Current transformer zero flux control method and system based on magnetic potential balance constraint

Abstract

The present application relates to the technical field of power electronic measurement, and discloses a current transformer zero magnetic flux control method and system based on magnetic potential balance constraint, wherein the current transformer zero magnetic flux control method based on magnetic potential balance constraint comprises: converting a current signal of a current transformer into a pulse sequence to generate an event-driven data format; processing the pulse sequence by using an event-driven wavelet decomposition algorithm to output a multi-band pulse stream; analyzing the multi-band pulse stream to calculate a coupling relationship between frequency components; based on the coupling relationship between the frequency components, using a pulse frequency coding algorithm to calculate magnetic potential balance compensation parameters; converting the magnetic potential balance compensation parameters into asynchronous control pulse sequences to generate compensation currents and output to a magnetic potential balance execution mechanism; the present application overcomes the inherent delay problem of the traditional synchronous clock sampling method by introducing a neuromorphic computing and event-driven processing mechanism, and realizes microsecond-level response control of the current transformer in a high-frequency switching power supply system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power electronic measurement, more particularly, it relates to a current transformer zero flux control method and system based on magnetic potential balance constraint. BACKGROUND

[0002] With the rapid development of power electronic technology, high-frequency switching power supply systems are widely used in various scenarios. In these systems, the current transformer as a key measuring element needs to respond to rapidly changing current signals within microseconds to ensure the stable operation and measurement accuracy of the system.

[0003] Currently, the traditional current transformer control method is mainly based on synchronous clock sampling technology and uses wavelet transform and graph neural network algorithms for signal processing. These methods have obvious technical defects in practical application: first, the signal processing delay is usually in milliseconds, which cannot meet the strict requirements of modern high-frequency power electronic systems for microsecond-level transient response; second, due to the large processing delay, the system cannot effectively capture and compensate high-frequency components, resulting in a decrease in the measurement accuracy of the current transformer and affecting the stability of the entire system; in addition, during the high-frequency switching process, the rapid change of current can easily cause the saturation of the magnetic core of the current transformer, resulting in measurement errors and seriously affecting the normal operation and safety of the system.

[0004] The prior art has not yet proposed an effective solution to the above problems, especially in the aspects of microsecond-level response time and prevention of magnetic core saturation, there is a clear technical gap, and new current transformer control technology needs to be developed. SUMMARY

[0005] The present application provides a current transformer zero flux control method and system based on magnetic potential balance constraint, which solves the technical problem that the current transformer cannot respond to rapidly changing current signals within microseconds in related technologies.

[0006] The present application provides a current transformer zero flux control method based on magnetic potential balance constraint, comprising:

[0007] Converting the current signal of the current transformer into a pulse sequence to generate an event-driven data format;

[0008] Processing the pulse sequence using an event-driven wavelet decomposition algorithm to output a multi-band pulse stream;

[0009] The analysis of the multi-band pulse stream, the calculation of the coupling relationship between the frequency components, includes: for each pair of frequency components, extracting the pulse generation time sequence; standardizing the time sequence to eliminate the difference in time scale of different frequency bands; calculating the time difference value after standardization and applying the time kernel function for weighting; normalizing the coupling weight of each frequency component to ensure that the total weight is 1; multiply the normalized coupling weight to obtain the coupling strength; and constructing a pulse timing parameter relationship graph model based on the coupling strength;

[0010] Based on the coupling relationship between the frequency components, the magnetic potential balance compensation parameter is calculated by using the pulse frequency coding algorithm, including: normalizing the magnetic flux data to standardize the magnetic flux value to the range of [-1, 1]; normalizing the pulse frequency code to make the numerical range of each frequency component consistent; combining the normalized magnetic flux data and the pulse frequency code to calculate the magnetic potential balance compensation parameter through a weighted nonlinear mapping function;

[0011] The magnetic potential balance compensation parameter is converted into an asynchronous control pulse sequence, a compensation current is generated and output to the magnetic potential balance execution mechanism.

[0012] Further, the step of converting the current signal of the current transformer into a pulse sequence includes:

[0013] Normalizing the original current signal to normalize the current amplitude to the range of [0, 1];

[0014] Applying a pulse density coding algorithm to convert the normalized current signal into a pulse sequence;

[0015] An adaptive threshold adjustment algorithm is executed, when the change rate of the current signal exceeds a preset threshold, the time window is dynamically reduced and the pulse generation frequency is increased.

[0016] Further, the event-driven wavelet decomposition algorithm uses a pulse neural network, which is composed of the following components:

[0017] Input layer: receiving pulse sequence and distributing to different processing channels;

[0018] Feature extraction layer: composed of multiple pulse neurons, each neuron has different time-frequency filter characteristics;

[0019] Frequency separation layer: grouping and clustering pulses with different frequency characteristics;

[0020] Output layer: generating a multi-band pulse stream that preserves the time-frequency characteristics of the original signal.

[0021] Further, the encoded pulse sequence is input to a neuromorphic processor for subsequent processing, including:

[0022] Timestamping the pulse sequence to ensure events are processed in the order they occurred;

[0023] Applying a pulse sparsification algorithm to remove redundant pulses and improve processing efficiency;

[0024] Distributing the processed pulse sequence to different processing units of the neuromorphic processor;

[0025] Achieving a response speed of microseconds through asynchronous parallel processing;

[0026] Generating a standardized event stream.

[0027] Further, using an event-driven wavelet decomposition algorithm to process the pulse sequence, outputting a multi-band pulse stream including:

[0028] Inputting the pulse sequence into the event-driven wavelet decomposition algorithm, which is based on pulse triggering and only performs calculations when the signal changes;

[0029] Performing pulse response transformation to convolve the input pulse with a pre-set wavelet-like basis function and extract the characteristics of different frequency components;

[0030] Applying frequency-selective filtering to distribute the pulse to different processing channels according to its time-frequency characteristics;

[0031] Aggregating the processing results to generate a multi-band pulse stream output, with each frequency band corresponding to a frequency component of the current signal.

[0032] Further, the step of converting the magnetic potential balance compensation parameter into an asynchronous control pulse sequence includes:

[0033] Normalizing the compensation parameter to eliminate differences in dimensions and numerical ranges between different parameters;

[0034] Mapping the normalized compensation parameter to the pulse amplitude and frequency space;

[0035] Through a nonlinear mapping function, generating the corresponding pulse density according to the size and rate of change of the parameter.

[0036] Further, the step of generating a compensation current includes:

[0037] Digitally filtering the input control pulse to remove high-frequency noise;

[0038] Data standardization of the pulse sequence to unify the time and frequency domain characteristics;

[0039] Applying a direct mapping function to convert the normalized pulse sequence into a compensation current amplitude;

[0040] The generated current amplitude is calibrated to ensure that it meets the input specifications of the actuator;

[0041] According to the pulse timing characteristics, the timing characteristics of the compensation current are adjusted.

[0042] Further, the step of generating a compensation current and outputting to the magnetic potential balance actuator includes:

[0043] Monitoring the real-time change of magnetic flux;

[0044] Calculate the deviation of the magnetic flux from zero;

[0045] Real-time adjustment of the compensation current makes the magnetic flux tend to zero.

[0046] Further, the generation process of the asynchronous control pulse sequence utilizes an asynchronous control event-driven model, which includes the following key components:

[0047] Pulse generator: dynamically generate control pulses according to compensation parameters;

[0048] Event queue manager: manage the ordering and scheduling of asynchronous pulse events;

[0049] Direct digital-to-analog mapping unit: convert pulses to analog compensation current;

[0050] Magnetic potential detection feedback unit: provide closed-loop control feedback.

[0051] The present application provides a current transformer zero flux control system based on magnetic potential balance constraint, which is used to execute the above-mentioned current transformer zero flux control method based on magnetic potential balance constraint, including:

[0052] Pulse conversion module: for converting current signal to pulse sequence and generating event-driven data;

[0053] Frequency band analysis module: for realizing event-driven wavelet decomposition and multi-band pulse stream output;

[0054] Frequency coupling calculation module: for processing multi-band pulse stream and constructing coupling relationship between frequency components;

[0055] Magnetic potential balance calculation module: for generating magnetic potential balance compensation parameters based on frequency coupling relationship;

[0056] Control execution module: for converting compensation parameters to control pulses and driving magnetic potential balance execution device.

[0057] The beneficial effects of the present application are that the traditional synchronous clock-based signal processing method needs to wait for a complete sampling period before starting processing, while the event-driven processing mechanism adopted by the present application triggers calculation only when the signal changes, fundamentally eliminating such delay;

[0058] By converting the current signal into a pulse sequence and using a pulse neural network for parallel processing, the system can respond to current changes within a microsecond time scale, improving response speed compared to traditional millisecond-level processing methods;

[0059] The present application can capture and respond to microsecond-level signal changes in real time, so it has better compensation ability for high-frequency components in high-frequency switching power supplies, and this enhanced compensation ability directly leads to effective prevention of core saturation;

[0060] Through the construction of the pulse timing parameter relationship diagram and the application of pulse frequency coding, the system can capture subtle changes in the current signal and make precise magnetic potential compensation based on these changes, thereby achieving more accurate magnetic potential balance and zero flux control;

[0061] The event-driven processing mechanism only activates calculation when the signal changes, which greatly reduces unnecessary calculation operations and reduces system power consumption;

[0062] Through the adaptive threshold mechanism and parameter adjustment algorithm, the present application can adapt to current transformers under different working conditions, improving the system's application range and stability. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 is a flowchart of the current transformer zero flux control method based on magnetic potential balance constraints in the present application;

[0064] Figure 2 is a broken line graph of response characteristic comparison in the pulse coding analysis process of current jump events;

[0065] Figure 3 is a broken line graph of the relationship between magnetic potential balance compensation parameters and magnetic flux;

[0066] Figure 4 is a column chart comparing the magnetic flux changes before and after compensation under different current conditions;

[0067] Figure 5 is a column chart comparing the response speed of different control methods;

[0068] Figure 6 is a column chart comparing the performance of the traditional synchronous control method and the method of the present application in terms of magnetic flux density reduction rate and iron loss reduction rate under different working frequencies. DETAILED DESCRIPTION

[0069] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that discussions of these implementations are merely provided to enable those skilled in the art to better understand so as to best use the subject matter described herein, and variations of elements discussed can be made by one skilled in the art, without departing from the scope of the present disclosure. Various examples can omit, substitute, or add various procedures or components as appropriate, and the descriptions and representations of various examples are used to provide textual descriptions for understanding and are not restrictive. Additionally, some examples described can be combined with other examples described.

[0070] A current transformer zero flux control method based on magnetic potential balance constraint is disclosed in at least one embodiment of the present application, as shown in Figure 1 The method comprises the following steps:

[0071] Step 1, convert the current signal of the current transformer into a pulse sequence to generate an event-driven data format;

[0072] This step includes the following operations:

[0073] Step 1.1, current signal sampling;

[0074] A high-speed analog-to-digital converter (ADC) is used to sample the secondary side current signal of the transformer to obtain an analog current signal.

[0075] Step 1.2, pulse density encoding;

[0076] The pulse density encoding algorithm is applied to convert the sampled analog current signal into a pulse sequence. The following operations are performed in detail:

[0077] The original current signal is normalized to normalize the current amplitude to the range of [0, 1] for subsequent processing;

[0078] The normalized signal amplitude is analyzed to establish a mapping relationship between the amplitude and the pulse density;

[0079] Generate pulse events with corresponding density on the timeline;

[0080] Mark the event sequence as a timestamp, address pair format.

[0081] Step 1.3, adjust the encoding parameters;

[0082] An adaptive threshold adjustment algorithm is executed to monitor the rate of change of the current signal in real time, and the encoding parameters are adjusted by the following steps:

[0083] Calculate the rate of change of the signal in a short time window;

[0084] Compare the calculation result with the preset threshold;

[0085] When the rate of change exceeds the threshold, the time window is dynamically reduced and the pulse generation frequency is increased;

[0086] When the rate of change decreases, the time window gradually returns to normal size.

[0087] As shown in Figure 2 the pulse encoding analysis process of the current jump event is shown, which shows that in the current jump event, as the current value increases sharply (from 50A to 250A), the system adaptively adjusts the time window and increases the number of generated pulses, which intuitively shows the fast response characteristics of the event-driven processing mechanism. Through the figure, it can be clearly observed that when the current changes sharply, the time window is dynamically reduced from 10μs to 2μs, and the number of generated pulses is increased accordingly, ensuring that the system can respond to current changes within microseconds.

[0088] Step 1.4, subsequent processing;

[0089] The encoded pulse sequence is input to the neuromorphic processor for subsequent processing. Specifically, it includes:

[0090] The pulse sequence is timestamped and sorted to ensure that events are processed in the order of occurrence;

[0091] The pulse sparsification algorithm is applied to remove redundant pulses and improve processing efficiency;

[0092] The processed pulse sequence is distributed to different processing units of the neuromorphic processor;

[0093] Through asynchronous parallel processing, the response speed of microseconds is realized;

[0094] The standardized event stream is generated to provide input for the next step of the wavelet decomposition algorithm.

[0095] Further, in the above pulse density encoding algorithm, for particularly fast changing current signals, a threshold prediction mechanism is used to generate pulses in advance when the signal reaches a certain percentage of the predicted value, further reducing processing delay.

[0096] Further, in the pulse frequency encoding, due to the dynamic adjustment of the time window size affecting the data distribution, the system introduces an adaptive normalization mechanism. This mechanism monitors the data distribution changes caused by window adjustment in real time, dynamically adjusts the normalization parameters, and ensures that the pulse sequences generated under different time window sizes have consistent statistical characteristics. This adaptive normalization mechanism improves the stability and accuracy of the system under fast changing signal conditions.

[0097] Further, to deal with high frequency components, the system performs power normalization on signals in different frequency channels to avoid bias when comparing high frequency and low frequency components directly. The power normalization process automatically adjusts the gain factor according to the frequency response characteristics to ensure the balance of different frequency components in energy distribution, while preserving the relative phase information of each frequency component, providing a more reliable data basis for subsequent frequency coupling analysis.

[0098] Step 2, processing the pulse sequence using an event-driven wavelet decomposition algorithm, outputting a multi-band pulse stream;

[0099] This step includes the following operations:

[0100] Step 2.1, pulse sequence input;

[0101] The pulse sequence is input into the event-driven wavelet decomposition algorithm, which is based on pulse triggering and only performs computation when the signal changes;

[0102] Step 2.2, pulse response transformation;

[0103] Performing pulse response transformation, the input pulse is convolved with a pre-set wavelet-like basis function to extract the characteristics of different frequency components;

[0104] Further, before performing pulse response transformation, the system performs amplitude normalization on the input pulse to map the amplitudes of pulses from different sources to a unified range. This normalization eliminates the amplitude differences of pulses from different signal sources, ensuring that the subsequent convolution operation can accurately extract the time-frequency characteristics without being affected by the original amplitude. The normalization process uses a dynamic threshold mechanism to automatically adjust the normalization parameters according to the pulse amplitude distribution in a short time window.

[0105] Step 2.3, frequency selective filtering;

[0106] Applying frequency selective filtering, the pulse is assigned to different processing channels according to its time-frequency characteristics;

[0107] Step 2.4, aggregate processing results;

[0108] Aggregating the processing results to generate a multi-band pulse stream output, each frequency band corresponding to a frequency component of the current signal.

[0109] Further, the above event-driven wavelet decomposition algorithm utilizes a spiking neural network model in hardware implementation. The spiking neural network model has the following components:

[0110] Input layer: receiving pulse sequence and assigning to different processing channels;

[0111] Feature extraction layer: composed of multiple pulse neurons, each neuron has different time-frequency filter characteristics, simulating wavelet basis functions;

[0112] Frequency separation layer: grouping and clustering pulses with different frequency characteristics;

[0113] Output layer: generating multi-band pulse stream, preserving the time-frequency characteristics of the original signal.

[0114] Step 3, analyze the multi-band pulse stream, calculate the coupling relationship between frequency components;

[0115] This step includes the following operations:

[0116] Step 3.1, receive the multi-band pulse stream;

[0117] Extract the time characteristic parameters of each frequency band, including pulse occurrence time, interval and density, etc.

[0118] The system adjusts the frequency-related weights of the multi-band pulse stream, taking into account the different physical meanings of different frequency components. The adjustment process includes: analyzing the energy contribution of each frequency component in the current signal, evaluating the impact of different frequencies on the saturation of the magnetic core, and establishing a frequency weight mapping relationship. Through this weight adjustment, it ensures that the actual importance of each frequency component can be reasonably reflected in the subsequent coupling relationship analysis, improving the practical value of the analysis results.

[0119] Step 3.2, execute the activity propagation algorithm;

[0120] Analyze the mutual influence between frequency components according to the following steps:

[0121] For each pair of frequency components, extract their pulse occurrence time sequence and ;

[0122] Standardize the time sequence to eliminate the differences in time scales of different frequency bands, and normalize the time difference to a relative time unit;

[0123] Calculate the normalized time difference and apply the time kernel function for weighting;

[0124] Normalize the coupling weights of each frequency component to ensure the sum of the weights is 1, and ensure the comparability of the weights of different frequency channels;

[0125] Multiply the normalized coupling weights to obtain the coupling strength ;

[0126] Repeat the above calculation process to obtain the complete coupling matrix. ​

[0127] ;

[0128] wherein, denotes the coupling strength between nodes and , and denote the spike firing times of nodes and , is a time kernel function, is a coupling weight; denotes the summation symbol.

[0129] Further, the time kernel function is defined as an exponential decay function:

[0130] ;

[0131] wherein and denote the spike firing times of nodes and , is a time kernel function; is the base of the natural logarithm, is a decay coefficient, used to control the decay rate of the influence of the time difference on the coupling strength. The larger the time difference, the smaller the coupling influence. The time kernel function ensures that the closer the time of the pulse events, the stronger the interaction between them, while the events with larger time intervals have less mutual influence.

[0132] Before performing the key coupling analysis, the system normalizes the calculated coupling strength matrix to ensure consistent comparison benchmarks for different frequency intervals. The normalization process includes: eliminating frequency-dependent bias, balancing the weights of high and low frequency components in coupling analysis, and adjusting the dynamic range of coupling strength to facilitate subsequent threshold screening. This normalization process improves the reliability and accuracy of cross-frequency interval coupling relationship analysis.

[0133] Step 3.3, key coupling analysis;

[0134] According to the calculated coupling matrix, perform the key coupling analysis algorithm:

[0135] Set the coupling strength threshold to filter weak coupling relationships;

[0136] Use clustering algorithm to group frequency components;

[0137] Calculate the centrality index within each group to identify key coupling nodes.

[0138] Step 3.4, output the results of key frequency coupling;

[0139] The results of the output key frequency coupling include frequency components and their mutual relations that have important influence on the saturation of the magnetic core.

[0140] Further, when performing the activity propagation algorithm, an adaptive time window technique is used, which can automatically adjust the time range of the correlation analysis according to the characteristics of different frequency components, thereby improving the accuracy of the coupling relation analysis.

[0141] Further, based on the above calculation results, the system implements a pulse timing parameter relation graph model, which is composed of the following components:

[0142] The graph structure layer is composed of nodes and edges, where the nodes represent the pulse characteristics of different frequency components, and the edges represent the coupling relations between the frequency components.

[0143] The timing characteristic storage unit records the time sequence information of each node.

[0144] The coupling matrix storage unit saves the calculated coupling strength between the frequency components.

[0145] The path analysis module is used to find the key nodes and coupling paths.

[0146] Step 4, based on the coupling relations between the frequency components, the magnetic potential balance compensation parameters are calculated using the pulse frequency coding algorithm.

[0147] This step includes the following operations:

[0148] Step 4.1, pulse frequency coding.

[0149] The pulse frequency coding algorithm is executed and processed according to the following steps:

[0150] The number of pulses of each frequency component is counted within a fixed time window.

[0151] The pulse number is divided by the window length to obtain the pulse frequency code:

[0152] ;

[0153] Where f1, f2, and fn represent the pulse frequencies of the 1st, 2nd, and nth frequency components, respectively. 、 、 is the total number of frequency components processed in the system.

[0154] The frequency code is combined with the key frequency coupling relation to generate an enhanced feature representation.

[0155] Step 4.2, collect magnetic flux data. ​

[0156] Collecting magnetic flux data And the following pretreatment:

[0157] The magnetic flux data is normalized to standardize the magnetic flux value to the range of [-1, 1];

[0158] The pulse frequency code is normalized to make the numerical range of each frequency component consistent;

[0159] The preprocessed magnetic flux data and pulse frequency code are combined and input into the magnetic potential balance calculation algorithm.

[0160] Before the magnetic potential balance difference calculation, the system performs dimensionless processing on all input parameters, especially the scale matching between physical quantities and calculation quantities. The specific processing includes: converting the magnetic flux data into dimensionless form, applying frequency response characteristic correction to the pulse frequency code, and establishing a unified calculation reference. This dimensionless ensures that different physical parameters can be reasonably combined in the calculation model, avoiding calculation bias caused by inconsistent dimensions.

[0161] Step 4.3, magnetic potential balance difference calculation;

[0162] Perform magnetic potential balance difference calculation according to the following relationship:

[0163] ;

[0164] Where, represents the magnetic potential balance difference; represents the magnetic flux; , , represent the first, second, and third frequency components of the pulse frequency, is the total number of frequency components processed in the system; is the magnetic potential balance difference calculation function.

[0165] As shown in Figure 3 , the magnetic potential balance compensation parameter calculation process is shown, which shows the relationship between the magnetic flux value and the calculated compensation parameter in the form of a line chart, showing how the system calculates the corresponding compensation parameter to achieve magnetic potential balance control as the magnetic flux increases. From the figure, it can be observed that the compensation parameter is always greater than the magnetic flux value, and the two show an approximate linear relationship, which reflects the stability of the system's magnetic potential balance calculation algorithm, ensuring that effective compensation control can be maintained throughout the magnetic flux change process.

[0166] Further, defined as a weighted nonlinear mapping function:

[0167] ;

[0168] wherein is a magnetic potential balance difference calculation function; denotes magnetic flux; denotes a summation symbol; is a weight coefficient of the th frequency component; , , denote the pulse frequency of the 1st, 2nd, and th frequency component, respectively, is the total number of frequency components processed in the system; is a nonlinear transformation function of the th frequency component, defined as:

[0169] ;

[0170] wherein is a pulse frequency of the th frequency component; is a nonlinear regulation coefficient.

[0171] The weighted nonlinear mapping function structure ensures that the linear relationship is dominant when the magnetic flux is low, while the influence of the nonlinear term is enhanced in the high magnetic flux region, which can accurately capture the complex relationship between the magnetic potential balance and the frequency components.

[0172] Step 4.4, compensation parameter adaptive calculation;

[0173] The compensation parameter adaptive calculation algorithm is executed to optimize the parameters through the following steps:

[0174] Calculate the trend of the change in the magnetic potential balance difference;

[0175] Adjust the compensation parameter based on the gradient descent method;

[0176] Adaptively adjust the learning rate according to the dynamic characteristics of the magnetic source;

[0177] Output the optimal compensation parameter.

[0178] Further, in the above magnetic potential balance difference calculation, segmented linear logic is used for fast calculation, and different calculation models are dynamically switched in the nonlinear region to quickly adapt to the current change range.

[0179] Further, the magnetic potential balance model used in this step is composed of the following components:

[0180] Input layer: receives the magnetic flux signal and the pulse frequency code;

[0181] Mapping layer: convert input features to a feature space suitable for computation;

[0182] Nonlinear computation unit: implement nonlinear mapping between magnetic potential balance difference and input features;

[0183] Parameter adaptive unit: adjust model parameters according to real-time magnetic source dynamic characteristics;

[0184] Output layer: generate compensation parameters.

[0185] Step 5, convert magnetic potential balance compensation parameters into asynchronous control pulse sequence, generate compensation current and output to magnetic potential balance execution mechanism;

[0186] This step includes the following operations:

[0187] Step 5.1, parameter to pulse conversion;

[0188] Execute parameter to pulse conversion algorithm, process according to the following steps:

[0189] Normalize the compensation parameters, map them to a standard range, and eliminate the differences in dimensions and numerical ranges between different parameters;

[0190] Map the normalized compensation parameters to the pulse amplitude and frequency space;

[0191] Through a nonlinear mapping function, generate the corresponding pulse density according to the size and rate of change of the parameters;

[0192] Arrange the pulse sequence in chronological order to form an asynchronous control pulse sequence;

[0193] Step 5.2, pulse to current mapping;

[0194] Apply pulse current mapping algorithm, perform the following calculation steps:

[0195] Digitally filter the input control pulse to remove high-frequency noise;

[0196] Standardize the pulse sequence data to unify the time and frequency domain characteristics;

[0197] Apply direct mapping function to convert the standardized pulse sequence to compensation current amplitude;

[0198] Calibrate the generated current amplitude to ensure it meets the input specifications of the execution mechanism;

[0199] Adjust the timing characteristics of the compensation current according to the pulse timing characteristics;

[0200] Step 5.3, zero flux control;

[0201] The zero flux control algorithm is executed to adjust the size and direction of the compensation current in real time:

[0202] The real-time change of the magnetic flux is monitored;

[0203] The deviation of the magnetic flux from zero is calculated;

[0204] The compensation current is adjusted so that the magnetic flux is closer and closer to zero;

[0205] Step 5.4, output compensation current;

[0206] The compensation current is output to the magnetic potential balance actuator to achieve zero flux control of the current transformer.

[0207] As shown in Figure 4 , the zero flux control effect verification result is shown, and the magnetic flux change before and after compensation under different current conditions is compared, which intuitively shows the effectiveness of the zero flux control method. From the figure, it can be clearly seen that the magnetic flux after compensation is much lower than the initial magnetic flux, which verifies that the system can maintain a magnetic flux suppression rate of more than 94% in the full current range, fully proving the effect of the method in preventing core saturation.

[0208] Further, in the pulse-to-current mapping algorithm, an adaptive frequency compensation technique is used, which can differentiate different frequency control pulses and improve the accuracy of the compensation current.

[0209] Further, the asynchronous control event-driven model in this step includes the following key components:

[0210] Pulse generator: dynamically generate control pulses according to compensation parameters;

[0211] Event queue manager: manage the ordering and scheduling of asynchronous pulse events;

[0212] Direct digital-to-analog mapping unit: no traditional digital-to-analog converter is needed, and the pulse is directly converted to analog compensation current;

[0213] Magnetic potential detection feedback unit: real-time detection of magnetic flux, providing closed-loop control feedback.

[0214] As shown in Figure 5 , the response time and settling time of the traditional synchronous control method and the method of the present application under different current step amplitudes are compared, which intuitively shows the superiority of the method of the present application in response speed. As can be seen from the figure, the response time and settling time of the method of the present application are about 90% lower than those of the traditional method, achieving microsecond-level response, which provides reliable guarantee for real-time control of high-frequency power electronic systems.

[0215] As shown in Figure 6As shown, the performance of the traditional synchronization control method and the method of the application in reducing the magnetic flux density and the iron loss is compared under different working frequencies, and the characteristics of the method of the application in maintaining stable superior performance under high-frequency conditions are shown. As shown in the figure, the magnetic flux density reduction rate of the traditional method is only 57% under high frequency (100 kHz), while the method of the application remains at about 93%, especially in terms of iron loss reduction, which always remains at a high level of more than 98%, and is almost unaffected even under extremely high frequency conditions, which is of great significance to the stable operation of high-frequency power electronic systems.

[0216] The current transformer zero-flux control system based on the magnetic potential balance constraint is used to execute the current transformer zero-flux control method based on the magnetic potential balance constraint, and comprises:

[0217] Pulse conversion module: for converting the current signal into a pulse sequence and generating event-driven data;

[0218] Frequency band analysis module: for realizing event-driven wavelet decomposition and multi-band pulse stream output;

[0219] Frequency coupling calculation module: for processing the multi-band pulse stream and constructing the coupling relationship between the frequency components;

[0220] Magnetic potential balance calculation module: for generating magnetic potential balance compensation parameters based on the frequency coupling relationship;

[0221] Control execution module: for converting the compensation parameters into control pulses and driving the magnetic potential balance execution device.

[0222] In this embodiment, the application provides an implementation example:

[0223] This embodiment is applied to a high-frequency fast charging system of an electric vehicle. The system uses a high-frequency DC-DC converter realized by a SiC device, and the switching frequency is as high as 100 kHz, the working voltage is 750 V, and the maximum charging current is 400 A. In this application scenario, the current transformer needs to accurately measure the high-frequency pulsating current and perform real-time control and protection. The traditional transformer is prone to core saturation at such a high frequency, which reduces the measurement accuracy and generates additional loss, while the method proposed in the application can effectively solve this problem.

[0224] In this example, a high-speed ADC with a sampling rate of 20 MHz is used to sample the charging current signal. Through a pulse density encoding algorithm, the analog current signal is converted into a pulse sequence. Taking the current jump at a certain time as an example, when the current jumps from 50 A to 250 A, the traditional transformer needs tens to hundreds of microseconds to respond stably, while after using the method, the pulse encoding example data of the current jump event is shown in Table 1:

[0225] Table 1: Pulse code example data of current jump event

[0226]

[0227] When the system detects that the current rate of change exceeds the preset threshold (>100 A / μs), it automatically narrows the time window from 10 μs to 2 μs and increases the pulse generation frequency, ensuring a fast response to current jumps within 2 μs.

[0228] In a high-frequency charging system, the current signal contains multiple frequency components, mainly including the base frequency (0 Hz DC component), the switching frequency (100 kHz), and its harmonic components. Through the event-driven wavelet decomposition algorithm, the system identifies five main frequency components: DC component, 20 kHz, 50 kHz, 100 kHz, and 150 kHz. The wavelet decomposition algorithm is based on pulse triggering and only performs when the signal changes, achieving efficient multi-band pulse stream output.

[0229] According to the multi-band pulse stream output by the wavelet decomposition algorithm, the system constructs a pulse timing parameter relationship diagram and a coupling strength matrix (normalized value) between frequency components as shown in Table 2:

[0230] Table 2: Coupling strength matrix (normalized value) between frequency components

[0231]

[0232] Analyzing the coupling strength matrix in Table 2, it can be seen that the 100 kHz switching frequency has a strong coupling relationship with the DC component and the 50 kHz component, indicating that the switching frequency has the most significant impact on the saturation of the magnetic core. This information is of great significance for subsequent magnetic potential compensation parameter calculation.

[0233] Based on the frequency coupling relationship obtained from the previous steps, the system calculates the weight of each frequency component and generates the magnetic potential balance compensation parameters. According to the description in step 4 of embodiment 1, the system collects magnetic flux data and performs normalization processing, standardizing the magnetic flux value to the range of [-1, 1] (shown as [0, 1] in Table 3 because the measured magnetic flux in this example is all positive). Then, the normalized magnetic flux data is combined with the pulse frequency code to calculate the compensation parameters through a weighted nonlinear mapping function. An example of magnetic potential balance compensation parameter calculation is shown in Table 3:

[0234] Table 3: Example of magnetic potential balance compensation parameter calculation

[0235]

[0236] The frequency component weight vector in the table The contribution degree corresponding to [DC component, 20 kHz, 50 kHz, 100 kHz, 150 kHz] is the result obtained based on the aforementioned coupling strength matrix analysis. Among them, the weight of the switching frequency (100 kHz) is the largest (0.45), which is consistent with the strong coupling relationship between the 100 kHz frequency component and other components in the coupling strength matrix.

[0237] The system generates an asynchronous control pulse sequence according to the calculated compensation parameters, and converts the pulse into a compensation current through direct digital-to-analog mapping. The process first normalizes the compensation parameters, then maps them to the pulse amplitude and frequency space, and finally generates the corresponding pulse density through a nonlinear mapping function. The compensation current in Table 4 is the result after processing by the nn direct digital-to-analog mapping unit, which does not require a traditional digital-to-analog converter and directly converts the pulse into an analog compensation current. To verify the zero flux control effect, the system monitors and records the magnetic flux in real time. The zero flux control effect verification data is shown in Table 4:

[0238] Table 4: Zero flux control effect verification data

[0239]

[0240] As can be seen from Table 4, the method effectively controls the magnetic flux and maintains a magnetic flux suppression rate of more than 94% in the full current range, effectively preventing core saturation.

[0241] The response speed comparison of different control methods is shown in Table 5:

[0242] Table 5: Response speed comparison of different control methods

[0243]

[0244] The data in Table 5 shows that compared with the traditional synchronous control method, the zero flux control method implemented by the invention improves the response speed by about 10 times, shortens the stable time by about 9 times, and reduces the overshoot by about 2 / 3, fully demonstrating the realization of microsecond-level response capability.

[0245] The core saturation prevention effect comparison of different control methods is shown in Table 6:

[0246] Table 6: Core saturation prevention effect comparison of different control methods

[0247]

[0248] Table 6 clearly shows the superiority of the method of the present application under high frequency conditions. As the working frequency increases, the flux control effect of the conventional method decreases sharply, while the method of the present application can still maintain a magnetic flux density reduction rate of more than 93% under high frequency conditions of 100 kHz, effectively preventing core saturation, while significantly reducing iron loss. This result fully demonstrates the applicability and effectiveness of the present application in high frequency power electronic systems.

[0249] The above describes the embodiments of the present application, but the embodiments are not limited to the specific implementation described above, which is only illustrative and not restrictive, and those skilled in the art can make more forms of equivalent embodiments under the inspiration of the embodiments, which are all within the protection of the embodiments.

Claims

1. A method for zero flux control of a current transformer based on magnetic potential balance constraint, characterized in that, The method comprises the following steps: Converting the current signal of the current transformer into a pulse sequence to generate an event-driven data format; Processing the pulse sequence by using an event-driven wavelet decomposition algorithm to output a multi-band pulse stream; Analyzing the multi-band pulse stream to calculate the coupling relationship between frequency components, including: for each pair of frequency components, extracting the pulse occurrence time sequence thereof; performing standardization processing on the time sequence to eliminate the difference in time scale of different frequency bands; calculating the time difference value after standardization and applying a time kernel function for weighting; performing normalization processing on the coupling weight of each frequency component to ensure that the total weight is 1; multiplying the normalized coupling weight to obtain the coupling strength; and constructing a pulse time sequence parameter relationship graph model based on the coupling strength; Based on the coupling relationship between the frequency components, the magnetic potential balance compensation parameter is calculated by using a pulse frequency coding algorithm, including: performing normalization processing on the magnetic flux data to standardize the magnetic flux value to the range of [-1, 1]; performing normalization processing on the pulse frequency code to make the numerical range of each frequency component consistent; combining the normalized magnetic flux data and the pulse frequency code to calculate the magnetic potential balance compensation parameter through a weighted nonlinear mapping function; Converting the magnetic potential balance compensation parameter into an asynchronous control pulse sequence to generate a compensation current and output the compensation current to a magnetic potential balance execution mechanism.

2. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, The step of converting the current signal of the current transformer into a pulse sequence comprises: Performing normalization processing on the original current signal to normalize the current amplitude to the range of [0, 1]; Applying a pulse density coding algorithm to convert the normalized current signal into a pulse sequence; Performing an adaptive threshold adjustment algorithm, when the change rate of the current signal exceeds a preset threshold, dynamically reducing the time window and increasing the pulse generation frequency.

3. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, The event-driven wavelet decomposition algorithm utilizes a pulse neural network, which is composed of the following components: Input layer: receiving the pulse sequence and distributing it to different processing channels; Feature extraction layer: composed of multiple pulse neurons, each neuron has different time-frequency filter characteristics; Frequency separation layer: grouping and clustering pulses with different frequency characteristics; Output layer: generating a multi-band pulse stream that retains the time-frequency characteristics of the original signal.

4. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, The coded pulse sequence is input into a neuromorphic processor for subsequent processing, including: Timestamp sorting of the pulse sequence to ensure that events are processed in the order of occurrence; Applying a pulse sparsification algorithm to remove redundant pulses and improve processing efficiency; Distributing the processed pulse sequence to different processing units of the neuromorphic processor; Achieving a response speed of microseconds through asynchronous parallel processing; Generating a standardized event stream.

5. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, Processing the pulse sequence by using an event-driven wavelet decomposition algorithm to output a multi-band pulse stream comprises: Inputting the pulse sequence into the event-driven wavelet decomposition algorithm, which is based on pulse triggering for calculation and is only executed when the signal changes; Performing pulse response transformation to convolve the input pulse with a preset wavelet-like basis function to extract the characteristics of different frequency components; Applying frequency-selective filtering to distribute the pulse to different processing channels according to its time-frequency characteristics; The aggregation result is used to generate a multi-band pulse stream output, each band corresponding to a frequency component of the current signal.

6. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, The step of converting the magnetic potential balance compensation parameter into an asynchronous control pulse sequence comprises: The compensation parameter is normalized to eliminate the dimensional and numerical range differences between different parameters; The normalized compensation parameter is mapped to the pulse amplitude and frequency space; The corresponding pulse density is generated according to the size and change rate of the parameter through a nonlinear mapping function.

7. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, The step of generating the compensation current comprises: The input control pulse is digitally filtered to remove high-frequency noise; The pulse sequence is data standardized to unify the time domain and frequency domain characteristics; A direct mapping function is applied to convert the standardized pulse sequence into a compensation current amplitude; The generated current amplitude is calibrated to ensure that it meets the input specifications of the actuator; The timing characteristics of the compensation current are adjusted according to the pulse timing characteristics.

8. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, The step of generating the compensation current and outputting it to the magnetic potential balance actuator comprises: The real-time change of the magnetic flux is monitored; The deviation of the magnetic flux from zero is calculated; The compensation current is adjusted in real time to make the magnetic flux approach zero.

9. The magnetic potential balance constraint based current transformer zero flux control method of claim 1, wherein, The generation process of the asynchronous control pulse sequence utilizes an asynchronous control event-driven model, which includes the following key components: Pulse generator: dynamically generates control pulses according to compensation parameters; Event queue manager: manages the ordering and scheduling of asynchronous pulse events; Direct digital-to-analog mapping unit: converts pulses to analog compensation current; Magnetic potential detection feedback unit: provides closed-loop control feedback.

10. A zero flux control system for a current transformer based on magnetic potential balance constraints, characterized in that, A method for performing the magnetic potential balance constraint-based current transformer zero flux control method of any one of claims 1-9, comprising: Pulse conversion module: for converting the current signal into a pulse sequence and generating event-driven data; Frequency band analysis module: for implementing event-driven wavelet decomposition and multi-band pulse stream output; Frequency coupling calculation module: for processing multi-band pulse stream and constructing the coupling relationship between frequency components; Magnetic potential balance calculation module: for generating magnetic potential balance compensation parameters based on the frequency coupling relationship; Control execution module: for converting the compensation parameters into control pulses and driving the magnetic potential balance execution device.

Images