Intelligent internet of things big data platform management system and method based on visual modeling

Abstract

The invention discloses an intelligent Internet of Things big data platform management system and method based on visual modeling, and relates to the technical field of Internet of Things data management, and the method comprises the following steps: collecting and storing equipment basic information, equipment operation parameters, maintenance information and configuration information, and carrying out the preprocessing of the collected data; constructing a visual platform supporting multi-dimensional interaction based on the preprocessed data, and extracting features through a feature extraction algorithm; through user interaction frequency calculation, model precision verification and complexity evaluation, quantitatively analyzing the performance of the visual intelligent Internet of Things big data model, and quantifying interaction behavior characteristics, modeling precision and structural complexity; visual intelligent Internet of Things big data model parameters are adjusted through an optimization formula, equipment operation state monitoring is carried out through a time axis and a screening tool, and intelligent Internet of Things big data platform management is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things data management, and specifically to a management system and method for an intelligent Internet of Things big data platform based on visual modeling. Background Art

[0002] In the current era of rapid digital and intelligent development, the wide application of Internet of Things technology in the industrial field has led to the continuous generation of a large amount of device data, and the intelligent Internet of Things big data platform has emerged as the times require. In many industrial scenarios, the efficient management of devices has become a key factor for enterprises to enhance their competitiveness and ensure stable operation.

[0003] However, there are many deficiencies in the existing management means of intelligent Internet of Things big data platforms. First of all, in the face of device data with wide sources, diverse types and large scales, traditional methods are difficult to quickly and accurately integrate the basic information, operation parameters, maintenance records and configuration details of different devices, resulting in serious data fragmentation and unable to provide a complete and reliable data source for subsequent analysis. Secondly, the visual modeling and performance analysis capabilities are weak. Most platforms fail to fully consider the actual needs of enterprises when constructing visual models, resulting in the disconnection between the models and business scenarios and unable to provide users with an intuitive and practical device status display and analysis interface. In terms of model performance evaluation, it is neither possible to accurately quantify the interaction behavior characteristics between users and the visual interface, making it difficult to judge the usage efficiency of platform functions by users, nor to effectively verify the representation accuracy of the model for complex device data and the rationality of its own structure, making the model optimization lack a scientific basis and unable to meet the growing device management needs. Summary of the Invention

[0004] The purpose of the present invention is to provide a management system and method for an intelligent Internet of Things big data platform based on visual modeling to solve the problems raised in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A management method for an intelligent Internet of Things big data platform based on visual modeling, the method comprising the following steps:

[0006] S100. Collect and store the basic information of devices, device operation parameters, maintenance information and configuration information, and preprocess the collected data;

[0007] S200. Based on the preprocessed data, construct a visual platform supporting multi-dimensional interaction, and extract features through a feature extraction algorithm;

[0008] S300. Through user interaction frequency calculation, model accuracy verification and complexity evaluation, quantitatively analyze the performance of the visual intelligent Internet of Things big data model, and quantify the interaction behavior characteristics, modeling accuracy and structural complexity;

[0009] S400. Adjust the parameters of the visual intelligent IoT big data model through an optimized formula, and monitor the device operation status through the timeline and filtering tools.

[0010] In step S100, the specific steps include:

[0011] S101. Collect device basic information, device operation parameters, maintenance information, and configuration information. Among them, the device basic information is collected through the device's built-in management interface, device manual, or communication with the device supplier to obtain the device model, brand, production date, and hardware specifications; the maintenance information is collected by collaborating with the device maintenance team to collect the device's maintenance records, including maintenance date, maintenance personnel, and maintenance content; the configuration information includes network configuration, communication protocol configuration, data collection frequency setting, and device working mode.

[0012] S102. Preprocess the collected data. Identify and remove abnormal data based on the statistical outlier detection method, eliminate data records with inconsistent data formats and duplicates, handle data missing values, and use the mean filling method according to the data characteristics. Apply the Min - Max normalization algorithm to uniformly convert data with different ranges and units into a standard scale.

[0013] In step S200, the specific steps include:

[0014] S201. Communicate with relevant enterprise personnel to obtain the enterprise's requirements for the visual intelligent IoT big data platform. According to the results of the requirements research, select Unity for visual modeling, MySQL for structured information storage, InfluxDB for time - series data management, and OPC - UA that supports industrial protocols and the IoT protocol MQTT as data collection tools, and design the overall architecture of the visual intelligent IoT big data platform. The overall architecture of the visual intelligent IoT big data platform includes a data collection layer responsible for connecting various IoT devices and collecting data, a data processing layer for cleaning, transforming, and aggregating data, a data storage layer for storing the structured information and time - series data of the devices, a visual modeling layer for constructing device and business scenario models, and an application layer for providing a device management function interface; among them, the application layer provides a device management function interface, including operations such as device addition, deletion, modification, and query, as well as a device status monitoring function. At the same time, users can understand the device operation status in real - time and view the device's historical data through the application layer interface. The visual modeling layer is used to construct the visual intelligent IoT big data model.

[0015] S202. Use the recurrent neural network RNN to extract the fluctuating feature data in the device operation parameters, and use the convolutional neural network CNN for feature recognition and analysis to extract device connection interruption data.

[0016] In step S300, the specific steps include:

[0017] S301. Obtain the records of the user's interaction with the visual interface from the application layer built in S201. The observation time period is [t0, t m , and the number of interactions occurring at time t i is I i . The calculation formula for the user interaction frequency UF is:

[0018]

[0019] where UF represents the user interaction frequency within the time period, m represents the number of time points between the start time and the end time, t m represents the end time of the observation time period, t0 represents the start time of the observation time period, I i represents the number of interactions between the user and the visual interface at time t i , and i represents the data subscript;

[0020] S302. Verify the accuracy of the established visual intelligent IoT big data model. The calculation formula for the accuracy of the visual intelligent IoT big data model is as follows: VMP = α1 * VRA + β1 * VTF + γ1 * VDF; where VMP represents the accuracy of the visual intelligent IoT big data model, VRA represents the degree of conformity between the representation of data by the visual intelligent IoT big data model and the actual data, VTF represents the detail resolution of the visual intelligent IoT big data model, VDF represents the visual data fluctuation fitting degree, which is mapped from the fluctuation feature data in the extracted device operation parameters through Fourier transform, and α1, β1, and γ1 respectively represent the corresponding weight coefficients, with the value range between 0 and 1, and α1 + β1 + γ1 = 1;

[0021] S303. Verify the structural complexity of the established visual intelligent IoT big data model. The formula for the structural complexity of the visual intelligent IoT big data model is as follows: VMC = α2 * VEN + β2 * VIL + γ2 * VRT + δ1 * VCI; where VMC represents the structural complexity of the visual intelligent IoT big data model, VEN represents the number of elements of the visual intelligent IoT big data model, VIL represents the interaction level of the visual intelligent IoT big data model, VRT represents the rendering time of the visual intelligent IoT big data model, VCI represents the impact degree of visual connection interruption, which is mapped from the extracted device connection interruption data through Fourier transform, and α2, β2, γ2, and δ1 respectively represent the corresponding weight coefficients, with the value range between 0 and 1, and α2 + β2 + γ2 + δ1 = 1.

[0022] In step S400, the specific steps include:

[0023] S401. Comprehensively balance the user interaction frequency, the accuracy and structural complexity of the visual intelligent IoT big data model, and construct an optimization index. The calculation formula is as follows:

[0024]

[0025] Among them, OPT ALL represents the optimization index, VMP represents the accuracy of the visual intelligent IoT big data model, VMC represents the structural complexity of the visual intelligent IoT big data model, p represents the number of user operation scenarios, w k represents the weight of the k-th user operation scenario, UF k represents the user interaction frequency under the k-th user operation scenario, represents the average value of the user interaction frequency under the user operation scenario, std(UF) represents the standard deviation of the user interaction frequency under the user operation scenario, q represents the number of resource types, Δr l represents the difference between the current value and the reference value of the utilization rate of the l-th resource, represents the reference value of the utilization rate of the l-th resource. In the visual intelligent IoT big data platform, users have different operation scenarios such as viewing real-time device data, analyzing historical data trends, and remotely controlling devices. By assigning different weights to different user operation scenarios, the relative importance of each scenario can be reflected in the overall calculation of the optimization index. During the operation of the visual intelligent IoT big data platform, various different types of resources are involved, including computing resources, network resources, storage resources, security protection resources, etc.;

[0026] S402. Collect historical data samples {OPT ALL1 , OPT ALL2 ,..., OPT ALLn} of the n optimization indexes of the visual intelligent IoT big data platform within a period of time, and calculate the average value of the historical data samples of the optimization index. The calculation formula is as follows:

[0027]

[0028] Among them, represents the average value of the historical data samples of the optimization index, OPT ALLi represents the historical data sample of the i-th optimization index, and n represents the number of historical data samples of the optimization index;

[0029] Calculate the standard deviation of the historical data samples of the optimization index. The calculation formula is as follows:

[0030]

[0031] Among them, σ represents the standard deviation of the historical data samples of the optimization index;

[0032] Based on the average value and standard deviation of the historical data samples of the optimization index, set the threshold of the optimization index, which is defined as follows: Among them, k represents the adjustment coefficient;

[0033] S403. Visual presentation: Set a time-axis slider that can be flexibly dragged to view the device operation data in different time periods. At the same time, provide a multi-dimensional screening tool to locate the target device data according to the region and device model.

[0034] A smart IoT big data platform management system based on visual modeling. The system includes a data collection module, a visual modeling module, a data analysis module, and an optimization monitoring module. The data collection module is used to collect and store device basic information, device operation parameters, maintenance information, and configuration information; the visual modeling module is used to preprocess the collected data and build a visual platform that supports multi-dimensional interaction based on the preprocessed data, and at the same time extract features through a feature extraction algorithm; the data analysis module quantitatively analyzes the performance of the visual smart IoT big data model through user interaction frequency calculation, model accuracy verification, and complexity evaluation, and quantifies the interaction behavior characteristics, modeling accuracy, and structural complexity; the optimization monitoring module adjusts the parameters of the visual smart IoT big data model through an optimization formula and monitors the device operation status through a time axis and a screening tool.

[0035] The data collection module includes a device information collection unit, an operation parameter collection unit, a maintenance information collection unit, and a network configuration collection unit. The device information collection unit collects the device model, brand, production date, and hardware specifications through the device's built-in management interface, device manual, or communication with the device supplier; the operation parameter collection unit is used to collect fluctuation feature data; the maintenance information collection unit collects the device's maintenance records, including maintenance date, maintenance personnel, and maintenance content, through cooperation with the device maintenance team; the network configuration collection unit is used to collect network configuration, communication protocol configuration, data collection frequency setting, and device working mode.

[0036] The visualization modeling module includes a data preprocessing unit, a framework design unit, and a feature extraction unit. The data preprocessing unit is used to eliminate outliers, duplicate records, and data with mismatched formats, and convert the data into a unified standard scale using the Min-Max algorithm. The framework design unit is used to communicate with relevant enterprise personnel to obtain the enterprise's requirements for the visualization intelligent IoT big data platform. According to the results of the requirements research, Unity is selected for visualization modeling, MySQL for structured information storage, InfluxDB for time series data management, and OPC-UA supporting industrial protocols and the IoT protocol MQTT as data collection tools to design the overall architecture of the visualization intelligent IoT big data platform, including a data collection layer, a data processing layer, a data storage layer, a visualization modeling layer, and an application layer. The feature extraction unit uses RNN to extract the fluctuation features in the device operation parameters and uses CNN to extract the feature data of device connection interruptions.

[0037] The data analysis module includes an interaction frequency calculation unit, a model accuracy verification unit, and a complexity evaluation unit. The interaction frequency calculation unit calculates the user interaction frequency according to the records of the user's interaction operations with the visualization interface obtained from the application layer. The model accuracy verification unit verifies the accuracy of the established visualization intelligent IoT big data model according to the model accuracy calculation formula. The complexity evaluation unit verifies the structural complexity of the established visualization intelligent IoT big data model according to the model structure complexity calculation formula.

[0038] The optimization monitoring module includes an index construction unit, a threshold setting unit, and a visualization presentation unit. The index construction unit is used to comprehensively balance the user interaction frequency, the accuracy and structural complexity of the visualization intelligent IoT big data model to construct an optimization index. The threshold setting unit is used to collect historical data samples of the optimization index of the visualization intelligent IoT big data platform over a period of time and calculate the threshold of the optimization index. The visualization presentation unit is used to set a time axis slider that can be flexibly dragged to view the device operation data at different times, and at the same time provide a multi-dimensional screening tool to locate the target device data by region and device model.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] 1. By quantitatively evaluating the model accuracy, user interaction frequency, and structural complexity, quantitatively analyzing the performance of the visualization intelligent IoT big data model, and at the same time, based on the comprehensive analysis of multi-dimensional data, providing a strong guarantee for the stable operation of the device;

[0041] 2. Construct a comprehensive optimization index, set the threshold according to historical data, adjust the parameters of the visualization intelligent IoT big data model through the optimization formula, and monitor the device operation status through the time axis and screening tool;

[0042] 3. The present invention collects device basic information, operation parameters, maintenance information, and configuration information, and uses preprocessing techniques to ensure the quality and availability of data, improve the efficiency of device data management, reduce the cumbersome processes and time costs of manual data processing, avoid data omissions and errors caused by manual operations, and ensure the integrity and accuracy of device information. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a schematic flowchart of the management method of the intelligent Internet of Things big data platform based on visual modeling of the present invention;

[0044] Figure 2 is a schematic structural diagram of the management system of the intelligent Internet of Things big data platform based on visual modeling of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0046] As Figure 1 - Figure 2 shown, the present invention provides a technical solution, a management method for an intelligent Internet of Things big data platform based on visual modeling, and the method includes the following steps:

[0047] S100. Collect and store device basic information, device operation parameters, maintenance information, and configuration information, and preprocess the collected data;

[0048] S200. Based on the preprocessed data, construct a visual platform that supports multi-dimensional interaction, and extract features through a feature extraction algorithm;

[0049] S300. Through user interaction frequency calculation, model accuracy verification, and complexity evaluation, quantitatively analyze the performance of the visual intelligent Internet of Things big data model, and quantify the interactive behavior characteristics, modeling accuracy, and structural complexity;

[0050] S400. Adjust the parameters of the visual intelligent Internet of Things big data model through an optimization formula, and monitor the device operation status through a time axis and a screening tool.

[0051] In step S100, the specific steps include:

[0052] S101. Collect the basic information of the device, device operation parameters, maintenance information, and configuration information. Among them, the basic information of the device is collected through the built-in management interface of the device, the device manual, or communication with the device supplier, including the model, brand, production date, and hardware specifications of the device; the maintenance information is collected by collaborating with the device maintenance team, including the maintenance records of the device, including the maintenance date, maintenance personnel, and maintenance content; the configuration information includes network configuration, communication protocol configuration, data collection frequency setting, and device working mode.

[0053] S102. Preprocess the collected data. Identify and remove abnormal data based on the statistical outlier detection method, eliminate data records with mismatched data formats and duplicates, process data missing values, and use the mean filling method according to the data characteristics. Apply the Min-Max normalization algorithm to uniformly convert data with different ranges and units into a standard scale.

[0054] In step S200, the specific steps include:

[0055] S201. Communicate with relevant enterprise personnel to obtain the enterprise's requirements for the visual intelligent Internet of Things big data platform. According to the results of the requirements research, select Unity for visual modeling, MySQL for structured information storage, InfluxDB for time series data management, and OPC-UA supporting industrial protocols and the Internet of Things protocol MQTT as data collection tools. Design the overall architecture of the visual intelligent Internet of Things big data platform. The overall architecture of the visual intelligent Internet of Things big data platform includes a data collection layer responsible for connecting various Internet of Things devices and collecting data, a data processing layer for cleaning, transforming, and aggregating data, a data storage layer for storing the structured information and time series data of the device, a visual modeling layer for constructing device and business scenario models, and an application layer for providing a device management function interface; among them, the application layer provides a device management function interface, including operations such as adding, deleting, modifying, and querying devices, as well as a device status monitoring function. At the same time, users can understand the running status of the device in real time and view the historical data of the device through the application layer interface. The visual modeling layer is used to construct a visual intelligent Internet of Things big data model.

[0056] S202. Use the recurrent neural network RNN to extract the fluctuating feature data in the device operation parameters, and use the convolutional neural network CNN for feature recognition and analysis to extract device connection interruption data.

[0057] In step S300, the specific steps include:

[0058] S301. Obtain the records of the user's interaction with the visual interface from the application layer built in S201. The observation time period is [t0, t m , and the number of interactions occurring at time t i is Ii , the calculation formula for the user interaction frequency UF is:

[0059]

[0060] Among them, UF represents the user interaction frequency within the time period, m represents the number of time points between the start time and the end time, t m represents the end time of the observation time period, t0 represents the start time of the observation time period, I i represents the number of interactions between the user and the visualization interface at time t i moment, and i represents the data subscript;

[0061] S302. Verify the accuracy of the established visual intelligent IoT big data model. The calculation formula for the accuracy of the visual intelligent IoT big data model is as follows: VMP = α1 * VRA + β1 * VTF + γ1 * VDF; where, VMP represents the accuracy of the visual intelligent IoT big data model, VRA represents the degree of conformity between the representation of the data by the visual intelligent IoT big data model and the actual data, VTF represents the detail resolution of the visual intelligent IoT big data model, VDF represents the fitting degree of visual data fluctuations, which is mapped from the fluctuation characteristic data in the extracted device operation parameters through Fourier transform, and α1, β1, and γ1 respectively represent the corresponding weight coefficients, with a value range between 0 and 1, and α1 + β1 + γ1 = 1;

[0062] S303. Verify the structural complexity of the established visual intelligent IoT big data model. The formula for the structural complexity of the visual intelligent IoT big data model is as follows: VMC = α2 * VEN + β2 * VIL + γ2 * VRT + δ1 * VCI; where, VMC represents the structural complexity of the visual intelligent IoT big data model, VEN represents the number of elements of the visual intelligent IoT big data model, VIL represents the interaction level of the visual intelligent IoT big data model, VRT represents the rendering time of the visual intelligent IoT big data model, VCI represents the impact degree of visual connection interruption, which is mapped from the extracted device connection interruption data through Fourier transform, and α2, β2, γ2, and δ1 respectively represent the corresponding weight coefficients, with a value range between 0 and 1, and α2 + β2 + γ2 + δ1 = 1.

[0063] In step S400, the specific steps include:

[0064] S401. Comprehensively balance the user interaction frequency, the accuracy and the structural complexity of the visual intelligent IoT big data model, and construct an optimization index. The calculation formula is as follows:

[0065]

[0066] Among them, OPTALL denotes the optimization metric, VMP denotes the accuracy of the visual intelligent IoT big data model, VMC denotes the structural complexity of the visual intelligent IoT big data model, p denotes the number of user operation scenarios, and w k denotes the weight of the k-th user operation scenario, and UF k denotes the user interaction frequency under the k-th user operation scenario, denotes the average value of the user interaction frequency under the user operation scenario, std(UF) denotes the standard deviation of the user interaction frequency under the user operation scenario, q denotes the number of resource types, and Δr l denotes the difference between the current value and the reference value of the utilization rate of the l-th resource, denotes the reference value of the utilization rate of the l-th resource. In the visual intelligent IoT big data platform, users have different operation scenarios such as viewing real-time device data, analyzing historical data trends, and remotely controlling devices. By assigning different weights to different user operation scenarios, the relative importance of each scenario can be reflected in the overall calculation of the optimization metric. During the operation of the visual intelligent IoT big data platform, various different types of resources are involved, including computing resources, network resources, storage resources, security protection resources, etc.;

[0067] S402. Collect historical data samples {OPT ALL1 , OPT ALL2 ,..., OPT ALLn} of the n optimization metrics of the visual intelligent IoT big data platform within a period of time, and calculate the average value of the historical data samples of the optimization metric. The calculation formula is as follows:

[0068]

[0069] where, denotes the average value of the historical data samples of the optimization metric, OPT ALLi denotes the i-th historical data sample of the optimization metric, and n denotes the number of historical data samples of the optimization metric;

[0070] Calculate the standard deviation of the historical data samples of the optimization metric. The calculation formula is as follows:

[0071]

[0072] where, σ denotes the standard deviation of the historical data samples of the optimization metric;

[0073] Set the threshold of the optimization metric according to the average value and standard deviation of the historical data samples of the optimization metric. The definition is as follows: where, k denotes the adjustment coefficient;

[0074] S403, visual presentation, set a timeline slider that can be flexibly dragged to view equipment operation data in different time periods, and provide multi-dimensional filtering tools to locate target equipment data by region and equipment model.

[0075] A smart IoT big data platform management system based on visual modeling, the system includes a data acquisition module, a visual modeling module, a data analysis module and an optimization monitoring module, the data acquisition module is used to collect and store basic information of equipment, equipment operating parameters, maintenance information and configuration information; the visual modeling module is used to preprocess the collected data, and build a visualization platform supporting multi-dimensional interaction based on the preprocessed data, and extract features through a feature extraction algorithm; the data analysis module quantitatively analyzes the performance of the visualized smart IoT big data model through user interaction frequency calculation, model accuracy verification and complexity evaluation, and quantifies the interactive behavior characteristics, modeling accuracy and structural complexity; the optimization monitoring module adjusts the parameters of the visualized smart IoT big data model through optimization formulas, and monitors the equipment operation status through a timeline and screening tools.

[0076] The data collection module includes an equipment information collection unit, an operation parameter collection unit, a maintenance information collection unit and a network configuration collection unit. The equipment information collection unit collects the model, brand, production date and hardware specifications of the equipment through the equipment's own management interface, equipment manual or communication with the equipment supplier; the operation parameter collection unit is used to collect fluctuation characteristic data; the maintenance information collection unit collects equipment maintenance records, including maintenance date, maintenance personnel and maintenance content, by collaborating with the equipment maintenance team; the network configuration collection unit is used to collect network configuration, communication protocol configuration, data collection frequency settings and equipment working mode.

[0077] The visual modeling module includes a data preprocessing unit, a framework design unit and a feature extraction unit. The data preprocessing unit is used to eliminate outliers, duplicate records and data that do not match the format, and use the Min-Max algorithm to convert the data into a unified standard scale; the framework design unit is used to communicate with relevant personnel of the enterprise to obtain the enterprise's needs for the visual smart IoT big data platform. According to the results of the demand survey, Unity is selected for visual modeling, MySQL is used for structured information storage, InfluxDB is used for time series data management, and OPC-UA that supports industrial protocols and MQTT, an IoT protocol, are used as data collection tools to design the overall architecture of the visual smart IoT big data platform, including data collection layer, data processing layer, data storage layer, visual modeling layer and application layer; the feature extraction unit uses RNN to extract fluctuation characteristics in device operating parameters, and uses CNN to extract feature data of device connection interruption.

[0078] The data analysis module includes an interaction frequency calculation unit, a model accuracy verification unit, and a complexity evaluation unit. The interaction frequency calculation unit calculates the user interaction frequency according to the records of the user's interaction operations with the visualization interface obtained from the application layer. The model accuracy verification unit verifies the accuracy of the established visual intelligent Internet of Things big data model according to the model accuracy calculation formula. The complexity evaluation unit verifies the structural complexity of the established visual intelligent Internet of Things big data model according to the model structure complexity calculation formula.

[0079] The optimization monitoring module includes an index construction unit, a threshold setting unit, and a visualization presentation unit. The index construction unit is used to comprehensively balance the user interaction frequency, the accuracy and structural complexity of the visual intelligent Internet of Things big data model, and construct optimization indexes. The threshold setting unit is used to collect historical data samples of the optimization indexes of the visual intelligent Internet of Things big data platform over a period of time and calculate the thresholds of the optimization indexes. The visualization presentation unit is used to set a time axis slider that can be flexibly dragged to view the device operation data at different times, and at the same time provide a multi-dimensional screening tool to locate the target device data according to the region and device model.

[0080] In the embodiment: Collect basic device information, device operation parameters, maintenance information, and configuration information. Remove outliers, duplicate records, and data with mismatched formats from the collected data, and use the Min-Max algorithm to convert the data into a unified standard scale. Build a visualization platform using Unity as needed. Build an intelligent device model in the visualization modeling layer. Extract fluctuating feature data through the Recurrent Neural Network (RNN), and it is found that there is a fluctuation cycle within every 30 minutes, and the fluctuation amplitude is between 0.08 and 0.12. Use the Convolutional Neural Network (CNN) to extract connection interruption data, and it is found that there has been 1 short connection interruption event in the past month, and the interruption time is about 8 seconds. Obtain the interaction operation records of users with the visualization interface within one month from the application layer. The observation period is from August 1, 2023, to August 31, 2023. A total of 350 interaction operations are counted. 20 interactions occurred on August 10. Calculated based on 31 days, the user interaction frequency UF is approximately 1.13 times per day. After verification, by comparing the error between the device operation data and the actual collected data, the degree of conformity VRA of the visualized intelligent Internet of Things big data model to the data and the actual data is 0.82. Evaluate the detail resolution VTF of the model based on the degree to which the model can display the changes in device operation parameters, and it is 0.78. Determine the visualized data fluctuation fitting degree VDF as 0.75 based on the fitting degree between the extracted fluctuating feature data after Fourier transform and the actual fluctuation situation. Set the weight coefficients α1 = 0.38, β1 = 0.3, γ1 = 0.32, then the accuracy VMP of the visualized intelligent Internet of Things big data model is 0.786. The number of elements VEN of the visualized intelligent Internet of Things big data model is 160. The user enters from the main interface to the device distribution interface, then to the specific device details interface, and finally to the parameter analysis interface, with a total of 3 layers, so the interaction level VIL is 3. The average rendering time VRT is 1.2 seconds. Evaluate the visualized connection interruption impact degree VCI as 0.45 based on the impact of the connection interruption on device operation and data collection. Set the weight coefficients α2 = 0.32, β2 = 0.28, γ2 = 0.2, δ1 = 0.2, then the structural complexity VMC of the visualized intelligent Internet of Things big data model is 52.37. There are 3 user operation scenarios: viewing the real-time status of the device, viewing the historical energy consumption data of the device, and remotely controlling the device. The weights are w1 = 0.42, w2 = 0.33, w3 = 0.25 respectively. The user interaction frequency UF1 in the scenario of viewing the real-time status of the device is 0.6 times per day, UF2 in the scenario of viewing the historical energy consumption data of the device is 0.35 times per day, and UF3 in the scenario of remotely controlling the device is 0.18 times per day. The average value of the user interaction frequency in the user operation scenario is 0.38 times per day, and the standard deviation is 0.15. During the operation of the platform, the current value of the computing resource utilization rate is 60%, and the baseline value is 45%, that is, Δr1 = 15%; the current value of the network resource utilization rate is 50%, and the baseline value is 35%, that is, Δr2 = 15%; the current value of the storage resource utilization rate is 70%, and the baseline value is 55%, that is, Δr3 = 15%; the current value of the security protection resource utilization rate is 80%, and the baseline value is 65%, that is, Δr4 = 15%. Then the optimization index OPT. ALL ≈0.43. Collect the historical data samples of 25 optimization indexes of the platform in the past two months, calculate the average value to be 0.46, and the standard deviation to be 0.03. Set the adjustment coefficient k = 1.8. Then the threshold T of the optimization index OPT = 0.406. Since the current optimization index 0.43 is greater than the threshold 0.406, it indicates that the platform is in good operating condition.

[0081] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A management method for a smart IoT big data platform based on visual modeling, characterized in that: The method includes the following steps: S100. Collect and store the basic information of the device, device operation parameters, maintenance information, and configuration information, and preprocess the collected data; S200. Based on the preprocessed data, build a visualization platform that supports multi-dimensional interaction, and extract features through a feature extraction algorithm; S300. Through user interaction frequency calculation, model accuracy verification, and complexity evaluation, quantitatively analyze the performance of the visualized intelligent IoT big data model, and quantify the interactive behavior characteristics, modeling accuracy, and structural complexity; S400. Adjust the parameters of the visualized intelligent IoT big data model through an optimization formula, and monitor the device operation status through a timeline and a filtering tool.

2. The management method of the intelligent IoT big data platform based on visual modeling according to claim 1, characterized in that: In step S100, the specific steps include: S101. Collect the basic information of the device, device operation parameters, maintenance information, and configuration information. Among them, the basic information of the device is collected through the device's built-in management interface, device manual, or communication with the device supplier to collect the device model, brand, production date, and hardware specifications; the maintenance information is collected by collaborating with the device maintenance team to collect the device's maintenance records, including maintenance date, maintenance personnel, and maintenance content; the configuration information includes network configuration, communication protocol configuration, data collection frequency setting, and device working mode; S102. Preprocess the collected data. Identify and remove abnormal data based on the statistical outlier detection method, eliminate data records with inconsistent data formats and duplicates, handle data missing values, and use the mean filling method according to the data characteristics. Apply the Min-Max normalization algorithm to uniformly convert data with different ranges and units into a standard scale.

3. The management method of the intelligent IoT big data platform based on visual modeling according to claim 2, wherein: In step S200, the specific steps include: S201. Communicate with relevant enterprise personnel to obtain the enterprise's requirements for the visualized intelligent IoT big data platform. According to the results of the requirements research, select Unity for visual modeling, MySQL for structured information storage, InfluxDB for time series data management, and OPC-UA that supports industrial protocols and the IoT protocol MQTT as data collection tools, and design the overall architecture of the visualized intelligent IoT big data platform. The overall architecture of the visualized intelligent IoT big data platform includes a data collection layer responsible for connecting various IoT devices and collecting data, a data processing layer for cleaning, transforming, and aggregating data, a data storage layer for storing the structured information and time series data of the device, a visual modeling layer for building device and business scenario models, and an application layer for providing a device management function interface; among them, the application layer provides a device management function interface, including operations such as adding, deleting, modifying, and querying devices, as well as a device status monitoring function. At the same time, users can understand the device operation status in real time and view the historical data of the device through the application layer interface. The visual modeling layer is used to build a visualized intelligent IoT big data model; S202. Use the recurrent neural network RNN to extract the fluctuation feature data in the device operation parameters, and use the convolutional neural network CNN for feature recognition and analysis to extract the device connection interruption data.

4. The management method of the intelligent IoT big data platform based on visual modeling according to claim 3, characterized in that: In step S300, the specific steps include: S301. Obtain the records of the user's interaction operations with the visualization interface from the application layer built in S201. The observation time period is [t0, t m , and the number of interactions occurring at time t i is I i . The calculation formula for the user interaction frequency UF is: Among them, UF represents the user interaction frequency within a time period, m represents the number of time points between the start time and the end time, t m represents the end time of the observation time period, t0 represents the start time of the observation time period, I i represents the number of interactions between the user and the visualization interface that occur at time t i moment, and i represents the data subscript; S302. Verify the accuracy of the established visual intelligent IoT big data model. The accuracy calculation formula of the visual intelligent IoT big data model is as follows: VMP = α1 * VRA + β1 * VTF + γ1 * VDF; where VMP represents the accuracy of the visual intelligent IoT big data model, VRA represents the degree of conformity between the representation of data by the visual intelligent IoT big data model and the actual data, VTF represents the detail resolution of the visual intelligent IoT big data model, VDF represents the visual data fluctuation fitting degree, which is mapped from the fluctuation characteristic data in the extracted device operation parameters through Fourier transform, and α1, β1, and γ1 respectively represent the corresponding weight coefficients, with the value range between 0 and 1, and α1 + β1 + γ1 = 1; S303. Verify the structural complexity of the established visual intelligent IoT big data model. The structural complexity formula of the visual intelligent IoT big data model is as follows: VMC = α2 * VEN + β2 * VIL + γ2 * VRT + δ1 * VCI; where VMC represents the structural complexity of the visual intelligent IoT big data model, VEN represents the number of elements of the visual intelligent IoT big data model, VIL represents the interaction level of the visual intelligent IoT big data model, VRT represents the rendering time of the visual intelligent IoT big data model, VCI represents the impact degree of visual connection interruption, which is mapped from the extracted device connection interruption data through Fourier transform, and α2, β2, γ2, and δ1 respectively represent the corresponding weight coefficients, with the value range between 0 and 1, and α2 + β2 + γ2 + δ1 = 1.

5. The management method of the intelligent IoT big data platform based on visual modeling according to claim 4, characterized in that: In step S400, the specific steps include: S401. Comprehensively balance the user interaction frequency, the accuracy and the structural complexity of the visual intelligent IoT big data model, and construct an optimization index. The calculation formula is as follows: Among them, OPT ALL represents the optimization index, VMP represents the accuracy of the visual intelligent IoT big data model, VMC represents the structural complexity of the visual intelligent IoT big data model, p represents the number of user operation scenarios, w k represents the weight of the k-th user operation scenario, UF k represents the user interaction frequency under the k-th user operation scenario, represents the average value of the user interaction frequency under the user operation scenario, std(UF) represents the standard deviation of the user interaction frequency under the user operation scenario, q represents the number of resource types, Δr l represents the difference between the current value and the reference value of the utilization rate of the l-th resource, represents the reference value of the utilization rate of the l-th resource; S402. Collect historical data samples of the n optimization metrics of the visual intelligent IoT big data platform over a period of time {OPT ALL1 , OPT ALL2 ,..., OPT ALLn}, and calculate the average value of the historical data samples of the optimization metrics. The calculation formula is as follows: Among them, represents the average value of the historical data samples of the optimization index, OPT ALLi represents the historical data sample of the i-th optimization index, and n represents the number of historical data samples of the optimization index; Calculate the standard deviation of the historical data samples of the optimization index. The calculation formula is as follows: where σ represents the standard deviation of the historical data samples of the optimization index; Set the threshold of the optimization metric according to the mean and standard deviation of the historical data samples of the optimization metric, which is defined as follows: where k represents the adjustment coefficient; S403. Visual presentation. Set a time-axis slider that can be flexibly dragged to view the device operation data in different time periods. At the same time, provide a multi-dimensional filtering tool to locate the target device data by region and device model.

6. The intelligent IoT big data platform management system based on visual modeling is characterized in that: The system includes a data collection module, a visual modeling module, a data analysis module, and an optimization monitoring module. The data collection module is used to collect and store device basic information, device operation parameters, maintenance information, and configuration information; The visual modeling module is used to preprocess the collected data, construct a visual platform supporting multi-dimensional interaction based on the preprocessed data, and extract features through a feature extraction algorithm; The data analysis module quantitatively analyzes the performance of the visual intelligent IoT big data model through user interaction frequency calculation, model accuracy verification, and complexity evaluation, and quantifies the interaction behavior characteristics, modeling accuracy, and structural complexity; the optimization monitoring module adjusts the parameters of the visual intelligent IoT big data model through an optimization formula and monitors the device operation status through a time axis and a filtering tool.

7. The intelligent IoT big data platform management system based on visual modeling according to claim 6, characterized in that: The data acquisition module includes a device information acquisition unit, an operating parameter acquisition unit, a maintenance information acquisition unit, and a network configuration acquisition unit. The device information acquisition unit collects the device model, brand, production date, and hardware specifications through the device's built-in management interface, device manual, or communication with the device supplier. The operating parameter acquisition unit is used to acquire fluctuation feature data. The maintenance information acquisition unit collects the device's maintenance records, including maintenance date, maintenance personnel, and maintenance content, by collaborating with the device maintenance team. The network configuration acquisition unit is used to acquire network configuration, communication protocol configuration, data acquisition frequency setting, and device working mode.

8. The intelligent IoT big data platform management system based on visual modeling according to claim 7, characterized in that: The visualization modeling module includes a data preprocessing unit, a framework design unit, and a feature extraction unit. The data preprocessing unit is used to remove outliers, duplicate records, and data with mismatched formats, and convert the data to a unified standard scale using the Min-Max algorithm. The framework design unit communicates with relevant enterprise personnel to obtain the enterprise's requirements for the visualization intelligent IoT big data platform. According to the results of the requirements research, Unity is selected for visualization modeling, MySQL for structured information storage, InfluxDB for time series data management, and OPC-UA, which supports industrial protocols, and MQTT, an IoT protocol, as data acquisition tools to design the overall architecture of the visualization intelligent IoT big data platform, including a data acquisition layer, a data processing layer, a data storage layer, a visualization modeling layer, and an application layer. The feature extraction unit uses RNN to extract the fluctuation features in the device operating parameters and uses CNN to extract the feature data of device connection interruptions.

9. The intelligent IoT big data platform management system based on visual modeling according to claim 8, wherein: The data analysis module includes an interaction frequency calculation unit, a model accuracy verification unit, and a complexity evaluation unit. The interaction frequency calculation unit calculates the user interaction frequency based on the records of the user's interaction operations with the visualization interface obtained from the application layer. The model accuracy verification unit verifies the accuracy of the established visualization intelligent IoT big data model according to the model accuracy calculation formula. The complexity evaluation unit verifies the structural complexity of the established visualization intelligent IoT big data model according to the model structure complexity calculation formula.

10. The intelligent IoT big data platform management system based on visual modeling according to claim 9, characterized in that: The optimization monitoring module includes an index construction unit, a threshold setting unit, and a visualization presentation unit. The index construction unit is used to comprehensively balance the user interaction frequency, the accuracy and structural complexity of the visualization intelligent IoT big data model, and construct an optimization index. The threshold setting unit is used to collect historical data samples of the optimization index of the visualization intelligent IoT big data platform over a period of time and calculate the threshold of the optimization index. The visualization presentation unit is used to set a time axis slider that can be flexibly dragged to view the device operation data at different time periods, and at the same time provides a multi-dimensional screening tool to locate the target device data by region and device model.

Images