensure the confidentiality and integrity of data transmission. For example, in Example 8, multiple sets of encrypted communication channels are established between the encrypted transmission module and the SIS system to ensure the secure transmission of abnormal types and related real-time flow data.

[0145] In this embodiment, the signaling channel refers to a channel in the encrypted communication channel that is specifically used to transmit signaling information. The signaling information includes information such as communication protocol, data format, encryption key, etc. For example, in embodiment 8, the encrypted communication channel includes a group of signaling channels for transmitting signaling information.

[0146] In this embodiment, the data encryption channel refers to a channel in the encrypted communication channel that is specifically used to transmit encrypted data. Encrypted data refers to data that has been encrypted to ensure the confidentiality and integrity of the data. For example, in Embodiment 8, the encrypted communication channel includes a data encryption channel for transmitting an encrypted encapsulation package.

[0147] In this embodiment, the SIS system key refers to the key used by the SIS system to encrypt and decrypt data. The key refers to the parameters used in the encryption and decryption algorithms to ensure the confidentiality and integrity of the data. For example, in Embodiment 8, the encryption transmission module uses the SIS system key to encrypt the encrypted hash of the encrypted package.

[0148] In this embodiment, the encryption index table refers to a table for storing index information of encrypted data. The index information includes information such as characteristic values ​​of the encrypted data and encrypted hashes. For example, in Embodiment 8, the encryption transmission module obtains the SIS system key and the encryption index table, and performs similarity detection on the characteristic values ​​of the encrypted encapsulation package in the encryption index table.

[0149] In this embodiment, the encrypted transmission ciphertext refers to data that has been encrypted to ensure the confidentiality and integrity of the data. For example, in Embodiment 8, the encrypted transmission module uses the SIS system key to encrypt the encrypted hash of the encrypted package to generate the encrypted transmission ciphertext.

[0150] The beneficial effect of the above technical solution is: by establishing multiple sets of encrypted communication channels, including signaling

[0151] The channel and data encryption channel ensure the confidentiality and integrity of data transmission. By encrypting the data packet and the encrypted index table, generating the encrypted hash and encrypted transmission ciphertext, the data security is further improved. Finally, the encrypted transmission ciphertext is loaded and inserted into the encrypted package, and uploaded to the SIS system through the encrypted communication channel. The system can effectively protect the security and integrity of industrial water flow data at the water source and ensure the reliability and confidentiality of data transmission.

Claims

1. The water source industrial water flow data acquisition system in network security mode is characterized by: include: Data acquisition module: used to collect industrial water flow data from water sources in real time and pre-process multi-source heterogeneous industrial water flow data; Data analysis module: used to analyze the pre-processed industrial water flow data, and to determine whether there is any abnormality in the real-time flow of industrial water at the water source based on the pre-built flow assessment model; Abnormal analysis module: used to obtain the real-time flow data of industrial water in the water source with abnormalities and perform cumulative analysis, determine the abnormal influencing factors based on the cumulative analysis results, and then determine the abnormal type, and send the abnormal type data to the encrypted transmission module; Encrypted transmission module: used to receive abnormal type data, self-generate encrypted channels and encrypted data, and upload the encrypted data to the SIS system.

2. The water source industrial water flow data acquisition system in network security mode as claimed in claim 1 is characterized in that: Data acquisition module, including: Data collection unit: collects time-series multi-source heterogeneous industrial water flow data based on a preset collection method, wherein the industrial water flow data includes water flow information, water quality index information, microbial index information, and organic pollutant index information; Preprocessing unit: normalize and preprocess the time-series multi-source heterogeneous industrial water flow data; Signal processor: obtains pre-processed industrial water flow data, and performs filtering and amplification processing on the industrial water flow data.

3. The water source industrial water flow data acquisition system in network security mode as claimed in claim 2 is characterized in that: The method for normalizing and preprocessing time-series multi-source heterogeneous industrial water flow data specifically includes: Loading time-series multi-source heterogeneous industrial water flow data; Analyze the meaning of industrial water flow data, determine the variable type of industrial water flow data, and process missing values ​​of industrial water flow data; the variable type of industrial water flow data includes quantitative variables and qualitative variables; Obtain the industrial water flow data after missing value processing, and normalize the industrial water flow data. The normalization method is to introduce dummy variables to factorize qualitative variables, and normalize and standardize quantitative variables. Load the normalized industrial water flow data, determine the data source types of the multiple groups of industrial water flow data, generate data source index nodes based on the industrial water flow data of the same data source type, generate operation logic indexes based on the data source index nodes, integrate the multiple groups of data source index nodes, and generate index node trees based on the multiple groups of data source index nodes.

4. The water source industrial water flow data acquisition system in network security mode as claimed in claim 1 is characterized in that: Data analysis modules include: Data modeling unit: obtain the normalized industrial water flow data stored in the index node tree, and build and train a flow evaluation model based on the industrial water flow data; Model processing unit: Analyze the pre-processed industrial water flow data, and determine whether there is any abnormality in the industrial water flow at the water source based on the pre-built flow assessment model.

5. The water source industrial water flow data acquisition system in network security mode as claimed in claim 4 is characterized in that: The method for constructing a traffic evaluation model includes: Obtain the normalized industrial water flow data stored in the index node tree, and divide the industrial water flow data into a training set and a test set, where the ratio of the training set to the test set is 7:3; Load the training set, perform dimensionality reduction processing on the training set, and obtain the feature matrix after dimensionality reduction, wherein the principal component analysis method is used to select features with a cumulative contribution of not less than 80% when performing dimensionality reduction processing on the training set; Performing clustering processing on the feature matrix after dimensionality reduction to determine multiple groups of combined evaluation predictors, wherein the combined evaluation predictors include at least one group of LSTM prediction models, GBRT prediction models, and SVR prediction models; Training the plurality of groups of evaluation predictors respectively, and adjusting the hyperparameters of the evaluation predictors during the training process, wherein the hyperparameters are adjusted by trial and error or Bayesian optimization to output the optimal evaluation predictor; Initialize the combined evaluation predictor again using the hyperparameters of the optimal evaluation predictor, calibrate the hyperparameters of the combined evaluation predictor based on the test set, and output the optimal evaluation predictor; Based on the optimal evaluation predictor, the industrial water flow of the water source is predicted, the prediction results are denormalized, the prediction values ​​of the test period are obtained, and it is judged whether the prediction values ​​reach the ideal prediction interval. If so, the flow evaluation model is output.

6. The water source industrial water flow data acquisition system in network security mode as claimed in claim 5 is characterized in that: The method of evaluating whether there is an abnormality in the industrial water flow at the water source based on the pre-built flow assessment model includes: The model processing unit establishes a connection with the index node tree and obtains the normalized industrial water flow data from the index node tree; The industrial water flow data is processed by dimension reduction to obtain the feature vector after dimension reduction. The industrial water flow data is (a1, a2, a3, ..., a n-1 ,a n ); Among them, H(a n ) represents the eigenvector of industrial water flow data after dimensionality reduction, K is the data type of industrial water flow data, For industrial water flow data conversion information, a n Enter information for industrial water flow data, Among them, J(a n ) is the feature recognition value of the feature recognition function: Among them, J(a n ) is the feature recognition value of the feature recognition function, b is the feature weight, a n Enter information for industrial water flow data, Among them, f(a) is the feature mapping function: Among them, ω is the characteristic decomposition coefficient, C is the characteristic mapping constant, and D is the number of external influencing characteristic factors when converting industrial water flow data; The predicted value of industrial water flow at the water source is determined based on the combined prediction function and the eigenvector after dimensionality reduction of industrial water flow data: Among them, λ is the transformation coefficient of the combined prediction function, and G is the data gradient converted by the traffic assessment model; The predicted value of the industrial water flow at the water source is compared with the preset flow range. If the predicted value of the industrial water flow at the water source is not within the preset flow range, it is determined that the real-time flow of the industrial water at the water source is abnormal.

7. The water source industrial water flow data acquisition system in network security mode as claimed in claim 1 is characterized in that: Abnormal analysis module, including: Data acquisition unit: acquires real-time flow-related data of industrial water at the water source within a preset time period when an abnormality exists; Frequency determination unit: grouping the real-time flow-related data of industrial water at the water source within a preset time period according to preset time intervals of different lengths, and determining the long-term frequency and short-term frequency of abnormal data of the real-time flow-related data of industrial water: Among them, f long is the long-term frequency of abnormal data in real-time traffic-related data, n is the total number of groups of real-time traffic-related data, m is the number of groups of real-time traffic-related data considered in the short term, and N i,j is the number of times the j-th type of anomaly occurs in the i-th group of time periods, w j is the preset weight of the j-th type of anomaly, α i is the duration weight corresponding to the i-th group of time periods, α g is the duration weight corresponding to the i-th group of time periods, k is the total number of abnormal types, E(t) is the environmental impact function of the abnormal frequency, and f short is the short-term frequency of abnormal data in real-time traffic-related data, N g,j is the number of times the j-th type of anomaly occurs in the g-th group of time periods, g = n-m + 1 is the m-th group of data selected from the n-th group of real-time traffic-related data, T i is the length of the i-th time period, T g is the length of the g-th group of time periods; Factor determination unit: Determine the abnormal impact factor based on the long-term abnormal frequency and short-term abnormal frequency of the real-time flow-related data of industrial water: Wherein, I is the abnormal impact factor determined based on the real-time flow-related data of industrial water; Type determination unit: compares the abnormal impact factor with the corresponding preset range, and then determines the abnormal type based on the preset range-abnormal type data table, and sends the abnormal type and related real-time traffic data to the encrypted transmission module.

8. The water source industrial water flow data acquisition system in network security mode as claimed in claim 1 is characterized in that: Receive abnormal data, self-generate encrypted channels and encrypt data, And upload the encrypted data to the SIS system, including: The encryption transmission module receives the abnormal type and related real-time flow data, and the encryption transmission module establishes multiple groups of encrypted communication channels with the SIS system, wherein the encrypted communication channels include a group of signaling channels and data encryption channels; Encrypt and encapsulate the abnormal type and related real-time traffic data through an encrypted data packet to obtain an encrypted encapsulation package; obtain the SIS system key and the encryption index table, perform similarity detection on the feature value of the encrypted encapsulation package in the encryption index table to obtain an encrypted hash, and use the SIS system key to encrypt the encrypted hash of the encrypted encapsulation package to generate an encrypted transmission ciphertext; Load the encrypted transmission ciphertext, insert the encrypted transmission ciphertext into the encrypted encapsulation package, and upload it to the SIS system through the encrypted communication channel.

Images