A low electromagnetic interference power adapter control method and system

Abstract

This invention belongs to the field of power adapter technology, specifically a low electromagnetic interference power adapter control method and system. First, it synchronously acquires multi-source signals from the entire operating link of the power adapter, completing signal preprocessing, EMI time-frequency domain feature extraction, interference source localization, and operating condition matching and classification. Then, based on the interference source localization results, noise contribution ranking, and the current operating condition label set, it completes suppression priority ranking, control parameter decoupling, multi-objective optimization, and control strategy timing scheduling. Next, based on the EMI control strategy timing instruction set executable for each operating condition, it sequentially executes switch-driven regulation, spread spectrum modulation, active noise cancellation, and passive filter network parameter matching. Based on this, the invention significantly reduces noise intensity from the source of interference, significantly improves EMI suppression, and eliminates the need for additional multi-stage filtering devices and shielding structures, ensuring the miniaturization and low cost advantages of the power adapter.

Description

Technical Field

[0001] This invention relates to the field of power adapter technology, specifically to a power adapter control method and system with low electromagnetic interference. Background Technology

[0002] With the rapid development of consumer electronics, new energy, and smart home industries, high frequency, miniaturization, and high power density have become the core development trends of power adapters. In particular, the large-scale application of wide-bandgap semiconductor devices such as gallium nitride (GaN) and silicon carbide (SiC) has significantly increased the switching frequency and power density of power adapters. However, at the same time, the high dv / dt and di / dt generated by high-frequency switching can lead to serious electromagnetic interference (EMI) problems, including conducted and radiated interference. This not only affects the operating stability of the adapter itself but also interferes with the normal operation of surrounding electronic equipment. Therefore, it is necessary to meet the limit requirements of domestic and international electromagnetic compatibility standards such as EN55032, FCC Part 15, and GB9254.

[0003] In existing technologies, EMI suppression solutions for power adapters are mainly passive, primarily relying on passive filtering components such as common-mode inductors, X capacitors, Y capacitors, and RC snubber circuits at the adapter's input and output terminals, while also using metal shielding to suppress radiated interference. Some solutions employ fixed-parameter spread spectrum modulation techniques to reduce noise peaks, but these solutions suffer from several unresolved technical drawbacks. For example, they focus primarily on end-stage suppression, failing to target noise sources, resulting in limited EMI suppression and sacrificing miniaturization advantages. Specifically, in traditional flyback power adapters, the high dv / dt (up to 45kV / μs) generated during the switching process of the transistor and the high di / dt generated by the diode's reverse recovery are the core noise sources, contributing over 70% of the overall conducted and radiated noise. Existing solutions only suppress voltage spikes through RC / RCD snubber circuits, failing to precisely control the dv / dt and di / dt of the switching action, and can only achieve end-stage suppression after noise generation through multi-stage filtering and shielding. To meet the radiated noise limit requirements, an additional 2-3 stages of common-mode filter circuits and a metal shield with a thickness of 0.2mm or more are required, which increases the overall size of the adapter by 15%-25% and the material cost by 10%-20%, which goes against the development trend of high power density and miniaturization. Summary of the Invention

[0004] To address the aforementioned technical problems of focusing primarily on end-point suppression without achieving targeted suppression of noise sources, resulting in limited EMI suppression effects and sacrificing miniaturization advantages, this invention provides the following technical solution:

[0005] A power adapter control method with low electromagnetic interference includes the following specific steps:

[0006] S1, Multi-dimensional electromagnetic interference perception and feature extraction under all working conditions: Synchronously collects multi-source signals from the entire working link of the power adapter, completes signal preprocessing, EMI time-frequency domain feature extraction, interference source localization and working condition matching and classification, and outputs interference source localization results, noise contribution ranking and current working condition label set;

[0007] S2, Adaptive EMI Suppression Strategy Generation and Multi-Objective Optimization: Based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction steps under all operating conditions, the suppression priority ranking, control parameter decoupling, multi-objective optimization and control strategy timing scheduling are completed, and the set of EMI control strategy timing instructions executable under different operating conditions is output.

[0008] S3, multi-link coordinated EMI targeted suppression execution:

[0009] S31, Adaptive control of switching transistor drive waveform: Based on the adaptive EMI suppression strategy generation and multi-objective optimization steps, the timing instruction set of the EMI control strategy is executed under different working conditions. Through the adjustable gate drive circuit, the drive waveform parameters of the switching transistor are adjusted in real time. At the same time, based on the optimized parameters, the increase in switching loss is minimized while suppressing noise, and the real-time control execution result of the switching transistor drive waveform and the drive parameter feedback value are output.

[0010] S32, Adaptive Spread Spectrum Modulation and Frequency Jitter Execution: Based on the real-time control execution result of the switching transistor drive waveform and the drive parameter feedback value output by the adaptive control step of the switching transistor drive waveform, adaptive spread spectrum modulation and frequency jitter control are performed on the PWM drive signal of the switching transistor. According to the optimized modulation depth, frequency range, and modulation period, the noise energy concentrated at a fixed frequency is dispersed to a wide frequency range. At the same time, it is synchronized with the drive waveform control action and outputs the spread spectrum modulation execution result of the PWM signal and the modulation parameter feedback value.

[0011] S33, Active Common-Mode / Differential-Mode Noise Adaptive Cancellation: Based on the spread spectrum modulation execution result and modulation parameter feedback value of the PWM signal output from the adaptive spread spectrum modulation and frequency jitter execution steps, the active noise cancellation circuit collects the common-mode and differential-mode noise signals of the current input and output circuits in real time, and generates a compensation signal with the same amplitude and opposite phase as the original noise signal according to the optimized compensation phase and compensation amplitude parameters. This compensation signal is injected into the primary and secondary circuits to cancel the conducted noise in real time, and outputs the active noise cancellation execution result and compensation parameter feedback value.

[0012] S34, Adjustable passive filter network adaptive matching: Based on the active noise cancellation execution result and compensation parameter feedback value output by the common-mode / differential-mode noise adaptive cancellation step, the inductor and capacitor parameters of the adjustable passive filter network are switched in real time through a semiconductor switch array to match the current operating conditions and residual noise characteristics, and output the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action.

[0013] S4, Full Closed-Loop EMI Effect Verification and Strategy Iterative Optimization: Based on the passive filter network parameter matching execution results and the full-link suppression action completion feedback signal output by the multi-link collaborative EMI targeted suppression execution steps, the system completes EMI suppression effect verification, strategy deviation identification, control parameter iterative correction, and full life cycle model update, and outputs the iteratively corrected control parameter set.

[0014] In a preferred embodiment of the low electromagnetic interference power adapter control method of the present invention, the specific steps of step S1 are as follows:

[0015] S11, Multi-source signal synchronous acquisition: Through a multi-channel synchronous ADC sampling circuit, multi-dimensional raw signals of the entire working link of the power adapter are synchronously acquired. The sampling clock is synchronized with the PWM drive signal of the adapter's main switch transistor, and the synchronously acquired multi-source raw time domain signal set is output.

[0016] S12, Original signal preprocessing and noise reduction: Based on the synchronous acquisition of multi-source raw time-domain signals output from the multi-source signal synchronous acquisition step, perform signal preprocessing and output a preprocessed high signal-to-noise ratio multi-source standard signal set.

[0017] S13, Joint extraction of electromagnetic interference features in the time and frequency domains: Based on the preprocessed high signal-to-noise ratio multi-source standard signal set output from the original signal preprocessing and noise reduction steps, joint time and frequency domain feature extraction is performed, and the time and frequency domain features of the signal are extracted by wavelet packet decomposition to locate the time domain occurrence time of noise; at the same time, the peak values ​​of dv / dt and di / dt of the switching node, the proportion of common-mode / differential-mode noise components, and operating condition characteristic parameters are extracted, and the EMI time and frequency domain characteristic parameter set and the operating condition characteristic parameter set are output;

[0018] S14, Interference Source Localization and Operating Condition Matching Classification: Based on the EMI time-frequency domain feature parameter set and operating condition feature parameter set output by the joint extraction steps of electromagnetic interference features in the time and frequency domains, a pre-trained lightweight gradient boosting decision tree classification model is used to accurately locate the noise interference source; at the same time, based on the operating condition feature parameters, the matching classification of the current operating condition is completed, and the interference source localization result, noise contribution ranking and current operating condition label set are output.

[0019] In a preferred embodiment of the low electromagnetic interference power adapter control method of the present invention, the specific steps of S2 are as follows:

[0020] S21, Interference Source Target Suppression Priority Ranking: Based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction steps under full operating conditions, combined with the noise limit requirements of the corresponding electromagnetic compatibility standards, the suppression priority of each interference source is ranked, and the suppression target value and limit margin requirements of each interference source are determined to form a priority ranking table, and the interference source suppression priority ranking table and suppression target parameter set are output.

[0021] S22, Device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation: Based on the interference source suppression priority ranking table and suppression target parameter set output by the interference source targeted suppression priority ranking step, combined with the iteratively corrected control parameter set output by the full closed-loop EMI effect verification and strategy iterative optimization step, device characteristic full life cycle adaptive compensation and transient condition prediction and strategy pre-generation processing are performed, and the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules is output;

[0022] S23, Decoupling of multiple control parameters and generation of initial strategy: Based on the device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation steps, the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules is decoupled from the multi-dimensional control parameters affecting EMI; at the same time, based on priority sorting, the initial suppression control parameter set corresponding to each interference source is generated, and the initial EMI suppression multi-dimensional control parameter set is output.

[0023] S24, Strategy optimization under multi-objective constraints: Based on the initial EMI suppression multi-dimensional control parameter set output by the multi-control parameter decoupling and initial strategy generation steps, a multi-objective optimization function and constraint conditions are constructed. An adaptive weighted particle swarm optimization algorithm is used to iteratively optimize the initial control parameter set and output the optimized EMI suppression optimal multi-dimensional control parameter set.

[0024] S25, Control Strategy Partition Mapping and Timing Scheduling: Based on the optimized multi-dimensional control parameter set for EMI suppression output by the strategy optimization steps under multi-objective constraints, the optimal control parameters are mapped to the corresponding operating condition partitions, establishing a one-to-one mapping table between operating conditions and control parameters; simultaneously, combined with the calibrated suppression target parameter set with device characteristic compensation benchmark and transient operating condition pre-control rules, timing scheduling rules for multiple control actions are formulated, the execution timing and synchronization triggering rules for each control parameter are set, and finally, timing control instructions that can be directly issued and executed are generated, outputting a set of EMI control strategy timing instructions executable for each operating condition.

[0025] In a preferred embodiment of the low electromagnetic interference power adapter control method of the present invention, the specific steps of step S4 are as follows:

[0026] S41, Real-time Verification of EMI Suppression Effect: Based on the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action output by the multi-link collaborative EMI targeted suppression execution steps, the comprehensive performance parameters are re-acquired through a multi-channel synchronous ADC sampling circuit, and the acquired noise signal is compared with the original noise signal to calculate the noise suppression amount and standard limit margin of each frequency band. At the same time, it is verified whether the comprehensive performance parameters meet the constraints, and the EMI suppression effect verification dataset and the power supply comprehensive performance parameter set are output.

[0027] S42, Control Strategy Effectiveness Evaluation and Deviation Identification: Based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output from the real-time EMI suppression effect verification step, compare them with the set post-calibration suppression target and constraints to evaluate whether the current control strategy's execution effect meets the standards; if there are cases of noise exceeding the standard or performance parameters exceeding the constraint range, identify the exceeding frequency band, the magnitude and direction of the deviation parameters, locate the cause of the deviation, and output a strategy effectiveness evaluation report, deviation parameter set, and deviation cause location results;

[0028] S43, Strategy Iteration Correction: Based on the strategy effect evaluation report, deviation parameter set and deviation cause location results output by the control strategy effect evaluation and deviation identification steps, execute the corresponding iterative optimization operation and output the iteratively corrected control parameter set;

[0029] S44, Fault Warning and Safety Protection Execution: Based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the real-time EMI suppression effect verification step, the strategy effect evaluation report, deviation parameter set and deviation cause location results output by the control strategy effect evaluation and deviation identification step, the abnormal state of the system is monitored in real time, and when an abnormal state is detected, the corresponding safety protection action is executed according to the abnormality level.

[0030] A low electromagnetic interference power adapter control system, comprising:

[0031] The full-condition, multi-dimensional electromagnetic interference sensing and feature extraction module synchronously collects multi-source signals from the entire working link of the power adapter, completes signal preprocessing, EMI time-frequency domain feature extraction, interference source localization and working condition matching and classification, and outputs interference source localization results, noise contribution ranking and current working condition label set;

[0032] The adaptive EMI suppression strategy generation and multi-objective optimization module, based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction module under all operating conditions, completes suppression priority ranking, control parameter decoupling, multi-objective optimization and control strategy timing scheduling, and outputs a set of EMI control strategy timing instructions that can be executed under different operating conditions.

[0033] The multi-link collaborative EMI targeted suppression execution module includes a switch drive waveform adaptive adjustment unit, an adaptive spread spectrum modulation and frequency jitter execution unit, an active common-mode / differential-mode noise adaptive cancellation unit, and an adjustable passive filter network adaptive matching unit.

[0034] The adaptive control unit for the switching transistor drive waveform, based on the timing instruction set of the EMI control strategy that can be executed under different operating conditions and generated by the adaptive EMI suppression strategy and the multi-objective optimization module, adjusts the driving waveform parameters of the switching transistor in real time through the adjustable gate drive circuit; at the same time, based on the optimized parameters, it minimizes the increase in switching losses while suppressing noise, and outputs the real-time control execution result of the switching transistor drive waveform and the feedback value of the drive parameters.

[0035] The adaptive spread spectrum modulation and frequency jitter execution unit, based on the real-time control execution result of the switching transistor driving waveform and the driving parameter feedback value output by the adaptive control unit of the switching transistor driving waveform, performs adaptive spread spectrum modulation and frequency jitter control on the PWM driving signal of the switching transistor. According to the optimized modulation depth, frequency range, and modulation period, the noise energy concentrated at a fixed frequency is dispersed to a wide frequency range. At the same time, it keeps synchronized with the driving waveform control action and outputs the spread spectrum modulation execution result of the PWM signal and the modulation parameter feedback value.

[0036] The active common-mode / differential-mode noise adaptive cancellation unit, based on the spread spectrum modulation execution result and modulation parameter feedback value of the PWM signal output by the adaptive spread spectrum modulation and frequency jitter execution unit, collects the common-mode and differential-mode noise signals of the current input and output circuits in real time through the active noise cancellation circuit, and generates a compensation signal with the same amplitude and opposite phase as the original noise according to the optimized compensation phase and compensation amplitude parameters. The compensation signal is injected into the primary and secondary circuits to cancel the conducted noise in real time, and outputs the active noise cancellation execution result and compensation parameter feedback value.

[0037] The adjustable passive filter network adaptive matching unit, based on the active noise cancellation execution result and compensation parameter feedback value output by the common-mode / differential-mode noise adaptive cancellation unit, switches the inductance and capacitance parameters of the adjustable passive filter network in real time through a semiconductor switch array to match the current operating conditions and residual noise characteristics, and outputs the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action.

[0038] The fully closed-loop EMI effect verification and strategy iteration optimization module, based on the passive filter network parameter matching execution results and the full-link suppression action completion feedback signal output by the multi-link collaborative EMI targeted suppression execution module, completes EMI suppression effect verification, strategy deviation identification, control parameter iterative correction and full life cycle model update, and outputs the iteratively corrected control parameter set.

[0039] As a preferred embodiment of the low electromagnetic interference power adapter control system described in this invention, the full-condition multi-dimensional electromagnetic interference sensing and feature extraction module includes:

[0040] The multi-source signal synchronous acquisition unit uses a multi-channel synchronous ADC sampling circuit to synchronously acquire multi-dimensional raw signals from the entire working link of the power adapter. The sampling clock is synchronized with the PWM drive signal of the adapter's main switch transistor, and the output is a set of synchronously acquired multi-source raw time-domain signals.

[0041] The raw signal preprocessing and noise reduction unit performs signal preprocessing based on the synchronously acquired multi-source raw time-domain signal set output by the multi-source signal synchronous acquisition unit, and outputs a preprocessed high signal-to-noise ratio multi-source standard signal set.

[0042] The electromagnetic interference feature extraction unit, based on the preprocessed high signal-to-noise ratio multi-source standard signal set output by the original signal preprocessing and noise reduction unit, performs joint time-domain and frequency-domain feature extraction. It also extracts the time-frequency domain features of the signal through wavelet packet decomposition to locate the time-domain occurrence time of the noise. Simultaneously, it extracts the peak values ​​of dv / dt and di / dt of the switching node, the proportion of common-mode / differential-mode noise components, and operating condition characteristic parameters, and outputs the EMI time-frequency domain feature parameter set and the operating condition feature parameter set.

[0043] The interference source localization and operating condition matching classification unit, based on the EMI time-frequency domain feature parameter set and operating condition feature parameter set output by the joint extraction unit of electromagnetic interference features in the time and frequency domains, accurately locates noise interference sources through a pre-trained lightweight gradient boosting decision tree classification model; at the same time, based on the operating condition feature parameters, it completes the matching and classification of the current operating condition, and outputs the interference source localization results, noise contribution ranking, and current operating condition label set.

[0044] As a preferred embodiment of the low electromagnetic interference power adapter control system described in this invention, the adaptive EMI suppression strategy generation and multi-objective optimization module includes:

[0045] The interference source targeted suppression priority ranking unit, based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction module under all operating conditions, and combined with the noise limit requirements of the corresponding electromagnetic compatibility standard, ranks the suppression priority of each interference source, determines the suppression target value and limit margin requirement of each interference source, forms a priority ranking table, and outputs the interference source suppression priority ranking table and suppression target parameter set.

[0046] The device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation unit, based on the interference source suppression priority ranking table and suppression target parameter set output by the interference source targeted suppression priority ranking unit, combined with the iteratively corrected control parameter set output by the full closed-loop EMI effect verification and strategy iteration optimization module, performs device characteristic full life cycle adaptive compensation and transient condition prediction and strategy pre-generation processing, and outputs a calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules;

[0047] The multi-control parameter decoupling and initial strategy generation unit, based on the device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation unit outputting the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules, decouples the multi-dimensional control parameters affecting EMI; at the same time, based on priority sorting, it generates the initial suppression control parameter set corresponding to each interference source and outputs the initial EMI suppression multi-dimensional control parameter set.

[0048] The strategy optimization unit under multi-objective constraints constructs a multi-objective optimization function and constraint conditions based on the initial EMI suppression multi-dimensional control parameter set output by the multi-control parameter decoupling and initial strategy generation unit. It then uses an adaptive weighted particle swarm optimization algorithm to iteratively optimize the initial control parameter set and outputs the optimized EMI suppression optimal multi-dimensional control parameter set.

[0049] The control strategy partition mapping and timing scheduling unit, based on the optimized multi-dimensional control parameter set for EMI suppression output by the strategy optimization unit under multi-objective constraints, maps the optimal control parameters to the corresponding operating condition partitions, establishing a one-to-one mapping table between operating conditions and control parameters. Simultaneously, combining the calibrated suppression target parameter set with device characteristic compensation reference and transient operating condition pre-control rules, it formulates timing scheduling rules for multiple control actions, sets the execution timing and synchronization triggering rules for each control parameter, and finally generates timing control instructions that can be directly issued and executed, outputting a set of EMI control strategy timing instructions executable for each operating condition.

[0050] As a preferred embodiment of the low electromagnetic interference power adapter control system described in this invention, the fully closed-loop EMI effect verification and strategy iteration optimization module includes:

[0051] The real-time EMI suppression effect verification unit, based on the passive filter network parameter matching execution result and the full-link suppression action completion feedback signal output by the multi-link collaborative EMI targeted suppression execution module, re-acquires the comprehensive performance parameters through a multi-channel synchronous ADC sampling circuit, compares the acquired noise signal with the original noise signal, calculates the noise suppression amount and standard limit margin of each frequency band, and verifies whether the comprehensive performance parameters meet the constraints, outputting the EMI suppression effect verification dataset and the power supply comprehensive performance parameter set;

[0052] The control strategy effectiveness evaluation and deviation identification unit compares the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the EMI suppression effect real-time verification unit with the set post-calibration suppression target and constraints to evaluate whether the current control strategy's execution effect meets the standards. If there are cases of noise exceeding the standard or performance parameters exceeding the constraint range, the unit identifies the frequency band exceeding the standard, the magnitude and direction of the deviation parameters, locates the cause of the deviation, and outputs a strategy effectiveness evaluation report, deviation parameter set, and deviation cause location results.

[0053] The strategy iteration and correction unit, based on the strategy effect evaluation report, deviation parameter set, and deviation cause location results output by the control strategy effect evaluation and deviation identification unit, performs corresponding iterative optimization operations and outputs the iteratively corrected control parameter set.

[0054] The fault warning and safety protection execution unit monitors the abnormal state of the system in real time based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the EMI suppression effect real-time verification unit, the strategy effect evaluation report, deviation parameter set and deviation cause location results output by the control strategy effect evaluation and deviation identification unit, and executes the corresponding safety protection action according to the abnormality level when an abnormal state is detected.

[0055] Compared with existing technologies:

[0056] 1. Through multi-dimensional signal sensing under all operating conditions and dynamic matching of adaptive control parameters, it can achieve dynamic EMI suppression with wide input voltage and full load range, taking into account both EMI compliance under all operating conditions and power conversion efficiency, thus resolving the inherent contradiction of fixed parameter schemes.

[0057] 2. By precisely controlling the waveform driven by the switching transistor, targeted suppression of noise sources is achieved. Combined with multi-link coordinated control of active noise cancellation along the propagation path and adaptive filtering at the end, it can significantly reduce noise intensity from the source of interference and significantly improve EMI suppression. No additional multi-stage filtering devices and shielding structures are required, ensuring the advantages of miniaturization and low cost of the power adapter.

[0058] 3. Through an adaptive optimization algorithm under multi-objective constraints, it can achieve coordinated optimization of multiple core indicators such as EMI suppression, conversion efficiency, and voltage regulation accuracy, thus solving the technical defects of strong coupling and compromise between various performance indicators. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the overall framework of the present invention;

[0060] Figure 2 This is a schematic diagram of the framework of the multi-dimensional electromagnetic interference sensing and feature extraction module under all working conditions of the present invention;

[0061] Figure 3 This is a schematic diagram of the adaptive EMI suppression strategy generation and multi-objective optimization module framework of the present invention;

[0062] Figure 4 This is a schematic diagram of the multi-link collaborative EMI targeted suppression execution module framework of the present invention;

[0063] Figure 5 This is a schematic diagram of the framework of the fully closed-loop EMI effect verification and strategy iteration optimization module of the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0065] This invention provides a low-electromagnetic-interference power adapter control method. Please refer to [link / reference]. Figure 1 The specific steps are as follows:

[0066] S1, Multi-dimensional electromagnetic interference perception and feature extraction under all working conditions: Synchronously collects multi-source signals from the entire working link of the power adapter, completes signal preprocessing, EMI time-frequency domain feature extraction, interference source localization and working condition matching and classification, and outputs interference source localization results, noise contribution ranking and current working condition label set;

[0067] The specific steps of S1 are as follows:

[0068] S11, Multi-source signal synchronous acquisition: Through a multi-channel synchronous ADC sampling circuit, multi-dimensional raw signals of the entire working link of the power adapter are synchronously acquired. The sampling clock is synchronized with the PWM drive signal of the adapter's main switch transistor to ensure that the time axis of all signals is aligned. The acquired signals include: AC voltage / current signal on the grid input side, voltage / current waveform of the main switch node, common-mode current signal of the transformer primary and secondary sides, DC voltage / current / conduction noise signal on the output side, spatial radiation noise signal, junction temperature signal of the switch transistor and rectifier diode, and current load power signal. The output is a set of synchronously acquired multi-source raw time-domain signals.

[0069] S12, Original Signal Preprocessing and Noise Reduction: Based on the synchronously acquired multi-source raw time-domain signal set output from the multi-source signal synchronous acquisition step, signal preprocessing is performed: sampling white noise is removed by moving average filtering, power frequency fundamental and harmonic interference is removed by wavelet transform, signal DC offset is eliminated by baseline correction, all signals are mapped to the standard amplitude range by linear normalization, and signal bad point removal and missing value completion are completed at the same time to obtain a standard signal with high signal-to-noise ratio and strict time axis alignment, and the preprocessed high signal-to-noise ratio multi-source standard signal set is output.

[0070] S13, Joint Time-Domain and Frequency-Domain Extraction of Electromagnetic Interference Features: Based on the preprocessed high signal-to-noise ratio multi-source standard signal set output from the original signal preprocessing and noise reduction steps, joint time-domain and frequency-domain feature extraction is performed: the time-domain signal is converted into a frequency-domain spectrum through Fast Fourier Transform (FFT), and the peak amplitude, center frequency, bandwidth, and limit margin of noise in each frequency band are extracted; the time-frequency domain features of the signal are extracted through wavelet packet decomposition to locate the time-domain occurrence time of the noise; at the same time, the peak values ​​of dv / dt and di / dt of the switching nodes, the proportion of common-mode / differential-mode noise components, the effective value of the input voltage, the load rate, the junction temperature of the device, and other operating condition characteristic parameters are extracted, and the EMI time-frequency domain feature parameter set and the operating condition feature parameter set are output;

[0071] S14, Interference Source Localization and Operating Condition Matching Classification: Based on the EMI time-frequency domain feature parameter set and operating condition feature parameter set output from the joint extraction steps of electromagnetic interference features in the time and frequency domains, a pre-trained lightweight gradient boosting decision tree classification model is used to accurately locate noise interference sources, identify core interference sources (switch turn-on action, switch turn-off action, rectifier diode reverse recovery, transformer leakage inductance resonance, input grid harmonics) and the noise contribution of each interference source; at the same time, based on the operating condition feature parameters, the current operating condition matching classification is completed, and corresponding operating condition labels are assigned (undervoltage condition, rated condition, overvoltage condition, no-load condition, light-load condition, half-load condition, full-load condition, overload condition), and the interference source localization results, noise contribution ranking, and current operating condition label set are output.

[0072] S2, Adaptive EMI Suppression Strategy Generation and Multi-Objective Optimization: Based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction steps under all operating conditions, the suppression priority ranking, control parameter decoupling, multi-objective optimization and control strategy timing scheduling are completed, and the set of EMI control strategy timing instructions executable under different operating conditions is output.

[0073] The specific steps of S2 are as follows:

[0074] S21, Interference Source Target Suppression Priority Ranking: Based on the interference source location results, noise contribution ranking, and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction steps under full operating conditions, combined with the noise limit requirements of the corresponding electromagnetic compatibility standards, the suppression priority of each interference source is ranked according to the core principles of "priority for exceeding the standard, priority for high contribution, and priority for operating condition sensitivity". At the same time, the suppression target value and limit margin requirements of each interference source are determined to form a priority ranking table, and the interference source suppression priority ranking table and suppression target parameter set are output.

[0075] S22, Device Characteristic Lifecycle Adaptive Compensation and Transient Condition Strategy Pre-generation: Based on the interference source suppression priority ranking table and suppression target parameter set output by the interference source targeted suppression priority ranking step, combined with the iteratively corrected control parameter set output by the full closed-loop EMI effect verification and strategy iteration optimization step, device characteristic lifecycle adaptive compensation and transient condition prediction and strategy pre-generation processing are performed: Device characteristic lifecycle adaptive compensation: Extract the characteristic drift law of core power devices such as switching transistors, rectifier diodes, transformers, and filter capacitors in the historical samples of the entire lifecycle, and combine it with the device junction temperature and running time data of the current operating condition to pre-compensate and calibrate the benchmark value of the suppression target parameter set to eliminate the EMI suppression effect attenuation caused by device temperature drift and aging; Transient condition prediction and strategy pre-generation: Based on the historical operating condition sample library, the triggering law and interference characteristics of transient conditions such as input voltage fluctuation and load jump are identified through the time-series prediction algorithm, and the transient conditions are predicted in combination with the current operating condition change trend. Corresponding emergency suppression strategies and parameter switching rules are pre-generated, and the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules is output;

[0076] S23, Decoupling of Multiple Control Parameters and Generation of Initial Strategy: Based on the device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation steps, a calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules is generated. This decouples the multi-dimensional control parameters affecting EMI and clarifies the mapping relationship between each control parameter and its corresponding interference source. The decoupled control parameters include: the turn-on slope, turn-off slope, and dead time parameters of the gate drive of the switching transistor; the spread spectrum modulation depth, frequency jitter range, and modulation period parameters of the PWM signal; the compensation phase and compensation amplitude parameters of active noise cancellation; and the inductor and capacitor switching parameters of the adjustable passive filter network. Simultaneously, based on priority sorting, an initial suppression control parameter set corresponding to each interference source is generated, and the initial EMI suppression multi-dimensional control parameter set is output.

[0077] S24, Strategy Optimization under Multi-Objective Constraints: Based on the initial EMI suppression multi-dimensional control parameter set output from the multi-control parameter decoupling and initial strategy generation steps, a multi-objective optimization function and constraints are constructed. The optimization objectives are "maximizing EMI noise limit margin, maximizing power conversion efficiency, and minimizing output voltage ripple". The constraints are "output voltage regulation accuracy ≤ ±1%, switching junction temperature ≤ 125℃, dynamic response time ≤ 100μs, and standby power consumption ≤ 50mW". An adaptive weighted particle swarm optimization algorithm is used to iteratively optimize the initial control parameter set, eliminate the coupling interference between various control parameters, and obtain the optimal control parameter set that takes into account both EMI suppression and overall power performance. The optimized EMI suppression optimal multi-dimensional control parameter set is output.

[0078] S25, Control Strategy Partition Mapping and Timing Scheduling: Based on the optimized multi-dimensional control parameter set for EMI suppression output from the strategy optimization steps under multi-objective constraints, the optimal control parameters are mapped to the corresponding operating condition partitions, establishing a one-to-one mapping table between operating conditions and control parameters. Simultaneously, combined with the calibrated suppression target parameter set with device characteristic compensation benchmark and transient operating condition pre-control rules, timing scheduling rules for multiple control actions are formulated. Following the order of "source control first, path suppression followed, and end filtering to finish," the execution timing and synchronization triggering rules of each control parameter are set to avoid multiple control actions being executed simultaneously, generating new beat frequency interference and resonant noise. Finally, timing control instructions that can be directly issued and executed are generated, outputting a set of EMI control strategy timing instructions executable for each operating condition.

[0079] S3, Multi-link Coordinated EMI Targeted Suppression Execution: Based on the adaptive EMI suppression strategy generation and multi-objective optimization step output, the working condition executable EMI control strategy timing instruction set is executed sequentially, including switch drive control, spread spectrum modulation, active noise cancellation and passive filter network parameter matching, and outputs the passive filter network parameter matching execution result and the full-link suppression action completion feedback signal.

[0080] The specific steps of S3 are as follows:

[0081] S31, Adaptive Control of Switching Transistor Drive Waveform: Based on the adaptive EMI suppression strategy and the multi-objective optimization steps, an executable EMI control strategy timing instruction set for different operating conditions is generated. Through an adjustable gate drive circuit, the drive waveform parameters of the switching transistor are adjusted in real time, including the turn-on slope, turn-off slope, and dead time. The peak values ​​of dv / dt and di / dt of the switching node are precisely controlled, reducing the intensity of the core interference source from the source of noise generation. At the same time, based on the optimized parameters, the increase in switching losses is minimized while suppressing noise, ensuring the power conversion efficiency. The real-time control execution result of the switching transistor drive waveform and the feedback value of the drive parameters are output.

[0082] S32, Adaptive Spread Spectrum Modulation and Frequency Jitter Execution: Based on the real-time control execution result of the switching transistor drive waveform and the drive parameter feedback value output by the adaptive control step of the switching transistor drive waveform, adaptive spread spectrum modulation and frequency jitter control are performed on the PWM drive signal of the switching transistor. According to the optimized modulation depth, frequency range, and modulation period, the noise energy concentrated at a fixed frequency is dispersed to a wide frequency range, which greatly reduces the noise peak value. At the same time, it is synchronized with the drive waveform control action to avoid beat frequency interference caused by frequency jitter and drive slope adjustment. The spread spectrum modulation execution result and modulation parameter feedback value of the PWM signal are output.

[0083] S33, Active Common-Mode / Differential-Mode Noise Adaptive Cancellation: Based on the spread spectrum modulation execution result and modulation parameter feedback value of the PWM signal output from the adaptive spread spectrum modulation and frequency jitter execution steps, the active noise cancellation circuit collects the common-mode and differential-mode noise signals of the current input and output circuits in real time, and generates a compensation signal with the same amplitude and opposite phase as the original noise signal according to the optimized compensation phase and compensation amplitude parameters. This compensation signal is injected into the primary and secondary circuits to cancel the conducted noise in real time, suppress the noise propagation path, and output the active noise cancellation execution result and compensation parameter feedback value.

[0084] S34, Adjustable Passive Filter Network Adaptive Matching: Based on the active noise cancellation execution result and compensation parameter feedback value output from the common-mode / differential-mode noise adaptive cancellation step, the inductance and capacitance parameters of the adjustable passive filter network are switched in real time through a semiconductor switch array to match the current operating conditions and residual noise characteristics. This complements the active noise cancellation of the preceding stage, providing final suppression of residual conducted and radiated noise. At the same time, it avoids the additional losses of the fixed filter network under light load conditions, improves the efficiency under all operating conditions, and outputs the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action.

[0085] S4, Full Closed-Loop EMI Effect Verification and Strategy Iterative Optimization: Based on the passive filter network parameter matching execution results and full-link suppression action completion feedback signals output by the multi-link collaborative EMI targeted suppression execution steps, the EMI suppression effect verification, strategy deviation identification, control parameter iterative correction and full life cycle model update are completed, and the iteratively corrected control parameter set is output.

[0086] The specific steps of S4 are as follows:

[0087] S41, Real-time Verification of EMI Suppression Effect: Based on the execution results of the passive filter network parameters matched with the feedback signal of the completion of the full-link suppression action, the noise signals of the adapter input side, output side, and space are re-acquired through a multi-channel synchronous ADC sampling circuit, as well as comprehensive performance parameters such as power conversion efficiency, output voltage regulation accuracy, device junction temperature, and dynamic response time. The acquired noise signals are compared with the original noise signals to calculate the noise suppression amount and standard limit margin of each frequency band. At the same time, it is verified whether the comprehensive performance parameters meet the constraints, and the EMI suppression effect verification dataset and the power supply comprehensive performance parameter set are output.

[0088] S42, Control Strategy Effectiveness Evaluation and Deviation Identification: Based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output from the real-time EMI suppression effect verification step, compare them with the set post-calibration suppression target and constraints to evaluate whether the execution effect of the current control strategy meets the standards; if there are cases of noise exceeding the standard or performance parameters exceeding the constraint range, identify the frequency band exceeding the standard, the magnitude and direction of the deviation parameters, locate the cause of the deviation (sudden change in operating conditions, parameter execution deviation, change in interference source, device characteristic drift), and output a strategy effectiveness evaluation report, deviation parameter set, and deviation cause location results;

[0089] S43, Strategy Iterative Correction: Based on the strategy effect evaluation report, deviation parameter set, and deviation cause location results output by the control strategy effect evaluation and deviation identification steps, perform the corresponding iterative optimization operation: If the current strategy effect meets the standard, store the current operating condition, control parameters, suppression effect, and device characteristic parameters into the full life cycle sample database as training samples for model optimization; if the effect does not meet the standard, iteratively correct the control parameter set according to the deviation parameters, regenerate the optimized control strategy, and output the iteratively corrected control parameter set;

[0090] S44, Fault Warning and Safety Protection Execution: Based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output from the real-time EMI suppression effect verification step, the strategy effect evaluation report, deviation parameter set and deviation cause location results output from the control strategy effect evaluation and deviation identification step, the abnormal state of the system is monitored in real time, including continuous noise exceeding the standard, device over-temperature, output overvoltage / overcurrent, input undervoltage / overvoltage, and device short circuit / open circuit faults; and when an abnormal state is detected, the corresponding safety protection action is executed according to the abnormality level, including power reduction operation, locking the PWM drive signal, and disconnecting the input relay, while generating corresponding fault warning codes and abnormal logs to achieve full life cycle safety protection of the system.

[0091] Please refer to a low electromagnetic interference power adapter control system. Figure 1 ,include:

[0092] The full-condition, multi-dimensional electromagnetic interference sensing and feature extraction module synchronously collects multi-source signals from the entire working link of the power adapter, completes signal preprocessing, EMI time-frequency domain feature extraction, interference source localization and working condition matching and classification, and outputs interference source localization results, noise contribution ranking and current working condition label set;

[0093] The adaptive EMI suppression strategy generation and multi-objective optimization module, based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction module under all operating conditions, completes suppression priority ranking, control parameter decoupling, multi-objective optimization and control strategy timing scheduling, and outputs a set of EMI control strategy timing instructions that can be executed under different operating conditions.

[0094] The multi-link collaborative EMI targeted suppression execution module, based on the working condition executable EMI control strategy timing instruction set output by the adaptive EMI suppression strategy generation and multi-objective optimization module, sequentially executes switch drive control, spread spectrum modulation, active noise cancellation and passive filter network parameter matching, and outputs the passive filter network parameter matching execution result and the full-link suppression action completion feedback signal.

[0095] The fully closed-loop EMI effect verification and strategy iteration optimization module, based on the passive filter network parameter matching execution results and the full-link suppression action completion feedback signal output by the multi-link collaborative EMI targeted suppression execution module, completes EMI suppression effect verification, strategy deviation identification, control parameter iterative correction and full life cycle model update, and outputs the iteratively corrected control parameter set.

[0096] Please see Figure 2 The full-condition multi-dimensional electromagnetic interference sensing and feature extraction module includes:

[0097] The multi-source signal synchronous acquisition unit uses a multi-channel synchronous ADC sampling circuit to synchronously acquire multi-dimensional raw signals from the entire working link of the power adapter. The sampling clock is synchronized with the PWM drive signal of the adapter's main switch transistor to ensure that the time axis of all signals is aligned. The acquired signals include: AC voltage / current signals on the grid input side, voltage / current waveforms of the main switch node, common-mode current signals of the transformer primary and secondary sides, DC voltage / current / conduction noise signals on the output side, spatial radiation noise signals, junction temperature signals of the switch transistor and rectifier diode, and current load power signals. The unit outputs a set of synchronously acquired multi-source raw time-domain signals.

[0098] The raw signal preprocessing and noise reduction unit performs signal preprocessing based on the synchronously acquired multi-source raw time-domain signal set output by the multi-source signal synchronous acquisition unit: removing sampling white noise through moving average filtering, removing power frequency fundamental and harmonic interference through wavelet transform, eliminating signal DC offset through baseline correction, mapping all signals to the standard amplitude range through linear normalization, and simultaneously completing signal defect elimination and missing value completion to obtain a standard signal with high signal-to-noise ratio and strict time axis alignment, and outputting a preprocessed high signal-to-noise ratio multi-source standard signal set.

[0099] The electromagnetic interference (EMI) feature extraction unit, based on the preprocessed high signal-to-noise ratio (SNR) multi-source standard signal set output by the original signal preprocessing and noise reduction unit, performs joint time-domain and frequency-domain feature extraction: It converts the time-domain signal into a frequency-domain spectrum using Fast Fourier Transform (FFT) to extract the peak amplitude, center frequency, bandwidth, and limit margin of noise in each frequency band; it also extracts the time-frequency domain features of the signal through wavelet packet decomposition to locate the time-domain occurrence of noise; simultaneously, it extracts the peak values ​​of dv / dt and di / dt of the switching nodes, the proportion of common-mode / differential-mode noise components, the effective value of the input voltage, load rate, device junction temperature, and other operating condition characteristic parameters, outputting an EMI time-frequency domain feature parameter set and an operating condition feature parameter set.

[0100] The interference source localization and operating condition matching classification unit, based on the EMI time-frequency domain feature parameter set and operating condition feature parameter set output by the joint extraction unit of electromagnetic interference features in the time and frequency domains, accurately locates noise interference sources through a pre-trained Lightweight Gradient Boosting Decision Tree (LightGBM) classification model. It identifies core interference sources (switch turn-on action, switch turn-off action, rectifier diode reverse recovery, transformer leakage inductance resonance, input grid harmonics) and the noise contribution of each interference source. Simultaneously, based on the operating condition feature parameters, it completes the matching and classification of the current operating condition, classifies the corresponding operating condition labels (undervoltage condition, rated condition, overvoltage condition, no-load condition, light-load condition, half-load condition, full-load condition, overload condition), and outputs the interference source localization results, noise contribution ranking, and the current operating condition label set.

[0101] Please see Figure 3 The adaptive EMI suppression strategy generation and multi-objective optimization module includes:

[0102] The interference source targeted suppression priority ranking unit, based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction module under all operating conditions, combined with the noise limit requirements of the corresponding electromagnetic compatibility standard, ranks the suppression priority of each interference source according to the core principle of "priority for exceeding the standard, priority for high contribution, and priority for operating condition sensitivity". At the same time, it determines the suppression target value and limit margin requirement of each interference source, forms a priority ranking table, and outputs the interference source suppression priority ranking table and suppression target parameter set.

[0103] The device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation unit, based on the interference source suppression priority ranking table and suppression target parameter set output by the interference source targeted suppression priority ranking unit, and combined with the iteratively corrected control parameter set output by the full closed-loop EMI effect verification and strategy iteration optimization module, performs device characteristic full life cycle adaptive compensation and transient condition prediction and strategy pre-generation processing: Device characteristic full life cycle adaptive compensation: Extracts the characteristic drift law of core power devices such as switching transistors, rectifier diodes, transformers, and filter capacitors in the historical samples of the full life cycle, and combines it with the device junction temperature and running time data of the current operating condition to pre-compensate and calibrate the benchmark value of the suppression target parameter set to eliminate the EMI suppression effect attenuation caused by device temperature drift and aging; Transient condition prediction and strategy pre-generation: Based on the historical operating condition sample library, the triggering law and interference characteristics of transient conditions such as input voltage fluctuations and load jumps are identified through time-series prediction algorithms, and transient conditions are predicted in combination with the current operating condition change trend. Corresponding emergency suppression strategies and parameter switching rules are pre-generated, and the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules is output;

[0104] The multi-control parameter decoupling and initial strategy generation unit, based on the device characteristic full-lifecycle adaptive compensation and transient condition strategy pre-generation unit outputting the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules, decouples the multi-dimensional control parameters affecting EMI and clarifies the mapping relationship between each control parameter and the corresponding interference source. The decoupled control parameters include: the turn-on slope, turn-off slope, and dead time parameters of the gate drive of the switching transistor; the spread spectrum modulation depth, frequency jitter range, and modulation period parameters of the PWM signal; the compensation phase and compensation amplitude parameters of active noise cancellation; and the inductor and capacitor switching parameters of the adjustable passive filter network. At the same time, based on priority sorting, it generates the initial suppression control parameter set corresponding to each interference source and outputs the initial EMI suppression multi-dimensional control parameter set.

[0105] The strategy optimization unit under multi-objective constraints constructs a multi-objective optimization function and constraints based on the initial EMI suppression multi-dimensional control parameter set output by the multi-control parameter decoupling and initial strategy generation unit. The optimization objectives are "maximizing EMI noise limit margin, maximizing power conversion efficiency, and minimizing output voltage ripple". The constraints are "output voltage regulation accuracy ≤ ±1%, switching junction temperature ≤ 125℃, dynamic response time ≤ 100μs, and standby power consumption ≤ 50mW". The adaptive weighted particle swarm optimization algorithm is used to iteratively optimize the initial control parameter set, eliminate the coupling interference between various control parameters, and obtain the optimal control parameter set that takes into account both EMI suppression and overall power performance. The optimized EMI suppression optimal multi-dimensional control parameter set is output.

[0106] The control strategy partition mapping and timing scheduling unit, based on the optimized multi-dimensional control parameter set for EMI suppression output by the strategy optimization unit under multi-objective constraints, maps the optimal control parameters to the corresponding operating condition partitions, establishing a one-to-one mapping table between operating conditions and control parameters. Simultaneously, combining the calibrated suppression target parameter set with device characteristic compensation reference and transient operating condition pre-control rules, it formulates timing scheduling rules for multiple control actions. Following the order of "source control first, path suppression second, and end-point filtering final," it sets the execution timing and synchronization triggering rules for each control parameter, avoiding the simultaneous execution of multiple control actions that generate new beat frequency interference and resonant noise. Finally, it generates timing control commands that can be directly issued and executed, outputting a set of EMI control strategy timing commands executable for each operating condition.

[0107] Please see Figure 4 The multi-link coordinated EMI targeted suppression execution module includes:

[0108] The adaptive control unit for the switching transistor drive waveform, based on the adaptive EMI suppression strategy and the multi-objective optimization module outputting the execution-mode EMI control strategy timing instruction set, adjusts the driving waveform parameters of the switching transistor in real time through an adjustable gate drive circuit. These parameters include the turn-on slope, turn-off slope, and dead time, precisely controlling the peak values ​​of dv / dt and di / dt of the switching node. This reduces the intensity of the core interference source from the noise source. Simultaneously, based on the optimized parameters, it minimizes the increase in switching losses while suppressing noise, ensuring power conversion efficiency. The unit outputs the real-time control execution result of the switching transistor drive waveform and the feedback value of the drive parameters.

[0109] The adaptive spread spectrum modulation and frequency jitter execution unit, based on the real-time control execution result of the switching transistor drive waveform and the drive parameter feedback value output by the adaptive control unit of the switching transistor drive waveform, performs adaptive spread spectrum modulation and frequency jitter control on the PWM drive signal of the switching transistor. According to the optimized modulation depth, frequency range, and modulation period, the noise energy concentrated at a fixed frequency is dispersed to a wide frequency range, which significantly reduces the noise peak value. At the same time, it keeps synchronized with the drive waveform control action to avoid beat frequency interference caused by frequency jitter and drive slope adjustment, and outputs the spread spectrum modulation execution result of the PWM signal and the modulation parameter feedback value.

[0110] The active common-mode / differential-mode noise adaptive cancellation unit, based on the spread spectrum modulation execution result and modulation parameter feedback value of the PWM signal output by the adaptive spread spectrum modulation and frequency jitter execution unit, collects the common-mode and differential-mode noise signals of the current input and output circuits in real time through the active noise cancellation circuit, and generates a compensation signal with the same amplitude but opposite phase as the original noise according to the optimized compensation phase and compensation amplitude parameters. The compensation signal is injected into the primary and secondary circuits to cancel the conducted noise in real time, suppress the noise propagation path, and output the active noise cancellation execution result and compensation parameter feedback value.

[0111] The adjustable passive filter network adaptive matching unit, based on the active noise cancellation execution result and compensation parameter feedback value output by the common-mode / differential-mode noise adaptive cancellation unit, switches the inductance and capacitance parameters of the adjustable passive filter network in real time through a semiconductor switch array. This matches the current operating conditions and residual noise characteristics, complementing the active noise cancellation in the preceding stage, and providing final suppression of residual conducted and radiated noise. Simultaneously, it avoids the additional losses of the fixed filter network under light load conditions, improving overall efficiency, and outputs the passive filter network parameter matching execution result and feedback signal indicating completion of the full-link suppression action.

[0112] Please see Figure 5 The fully closed-loop EMI effect verification and strategy iteration optimization module includes:

[0113] The real-time EMI suppression effect verification unit, based on the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action output by the multi-link collaborative EMI targeted suppression execution module, re-acquires noise signals from the adapter input side, output side, and space through a multi-channel synchronous ADC sampling circuit, as well as comprehensive performance parameters such as power conversion efficiency, output voltage regulation accuracy, device junction temperature, and dynamic response time. It compares the acquired noise signals with the original noise signals, calculates the noise suppression amount and standard limit margin for each frequency band, and verifies whether the comprehensive performance parameters meet the constraints. It outputs the EMI suppression effect verification dataset and the power supply comprehensive performance parameter set.

[0114] The control strategy effectiveness evaluation and deviation identification unit compares the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the EMI suppression effect real-time verification unit with the set post-calibration suppression target and constraints to evaluate whether the execution effect of the current control strategy meets the standards. If there are cases of noise exceeding the standard or performance parameters exceeding the constraint range, the unit identifies the frequency band exceeding the standard, the magnitude and direction of the deviation parameters, locates the cause of the deviation (sudden change in operating conditions, parameter execution deviation, change in interference source, device characteristic drift), and outputs a strategy effectiveness evaluation report, deviation parameter set, and deviation cause location results.

[0115] The strategy iteration and correction unit, based on the strategy effect evaluation report, deviation parameter set, and deviation cause location results output by the control strategy effect evaluation and deviation identification unit, performs corresponding iterative optimization operations: if the current strategy effect meets the target, the current operating condition, control parameters, suppression effect, and device characteristic parameters are stored in the full life cycle sample database as training samples for model optimization; if the effect does not meet the target, the control parameter set is iteratively corrected according to the deviation parameters, and the optimized control strategy is regenerated; the iteratively corrected control parameter set is output.

[0116] The fault warning and safety protection execution unit, based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the EMI suppression effect real-time verification unit, the strategy effect evaluation report, deviation parameter set and deviation cause location results output by the control strategy effect evaluation and deviation identification unit, monitors the abnormal state of the system in real time, including continuous noise exceeding the standard, device over-temperature, output overvoltage / overcurrent, input undervoltage / overvoltage, and device short circuit / open circuit faults. When an abnormal state is detected, it executes the corresponding safety protection actions according to the abnormality level, including power reduction operation, locking the PWM drive signal, and disconnecting the input relay. At the same time, it generates corresponding fault warning codes and abnormal logs to achieve full life cycle safety protection of the system.

[0117] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A low electromagnetic interference power adapter control method, characterized by, The specific steps are as follows: S1, Multi-dimensional electromagnetic interference perception and feature extraction under all working conditions: Synchronously collects multi-source signals from the entire working link of the power adapter, completes signal preprocessing, EMI time-frequency domain feature extraction, interference source localization and working condition matching and classification, and outputs interference source localization results, noise contribution ranking and current working condition label set; S2, Adaptive EMI Suppression Strategy Generation and Multi-Objective Optimization: Based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction steps under all operating conditions, the suppression priority ranking, control parameter decoupling, multi-objective optimization and control strategy timing scheduling are completed, and the set of EMI control strategy timing instructions executable under different operating conditions is output. S3, multi-link coordinated EMI targeted suppression execution: S31, Adaptive control of switching transistor drive waveform: Based on the adaptive EMI suppression strategy generation and multi-objective optimization steps, the timing instruction set of the EMI control strategy is executed under different working conditions. Through the adjustable gate drive circuit, the drive waveform parameters of the switching transistor are adjusted in real time. At the same time, based on the optimized parameters, the increase in switching loss is minimized while suppressing noise, and the real-time control execution result of the switching transistor drive waveform and the drive parameter feedback value are output. S32, Adaptive Spread Spectrum Modulation and Frequency Jitter Execution: Based on the real-time control execution result of the switching transistor drive waveform and the drive parameter feedback value output by the adaptive control step of the switching transistor drive waveform, adaptive spread spectrum modulation and frequency jitter control are performed on the PWM drive signal of the switching transistor. According to the optimized modulation depth, frequency range, and modulation period, the noise energy concentrated at a fixed frequency is dispersed to a wide frequency range. At the same time, it is synchronized with the drive waveform control action and outputs the spread spectrum modulation execution result of the PWM signal and the modulation parameter feedback value. S33, Active Common-Mode / Differential-Mode Noise Adaptive Cancellation: Based on the spread spectrum modulation execution result and modulation parameter feedback value of the PWM signal output from the adaptive spread spectrum modulation and frequency jitter execution steps, the active noise cancellation circuit collects the common-mode and differential-mode noise signals of the current input and output circuits in real time, and generates a compensation signal with the same amplitude and opposite phase as the original noise signal according to the optimized compensation phase and compensation amplitude parameters. This compensation signal is injected into the primary and secondary circuits to cancel the conducted noise in real time, and outputs the active noise cancellation execution result and compensation parameter feedback value. S34, Adjustable passive filter network adaptive matching: Based on the active noise cancellation execution result and compensation parameter feedback value output by the common-mode / differential-mode noise adaptive cancellation step, the inductor and capacitor parameters of the adjustable passive filter network are switched in real time through a semiconductor switch array to match the current operating conditions and residual noise characteristics, and output the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action. S4, Full Closed-Loop EMI Effect Verification and Strategy Iterative Optimization: Based on the passive filter network parameter matching execution results and the full-link suppression action completion feedback signal output by the multi-link collaborative EMI targeted suppression execution steps, the system completes EMI suppression effect verification, strategy deviation identification, control parameter iterative correction, and full life cycle model update, and outputs the iteratively corrected control parameter set.

2. The low electromagnetic interference power adapter control method of claim 1, wherein, The specific steps of S1 are as follows: S11, Multi-source signal synchronous acquisition: Through a multi-channel synchronous ADC sampling circuit, multi-dimensional raw signals of the entire working link of the power adapter are synchronously acquired. The sampling clock is synchronized with the PWM drive signal of the adapter's main switch transistor, and the synchronously acquired multi-source raw time domain signal set is output. S12, Original signal preprocessing and noise reduction: Based on the synchronous acquisition of multi-source raw time-domain signals output from the multi-source signal synchronous acquisition step, perform signal preprocessing and output a preprocessed high signal-to-noise ratio multi-source standard signal set. S13, Joint extraction of electromagnetic interference features in the time and frequency domains: Based on the preprocessed high signal-to-noise ratio multi-source standard signal set output from the original signal preprocessing and noise reduction steps, joint time and frequency domain feature extraction is performed, and the time and frequency domain features of the signal are extracted by wavelet packet decomposition to locate the time domain occurrence time of noise; at the same time, the peak values ​​of dv / dt and di / dt of the switching node, the proportion of common-mode / differential-mode noise components, and operating condition characteristic parameters are extracted, and the EMI time and frequency domain characteristic parameter set and the operating condition characteristic parameter set are output; S14, Interference Source Localization and Operating Condition Matching Classification: Based on the EMI time-frequency domain feature parameter set and operating condition feature parameter set output by the joint extraction steps of electromagnetic interference features in the time and frequency domains, a pre-trained lightweight gradient boosting decision tree classification model is used to accurately locate the noise interference source; at the same time, based on the operating condition feature parameters, the matching classification of the current operating condition is completed, and the interference source localization result, noise contribution ranking and current operating condition label set are output.

3. The low electromagnetic interference power adapter control method according to claim 1, characterized in that, The specific steps of S2 are as follows: S21, Interference Source Target Suppression Priority Ranking: Based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction steps under full operating conditions, combined with the noise limit requirements of the corresponding electromagnetic compatibility standards, the suppression priority of each interference source is ranked, and the suppression target value and limit margin requirements of each interference source are determined to form a priority ranking table, and the interference source suppression priority ranking table and suppression target parameter set are output. S22, Device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation: Based on the interference source suppression priority ranking table and suppression target parameter set output by the interference source targeted suppression priority ranking step, combined with the iteratively corrected control parameter set output by the full closed-loop EMI effect verification and strategy iterative optimization step, device characteristic full life cycle adaptive compensation and transient condition prediction and strategy pre-generation processing are performed, and the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules is output; S23, Decoupling of multiple control parameters and generation of initial strategy: Based on the device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation steps, the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules is decoupled from the multi-dimensional control parameters affecting EMI; at the same time, based on priority sorting, the initial suppression control parameter set corresponding to each interference source is generated, and the initial EMI suppression multi-dimensional control parameter set is output. S24, Strategy optimization under multi-objective constraints: Based on the initial EMI suppression multi-dimensional control parameter set output by the multi-control parameter decoupling and initial strategy generation steps, a multi-objective optimization function and constraint conditions are constructed. An adaptive weighted particle swarm optimization algorithm is used to iteratively optimize the initial control parameter set and output the optimized EMI suppression optimal multi-dimensional control parameter set. S25, Control Strategy Partition Mapping and Timing Scheduling: Based on the optimized multi-dimensional control parameter set for EMI suppression output by the strategy optimization steps under multi-objective constraints, the optimal control parameters are mapped to the corresponding operating condition partitions, establishing a one-to-one mapping table between operating conditions and control parameters; simultaneously, combined with the calibrated suppression target parameter set with device characteristic compensation benchmark and transient operating condition pre-control rules, timing scheduling rules for multiple control actions are formulated, the execution timing and synchronization triggering rules for each control parameter are set, and finally, timing control instructions that can be directly issued and executed are generated, outputting a set of EMI control strategy timing instructions executable for each operating condition.

4. The low electromagnetic interference power adapter control method according to claim 1, characterized in that, The specific steps of S4 are as follows: S41, Real-time Verification of EMI Suppression Effect: Based on the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action output by the multi-link collaborative EMI targeted suppression execution steps, the comprehensive performance parameters are re-acquired through a multi-channel synchronous ADC sampling circuit, and the acquired noise signal is compared with the original noise signal to calculate the noise suppression amount and standard limit margin of each frequency band. At the same time, it is verified whether the comprehensive performance parameters meet the constraints, and the EMI suppression effect verification dataset and the power supply comprehensive performance parameter set are output. S42, Control Strategy Effectiveness Evaluation and Deviation Identification: Based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output from the real-time EMI suppression effect verification step, compare them with the set post-calibration suppression target and constraints to evaluate whether the current control strategy's execution effect meets the standards; if there are cases of noise exceeding the standard or performance parameters exceeding the constraint range, identify the exceeding frequency band, the magnitude and direction of the deviation parameters, locate the cause of the deviation, and output a strategy effectiveness evaluation report, deviation parameter set, and deviation cause location results; S43, Strategy Iteration Correction: Based on the strategy effect evaluation report, deviation parameter set and deviation cause location results output by the control strategy effect evaluation and deviation identification steps, execute the corresponding iterative optimization operation and output the iteratively corrected control parameter set; S44, Fault Warning and Safety Protection Execution: Based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the real-time EMI suppression effect verification step, the strategy effect evaluation report, deviation parameter set and deviation cause location results output by the control strategy effect evaluation and deviation identification step, the abnormal state of the system is monitored in real time, and when an abnormal state is detected, the corresponding safety protection action is executed according to the abnormality level.

5. A low electromagnetic interference power adapter control system, characterized in that, include: The full-condition, multi-dimensional electromagnetic interference sensing and feature extraction module synchronously collects multi-source signals from the entire working link of the power adapter, completes signal preprocessing, EMI time-frequency domain feature extraction, interference source localization and working condition matching and classification, and outputs interference source localization results, noise contribution ranking and current working condition label set; The adaptive EMI suppression strategy generation and multi-objective optimization module, based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction module under all operating conditions, completes suppression priority ranking, control parameter decoupling, multi-objective optimization and control strategy timing scheduling, and outputs a set of EMI control strategy timing instructions that can be executed under different operating conditions. The multi-link collaborative EMI targeted suppression execution module includes a switch drive waveform adaptive adjustment unit, an adaptive spread spectrum modulation and frequency jitter execution unit, an active common-mode / differential-mode noise adaptive cancellation unit, and an adjustable passive filter network adaptive matching unit. The adaptive control unit for the switching transistor drive waveform, based on the timing instruction set of the EMI control strategy that can be executed under different operating conditions and generated by the adaptive EMI suppression strategy and the multi-objective optimization module, adjusts the driving waveform parameters of the switching transistor in real time through the adjustable gate drive circuit; at the same time, based on the optimized parameters, it minimizes the increase in switching losses while suppressing noise, and outputs the real-time control execution result of the switching transistor drive waveform and the feedback value of the drive parameters. The adaptive spread spectrum modulation and frequency jitter execution unit, based on the real-time control execution result of the switching transistor driving waveform and the driving parameter feedback value output by the adaptive control unit of the switching transistor driving waveform, performs adaptive spread spectrum modulation and frequency jitter control on the PWM driving signal of the switching transistor. According to the optimized modulation depth, frequency range, and modulation period, the noise energy concentrated at a fixed frequency is dispersed to a wide frequency range. At the same time, it keeps synchronized with the driving waveform control action and outputs the spread spectrum modulation execution result of the PWM signal and the modulation parameter feedback value. The active common-mode / differential-mode noise adaptive cancellation unit, based on the spread spectrum modulation execution result and modulation parameter feedback value of the PWM signal output by the adaptive spread spectrum modulation and frequency jitter execution unit, collects the common-mode and differential-mode noise signals of the current input and output circuits in real time through the active noise cancellation circuit, and generates a compensation signal with the same amplitude and opposite phase as the original noise according to the optimized compensation phase and compensation amplitude parameters. The compensation signal is injected into the primary and secondary circuits to cancel the conducted noise in real time, and outputs the active noise cancellation execution result and compensation parameter feedback value. The adjustable passive filter network adaptive matching unit, based on the active noise cancellation execution result and compensation parameter feedback value output by the common-mode / differential-mode noise adaptive cancellation unit, switches the inductance and capacitance parameters of the adjustable passive filter network in real time through a semiconductor switch array to match the current operating conditions and residual noise characteristics, and outputs the passive filter network parameter matching execution result and the feedback signal of the completion of the full-link suppression action. The fully closed-loop EMI effect verification and strategy iteration optimization module, based on the passive filter network parameter matching execution results and the full-link suppression action completion feedback signal output by the multi-link collaborative EMI targeted suppression execution module, completes EMI suppression effect verification, strategy deviation identification, control parameter iterative correction and full life cycle model update, and outputs the iteratively corrected control parameter set.

6. A low electromagnetic interference power adapter control system according to claim 5, characterized in that, The full-condition multi-dimensional electromagnetic interference sensing and feature extraction module includes: The multi-source signal synchronous acquisition unit uses a multi-channel synchronous ADC sampling circuit to synchronously acquire multi-dimensional raw signals from the entire working link of the power adapter. The sampling clock is synchronized with the PWM drive signal of the adapter's main switch transistor, and the output is a set of synchronously acquired multi-source raw time-domain signals. The raw signal preprocessing and noise reduction unit performs signal preprocessing based on the synchronously acquired multi-source raw time-domain signal set output by the multi-source signal synchronous acquisition unit, and outputs a preprocessed high signal-to-noise ratio multi-source standard signal set. The electromagnetic interference feature extraction unit, based on the preprocessed high signal-to-noise ratio multi-source standard signal set output by the original signal preprocessing and noise reduction unit, performs joint time-domain and frequency-domain feature extraction. It also extracts the time-frequency domain features of the signal through wavelet packet decomposition to locate the time-domain occurrence time of the noise. Simultaneously, it extracts the peak values ​​of dv / dt and di / dt of the switching node, the proportion of common-mode / differential-mode noise components, and operating condition characteristic parameters, and outputs the EMI time-frequency domain feature parameter set and the operating condition feature parameter set. The interference source localization and operating condition matching classification unit, based on the EMI time-frequency domain feature parameter set and operating condition feature parameter set output by the joint extraction unit of electromagnetic interference features in the time and frequency domains, accurately locates noise interference sources through a pre-trained lightweight gradient boosting decision tree classification model; at the same time, based on the operating condition feature parameters, it completes the matching and classification of the current operating condition, and outputs the interference source localization results, noise contribution ranking, and current operating condition label set.

7. A low electromagnetic interference power adapter control system according to claim 5, characterized in that, The adaptive EMI suppression strategy generation and multi-objective optimization module includes: The interference source targeted suppression priority ranking unit, based on the interference source location results, noise contribution ranking and current operating condition label set output by the multi-dimensional electromagnetic interference sensing and feature extraction module under all operating conditions, and combined with the noise limit requirements of the corresponding electromagnetic compatibility standard, ranks the suppression priority of each interference source, determines the suppression target value and limit margin requirement of each interference source, forms a priority ranking table, and outputs the interference source suppression priority ranking table and suppression target parameter set. The device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation unit, based on the interference source suppression priority ranking table and suppression target parameter set output by the interference source targeted suppression priority ranking unit, combined with the iteratively corrected control parameter set output by the full closed-loop EMI effect verification and strategy iteration optimization module, performs device characteristic full life cycle adaptive compensation and transient condition prediction and strategy pre-generation processing, and outputs a calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules; The multi-control parameter decoupling and initial strategy generation unit, based on the device characteristic full life cycle adaptive compensation and transient condition strategy pre-generation unit outputting the calibrated suppression target parameter set with device characteristic compensation benchmark and transient condition pre-control rules, decouples the multi-dimensional control parameters affecting EMI; at the same time, based on priority sorting, it generates the initial suppression control parameter set corresponding to each interference source and outputs the initial EMI suppression multi-dimensional control parameter set. The strategy optimization unit under multi-objective constraints constructs a multi-objective optimization function and constraint conditions based on the initial EMI suppression multi-dimensional control parameter set output by the multi-control parameter decoupling and initial strategy generation unit. It then uses an adaptive weighted particle swarm optimization algorithm to iteratively optimize the initial control parameter set and outputs the optimized EMI suppression optimal multi-dimensional control parameter set. The control strategy partition mapping and timing scheduling unit, based on the optimized multi-dimensional control parameter set for EMI suppression output by the strategy optimization unit under multi-objective constraints, maps the optimal control parameters to the corresponding operating condition partitions, establishing a one-to-one mapping table between operating conditions and control parameters. Simultaneously, combining the calibrated suppression target parameter set with device characteristic compensation reference and transient operating condition pre-control rules, it formulates timing scheduling rules for multiple control actions, sets the execution timing and synchronization triggering rules for each control parameter, and finally generates timing control instructions that can be directly issued and executed, outputting a set of EMI control strategy timing instructions executable for each operating condition.

8. A low electromagnetic interference power adapter control system according to claim 5, characterized in that, The fully closed-loop EMI performance verification and strategy iteration optimization module includes: The real-time EMI suppression effect verification unit, based on the passive filter network parameter matching execution result and the full-link suppression action completion feedback signal output by the multi-link collaborative EMI targeted suppression execution module, re-acquires the comprehensive performance parameters through a multi-channel synchronous ADC sampling circuit, compares the acquired noise signal with the original noise signal, calculates the noise suppression amount and standard limit margin of each frequency band, and verifies whether the comprehensive performance parameters meet the constraints, outputting the EMI suppression effect verification dataset and the power supply comprehensive performance parameter set; The control strategy effectiveness evaluation and deviation identification unit compares the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the EMI suppression effect real-time verification unit with the set post-calibration suppression target and constraints to evaluate whether the current control strategy's execution effect meets the standards. If there are cases of noise exceeding the standard or performance parameters exceeding the constraint range, the unit identifies the frequency band exceeding the standard, the magnitude and direction of the deviation parameters, locates the cause of the deviation, and outputs a strategy effectiveness evaluation report, deviation parameter set, and deviation cause location results. The strategy iteration and correction unit, based on the strategy effect evaluation report, deviation parameter set, and deviation cause location results output by the control strategy effect evaluation and deviation identification unit, performs corresponding iterative optimization operations and outputs the iteratively corrected control parameter set. The fault warning and safety protection execution unit monitors the abnormal state of the system in real time based on the EMI suppression effect verification dataset and power supply comprehensive performance parameter set output by the EMI suppression effect real-time verification unit, the strategy effect evaluation report, deviation parameter set and deviation cause location results output by the control strategy effect evaluation and deviation identification unit, and executes the corresponding safety protection action according to the abnormality level when an abnormal state is detected.

Images