An Internet of Things-based terminal security intelligent monitoring system and method

Abstract

The present invention discloses a terminal security intelligent monitoring system and method based on the Internet of Things, relating to the technical field of terminal security intelligent monitoring, and comprising the following steps: during the operation of the frequency converter, converting data that the rectified direct current enters the inverter and is converted into alternating current of the required frequency through an insulated gate bipolar transistor is acquired in real time; features are extracted from the conversion data, and abnormal analysis is performed on the extracted features. The present invention uses a deep learning model to extract features and perform abnormal analysis on the conversion data, accurately identify potential abnormalities in the frequency conversion process of the frequency converter, ensure timely processing, prevent the fan from affecting the train ventilation and air conditioning systems. By comparing the actual and expected frequency curves, an expected deviation index is generated to provide a maintenance guide, reduce faults and downtime, improve the efficiency and lifespan of the frequency converter, ensure the normal operation of the train system, and improve the operation efficiency and passenger satisfaction.

Description

Technical Field

[0001] The present invention relates to the technical field of terminal security intelligent monitoring, and in particular to a terminal security intelligent monitoring system and method based on the Internet of Things. Background Art

[0002] The terminal security intelligent monitoring based on the Internet of Things refers to connecting the terminal equipment (such as the variable frequency axial flow fan of the train) to the Internet of Things platform through sensors and communication equipment, collecting the operating status data of the equipment (such as temperature, vibration, speed, etc.) in real time, and using intelligent algorithms to analyze and process the data to achieve remote monitoring of the equipment, fault warning and maintenance suggestions. For the variable frequency axial flow fan of the train, the intelligent monitoring of the Internet of Things can monitor its working condition in real time, detect potential faults in time, optimize the fan operating parameters, reduce manual inspections, and improve the reliability and safety of the equipment, thereby ensuring the stable operation of the train ventilation system.

[0003] The train variable frequency axial flow fan is a device specially used in the train ventilation system. It uses variable frequency technology to adjust the fan speed to meet different ventilation needs. This fan controls the air volume and air pressure by changing the motor frequency to achieve energy saving and efficient operation. The design of the axial flow fan allows the air to flow along the axial direction, with the advantages of compact structure, light weight and low noise. It is mainly used in the train air conditioning system, carriage ventilation and heat dissipation of electronic equipment. Frequency conversion control can not only improve the stability and reliability of the system, but also extend the service life of the equipment, reduce maintenance costs, and improve the comfort and safety of passengers.

[0004] Insulated gate bipolar transistor (IGBT) is a key power switching device in the inverter, and its role is to efficiently switch current during rectification and inversion. IGBT combines the high current density of bipolar transistors and the high input impedance characteristics of insulated gate field effect transistors (MOSFET), and can switch quickly at high voltage and high current to control the AC frequency and voltage output by the inverter, thereby accurately adjusting the speed and torque of the motor and achieving efficient frequency conversion and power transmission.

[0005] The prior art has the following deficiencies:

[0006] In the process of using frequency conversion technology to adjust the fan speed of the train variable frequency axial flow fan, if there is a potential abnormality when the insulated gate bipolar transistor converts the rectified DC power into the required frequency AC power, resulting in the inverter being unable to perform efficient frequency conversion, the existing technology is usually unable to intelligently identify the potential abnormality. The inverter is unable to efficiently perform frequency conversion, which may cause the fan to be unable to adjust the speed according to ventilation needs. When this happens, it will seriously affect the ventilation and air-conditioning system in the train and reduce the comfort of passengers.

[0007] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0008] The purpose of the present invention is to provide a terminal security intelligent monitoring system and method based on the Internet of Things, which uses a deep learning model to extract features and analyze abnormalities of conversion data acquired in real time, accurately identify potential abnormalities of the frequency converter during the frequency conversion process, and solve the defect of being unable to intelligently identify abnormalities, ensure timely handling of abnormalities, prevent the fan from affecting the ventilation and air-conditioning system of the train, and improve system stability and passenger comfort. By comparing the actual frequency conversion curve with the expected frequency curve, an expected deviation index is generated, and the system can analyze and provide maintenance guidance in real time, ensure timely and accurate abnormal maintenance, reduce failures and downtime, improve the efficiency and life of the frequency converter, ensure the normal operation of the train system, and improve operational efficiency and passenger satisfaction, so as to solve the problems in the above-mentioned background technology.

[0009] In order to achieve the above object, the present invention provides the following technical solution: a terminal security intelligent monitoring method based on the Internet of Things, comprising the following steps:

[0010] During the operation of the inverter, the conversion data of the rectified DC power entering the inverter and being converted into the AC power of the required frequency through the insulated gate bipolar transistor is obtained in real time;

[0011] Extract features from the converted data and perform anomaly analysis on the extracted features. Input the feature data after anomaly analysis into a pre-trained deep learning model, and use the deep learning model to evaluate the process of converting DC power into AC power of the required frequency and identify potential anomalies.

[0012] After identifying the potential anomaly, the actual frequency conversion curve of the frequency converter when performing frequency conversion within a period of time is obtained, and the expected frequency curve required when the frequency converter performs frequency conversion within the same period of time is obtained;

[0013] Compare and analyze the actual frequency conversion curve with the expected frequency curve, and conduct real-time analysis of the deviations and anomalies in the frequency conversion process of the inverter;

[0014] Based on the results of the comparison and analysis, maintenance guidance is provided for the abnormal process when DC power is converted into AC power of the required frequency, and abnormal maintenance is performed in a timely manner.

[0015] Preferably, features are extracted from the conversion data, and the extracted features include circuit harmonics in the alternating current and switching durations of the insulated gate bipolar transistors. After abnormal analysis of the circuit harmonics in the alternating current and the switching durations of the insulated gate bipolar transistors, harmonic distortion reference values ​​and switching time difference reference values ​​are generated. The harmonic distortion reference values ​​and switching time difference reference values ​​after abnormal analysis are input into a pre-trained deep learning model to generate an AC conversion coefficient. The process of converting direct current into alternating current of a required frequency is evaluated using the AC conversion coefficient to identify potential conversion anomalies.

[0016] Preferably, after performing abnormal analysis on circuit harmonics in the alternating current, the steps of generating a harmonic distortion reference value are as follows:

[0017] In the detection window, the AC waveform is sampled at high frequency to obtain a set of discrete data points. The sampled signal is expressed as ,in n =1, 2, 3, 4, ..., N -1, N is a positive integer;

[0018] Calculate the root mean square value of the original AC signal. The calculation expression is: , where Represents the RMS value of the original AC signal;

[0019] Use a bandpass filter bank to separate the harmonic components of different frequencies and k The sampled values ​​of the subharmonic components are To express;

[0020] Calculate the root mean square value of each harmonic component. The calculation expression is: , where Indicates k The RMS value of the subharmonics, k Indicates the number or order of harmonics;

[0021] Calculate the total harmonic distortion, the expression is: ,in, is the fundamental component, K is the upper limit of the harmonic order considered, is the total harmonic distortion;

[0022] Calculate the harmonic distortion reference value. The calculation expression is: , where Indicates the harmonic distortion reference value, Represents weighted total harmonic distortion, and the calculation expression is: ,in, represents the weight function, .

[0023] Preferably, after performing abnormal analysis on the switching time of the insulated gate bipolar transistor in the alternating current, the step of generating the switching time difference reference value is as follows:

[0024] Collect the switching duration data of IGBT in real time within the detection window, including rise time , Fall time , Turn-on delay time and turn-off delay time ;

[0025] Use the median absolute deviation to remove outliers. The specific formula is: ,in, For the collected switching time data, the criteria for judging abnormal values ​​are: , for those that meet this condition , identified as outliers and removed from the data set, where Indicates the threshold set when judging abnormal values;

[0026] Short-time Fourier transform is used to analyze the time-varying characteristics of the switching duration, extract the instantaneous frequency and amplitude changes, and set is the switching time signal, then the expression of short-time Fourier transform is ,in, Represents the time t and frequency f The signal on The Fourier transform result is specifically explained as It is the spectrum representation of the signal in a specific time window, that is, at a certain time t, different frequency components f The amplitude and phase information of is a window function used to capture a small segment of the signal at time t. is the original signal, is a complex exponential function, which is used to perform Fourier transform on the signal. j is an imaginary unit;

[0027] Extract features from the short-time Fourier transform results, including instantaneous frequency offset and amplitude change, and define instantaneous frequency offset and amplitude changes for: , , Indicates that at time t, the signal After short-time Fourier transform processing, the frequency f The amplitude of Indicates that at time point t-1, the signal After short-time Fourier transform processing, the frequency fThe amplitude at

[0028] Use the characteristic entropy to calculate the switching time difference uncertainty index. The characteristic entropy represents the uncertainty of the system state. Let be the probability distributions of the instantaneous frequency offset and amplitude change, then the characteristic entropy H The calculation expression is: ;

[0029] The comprehensive characteristic entropy H , the instantaneous frequency offset and the amplitude change , calculate the switching time difference reference value. The calculation expression is: , where, represents the switching time difference reference value, represents the result of summing up the instantaneous frequency offset within the detection window, represents the result of summing up the amplitude change within the detection window, , , are respectively H , , The weight coefficients of , and satisfy

[0030] Preferably, within the detection window, compare and analyze the AC conversion coefficient generated when the rectified DC power enters the inverter and is converted into the required frequency AC power by an insulated gate bipolar transistor with a preset AC conversion coefficient reference threshold. The results of the comparison and analysis are as follows:

[0031] If the AC conversion coefficient is greater than or equal to the AC conversion coefficient reference threshold, generate a potential abnormal signal;

[0032] If the AC conversion coefficient is less than the AC conversion coefficient reference threshold, generate a normal conversion signal.

[0033] Preferably, compare and analyze the actual frequency conversion curve with the expected frequency curve to generate an expected deviation index. The specific steps are as follows:

[0034] Obtain the actual frequency conversion curve and the expected frequency curve. Represent the actual frequency conversion curve with and represent the expected frequency curve with , , , where, and represent the actual and expected frequency values at the same time point respectively, m is the number of sampling points, uAn index representing a time point;

[0035] The frequency data is normalized, and the normalization formula is: , ,in, and are respectively the actual frequency value and the expected frequency value after normalization, Indicates the minimum value in the actual frequency conversion curve, Indicates the maximum value in the actual frequency conversion curve. Similarly, represents the minimum value in the expected frequency curve, Shows the maximum value in the expected frequency curve;

[0036] Calculate the cosine similarity between the actual frequency conversion curve and the expected frequency curve. The calculation expression is:

[0037] , where is the cosine similarity, with a value range of [-1, 1];

[0038] Calculate the expected deviation index, the calculation expression is: , where Represents the expected deviation index.

[0039] Preferably, the expected deviation index generated by comparing the actual frequency conversion curve with the expected frequency curve is compared with a preset expected deviation index reference threshold, and the result of the comparison analysis is as follows:

[0040] If the expected deviation index is greater than or equal to the expected deviation index reference threshold, a deviation signal is generated, indicating that there is a deviation between the actual frequency conversion curve and the expected frequency curve that exceeds the allowable range and needs to be processed in time;

[0041] If the expected deviation index is less than the expected deviation index reference threshold, an expected compliance signal is generated, indicating that the deviation between the actual frequency conversion curve and the expected frequency curve is within the allowable range and no processing is required.

[0042] A terminal security intelligent monitoring system based on the Internet of Things, including a data acquisition module, a feature extraction and anomaly detection module, a frequency curve acquisition module, a comparison and deviation analysis module and a maintenance guidance module;

[0043] The data acquisition module acquires the conversion data of the rectified direct current into the required frequency alternating current through the insulated gate bipolar transistor after entering the inverter in real time during the operation of the inverter;

[0044] The feature extraction and anomaly detection module extracts features from the converted data and performs anomaly analysis on the extracted features. The feature data after anomaly analysis is input into a pre-trained deep learning model, and the deep learning model is used to evaluate the process of converting DC power into AC power of the required frequency and identify potential anomalies.

[0045] The frequency curve acquisition module, after identifying the potential anomaly, acquires the actual frequency conversion curve when the frequency converter performs frequency conversion within a period of time, and simultaneously acquires the expected frequency curve required when the frequency converter performs frequency conversion within the same period of time;

[0046] The comparison and deviation analysis module compares and analyzes the actual frequency conversion curve with the expected frequency curve, and performs real-time analysis on the deviations and anomalies in the frequency conversion process of the inverter;

[0047] The maintenance guidance module provides maintenance guidance for abnormal processes when converting DC power into AC power of required frequency based on the results of comparison and analysis, and performs abnormal maintenance in a timely manner.

[0048] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0049] The present invention utilizes a deep learning model to perform feature extraction and anomaly analysis on conversion data acquired in real time, accurately identifying potential abnormalities of the frequency converter during the frequency conversion process, solving the defect of the prior art that potential abnormalities cannot be intelligently identified, ensuring that the frequency converter can promptly detect and handle abnormalities during operation, and preventing the fan from affecting the ventilation and air-conditioning system in the train due to the inability to adjust the speed, thereby improving the stability of the system and the comfort of passengers.

[0050] After identifying potential abnormalities, the present invention generates an expected deviation index by acquiring and comparing the actual frequency conversion curve with the expected frequency curve, and compares it with the expected deviation index reference threshold. The system can analyze the deviations and abnormalities of the frequency converter in real time during the frequency conversion process, and provide maintenance guidance for the abnormal process of converting direct current into alternating current of the required frequency according to the comparison results, to ensure timely and accurate abnormal maintenance. Through this precise maintenance, the system can effectively prevent the occurrence of faults, reduce downtime and maintenance costs, and improve the operating efficiency and life of the frequency converter. When the expected compliance signal appears, the system prompts that the maintenance is successful, to ensure the effectiveness and reliability of the maintenance process, to further ensure the normal operation of the train ventilation and air-conditioning system, and to improve the overall operating efficiency and passenger satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0052] Figure 1 The present invention is a method flow chart of a terminal security intelligent monitoring method based on the Internet of Things.

[0053] Figure 2 The present invention is a module schematic diagram of a terminal security intelligent monitoring system based on the Internet of Things. DETAILED DESCRIPTION

[0054] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0055] The present invention provides Figure 1 The terminal security intelligent monitoring method based on the Internet of Things shown includes the following steps:

[0056] During the operation of the inverter, the conversion data of the rectified DC power entering the inverter and being converted into the AC power of the required frequency through the insulated gate bipolar transistor is obtained in real time;

[0057] During the operation of the inverter, the conversion data of the rectified DC power entering the inverter and being converted into the required frequency AC power through the insulated gate bipolar transistor is obtained in real time, including multiple key parameters. These parameters mainly include input and output voltage, current value, output frequency, switching state of the insulated gate bipolar transistor, temperature, and power loss during the conversion process. Among them, the input and output voltage and current values ​​can reflect the load condition and power transmission efficiency of the circuit, and the output frequency is directly related to the speed and ventilation effect of the fan. The switching state monitoring of the insulated gate bipolar transistor is to evaluate whether it is working properly, and the switching speed and switching loss are key performance indicators. In addition, temperature data can help monitor the heat dissipation of the insulated gate bipolar transistor and other electronic components to prevent overheating failures.

[0058] Extract features from the converted data and perform anomaly analysis on the extracted features. Input the feature data after anomaly analysis into a pre-trained deep learning model, and use the deep learning model to evaluate the process of converting DC power into AC power of the required frequency and identify potential anomalies.

[0059] Features are extracted from the conversion data, the extracted features include circuit harmonics in the AC power and the switching time of the insulated gate bipolar transistor. After abnormal analysis of the circuit harmonics in the AC power and the switching time of the insulated gate bipolar transistor, harmonic distortion reference values ​​and switching time difference reference values ​​are generated. The harmonic distortion reference values ​​and switching time difference reference values ​​after abnormal analysis are input into a pre-trained deep learning model to generate an AC conversion coefficient. The AC conversion coefficient is used to evaluate the process of converting DC power into AC power of the required frequency to identify potential conversion anomalies.

[0060] When the rectified DC is converted to AC of the required frequency through an IGBT, if there is harmonic distortion in the AC, it will cause potential abnormalities in the frequency conversion of the inverter, making it impossible to perform efficient frequency conversion. Harmonic distortion is the waveform distortion of high-order harmonic currents and voltages caused by nonlinear loads or switching actions. These harmonics will interfere with the normal operation of the inverter. Specifically, harmonic distortion will increase the switching loss and electromagnetic interference of the IGBT, leading to thermal management problems and equipment overheating, which in turn affects the reliability and life of the IGBT. In addition, harmonic distortion will lead to a decrease in power quality, increase reactive power in the power grid, reduce system efficiency, and ultimately affect the control accuracy and stability of the inverter. Due to these factors, harmonic distortion will cause the inverter to be unable to effectively adjust the output frequency and voltage, unable to meet the load requirements, and may even cause system failures in severe cases.

[0061] After performing anomaly analysis on the circuit harmonics in the AC power, the steps to generate the harmonic distortion reference value are as follows:

[0062] In the detection window, the AC waveform is sampled at high frequency to obtain a set of discrete data points. The sampled signal is expressed as ,in n =1, 2, 3, 4, ..., N -1, N is a positive integer;

[0063] Calculate the root mean square value of the original AC signal. The calculation expression is: , where Represents the RMS value of the original AC signal;

[0064] Use a bandpass filter bank to separate the harmonic components of different frequencies and k The sampled values ​​of the subharmonic components are To express;

[0065] Calculate the root mean square value of each harmonic component. The calculation expression is: , where Indicates kThe RMS value of the subharmonics, k Indicates the number or order of harmonics;

[0066] Calculate the total harmonic distortion, the expression is: ,in, is the fundamental component, K is the upper limit of the harmonic order considered, is the total harmonic distortion;

[0067] Calculate the harmonic distortion reference value. The calculation expression is: , where Indicates the harmonic distortion reference value, Represents weighted total harmonic distortion, and the calculation expression is: ,in, Show weight function (introducing weight function is to amplify the influence of higher harmonics). .

[0068] It can be seen from the harmonic distortion reference value that when the rectified DC power is converted into AC power of the required frequency through the insulated gate bipolar transistor, within the detection window, the larger the harmonic distortion reference value generated after abnormal analysis of the circuit harmonics in the AC power, the greater the hidden danger of potential abnormalities in this process. This means that the insulated gate bipolar transistor may be subject to more harmonic interference during the conversion process, resulting in a decrease in the output power quality, reduced system efficiency, and may even cause problems such as equipment overheating and increased switching losses. Conversely, the smaller the harmonic distortion reference value, the less harmonic interference, the more stable and efficient the insulated gate bipolar transistor is in the process of converting DC power to AC power, and the less hidden danger of potential abnormalities.

[0069] When the rectified DC power enters the inverter and is converted into AC power of the required frequency through the insulated gate bipolar transistor, the switching time of the insulated gate bipolar transistor varies greatly, which will lead to potential abnormalities in the frequency conversion of the inverter and make it impossible to perform efficient frequency conversion. Specifically, the switching time of the insulated gate bipolar transistor includes the rise time, fall time, turn-on delay time and turn-off delay time. If these time parameters vary greatly in each switching cycle, it will directly affect the switching behavior of the insulated gate bipolar transistor and the output performance of the inverter. First, the difference in switching time will cause the asymmetry of the voltage and current waveforms, affecting the precise control of the inverter output AC power. Since frequency conversion depends on the precise switching operation of the insulated gate bipolar transistor, the inconsistency of the switching time will cause fluctuations and instability in the output frequency, and the required precise AC waveform cannot be generated.

[0070] Secondly, large differences in switching time will also cause a decrease in power conversion efficiency. Ideally, IGBTs should minimize switching losses when switching quickly, but due to inconsistent switching times, switching losses increase, resulting in additional heat generation. Switching losses include turn-on losses and turn-off losses. When the switching time difference is large, these losses will be unevenly distributed, which may cause local overheating and further reduce the overall efficiency. In the long run, this uneven heat distribution will accelerate the aging and failure of IGBTs and shorten the life of the equipment. In addition, when the switching time difference is large, the response speed of IGBTs is inconsistent, which will also have an adverse effect on the dynamic performance of the inverter. The inverter needs to respond quickly to load changes to maintain stable operation, but the switching time difference will delay or advance the switching action of some IGBTs, resulting in a slow or uncoordinated response of the system to load changes, and the output frequency and power cannot be adjusted in time, affecting the dynamic performance and stability of the system. Finally, large switching time differences may make it difficult to optimize the control algorithm of the inverter. The inverter's control system relies on preset switching time parameters to adjust the output frequency and voltage. However, when the actual switching time deviates from these parameters, the control system has difficulty effectively compensating for these deviations, resulting in reduced control accuracy and inability to achieve efficient frequency conversion.

[0071] In summary, the large difference in the switching time of the insulated gate bipolar transistor will lead to potential anomalies in the frequency conversion of the inverter, affecting the power conversion efficiency, dynamic performance and control accuracy, and failing to achieve efficient frequency conversion.

[0072] After performing abnormal analysis on the switching time of an insulated gate bipolar transistor in AC power, the steps for generating a reference value of the switching time difference are as follows:

[0073] Collect the switching duration data of IGBT in real time within the detection window, including rise time , Fall time , Turn-on delay time and turn-off delay time ;

[0074] Use the median absolute deviation to remove outliers. The specific formula is: ,in, For the collected switching time data, the criteria for judging abnormal values ​​are: , for those that meet this condition , identified as outliers and removed from the data set, where Indicates the threshold set when judging outliers, usually 1.4826 (a constant under approximate normal distribution);

[0075] Short-time Fourier transform is used to analyze the time-varying characteristics of the switching duration, extract the instantaneous frequency and amplitude changes, and set is the switching time signal, then the expression of short-time Fourier transform is ,in, Represents the time t and frequency f The signal on The Fourier transform result is specifically explained as It is the spectrum representation of the signal in a specific time window, that is, at a certain time t, different frequency components f The amplitude and phase information of is a window function, such as a Hamming window, which is used to capture a small segment of the signal at time t. is the original signal, is a complex exponential function, which is used to perform Fourier transform on the signal. j is an imaginary unit;

[0076] Extract features from the short-time Fourier transform results, including instantaneous frequency offset and amplitude change, and define instantaneous frequency offset and amplitude changes for: , , Indicates that at time t, the signal After short-time Fourier transform processing, the frequency f The amplitude of Indicates that at time point t-1, the signal After short-time Fourier transform processing, the frequency f The amplitude of

[0077] The characteristic entropy is used to calculate the uncertainty index of the switching time difference. The characteristic entropy represents the uncertainty of the system state. is the probability distribution of instantaneous frequency offset and amplitude change, then the characteristic entropy H The calculation expression is: ;

[0078] It should be noted that It can be estimated by the normalized histogram of instantaneous frequency offset and amplitude variation.

[0079] Comprehensive feature entropy H , instantaneous frequency deviation And the amplitude changes , calculate the reference value of the switching time difference, the calculation expression is: , where Indicates the reference value of the switching time difference, Indicates the instantaneous frequency deviation The result of summing within the detection window is: Indicates the change in amplitude The result of summing within the detection window is: , , They are H , , The weight coefficient of , , , It can be adjusted according to the actual application.

[0080] It can be seen from the switching time difference reference value that the larger the performance value of the switching time difference reference value, the greater the potential abnormal hidden danger of the insulated gate bipolar transistor when converting the rectified direct current into the required frequency alternating current, and vice versa. The specific explanation is that the switching time difference reference value quantifies the degree of inconsistency and fluctuation of the insulated gate bipolar transistor during the switching process by analyzing the switching time of the insulated gate bipolar transistor in the detection window. When the switching time difference reference value is large, it means that there are significant differences and instabilities in the switching time of the insulated gate bipolar transistor, which may lead to inaccurate frequency conversion and reduced efficiency, thereby increasing the risk of system failure and performance degradation; on the contrary, when the switching time difference reference value is small, it means that the switching time of the insulated gate bipolar transistor is more consistent and stable, the frequency conversion process is smoother, the system operation is more reliable, and the potential abnormal hidden danger is smaller.

[0081] The deep learning model is not specifically limited here, and can achieve the harmonic distortion reference value and switching time difference reference value Perform comprehensive analysis to generate AC conversion coefficients In order to realize the technical solution of the present invention, the present invention provides a specific implementation method;

[0082] AC conversion factor The resulting calculation formula is:

[0083]

[0084] , where , They are harmonic distortion reference values and switching time difference reference value The preset scaling factor of , Both are greater than 0.

[0085] It can be seen from the AC conversion coefficient that within the detection window, the larger the harmonic distortion reference value and the switching time difference reference value generated when the rectified DC power enters the inverter and is converted into the required frequency AC power through the insulated gate bipolar transistor, that is, the larger the performance value of the generated AC conversion coefficient, the greater the potential abnormal risk of the insulated gate bipolar transistor when converting the rectified DC power into the required frequency AC power. Conversely, the smaller the potential abnormal risk of the insulated gate bipolar transistor when converting the rectified DC power into the required frequency AC power.

[0086] Within the detection window, the AC conversion coefficient generated when the rectified DC power enters the inverter and is converted into the required frequency AC power through the insulated gate bipolar transistor is compared and analyzed with the preset AC conversion coefficient reference threshold. The results of the comparison and analysis are as follows:

[0087] If the AC conversion coefficient is greater than or equal to the AC conversion coefficient reference threshold, a potential abnormality signal is generated, indicating that there is a potential abnormality when the rectified DC power enters the inverter and is converted into the required frequency AC power through the insulated gate bipolar transistor;

[0088] If the AC conversion coefficient is less than the AC conversion coefficient reference threshold, a normal conversion signal is generated, indicating that the rectified DC power can be efficiently converted when it enters the inverter and is converted into AC power of the required frequency through the insulated gate bipolar transistor.

[0089] After identifying the potential anomaly, the actual frequency conversion curve of the frequency converter when performing frequency conversion within a period of time is obtained, and the expected frequency curve required when the frequency converter performs frequency conversion within the same period of time is obtained;

[0090] The actual frequency conversion curve of the inverter during frequency conversion can be obtained by installing a high-precision data acquisition system inside the inverter. The system includes devices such as voltage sensors, current sensors, and frequency meters to monitor the voltage and current signals at the output of the inverter in real time. Through the data of these sensors, the output frequency of the inverter at different time points can be accurately recorded using sampling and data processing technology. In order to ensure the accuracy and reliability of the data, the data acquisition system needs to have high sampling rate and low noise characteristics, and be able to work stably in a complex electromagnetic environment. At the same time, the data recording device should have sufficient storage capacity and computing power to process and store large amounts of real-time data. By subsequent analysis of these data, the frequency conversion curve of the inverter during operation can be drawn to show the frequency changes of the inverter under different loads and working conditions.

[0091] The expected frequency curve required for the frequency converter to perform frequency conversion is usually determined by the design parameters and operating requirements of the fan. First, it is necessary to refer to the technical specifications and performance curves provided by the fan manufacturer. These documents usually contain the performance data of the fan at different operating points (such as different speeds, air volumes, and wind pressure conditions). Combined with the actual needs of the train air-conditioning system, engineers can use simulation software to perform detailed system simulations to simulate the operating performance of the fan under different environmental conditions and loads. Through simulation and experimental data, the expected frequency curve of the fan under various working conditions is drawn. In addition, the expected frequency curve can be optimized by statistically analyzing the historical operating data of the fan in combination with the actual operating data to more accurately reflect the performance requirements of the fan in the actual operating environment. The expected frequency curve obtained in this way can provide an important reference for the optimized control and performance evaluation of the fan.

[0092] Compare and analyze the actual frequency conversion curve with the expected frequency curve, and conduct real-time analysis of the deviations and anomalies in the frequency conversion process of the inverter;

[0093] Compare and analyze the actual frequency conversion curve with the expected frequency curve to generate the expected deviation index. The specific steps are as follows:

[0094] Get the actual frequency conversion curve and the expected frequency curve, and use the actual frequency conversion curve Indicates that the expected frequency curve is expressed as express, , ,in, and Respectively at the same time point The actual and expected frequency values ​​on m is the number of sampling points, u An index representing a time point;

[0095] The frequency data is normalized (to make the data comparable on the same scale), and the normalization formula is: , ,in, and are respectively the actual frequency value and the expected frequency value after normalization, Indicates the minimum value in the actual frequency conversion curve (referring to the actual frequency conversion curve over a period of time). Indicates the maximum value in the actual frequency conversion curve (referring to the actual frequency conversion curve over a period of time). Similarly, Indicates the minimum value in the expected frequency curve (referring to the expected frequency curve over a period of time). Indicates the maximum value in the expected frequency curve (referring to the expected frequency curve over a period of time);

[0096] Calculate the cosine similarity between the actual frequency conversion curve and the expected frequency curve. The calculation expression is:

[0097] , where is the cosine similarity, with a value range of [-1, 1];

[0098] Calculate the expected deviation index, the calculation expression is: , where Represents the expected deviation index.

[0099] It can be seen from the expected deviation index that the larger the performance value of the expected deviation index, the greater the deviation between the actual frequency conversion curve when the inverter performs frequency conversion and the expected frequency curve required when the inverter performs frequency conversion, which means that the potential abnormality of the insulated gate bipolar transistor when converting the rectified DC power to the required frequency AC power is more serious. A larger expected deviation index means that the error and distortion in the frequency conversion process are greater, which makes the inverter unable to perform efficient frequency conversion, seriously affecting the stability and efficiency of the system.

[0100] The expected deviation index generated by comparing the actual frequency conversion curve with the expected frequency curve is compared with the preset expected deviation index reference threshold. The results of the comparison analysis are as follows:

[0101] If the expected deviation index is greater than or equal to the expected deviation index reference threshold, a deviation signal is generated, indicating that there is a deviation between the actual frequency conversion curve and the expected frequency curve that exceeds the allowable range and needs to be processed in time;

[0102] If the expected deviation index is less than the expected deviation index reference threshold, an expected compliance signal is generated, indicating that the deviation between the actual frequency conversion curve and the expected frequency curve is within the allowable range and does not need to be processed, which may be a sudden accidental situation.

[0103] Provide maintenance guidance for abnormal processes when converting DC power into AC power of required frequency based on the results of comparison and analysis, and perform abnormal maintenance in a timely manner;

[0104] The degree of deviation in the frequency conversion process reflected by the expected deviation index is used to guide the maintenance and optimization of the inverter and the IGBT. The larger the expected deviation index, the greater the deviation between the actual frequency conversion curve and the expected frequency curve, which means that there is a more serious abnormality in the system. By analyzing the size and trend of the expected deviation index, specific abnormal conditions, such as harmonic distortion, abnormal IGBT switching time, etc., can be identified and their severity can be determined. Based on this information, maintenance personnel can take corresponding measures, such as checking and replacing faulty IGBTs, adjusting control parameters, optimizing the cooling system, and improving filter performance, to restore and improve the performance and stability of the inverter. At the same time, regular monitoring and analysis of the expected deviation index can prevent potential faults, extend the service life of equipment, and ensure that the inverter and fan system operate in the best condition, thereby achieving an efficient and reliable DC to AC conversion process. This data-driven maintenance strategy helps to detect and solve problems in a timely manner, reduce downtime, and improve the overall efficiency and safety of the system.

[0105] During the maintenance process, when the expected deviation index is less than the expected deviation index reference threshold, that is, when the expected compliance signal is generated, it indicates that the maintenance is successful.

[0106] The present invention utilizes a deep learning model to perform feature extraction and anomaly analysis on conversion data acquired in real time, accurately identifying potential abnormalities of the frequency converter during the frequency conversion process, solving the defect of the prior art that potential abnormalities cannot be intelligently identified, ensuring that the frequency converter can promptly detect and handle abnormalities during operation, and preventing the fan from affecting the ventilation and air-conditioning system in the train due to the inability to adjust the speed, thereby improving the stability of the system and the comfort of passengers.

[0107] After identifying potential abnormalities, the present invention generates an expected deviation index by acquiring and comparing the actual frequency conversion curve with the expected frequency curve, and compares it with the expected deviation index reference threshold. The system can analyze the deviations and abnormalities of the frequency converter in real time during the frequency conversion process, and provide maintenance guidance for the abnormal process of converting direct current into alternating current of the required frequency according to the comparison results, to ensure timely and accurate abnormal maintenance. Through this precise maintenance, the system can effectively prevent the occurrence of faults, reduce downtime and maintenance costs, and improve the operating efficiency and life of the frequency converter. When the expected compliance signal appears, the system prompts that the maintenance is successful, to ensure the effectiveness and reliability of the maintenance process, to further ensure the normal operation of the train ventilation and air-conditioning system, and to improve the overall operating efficiency and passenger satisfaction.

[0108] The present invention provides Figure 2The terminal security intelligent monitoring system based on the Internet of Things shown includes a data acquisition module, a feature extraction and anomaly detection module, a frequency curve acquisition module, a comparison and deviation analysis module, and a maintenance guidance module;

[0109] The data acquisition module acquires the conversion data of the rectified direct current into the required frequency alternating current through the insulated gate bipolar transistor after entering the inverter in real time during the operation of the inverter;

[0110] The feature extraction and anomaly detection module extracts features from the converted data and performs anomaly analysis on the extracted features. The feature data after anomaly analysis is input into a pre-trained deep learning model, and the deep learning model is used to evaluate the process of converting DC power into AC power of the required frequency and identify potential anomalies.

[0111] The frequency curve acquisition module, after identifying the potential anomaly, acquires the actual frequency conversion curve when the frequency converter performs frequency conversion within a period of time, and simultaneously acquires the expected frequency curve required when the frequency converter performs frequency conversion within the same period of time;

[0112] The comparison and deviation analysis module compares and analyzes the actual frequency conversion curve with the expected frequency curve, and performs real-time analysis on the deviations and anomalies in the frequency conversion process of the inverter;

[0113] The maintenance guidance module provides maintenance guidance for abnormal processes when converting DC power into AC power of required frequency based on the results of comparison and analysis, and performs abnormal maintenance in a timely manner;

[0114] An embodiment of the present invention provides a terminal security intelligent monitoring method based on the Internet of Things, which is implemented by the above-mentioned terminal security intelligent monitoring system based on the Internet of Things. The specific method and process of a terminal security intelligent monitoring system based on the Internet of Things are detailed in the embodiment of the above-mentioned terminal security intelligent monitoring method based on the Internet of Things, which will not be repeated here.

[0115] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0116] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0117] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A terminal security intelligent monitoring method based on the Internet of Things, characterized in that: The following steps are involved: During the operation of the inverter, the conversion data of the rectified DC power entering the inverter and being converted into the AC power of the required frequency through the insulated gate bipolar transistor is obtained in real time; Extract features from the converted data and perform anomaly analysis on the extracted features. Input the feature data after anomaly analysis into a pre-trained deep learning model, and use the deep learning model to evaluate the process of converting DC power into AC power of the required frequency and identify potential anomalies. Extract features from the conversion data, the extracted features include circuit harmonics in the alternating current and the switching time of the insulated gate bipolar transistor, generate harmonic distortion reference values ​​and switching time difference reference values ​​after abnormal analysis of the circuit harmonics in the alternating current and the switching time of the insulated gate bipolar transistor, input the harmonic distortion reference values ​​and switching time difference reference values ​​after abnormal analysis into the following formula to generate an alternating current conversion coefficient, evaluate the process of converting direct current into alternating current of the required frequency through the alternating current conversion coefficient, and identify potential conversion abnormalities; AC conversion factor The resulting calculation formula is: , where , They are harmonic distortion reference values and switching time difference reference value The preset scaling factor of , All are greater than 0; After identifying the potential anomaly, the actual frequency conversion curve of the frequency converter when performing frequency conversion within a period of time is obtained, and the expected frequency curve required when the frequency converter performs frequency conversion within the same period of time is obtained; Compare and analyze the actual frequency conversion curve with the expected frequency curve, and conduct real-time analysis of the deviations and anomalies in the frequency conversion process of the inverter; Compare and analyze the actual frequency conversion curve with the expected frequency curve to generate the expected deviation index. The specific steps are as follows: Get the actual frequency conversion curve and the expected frequency curve, and use the actual frequency conversion curve Indicates that the expected frequency curve is expressed as express, , ,in, and Respectively at the same time point The actual and expected frequency values ​​on m is the number of sampling points, u An index representing a time point; The frequency data is normalized, and the normalization formula is: , ,in, and are respectively the actual frequency value and the expected frequency value after normalization, Indicates the minimum value in the actual frequency conversion curve, Indicates the maximum value in the actual frequency conversion curve. Similarly, represents the minimum value in the expected frequency curve, represents the maximum value in the expected frequency curve; Calculate the cosine similarity between the actual frequency conversion curve and the expected frequency curve. The calculation expression is: , where is the cosine similarity, with a value range of [-1, 1]; Calculate the expected deviation index, the calculation expression is: In the formula, represents the expected deviation index; Based on the results of the comparison and analysis, maintenance guidance is provided for the abnormal process when DC power is converted into AC power of the required frequency, and abnormal maintenance is performed in a timely manner.

2. According to the method of claim 1, the terminal security intelligent monitoring method based on the Internet of Things is characterized in that: After performing anomaly analysis on the circuit harmonics in the AC power, the steps to generate the harmonic distortion reference value are as follows: In the detection window, the AC waveform is sampled at high frequency to obtain a set of discrete data points. The sampled signal is expressed as ,in n =1, 2, 3, 4, ..., N -1, N is a positive integer; Calculate the root mean square value of the original AC signal. The calculation expression is: , where Represents the RMS value of the original AC signal; Use a bandpass filter bank to separate the harmonic components of different frequencies and k The sampled values ​​of the subharmonic components are To express; Calculate the root mean square value of each harmonic component. The calculation expression is: , where Indicates k The RMS value of the subharmonics, k Indicates the number or order of harmonics; Calculate the total harmonic distortion, the expression is: ,in, is the fundamental component, K is the upper limit of the harmonic order considered, is the total harmonic distortion; Calculate the harmonic distortion reference value. The calculation expression is: , where Indicates the harmonic distortion reference value, Represents weighted total harmonic distortion, and the calculation expression is: ,in, represents the weight function, .

3. According to the method of claim 1, the terminal security intelligent monitoring method based on the Internet of Things is characterized in that: After performing abnormal analysis on the switching time of an insulated gate bipolar transistor in AC power, the steps for generating a reference value of the switching time difference are as follows: Collect the switching duration data of IGBT in real time within the detection window, including rise time , Fall time , Turn-on delay time and turn-off delay time ; Use the median absolute deviation to remove outliers. The specific formula is: ,in, For the collected switching time data, the criteria for judging abnormal values ​​are: , for those that meet this condition , identified as outliers and removed from the data set, where Indicates the threshold value set when judging abnormal values; Short-time Fourier transform is used to analyze the time-varying characteristics of the switching duration, extract the instantaneous frequency and amplitude changes, and set is the switching time signal, then the expression of short-time Fourier transform is ,in, Represents the time t and frequency f The signal on The Fourier transform result is specifically explained as It is the spectrum representation of the signal in the time window, that is, at a certain time t, different frequency components f The amplitude and phase information of is a window function used to capture a small segment of the signal at time t. is the original signal, is a complex exponential function, which is used to perform Fourier transform on the signal. j is an imaginary unit; Extract features from the short-time Fourier transform results, including instantaneous frequency offset and amplitude change, and define instantaneous frequency offset and amplitude changes for: , , Indicates that at time t, the signal After short-time Fourier transform processing, the frequency f The amplitude of Indicates that at time point t-1, the signal After short-time Fourier transform processing, the frequency f The amplitude of The characteristic entropy is used to calculate the uncertainty index of the switching time difference. The characteristic entropy represents the uncertainty of the system state. is the probability distribution of instantaneous frequency offset and amplitude change, then the characteristic entropy H The calculation expression is: ; Comprehensive feature entropy H , instantaneous frequency deviation And the amplitude changes , calculate the reference value of the switching time difference, the calculation expression is: , where Indicates the reference value of the switching time difference, Indicates the instantaneous frequency deviation The result of summing within the detection window is: Indicates the change in amplitude The result of summing within the detection window is: , , They are H , , The weight coefficient of .

4. The method for intelligent terminal security monitoring based on the Internet of Things according to claim 1 is characterized in that: Within the detection window, the AC conversion coefficient generated when the rectified DC power enters the inverter and is converted into the required frequency AC power through the insulated gate bipolar transistor is compared and analyzed with the preset AC conversion coefficient reference threshold. The results of the comparison and analysis are as follows: If the AC conversion factor is greater than or equal to the AC conversion factor reference threshold, a potential abnormal signal is generated; If the AC conversion coefficient is less than the AC conversion coefficient reference threshold, a normal conversion signal is generated.

5. The method for intelligent terminal security monitoring based on the Internet of Things according to claim 4 is characterized in that: The expected deviation index generated by comparing the actual frequency conversion curve with the expected frequency curve is compared with the preset expected deviation index reference threshold. The results of the comparison analysis are as follows: If the expected deviation index is greater than or equal to the expected deviation index reference threshold, a deviation signal is generated, indicating that there is a deviation between the actual frequency conversion curve and the expected frequency curve that exceeds the allowable range and needs to be processed in time; If the expected deviation index is less than the expected deviation index reference threshold, an expected compliance signal is generated, indicating that the deviation between the actual frequency conversion curve and the expected frequency curve is within the allowable range and no processing is required.

6. A terminal security intelligent monitoring system based on the Internet of Things, used to implement the terminal security intelligent monitoring method based on the Internet of Things as described in any one of claims 1 to 5, characterized in that: It includes data acquisition module, feature extraction and anomaly detection module, frequency curve acquisition module, comparison and deviation analysis module and maintenance guidance module; The data acquisition module acquires the conversion data of the rectified direct current into the required frequency alternating current through the insulated gate bipolar transistor after entering the inverter in real time during the operation of the inverter; The feature extraction and anomaly detection module extracts features from the converted data and performs anomaly analysis on the extracted features. The feature data after anomaly analysis is input into a pre-trained deep learning model, and the deep learning model is used to evaluate the process of converting DC power into AC power of the required frequency and identify potential anomalies. The frequency curve acquisition module, after identifying the potential anomaly, acquires the actual frequency conversion curve when the frequency converter performs frequency conversion within a period of time, and simultaneously acquires the expected frequency curve required when the frequency converter performs frequency conversion within the same period of time; The comparison and deviation analysis module compares and analyzes the actual frequency conversion curve with the expected frequency curve, and performs real-time analysis on the deviations and anomalies in the frequency conversion process of the inverter; The maintenance guidance module provides maintenance guidance for abnormal processes when converting DC power into AC power of required frequency based on the results of comparison and analysis, and performs abnormal maintenance in a timely manner.

Images