A phase difference detection method based on double checking

Abstract

The application belongs to the technical field of power system synchronizing device, and specifically provides a phase difference detection method based on double check, which synchronously collects vibration signals of the environment where the synchronizing device is located through external sensors, actively compensates the vibration phase error of the hardware detection channel, performs time-frequency analysis and multi-bright-point echo feature extraction on the voltage signal, identifies and quantifies the interference phase features introduced by the non-ideal features of the hardware structure, and feeds back to the software detection module for adaptive correction, dynamically generates fusion weights according to real-time working condition parameters including frequency difference, harmonic distortion rate, signal-to-noise ratio and vibration amplitude, weights and fuses the two-way phase difference after compensation and correction, and evaluates the health state of the system according to the fusion residual error. The application solves the problems of lack of depth cooperation between hardware and software modules and poor environmental adaptability, and has the advantages of high precision, strong robustness and good adaptive ability.

Description

Technical Field

[0001] This invention relates to the field of power system synchronization device technology, specifically to a phase difference detection method based on dual verification. Background Technology

[0002] In power systems, the parallel operation of two independent power systems (such as a generator and a power grid) requires precise detection of the voltage difference, frequency difference, and phase difference between the system-side voltage and the target-side voltage using a synchronizing device. Among these, accurate detection of the phase difference is crucial to ensuring successful synchronous closing and preventing asynchronous closing accidents.

[0003] Existing phase difference detection methods are mainly divided into two categories: hardware detection methods and software detection methods. Hardware detection methods typically obtain the phase difference directly by comparing the zero-crossing points of the voltage sine wave. While simple in principle and fast in response, the detection accuracy depends entirely on the accuracy of a single zero-crossing point. In practical applications, external electromagnetic interference can easily cause jitter in the zero-crossing signal, high harmonic content in the power grid can cause the fundamental wave zero-crossing point to shift, and even hardware circuit malfunctions can directly lead to phase measurement errors, resulting in low reliability.

[0004] Software-based detection methods calculate the fundamental phase by performing a discrete Fourier transform on the voltage signal. This method utilizes data from the entire cycle, offering strong anti-interference capabilities, but its calculation accuracy is highly dependent on the stability of the signal frequency. When a frequency difference exists between the system side and the target side (…),… When the frequency difference increases, the traditional discrete Fourier transform method will produce significant calculation errors due to spectral leakage, and the error will increase with the frequency difference.

[0005] The system directly calculates the difference between hardware and software detection results and locks the circuit breaker when the difference exceeds the limit. This approach is merely a simple comparison and selection of results, representing an open-loop "double comparison" rather than a closed-loop "double verification." It cannot compensate for dynamically changing interference in real time, nor can it perform deep fusion and collaborative correction when a single detection method fails. This results in a lack of adaptability to environmental factors (such as vibration and noise) and the system's own condition (such as hardware aging). The long-term stability and robustness under complex operating conditions of the measurements still need improvement. Summary of the Invention

[0006] This invention proposes a phase difference detection method based on dual verification. Addressing the problems of existing phase difference detection methods for synchronization devices, such as the lack of deep collaboration and feedback mechanisms between hardware and software detection modules, leading to significant impacts on measurement accuracy from complex operating conditions and poor system adaptability, a closed-loop detection system integrating real-time vibration noise compensation, multi-brightness echo feature decoupling, and dynamic fusion decision-making is constructed. Vibration signals from the environment surrounding the synchronization device are synchronously acquired via external sensors, and an adaptive filter is used to actively compensate for vibration phase errors in the hardware detection channel. Time-frequency analysis and multi-brightness echo feature extraction are performed on the voltage signal to identify and quantify interference phase features introduced by non-ideal features such as edges and valves in the hardware structure of the synchronization device, and these features are fed back to the software detection module for adaptive correction. Based on real-time operating parameters including frequency difference, harmonic distortion rate, signal-to-noise ratio, and vibration amplitude, fusion weights are dynamically generated. The compensated and corrected hardware and software phase differences are weighted and fused to output the final phase difference. The system health status is assessed based on the fusion residual. This is achieved through "active compensation." Feature decoupling The "dynamic fusion" closed-loop collaborative mechanism organically combines environmental perception, structural feature extraction and dual-channel adaptive correction, which significantly improves the accuracy, robustness and adaptability of phase difference detection in complex environments, and provides a new technical approach for high-reliability grid connection and closing of synchronous devices.

[0007] The technical solution adopted by this invention is as follows: This solution provides a phase difference detection method based on dual verification, applied to a synchronization device. The method includes the following steps: Step A1: Collect the system-side voltage signal and the object-side voltage signal of the synchronizing device, and obtain the preprocessed system-side voltage signal and object-side voltage signal after preprocessing; set up an external sensor to synchronously collect the vibration signal of the environment where the synchronizing device is located. Step A2: Set up hardware detection channel and software detection channel, and input the preprocessed system-side voltage signal and object-side voltage signal in parallel into the hardware detection channel and software detection channel to generate hardware phase difference and initial software phase difference respectively; Step A3: Based on the vibration signal, construct a vibration compensation model, calculate the phase error compensation amount caused by vibration, and feed it back to the hardware detection channel in real time to dynamically compensate the hardware phase difference and generate the compensated hardware phase difference. Step A4: Extract multi-brightness echo features from the preprocessed system-side voltage signal and the object-side voltage signal to obtain interference phase features; identify and quantify the interference phase features, correct the initial software phase difference based on the interference phase features, and generate the corrected software phase difference; Step A5: Calculate the current operating parameters in real time, dynamically generate the first fusion weight and the second fusion weight based on the current operating parameters, and perform weighted fusion on the compensated hardware phase difference and the corrected software phase difference to generate the final phase difference.

[0008] Further, in step A1, the system-side voltage signal and the object-side voltage signal are preprocessed, specifically including: setting up a signal conditioning circuit, an analog-to-digital converter, and a zero-crossing comparator; the signal conditioning circuit filters, amplifies, and adjusts the amplitude of the system-side voltage signal and the object-side voltage signal to match the input range of the analog-to-digital converter and the zero-crossing comparator. External sensors are installed inside or near the synchronizing device chassis to collect vibration signals from the environment where the synchronizing device is located.

[0009] Furthermore, in step A2, the hardware detection channel employs a detection method based on a zero-crossing comparator. Specifically, the preprocessed system-side voltage signal and the target-side voltage signal are converted into square wave signals by a zero-crossing comparator. The hardware phase difference is calculated by measuring the time difference between the rising or falling edges of the two square wave signals and combining it with the system's rated frequency. The software detection channel employs a quasi-software detection method based on Discrete Fourier Transform. Specifically, the preprocessed system-side voltage signal and the target-side voltage signal are sampled synchronously using analog and digital methods. The real and imaginary parts of the fundamental phasors of the preprocessed system-side voltage signal and the target-side voltage signal are calculated using Discrete Fourier Transform, and the initial software phase difference is obtained through arctangent calculation.

[0010] Furthermore, step A3 involves constructing a vibration compensation model, specifically including the following steps: Step A31: Establish a transfer function model between the vibration signal and the zero-crossing phase jitter in the hardware detection channel. The transfer function model is a linear time-invariant system, and the system function is obtained through the system identification method. Step A32: Set up an adaptive filter; use the adaptive filter to process the vibration signal in real time, with the goal of minimizing the residual phase noise at the hardware phase difference output terminal, and adaptively adjust the filter coefficients to generate real-time compensation. Step A33: The real-time compensation amount is superimposed on the zero-crossing comparator input of the hardware detection channel in the form of negative feedback or directly subtracted from the hardware phase difference to realize closed-loop vibration compensation and generate the compensated hardware phase difference.

[0011] Furthermore, in step A4, multi-brightness echo feature extraction is performed on the preprocessed system-side voltage signal and the object-side voltage signal. Specific steps include: Step A41: Perform time-frequency analysis on the preprocessed system-side voltage signal and object-side voltage signal, separate the signal components arriving at different times in the time-frequency domain, and identify the echo components from multiple scattering bright spots in the hardware structure of the synchronous device. Step A42: Calculate the relative propagation path difference between echo components of different scattering bright spots based on the physical geometry of the synchronous device; Step A43: Based on the relative propagation path difference, quantify the interference pattern between the echoes of each scattering bright spot, and extract the dominant frequency and amplitude of the interference pattern as interference phase features.

[0012] Furthermore, the initial software phase difference is corrected based on the interference phase characteristics. Specifically, the product of the interference phase characteristics and the preset correction coefficient is subtracted from the initial software phase difference to obtain the corrected software phase difference.

[0013] Furthermore, the specific steps in step A5 for generating the final phase difference include: Step A51: Calculate the current system operating parameters in real time, including absolute value of frequency difference, total harmonic distortion rate, signal-to-noise ratio, and vibration amplitude; Step A52: Input the operating condition parameters into the pre-trained dynamic weight generation model, output the first fusion weight of the compensated hardware phase difference and the second fusion weight of the corrected software phase difference, and the sum of the first fusion weight and the second fusion weight is 1; Step A53: Based on the first fusion weight and the second fusion weight, perform weighted fusion on the compensated hardware phase difference and the corrected software phase difference to generate the final phase difference.

[0014] Furthermore, in the dynamic weight generation model, the first fusion weight is increased when the absolute value of the frequency difference or the total harmonic distortion rate increases; and the second fusion weight is increased when the vibration amplitude increases or the signal-to-noise ratio decreases.

[0015] Furthermore, following step A5, step A6 is included: calculating the fusion residual between the compensated hardware phase difference and the corrected software phase difference, and assessing the health status of the hardware detection channel or the software detection channel based on the time series change trend of the fusion residual. When the fusion residual continues to increase and exceeds a first preset threshold, it is determined that the corresponding channel has experienced performance degradation; when the fusion residual step exceeds a second preset threshold, it is determined that a hardware fault has occurred, an alarm signal is issued, and the synchronous closing command is blocked.

[0016] Compared with the prior art, the beneficial effects of the present invention are: (1) By setting external sensors to collect vibration signals of the environment where the synchronizing device is located in real time, and constructing a closed-loop vibration compensation model based on an adaptive filter, the real-time compensation amount is superimposed on the zero-crossing comparator input of the hardware detection channel in the form of negative feedback or directly subtracted from the hardware phase difference, thereby realizing the active suppression of zero-crossing phase jitter caused by vibration. This effectively solves the problem of the sharp drop in accuracy of traditional hardware detection methods under harsh environments such as electromagnetic interference and mechanical vibration, and significantly improves the robustness and stability of hardware phase difference detection, providing a basic guarantee for the reliable operation of the synchronizing device under complex field conditions; (2) By performing time-frequency analysis and multi-brightness echo feature extraction on the preprocessed system-side voltage signal and object-side voltage signal, interference phase features introduced by non-ideal features such as end faces, edges, and valves in the hardware structure of the synchronous device are identified and quantified, and fed back to the software detection module for adaptive correction. The structural interference information contained in the hardware detection channel is used inversely to the software detection algorithm, realizing deep information interaction and synergistic effect between the hardware and software detection modules, rather than simple independent work or result comparison, thus overcoming the inherent technical bias of those skilled in the art that dual detection modules can only perform simple comparisons; (3) By calculating operating parameters such as absolute value of frequency difference, total harmonic distortion rate, signal-to-noise ratio, and vibration amplitude in real time, and inputting them into a pre-trained dynamic weight generation model, the first fusion weight of the compensated hardware phase difference and the second fusion weight of the corrected software phase difference are adaptively generated, and the two phase differences are weighted and fused. This enables the system to always select or emphasize the more reliable detection information source under different environmental conditions, solving the problem that the existing dual comparison scheme cannot dynamically optimize the measurement strategy according to real-time operating conditions and cannot adaptively adjust when the frequency difference or harmonics change, significantly improving the accuracy and adaptability of phase difference detection under complex operating conditions; (4) The fusion residual between the compensated hardware phase difference and the corrected software phase difference is calculated, and the health status of the hardware detection channel or the software detection channel is evaluated based on the time series change trend of the residual. When the fusion residual continues to increase and exceeds the first preset threshold, performance degradation is determined. When the fusion residual step exceeds the second preset threshold, hardware fault is determined and an alarm and a blocking synchronous closing command are issued. By utilizing the collaborative feedback of two independent detection channels to form a closed-loop self-diagnostic capability, asynchronous closing accidents caused by performance degradation or hardware faults of a single detection channel are effectively avoided, and the safety of the synchronization device is greatly improved. Attached Figure Description

[0017] Figure 1 This is a schematic flowchart of a phase difference detection method based on dual verification according to the present invention; The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] Example: Please see Figure 1 This embodiment presents a phase difference detection method based on dual verification, applied to a synchronization device, for detecting system-side voltage. and object-side voltage The phase difference is detected with high precision and high robustness.

[0020] The specific steps include: Step A1, Signal Acquisition and Preprocessing: The system includes a signal conditioning circuit, an analog-to-digital converter, a zero-crossing comparator, an external sensor, and a voltage transformer. The external sensor is a three-axis MEMS accelerometer, installed inside the synchronous device chassis.

[0021] The system-side voltage signal of the synchronizing device is collected in real time using a voltage transformer. and object-side voltage signal The system-side voltage signal is conditioned by a signal conditioning circuit. and object-side voltage signal The signal is filtered, amplified, and its amplitude adjusted to match the input range of the analog-to-digital converter and the zero-crossing comparator, resulting in the preprocessed system-side voltage signal. and the preprocessed object-side voltage signal .

[0022] Vibration signals from the environment surrounding the synchronous device are collected synchronously using external sensors. The vibration sampling rate is 100 times the system's rated frequency. That is, for a system rated frequency of 50Hz, the vibration sampling rate is set to 5kHz to capture high-frequency vibration components.

[0023] Step A2, Dual-channel parallel detection: Set up hardware and software detection channels. The preprocessed system-side voltage signal... and the preprocessed object-side voltage signal The data is fed into both the hardware testing channel and the software testing channel in parallel.

[0024] Hardware detection channel: Employs a detection method based on zero-crossing comparators. Specifically, it uses the pre-processed system-side voltage signal... and the preprocessed object-side voltage signal The signals are converted into square wave signals by zero-crossing comparators. and Measure the time difference between the rising (or falling) edges of two square wave signals. Combined with the system's rated frequency (50Hz), the hardware phase difference was calculated. : ; Software detection channel: Employs a quasi-software phase difference detection method. This is applied to the pre-processed system-side voltage signal. and the preprocessed object-side voltage signal Perform analog-to-digital conversion at sampling frequency ( Synchronous sampling is performed. Data is collected within one cycle (10 cycles, 200 samples per cycle, totaling 2000 samples). Point data. The preprocessed system-side voltage signal is calculated using Discrete Fourier Transform. and the preprocessed object-side voltage signal The fundamental phasor.

[0025] For the system-side voltage, the discrete sampling sequence is: ( The real part of the fundamental phasor and the virtual part The calculation is as follows: ; ; For the object-side voltage, the discrete sampling sequence is: ( The real part of the fundamental phasor and the virtual part The calculation is as follows: ; ; The initial software phase difference for: ; Let the system-side voltage frequency be... The object-side voltage frequency is When there is a frequency difference between the two At this time, the calculation results contain approximate errors, which need to be corrected in subsequent steps.

[0026] Step A3: Hardware phase difference closed-loop compensation based on vibration signal: To improve the accuracy of hardware testing, this step introduces vibration signals for closed-loop feedback compensation.

[0027] Phase error caused by vibration is defined as hardware phase difference measurement error. This error is caused by zero-crossing phase jitter in the hardware detection channel. (Establish vibration signal) Error in hardware phase difference measurement The transfer function model between them. The transfer function model approximates a linear time-invariant system, and the system function is... The system identification method is used to obtain the model, which includes: applying a preset sweep frequency vibration excitation to the synchronizing device, acquiring the phase difference output of the hardware detection channel, and fitting the input and output data using a least squares or subspace identification algorithm to obtain a transfer function model describing the vibration signal to phase jitter. The vibration acceleration is characterized, the velocity is obtained by integrating the vibration acceleration, the velocity causes displacement, the displacement causes changes in optical path difference, and thus causes phase jitter.

[0028] To achieve real-time compensation, an adaptive filter is constructed; in this embodiment, a minimum mean square filter is used. The input to the filter is the vibration signal. The output is the real-time compensation amount. With the objective of minimizing the residual phase noise at the hardware phase difference output, the coefficient update formula for the minimum mean square filter is: ; in, These are the filter coefficients. Step size factor The error signal after compensation. These are discrete sampled values ​​of the vibration signal. Error signal. It is obtained from the high-frequency components of the hardware phase difference output.

[0029] The generated real-time compensation amount In the form of negative feedback, directly with the hardware phase difference obtained in step A2 Subtract to obtain the compensated hardware phase difference : ; Step A4: Adaptive software phase difference correction based on multi-brightness echo characteristic decoupling: The metal structures, connectors, and relays inside the synchronous device will generate a "multiple bright spots" echo effect similar to that in sonar under high-frequency transient signal excitation, which will distort the waveform of the voltage signal and thus affect the accuracy of software detection based on discrete Fourier transform.

[0030] This step first processes the preprocessed system-side voltage signal. With object-side voltage signal Wavelet transform was performed to separate the signal components arriving at different times in the time-frequency domain. These components correspond to reflections or scattering from different parts of the hardware structure of the synchronous device. Through analysis, bright spots on the end face, edges, and valves (relays) were identified.

[0031] Based on the physical geometry of the synchronous device, estimate the relative propagation path difference of echoes from different scattering bright spots. The relative propagation path difference between end-face bright spots and edge bright spots. Approximately: ; in, The characteristic length of the device; The characteristic radius; The signal incidence equivalent angle characterizes the equivalent geometric relationship of the propagation path of the voltage signal between different structures (such as end faces and edges) within the synchronous device. The specific values ​​are determined by applying standard test signals to devices of the same model and time in advance and analyzing their frequency response characteristics.

[0032] Based on relative propagation path difference The interference pattern between echoes from different scattering points was calculated. This interference pattern, in the frequency domain, exhibits periodic enhancement or attenuation at specific frequency points. The dominant frequency and amplitude of this interference pattern were extracted as interference phase characteristics. .

[0033] Utilizing the extracted interference phase features For the initial software phase difference Perform correction to obtain the corrected software phase difference. The correction model is: ; in, The preset correction coefficients characterize the degree of influence of the interference phase on the final phase difference. Feedback and coordination between the hardware signal (containing structural interference information) and the software algorithm (used to correct its own errors) are achieved.

[0034] Step A5: Dynamic Cooperative Evaluation and Phase Difference Fusion The compensated hardware phase difference and the corrected software phase difference are dynamically fused.

[0035] Real-time calculation of current system operating parameters, including absolute frequency difference, total harmonic distortion, signal-to-noise ratio, and vibration amplitude: Absolute value of frequency difference: ,in For system-side voltage frequency, The voltage frequency on the object side.

[0036] Total harmonic distortion: ,in The highest harmonic order, This is the effective value of the fundamental voltage. For the first Effective value of subharmonic voltage.

[0037] Signal-to-noise ratio: ,in For signal power, This represents noise power.

[0038] Vibration amplitude: ,in This represents the root mean square operation.

[0039] These parameters are input into a pre-trained dynamic weight generation model. The dynamic weight generation model can be a simple fuzzy logic controller or a shallow neural network trained offline. In this embodiment, a shallow neural network is used as the dynamic weight generation model, and the pre-training process includes: Constructing a training sample set: During the offline debugging phase of the device, multiple sets of historical data under different operating conditions were collected. Each set of data includes the absolute value of the frequency difference. Total harmonic distortion Vibration amplitude Signal-to-noise ratio And use it as input features; for each input group, test separately. Calculate the final phase difference after weighted fusion of hardware and software phase differences, and calculate the fluctuation (e.g., standard deviation) of the final phase difference across multiple measurements. Select the value that minimizes the fluctuation. As labels for this group of samples At the same time, .

[0040] Constructing a shallow neural network: The input layer has 4 nodes, corresponding to the normalized operating parameters; the hidden layer has 3 nodes with ReLU activation; the output layer has 1 node with Sigmoid activation, and the output value... This is the first fusion weight. Define network parameters: including the weight matrix from the input layer to the hidden layers. Bias vector and the weight vector from the hidden layer to the output layer. Bias scalar .

[0041] Training the network: Normalize the input features and input them into the network, then perform forward computation to obtain... The mean squared error loss function is used. ,in For batch size, The sample labels are used. Gradient descent is used to iteratively update the network parameters. Continue training until the loss converges. After training is complete, save the final network parameters.

[0042] In actual operation, the normalized operating parameters calculated in real time are input into the neural network, and the first fusion weight is obtained through forward calculation. and order The dynamic weight generation model outputs the first fused weight after compensating for the hardware phase difference. The second fusion weight of the software phase difference after correction And satisfy .

[0043] During the pre-training process of the dynamic weight generation model, the determination of training sample labels follows the following quantification rules: For the absolute value of frequency difference Or total harmonic distortion For larger operating conditions, prioritize larger options. As a label. Specifically, in In the process, several candidate values ​​that minimize the final phase difference fluctuation (standard deviation) are selected, and the maximum value is chosen as the label for this operating condition. .

[0044] For vibration amplitude Larger or signal-to-noise ratio For lower operating conditions, prioritize smaller options. (i.e., larger) () as a tag. Specifically, in After filtering out several candidate values ​​that minimize the degree of fluctuation, the minimum value is selected as the label.

[0045] For general operating conditions, the option that minimizes the fluctuation should be selected directly. As a label.

[0046] Using the labeling rules described above, the trained neural network automatically learns the monotonic mapping relationship between operating parameters and fusion weights: or The larger the output The larger; The larger or The lower the output The smaller (i.e.) The larger (the larger).

[0047] Based on the first fusion weight and the second fusion weight, the compensated hardware phase difference and the corrected software phase difference are weighted and fused to generate the final phase difference. : ; Calculate the fusion residual between the compensated hardware phase difference and the corrected software phase difference: ; Step A6, System Health Diagnosis and Alarms (Optional): By recording and analyzing residuals over a long period of time The time series. Under normal circumstances, the fused residuals... It should fluctuate slightly around zero. If a fusion residual is detected... If the value continues to increase and exceeds the first preset threshold, it indicates that the performance of one of the detection channels may be degraded. The first preset threshold is set as follows: under normal operating conditions of the synchronous device (no faults, no strong external interference), the fusion residual is continuously measured and its standard deviation is calculated. ,Pick or As the first preset threshold. If If the value exceeds the first preset threshold for multiple consecutive measurement cycles (e.g., 10 cycles), the performance is deemed to have degraded.

[0048] If the fusion residual If a sudden step jump occurs and exceeds the second preset threshold, a hardware fault is determined in the corresponding channel. The system immediately issues an alarm signal and blocks the synchronous closing command. The second preset threshold is set to twice the first preset threshold under normal operating conditions. The fused residual of the two independent channels is used as a feedback signal to achieve closed-loop self-diagnosis of the entire measurement system.

[0049] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A phase difference detection method based on dual verification, applied to a synchronization device, characterized in that, The phase difference detection method includes the following steps: Step A1: Collect the system-side voltage signal and the object-side voltage signal of the synchronizing device, and obtain the preprocessed system-side voltage signal and object-side voltage signal after preprocessing; set up external sensors to synchronously collect the vibration signal of the environment where the synchronizing device is located; Step A2: Set up hardware detection channel and software detection channel, and input the preprocessed system-side voltage signal and object-side voltage signal in parallel into the hardware detection channel and software detection channel to generate hardware phase difference and initial software phase difference respectively; Step A3: Based on the vibration signal, construct a vibration compensation model, calculate the phase error compensation amount caused by vibration, and feed it back to the hardware detection channel in real time to dynamically compensate the hardware phase difference and generate the compensated hardware phase difference. Step A4: Extract multi-brightness echo features from the preprocessed system-side voltage signal and the object-side voltage signal to obtain interference phase features; identify and quantify the interference phase features, correct the initial software phase difference based on the interference phase features, and generate the corrected software phase difference; Step A5: Calculate the current operating parameters in real time, dynamically generate the first fusion weight and the second fusion weight based on the current operating parameters, and perform weighted fusion on the compensated hardware phase difference and the corrected software phase difference to generate the final phase difference.

2. The phase difference detection method based on dual verification according to claim 1, characterized in that: The system-side voltage signal and the object-side voltage signal are preprocessed, specifically including: setting up a signal conditioning circuit, an analog-to-digital converter, and a zero-crossing comparator; the system-side voltage signal and the object-side voltage signal are filtered, amplified, and their amplitudes are adjusted by the signal conditioning circuit to match the input range of the analog-to-digital converter and the zero-crossing comparator.

3. The phase difference detection method based on dual verification according to claim 2, characterized in that: The preprocessed system-side voltage signal and object-side voltage signal are converted into square wave signals by a zero-crossing comparator through a hardware detection channel. The time difference between the two square wave signals is measured, and the hardware phase difference is calculated in combination with the system's rated frequency. The preprocessed system-side voltage signal and the object-side voltage signal are sampled synchronously using analog and digital methods through the software detection channel. The real and imaginary parts of the fundamental phasors of the preprocessed system-side voltage signal and the object-side voltage signal are calculated using discrete Fourier transform. The initial software phase difference is obtained through arctangent operation.

4. The phase difference detection method based on dual verification according to claim 3, characterized in that: The specific steps for constructing the vibration compensation model include: establishing a transfer function model between the vibration signal and the zero-crossing phase jitter in the hardware detection channel; setting an adaptive filter to process the vibration signal in real time and generate a real-time compensation quantity; and fusing the real-time compensation quantity with the hardware phase difference in the form of negative feedback to generate the compensated hardware phase difference.

5. The phase difference detection method based on dual verification according to claim 4, characterized in that: The preprocessed system-side voltage signal and object-side voltage signal are subjected to multi-bright-point echo feature extraction. The specific steps include: performing time-frequency analysis on the preprocessed system-side voltage signal and object-side voltage signal to separate the signal components arriving at different times and identify the echo components from multiple scattering bright spots in the hardware structure of the synchronous device; calculating the relative propagation path difference between the echo components of different scattering bright spots; quantizing the interference pattern between the echoes based on the relative propagation path difference, and extracting the dominant frequency and amplitude of the interference pattern as interference phase features.

6. The phase difference detection method based on dual verification according to claim 5, characterized in that: The initial software phase difference is corrected based on the interference phase characteristics. Specifically, the product of the interference phase characteristics and the preset correction coefficient is subtracted from the initial software phase difference to obtain the corrected software phase difference.

7. The phase difference detection method based on dual verification according to claim 1, characterized in that: The generation of the first and second fusion weights specifically includes: inputting the operating parameters into the pre-trained dynamic weight generation model, and outputting the first fusion weight of the compensated hardware phase difference and the second fusion weight of the corrected software phase difference; the operating parameters include the absolute value of the frequency difference, the total harmonic distortion rate, the signal-to-noise ratio, and the vibration amplitude; in the dynamic weight generation model, when the absolute value of the frequency difference or the total harmonic distortion rate increases, the first fusion weight is increased; when the vibration amplitude increases or the signal-to-noise ratio decreases, the second fusion weight is increased.

8. The phase difference detection method based on dual verification according to claim 6, characterized in that: It also includes step A6: calculating the fusion residual between the compensated hardware phase difference and the corrected software phase difference, and evaluating the health status of the hardware detection channel and the software detection channel based on the time series change trend of the fusion residual.

9. The phase difference detection method based on dual verification according to claim 8, characterized in that: When the fusion residual continuously increases beyond the first preset threshold, it is determined that the corresponding channel has experienced performance degradation. When the fusion residual step exceeds the second preset threshold, it is determined that a hardware fault has occurred, and an alarm and a blocking synchronous closing command are issued.

Images