Online Measurement Method and System for Resonant Circuit Quality Factor

Abstract

This invention relates to the field of electronic measurement technology and discloses an online measurement method and system for the quality factor of resonant circuits. The method involves injecting a small-amplitude sinusoidal disturbance signal into the excitation end of the resonant circuit under test using weak electromagnetic coupling, extracting the amplitude and phase of the disturbance response using lock-in amplification technology, performing Lorentz fitting on the frequency response curve to calculate the negative 3 dB bandwidth, outputting the quality factor in real time from the ratio of the resonant frequency to the bandwidth, introducing cross-correlation narrowband filtering to compress the equivalent noise bandwidth to below 1 Hz, and having an automatic resonant frequency tracking function.

Description

Technical Field

[0001] This invention relates to the field of electronic measurement technology, specifically to an online method and system for measuring the quality factor of resonant circuits for wireless power transmission systems and radio frequency circuit debugging applications. Background Technology

[0002] The quality factor is a core parameter characterizing the frequency selectivity and energy storage efficiency of a resonant circuit. It is defined as the ratio of the resonant frequency to the negative 3 dB bandwidth, directly reflecting the ratio of stored to dissipated energy within a single oscillation cycle. In wireless power transmission systems, the quality factor of the resonant circuits at both the transmitting and receiving ends directly determines the energy transmission efficiency and transmission distance. In RF communication front-ends, the quality factor of filters and oscillators affects frequency band selectivity and phase noise performance. In precision sensor applications, the quality factor of resonant sensors is closely related to measurement sensitivity and resolution. As wireless charging technology evolves towards high-power and dynamic alignment scenarios, the real-time fluctuation of the quality factor caused by factors such as temperature changes, load switching, and positional offset during resonant coil operation is becoming increasingly prominent, making online real-time monitoring of the quality factor increasingly urgent.

[0003] Existing quality factor (QF) measurement methods mainly fall into two categories: offline measurement using impedance analyzers and frequency scanning network analysis. Offline impedance analyzer measurement, represented by precision instruments such as the Agilent E4990A, works by completely disconnecting the resonant circuit under test from the system and connecting it to the instrument port. Impedance spectra are obtained through precise excitation and response detection, and the QF is then calculated. While this method offers high accuracy, it requires interrupting system operation for each measurement, making it unsuitable for normal circuit operation. Furthermore, the instruments are bulky and expensive, making them unsuitable for online production line testing and embedded monitoring scenarios. Frequency scanning network analysis utilizes a vector network analyzer to perform S-parameter scanning at the port under test, then indirectly calculates the QF using the 3dB bandwidth method or circle fitting. Although it offers some online capability in laboratory environments, the coaxial connectors or probes at the test port introduce additional contact resistance and parasitic capacitance, altering the inherent QF of the resonant circuit and causing a systematic deviation between the measured and true values. Additionally, the scanning speed of the network analyzer is limited by the intermediate frequency bandwidth setting, with typical scan rates only a few tens of times per second, making it difficult to track rapid dynamic changes in the QF at the millisecond level.

[0004] Chinese patent application CN119438707A discloses an impedance measurement method based on a KAN neural network and a continuous disturbance injection device. This method injects a sinusoidal disturbance signal into the grid connection point of the system under test using a continuous disturbance injection device, and utilizes the nonlinear approximation capability of the KAN neural network to perform spectrum identification to obtain the impedance amplitude and phase at different frequencies. Finally, the impedance measurement results are visualized. However, this scheme has the following shortcomings: First, the disturbance signal is directly connected in series to the grid connection point of the system via the disturbance injection device rather than being superimposed through weak coupling. The device connection itself changes the impedance characteristics of the system under test, significantly affecting the normal operation of the system. Second, the training and optimization process of the KAN neural network requires a large number of iterative calculations and the convergence time is uncertain, making it difficult to achieve millisecond-level real-time output updates. Third, this scheme uses impedance amplitude and phase as the final output, does not involve the direct calculation of the quality factor, and does not provide an automatic tracking mechanism for resonant frequency drift.

[0005] Therefore, there is an urgent need in the existing technology for an online quality factor measurement scheme that can be implemented under normal operating conditions of the resonant circuit, has a high measurement refresh rate, strong anti-interference ability, and automatic resonant frequency tracking function, in order to meet the urgent needs of practical application scenarios such as dynamic optimization of wireless power transmission systems, online debugging of radio frequency circuits, and status monitoring of industrial resonant equipment. Summary of the Invention

[0006] In view of the technical problems in the prior art, such as the need to disconnect the circuit and perform offline measurement of the quality factor of the resonant circuit, low measurement refresh rate, and inability to adaptively track when the resonant frequency drifts, the purpose of this invention is to provide an online quality factor measurement scheme that can be implemented under the normal working condition of the resonant circuit, with high refresh rate and strong anti-interference capability.

[0007] To achieve the above objectives, this invention proposes an online method for measuring the quality factor of a resonant circuit, comprising the following steps:

[0008] The process includes several steps: First, a small-amplitude sinusoidal disturbance signal is injected into the excitation terminal of the resonant circuit under test via electromagnetic coupling. The amplitude of the disturbance signal is controlled below a preset proportion of the main signal amplitude, and the frequency of the disturbance signal is scanned point-by-point within a preset frequency range near the resonant frequency at preset steps. Second, a lock-in amplifier response extraction step uses the disturbance signal as a reference signal and employs phase-sensitive detection and low-pass filtering to extract the amplitude and phase components of the resonant current response signal that are at the same frequency as the disturbance signal. Third, a frequency response curve fitting and bandwidth calculation step arranges the response amplitude values ​​at each disturbance frequency point in frequency order to construct a frequency response curve. The process involves: Lorentz function fitting to determine the resonant peak frequency and peak amplitude, and calculating the negative 3 dB bandwidth; real-time quality factor calculation and resonant frequency tracking steps, where the quality factor is calculated in real time based on the ratio of the resonant peak frequency to the negative 3 dB bandwidth, and the scanning center frequency is updated when the resonant peak frequency drift exceeds a preset drift threshold; and correlation detection, narrowband filtering, and multi-interface output steps, where narrowband filtering based on cross-correlation is applied to the resonant current response signal to compress the equivalent noise bandwidth to below a preset noise bandwidth threshold, and the time constant of the lock-in amplifier and the filter order are adaptively adjusted, and the quality factor measurement results are output through analog and digital communication interfaces respectively.

[0009] This invention also proposes an online measurement system for the quality factor of resonant circuits, including a disturbance signal coupling injection module, a lock-in amplifier response extraction module, a frequency response fitting and bandwidth calculation module, a quality factor calculation and resonant frequency tracking module, and a correlation detection narrowband filtering and multi-interface output module. Each module corresponds one-to-one with the steps of the above method. Specifically, the disturbance signal coupling injection module includes a direct digital synthesizer, a power adjustment unit, and a coupling loop; the lock-in amplifier response extraction module includes a non-contact current sensor, a preamplifier, and a digital lock-in amplifier; the quality factor calculation and resonant frequency tracking module feeds back an updated scanning center frequency to the disturbance signal coupling injection module when a resonant frequency drift is detected, forming a closed-loop tracking path; the correlation detection narrowband filtering and multi-interface output module adaptively adjusts the parameters of the lock-in amplifier response extraction module according to the noise environment, while providing both analog and digital communication dual-interface outputs. These modules constitute a deeply coupled and collaborative architecture combining a forward signal processing chain and a feedback control loop.

[0010] Compared with existing technologies, this invention has the following advantages: By injecting a micro-amplitude perturbation signal at the excitation end of the resonant circuit in a weak coupling manner, the perturbation amplitude is controlled to be less than one percent of the main signal, ensuring that the measurement process does not affect the normal operation of the resonant circuit, thus achieving true online measurement; by using lock-in amplification technology combined with correlation detection narrowband filtering, the equivalent noise bandwidth is compressed to below 1Hz, enabling the weak perturbation response signal to be reliably extracted from the noise of a strong electromagnetic interference environment; the quality factor refresh rate can reach 1000 times per second, which can capture rapid dynamic changes caused by load switching and position offset; it has an automatic resonant frequency tracking function, which automatically adjusts the scanning center frequency to maintain measurement validity when the resonant point drifts due to temperature changes or component aging; the quality factor measurement range covers 10 to 10000 with a relative error controlled within 2%, meeting a wide range of measurement needs from low-Q RF filters to high-Q superconducting cavities. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating the online measurement method for the quality factor of a resonant circuit provided in an embodiment of the present invention.

[0012] Figure 2 This is a schematic diagram of the architecture of the online measurement system for the quality factor of a resonant circuit provided in an embodiment of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be noted that the following embodiments are only for explaining the invention and are not intended to limit the scope of protection of the invention.

[0014] See Figure 1 This invention provides an online method for measuring the quality factor of resonant circuits. This method addresses the need for real-time monitoring of the quality factor of resonant circuits in wireless power transmission systems and RF circuit debugging. It involves injecting a small-amplitude disturbance signal into the excitation end of the resonant circuit under test using weak electromagnetic coupling. This is combined with lock-in amplification technology and related narrowband filtering technology to extract the weak disturbance response signal, and then calculating the quality factor by fitting a frequency response curve. This method features no circuit disconnection required, high refresh rate, and strong anti-interference capability, meeting the practical needs of continuous quality factor monitoring under dynamic operating conditions. The method includes the following five core steps.

[0015] Step S1: Coupling and Injecting a Micro-Amplitude Sinusoidal Perturbation Signal. The core purpose of this step is to superimpose a frequency-controllable sinusoidal perturbation signal onto the resonant circuit under test in a manner that has minimal impact on its operating state. This allows subsequent steps to extract quality factor information by analyzing the resonant response excited by the perturbation signal.

[0016] In one specific embodiment, the resonant circuit under test is the transmitting resonant coil of a wireless power transmission system operating at 85kHz. This resonant coil consists of a Litz wire wound coil with an inductance of 47μH and a thin-film resonant capacitor with a capacitance of 75nF connected in series. During normal operation, the peak current of the main signal is approximately 15A. The injection of the disturbance signal is achieved using a coupling loop. Specifically, a single-turn coupling loop is placed approximately 5mm from the surface of the resonant coil. The diameter of this coupling loop is approximately 1 / 10 of the diameter of the resonant coil, ensuring that the coupling coefficient is controlled within a weak coupling state between 0.001 and 0.01. Preferably, in this embodiment, the coupling coefficient is approximately 0.005. At this value, the equivalent current amplitude induced in the resonant circuit by the injected disturbance signal is approximately 0.5% of the main signal amplitude, far below the upper limit requirement of 1%, thus ensuring that the disturbance injection does not significantly change the inherent quality factor and operating point of the resonant circuit.

[0017] The perturbation signal is generated by a high-precision direct digital synthesizer with a frequency resolution of 0.01 Hz and a frequency stability better than ±0.1 ppm. In one embodiment of the invention, the scanning range of the perturbation frequency is set to ±10% of the resonant frequency, i.e., for a resonant frequency of 85 kHz, the scanning range is 76.5 kHz to 93.5 kHz, with a total scanning bandwidth of 17 kHz. The frequency step is set to 0.2% of the scanning range, i.e., approximately 34 Hz, meaning that a complete scan contains approximately 500 frequency points. Preferably, the scanning method adopts a linearly increasing mode from low frequency to high frequency, and the dwell time of each frequency point is set to a duration of not less than 5 times the time constant of the lock-in amplifier to ensure that the output signal of the lock-in amplifier reaches a steady state. In different application scenarios, the preset frequency range can be adjusted according to the estimated value of the quality factor of the resonant circuit under test: For narrowband resonant circuits with high quality factors (e.g., Q greater than 1000), the half-power bandwidth of the resonant peak is narrow, and the scanning range is set to ±5% of the resonant frequency to fully cover the resonant peak and ensure sufficient frequency resolution. For example, for a resonant frequency of 85kHz, the scanning range is 80.75kHz to 89.25kHz, and the total scanning bandwidth is 8.5kHz; For broadband resonant circuits with low quality factors (e.g., Q less than 50), the half-power bandwidth of the resonant peak is wide, and the scanning range needs to be expanded to ±15% of the resonant frequency to fully cover the entire effective response region of the resonant peak. For example, for a resonant frequency of 85kHz, the scanning range is 72.25kHz to 97.75kHz, and the total scanning bandwidth is 25.5kHz. Accordingly, the frequency step size is selected between 0.1% and 0.5% of the scanning range, depending on the scanning accuracy requirements: a fine step size of 0.1% is used in high-precision measurement scenarios to obtain denser sampling points and thus improve the Lorentz fitting accuracy. For example, when the scanning range is 8.5kHz, the step size is 8.5Hz, and a complete scan contains approximately 1000 frequency points; a coarse step size of 0.5% is used in fast measurement scenarios to shorten the scanning time. For example, when the scanning range is 25.5kHz, the step size is 127.5Hz, and a complete scan contains approximately 200 frequency points.

[0018] In this embodiment, the output power of the disturbance signal generator is controlled between -20dBm and 0dBm. The specific power level is automatically adjusted according to the main signal power of the resonant circuit under test to meet the constraint that the disturbance amplitude does not exceed 1% of the main signal. Preferably, when the system starts up, it first performs a rapid sampling evaluation of the main signal amplitude of the resonant circuit under test, and sets the upper limit of the transmission power of the disturbance signal based on the evaluation result. During the disturbance signal injection process, the actual injected disturbance current amplitude is monitored in real time through the current sampling resistor set in the coupling loop. If the disturbance current amplitude is detected to exceed the preset upper limit, the disturbance signal power is automatically attenuated. This closed-loop amplitude control mechanism ensures that the disturbance effect can be controlled within an acceptable range under different operating conditions.

[0019] It is worth noting that the weak electromagnetic coupling injection method used in this invention differs fundamentally from the method in CN119438707A, which directly connects the disturbance injection device at the grid connection point. The weak coupling method does not require any additional components in series or parallel in the signal path of the resonant circuit, thus avoiding the introduction of additional contact resistance and parasitic parameters, fundamentally preventing interference from the measurement probe connection on the measured quality factor. Furthermore, the weak coupling injection method offers the advantage of electrical isolation; the disturbance signal generator is connected to the resonant circuit under test via magnetic field coupling rather than wires, ensuring measurement safety even in high-voltage resonant circuits.

[0020] Preferably, in this embodiment, the coupling loop is a single-turn copper loop with a diameter of approximately 25 mm and a conductor cross-sectional area of ​​1 mm². 2 The loop plane is placed parallel to the surface of the resonant coil. The coupling loop is connected to the disturbance signal generator via a 50Ω coaxial cable, the length of which should not exceed 1m to reduce signal transmission loss. A 47Ω matching resistor is connected in series at the connection point between the coupling loop and the coaxial cable to achieve source impedance matching and avoid the influence of reflected waves on the waveform quality of the disturbance signal. Experimental verification shows that, under the above configuration, the total harmonic distortion of the disturbance signal injected into the resonant circuit is better than -60dB, ensuring that the spectral purity of the disturbance signal meets the reference signal quality requirements of the subsequent lock-in amplifier.

[0021] Step S2: Lock-in Amplification Response Extraction. The core function of this step is to accurately extract the amplitude and phase information at the same frequency as the disturbance signal from the weak disturbance response signal that is submerged by noise in the resonant circuit under test.

[0022] The lock-in amplifier (LPA) is the core signal processing device in this step, and its operating principle is based on the orthogonality of sinusoidal functions. Specifically, the LPA performs phase-sensitive detection on the input resonant current response signal and the reference signal, that is, multiplies the response signal by the in-phase and quadrature components of the reference signal, respectively. Then, a low-pass filter is used to eliminate higher harmonic components, thereby obtaining in-phase output X and quadrature output Y of the response components with the same frequency as the reference signal. The amplitude of the response signal is determined by… The phase of the response signal relative to the reference signal is calculated to be determined by... Obtained through calculation.

[0023] In one embodiment of the present invention, the resonant current response signal is acquired using a non-contact current sensor, specifically a Rogowski coil or a Hall effect current sensor, which is installed near the current path of the resonant circuit without disconnecting the circuit. Preferably, this embodiment uses a Rogowski coil with a sensitivity of 100mV / A and a bandwidth covering 10kHz to 10MHz, meeting the measurement requirements for the 85kHz resonant frequency and its vicinity. The voltage signal output by the Rogowski coil is amplified by a preamplifier and then input to the signal input terminal of a lock-in amplifier.

[0024] The reference signal of the lock-in amplifier is directly taken from the synchronous output port of the disturbance signal generator in step S1, ensuring a strict frequency and phase lock-in relationship between the reference signal and the actual injected disturbance signal. In this embodiment, the initial time constant of the lock-in amplifier is set to 1ms, corresponding to an equivalent noise bandwidth of approximately 125Hz, and the filter order is set to 2nd order, i.e., using an attenuation slope of 12dB / oct. This setting provides sufficient noise suppression capability while ensuring a quality factor refresh rate of 1000 times / s.

[0025] In the actual measurement process, the measurement procedure at each scanned frequency point is as follows: The disturbance frequency controller simultaneously sends the current frequency value to the reference input terminals of the disturbance signal generator and the lock-in amplifier. The disturbance signal generator outputs a sinusoidal disturbance signal at the current frequency and injects it into the resonant circuit through a coupling loop. The lock-in amplifier waits for a settling time of 5 times the time constant before sampling the X and Y outputs of the current frequency point, and stores the calculated amplitude value R and phase value θ in the frequency response data buffer. After all frequency points have been scanned, the frequency response data buffer contains a complete set of frequency-amplitude-phase data triplets for subsequent curve fitting steps.

[0026] Preferably, to improve the measurement accuracy at a single frequency point, multiple samples can be taken at each frequency point and the average value can be calculated. In this embodiment, each frequency point is sampled four times consecutively and the arithmetic average is taken, which effectively reduces the random noise of a single sample by about 6dB. This step also includes an overload detection mechanism: when the input signal amplitude of the lock-in amplifier exceeds 80% of the full scale, the gain of the preamplifier is automatically reduced to avoid signal clipping distortion.

[0027] Furthermore, this invention optimizes the digital implementation of the lock-in amplifier. In traditional analog lock-in amplifiers, the reference signal channel of the phase-sensitive detector suffers from inherent errors such as phase drift and amplitude imbalance. However, the digital lock-in amplifier used in this embodiment digitizes the input signal at a sampling rate of at least 1 MSPS using a high-speed analog-to-digital converter, and then implements numerical multiplication and digital low-pass filtering in the FPGA, thereby completely eliminating the temperature drift and aging problems of analog devices. The digital low-pass filter adopts a fourth-order IIR Butterworth structure, and its cutoff frequency is determined by the time constant parameter, specifically, the cutoff frequency equals the reciprocal of the time constant divided by 2π. With a time constant of 1 ms, the cutoff frequency of the low-pass filter is approximately 159 Hz, providing approximately 64 dB of suppression depth for the 170 kHz second harmonic component with a resonant frequency of 85 kHz.

[0028] In one embodiment of the present invention, the lock-in amplifier is further equipped with a harmonic suppression function for the reference signal. When there are high-order harmonic components of the main signal in the resonant circuit under test, and their frequencies are close to certain scanning frequency points, these harmonic components may leak into the output of the lock-in amplifier, leading to measurement errors. To address this, the system introduces a pure sine reference generation technique in the reference signal channel. A pure sine reference signal with a total harmonic distortion better than -80dB is generated using a lookup table method, ensuring that the phase-sensitive detection process extracts only the fundamental component, thus effectively suppressing harmonic leakage.

[0029] Step S3: Frequency response curve fitting and negative 3 dB bandwidth calculation. This step uses the frequency response data obtained in step S2 to accurately determine the characteristic parameters of the resonant peak through a curve fitting algorithm, thereby calculating the negative 3 dB bandwidth required for the quality factor.

[0030] The frequency response curve of the resonant circuit near the resonant frequency exhibits a typical Lorentz shape, and its amplitude response can be expressed as:

[0031] ,

[0032] in: For frequency The response amplitude at that point is expressed in V. The maximum response amplitude at the resonance peak is expressed in V. The load quality factor is dimensionless. The resonant peak frequency is expressed in Hz. This formula reflects the balance between energy storage and dissipation at different frequencies in a resonant circuit. The higher the quality factor, the sharper the resonant peak and the better the frequency selectivity.

[0033] This invention employs a weighted least squares algorithm to fit the Lorentz function to the frequency response data. Preferably, the weight allocation strategy is as follows: data points closer to the resonance peak are assigned higher weights, and data points farther from the resonance peak are assigned lower weights. Specifically, the weight function is defined as: ,

[0034] in: For the first The fitting weights for each frequency point are dimensionless and range from 0 to 1. For the first The measured response amplitude at each frequency point is expressed in V. In this embodiment, the total number of scanning frequency points is... This weighting method makes the fitting algorithm pay more attention to data points with high signal-to-noise ratio near the resonance peak, reducing the adverse impact of low signal-to-noise ratio data points far from the resonance peak on the fitting accuracy.

[0035] The fitting process employs the Levenberg-Marquardt iterative optimization algorithm, with... , and These are three parameters to be optimized. The method for determining the initial values ​​of the iteration is as follows: The initial value is taken as the frequency point with the largest response amplitude. The initial value is taken as the maximum amplitude value. The initial value is taken as an empirical estimate of 100. The convergence criterion for the iteration is that the relative changes of the three parameters in two consecutive iterations are all less than 100. In this embodiment, under the condition of 500 scanning frequency points, the Levenberg-Marquardt algorithm typically converges within 10 to 20 iterations, with a computation time of approximately 0.2 ms.

[0036] After the fitting is completed, based on the fitted result and Calculate the negative 3 dB bandwidth. For a Lorentz-type frequency response curve, the negative 3 dB bandwidth... It can be calculated using the following formula:

[0037] ,

[0038] in: The bandwidth is negative 3 dB, measured in Hz. The resonant peak frequency is expressed in Hz. Let $\mathbf{v}$ be the load quality factor, which is dimensionless. This formula shows the inverse relationship between the quality factor and bandwidth: the higher the quality factor, the narrower the bandwidth.

[0039] Preferably, this step also includes a fitting quality evaluation mechanism. The root mean square value of the fitting residuals... Defined as:

[0040] ,

[0041] in: For the first The measured response amplitude at each frequency point is expressed in V. For the fitted function at the th The predicted values ​​at each frequency point are expressed in V. When the root mean square value of the fitted residual exceeds the peak amplitude... At 5%, that is If the system determines that the current fitting accuracy is insufficient, it automatically halves the scanning frequency step size and increases the frequency sampling density near the resonance peak before re-executing the scanning and fitting process until the fitting residual meets the requirements. This adaptive accuracy control strategy ensures reliable fitting results across different quality factor ranges.

[0042] It is important to note that for broadband resonant circuits with low quality factors (e.g., Q < 30), the large negative 3 dB bandwidth means that the scanning frequency window may contain many data points far from the resonance peak. In this case, the Lorentz fitting converges faster, but the fitting accuracy is more sensitive to noise. To address this, this invention adds a data pre-screening step before fitting: first, the frequency point corresponding to the maximum response amplitude is roughly identified; then, only data points centered at this frequency point and within a range three times the estimated bandwidth are retained for fitting, while low signal-to-noise ratio data points far from the resonance peak are discarded. This strategy improves the fitting accuracy of low-Q circuits by approximately 30% without increasing computational complexity.

[0043] Conversely, for narrowband resonant circuits with extremely high quality factors (e.g., Q>5000), the negative 3 dB bandwidth is only on the order of several hertz, and the spacing between scanning frequency points may be larger than the bandwidth itself. In this case, the system automatically switches to a fine scanning mode: first, a coarse scan is performed with a large step (e.g., 100 Hz) to locate the approximate position of the resonant peak; then, a fine scan is performed within the range of ±2 times the estimated bandwidth near the resonant peak with a step of 1 / 10 of the bandwidth. This two-stage scanning strategy takes into account both the need for wide-range search and fine bandwidth measurement, ensuring full-range measurement coverage from low Q to high Q.

[0044] Step S4: Real-time calculation of quality factor and automatic tracking of resonant frequency.

[0045] This step calculates the quality factor in real time based on the fitting parameters obtained in step S3, and automatically adjusts subsequent scanning parameters when the resonant frequency drifts to maintain the validity and accuracy of the measurement.

[0046] The formula for calculating the quality factor is:

[0047] ,

[0048] in: The quality factor is dimensionless, and the measurement range of this system is 10 to 10000. The resonant peak frequency is expressed in Hz. The bandwidth is negative 3 dB, in Hz. Equivalently, it is directly given by the Lorentz fit in step S3. The parameter is the desired quality factor value. In one embodiment of the present invention, the quality factor calculation result is smoothed by a digital moving average filter before being output, and the moving window length is set to 10 measurement cycles to further suppress quality factor fluctuations caused by measurement noise.

[0049] Automatic resonant frequency tracking is one of the key features of this invention. In the actual operation of wireless power transmission systems, the resonant frequency may drift due to the following factors: temperature changes cause changes in the permeability of the core material, leading to changes in inductance; load switching causes changes in the reflected impedance, leading to changes in the equivalent capacitance; and the relative position offset of the transmitting and receiving coils causes changes in mutual inductance, leading to a shift in the system's resonant point. If the above drift exceeds the scanning frequency window, it will cause the quality factor measurement to fail.

[0050] The resonant frequency tracking mechanism of this invention employs the following strategy: after each quality factor measurement, the resonant peak frequency obtained in the current measurement cycle is... Resonant peak frequency with the previous measurement cycle Compare and calculate the frequency drift:

[0051] ,

[0052] in: The resonant frequency drift between two adjacent measurement periods, in Hz; superscript Indicates the current measurement period, superscript Indicates the previous measurement period. When Exceeding the preset drift threshold When the system detects a significant shift in the resonant frequency, it updates the scanning center frequency to the currently measured resonant peak frequency. .

[0053] Preferably, the preset drift threshold is set to a value between 10% and 30% of the current negative 3 dB bandwidth. The specific value is determined based on the stability of the operating environment: when the resonant circuit operates in a stable environment with slow temperature changes and infrequent load switching, the preset drift threshold can be set to 30% of the current negative 3 dB bandwidth to reduce unnecessary tracking actions and thus reduce system computational overhead; when the resonant circuit operates in a drastic environment with frequent load switching or rapid temperature fluctuations, the preset drift threshold is tightened to 10% of the current negative 3 dB bandwidth to improve tracking sensitivity and ensure that the resonant peak is always within the scanning window. In the typical operating conditions of this embodiment, the preset drift threshold is set to 20% of the current negative 3 dB bandwidth, i.e. The rationale for this setting is that when the frequency drift is less than 20% of the bandwidth, the resonance peak is still completely contained within the scanning window, and the fitting accuracy of the quality factor will not decrease significantly; when the drift exceeds this threshold, the resonance peak may partially move out of the scanning window, and the scanning center frequency must be updated to ensure that the resonance peak is always located in the center region of the scanning window.

[0054] Furthermore, to address extreme scenarios involving rapid and significant drift in the resonant frequency, this invention employs a two-stage tracking strategy. The first stage is conventional tracking: when... At this stage, only the scanning center frequency is updated while the scanning range remains unchanged. The second stage is fast tracking: when Simultaneously, while updating the scanning center frequency, the scanning range is temporarily expanded to 1.5 times the original and the frequency step size is increased to 2 times the original to quickly capture the position of the resonance peak after a large drift. Once the resonance peak stabilizes again, the original scanning parameters are automatically restored. In this embodiment, the switching criterion for the two-level tracking strategy also considers the consistency of continuous drift directions: when the frequency drift direction is the same for three consecutive measurement cycles, the system predicts that the resonance frequency is in a monotonic change process and will pre-offset the scanning window by half a bandwidth in the current drift direction to adapt to the upcoming further drift. This predictive tracking mechanism enables the system to have better tracking performance in slow, monotonic drift scenarios such as gradual temperature changes.

[0055] In one embodiment of the present invention, to avoid false triggering during frequency tracking, the system also introduces a drift confirmation mechanism. Specifically, when the first detection... Instead of immediately updating the scan center frequency, the system verifies whether the drift direction and drift amount remain consistent over the next two consecutive measurement cycles. Only when three consecutive checks confirm that the drift exceeds the threshold is the scan parameters officially updated. This mechanism effectively avoids erroneous tracking actions caused by transient noise spikes or fitting outliers, improving the system's tracking robustness in noisy environments.

[0056] This step, together with step S1, forms a closed-loop feedback: the updated scan center frequency output in step S4 directly affects the frequency scan start parameter of the disturbance frequency controller in step S1, causing the scan window of the next measurement cycle to automatically align with the new resonant peak position. This closed-loop mechanism ensures that the system can still maintain reliable quality factor measurement even under dynamic operating conditions where the resonant frequency continues to drift.

[0057] Step S5: Correlation Detection Narrowband Filtering and Multi-Interface Output. This step achieves two main functions: first, it further improves the signal-to-noise ratio of the measurement system in strong electromagnetic interference environments through narrowband filtering technology based on cross-correlation calculations; second, it outputs the quality factor measurement results through multiple interface formats to adapt to different upper-level system requirements.

[0058] In the actual operating environment of wireless power transmission systems, strong electromagnetic interference exists near the resonant circuit. The amplitude of the main signal may be 40dB to 60dB higher than the disturbance response signal. In addition, noise sources such as high-frequency harmonics of the switching transistor, power supply ripple, and environmental radio frequency interference often drown out the weak disturbance response signal. Although the lock-in amplifier in step S2 provides a certain degree of noise suppression capability, it may still be insufficient under extreme interference conditions.

[0059] This invention introduces a narrowband filtering technique based on cross-correlation as a supplement to the lock-in amplifier. Its basic principle is: to filter the resonant current response signal... Local copy of the disturbance signal Perform cross-correlation calculations:

[0060] ,

[0061] in: It is a cross-correlation function. The time delay variable is expressed in seconds (s). The integration time is expressed in seconds (s). In this embodiment... Take 1 second; This is the resonant current response signal, measured in amperes (A). The signal is a local copy of the perturbation signal, a dimensionless normalized signal. Since the cross-correlation operation only retains the frequency components common to the two signals, and the noise is uncorrelated with the perturbation signal, the noise is greatly suppressed after the cross-correlation operation.

[0062] The cross-power spectrum is obtained by performing a Fourier transform on the cross-correlation function. :

[0063] ,

[0064] in: Cross-power spectrum, for The conjugate of the spectrum, for The spectrum. The perturbation frequency is extracted from the cross-power spectrum. Spectral peak amplitude at and phase This refers to the amplitude and phase of the disturbance response signal after narrowband filtering. When the integration time... At time s, the equivalent noise bandwidth is approximately Hz, compared to the time constant of a lock-in amplifier The equivalent noise bandwidth at ms is reduced by approximately 21 dB at 125 Hz.

[0065] Preferably, the present invention also incorporates an adaptive adjustment mechanism for the lock-in amplifier parameters. The system estimates the noise power spectral density of the current measurement environment in real time. During the interval between each scan, the perturbation signal injection is paused and a pure noise signal is acquired. The power spectrum of this noise signal is then estimated to obtain... Based on the target signal-to-noise ratio (Set to 20dB in this embodiment) and the current noise power spectral density Calculate the required equivalent noise bandwidth. :

[0066] ,

[0067] in: Equivalent noise bandwidth, in Hz; The power of the disturbance response signal, in units of / Hz; Noise power spectral density, in units of / Hz; The target signal-to-noise ratio is expressed in dB. This is based on the calculated... Determine the minimum time constant of the lock-in amplifier. Under the constraint of ensuring a quality factor refresh rate of not less than 100 times / s, a time constant of [value missing] is selected. to Within the range of values, the filter order is set from 2nd to 4th order. The higher the order, the stronger the noise suppression but the slower the response speed.

[0068] Regarding multi-interface output, this invention provides both analog output and digital communication output interfaces. The analog interface outputs a 0-10V voltage signal or a 4-20mA current signal, with a linear mapping relationship between the quality factor value and the output voltage or current. Specifically, when the quality factor measurement range is set to 10 to 10000, the quality factor value... With output voltage The mapping relationship between them is as follows:

[0069] ,

[0070] in: This is an analog output voltage, measured in volts (V), ranging from 0 to 10V. This is a dimensionless measurement of the quality factor. This is the lower limit of the measurement range; This is the upper limit of the measurement range; V represents the full-scale range of the output voltage; V represents the zero-point offset of the output voltage. The update rate of this analog output interface is consistent with the quality factor refresh rate, reaching 1000 times / s.

[0071] The digital communication interface supports three protocols: UART, SPI, and I2C, which can be switched via a protocol selection jumper or software configuration. The data packets output by the digital interface include the following fields: quality factor value (32-bit floating-point number), resonant frequency value (32-bit floating-point number, unit Hz), negative 3 dB bandwidth value (32-bit floating-point number, unit Hz), and measurement timestamp (32-bit unsigned integer, unit ms). Preferably, the UART interface baud rate is set to 115200 bps, the single data packet length is 20 bytes, and the maximum data output rate is 5760 packets / s, far exceeding the quality factor refresh rate requirement. The SPI interface supports a clock frequency from 1MHz to 10MHz, adopts full-duplex communication mode, and the master device can read the latest measurement results or configure system parameters via SPI commands. The I2C interface device address can be configured via hardware pins, supporting both standard and fast communication rates.

[0072] In practical applications, the three digital communication interfaces mentioned above can be flexibly selected according to the interface type of the host system. Preferably, when the host system is an industrial PLC, the analog 4-20mA output interface is recommended because of its strong anti-interference capability and long transmission distance; when the host system is an embedded microcontroller, the SPI or I2C interface is recommended to achieve high-speed data transmission and parameter configuration; when the host system is PC-based test software, the UART interface is recommended, and virtual serial communication with the PC can be achieved through a USB-to-UART bridge chip.

[0073] In one embodiment of the present invention, the above-described method was subjected to system-level performance verification. The test environment was an 85kHz magnetically coupled resonant wireless charging system with a rated power of 3kW, and a Keysight E4990A standard impedance analyzer was used as the reference benchmark for quality factor measurement. The test consisted of two parts: static accuracy testing and dynamic tracking testing.

[0074] In the static accuracy test, nine groups of standard resonant circuits with quality factors of 10, 50, 100, 200, 500, 1000, 2000, 5000, and 8000 were measured, with each group measured 100 times consecutively. The test results show that the relative deviation between the quality factor measurement value of the method of this invention and the E4990A reference value does not exceed ±2.0% in all test groups, meeting the design requirements. In the extremely low Q test group with a quality factor of 10, due to the relatively flat resonance peak, the typical relative deviation is ±1.8% to ±2.0%. Reliable measurement results can still be obtained by increasing the number of scan frequency points and adopting a data pre-screening strategy. In the medium-low Q range with a quality factor of 50 to 500, the typical relative deviation is ±0.8% to ±1.2%, demonstrating excellent measurement repeatability. In the high Q range with a quality factor of 1000 to 5000, the typical relative deviation is ±1.0% to ±1.8%. In the extremely high Q test group with a quality factor of 8000, the system automatically switches to a two-stage fine scanning mode, with a typical relative deviation of ±1.5% to ±1.9%, verifying the system's measurement capability for extremely high quality factors. The standard deviation (repeatability index) of 100 repeated measurements is better than 0.5% in all test groups.

[0075] In the dynamic tracking test, the resonant frequency of the resonant circuit was changed at a rate of twice per second using an adjustable capacitor, causing the resonant frequency to periodically change between 80kHz and 90kHz. Simultaneously, a variable resistor was used to adjust the circuit loss, causing the quality factor to dynamically fluctuate between 100 and 800. Test results show that the method of this invention can continuously output quality factor measurements during the dynamic change of the resonant frequency, with a stable refresh rate of 1000 times / s, a resonant frequency tracking delay of no more than two measurement cycles (2ms), and a dynamic measurement error of no more than ±3%. In contrast, traditional impedance analyzers are completely unable to perform measurements under the above dynamic conditions, and vector network analyzers can only achieve a scanning speed of approximately 10 times / s with a measurement error as high as ±8%.

[0076] In the anti-interference performance test, the quality factor of this invention was measured under the full-power (3kW) operation of the wireless charging system. At this time, the electromagnetic field strength near the measurement probe was approximately 50 A / m, and the ambient noise signal was about 50 dB higher than the disturbance response signal. Without the cross-correlation narrowband filter function, the standard deviation of the quality factor measurement fluctuation was approximately 8%. After enabling the cross-correlation narrowband filter function (integration time set to 0.5 s), the standard deviation of the quality factor measurement fluctuation dropped to below 0.6%, with a corresponding quality factor refresh rate of 200 times / s, confirming the significant improvement effect of the cross-correlation narrowband filter technology in strong interference environments. Furthermore, with the integration time increased to 1 s, the equivalent noise bandwidth is compressed to 1 Hz, and the standard deviation of the quality factor fluctuation can be further reduced to 0.3%, but the refresh rate correspondingly drops to 100 times / s. Users can select the appropriate integration time parameter according to their actual needs for accuracy and speed.

[0077] In the resonant frequency tracking performance test, the receiving coil of the wireless charging system was driven by a stepper motor to move horizontally at a speed of 5 mm / s, causing a continuous change in the mutual inductance between the transmitting and receiving coils, which in turn led to resonant frequency drift. The test results showed that, under the condition that the displacement range of the receiving coil was ±30 mm, the maximum drift of the resonant frequency was approximately 2.5 kHz (approximately ±3% relative to the nominal resonant frequency of 85 kHz). The two-stage tracking strategy of this invention successfully tracked the entire drift process, and the measurement error of the quality factor during the tracking process did not exceed ±2.5%, with no tracking loss or mistracking occurring.

[0078] In summary, the method of this invention meets or exceeds the design requirements in key performance indicators such as static accuracy, dynamic tracking, anti-interference capability, and measurement range, verifying the feasibility and superiority of the online measurement scheme for the quality factor of resonant circuits based on micro-amplitude perturbation injection and phase-locked amplification technology. Compared with existing offline measurement methods based on impedance analyzers and online impedance analysis methods based on neural networks, the core advantages of this invention are: measurement can be performed without disconnecting the resonant circuit; the measurement refresh rate is more than two orders of magnitude higher than traditional schemes; the built-in automatic resonant frequency tracking mechanism enables the system to adapt to dynamic operating conditions such as temperature changes, load switching, and position offset; and the signal processing architecture based on phase-locked amplification and cross-correlation detection is far superior to deep learning network-based schemes in terms of algorithm complexity and real-time performance, making it more suitable for embedded system integration and mass production line deployment.

[0079] See Figure 2This invention also provides an online measurement system for the quality factor of resonant circuits. The overall design goal of this system is to achieve high-speed, real-time online measurement of the quality factor, while possessing low power consumption and small size embedded integration characteristics. The system includes five functional modules, corresponding one-to-one with the five steps in the above method embodiments. These modules are interconnected via a high-speed parallel bus and control signal lines.

[0080] The disturbance signal coupling injection module 1 comprises three sub-units: a direct digital synthesizer, a power adjustment unit, and a coupling loop. The direct digital synthesizer generates a high-precision sinusoidal disturbance signal with controllable frequency, achieving a frequency resolution of 0.01 Hz and a frequency stability better than ±0.1 ppm. The power adjustment unit receives scan center frequency update commands from the quality factor calculation and resonant frequency tracking module 4, as well as disturbance amplitude control commands from the correlation detection narrowband filtering and multi-interface output module 5, dynamically adjusting the output power of the disturbance signal. The coupling loop injects the disturbance signal into the excitation terminal of the resonant circuit under test using weak electromagnetic coupling, with the coupling coefficient controlled between 0.001 and 0.01.

[0081] The lock-in amplifier response extraction module 2 comprises three sub-units: a non-contact current sensor, a preamplifier, and a digital lock-in amplifier. The non-contact current sensor picks up the current signal in the resonant circuit and converts it into a voltage signal. The preamplifier amplifies and impedance-matches the voltage signal. The digital lock-in amplifier uses the disturbance signal as a reference signal to perform phase-sensitive detection and low-pass filtering operations, outputting the in-phase component X and quadrature component Y at each scan frequency point. The time constant and filter order of this module can be adaptively adjusted by the correlation detection narrowband filtering and multi-interface output module 5.

[0082] The frequency response fitting and bandwidth calculation module 3 receives amplitude and phase data for all frequency points output by the lock-in amplification response extraction module 2, executes a weighted least squares Lorentz function fitting algorithm, and outputs three key parameters: resonant peak frequency, peak amplitude, and negative 3 dB bandwidth. This module has a built-in fitting quality evaluation function; when the fitting residual exceeds the standard, it automatically sends a command to the perturbation signal coupling injection module 1 to increase the scan density. The core computation of this module is implemented by an embedded digital signal processor, and a single fitting operation takes approximately 0.2 ms.

[0083] The quality factor calculation and resonant frequency tracking module 4 receives the resonant peak frequency and negative 3 dB bandwidth parameters from the frequency response fitting and bandwidth calculation module 3, calculates the quality factor, and executes the resonant frequency drift detection and two-stage tracking strategy. This module outputs the quality factor measurement, resonant frequency measurement, and negative 3 dB bandwidth measurement. When the resonant frequency drift exceeds a preset threshold, it sends a scan center frequency update command to the disturbance signal coupling injection module 1, forming a closed-loop feedback path from the quality factor calculation and resonant frequency tracking module 4 to the disturbance signal coupling injection module 1.

[0084] The correlation detection narrowband filter and multi-interface output module 5 performs dual functions. Its narrowband filter subunit performs cross-correlation calculations and cross-power spectrum analysis on the resonant current response signal, compressing the equivalent noise bandwidth to below 1Hz. Based on the real-time noise power spectral density estimation results, it adaptively adjusts the time constant and filter order of the lock-in amplifier response extraction module 2, forming a parameter optimization path from the correlation detection narrowband filter and multi-interface output module 5 to the lock-in amplifier response extraction module 2. Its multi-interface output subunit provides analog interfaces (0-10V voltage or 4-20mA current) and digital communication interfaces (UART / SPI / I2C), outputting the quality factor measurement results in a format adapted to the requirements of the host system.

[0085] The data flow and control flow among the five modules form a deeply coupled collaborative architecture: the disturbance signal coupling injection module 1 outputs the disturbance signal → the lock-in amplification response extraction module 2 extracts the response → the frequency response fitting and bandwidth calculation module 3 fits the curve → the quality factor calculation and resonant frequency tracking module 4 calculates the quality factor → the quality factor calculation and resonant frequency tracking module 4 feeds back to update the scanning center frequency of the disturbance signal coupling injection module 1, forming a closed loop of forward signal processing and feedback tracking; simultaneously, the correlation detection narrowband filtering and multi-interface output module 5 adaptively adjusts the parameters of the lock-in amplification response extraction module 2 according to the noise environment and controls the disturbance amplitude of the disturbance signal coupling injection module 1, forming a horizontal parameter optimization path. This multi-dimensional coupling architecture enables the system to autonomously coordinate the working parameters of each module when facing complex dynamic conditions, achieving an optimal balance between measurement accuracy, refresh rate, and anti-interference capability.

[0086] In terms of hardware implementation, in one embodiment of the present invention, the entire measurement system is integrated on a four-layer printed circuit board measuring 120mm × 80mm. The direct digital synthesizer in the disturbance signal coupling injection module 1 uses the AD9834 chip, with a frequency resolution of 0.28Hz / bit and a 75MHz reference clock. The analog-to-digital converter in the lock-in amplifier response extraction module 2 uses the ADS8688 chip, with 16-bit resolution and a sampling rate of 500kSPS, meeting the oversampling requirements for signals near 85kHz. The computational functions of the frequency response fitting and bandwidth calculation module 3 and the quality factor calculation and resonant frequency tracking module 4 are implemented by an STM32H743 microprocessor with a main frequency of 480MHz and a built-in double-precision floating-point unit, capable of completing Lorentz fitting calculations for 500 data points within 0.2ms. The cross-correlation calculations in the correlation detection narrowband filtering and multi-interface output module 5 are accelerated by an FPGA coprocessor, and the analog output interface is implemented using a 16-bit digital-to-analog converter. The entire system is powered by 5V DC and consumes approximately 2.5W, making it suitable for embedded integrated applications.

[0087] Preferably, the system calibration process is as follows: Before initial use, the system is calibrated at two points using a standard resonant circuit with a known quality factor (the quality factor calibration standard is pre-calibrated by an impedance analyzer). The measurement gain and offset at the low-Q end (quality factor approximately 50) and the high-Q end (quality factor approximately 2000) are calibrated respectively. Calibration data is stored in on-chip non-volatile memory and automatically retrieved during subsequent measurements. The system also provides an automatic self-test function, automatically detecting the operating status of each module after power-on and outputting the self-test result code on the digital interface. When a module malfunctions or experiences performance degradation, the system outputs the corresponding fault code through the digital interface and a preset fault indication voltage (e.g., 0V or full-scale voltage) at the analog output port, allowing the host system to respond promptly.

[0088] Furthermore, the system embodiments of this invention also include a measurement mode selection function. The refresh rate of the quality factor can be flexibly configured within the range of 100 times / s to 1000 times / s according to measurement requirements. Users can configure three measurement modes via a digital interface: high-speed mode, with a quality factor refresh rate of 1000 times / s, suitable for scenarios requiring the capture of rapidly changing dynamics; in this mode, the cross-correlation narrowband filter function is turned off to ensure speed. Standard mode, with a quality factor refresh rate of 200 times / s, the cross-correlation narrowband filter function is enabled, and the integration time is set to 0.5s, balancing speed and accuracy. High-precision mode, with a quality factor refresh rate of 100 times / s, the cross-correlation narrowband filter function is enabled, and the integration time is set to 1s, suitable for measurement scenarios requiring a balance between accuracy and response speed under static or slowly changing conditions. Users can further increase the cross-correlation narrowband filter integration time to 2s to reduce the refresh rate to 50 times / s to obtain higher measurement accuracy. The three modes can be switched in real time during operation, and the system automatically adjusts the operating parameters of each module to adapt to the new mode requirements.

[0089] The embodiments of the present invention are not limited to the specific embodiments described above. Those skilled in the art can make various equivalent changes or substitutions based on the technical solutions of the present invention, and all such changes or substitutions should be included within the protection scope of the present invention.

Claims

1. An online method for measuring the quality factor of a resonant circuit, characterized in that, Includes the following steps: Micro-amplitude sinusoidal disturbance signal coupling injection step: A micro-amplitude sinusoidal disturbance signal is injected into the excitation end of the resonant circuit under test by electromagnetic coupling. The amplitude of the micro-amplitude sinusoidal disturbance signal is controlled below a preset proportion of the amplitude of the main signal. The frequency of the micro-amplitude sinusoidal disturbance signal is scanned point by point in a preset frequency range near the resonant frequency by a preset step amount. Lock-in amplifier response extraction steps: Using the micro-amplitude sinusoidal disturbance signal as a reference signal, a lock-in amplifier is used to extract the amplitude and phase components of the resonant current response signal that are at the same frequency as the micro-amplitude sinusoidal disturbance signal through phase-sensitive detection and low-pass filtering, so as to obtain the response amplitude value and response phase value at each disturbance frequency point. Frequency response curve fitting and bandwidth calculation steps: Construct a frequency response curve by taking the response amplitude values ​​obtained at each disturbance frequency point in frequency order, perform Lorentz function fitting on the frequency response curve to determine the resonant peak frequency and peak amplitude, and calculate the difference between the upper and lower sidebands corresponding to the point where the amplitude drops to -3 dB of the peak as the negative 3 dB bandwidth. Real-time calculation of quality factor and resonant frequency tracking steps: The quality factor is calculated in real time based on the ratio of the resonant peak frequency to the negative 3 dB bandwidth. When the drift of the resonant peak frequency exceeds the preset drift threshold, the drifted resonant peak frequency is fed back to the micro-amplitude sinusoidal disturbance signal coupling and injection step to update the scanning center frequency. Related detection narrowband filtering and multi-interface output steps: Apply narrowband filtering based on cross-correlation operation to the resonant current response signal to compress the equivalent noise bandwidth to below a preset threshold, adaptively adjust the time constant and filter order of the lock-in amplifier according to the noise power spectral density, and output the quality factor measurement result through analog interface and digital communication interface.

2. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The amplitude of the micro-amplitude sinusoidal disturbance signal is controlled to be less than 1% of the amplitude of the main signal, the preset frequency range is ±5% to ±15% of the resonant frequency, the preset step is 0.1% to 0.5% of the preset frequency range, and the frequency stability of the micro-amplitude sinusoidal disturbance signal is better than ±0.1ppm.

3. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The refresh rate of the quality factor is from 100 times / s to 1000 times / s, and the measurement range of the quality factor covers 10 to 10000.

4. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The preset drift threshold is 10% to 30% of the current negative 3 dB bandwidth. When the resonant peak frequency drift exceeds the preset drift threshold, the update amount of the scan center frequency is equal to the drift amount, and a complete frequency scan cycle is re-executed after the update.

5. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The Lorentz function fitting uses a weighted least squares algorithm, where frequency points closer to the resonant peak are assigned a greater fitting weight than those farther from the resonant peak. Furthermore, when the root mean square of the fitting residual exceeds a preset residual ratio of the peak amplitude, the number of scanning frequency points is automatically increased to improve the fitting accuracy.

6. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The narrowband filtering process based on cross-correlation includes: performing cross-correlation operation on the resonant current response signal and a local copy of the micro-amplitude sinusoidal disturbance signal; performing frequency domain transformation on the cross-correlation operation result to obtain the cross-power spectrum; extracting the spectral peak amplitude and phase at the disturbance frequency from the cross-power spectrum as the filtered response signal; and compressing the equivalent noise bandwidth to below 1 Hz.

7. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The adaptive adjustment of the time constant and filter order of the lock-in amplifier includes: real-time estimation of the noise power spectral density of the current measurement environment, calculation of the minimum required time constant based on the target signal-to-noise ratio requirement and the current noise power spectral density level, selection of the maximum available value of the time constant under the constraint that the quality factor refresh rate is not lower than the preset minimum refresh rate, and setting the filter order to 2nd to 4th order.

8. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The electromagnetic coupling method is implemented using a coupling loop or a coupling probe. The coupling coefficient of the coupling loop is controlled between 0.001 and 0.01 to ensure that the injected disturbance signal does not affect the inherent quality factor of the resonant circuit under test.

9. The method for online measurement of the quality factor of a resonant circuit according to claim 1, characterized in that, The analog interface outputs a 0-10V or 4-20mA linear mapping signal for the quality factor. The digital communication interface supports UART, SPI, or I2C protocols, and the output includes the quality factor value, resonant frequency value, negative 3 dB bandwidth value, and measurement timestamp.

10. An online quality factor measurement system for resonant circuits, used to implement the online quality factor measurement method for resonant circuits according to any one of claims 1-9, characterized in that, include: The disturbance signal coupling injection module is used to inject a micro-amplitude sinusoidal disturbance signal into the excitation end of the resonant circuit under test by electromagnetic coupling. The amplitude of the micro-amplitude sinusoidal disturbance signal is controlled below a preset proportion of the amplitude of the main signal, and the frequency of the micro-amplitude sinusoidal disturbance signal is scanned point by point within a preset frequency range near the resonant frequency. The phase-locked amplification response extraction module is used to extract the amplitude and phase components of the resonant current response signal that are at the same frequency as the disturbance signal by using the micro-amplitude sinusoidal disturbance signal as a reference signal and through phase-sensitive detection and low-pass filtering. The frequency response fitting and bandwidth calculation module is used to construct a frequency response curve based on the response amplitude value at each disturbance frequency point, perform Lorentz function fitting on the frequency response curve to determine the resonant peak frequency and peak amplitude, and calculate the negative 3 dB bandwidth. The quality factor calculation and resonant frequency tracking module is used to calculate the quality factor in real time based on the ratio of the resonant peak frequency to the negative 3 dB bandwidth, and to update the scanning center frequency of the micro-amplitude sinusoidal disturbance signal coupling injection module when the resonant peak frequency drift exceeds the preset drift threshold. The correlation detection narrowband filter and multi-interface output module is used to apply narrowband filtering based on cross-correlation operation to the resonant current response signal to compress the equivalent noise bandwidth, and adaptively adjust the time constant and filter order of the lock-in amplification response extraction module according to the noise power spectral density. At the same time, the quality factor measurement results are output through analog interface and digital communication interface respectively.

Images