Intelligent crystal resonator control method based on adaptive frequency compensation

Abstract

The invention belongs to the technical field of crystal resonator management and control, and particularly relates to a crystal resonator intelligent control method based on adaptive frequency compensation, which comprises the steps of data monitoring and acquisition, feature extraction and fusion, adaptive compensation decision, dynamic execution adjustment and closed-loop calibration analysis. According to the invention, environmental parameters are monitored in real time through multi-sensor fusion, an optimal compensation strategy is dynamically generated in combination with a machine learning algorithm, adaptive frequency compensation and automatic precise control of the crystal resonator are realized, long-term stability and reliability of the crystal resonator in a complex environment are significantly improved, and the service life of the crystal resonator is prolonged. The operation performance of the crystal resonator can be ensured, and the monitoring personnel can take corresponding improvement and optimization measures in time by reasonably analyzing the acquisition monitoring abnormal performance of the crystal resonator and the power consumption condition of the crystal resonator and giving an alarm in time. The operation performance of the crystal resonator is further ensured; and the operation supervision difficulty is obviously reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of crystal resonator control, and specifically to an intelligent control method for crystal resonators based on adaptive frequency compensation. Background Art

[0002] A crystal resonator is an electronic component based on quartz crystals or other piezoelectric materials. Its core principle is to utilize the piezoelectric effect to achieve the mutual conversion between mechanical vibration and electrical signals. Its high-precision and high-stability frequency output is the cornerstone of technologies such as communication, navigation, and computing;

[0003] Traditional frequency compensation methods for crystal resonators generally adopt fixed compensation strategies or simple linear models, which cannot dynamically adapt to environmental changes and cope with non-linear disturbances, resulting in insufficient accuracy, making it difficult to achieve adaptive frequency compensation and automated precise control of crystal resonators, and unable to reasonably analyze and promptly alarm the abnormal manifestations of acquisition and monitoring of crystal resonators and the power consumption status of crystal resonators, making it difficult to ensure the operating performance of crystal resonators and significantly reducing the difficulty of its operation supervision;

[0004] In view of the above technical deficiencies, a solution is proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent control method for crystal resonators based on adaptive frequency compensation, which solves the problems that the prior art cannot dynamically adapt to environmental changes and cope with non-linear disturbances, making it difficult to achieve adaptive frequency compensation and automated precise control of crystal resonators, and unable to reasonably analyze and promptly alarm the abnormal manifestations of acquisition and monitoring of crystal resonators and the power consumption status of crystal resonators, making it difficult to ensure the operating performance of crystal resonators and significantly reducing the difficulty of its operation supervision.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] An intelligent control method for crystal resonators based on adaptive frequency compensation, comprising the following steps:

[0008] Step 1: A multi-parameter perception and transmission module collects the working environment parameters and output signal characteristics of the crystal resonator in real time, and the original signal is transmitted to the feature extraction and fusion module through the SPI / I2C bus;

[0009] Step 2: The feature extraction and fusion module processes the original signal to obtain a standardized feature vector, and the standardized feature vector is transmitted to the adaptive compensation decision module through the DMA channel;

[0010] Step 3: The adaptive compensation decision module trains a lightweight neural network model based on the historical dataset to learn the mapping relationship between environmental parameters and the compensation amount. The current feature vector and the historical data of the previous ten time steps are input into the model, and the model analyzes the input data and outputs a compensation parameter combination.

[0011] Step 4: The dynamic execution adjustment module converts the compensation voltage into an analog signal through a 12-bit DAC, applies it to the voltage-controlled end of the resonator, synchronously adjusts the load capacitance and the loop gain, and outputs a frequency signal to the phase-locked loop.

[0012] Step 5: The closed-loop calibration analysis module monitors the instantaneous deviation of the output frequency through the phase-locked loop, and triggers recalibration when the instantaneous deviation of the output frequency exceeds ±0.5 ppm.

[0013] Further, in Step 1, the multi-parameter perception and transmission module includes a temperature sensor, a three-axis vibration sensor, a voltage monitoring sensor, and a frequency counter; among them, the temperature sensor is used to monitor the surface temperature and the environmental temperature of the crystal resonator, with an accuracy of ±0.05 °C;

[0014] The three-axis vibration sensor is used to capture the axial vibration spectra of the X, Y, and Z axes of the crystal resonator, with a bandwidth of 0 - 15 kHz; the voltage monitoring sensor is used to detect the driving voltage fluctuation of the crystal resonator, with a resolution of 1 mV; the frequency counter is used to measure the instantaneous value of the output frequency, with an accuracy of ±0.1 ppm.

[0015] Further, in Step 2, the processing process of the feature extraction and fusion module includes:

[0016] Noise suppression: Perform Kalman filtering and moving average processing on the original signal;

[0017] Feature generation: Extract time-domain features including the temperature change rate, the proportion of the main vibration frequency energy, and the peak-to-peak value of the voltage ripple. Analyze the harmonic distribution of the vibration spectrum through fast Fourier transform, and calculate the aging factor;

[0018] Data fusion: Integrate multi-dimensional features into a standardized feature vector, with the dimension ≤ 20.

[0019] Further, in Step 3, the compensation parameter combination includes a voltage compensation amount, dynamic PID parameters, and a filter cut-off frequency; among them, the adjustment range of the voltage compensation amount is 0 - 3.3 V, with a step of 1 mV; the filter cut-off frequency is dynamically adjusted according to the vibration spectrum.

[0020] Further, in Step 5, when recalibration is triggered, the recalibration process is as follows:

[0021] Freeze the current compensation parameters and collect frequency fluctuation data within 5 ms;

[0022] Determine the main interference source through fast Fourier analysis;

[0023] Generate a correction coefficient and feedback it to the adaptive compensation decision module to update the compensation parameters for the next cycle.

[0024] Furthermore, the adaptive compensation decision module, the dynamic execution adjustment module, and the closed-loop calibration analysis module are all communicatively connected to the supervision terminal. The adaptive compensation decision module, the dynamic execution adjustment module, and the closed-loop calibration analysis module send compensation decision information, dynamic execution adjustment information, and closed-loop calibration analysis information to the supervision terminal for display, and the supervision terminal is communicatively connected to the acquisition cooperation monitoring module;

[0025] The acquisition cooperation monitoring module analyzes the acquisition cooperation performance of all sensors in the multi-parameter perception transmission module, generates an acquisition cooperation alarm signal or an acquisition cooperation normal signal accordingly, and sends the acquisition cooperation alarm signal or the acquisition cooperation normal signal to the supervision terminal for display. When the supervision terminal receives the acquisition cooperation alarm signal, it issues a corresponding early warning.

[0026] Furthermore, the specific analysis process of the acquisition cooperation monitoring module is as follows:

[0027] Obtain all the sensors involved in the multi-parameter perception transmission module, mark the corresponding sensor as i, and i is a natural number greater than 1; collect the production date of sensor i and mark the time interval between it and the current date as the production time characteristic value, and mark the total running time of sensor i in the historical stage as the running time characteristic value; and take the current moment as the end moment and trace back for a tracing period with a set duration of T1. Calculate the ratio of the number of times sensor i fails within the tracing period to the running time of sensor i within the tracing period to obtain the monitoring stability difference value;

[0028] Calculate the perception hidden danger value by performing a weighted sum calculation on the production time characteristic value, the running time characteristic value, and the monitoring stability difference value, and compare the perception hidden danger value with the corresponding preset perception hidden danger threshold. If the perception hidden danger value exceeds the corresponding preset perception hidden danger threshold, mark sensor i as a blocked sensor;

[0029] If the perception hidden danger value does not exceed the corresponding preset perception hidden danger threshold, obtain the number of occurrences where the acquisition frequency of sensor i within a unit time is not within the corresponding preset acquisition frequency range and mark it as the acquisition frequency difference value. Compare the acquisition frequency difference value with the corresponding preset acquisition frequency difference threshold. If the acquisition frequency difference value exceeds the corresponding preset acquisition frequency difference threshold, mark sensor i as a blocked sensor;

[0030] If the collected frequency deviation value does not exceed the corresponding preset frequency deviation threshold, when the collection frequency is not within the corresponding preset collection frequency range, obtain the frequency deviation amplitude data, calculate the average value of all the frequency deviation amplitude data within a unit time to obtain the frequency deviation performance value, and mark the frequency deviation amplitude data with the largest value within a unit time as the frequency deviation table amplitude value;

[0031] Calculate the collection monitoring value by performing weighted summation of the collected frequency deviation value, the frequency deviation performance value, and the frequency deviation table amplitude value, and compare the collection monitoring value with the corresponding preset collection monitoring threshold. If the collection monitoring value exceeds the corresponding preset collection monitoring threshold, mark the sensor i as a blocked sensor; if there is a blocked sensor in the multi-parameter perception transmission module, generate a collection cooperation alarm signal; if there is no blocked sensor in the multi-parameter perception transmission module, generate a collection cooperation normal signal.

[0032] Further, the supervision terminal is communicatively connected to the multi-mode switching module. The multi-mode switching module comprehensively analyzes and calculates the environmental complexity index based on the temperature change rate, vibration energy, and frequency deviation, and dynamically switches modes based on the environmental complexity index, including the low-power mode, the balanced mode, and the high-performance mode; among them, the low-power mode only maintains the core functions, the balanced mode maintains the benchmark operating state, and the high-performance mode maintains full-load operation.

[0033] Further, the supervision terminal is communicatively connected to the power consumption management evaluation module. The supervision terminal sends the collection cooperation normal signal to the power consumption management evaluation module. The power consumption management evaluation module analyzes the power consumption status of the crystal resonator within a unit time when receiving the collection cooperation normal signal, generates a power consumption management qualified signal or a power consumption management alarm signal through the analysis, and sends the power consumption management qualified signal or the power consumption management alarm signal to the supervision terminal for display. When the supervision terminal receives the power consumption management alarm signal, it issues a corresponding early warning.

[0034] Further, the specific analysis process of the power consumption management evaluation module is as follows:

[0035] Set a number of detection periods within a unit time, collect the energy consumption of the crystal resonator in the corresponding detection periods and compare it with the corresponding preset energy consumption standard value. If the energy consumption exceeds the preset energy consumption standard value, mark the corresponding detection period as a power consumption management abnormal period; obtain the number of power consumption management abnormal periods within a unit time and calculate the ratio with the total number of detection periods to obtain the power consumption management abnormal characteristic value. Compare the power consumption management abnormal characteristic value with the preset power consumption management abnormal characteristic threshold. If the power consumption management abnormal characteristic value exceeds the preset power consumption management abnormal characteristic threshold, generate a power consumption management alarm signal;

[0036] If the abnormal characteristic value of the consumption tube does not exceed the preset abnormal characteristic threshold of the consumption tube, then calculate the ratio of the energy consumption in the corresponding detection period to the corresponding preset energy consumption standard value to obtain the measured energy consumption ratio, calculate the average value of the measured energy consumption ratios in all detection periods to obtain the energy consumption analysis value, and mark the measured energy consumption ratio with the largest value as the abnormal energy consumption value;

[0037] Calculate the power consumption management evaluation value by weighted summation of the abnormal characteristic value of the consumption tube, the energy consumption analysis value and the abnormal energy consumption value, and compare the power consumption management evaluation value with the preset power consumption management evaluation threshold. If the power consumption management evaluation value exceeds the preset power consumption management evaluation threshold, generate a power consumption management alarm signal; if the power consumption management evaluation value does not exceed the preset power consumption management evaluation threshold, generate a power consumption management qualified signal.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] 1. In the present invention, by integrating data monitoring and acquisition, feature extraction and fusion, adaptive compensation decision-making, dynamic execution and adjustment, and closed-loop calibration and analysis, it is possible to monitor the environmental parameters of the crystal resonator in real time and dynamically generate an optimal compensation strategy through analysis, realizing the adaptive frequency compensation and automatic precise control of the crystal resonator, significantly improving the long-term stability and reliability of the resonator in a complex environment, being beneficial to ensuring the operating performance of the crystal resonator, and having a high level of intelligence;

[0040] 2. In the present invention, by analyzing the acquisition cooperation performance of all sensors in the multi-parameter perception and transmission module through the acquisition cooperation monitoring module, when generating an acquisition cooperation alarm signal, check, repair or replace the corresponding sensor, which is beneficial to maintaining the accuracy of the compensation analysis result and the control stability of the crystal resonator, and when generating an acquisition cooperation normal signal, analyze the power consumption status of the crystal resonator through the power consumption management evaluation module, and make reasonable improvement and optimization measures when generating a power consumption management alarm signal, ensuring the safe, stable and energy-saving operation of the crystal resonator, and further reducing the operation supervision difficulty of the crystal resonator. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] For the convenience of those skilled in the art to understand, the present invention will be further described below with reference to the accompanying drawings;

[0042] Figure 1 It is the flowchart of the method in the first embodiment of the present invention;

[0043] Figure 2 It is the system block diagram in the first embodiment of the present invention;

[0044] Figure 3 It is the system block diagram in the second and third embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] Example 1: Figure 1-2 As shown, the crystal resonator intelligent control method based on adaptive frequency compensation proposed by the present invention includes the following steps:

[0047] Step 1: The multi-parameter sensing and transmission module collects the working environment parameters and output signal characteristics of the crystal resonator in real time, and transmits the original signal to the feature extraction and fusion module via the SPI / I2C bus. It should be noted that the multi-parameter sensing and transmission module includes a temperature sensor, a three-axis vibration sensor, a voltage monitoring sensor, and a frequency counter. Among them, the temperature sensor is used to monitor the surface temperature and ambient temperature of the crystal resonator with an accuracy of ±0.05°C.

[0048] The three-axis vibration sensor is used to capture the axial vibration spectrum of the crystal resonator's X, Y, and Z axes, with a bandwidth of 0-15kHz. The voltage monitoring sensor is used to detect the driving voltage fluctuation of the crystal resonator with a resolution of 1mV. The frequency counter is used to measure the instantaneous value of the output frequency with an accuracy of ±0.1ppm.

[0049] Step 2: The feature extraction and fusion module processes the original signal to obtain a standardized feature vector, which is then transmitted to the adaptive compensation decision module via the DMA channel. The specific processing is as follows:

[0050] Noise suppression: Kalman filtering and sliding average processing are performed on the original signal to eliminate high-frequency interference and transient noise;

[0051] Feature generation: Extract time-domain features such as temperature change rate, vibration main frequency energy ratio, and voltage ripple peak-to-peak value. Analyze the harmonic distribution of the vibration spectrum through fast Fourier transform (FFT) and calculate the aging factor (based on the weighted value of accumulated operating time and temperature stress).

[0052] Data fusion: Integrate multidimensional features into standardized feature vectors with dimensions ≤ 20 for subsequent decision making.

[0053] Step 3: The adaptive compensation decision module trains a lightweight neural network model based on historical data sets (including temperature, vibration, frequency deviation, etc.) to learn the mapping relationship between environmental parameters and compensation amounts. The current feature vector and historical data of the previous ten time steps are input to the model. The model analyzes the input data and outputs a compensation parameter combination.

[0054] Among them, the compensation parameter combination includes the voltage compensation amount (adjustment range 0 to 3.3V, step 1mV), the dynamic PID parameters (weight distribution of the proportional, integral, and differential terms), and the filter cut-off frequency (dynamically adjusted according to the vibration spectrum).

[0055] Step 4: The dynamic execution adjustment module converts the compensation voltage into an analog signal through a 12-bit DAC, applies it to the voltage-controlled terminal of the resonator, synchronously adjusts the load capacitance and the loop gain to ensure the stability of the resonance point, and outputs a frequency signal to the phase-locked loop (PLL) to further suppress phase noise;

[0056] Furthermore, the dynamic execution adjustment module mainly involves a voltage-controlled oscillation circuit (receiving the compensation voltage and adjusting the output frequency of the crystal resonator), a programmable load capacitance array (switching the capacitance value according to the instruction, range 10 - 30pF, step 0.5pF), and a digital potentiometer (adjusting the feedback loop gain to match the PID parameters).

[0057] Step 5: The closed-loop calibration analysis module monitors the instantaneous deviation of the output frequency through the phase-locked loop, and triggers re-calibration when the instantaneous deviation of the output frequency exceeds ±0.5ppm; and when re-calibration is triggered, the re-calibration process is as follows:

[0058] Freeze the current compensation parameters and collect the frequency fluctuation data within 5ms;

[0059] Determine the main interference source (such as temperature or vibration dominance) through fast Fourier analysis;

[0060] Generate a correction coefficient and feedback it to the adaptive compensation decision module to update the compensation parameters for the next cycle; and write the calibration result into the non-volatile memory for long-term aging compensation.

[0061] It should be noted that the adaptive compensation decision module, the dynamic execution adjustment module, and the closed-loop calibration analysis module are all communicatively connected to the supervision terminal. The adaptive compensation decision module, the dynamic execution adjustment module, and the closed-loop calibration analysis module send the compensation decision information, the dynamic execution adjustment information, and the closed-loop calibration analysis information to the supervision terminal for display, so as to facilitate the user to master the relevant information in detail and perform manual intervention control in a timely manner, further ensuring the operation effect of the crystal resonator.

[0062] Furthermore, the supervision terminal is communicatively connected to the multi-mode switching module. The multi-mode switching module comprehensively analyzes and calculates the environmental complexity index based on the temperature change rate, the vibration energy, and the frequency deviation, and dynamically switches modes based on the environmental complexity index, including the low-power mode, the balanced mode, and the high-performance mode; and sends the switching information to the supervision terminal to facilitate the supervisor to master the mode information in detail and perform manual intervention control in a timely manner according to needs; it should be noted that the calculation formula of the environmental complexity index is as follows:

[0063] ; where, ECI represents the environmental complexity index, |dT / dt| represents the temperature change rate, dT represents the temperature change amount, dt represents the duration; RMS represents the vibration energy, |Δf| represents the frequency deviation;

[0064] Among them, in the low-power mode (preferably, when ECI < 1), only the core functions are maintained (for example, only the core sensors and the clock circuit are powered); in the balanced mode (preferably, when 1 ≤ ECI < 3), the reference operating state is maintained and the redundant computing units are turned off; in the high-performance mode (preferably, when ECI ≥ 3), full-load operation is maintained; through the dynamic switching management of the modes, the power consumption is reduced while ensuring the operation effect of the crystal resonator.

[0065] The present invention realizes the frequency deviation of the crystal resonator within ±0.3 ppm and the phase noise improvement exceeding 10 dBc / Hz@1 kHz in the wide temperature range of -40°C to 85°C through real-time monitoring of environmental parameters by multi-sensor fusion and dynamically generating an optimal compensation strategy in combination with a machine learning algorithm. At the same time, it has a fast dynamic response ability of less than 10 ms, significantly improving the long-term stability and reliability of the resonator in a complex environment, which is beneficial to ensuring the operation performance of the crystal resonator and reducing the operation supervision difficulty.

[0066] Example 2: As Figure 3 shown, the difference between this embodiment and Embodiment 1 is that the supervision terminal is communicatively connected to the acquisition and cooperation monitoring module, and the acquisition and cooperation monitoring module analyzes the acquisition and cooperation performance of all sensors in the multi-parameter perception and transmission module, generates an acquisition and cooperation alarm signal or an acquisition and cooperation normal signal accordingly, and sends the acquisition and cooperation alarm signal or the acquisition and cooperation normal signal to the supervision terminal for display;

[0067] When the supervision terminal receives the acquisition and cooperation alarm signal, it issues a corresponding warning to remind the supervisor to check, repair or replace the corresponding sensor to ensure the monitoring and acquisition performance of the crystal resonator, which is beneficial to maintaining the accuracy and control stability of the compensation analysis result of the crystal resonator; the specific analysis process of the acquisition and cooperation monitoring module is as follows:

[0068] All sensors involved in the multi-parameter perception and transmission module are obtained, and the corresponding sensors are marked as i, and i is a natural number greater than 1; the interval duration between the production date of sensor i and the current date is marked as the production time characteristic value, and the total operating duration of sensor i in the historical stage is marked as the operating time characteristic value; and taking the current moment as the end moment and tracing back for a tracing period with a set duration of T1, the ratio of the number of times sensor i fails within the tracing period to the operating duration of sensor i within the tracing period is calculated to obtain the monitoring stability and abnormality value;

[0069] The perceived hidden danger value is calculated by weighted summation of the production-time characteristic value, the operation-time characteristic value, and the monitoring stability deviation value, that is, corresponding preset weight coefficients are assigned to the production-time characteristic value, the operation-time characteristic value, and the monitoring stability deviation value respectively, and the production-time characteristic value, the operation-time characteristic value, and the monitoring stability deviation value are multiplied by the corresponding preset weight coefficients respectively, and the sum of the three groups of product results is marked as the perceived hidden danger value; moreover, the larger the value of the perceived hidden danger value, the worse the comprehensive quality condition of the sensor i;

[0070] The perceived hidden danger value is numerically compared with the corresponding preset perceived hidden danger threshold. If the perceived hidden danger value exceeds the corresponding preset perceived hidden danger threshold, it indicates that the comprehensive quality condition of the sensor i is poor and it is not conducive to ensuring its monitoring and acquisition performance, then the sensor i is marked as a blocked sensor;

[0071] If the perceived hidden danger value does not exceed the corresponding preset perceived hidden danger threshold, it indicates that the comprehensive quality condition of the sensor i is good, then the occurrence times when the acquisition frequency of the sensor i within the unit time is not within the corresponding preset acquisition frequency range are obtained and marked as the acquisition frequency deviation value, and the acquisition frequency deviation value is numerically compared with the corresponding preset acquisition frequency deviation threshold. If the acquisition frequency deviation value exceeds the corresponding preset acquisition frequency deviation threshold, it indicates that the acquisition frequency execution of the sensor i is unstable initially, then the sensor i is marked as a blocked sensor;

[0072] If the acquisition frequency deviation value does not exceed the corresponding preset acquisition frequency deviation threshold, then the frequency deviation amplitude data is obtained when the acquisition frequency is not within the corresponding preset acquisition frequency range, the mean value of all the frequency deviation amplitude data within the unit time is calculated to obtain the frequency deviation performance value, and the frequency deviation amplitude data with the largest value within the unit time is marked as the frequency deviation table amplitude value;

[0073] The acquisition monitoring value is calculated by weighted summation of the acquisition frequency deviation value, the frequency deviation performance value, and the frequency deviation table amplitude value, that is, corresponding preset weight coefficients are assigned to the acquisition frequency deviation value, the frequency deviation performance value, and the frequency deviation table amplitude value respectively, and the acquisition frequency deviation value, the frequency deviation performance value, and the frequency deviation table amplitude value are multiplied by the corresponding preset weight coefficients respectively, and the sum of the three groups of product results is marked as the acquisition monitoring value; moreover, the larger the value of the acquisition monitoring value, the worse the comprehensive acquisition frequency execution performance of the sensor i within the unit time;

[0074] The acquisition monitoring value is numerically compared with the corresponding preset acquisition monitoring threshold. If the acquisition monitoring value exceeds the corresponding preset acquisition monitoring threshold, it indicates that the comprehensive acquisition frequency execution performance of the sensor i within the unit time is poor, then the sensor i is marked as a blocked sensor;

[0075] If the memristive sensor is involved in the multi-parameter perception and transmission module, it indicates that the acquisition and monitoring cooperation performance of the multi-parameter perception and transmission module is poor, and then a collection cooperation alarm signal is generated; if the memristive sensor is not involved in the multi-parameter perception and transmission module, it indicates that the acquisition and monitoring cooperation performance of the multi-parameter perception and transmission module is good, and then a collection cooperation normal signal is generated.

[0076] Embodiment 3: As Figure 3 shown, the difference between this embodiment and Embodiment 1 and Embodiment 2 is that the supervision terminal is communicatively connected to the power consumption management evaluation module. The supervision terminal sends the collection cooperation normal signal to the power consumption management evaluation module. The power consumption management evaluation module analyzes the power consumption status of the crystal resonator within a unit time when receiving the collection cooperation normal signal, and generates a power consumption management qualified signal or a power consumption management alarm signal through the analysis;

[0077] and sends the power consumption management qualified signal or the power consumption management alarm signal to the supervision terminal for display. When the supervision terminal receives the power consumption management alarm signal, it issues a corresponding warning to remind the supervision personnel to promptly conduct cause investigation and analysis and make reasonable improvement and optimization measures to avoid the power consumption of the crystal resonator exceeding the standard, ensure the safe, stable and energy-saving operation of the crystal resonator, and further reduce the operation supervision difficulty of the crystal resonator. The specific analysis process of the power consumption management evaluation module is as follows:

[0078] Set several detection periods within a unit time, where the duration of all detection periods is the same; collect the energy consumption of the crystal resonator in the corresponding detection period and compare it with the corresponding preset energy consumption standard value. If the energy consumption exceeds the preset energy consumption standard value, it indicates that the energy consumption in the corresponding detection period is high and does not meet the power consumption management requirements, and then mark the corresponding detection period as an abnormal power consumption management period;

[0079] Obtain the number of abnormal power consumption management periods within a unit time and calculate the ratio with the total number of detection periods to obtain the abnormal power consumption management characteristic value. Compare the abnormal power consumption management characteristic value with the preset abnormal power consumption management characteristic threshold. If the abnormal power consumption management characteristic value exceeds the preset abnormal power consumption management characteristic threshold, it indicates that the power consumption management performance of the crystal resonator within a unit time is poor, and then a power consumption management alarm signal is generated;

[0080] If the abnormal power consumption management characteristic value does not exceed the preset abnormal power consumption management characteristic threshold, then calculate the ratio of the energy consumption in the corresponding detection period to the corresponding preset energy consumption standard value to obtain the energy consumption measurement value, calculate the average value of the energy consumption measurement values of all detection periods to obtain the energy consumption analysis value, and mark the energy consumption measurement value with the largest value as the energy consumption abnormal value;

[0081] The power consumption management evaluation value is obtained by calculating the weighted sum of the abnormal characteristic value of the consumption tube, the energy consumption analysis value, and the energy consumption abnormal value; that is, corresponding preset weight coefficients are assigned to the abnormal characteristic value of the consumption tube, the energy consumption analysis value, and the energy consumption abnormal value respectively, and the abnormal characteristic value of the consumption tube, the energy consumption analysis value, and the energy consumption abnormal value are multiplied by the corresponding preset weight coefficients respectively, and the sum value of the three groups of product results is marked as the power consumption management evaluation value; moreover, the larger the value of the power consumption management evaluation value, the worse the comprehensive power consumption management performance of the crystal resonator per unit time.

[0082] The power consumption management evaluation value is numerically compared with the preset power consumption management evaluation threshold. If the power consumption management evaluation value exceeds the preset power consumption management evaluation threshold, indicating that the comprehensive power consumption management performance of the crystal resonator per unit time is poor, a power consumption management alarm signal is generated; if the power consumption management evaluation value does not exceed the preset power consumption management evaluation threshold, indicating that the comprehensive power consumption management performance of the crystal resonator per unit time is good, a power consumption management qualified signal is generated.

[0083] The working principle of the present invention: When in use, the multi-parameter perception and transmission module is used to collect the working environment parameters and output signal characteristics of the crystal resonator in real time. The feature extraction and fusion module processes the original signal to obtain a standardized feature vector. The adaptive compensation decision module analyzes the input data and outputs a compensation parameter combination. The dynamic execution adjustment module executes corresponding compensation adjustment measures. The closed-loop calibration analysis module monitors the instantaneous deviation of the output frequency and determines whether to trigger re-calibration, realizing the adaptive frequency compensation and automatic precise control of the crystal resonator, improving the long-term stability and reliability of the crystal resonator in a complex environment, and through reasonably analyzing and timely alarming the abnormal performance of the acquisition and monitoring of the crystal resonator and the power consumption status of the crystal resonator, it is beneficial for the supervisors to make corresponding improvement and optimization measures in time, further ensuring the operation performance of the crystal resonator and significantly reducing the operation supervision difficulty, with a high level of intelligence.

[0084] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The determination of the threshold in the technical solution is based on the data mean obtained through training with a large number of data dimensions. The preferred embodiments do not elaborate on all the details, nor do they limit the present invention to only the specific implementation manners. Obviously, according to the content of this specification, many modifications and changes can be made. This specification selects and specifically describes these embodiments to better explain the principle and practical application of the present invention, so that those skilled in the art in the technical field can well understand and utilize the present invention. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. An intelligent control method for a crystal resonator based on adaptive frequency compensation, characterized in that It includes the following steps: Step 1: The multi-parameter perception and transmission module collects the working environment parameters and output signal characteristics of the crystal resonator in real time, and the original signal is transmitted to the feature extraction and fusion module through the SPI / I2C bus; Step 2: The feature extraction and fusion module processes the original signal to obtain a standardized feature vector, and the standardized feature vector is transmitted to the adaptive compensation decision module through the DMA channel; Step 3: The adaptive compensation decision module trains a lightweight neural network model based on the historical data set, learns the mapping relationship between the environment parameters and the compensation amount, inputs the current feature vector and the historical data of the previous ten time steps into the model, and the model analyzes the input data and outputs a compensation parameter combination; Step 4: The dynamic execution and adjustment module converts the compensation voltage into an analog signal through a 12-bit DAC, applies it to the voltage control terminal of the resonator, synchronously adjusts the load capacitance and the loop gain, and outputs a frequency signal to the phase-locked loop; Step 5: The closed-loop calibration analysis module monitors the instantaneous deviation of the output frequency through the phase-locked loop, and triggers re-calibration when the instantaneous deviation of the output frequency exceeds ±0.5ppm.

2. The intelligent control method of a crystal resonator based on adaptive frequency compensation according to claim 1, wherein In Step 1, the multi-parameter perception and transmission module includes a temperature sensor, a three-axis vibration sensor, a voltage monitoring sensor, and a frequency counter.

3. The intelligent control method of a crystal resonator based on adaptive frequency compensation according to claim 1, wherein, In Step 2, the processing process of the feature extraction and fusion module includes: Noise suppression: Perform Kalman filtering and moving average processing on the original signal; Feature generation: Extract time-domain features including the temperature change rate, the proportion of the main vibration frequency energy, and the peak-to-peak value of the voltage ripple, analyze the harmonic distribution of the vibration spectrum through fast Fourier transform, and calculate the aging factor; Data fusion: Integrate multi-dimensional features into a standardized feature vector.

4. The intelligent control method for a crystal resonator based on adaptive frequency compensation according to claim 1, wherein In Step 3, the compensation parameter combination includes a voltage compensation amount, dynamic PID parameters, and a filter cut-off frequency.

5. The intelligent control method of a crystal resonator based on adaptive frequency compensation according to claim 1, characterized in that In Step 5, when re-calibration is triggered, the re-calibration process is as follows: Freeze the current compensation parameters, collect frequency fluctuation data within 5ms; Determine the main interference source through fast Fourier analysis; Generate a correction coefficient and feedback it to the adaptive compensation decision module to update the compensation parameters for the next cycle.

6. The intelligent control method for a crystal resonator based on adaptive frequency compensation according to claim 1, wherein The adaptive compensation decision module, the dynamic execution and adjustment module, and the closed-loop calibration analysis module are all communicatively connected to the supervision terminal. The adaptive compensation decision module, the dynamic execution and adjustment module, and the closed-loop calibration analysis module send compensation decision information, dynamic execution and adjustment information, and closed-loop calibration analysis information to the supervision terminal for display, and the supervision terminal is communicatively connected to the acquisition and cooperation monitoring module; The acquisition and cooperation monitoring module analyzes the acquisition and cooperation performance of all sensors in the multi-parameter perception and transmission module, and generates an acquisition and cooperation alarm signal or an acquisition and cooperation normal signal accordingly. When the supervision terminal receives the acquisition and cooperation alarm signal, it issues a corresponding warning.

7. The intelligent control method of a crystal resonator based on adaptive frequency compensation according to claim 6, wherein The specific analysis process of the acquisition and cooperation monitoring module is as follows: Obtain all the sensors involved in the multi-parameter perception and transmission module. If there is a memristive sensor involved in the multi-parameter perception and transmission module, generate an acquisition and cooperation alarm signal; otherwise, generate an acquisition and cooperation normal signal.

8. The intelligent control method of a crystal resonator based on adaptive frequency compensation according to claim 6, characterized in that The supervision terminal is communicatively connected to the multi-mode switching module, and the multi-mode switching module dynamically switches modes based on the environmental complexity index, including a low-power mode, a balanced mode, and a high-performance mode.

9. The intelligent control method for a crystal resonator based on adaptive frequency compensation according to claim 8, wherein, The power consumption management evaluation module for the communication connection of the supervision terminal. When the power consumption management evaluation module receives the normal signal for collection cooperation, it analyzes the power consumption status of the crystal resonator within a unit time, generates a qualified signal for power consumption management or an alarm signal for power consumption management through the analysis. When the supervision terminal receives the alarm signal for power consumption management, it issues a corresponding early warning.

10. The intelligent control method of a crystal resonator based on adaptive frequency compensation according to claim 9, wherein The specific analysis process of the power consumption management evaluation module is as follows: Obtain the number of abnormal power consumption management periods within a unit time and calculate the ratio with the total number of detection periods to obtain the abnormal power consumption management characteristic value. If the abnormal power consumption management characteristic value exceeds the preset abnormal power consumption management characteristic threshold, an alarm signal for power consumption management is generated; if the abnormal power consumption management characteristic value does not exceed the preset abnormal power consumption management characteristic threshold, the power consumption management evaluation value is calculated by weighted summation of the abnormal power consumption management characteristic value, the power consumption analysis value, and the power consumption difference value. If the power consumption management evaluation value exceeds the preset power consumption management evaluation threshold, an alarm signal for power consumption management is generated; otherwise, a qualified signal for power consumption management is generated.

Images