Electromagnetic interference suppression method and device based on high-frequency GaN switching characteristics

Abstract

The invention relates to an electromagnetic interference suppression method and device based on high-frequency GaN switching characteristics. The method comprises: in a power conversion circuit constructed based on a GaN switching device, edge analysis processing is performed on a driving signal of the GaN switching device to obtain edge feature data, and the edge feature data comprises a rising edge change rate and a falling edge change rate in a plurality of continuous switching periods; determining an equivalent network model constructed based on parasitic parameters in the power conversion circuit, and performing frequency domain response calculation on the equivalent network model according to the edge feature data, interference frequency band distribution information caused by the edge change is obtained, analysis processing based on spectrum energy adjustment is carried out on a driving signal of the GaN switching device according to the interference frequency band distribution information, and an adjustment control instruction used for restraining key frequency band interference is obtained. By adopting the method, an effective suppression strategy oriented to a specific interference frequency band can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of high-frequency power conversion, and in particular to a method and device for suppressing electromagnetic interference based on high-frequency GaN switching characteristics. Background Art

[0002] In the field of high-frequency power conversion technology, electric energy regulation is involved based on GaN switching devices. Due to the high-speed switching and steep edge characteristics of GaN devices, they are prone to introduce additional high-frequency electromagnetic interference, and thus this interference needs to be effectively suppressed.

[0003] Related electromagnetic interference suppression methods usually configure filters or shielding structures at the output end to absorb and block interference energy. However, such methods rely on fixed physical structures and are difficult to dynamically adjust according to the interference frequency characteristics, resulting in limited interference control capabilities of the system within a specific frequency band. Summary of the Invention

[0004] Based on this, it is necessary to provide an electromagnetic interference suppression method, device, computer equipment and computer-readable storage medium based on high-frequency GaN switching characteristics to address the above technical problems, so as to implement an effective suppression strategy for specific interference frequency bands.

[0005] In a first aspect, the present application provides an electromagnetic interference suppression method based on high-frequency GaN switching characteristics, comprising: In a power conversion circuit constructed based on a GaN switching device, edge analysis processing is performed on a driving signal of the GaN switching device to obtain edge characteristic data, wherein the edge characteristic data includes a rising edge change rate and a falling edge change rate within multiple consecutive switching cycles; Determining an equivalent network model constructed based on parasitic parameters in the power conversion circuit, performing frequency domain response calculation on the equivalent network model according to the edge characteristic data, and obtaining interference frequency band distribution information caused by edge changes; According to the interference frequency band distribution information, the driving signal of the GaN switching device is analyzed and processed based on spectrum energy adjustment to obtain an adjustment control instruction for suppressing interference in a key frequency band.

[0006] In a second aspect, the present application further provides an electromagnetic interference suppression device based on high-frequency GaN switching characteristics, comprising: An edge analysis module is configured to perform edge analysis on a drive signal of a GaN switching device in a power conversion circuit based on the GaN switching device to obtain edge characteristic data, wherein the edge characteristic data includes a rising edge change rate and a falling edge change rate within a plurality of consecutive switching cycles; a response analysis module, configured to determine an equivalent network model constructed based on parasitic parameters in the power conversion circuit, perform frequency domain response calculation on the equivalent network model according to the edge characteristic data, and obtain interference frequency band distribution information caused by edge changes; An instruction generation module is used to perform spectrum energy adjustment-based analysis processing on the driving signal of the GaN switching device according to the interference frequency band distribution information, so as to obtain an adjustment control instruction for suppressing interference in the key frequency band.

[0007] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above steps when executing the computer program.

[0008] In a fourth aspect, the present application also provides a computer-readable storage medium on which a computer program is stored, and the computer program implements the above steps when executed by a processor.

[0009] The above-mentioned electromagnetic interference suppression method, device, computer equipment and computer-readable storage medium based on high-frequency GaN switching characteristics, first, edge characteristic data is obtained by edge analysis processing of the driving signal of the GaN switching device, thereby realizing quantitative expression of the edge excitation at the time domain level; secondly, the equivalent network model constructed based on parasitic parameters is subjected to frequency domain response processing through the edge characteristic data to obtain interference frequency band distribution information, thereby realizing characteristic expression of interference energy at the frequency domain level; thirdly, the driving signal waveform structure is adjusted according to the interference frequency band distribution information to obtain adjustment control instructions for suppressing key frequency band interference, thereby realizing energy reconstruction and transfer control of the key frequency band; based on this, through the linkage mechanism of time-frequency feature extraction of the driving signal edge and frequency domain modeling analysis, a controllable mapping relationship between high-frequency edge excitation and interference frequency output is effectively established, thereby realizing an effective suppression strategy for specific interference frequency bands in the circuit scenario where GaN switching devices are introduced for power regulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0011] Figure 1 1 is a flow chart of an electromagnetic interference suppression method based on high-frequency GaN switching characteristics in one embodiment; Figure 2FIG. 4 is a structural block diagram of an electromagnetic interference suppression device based on high-frequency GaN switching characteristics in one embodiment. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0013] In one embodiment, Figure 1 As shown, a method for suppressing electromagnetic interference based on the high-frequency GaN switching characteristics is provided. This embodiment uses the method applied to a server as an example. It is understood that the method can also be applied to a terminal, or to a system including a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps S101 to S103.

[0014] In step S101, in a power conversion circuit constructed based on a GaN switching device, edge analysis processing is performed on a driving signal of the GaN switching device to obtain edge characteristic data, where the edge characteristic data includes a rising edge change rate and a falling edge change rate in multiple consecutive switching cycles.

[0015] Among them, the power conversion circuit refers to an electronic circuit structure used to boost, buck, convert AC / DC or perform other forms of power modulation processing on the input power supply; the GaN switching device refers to a high-speed power switching device made based on GaN (gallium nitride) material, which is used to implement high-frequency, high-efficiency current control functions in the electronic circuit structure; the driving signal of the GaN switching device refers to an external control voltage signal used to control the on and off states of the GaN switching device.

[0016] Among them, edge characteristic data represents the quantitative description data obtained by analyzing the edge change process of the driving signal of the GaN switching device. The rising edge change rate and falling edge change rate respectively represent the voltage change per unit time when the voltage of the driving signal changes from low to high (rising edge) or from high to low (falling edge) during the level change process.

[0017] For example, in a power conversion circuit constructed based on a GaN switching device, the driving signal waveform of the GaN switching device is sampled with high time resolution, and the voltage change rate of the rising edge and the falling edge in each switching cycle is calculated by numerical differentiation, that is, the slope of each edge is extracted, and its statistical results in multiple consecutive switching cycles are recorded to obtain edge characteristic data within a certain time range; the edge characteristic data includes the rising edge change rate and the falling edge change rate corresponding to multiple consecutive switching cycles, which serves as descriptive information of the driving signal characteristics of the GaN switching device during the continuous switching process.

[0018] Step S102 : determining an equivalent network model constructed based on parasitic parameters in the power conversion circuit, performing frequency domain response calculation on the equivalent network model according to edge characteristic data, and obtaining interference frequency band distribution information caused by edge variation.

[0019] Among them, parasitic parameters refer to the non-ideal additional electrical characteristic parameters caused by the structural characteristics, packaging form and connection method of the GaN switching device itself in the power conversion circuit. They are used to describe the non-functional coupling effects generated during high-frequency switching.

[0020] Among them, the equivalent network model represents an electrical equivalent network structure constructed by abstractly modeling the device structure, connection method and parasitic coupling effects contained in the actual circuit, which is used to simulate and calculate the response behavior of the circuit under the excitation of the GaN switching device in the frequency domain. The interference frequency band distribution information represents the spectrum analysis data formed based on the response results of the equivalent network model to edge excitation in different frequency ranges, and is used to reflect the distribution characteristics of the interference energy caused by high-frequency switching operations in the frequency domain.

[0021] For example, first, the parasitic parameters introduced by the GaN switching device in the power conversion circuit are determined, such as the pin length, lead layout, and geometric distribution between the internal wiring structure of the device and the PCB contact points of the GaN switching device, which will introduce non-negligible parasitic inductance and parasitic capacitance during the operation of the device. Secondly, the GaN switching device is regarded as an excitation source, and the parasitic parameters introduced by it are used as passive components in the electrical equivalent network structure. The coupling relationship of different physical paths composed of various components in the power conversion circuit is then represented in the form of a combination of inductance, capacitance, and resistance, thereby constructing a complete equivalent network model. Next, the edge characteristic data is used as edge excitation and applied to the equivalent network model. Then, the frequency domain response analysis of the equivalent network model is performed to solve its excitation response at different frequencies, and finally the spectrum distribution of the high-frequency electromagnetic interference caused by the edge excitation of the GaN switching device is calculated, that is, the interference frequency band distribution information. Specifically, this interference frequency band distribution information reflects the frequency components generated by different slope changes over multiple switching cycles and their excitation strength in the corresponding parasitic paths. This allows the concentrated distribution intervals and amplitude weights of each major interference frequency band to be determined. This establishes a clear connection between the time-domain variation of the GaN switch device and the frequency-domain response of the circuit, enabling the driver signal's time-domain behavior to be quantitatively derived from its high-frequency interference performance.

[0022] Step S103 : performing spectrum energy adjustment-based analysis processing on the driving signal of the GaN switch device according to the interference frequency band distribution information to obtain an adjustment control instruction for suppressing interference in the key frequency band.

[0023] Among them, the adjustment control instructions represent control parameters generated based on the interference frequency band distribution information for adjusting the drive signal structure of the GaN switching device, thereby changing the shape of the drive waveform in the actual circuit to reduce the interference energy output in a specific frequency band; for example, when the interference in a certain frequency band is strong, a set of timing adjustment signals is generated to reduce the rising edge change rate or introduce a control buffer to achieve effective management of electromagnetic interference.

[0024] For example, the interference frequency band distribution information is matched with the waveform characteristics of the driving signal to identify which edge slopes or transition processes cause strong electromagnetic interference within a specific frequency band, and a waveform adjustment method is set based on this to weaken the excitation components of this specific frequency band. Specifically, while maintaining the switching logic unchanged, a controllable extension process is introduced to the rising and falling edges of the driving signal. That is, by increasing the transition time of the driving signal edge, the slope is reduced to weaken its high-frequency component; furthermore, under the condition that the overall period and duty cycle of the driving signal remain stable, the interference energy concentrated in the specific frequency band is dispersed to other wider frequency ranges with less impact on the overall function of the system; in addition, the responsiveness of the switching characteristics of the GaN switching device itself to waveform changes must be comprehensively considered to ensure that the adjusted signal remains within its driving level range and does not cause false turn-on or false turn-off phenomena. Ultimately, the aforementioned strategy for adjusting the spectral energy of the driving signal is used to generate a regulation control instruction for suppressing interference in the critical frequency band. During the actual driving process, this regulation control instruction enables the GaN switching device to effectively weaken the interference output in the critical frequency band while performing power conversion, thereby achieving control of electromagnetic compatibility performance at the system level.

[0025] In the above-mentioned electromagnetic interference suppression method based on high-frequency GaN switching characteristics, first, edge feature data is obtained based on edge analysis processing of the driving signal of the GaN switching device, thereby realizing quantitative expression of the edge excitation at the time domain level; secondly, the equivalent network model constructed based on parasitic parameters is subjected to frequency domain response processing through the edge feature data to obtain interference frequency band distribution information, thereby realizing characteristic expression of interference energy at the frequency domain level; thirdly, the driving signal is analyzed and processed based on spectrum energy adjustment according to the interference frequency band distribution information to obtain adjustment control instructions for suppressing interference in the key frequency band, thereby realizing energy reconstruction and transfer control of the key frequency band; based on this, through the linkage mechanism of time-frequency feature extraction of the driving signal edge and frequency domain modeling analysis, a controllable mapping relationship between high-frequency edge excitation and interference frequency output is effectively established, thereby realizing an effective suppression strategy for specific interference frequency bands in the circuit scenario where GaN switching devices are introduced for power regulation.

[0026] In an exemplary embodiment, edge analysis processing is performed on a driving signal of a GaN switching device to obtain edge characteristic data, including steps S201 to S203.

[0027] Step S201 : performing equal-interval sampling processing on the driving signal of the GaN switching device in a plurality of consecutive switching cycles to obtain a sampling voltage sequence representing the waveform variation process of each switching cycle.

[0028] The sampled voltage sequence represents a set of voltage values ​​arranged in time sequence, obtained by sampling the driving signal of the GaN switching device at equal time intervals in each switching cycle.

[0029] For example, first, sufficient sampling points are placed within each switching cycle to ensure that the rising edge, stable high level, falling edge, and stable low level segments of the waveform are fully recorded. Second, the sampling points need to be evenly spaced, and the sampling frequency must be significantly higher than the maximum rate of change of the drive signal to meet the basic requirements of signal restoration accuracy. Based on this, during the sampling process, several sets of voltage and time sequences, namely sampled voltage sequences, can be formed over multiple consecutive switching cycles. Each set of sampled voltage sequences can reflect the changing trajectory of the drive signal within a complete switching cycle, including key information such as signal edge steepness, duration, and transition shape.

[0030] Step S202 , performing slope fitting processing on the rising edge segment and the falling edge segment in each switching cycle according to the sampled voltage sequence, to obtain a first edge slope vector representing the rising edge change rate and a second edge slope vector representing the falling edge change rate.

[0031] Among them, the first edge slope vector represents the set of rising edge change rates obtained after slope fitting is performed on the rising edge segment of the drive signal in each switching cycle within multiple switching cycles; the second edge slope vector represents the set of falling edge change rates obtained after slope fitting is performed on the falling edge segment of the drive signal in each cycle within multiple switching cycles.

[0032] For example, first, a rising edge corresponds to the process in which the drive signal transitions from a low level to a high level, while a falling edge corresponds to the process in which the drive signal transitions from a high level to a low level. Thus, by setting a voltage threshold interval, the starting and ending positions of each transition can be identified, thereby extracting the corresponding rising edge segment and falling edge segment from each set of sampled voltage sequences. On this basis, slope fitting is performed on the voltage-time data of each edge segment. For example, a linear regression method is used to fit a linear curve with a stable slope using the least squares error criterion, and the slope of the linear curve is used as a representative value of the rate of change of the corresponding edge segment. Ultimately, a rising edge change slope value and a falling edge change slope value are obtained in each switching cycle, thereby forming a first edge slope vector corresponding to each rising edge change slope value and a second edge slope vector corresponding to each falling edge change slope value. Each element in the two edge slope vectors corresponds to the edge change slope value of a specific switching cycle.

[0033] Step S203 : performing period balancing processing based on discrete distribution characteristics on the edge change rates in all switching cycles according to the first edge slope vector and the second edge slope vector to obtain edge feature data.

[0034] The discrete distribution characteristic represents the statistical distribution of the values ​​in the first edge slope vector and the second edge slope vector, and is used to analyze the discreteness and regularity of the edge change rate in the numerical space during each switching cycle. For example, standard deviation, range, or frequency statistics can be used to determine whether the edge change is stable and concentrated or highly fluctuating.

[0035] For example, in the actual switching process, the edge change rate of the GaN switching device in different switching cycles may have slight differences due to factors such as temperature, voltage fluctuations or inconsistent driver response. Therefore, these differences can be processed by period equalization through statistical means to obtain a representative and stable edge feature representation. Specifically, in the first edge slope vector and the second edge slope vector, the discrete distribution characteristics of the edge change slope values ​​in all switching cycles are analyzed in terms of dimensional indicators such as average slope, maximum slope or standard deviation, and the edge slope vector is normalized, distributed fitted or mean filtered, so that the edge feature data finally formed can reflect the change trend in the whole cycle, so as to eliminate the value range offset caused by measurement error under different switching cycles.

[0036] In this embodiment, first, the driving signal of the GaN switching device within multiple consecutive switching cycles is sampled at equal intervals to obtain a sampling voltage sequence that completely characterizes the periodic waveform changes, thereby ensuring the time domain continuity of the edge characteristics; secondly, the rising edge segment and the falling edge segment in the sampling voltage sequence are slope fitted to convert the edge change process into a quantifiable edge slope vector, thereby improving the expression accuracy of the edge characteristics; thirdly, the first and second edge slope vectors and the second edge slope vector are subjected to periodic equalization based on discrete distribution characteristics to obtain a data representation that stably characterizes the edge characteristics of the entire cycle; based on this, by constructing a quantifiable and equalizable edge characteristic data representation covering the entire cycle, the structured extraction and refined expression of the high-frequency switch driving signal waveform at the time domain level are achieved.

[0037] In an exemplary embodiment, frequency domain response calculation is performed on the equivalent network model according to edge characteristic data to obtain interference frequency band distribution information caused by edge changes, including steps S301 to S303.

[0038] Step S301 : constructing an excitation function with an asymmetric time slope expression according to the rising edge change rate and the falling edge change rate in the edge feature data.

[0039] The excitation function represents a mathematical function that models the rising edge and the falling edge of the signal with different change slopes on the time axis, so as to simulate the input of the GaN switching device driving signal to the equivalent network model.

[0040] For example, the edge characteristic data records the rising edge change rate and the falling edge change rate respectively, but the actual GaN switching device's drive signal waveform often has obvious asymmetry in the slope of the two edges, that is, the rising edge change rate and the falling edge change rate show structural differences in amplitude and duration; based on this, it is necessary to construct a set of excitation functions with asymmetric time slope expression based on the rising edge change rate and the falling edge change rate in the edge characteristic data. This excitation function is constructed using a time rate segmentation method, using the rising edge change rate to model a time function in the rising edge segment and the falling edge change rate to model another time function in the falling edge segment; by splicing these two time functions in chronological order and maintaining the continuity of the signal amplitude and the differentiability on the time axis, an excitation function representative of the actual drive signal waveform is formed.

[0041] Step S302: input the excitation function into the equivalent network model, perform frequency domain response calculation on the equivalent network model under the action of the excitation function within a preset frequency range, and obtain the response amplitude of each frequency point to the excitation function.

[0042] For example, after applying the excitation function to the equivalent network model, the network response is solved in the frequency domain. Specifically, the excitation function is converted into a frequency domain expression and multiplied with the transfer characteristic function of the equivalent network model in the complex frequency domain to solve the system output amplitude response at each preset frequency point in the preset frequency range. The preset frequency range usually covers a range of several times the typical frequency variation of the driving signal to capture the main energy distribution path of high-frequency excitation in the network structure. Based on this, during the entire frequency domain analysis process, each frequency point will obtain a corresponding response amplitude to reflect the degree to which the corresponding frequency component is amplified, weakened, or remains unchanged in the network structure, thereby achieving a quantified result of the interference energy in the frequency dimension.

[0043] Step S303 : performing multi-scale clustering processing based on amplitude trend characteristics on each frequency point according to the response amplitude of each frequency point to the excitation function, and obtaining the interference frequency band distribution information caused by the edge change.

[0044] Among them, the amplitude trend feature indicates the changing trend of the numerical distribution of the response amplitude corresponding to different frequency points after the frequency domain response calculation of the equivalent network model, which is used to reflect the aggregation, diffusion or transition state of the interference energy in the frequency dimension; for example, in the process of frequency from low to high, the response amplitude shows a gradual increase, high value maintenance or a sharp decrease.

[0045] For example, the larger the response amplitude of a frequency point, the stronger the excitation effect on that frequency point. Therefore, in order to identify the frequency bands where interference energy is primarily concentrated, it is necessary to analyze the amplitude trend characteristics of these response amplitudes. Specifically, the gradient of the response amplitudes of all frequency points must first be calculated to obtain the derivative characteristics of the response amplitude as it changes with frequency. Then, combining the original amplitude range and the derivative characteristics, the amplitude trend characteristics of each frequency point are extracted from multidimensional feature indicators such as the amplitude change direction, change amplitude, and local change density. Next, all frequency points are clustered based on these amplitude trend characteristics, and frequency points with similar amplitude trend characteristics are divided into the same group. The boundaries of each group are represented as continuous frequency ranges. Finally, during the classification process, frequency bands with significant amplitude changes, strong clustering, and a continuous rise or high-value maintenance state are identified as interference frequency bands where interference is concentrated. The start and end boundaries, center frequency, and average response amplitude of the interference frequency band are used as components of the interference frequency band distribution information.

[0046] In this embodiment, first, an excitation function with an asymmetric time slope expression is constructed based on the rising edge change rate and the falling edge change rate, thereby realizing the true expression of the asymmetric edge excitation behavior; secondly, the excitation function is input into the equivalent network model for frequency domain response calculation, thereby adaptively obtaining the response amplitude of the circuit to the edge change at different frequency points; thirdly, the amplitude trend characteristics of the frequency point response amplitude are subjected to multi-scale clustering processing, thereby identifying the distribution information of the interference frequency band caused by the edge change; based on this, by constructing an excitation function that matches the edge change and combining it with frequency response clustering analysis, the structured identification of the interference frequency band can be accurately and adaptively achieved.

[0047] In an exemplary embodiment, the frequency domain response calculation is performed on the equivalent network model under the action of the excitation function within a preset frequency range to obtain the response amplitude of each frequency point to the excitation function, including steps S401 to S403.

[0048] Step S401 : determining a rate weight coefficient related to the edge change rate, performing weighted discrete Fourier analysis on the excitation function according to the rate weight coefficient, and obtaining a weighted frequency domain excitation component corresponding to each frequency point.

[0049] Among them, the rate weight coefficient represents the weighting factor calculated according to the change rate of the rising edge and the falling edge in the excitation function, which is used to highlight the impact of different time segments on the frequency domain energy distribution; for example, the time period with a larger edge change rate corresponds to a larger rate weight coefficient to enhance its contribution to the frequency domain analysis.

[0050] The weighted frequency domain excitation component representation refers to a set of frequency components obtained by weighted calculation based on the rate weight coefficient during the Fourier transform process of the excitation function, and is used to represent the energy distribution of different frequency components in the excitation signal.

[0051] For example, based on the rate of change of the rising and falling edges in the excitation function, the amplitude of the change in the numerical space is extracted, and a set of rate weight coefficients is defined based on this; the rate weight coefficient is used to characterize the energy contribution of the excitation signal in different time periods in the time domain. The larger the slope of the change, the stronger the excitation ability of the frequency domain response, so a larger rate weight coefficient should be assigned. Furthermore, the rate weight coefficient is used as a weighting factor to be embedded in the Fourier transform calculation process, so that the frequency domain decomposition based on the excitation function not only depends on its overall waveform shape, but also reflects the influence of its local slope characteristics; specifically, the excitation function is projected onto a set of basis function spaces within a preset frequency range in the form of a weighted discrete Fourier transform, and the rate weighted result based on the corresponding rate weight coefficient is superimposed on each frequency component, and finally a weighted frequency domain excitation component containing the frequency component and its energy contribution relationship is formed to amplify the energy performance of the time period with a faster edge change rate in the frequency domain.

[0052] Step S402 : After the weighted frequency domain excitation component is applied to the equivalent network model, frequency domain response calculation is performed within a preset frequency range to obtain the initial response amplitude corresponding to each frequency point.

[0053] Exemplarily, the weighted frequency domain excitation component is input into the equivalent network model, indicating that the network structure is excited by the corresponding weighted excitation signal at each frequency point within the preset frequency range. Subsequently, a frequency scanning calculation is performed on each response node in the equivalent network model. By solving the product of the circuit's transfer function in the complex frequency domain and the input frequency component, the system output response corresponding to each frequency point is obtained. The output result of the equivalent network model reflects an enhanced response to the high-rate excitation area, that is, the energy performance in the frequency domain of the time period with a faster edge change rate is amplified. Based on this, after traversing the entire frequency range, a set of frequency point-initial response amplitude pairs can be finally obtained to describe the response characteristics of the circuit structure to the weighted excitation signal at different frequencies.

[0054] Step S403: determining a response sensitivity factor related to the response amplitude, performing sensitivity weighting processing on the initial response amplitude corresponding to each frequency point according to the response sensitivity factor, and obtaining the response amplitude of each frequency point to the excitation function.

[0055] Among them, the response sensitivity factor represents a weighting factor set based on the coupling degree, electromagnetic discharge capability and energy concentration effect of each structural unit in the equivalent network model in different frequency ranges. It is used to adjust the initial response amplitude of each frequency point so that it reflects the sensitivity of the circuit structure to external excitation at that frequency point. For example, when the parasitic structure has a resonance effect at a certain frequency, the corresponding sensitivity factor is set to a larger value to reflect its stronger interference amplification effect.

[0056] For example, this set of response sensitivity factors is associated with the initial response amplitude at each frequency point, and a weighted calculation operation is performed. That is, the initial response amplitude at each frequency point is multiplied by the corresponding response sensitivity factor to obtain the final response amplitude at each frequency point. Based on this, the response amplitude obtained after weighted processing not only reflects the input characteristics and slope distribution of the excitation function, but also incorporates the frequency response tendency of the circuit itself, making the output result more representative and directional in terms of numerical value. The resulting set of frequency point-response amplitude pairs constitutes the true response characteristics of the circuit structure in the frequency domain under the corresponding excitation signal.

[0057] In this embodiment, first, a rate weight coefficient is introduced into the excitation function and a weighted Fourier analysis is performed to obtain a weighted frequency domain excitation component corresponding to each frequency point, thereby enhancing the expression of the influence of the edge change rate on the frequency domain energy distribution; secondly, the weighted frequency domain excitation component is input into the equivalent network model and a frequency domain response calculation is performed to obtain the initial response amplitude corresponding to each frequency point, and then a response sensitivity factor is introduced into the initial response amplitude and a weighted processing is performed to obtain the response amplitude of each frequency point, thereby improving the reflection accuracy of the frequency response result on the actual interference sensitivity of the circuit structure.

[0058] In an exemplary embodiment, according to the response amplitude of each frequency point to the excitation function, multi-scale clustering processing based on amplitude trend characteristics is performed on each frequency point to obtain interference frequency band distribution information caused by edge changes, including steps S501 to S503.

[0059] Step S501 , performing first-order difference and sliding window mean difference processing on the response amplitude of each frequency point to obtain the local change rate of each frequency point, and marking the frequency point with a local change rate greater than a preset threshold as a candidate frequency point with significant amplitude change.

[0060] The local change rate of a frequency point represents the rate of change of the response amplitude of a certain frequency point relative to its adjacent frequency points, and is used to measure the degree of mutation or fluctuation of the frequency point in the spectrum response.

[0061] For example, first, the response amplitudes corresponding to all frequency points are processed by first-order difference, and the change in amplitude between adjacent frequency points is calculated to reflect the change in the slope of the spectral response curve; secondly, combined with the sliding window mean difference processing, the mean sequence is calculated for each frequency point in its local neighborhood, and the mean sequence is further differentiated to capture the fluctuation characteristics of the spectral response curve on a longer scale. That is, these two processing methods together constitute a mechanism for extracting the local change rate of the response amplitude change. The first-order difference processing method can be used to identify mutation points, and the sliding window mean difference processing method can be used to identify structural turning points in slow changes, thereby obtaining the local change rate of each frequency point. Subsequently, the local change rate of all frequency points is compared with a preset threshold, and the frequency points whose local change rate exceeds the preset threshold are marked as candidate frequency points, as the locations that show a strong response to the change of the excitation function in the frequency domain response.

[0062] Step S502 : constructing local amplitude trend vectors for adjacent candidate frequency points, and performing similarity matching processing on each local amplitude trend vector according to multiple window scales to obtain a target frequency point set that satisfies cross-scale trend consistency.

[0063] The local amplitude trend vector represents a set of vectors consisting of the response amplitude sequences of a candidate frequency point and its adjacent candidate frequency points at different window scales, centered on a certain candidate frequency point. This vector describes the amplitude variation pattern of the candidate frequency point at different window scales, where different window scales are determined based on the number of adjacent candidate frequency points. For example, if a candidate frequency point shows a rapid rise in a small-scale window but maintains an upward trend in a large-scale window, its local amplitude trend vector can reflect its trend stability.

[0064] The target frequency point set refers to the set of frequency points that meet the multi-scale trend consistency selected from the candidate frequency points, and is used to characterize the frequency positions that show significant response changes at multiple window scales.

[0065] For example, first, with a candidate frequency point as the center, different window scales are determined based on different numbers of adjacent candidate frequency points, and the local amplitude trend vector corresponding to the candidate frequency point is constructed based on the response amplitude sequence of the candidate frequency point and the adjacent candidate frequency points at different window scales. Furthermore, after constructing the local amplitude trend vectors corresponding to each candidate frequency point, these local amplitude trend vectors are input into the similarity matching algorithm and compared one by one for multiple window scales to determine whether adjacent candidate points have trend synchronization phenomena at multiple window scales, that is, whether they show amplitude change patterns with consistent directions and similar amplitude change forms at multiple window scales. Based on this, candidate frequency points that only undergo isolated changes at a certain window scale but lack overall trend support can be eliminated, while retaining those candidate frequency points that show stable response trends at different window scales. Finally, through similarity matching and consistency screening, a set of target frequency points that meet cross-scale trend consistency is obtained. The target frequency points in the target frequency point set are manifestations of stable concentration of interference energy in the spectrum space.

[0066] Step S503: Obtaining the interference frequency band distribution information caused by the edge change according to the frequency range covered by the target frequency point set.

[0067] For example, first, a continuity analysis is performed on the target frequency point set to identify the segments with close frequency intervals, and the starting frequency and ending frequency of each segment are determined to construct one or more frequency intervals. Each frequency interval represents an interference frequency band where an edge excitation causes a significantly concentrated response in the equivalent network model. Subsequently, the response amplitudes of the target frequency points in each frequency interval are statistically processed to extract the amplitude indicators of the frequency interval, such as the average amplitude, maximum response point or rate of change, etc., to further characterize the energy level and impact intensity of the frequency interval. Finally, the starting frequency, ending frequency and corresponding amplitude indicators of all frequency intervals are integrated to form complete interference frequency band distribution information.

[0068] In this embodiment, first, the frequency point response amplitude is processed by first-order difference and sliding window mean difference, so as to identify candidate frequency points with significant amplitude changes in multiple dimensions; secondly, the corresponding local amplitude trend vector is constructed with each candidate frequency point as the center and multi-scale similarity matching is performed, so as to accurately screen out the target frequency point set that meets the cross-scale trend consistency, thereby obtaining the interference frequency band distribution information caused by edge changes; based on this, by introducing the multi-scale trend extraction and clustering processing mechanism, the high-precision conversion of the frequency response results into the structural expression of the interference frequency band is achieved.

[0069] In an exemplary embodiment, the driving signal of the GaN switching device is analyzed and processed based on spectrum energy adjustment according to the interference frequency band distribution information to obtain an adjustment control instruction for suppressing key frequency band interference, including steps S601 to S603.

[0070] In step S601 , the key frequency band to be suppressed in the interference frequency band distribution information is used as the target frequency band, and the optimal energy attenuation path of the adjustable frequency component in the driving signal of the GaN switch device is calculated by combining the center frequency and equivalent bandwidth of the target frequency band.

[0071] Among them, the adjustable frequency component represents the frequency component in the spectrum structure of the GaN switching device drive signal, which can be used to adjust the energy intensity through edge adjustment, pulse width perturbation, etc.; the optimal energy attenuation path represents the spectrum energy adjustment strategy for the adjustable frequency component, so as to minimize the energy distribution of the drive signal in the critical frequency band without changing the signal function logic, for example, by slowing down the edge change or perturbing the pulse period to transfer the energy density.

[0072] For example, first, the key frequency band to be suppressed is extracted from the interference frequency band distribution information as the target frequency band, and the center frequency and equivalent bandwidth of the target frequency band are obtained, wherein the center frequency is used to describe the frequency position where the interference energy is most concentrated in the target frequency band, and the equivalent bandwidth is used to describe the actual coverage range of the target frequency band in the frequency domain, thereby comprehensively representing the spectrum structure of the target frequency band. Then, according to the energy distribution characteristics of the adjustable frequency component in the driving signal, the adjustment space corresponding to key time domain parameters such as edge slope, pulse duration and repetition frequency is determined, and based on the spectrum structure of the target frequency band, an optimal energy attenuation path is constructed to guide the signal energy to transfer to the non-sensitive frequency band. This optimal energy attenuation path aims to achieve spectrum reconstruction. Without changing the signal function logic, it achieves maximum attenuation of the energy density of the key frequency band in the frequency domain by fine-tuning the edge rate and pulse shape.

[0073] Step S602: performing parameter matching processing on the target frequency band according to the optimal energy attenuation path to obtain a set of adjustment parameters of the target frequency band based on the driving signal, wherein the adjustment parameters include an edge rising rate control amount, an edge falling rate control amount, and a pulse width disturbance factor.

[0074] The adjustment parameter set represents a set of control parameters for regulating the waveform structure of the driving signal of the GaN switching device, so as to reconstruct the driving signal shape in the time domain to achieve active adjustment of the spectral characteristics.

[0075] Among them, the edge rising rate control amount represents the control parameter used to adjust the rising edge change rate in the driving signal, the edge falling rate control amount represents the control parameter used to adjust the falling edge change rate in the driving signal, and the pulse width perturbation factor represents the control parameter used to slightly perturb the duration of each pulse in the driving signal.

[0076] For example, the optimal energy attenuation path reflects the energy attenuation direction and amplitude adjustment degree of the adjustable frequency component in the driving signal within the target frequency band, thereby clarifying which frequency components should be weakened and which frequency components should guide energy dispersion. On this basis, by establishing a mapping relationship between the optimal energy attenuation path and the time domain structure of the driving signal, the edge change behavior and pulse timing characteristics corresponding to these frequency components are identified. For example, the energy concentration of the frequency component in the high frequency band is usually related to a larger edge change slope, so the high frequency component can be weakened by slowing down the edge rise rate and fall rate; for another example, the local accumulation of spectral energy is related to the coupling structure between the pulse duration and the repetition period, so it is necessary to introduce a pulse width perturbation factor. Under the premise of maintaining periodic logic, the pulse duration is slightly adjusted to break up the spectral energy concentration. Through the above-mentioned mapping process from frequency domain to time domain, a set of adjustment parameters including edge rise rate control amount, edge fall rate control amount and pulse width perturbation factor can be obtained.

[0077] Step S603: performing control logic coding processing on the adjustment parameter set to obtain an adjustment control instruction for suppressing interference in the key frequency band.

[0078] Exemplarily, the control logic encoding process is performed on the adjustment parameter set to parameterize its time domain structure and encapsulate it into a logical command format that can be recognized by the controller; specifically, during the encoding process, the adjustment parameter fields such as the edge rise rate control amount, the edge fall rate control amount and the pulse width disturbance factor in the adjustment parameter set are extracted, and these adjustment parameter fields are encoded and converted according to the preset control protocol structure, thereby mapping them into adjustment control instructions that can be sent to the controller.

[0079] In this embodiment, first, the optimal energy attenuation path of the adjustable frequency component in the driving signal is constructed based on the center frequency and equivalent bandwidth of the target frequency band, so as to clarify the direction and strategy of the spectrum energy adjustment; secondly, the target frequency band is parameter matched according to the energy attenuation path to obtain a set of adjustment parameters for signal structure adjustment at the level of establishing a frequency domain to time domain mapping relationship; thirdly, the adjustment parameter set is control logic encoded to generate executable adjustment control instructions. Based on this, directional suppression of key interference frequency bands is achieved through a combination of spectrum guidance, parameter mapping and logic coding.

[0080] In an exemplary embodiment, after obtaining the adjustment control instruction for suppressing interference in the key frequency band, the method further includes steps S701 and S702.

[0081] Step S701 : determining the parameter fields in the adjustment control instruction, performing structured allocation and trigger identification tagging processing on the parameter fields according to a preset register mapping rule, and obtaining a reconstructed adjustment instruction including a rising edge adjustment field, a falling edge adjustment field, and a pulse width disturbance field.

[0082] The register mapping rule represents a mapping specification for allocating each parameter field in the regulation control instruction to the hardware register address space, so as to establish a one-to-one correspondence between each parameter field and the physical register.

[0083] Among them, the reconstructed adjustment instruction represents a control instruction with a precise mapping relationship in the hardware register space, which is formed by structured analysis and identification of the parameter fields in the adjustment control instruction. It defines the characteristic behaviors such as the rising edge, falling edge and pulse width of the GaN switching device drive signal at the level of the underlying instruction configuration.

[0084] For example, first, the parameter fields in the adjustment control instruction are structured and allocated according to the preset register mapping rules, that is, the mapping relationship in the register address space is specified for each parameter field, so that when the controller accesses the parameter field, it can accurately correspond to the specific location of the register. Secondly, the trigger identifier is marked on each parameter field, that is, a corresponding trigger identifier is attached to each parameter field to activate the corresponding signal adjustment operation in each switching cycle or specific state. Through the above processing, the abstract adjustment control instruction is finally converted into a reconstructed adjustment instruction containing a rising edge adjustment field, a falling edge adjustment field and a pulse width disturbance field, forming a bottom-level instruction configuration with a clear structure and a clear execution path, and establishing a basic data interface for subsequent register writing and dynamic drive generation.

[0085] Step S702 : In each switching cycle, the signal reconstruction logic is triggered based on the mapping content of the reconstruction adjustment instruction in the register and the response characteristics of the key frequency band to dynamically generate a reconstructed driving signal with the ability to suppress interference in the key frequency band.

[0086] The reconstructed driving signal refers to a new driving signal waveform generated by adjusting the edge shape and pulse time structure under the action of the reconstruction adjustment instruction, which is used to replace the original driving signal to reduce the interference energy distribution in the key frequency band.

[0087] For example, first, during each switching cycle, the controller reads the parameter fields in the register that are valid for the current switching cycle, and uses these parameter fields as the input basis for reconstructing the drive signal during the current switching cycle. Based on this, the controller determines whether the current period is the most sensitive to the circuit response in the critical frequency band, such as whether the critical frequency band causes strong resonance, electromagnetic radiation, or interference conduction in the circuit structure during the current period. If so, the signal reconstruction operation of the drive signal is immediately triggered. Specifically, the slopes of the rising and falling edges are adjusted based on the rising edge adjustment field and the falling edge adjustment field to make the transition behavior of the signal edge more gentle, thereby reducing its excitation intensity in the high-frequency domain. Furthermore, the original pulse duration is slightly extended or shortened based on the pulse width perturbation field, so that the spectrum energy is diffused from the concentrated area to the adjacent non-critical frequency band. Based on this, the signal reconstruction process enables the output signal to have dynamic spectrum control capabilities without changing the basic timing structure of the drive signal, and can actively avoid the interference frequency band and weaken its energy distribution, thereby achieving signal-level interference suppression. Ultimately, the entire execution process is repeated within each switching cycle, thereby periodically generating a reconstructed drive signal suitable for the corresponding switching cycle. Based on this, a signal conditioning mechanism with periodic response and automatic adjustment is formed to ensure that stable electromagnetic compatibility performance is always maintained under different interference response states.

[0088] In this embodiment, first, the parameter fields in the regulation control instruction are structured and assigned and triggered with identification markings according to the register mapping rules to obtain a reconstructed regulation instruction, thereby realizing accurate mapping of the instruction configuration in the hardware register space; secondly, combined with the mapping content in the register and the response characteristics of the key frequency band, the signal reconstruction logic is triggered to realize the dynamic generation and regulation control of the reconstructed drive signal in each switching cycle. Based on this, the closed-loop execution of the instruction in the underlying control path is realized, thereby enhancing the real-time and adaptability of the interference suppression strategy at the cycle level.

[0089] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0090] Based on the same inventive concept, embodiments of the present application also provide an electromagnetic interference suppression device based on high-frequency GaN switching characteristics for implementing the aforementioned electromagnetic interference suppression method based on high-frequency GaN switching characteristics. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more embodiments of the electromagnetic interference suppression device based on high-frequency GaN switching characteristics provided below can be found in the above-mentioned limitations of the electromagnetic interference suppression method based on high-frequency GaN switching characteristics, and will not be repeated here.

[0091] In an exemplary embodiment, Figure 2 As shown, an electromagnetic interference suppression device based on high-frequency GaN switching characteristics is provided, including: an edge analysis module 201, a response analysis module 202 and an instruction generation module 203, wherein: An edge analysis module 201 is configured to perform edge analysis on a driving signal of a GaN switching device in a power conversion circuit based on the GaN switching device to obtain edge characteristic data, wherein the edge characteristic data includes a rising edge change rate and a falling edge change rate within a plurality of consecutive switching cycles; The response analysis module 202 is used to determine an equivalent network model constructed based on parasitic parameters in the power conversion circuit, perform frequency domain response calculation on the equivalent network model based on edge characteristic data, and obtain information on the frequency band distribution of interference caused by edge changes; The instruction generation module 203 is used to perform spectrum energy adjustment-based analysis processing on the driving signal of the GaN switch device according to the interference frequency band distribution information, and obtain an adjustment control instruction for suppressing interference in the key frequency band.

[0092] In an exemplary embodiment, the edge analysis module 201 is further used to: perform equally spaced sampling processing on the driving signal of the GaN switching device in multiple consecutive switching cycles to obtain a sampling voltage sequence that characterizes the waveform change process of each switching cycle; perform slope fitting processing on the rising edge segment and the falling edge segment in each switching cycle according to the sampling voltage sequence to obtain a first edge slope vector that characterizes the rising edge change rate and a second edge slope vector that characterizes the falling edge change rate; and perform period balancing processing on the edge change rate in all switching cycles based on discrete distribution characteristics according to the first edge slope vector and the second edge slope vector to obtain edge feature data.

[0093] In an exemplary embodiment, the response analysis module 202 is also used to: construct an excitation function with an asymmetric time slope expression based on the rising edge change rate and the falling edge change rate in the edge feature data; input the excitation function into the equivalent network model, and perform frequency domain response calculation on the equivalent network model under the action of the excitation function within a preset frequency range to obtain the response amplitude of each frequency point to the excitation function; and perform multi-scale clustering processing based on amplitude trend characteristics on each frequency point according to the response amplitude of each frequency point to the excitation function to obtain the interference frequency band distribution information caused by edge changes.

[0094] In an exemplary embodiment, the response analysis module 202 is also used to: determine a rate weight coefficient related to the edge change rate, perform weighted discrete Fourier analysis on the excitation function according to the rate weight coefficient, and obtain a weighted frequency domain excitation component corresponding to each frequency point; after applying the weighted frequency domain excitation component to the equivalent network model, perform frequency domain response calculation within a preset frequency range to obtain the initial response amplitude corresponding to each frequency point; determine a response sensitivity factor related to the response amplitude, perform sensitivity weighted processing on the initial response amplitude corresponding to each frequency point according to the response sensitivity factor, and obtain the response amplitude of each frequency point to the excitation function.

[0095] In an exemplary embodiment, the response analysis module 202 is further used to: perform first-order difference and sliding window mean difference processing on the response amplitude of each frequency point to obtain the local change rate of each frequency point, and mark the frequency point with a local change rate greater than a preset threshold as a candidate frequency point with significant amplitude change; construct a local amplitude trend vector for adjacent candidate frequency points, perform similarity matching processing on each local amplitude trend vector according to multiple window scales, and obtain a target frequency point set that meets cross-scale trend consistency; and obtain interference frequency band distribution information caused by edge changes based on the frequency range covered by the target frequency point set.

[0096] In an exemplary embodiment, the instruction generation module 203 is also used to: take the key frequency band to be suppressed in the interference frequency band distribution information as the target frequency band, combine the center frequency and equivalent bandwidth of the target frequency band, and calculate the optimal energy attenuation path of the adjustable frequency component in the driving signal of the GaN switching device; according to the optimal energy attenuation path, perform parameter matching processing on the target frequency band to obtain a set of adjustment parameters of the target frequency band based on the driving signal, the adjustment parameters including the edge rise rate control amount, the edge fall rate control amount and the pulse width disturbance factor; perform control logic coding processing on the adjustment parameter set to obtain an adjustment control instruction for suppressing interference in the key frequency band.

[0097] In an exemplary embodiment, the device also includes an execution module, which is used to: determine the parameter fields in the adjustment control instruction, perform structured allocation and trigger identification labeling processing on the parameter fields according to preset register mapping rules, and obtain a reconstructed adjustment instruction including a rising edge adjustment field, a falling edge adjustment field and a pulse width disturbance field; in each switching cycle, combined with the mapping content of the reconstructed adjustment instruction in the register and the response characteristics of the key frequency band, trigger the signal reconstruction logic to dynamically generate a reconstructed drive signal with the ability to suppress interference in the key frequency band.

[0098] Each module in the aforementioned electromagnetic interference suppression device based on high-frequency GaN switching characteristics can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0099] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in any of the above embodiments when executing the computer program.

[0100] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in any of the above embodiments are implemented.

[0101] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0102] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0103] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for suppressing electromagnetic interference based on high-frequency GaN switching characteristics, characterized in that: The method comprises: In a power conversion circuit constructed based on a GaN switching device, edge analysis processing is performed on a driving signal of the GaN switching device to obtain edge characteristic data, wherein the edge characteristic data includes a rising edge change rate and a falling edge change rate within multiple consecutive switching cycles; Determining an equivalent network model constructed based on parasitic parameters in the power conversion circuit, performing frequency domain response calculation on the equivalent network model according to the edge characteristic data, and obtaining interference frequency band distribution information caused by edge changes; According to the interference frequency band distribution information, the driving signal of the GaN switching device is subjected to analysis processing based on spectrum energy adjustment to obtain an adjustment control instruction for suppressing interference in a key frequency band.

2. The method according to claim 1, characterized in that The performing edge analysis on the driving signal of the GaN switching device to obtain edge characteristic data includes: Performing equal-interval sampling processing on the driving signal of the GaN switching device in multiple consecutive switching cycles to obtain a sampling voltage sequence representing the waveform change process of each switching cycle; Perform slope fitting processing on the rising edge segment and the falling edge segment in each switching cycle according to the sampled voltage sequence, to obtain a first edge slope vector representing the rising edge change rate and a second edge slope vector representing the falling edge change rate; According to the first edge slope vector and the second edge slope vector, a period balancing process based on discrete distribution characteristics is performed on the edge change rate in all switching cycles to obtain edge feature data.

3. The method according to claim 1, characterized in that The performing frequency domain response calculation on the equivalent network model according to the edge feature data to obtain interference frequency band distribution information caused by edge changes includes: Constructing an excitation function with an asymmetric time slope expression according to the rising edge change rate and the falling edge change rate in the edge feature data; Inputting the excitation function into the equivalent network model, performing frequency domain response calculation on the equivalent network model under the action of the excitation function within a preset frequency range, and obtaining the response amplitude of each frequency point to the excitation function; According to the response amplitude of each frequency point to the excitation function, a multi-scale clustering process based on amplitude trend characteristics is performed on each frequency point to obtain the interference frequency band distribution information caused by edge changes.

4. The method according to claim 3, characterized in that The equivalent network model under the action of the excitation function is subjected to frequency domain response calculation within a preset frequency range to obtain the response amplitude of each frequency point to the excitation function, including: Determining a rate weight coefficient related to the edge change rate, performing weighted discrete Fourier analysis on the excitation function according to the rate weight coefficient to obtain a weighted frequency domain excitation component corresponding to each frequency point; After applying the weighted frequency domain excitation component to the equivalent network model, frequency domain response calculation is performed within a preset frequency range to obtain the initial response amplitude corresponding to each frequency point; A response sensitivity factor related to the response amplitude is determined, and sensitivity weighting processing is performed on the initial response amplitude corresponding to each frequency point according to the response sensitivity factor to obtain the response amplitude of each frequency point to the excitation function.

5. The method according to claim 3, characterized in that The step of performing multi-scale clustering processing on each frequency point based on the amplitude trend characteristics according to the response amplitude of each frequency point to the excitation function to obtain the interference frequency band distribution information caused by the edge change includes: Perform first-order difference and sliding window mean difference processing on the response amplitude of each frequency point to obtain the local change rate of each frequency point, and mark the frequency point with a local change rate greater than a preset threshold as a candidate frequency point with significant amplitude change; Construct local amplitude trend vectors for adjacent candidate frequency points, perform similarity matching on each local amplitude trend vector according to multiple window scales, and obtain a set of target frequency points that meet cross-scale trend consistency; According to the frequency range covered by the target frequency point set, the interference frequency band distribution information caused by the edge change is obtained.

6. The method according to claim 1, characterized in that The analyzing and processing the driving signal of the GaN switching device based on spectrum energy adjustment according to the interference frequency band distribution information to obtain an adjustment control instruction for suppressing interference in a key frequency band includes: Taking the key frequency band to be suppressed in the interference frequency band distribution information as the target frequency band, and combining the center frequency and equivalent bandwidth of the target frequency band, calculating the optimal energy attenuation path of the adjustable frequency component in the driving signal of the GaN switching device; performing parameter matching processing on the target frequency band according to the optimal energy attenuation path to obtain a set of adjustment parameters of the target frequency band based on the driving signal, wherein the adjustment parameters include an edge rising rate control amount, an edge falling rate control amount, and a pulse width perturbation factor; The control logic encoding process is performed on the adjustment parameter set to obtain an adjustment control instruction for suppressing interference in a key frequency band.

7. The method according to claim 1, characterized in that After obtaining the adjustment control instruction for suppressing interference in the key frequency band, the method further includes: Determining parameter fields in the adjustment control instruction, performing structured allocation and trigger identification tagging processing on the parameter fields according to a preset register mapping rule, and obtaining a reconstructed adjustment instruction including a rising edge adjustment field, a falling edge adjustment field, and a pulse width disturbance field; In each switching cycle, the signal reconstruction logic is triggered in combination with the mapping content of the reconstruction adjustment instruction in the register and the response characteristics of the key frequency band to dynamically generate a reconstructed driving signal with the ability to suppress interference in the key frequency band.

8. An electromagnetic interference suppression device based on high-frequency GaN switching characteristics, characterized in that: The device comprises: An edge analysis module is configured to perform edge analysis on a drive signal of a GaN switching device in a power conversion circuit based on the GaN switching device to obtain edge characteristic data, wherein the edge characteristic data includes a rising edge change rate and a falling edge change rate within a plurality of consecutive switching cycles; a response analysis module, configured to determine an equivalent network model constructed based on parasitic parameters in the power conversion circuit, perform frequency domain response calculation on the equivalent network model according to the edge characteristic data, and obtain interference frequency band distribution information caused by edge changes; An instruction generation module is used to perform spectrum energy adjustment-based analysis processing on the driving signal of the GaN switching device according to the interference frequency band distribution information, so as to obtain an adjustment control instruction for suppressing interference in a key frequency band.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images