Method and device for tracing real-time monitoring of electromagnetic interference spectrum of frequency converter

Abstract

The application provides a variable frequency converter electromagnetic interference spectrum real-time monitoring traceability method and device, relates to the technical field of variable frequency converters, and comprises the following steps: establishing a mapping relationship between variable frequency converter operating state parameters and electromagnetic interference data of magnetic leakage points; taking the variable frequency converter operating state parameters as an excitation source to inject an electromagnetic simulation space, and generating predicted interference spectrum and electromagnetic field distribution information in real time; according to the mapping relationship, the predicted interference spectrum and electromagnetic field distribution information are matched and compared with electromagnetic interference data, and are compared with a preset interference fingerprint library; according to the interference spectrum characteristics, the spatial magnetic field distribution of the electromagnetic interference data is combined to locate the field strength spatial coordinates, and based on the interference spectrum characteristics and the field strength spatial coordinates, as well as the predicted interference spectrum and electromagnetic field distribution information, an interference spectrum traceability result is generated. Through the application, the technical problem of low variable frequency converter electromagnetic interference recognition accuracy in the prior art can be solved, and the technical effect of improving electromagnetic interference recognition accuracy is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of frequency converters, and particularly relates to a method and device for real-time monitoring and tracing of electromagnetic interference spectrum of a frequency converter. BACKGROUND

[0002] With the rapid development of power electronics technology and intelligent manufacturing technology, frequency converters, as the core devices for regulating motor speed and improving energy efficiency, are widely used in industrial automation, rail transportation, new energy equipment, high-end manufacturing and other fields. However, due to the high-frequency switching characteristics of power devices and the complexity of circuit topology, strong electromagnetic interference signals are generated during the operation of the frequency converter.

[0003] At present, the existing electromagnetic interference monitoring and analysis methods mainly rely on traditional spectrum analyzers, near-field probes and test receivers and other means. Such methods can accurately obtain the electromagnetic interference signals of the frequency converter in the laboratory environment, but have obvious limitations in complex industrial sites. First, traditional methods often require manual point-by-point scanning, which is low in efficiency and difficult to realize real-time dynamic monitoring. Secondly, such methods can only obtain the frequency domain information of the interference signal, and lack effective association with the running state and spatial distribution characteristics of the frequency converter, which leads to the inability to fully reflect the formation mechanism and propagation path of the interference.

[0004] In summary, the existing technology has the technical problem that due to the dependence of electromagnetic interference monitoring means on manual operation and the lack of multi-dimensional data fusion, real-time and accurate identification of interference sources and analysis of propagation paths cannot be realized in complex industrial environments, further affecting the electromagnetic compatibility improvement and overall operation safety and stability of the frequency converter system. SUMMARY

[0005] The purpose of the present application is to provide a method and device for real-time monitoring and tracing of electromagnetic interference spectrum of a frequency converter, to solve the technical problem in the prior art that due to the dependence of electromagnetic interference monitoring means on manual operation and the lack of multi-dimensional data fusion, real-time and accurate identification of interference sources and analysis of propagation paths cannot be realized in complex industrial environments, further affecting the electromagnetic compatibility improvement and overall operation safety and stability of the frequency converter system.

[0006] In view of the above problems, the present application provides a method and device for real-time monitoring and tracing of electromagnetic interference spectrum of a frequency converter.

[0007] In a first aspect, the application provides a variable frequency drive electromagnetic interference spectrum real-time monitoring and tracing method, which is realized by a variable frequency drive electromagnetic interference spectrum real-time monitoring and tracing device, and includes the following steps: synchronously collecting variable frequency drive operation state parameters and electromagnetic interference data of leakage points, establishing a mapping relationship between the variable frequency drive operation state parameters and the electromagnetic interference data according to time stamps; constructing an electromagnetic simulation space based on the physical structure and circuit parameters of the variable frequency drive, and injecting the variable frequency drive operation state parameters into the electromagnetic simulation space as an excitation source to generate predicted interference spectrum and electromagnetic field distribution information in real time; according to the mapping relationship, matching and comparing the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data, and comparing them with a preset interference fingerprint library to identify interference spectrum characteristics; and positioning the field strength spatial coordinates based on the spatial magnetic field distribution of the electromagnetic interference data according to the interference spectrum characteristics, and generating an interference spectrum tracing result based on the interference spectrum characteristics and the field strength spatial coordinates, as well as the predicted interference spectrum and electromagnetic field distribution information.

[0008] Preferably, the variable frequency drive electromagnetic interference spectrum real-time monitoring and tracing method further includes: analyzing electromagnetic energy leakage paths of input power supply lines, output motor cable lines and external connected devices of the variable frequency drive with the variable frequency drive position as the center, identifying leakage magnetic field transmission probabilities and leakage magnetic coefficients of path elements; grading the importance of the path elements according to the leakage magnetic field transmission probabilities and the leakage magnetic coefficients, and planning the layout density and layout distance step length of the sensors according to the grading results, wherein the higher the element grade is, the higher the layout density is and the smaller the distance step length is; deploying a sensor array according to the layout planning scheme, and recording three-dimensional spatial coordinates of each sensor to construct a sensor layout positioning coordinate system.

[0009] Preferably, the variable frequency drive electromagnetic interference spectrum real-time monitoring and tracing method further includes: collecting operation state parameters of the variable frequency drive in real time, including at least output frequency, load torque and DC bus voltage; synchronously acquiring electromagnetic interference data collected by the sensor array; applying synchronous time stamps to all operation state parameters and electromagnetic interference data based on a unified clock source; according to the time stamps, performing space-time mapping binding on operation state parameters, electromagnetic interference data and corresponding sensor spatial coordinates of interference electromagnetic data at the same time, constructing a three-dimensional mapping relationship, and taking the three-dimensional mapping relationship as the mapping relationship between the variable frequency drive operation state parameters and the electromagnetic interference data.

[0010] Preferably, the variable frequency converter electromagnetic interference spectrum real-time monitoring tracing method further comprises: synchronously collecting operation state parameters of two or more variable frequency converters, and collecting mixed electromagnetic interference monitoring signals through electromagnetic sensors arranged at at least two different spatial positions; performing signal demixing on the mixed electromagnetic interference monitoring signals through a blind source separation algorithm to obtain a plurality of independent interference components corresponding to potential interference data; and performing interference source positioning on the demixed plurality of independent interference components by using preset characteristic changes of a target variable frequency converter to locate spectrum characteristic changes, and establishing a mapping relationship between the variable frequency converter operation state parameters and the electromagnetic interference data.

[0011] Preferably, the variable frequency converter electromagnetic interference spectrum real-time monitoring tracing method further comprises: controlling the operation state parameters of a target variable frequency converter to have preset characteristic changes, and monitoring spectrum changes of the independent interference components, wherein the preset characteristic changes have an identification spectrum characteristic; and performing spectrum change identification on the independent interference components by using the identification spectrum characteristic, and marking the independent interference components having corresponding spectrum changes in the spectrum as originating from the target variable frequency converter.

[0012] Preferably, the variable frequency converter electromagnetic interference spectrum real-time monitoring tracing method further comprises: constructing a signal matrix from the mixed electromagnetic interference monitoring signals collected by the electromagnetic sensors at different spatial positions; inputting the signal matrix into a blind source separation model, performing separation matrix optimization based on the signal matrix through the blind source separation model, solving a source signal estimation matrix, and outputting each row vector in the source signal estimation matrix as the plurality of independent interference components.

[0013] Preferably, the variable frequency converter electromagnetic interference spectrum real-time monitoring tracing method further comprises: performing spatial alignment on the electromagnetic interference data by using a predicted interference spectrum and electromagnetic field distribution information, performing frame-by-frame difference comparison based on the aligned interference spectrum data to obtain a difference comparison result including an abnormal frequency band; and performing similarity comparison on the spectrum data of the abnormal frequency band and reference spectrum in the preset interference fingerprint library, and determining a spectrum characteristic satisfying a comparison confidence requirement as the identified interference spectrum characteristic.

[0014] Preferably, the real-time monitoring and tracing method for electromagnetic interference spectrum of frequency converters further includes: simultaneously collecting interference spectrum data, spatial field strength distribution data, and frequency converter operating condition data at a known interference source, and associating them with the physical fault information of the interference source to establish a multimodal data sample; extracting spectral features, spatial location features, and operating condition features from the multimodal data sample, and labeling them with corresponding fault type tags to construct an initial interference fingerprint; injecting the physical fault through digital twin model simulation to generate simulated spectrum and spatial distribution data, and performing data enhancement and expansion on the initial interference fingerprint; storing multiple initial interference fingerprints in a database to obtain the preset interference fingerprint database; wherein, the difference spectral features are identified by comparing the interference fingerprints between faults, and the incremental confidence spectrum labeling of the interference fingerprint is performed using the difference spectral features.

[0015] Preferably, the real-time monitoring and tracing method for electromagnetic interference spectrum of the frequency converter further includes: extracting sensor monitoring data deployed at different spatial locations within the abnormal frequency band according to the abnormal frequency band corresponding to the interference spectrum characteristics, and forming a magnetic field strength observation vector; establishing a field strength distribution likelihood function with the spatial coordinates and emission intensity of the potential interference source as variables based on the electromagnetic wave propagation attenuation characteristics; using the magnetic field strength observation vector as input and the field strength distribution likelihood function as a basis, constructing a maximum likelihood estimation function for iterative solution, solving for the optimal spatial coordinate solution that maximizes the function value, and obtaining the field strength spatial coordinates, which are used to represent the spatial location of the physical interference source corresponding to the interference spectrum characteristics.

[0016] Secondly, this application also provides a real-time monitoring and tracing device for electromagnetic interference spectrum of frequency converters, used to execute the real-time monitoring and tracing method for electromagnetic interference spectrum of frequency converters as described in the first aspect, including: a mapping relationship establishment module, used to synchronously collect frequency converter operating status parameters and electromagnetic interference data at leakage magnetic points, and establish a mapping relationship between frequency converter operating status parameters and electromagnetic interference data according to the collection timestamp; an information generation module, used to construct an electromagnetic simulation space based on the physical structure and circuit parameters of the frequency converter, and inject the frequency converter operating status parameters as an excitation source into the electromagnetic simulation space to generate predicted interference spectrum and electromagnetic field distribution information in real time; an interference spectrum feature identification module, used to match and compare the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data according to the mapping relationship, and compare with a preset interference fingerprint database to identify interference spectrum features; and an interference spectrum tracing result generation module, used to locate the field strength spatial coordinates based on the interference spectrum features and the spatial magnetic field distribution of the electromagnetic interference data, and generate interference spectrum tracing results based on the interference spectrum features, the field strength spatial coordinates, and the predicted interference spectrum and electromagnetic field distribution information.

[0017] The technical solution provided in this application has at least the following technical effects or advantages: by achieving the technical goal of establishing a multi-dimensional mapping relationship between operating status parameters and electromagnetic interference data and constructing a visual source tracing model, it achieves the technical effect of being able to quickly locate electromagnetic interference sources, accurately analyze propagation paths, and provide decision-making basis for interference suppression under complex operating conditions.

[0018] The above description is merely an overview of the technical solution of this application. To enable a clearer understanding of the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the method for real-time monitoring and tracing the electromagnetic interference spectrum of frequency converters in this application.

[0021] Figure 2 This is a schematic diagram of the structure of the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter in this application.

[0022] Figure labeling: Module 1 for establishing mapping relationship, Module 2 for information generation, Module 3 for interference spectrum feature identification, and Module 4 for generating interference spectrum source tracing results. Detailed Implementation

[0023] This application provides a method and device for real-time monitoring and tracing the electromagnetic interference spectrum of frequency converters. It addresses the technical problem in existing technologies where electromagnetic interference monitoring relies on manual operation and lacks multi-dimensional data fusion, making it impossible to achieve real-time and accurate interference source identification and propagation path analysis in complex industrial environments. This further impacts the electromagnetic compatibility improvement and overall operational safety and stability of frequency converter systems. The application achieves the technical goal of establishing a multi-dimensional mapping relationship between operating status parameters and electromagnetic interference data and constructing a visualized tracing model. This enables rapid location of electromagnetic interference sources, accurate analysis of propagation paths, and provision of decision-making basis for interference suppression under complex operating conditions.

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0025] Example 1, please refer to the appendix. Figure 1 This application provides a method for real-time monitoring and tracing the electromagnetic interference spectrum of a frequency converter, which is applied to a real-time monitoring and tracing device for electromagnetic interference spectrum of a frequency converter, and specifically includes the following steps:

[0026] S1: Synchronously collect inverter operating status parameters and electromagnetic interference data at leakage magnetic points, and establish a mapping relationship between inverter operating status parameters and electromagnetic interference data according to the collection timestamps.

[0027] Specifically, synchronously acquiring inverter operating status parameters and electromagnetic interference data at leakage points means simultaneously recording the inverter's operating status information and electromagnetic interference signals collected by surrounding sensors at the same time. Inverter operating status parameters include key indicators such as output frequency, load torque, and DC bus voltage, used to describe the inverter's operation at a given moment. Electromagnetic interference data at leakage points refers to interference signals collected at locations around the inverter where magnetic fields may leak, such as a sensor measuring a magnetic field strength of 0.3 Tesla at a certain point. Simultaneous acquisition ensures that operating status and interference data remain consistent over time.

[0028] By aligning the data acquisition timestamps, a mapping relationship is established between the inverter's operating status parameters and electromagnetic interference data. This involves adding precise timestamps to all acquired data, enabling matching of different types of data at the same time. For example, if the inverter's output frequency is 50 Hz and its DC bus voltage is 600 volts, while the sensor acquires a magnetic field strength of 0.3 Tesla, establishing a mapping relationship between the two through timestamp alignment helps researchers understand the electromagnetic interference characteristics under different operating conditions.

[0029] S2: Construct an electromagnetic simulation space based on the physical structure and circuit parameters of the frequency converter, and inject the operating state parameters of the frequency converter as an excitation source into the electromagnetic simulation space to generate predicted interference spectrum and electromagnetic field distribution information in real time.

[0030] Specifically, an electromagnetic simulation space is constructed based on the physical structure and circuit parameters of the frequency converter. This involves building a virtual electromagnetic environment in a computer based on the actual hardware components and electrical characteristics of the frequency converter. The physical structure includes the casing, power modules, input / output terminals, and cable routing, while the circuit parameters encompass values ​​such as resistance, inductance, capacitance, and switching frequency. The electromagnetic simulation space is used to simulate the propagation characteristics of electromagnetic waves inside and outside the frequency converter; for example, a magnetic field leakage point may form at a cable bend.

[0031] The inverter's operating status parameters are injected into the electromagnetic simulation space as an excitation source. This means that key parameters measured during actual operation are input into the simulation model as driving conditions for electromagnetic disturbances. Operating status parameters include output frequency, load torque, and DC bus voltage. For example, when the output frequency is 50 Hz, the load torque is 20 Nm, and the DC bus voltage is 600 V, the simulation model is driven to calculate the electromagnetic field distribution. The injection of the excitation source applies realistic energy input to the model, thereby generating electromagnetic interference behavior consistent with actual conditions. After inputting the parameters, the intensity distribution of electromagnetic waves at different frequencies and the field strength distribution in space are quickly output, generating predicted interference spectrum and electromagnetic field distribution information in real time. The interference spectrum refers to the frequency domain performance of the interference signal, while the electromagnetic field distribution information shows the distribution of magnetic or electric field strength at different locations.

[0032] S3: Based on the mapping relationship, the predicted interference spectrum and electromagnetic field distribution information are matched and compared with the electromagnetic interference data, and compared with the preset interference fingerprint database to identify the interference spectrum characteristics.

[0033] Specifically, by establishing a spatiotemporal mapping relationship, the simulation-generated predictions and sensor-measured data are analyzed at the same time and spatial location. The mapping relationship binds operating parameters, electromagnetic interference signals, and spatial coordinates together. The predicted interference spectrum represents the distribution of interference signals at different frequencies, while the electromagnetic field distribution information refers to the distribution of electric and magnetic field strengths at different points in space. Through matching and comparison, differences or consistency between predictions and measurements can be identified. For example, in a 200 Hz frequency band, the predicted interference intensity is 0.25 Tesla, while the measured intensity is 0.27 Tesla. Comparison can determine whether the difference is within a reasonable range.

[0034] The matched spectrum data is then compared with existing interference fingerprints in the database. The preset interference fingerprint database stores the characteristic spectra of various typical faults and interference sources; each fingerprint contains frequency characteristics, amplitude characteristics, and possible spatial characteristics. The database then determines whether the actually collected spectrum characteristics match those already in the database. If the similarity reaches a certain threshold, the type of interference can be confirmed.

[0035] S4: Based on the interference spectrum characteristics, combined with the spatial magnetic field distribution of electromagnetic interference data, locate the spatial coordinates of the field strength, and generate interference spectrum source tracing results based on the interference spectrum characteristics, the spatial coordinates of the field strength, and the predicted interference spectrum and electromagnetic field distribution information.

[0036] Specifically, based on the interference spectrum characteristics and the spatial magnetic field distribution of electromagnetic interference data, the spatial coordinates of the field strength are located. That is, based on the identified interference spectrum characteristics, the frequency bands exhibiting abnormal behavior in the spectrum are found, and then, combined with the magnetic field strength data measured by sensors at different locations, the specific spatial location of the interference source is calculated. Interference spectrum characteristics refer to signal patterns that exhibit abnormal enhancement or changes at specific frequencies, while spatial magnetic field distribution refers to the distribution of magnetic field strength collected at multiple locations.

[0037] Based on the interference spectrum characteristics and field strength spatial coordinates, as well as the predicted interference spectrum and electromagnetic field distribution information, interference spectrum source tracing results are generated. That is, after determining the spatial location of the interference source, the spatial location is combined with the interference spectrum characteristics, and then compared with the interference spectrum and electromagnetic field distribution results predicted by simulation, thus obtaining a complete interference source tracing analysis. The predicted interference spectrum and electromagnetic field distribution information comes from the electromagnetic simulation space, providing a theoretical reference for actual observation. The interference spectrum source tracing results combine observation, location, and simulation to determine the source of the interference signal and its formation cause.

[0038] Furthermore, this application also includes: taking the inverter location as the center, analyzing the electromagnetic energy leakage paths of its input power line, output motor cable, and external connection equipment, and identifying the leakage flux transmission probability and leakage flux coefficient of each path element; classifying the importance of the path elements according to the leakage flux transmission probability and leakage flux coefficient, and planning the sensor deployment density and deployment distance step size according to the classification results, wherein the higher the element level, the higher the deployment density and the smaller the distance step size; deploying the sensor array according to the deployment plan, and recording the three-dimensional spatial coordinates of each sensor to construct a sensor deployment positioning coordinate system.

[0039] Specifically, taking the physical installation point of the frequency converter as a reference point, the analysis focuses on the power input lines of the frequency converter, the output cables of the drive motor, and the external equipment connected to the frequency converter. Electromagnetic energy leakage paths refer to the channels through which electromagnetic waves or magnetic fields may diffuse outward along wires, joints, or equipment housings. This helps identify lines or equipment that easily radiate or transmit electromagnetic interference. The probability of leakage flux propagation indicates the likelihood that a component will leak its internal magnetic field into the external space, while the leakage flux coefficient reflects the strength of the magnetic field leakage.

[0040] Then, the importance of path elements is graded based on the probability of electromagnetic leakage propagation and the electromagnetic leakage coefficient, classifying them into different levels, such as high, medium, and low, thereby identifying the more critical elements in the electromagnetic interference tracing process. Planning the sensor deployment density determines the spatial density of the sensors, while the deployment distance step size represents the distance between adjacent sensors. A higher path element level indicates a greater contribution to electromagnetic leakage, corresponding to higher monitoring requirements; therefore, a higher sensor deployment density and a smaller step size are needed.

[0041] Next, after completing the deployment plan, the sensor array needs to be deployed. This involves placing multiple sensors in a three-dimensional space according to certain rules and layouts to simultaneously monitor electromagnetic data from different directions and locations. The three-dimensional spatial coordinates of each sensor need to be recorded. For example, the position of a sensor might be 0.5 meters in the x-direction, 0.3 meters in the y-direction, and 1 meter in the z-direction. By recording the position data, a sensor deployment positioning coordinate system can be constructed to identify the monitoring point position corresponding to each sensor, ensuring that the collected interference data accurately corresponds to the physical spatial location.

[0042] Furthermore, this application also includes: real-time acquisition of the inverter's operating status parameters, including at least output frequency, load torque, and DC bus voltage; synchronous acquisition of electromagnetic interference data collected by the sensor array; spoofing all operating status parameters and electromagnetic interference data with a unified clock source; and, based on the timestamps, spatiotemporally mapping and binding the operating status parameters, electromagnetic interference data, and sensor spatial coordinates corresponding to the interference electromagnetic data at the same moment to construct a three-dimensional mapping relationship, which serves as the mapping relationship between the inverter's operating status parameters and electromagnetic interference data.

[0043] Specifically, the inverter's operating status parameters are collected in real time, including output frequency, load torque, and DC bus voltage. Output frequency represents the frequency at which the inverter supplies electrical energy to the motor; for example, an output frequency of 50 Hz corresponds to the motor's standard operating condition. Load torque is the torque required by the motor during operation, directly reflecting the load size; for example, the torque may be only 5 Nm under no-load conditions, but could rise to 50 Nm under full load. DC bus voltage is the voltage value of the inverter's internal DC circuit, determining the energy basis for the entire power conversion.

[0044] Next, electromagnetic interference data collected by the sensor array is acquired simultaneously. This means that multiple sensors distributed around the inverter simultaneously collect electromagnetic signals and record the electromagnetic interference generated by the inverter during operation. A sensor array is a collection of sensors arranged in space according to certain rules, which monitors the interference field strength at different points in space at the same time. For example, deploying three sensors on the left, above, and directly in front of the inverter can acquire interference data from different directions in three-dimensional space.

[0045] Then, based on a unified clock source, all operating status parameters and electromagnetic interference data are timestamped synchronously. The unified clock source can be a high-precision clock signal generator, ensuring that all collected data is recorded under the same time base. The timestamp is the time identifier attached to the data recording. For example, at 0.01 seconds, the inverter's output frequency is 45 Hz, and sensor 1 records an electromagnetic interference intensity of 0.2 Tesla. Thus, all data can be matched one-to-one in chronological order.

[0046] Finally, based on the timestamp, the operating status parameters, electromagnetic interference data, and sensor spatial coordinates at the same moment are bound together. That is, within the same time dimension, a corresponding relationship is established between the operating status data and the electromagnetic interference data at the spatial location, and then these data are extended into a three-dimensional mapping relationship through three-dimensional spatial coordinates. For example, at 0.05 seconds, the output frequency is 48 Hz, and the corresponding interference intensity at the coordinates x = 1 meter, y = 0.5 meters, z = 0.3 meters is recorded as 0.3 Tesla.

[0047] Furthermore, this application also includes: synchronously acquiring operating status parameters of two or more frequency converters, and acquiring mixed electromagnetic interference monitoring signals through electromagnetic sensors deployed at at least two different spatial locations; demixing the mixed electromagnetic interference monitoring signals using a blind source separation algorithm to obtain several independent interference components corresponding to potential interference data; using preset characteristic changes of the target frequency converter to locate spectral characteristic changes, identifying the interference sources of the demixed several independent interference components, and establishing a mapping relationship between the frequency converter operating status parameters and electromagnetic interference data.

[0048] Specifically, the system synchronously collects operating status parameters from two or more frequency converters, meaning it records key operating information from multiple frequency converters simultaneously within the same time period, including output frequency, load torque, and DC bus voltage. It also collects mixed electromagnetic interference monitoring signals using electromagnetic sensors deployed at at least two different spatial locations. For example, sensors are installed at different locations around the frequency converters, such as one on the left and one above. Each sensor simultaneously collects electromagnetic signals from its own location. Since multiple frequency converters operate simultaneously, the collected signals are a mixture of interference sources.

[0049] Then, the mixed electromagnetic interference monitoring signal is demixed using a blind source separation algorithm. This involves separating the signals from multiple superimposed interference sources using statistical methods, thereby obtaining the independent interference component corresponding to each potential interference source. The blind source separation algorithm does not rely on prior information about the interference sources; it separates different signals by analyzing the statistical independence between them. For example, if two frequency converters simultaneously generate interference, one in the 100 Hz band and the other in the 150 Hz band, blind source separation can yield independent signal components at 100 Hz and 150 Hz respectively.

[0050] Next, the target inverter's preset characteristic changes are used to locate spectral characteristic changes. This is achieved by actively adjusting the target inverter's operating parameters, such as changing the output frequency or load torque, to produce identifiable characteristic changes in the spectrum. The response of independent interference components is then observed. The sources of the several independent interference components obtained after demixing are identified to determine the inverter to which the interference component belongs. This establishes a mapping relationship between inverter operating state parameters and electromagnetic interference data, meaning that at a specific time point and frequency, the inverter's operating state corresponds one-to-one with the electromagnetic interference it generates.

[0051] Furthermore, this application also includes: controlling the operating state parameters of the target frequency converter to undergo preset characteristic changes, monitoring the spectral changes of the independent interference components, wherein the preset characteristic changes have spectral identification features; using the spectral identification features to identify the spectral changes of the independent interference components, and marking the independent interference components that show corresponding spectral changes in the spectrum as originating from the target frequency converter.

[0052] Specifically, the operating parameters of the target frequency converter are controlled to undergo preset characteristic changes, such as adjusting the output frequency, load torque, or DC bus voltage, so that the operating parameters produce identifiable changes in the electromagnetic interference spectrum. Preset characteristic changes refer to pre-planned parameter adjustment amplitudes or patterns. Monitoring the spectral changes of independent interference components involves observing the frequency domain response of the independent interference signals obtained through blind source separation while adjusting the frequency converter parameters. For example, a frequency band with an original amplitude of 0.2 Tesla may increase to 0.25 Tesla after adjustment, thus forming identifiable spectral characteristics.

[0053] Then, the independent interference components are identified by utilizing spectral characteristics. This involves matching the monitored spectral changes with preset features to pinpoint how the independent interference components respond to adjustments in the inverter's parameters. Spectral change identification means analyzing the amplitude changes of the signal along the frequency axis. For example, if the amplitude increases from 0.2 Tesla to 0.25 Tesla in the 50 Hz band, the corresponding independent interference components can be identified. Independent interference components exhibiting corresponding spectral changes are marked as originating from the target inverter, confirming that the independent interference components are generated by the actively adjusted inverter, rather than from other equipment or environmental noise.

[0054] Furthermore, this application also includes: constructing a signal matrix by collecting multiple mixed electromagnetic interference signal measurement signals from electromagnetic sensors at different spatial locations; inputting the signal matrix into a blind source separation model, optimizing the separation matrix based on the signal matrix through the blind source separation model, solving the source signal estimation matrix, and outputting each row vector in the source signal estimation matrix as the several independent interference components.

[0055] Specifically, a signal matrix is ​​constructed by combining multiple mixed electromagnetic interference signals collected by electromagnetic sensors from different spatial locations. This involves arranging the electromagnetic signals collected by each sensor within the same time period into a two-dimensional table, where each row represents the data sequence of one sensor, and each column represents the collected values ​​from all sensors at the same moment. The presence of multiple mixed electromagnetic interference signals means that due to the simultaneous presence of multiple frequency converters or interference sources, the signal collected by each sensor is the result of the superposition of interference signals. For example, the signal amplitudes collected by three sensors at a certain moment might be 0.2 Tesla, 0.3 Tesla, and 0.25 Tesla, respectively. Arranging these values ​​in a matrix forms the signal matrix.

[0056] Then, the signal matrix is ​​input into the blind source separation model and provided to the algorithm model for processing. The blind source separation model is a mathematical and statistical tool that uses a large number of historical mixed electromagnetic interference signals and known independent interference source components as training samples. It is trained on a preset separation network with the goal of maximizing the statistical independence of the output components, and is able to separate independent source signals from mixed signals.

[0057] A blind source separation model is a separation matrix W capable of performing separation calculations. It can be obtained through two methods: direct solution based on traditional optimization algorithms and offline training followed by deployment of a neural network model. The method based on direct solution using traditional optimization algorithms is for each new batch of observed signals X(t). An optimization algorithm (such as FastICA or IVA) is run in real-time on-site to calculate a separation matrix W suitable for the current measurement. When new data X(t) is input, a random matrix is ​​initialized, and then iterative calculation is performed based on natural gradient descent until convergence is obtained. The method based on offline training followed by deployment of a neural network model involves collecting mixed signals X(t) under various operating conditions, while simultaneously measuring or simulating real, clean source signals S(t) as "labeled data." A network structure is then selected, such as a separation network based on an encoder-decoder structure. The error between the separated output and the real source signal (e.g., mean square error) is used as the loss function. The network is iteratively trained with preset data, and the network parameters are adjusted to minimize the loss function. Finally, the observed signal X(t) is input into the network, and the separation result Š(t) is output.

[0058] Each row in the source signal estimation matrix corresponds to an independent sequence of interference components. The vectors of each row in the source signal estimation matrix are output as several independent interference components. That is, each row of the matrix is ​​treated as a time-series signal of an independent interference source. For example, if the amplitude of the original mixed signal increases from 0.2 Tesla to 0.5 Tesla, after separation, the first row might represent an increase in the interference signal of strain gauge A from 0.1 Tesla to 0.3 Tesla, and the second row might represent an increase in the interference signal of strain gauge B from 0.1 Tesla to 0.2 Tesla, thus achieving the differentiation of interference sources.

[0059] Furthermore, this application also includes: spatially aligning the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data; performing frame-by-frame differential comparison based on the aligned interference spectrum data to obtain differential comparison results, including abnormal frequency bands; and performing similarity comparison between the spectrum data of the abnormal frequency bands and the reference spectrum in the preset interference fingerprint database, and determining the spectrum features that meet the comparison confidence requirements as the identified interference spectrum features.

[0060] Specifically, spatial alignment is achieved by using the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data. This involves matching the predicted results obtained through simulation with the actual data collected by sensors in terms of spatial location and frequency distribution. The interference spectrum refers to the intensity distribution of electromagnetic interference signals within different frequency ranges, while the electromagnetic field distribution information refers to the electric or magnetic field strength at different spatial points. Spatial alignment ensures a one-to-one correspondence between the virtual simulation results and the actual measurement results. For example, at a point 0.5 meters from the inverter, the simulated predicted magnetic field is 0.3 Tesla, while the actual sensor measurement is 0.28 Tesla. Spatial alignment facilitates subsequent discrepancy analysis.

[0061] Frame-by-frame differential comparison is performed based on the aligned interference spectrum data to obtain the differential comparison results, including anomalous frequency bands. This involves comparing the predicted and measured results frame by frame along the time axis and calculating the differences. Frame-by-frame comparison, which compares the video frame by frame, reflects the differences in interference over time. The differential comparison results will mark frequency ranges with significant differences as anomalous frequency bands.

[0062] The spectral data of the abnormal frequency band is compared with the reference spectrum in a pre-set interference fingerprint database. This involves comparing the identified abnormal frequency band with stored typical interference characteristics. The interference fingerprint database is a database that stores the characteristic spectra of various known interference sources. The similarity comparison is used to determine whether the abnormal frequency band matches known interference. For example, if the abnormal frequency band is between 200 Hz and 250 Hz and has 90% similarity to an interference fingerprint in the database caused by cable insulation aging, it indicates that the abnormality may be related to that fault.

[0063] Spectral features that meet the comparison confidence requirements are identified as interference spectral features. That is, only when the similarity reaches a certain confidence level can an abnormal frequency band be confirmed to belong to a certain interference source. The confidence requirement is the confidence threshold, for example, set at 90%. If the comparison result is higher than 90%, the frequency band is considered to indeed originate from a specific interference source; otherwise, it is only used as reference information. For example, when an abnormal frequency band has a 95% similarity to the fingerprint of a known fault, it can be confirmed that it is a manifestation of that fault.

[0064] Furthermore, this application also includes: simultaneously collecting interference spectrum data, spatial field strength distribution data, and inverter operating condition data on a known interference source, and associating them with the physical fault information of the interference source to establish a multimodal data sample; extracting spectral features, spatial location features, and operating condition features from the multimodal data sample, and labeling them with corresponding fault type tags to construct an initial interference fingerprint; injecting the physical fault through digital twin model simulation to generate simulated spectrum and spatial distribution data, and performing data enhancement and expansion on the initial interference fingerprint; storing multiple initial interference fingerprints in a database to obtain the preset interference fingerprint database; wherein, based on the comparison and identification of difference spectral features between interference fingerprints of fault obstacles, incremental confidence spectrum labeling of interference fingerprints is performed using the difference spectral features.

[0065] Specifically, interference spectrum data, spatial field strength distribution data, and inverter operating condition data are simultaneously collected from known interference sources. That is, when the existence of an interference source is known, the interference behavior in the frequency domain, the spatial distribution of electromagnetic field strength, and the inverter's operating status are simultaneously acquired. Interference spectrum data is a curve describing the energy distribution of the interference at different frequencies; spatial field strength distribution data is the intensity of the magnetic or electric field at different locations; and inverter operating condition data includes output frequency, load size, and voltage. Correlating the interference spectrum data, spatial field strength distribution data, and inverter operating condition data with the physical fault information of the interference source can form a multi-dimensional joint sample, serving as a multi-modal data sample. For example, in the case of a cable insulation failure, the corresponding spectral anomalies, the location of the spatial magnetic field enhancement, and the corresponding operating load conditions can be recorded.

[0066] Spectral features, spatial location features, and operating condition features are extracted from multimodal data samples and labeled with corresponding fault type tags; this process involves feature extraction from the collected data. Spectral features refer to significant peaks or abnormal frequency bands in the frequency distribution, spatial location features are the spatial variation patterns of the interference field strength, and operating condition features are the parameter performance of the frequency converter under specific operating conditions. By mapping the extracted features to the actual fault type, a complete interference feature description can be obtained, serving as an initial interference fingerprint that represents the characteristics of the fault.

[0067] By simulating physical faults using digital twin models, simulated spectral and spatial distribution data are generated, enhancing and expanding the initial interference fingerprint. Specifically, digital twin technology maps a virtual model of the frequency converter to the actual equipment and simulates the same physical fault, thus obtaining more spectral and spatial distribution data. This allows for verification of the initial interference fingerprint and also expands the sample size, improving its generalization ability. For example, if an anomaly is known to occur in the 10Hz to 20Hz range, the simulation can simulate the response of this frequency band under different load conditions, obtaining multiple spectral representations of varying intensities, making the initial fingerprint more complete.

[0068] Multiple initial interference fingerprints are stored in a database to obtain a preset interference fingerprint database. This involves centrally storing the established and expanded interference fingerprints to form a database. This database contains spectral characteristics, spatial characteristics, and operating condition characteristics corresponding to different fault types for subsequent identification.

[0069] By comparing and identifying the spectral differences between interference fingerprints from different faults, incremental confidence spectral annotation of the interference fingerprints is performed. This involves comparing existing fingerprints to identify the spectral differences. These differences allow for more precise annotation of the interference fingerprints, increasing the confidence level of fingerprint recognition. For example, in two types of faults, one may have a peak at 200 Hz, and the other at 250 Hz. Comparison can identify these differences, enhancing the database's discriminative power.

[0070] Furthermore, this application also includes: extracting sensor monitoring data deployed at different spatial locations within the abnormal frequency band according to the abnormal frequency band corresponding to the interference spectrum characteristics, and constructing a magnetic field strength observation vector; establishing a field strength distribution likelihood function with the spatial coordinates and emission intensity of the potential interference source as variables based on the electromagnetic wave propagation attenuation characteristics; constructing a maximum likelihood estimation function based on the magnetic field strength observation vector as input and the field strength distribution likelihood function, and iteratively solving the optimal spatial coordinate solution that maximizes the function value, thereby obtaining the field strength spatial coordinates, which are used to represent the spatial location of the physical interference source corresponding to the interference spectrum characteristics.

[0071] Specifically, based on the abnormal frequency bands corresponding to the characteristics of the interference spectrum, monitoring data from sensors deployed at different spatial locations within these abnormal frequency bands are extracted to form a magnetic field strength observation vector. That is, after identifying the abnormal frequency bands in the interference spectrum, the magnetic field strength values ​​for those bands are extracted from data collected by sensors distributed at different spatial points and combined into a vector. An abnormal frequency band refers to a frequency range in the spectrum exhibiting abnormal enhancement or attenuation, while the sensor monitoring data consists of the field strength values ​​measured by each sensor within that frequency band. The magnetic field strength observation vector is formed by arranging and combining scattered data into a mathematical vector. For example, in the 200 Hz abnormal frequency band, if sensors at positions of 1 meter, 2 meters, and 3 meters measure field strengths of 0.3 Tesla, 0.2 Tesla, and 0.1 Tesla, respectively, this constitutes an observation vector.

[0072] Based on the attenuation characteristics of electromagnetic wave propagation, a likelihood function for the field strength distribution is established, using the spatial coordinates and emission intensity of potential interference sources as variables. This utilizes the physical law that electromagnetic wave intensity decreases with increasing distance during propagation to build a mathematical model. The attenuation characteristics of electromagnetic wave propagation indicate that electromagnetic field strength is inversely proportional to the square of the distance or a more complex function. Spatial coordinates represent the possible location parameters of the potential interference source, while emission intensity represents the initial energy of the source signal. The field strength distribution likelihood function estimates the probability of a certain field strength distribution occurring at different locations using multiple variables. For example, if 0.3 Tesla is observed at a distance of 1 meter, the theoretical prediction at a distance of 2 meters should be 0.15 Tesla.

[0073] Using the observed magnetic field strength vector as input and the field strength distribution likelihood function as a foundation, a maximum likelihood estimation function is constructed for iterative solution. This involves substituting the observed sensor data into the likelihood function and using maximum likelihood estimation to find the parameter solution that best matches the observed results. Maximum likelihood estimation is a statistical method that adjusts variables in the model to maximize the probability of the observed data occurring. Iterative solution means continuously updating the hypothesized source location and emission intensity, gradually approximating the true solution. For example, initially assuming the interference source is at 2 meters, the predicted value differs significantly from the observed value. After multiple iterations, when the position is adjusted to 1.2 meters, the function value reaches its maximum, indicating that the source location is likely at 1.2 meters.

[0074] Solving for the optimal spatial coordinate solution that maximizes the function value yields the field strength spatial coordinates, which represent the spatial location of the physical interference source corresponding to the interference spectrum characteristics. In other words, the optimal solution obtained through iteration can ultimately determine the location of the interference source in space. The field strength spatial coordinates are the three-dimensional coordinate values ​​of the optimal solution, corresponding to the actual spatial location of the interference source, thereby identifying interference characteristics on the spectrum and determining the location of the interference source.

[0075] In summary, the real-time monitoring and tracing method for electromagnetic interference spectrum of frequency converters provided in this application has the following technical effects: by achieving the technical goal of establishing a multi-dimensional mapping relationship between operating status parameters and electromagnetic interference data and constructing a visual tracing model, it can quickly locate the electromagnetic interference source under complex operating conditions, accurately analyze the propagation path, and provide a basis for decision-making for interference suppression.

[0076] Example 2: Based on the same inventive concept as the real-time monitoring and tracing method for electromagnetic interference spectrum of frequency converters in the foregoing examples, this application also provides a real-time monitoring and tracing device for electromagnetic interference spectrum of frequency converters. Please refer to the appendix. Figure 2The system includes: a mapping relationship establishment module 1, used to synchronously collect inverter operating status parameters and electromagnetic interference data at leakage magnetic points, and establish a mapping relationship between inverter operating status parameters and electromagnetic interference data according to the collection timestamps; an information generation module 2, used to construct an electromagnetic simulation space based on the physical structure and circuit parameters of the inverter, and inject the inverter operating status parameters as an excitation source into the electromagnetic simulation space to generate predicted interference spectrum and electromagnetic field distribution information in real time; an interference spectrum feature identification module 3, used to match and compare the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data according to the mapping relationship, and compare it with a preset interference fingerprint database to identify interference spectrum features; and an interference spectrum tracing result generation module 4, used to locate the field strength spatial coordinates based on the interference spectrum features and the spatial magnetic field distribution of the electromagnetic interference data, and generate interference spectrum tracing results based on the interference spectrum features, field strength spatial coordinates, and predicted interference spectrum and electromagnetic field distribution information.

[0077] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter is also used to: analyze the electromagnetic energy leakage paths of the input power line, output motor cable and external connected equipment centered on the location of the frequency converter, and identify the leakage flux transmission probability and leakage flux coefficient of each path element; classify the importance of the path elements according to the leakage flux transmission probability and leakage flux coefficient, and plan the sensor deployment density and deployment distance step size according to the classification results, wherein the higher the element level, the higher the deployment density and the smaller the distance step size; deploy the sensor array according to the deployment plan, and record the three-dimensional spatial coordinates of each sensor to construct a sensor deployment positioning coordinate system.

[0078] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter is also used for: real-time acquisition of the operating status parameters of the frequency converter, including at least the output frequency, load torque and DC bus voltage; synchronous acquisition of electromagnetic interference data collected by the sensor array; spoofing all operating status parameters and electromagnetic interference data with a unified clock source; and, based on the timestamps, spatiotemporally mapping and binding the operating status parameters, electromagnetic interference data and the sensor spatial coordinates corresponding to the interference electromagnetic data at the same moment to construct a three-dimensional mapping relationship, which serves as the mapping relationship between the operating status parameters and electromagnetic interference data of the frequency converter.

[0079] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter is also used for: synchronously collecting operating status parameters of two or more frequency converters, and collecting mixed electromagnetic interference monitoring signals through electromagnetic sensors deployed at least two different spatial locations; demixing the mixed electromagnetic interference monitoring signals using a blind source separation algorithm to obtain several independent interference components corresponding to potential interference data; locating spectral feature changes using preset feature changes of the target frequency converter, identifying the source of interference in the demixed several independent interference components, and establishing a mapping relationship between the frequency converter operating status parameters and electromagnetic interference data.

[0080] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the inverter is also used to: control the operating status parameters of the target inverter to undergo preset characteristic changes, monitor the spectrum changes of the independent interference components, wherein the preset characteristic changes have spectral identification features; use the spectral identification features to identify the spectrum changes of the independent interference components, and mark the independent interference components that show corresponding spectrum changes in the spectrum as originating from the target inverter.

[0081] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter is also used to: construct a signal matrix from multiple mixed electromagnetic interference signal measurement signals collected by electromagnetic sensors from different spatial locations; input the signal matrix into a blind source separation model, optimize the separation matrix based on the signal matrix through the blind source separation model, solve the source signal estimation matrix, and output each row vector in the source signal estimation matrix as the several independent interference components.

[0082] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter is also used for: spatially aligning the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data; performing frame-by-frame differential comparison based on the aligned interference spectrum data to obtain differential comparison results, including abnormal frequency bands; and performing similarity comparison between the spectrum data of the abnormal frequency bands and the reference spectrum in the preset interference fingerprint database, and determining the spectrum features that meet the comparison confidence requirements as the identified interference spectrum features.

[0083] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter is also used for: synchronously collecting interference spectrum data, spatial field strength distribution data, and frequency converter operating condition data on a known interference source, and associating them with the physical fault information of the interference source to establish a multi-modal data sample; extracting spectral features, spatial location features, and operating condition features from the multi-modal data sample, and labeling them with corresponding fault type tags to construct an initial interference fingerprint; simulating and injecting the physical fault through a digital twin model to generate simulated spectrum and spatial distribution data, and performing data enhancement and expansion on the initial interference fingerprint; storing multiple initial interference fingerprints in a database to obtain the preset interference fingerprint database; wherein, based on the comparison and identification of difference spectral features between interference fingerprints of fault obstacles, incremental confidence spectrum labeling of the interference fingerprint is performed using the difference spectral features.

[0084] Furthermore, the real-time monitoring and tracing device for electromagnetic interference spectrum of the frequency converter is also used for: extracting sensor monitoring data deployed at different spatial locations within the abnormal frequency band according to the abnormal frequency band corresponding to the interference spectrum characteristics, and forming a magnetic field strength observation vector; establishing a field strength distribution likelihood function with the spatial coordinates and emission intensity of the potential interference source as variables based on the electromagnetic wave propagation attenuation characteristics; constructing a maximum likelihood estimation function based on the magnetic field strength observation vector as input and the field strength distribution likelihood function, and iteratively solving the optimal spatial coordinate solution that maximizes the function value, thereby obtaining the field strength spatial coordinates, which are used to represent the spatial location of the physical interference source corresponding to the interference spectrum characteristics.

[0085] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The inverter electromagnetic interference spectrum real-time monitoring and tracing method and specific examples in the aforementioned embodiment one are also applicable to the inverter electromagnetic interference spectrum real-time monitoring and tracing device in this embodiment. Through the foregoing detailed description of the inverter electromagnetic interference spectrum real-time monitoring and tracing method, those skilled in the art can clearly understand the inverter electromagnetic interference spectrum real-time monitoring and tracing device in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0086] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0087] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for real-time monitoring and tracing the source of electromagnetic interference spectrum in frequency converters, characterized in that, include: Synchronously collect inverter operating status parameters and electromagnetic interference data at leakage magnetic points, and establish a mapping relationship between inverter operating status parameters and electromagnetic interference data according to the collection timestamps; An electromagnetic simulation space is constructed based on the physical structure and circuit parameters of the frequency converter, and the operating status parameters of the frequency converter are injected into the electromagnetic simulation space as an excitation source to generate predicted interference spectrum and electromagnetic field distribution information in real time. Based on the mapping relationship, the predicted interference spectrum and electromagnetic field distribution information are matched and compared with the electromagnetic interference data, and compared with a preset interference fingerprint database to identify interference spectrum characteristics. Based on the interference spectrum characteristics, the spatial coordinates of the field strength are located by combining the spatial magnetic field distribution of the electromagnetic interference data. Based on the interference spectrum characteristics and the field strength spatial coordinates, as well as the predicted interference spectrum and electromagnetic field distribution information, the interference spectrum source tracing result is generated. Synchronously collect inverter operating status parameters and electromagnetic interference data at leakage magnetic points, including: Taking the inverter location as the center, the electromagnetic energy leakage paths of its input power line, output motor cable and external connection equipment are analyzed to identify the leakage flux transmission probability and leakage flux coefficient of each path element. The importance of the path elements is classified according to the leakage flux transmission probability and leakage flux coefficient, and the deployment density and deployment distance step of the sensors are planned according to the classification results. The higher the element level, the higher the deployment density and the smaller the distance step. Deploy the sensor array according to the deployment plan, record the three-dimensional spatial coordinates of each sensor, and construct a sensor deployment positioning coordinate system; Synchronously collect inverter operating status parameters and electromagnetic interference data at leakage flux points, and establish a mapping relationship between inverter operating status parameters and electromagnetic interference data according to the collection timestamps, including: Real-time acquisition of the inverter's operating status parameters, including at least output frequency, load torque, and DC bus voltage; Simultaneously acquire electromagnetic interference data collected by the sensor array; All operating status parameters and electromagnetic interference data are time-stamped based on a unified clock source. Based on the timestamp, the operating status parameters, electromagnetic interference data, and sensor spatial coordinates corresponding to the interference electromagnetic data at the same moment are spatiotemporally mapped and bound to construct a three-dimensional mapping relationship, which serves as the mapping relationship between the inverter's operating status parameters and electromagnetic interference data. Establishing a mapping relationship between inverter operating status parameters and electromagnetic interference data based on the collected timestamps also includes: Simultaneously collect operating status parameters of two or more frequency converters, and collect mixed electromagnetic interference monitoring signals through electromagnetic sensors deployed in at least two different spatial locations; The hybrid electromagnetic interference monitoring signal is demixed using a blind source separation algorithm to obtain several independent interference components corresponding to the potential interference data. By utilizing the preset characteristic changes of the target frequency converter to locate the spectral characteristic changes, the interference sources of several independent interference components in the demixing are determined, and a mapping relationship between the operating status parameters of the frequency converter and electromagnetic interference data is established.

2. The method for real-time monitoring and tracing the electromagnetic interference spectrum of a frequency converter according to claim 1, characterized in that, By utilizing preset characteristic changes in the target frequency converter to locate spectral characteristic changes, the source of interference is determined for several independent interference components after demixing, including: The operating status parameters of the target frequency converter are controlled to undergo preset characteristic changes, and the spectral changes of the independent interference components are monitored, wherein the preset characteristic changes have spectral identification features; The spectral characteristics are used to identify the spectral changes of independent interference components, and the independent interference components that show corresponding spectral changes in the spectrum are marked as originating from the target frequency converter.

3. The method for real-time monitoring and tracing the electromagnetic interference spectrum of a frequency converter according to claim 1, characterized in that, The hybrid electromagnetic interference monitoring signal is demixed using a blind source separation algorithm to obtain several independent interference components corresponding to the potential interference data, including: The multi-channel mixed electromagnetic interference signal collection from electromagnetic sensors at different spatial locations is used to form a signal matrix; The signal matrix is ​​input into the blind source separation model. The blind source separation model optimizes the separation matrix based on the signal matrix to solve for the source signal estimation matrix. Each row vector in the source signal estimation matrix is ​​output as the several independent interference components.

4. The method for real-time monitoring and tracing the electromagnetic interference spectrum of a frequency converter according to claim 1, characterized in that, Based on the mapping relationship, the predicted interference spectrum and electromagnetic field distribution information are matched and compared with the electromagnetic interference data, and compared with a preset interference fingerprint database to identify interference spectrum characteristics, including: Spatially align the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data, and perform frame-by-frame differential comparison based on the aligned interference spectrum data to obtain differential comparison results, including abnormal frequency bands. The spectrum data of the abnormal frequency band is compared with the reference spectrum in the preset interference fingerprint database to determine the spectrum features that meet the comparison confidence requirements as the identified interference spectrum features.

5. The method for real-time monitoring and tracing the electromagnetic interference spectrum of a frequency converter according to claim 4, characterized in that, The spectral data of the abnormal frequency band is compared with the reference spectrum in the preset interference fingerprint database, which includes the following steps: At known interference sources, interference spectrum data, spatial field strength distribution data, and inverter operating condition data are collected synchronously and correlated with the physical fault information of the interference source to establish multimodal data samples. Spectral features, spatial location features, and operating condition features are extracted from the multimodal data samples, and corresponding fault type labels are labeled to construct an initial interference fingerprint; The physical fault is injected through digital twin model simulation, generating simulated spectrum and spatial distribution data, and the initial interference fingerprint is augmented and expanded. Multiple initial interference fingerprints are stored in a database to obtain the preset interference fingerprint database; Specifically, the differential spectral features are identified by comparing the interference fingerprints between faults and obstacles, and the incremental confidence spectrum labeling of the interference fingerprints is performed using the differential spectral features.

6. The method for real-time monitoring and tracing the electromagnetic interference spectrum of a frequency converter according to claim 4, characterized in that, Based on the aforementioned interference spectrum characteristics, and combined with the spatial magnetic field distribution of electromagnetic interference data, the spatial coordinates of the field strength are located, including: Based on the abnormal frequency bands corresponding to the interference spectrum characteristics, sensor monitoring data deployed at different spatial locations within the abnormal frequency bands are extracted to form a magnetic field strength observation vector. Based on the electromagnetic wave propagation attenuation characteristics, a field strength distribution likelihood function is established with the spatial coordinates and emission intensity of the potential interference source as variables. Using the magnetic field strength observation vector as input and the field strength distribution likelihood function as a basis, a maximum likelihood estimation function is constructed and iteratively solved to find the optimal spatial coordinate solution that maximizes the function value, thereby obtaining the field strength spatial coordinates, which are used to represent the spatial location of the physical interference source corresponding to the interference spectrum characteristics.

7. A real-time monitoring and tracing device for electromagnetic interference spectrum of frequency converters, characterized in that, The steps for implementing the real-time monitoring and tracing method for electromagnetic interference spectrum of a frequency converter according to any one of claims 1 to 6 include: The mapping relationship establishment module is used to synchronously collect inverter operating status parameters and electromagnetic interference data of leakage magnetic points, and establish a mapping relationship between inverter operating status parameters and electromagnetic interference data according to the collection timestamp. The information generation module is used to construct an electromagnetic simulation space based on the physical structure and circuit parameters of the frequency converter, and inject the operating status parameters of the frequency converter as an excitation source into the electromagnetic simulation space to generate predicted interference spectrum and electromagnetic field distribution information in real time. The interference spectrum feature identification module is used to match and compare the predicted interference spectrum and electromagnetic field distribution information with the electromagnetic interference data based on the mapping relationship, and compare it with a preset interference fingerprint database to identify interference spectrum features. The interference spectrum tracing result generation module is used to generate interference spectrum tracing results based on the interference spectrum characteristics, combined with the spatial magnetic field distribution of electromagnetic interference data to locate the field strength spatial coordinates, and based on the interference spectrum characteristics, field strength spatial coordinates, and predicted interference spectrum and electromagnetic field distribution information.

Images