Atomic clock atomic air chamber filling method

Abstract

The invention discloses a filling method for an atomic air chamber of an atomic clock. The method comprises the following steps: cleaning an atomic gas chamber by adopting an acetone solution, a hydrochloric acid solution and supercritical CO2 in sequence; the cleaned atomic gas chamber is put into a drying oven with ultraviolet light irradiation to be baked, and a pulse type vacuum-inert gas exchange environment is introduced into the atomic gas chamber through a connector of the atomic gas chamber, so that a passivation layer is formed on the inner wall of the atomic gas chamber; the inflation system inflates alkali metal steam and buffer gas into the cooled atomic gas chamber, the buffer gas is neon gas and carbon monoxide, and in the inflation process of the atomic gas chamber, the proportion of the neon gas and the carbon monoxide is adjusted based on the inflation amount of the alkali metal in the atomic gas chamber and the temperature of the atomic gas chamber; and sealing the atomic gas chamber under the condition that the alkali metal steam and the buffer gas in the atomic gas chamber reach set values. The problem that in the prior art, the filling precision of the atomic gas chamber is low due to the fact that the filling process of the atomic gas chamber is simple and extensive is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of atomic clock, in particular, to a method for filling an atomic gas cell of an atomic clock. BACKGROUND

[0002] One of the core components of an atomic clock is an atomic gas cell, which needs to encapsulate alkali metal atoms (such as rubidium, cesium) and buffer gas at the same time. The filling process of the atomic gas cell directly determines the atomic coherence, signal-to-noise ratio and long-term stability. In the prior art, when filling the atomic gas cell, a rough operation is usually adopted by using manual experience and single parameter feedback, and then problems such as alkali metal adsorption and ratio drift occur. That is, due to the simple and rough filling process of the atomic gas cell in the prior art, the filling precision of the atomic gas cell is low.

[0003] At present, no effective solution has been proposed for the above problems. SUMMARY

[0004] The embodiments of the present application provide a method for filling an atomic gas cell of an atomic clock to at least solve the problem of low filling precision of the atomic gas cell due to the simple and rough filling process of the atomic gas cell in the prior art.

[0005] According to an aspect of an embodiment of the present application, a method for filling an atomic gas cell of an atomic clock is provided, comprising: sequentially using acetone solution, hydrochloric acid solution and supercritical CO2 to clean the atomic gas cell; placing the cleaned atomic gas cell into an oven provided with ultraviolet light irradiation, baking, and introducing a pulsed vacuum-inert gas exchange environment into the atomic gas cell through an interface of the atomic gas cell to form a passivation layer on the inner wall of the atomic gas cell, wherein the wavelength of the ultraviolet light irradiation is 254 mm, and the pulsed vacuum-inert gas exchange environment refers to circulating vacuum and inert gas filling in the atomic gas cell; placing the cooled atomic gas cell into a gas filling system, the gas filling system is used to fill alkali metal vapor and buffer gas into the atomic gas cell, the buffer gas is neon and carbon monoxide, and the proportion between the neon and the carbon monoxide filled into the atomic gas cell is adjusted based on the amount of the alkali metal filled into the atomic gas cell and the temperature of the atomic gas cell during the filling process of the atomic gas cell; sealing the atomic gas cell when the alkali metal vapor and the buffer gas in the atomic gas cell reach a set value; and taking out the atomic gas cell from the gas filling system to complete the filling.

[0006] Optionally, in the embodiments of the present application, the supercritical CO2 cleaning process comprises: collecting images of the inner wall of the atomic cell in real time through the optical fiber endoscope, with an interval of less than or equal to 3 seconds; using a convolutional neural network to identify the residue distribution characteristics of the collected images of the inner wall of the atomic cell; dynamically adjusting the cleaning parameters of the supercritical CO2 according to the identification result; and cleaning based on the cleaning parameters to remove nanoscale silicate and metal oxide residues on the inner wall of the atomic cell, wherein the pressure of the supercritical CO2 in the cleaning parameters is adjusted within the range of 10-30 MPa, the temperature is adjusted within the range of 40-60°C, and the cleaning time is adjusted within the range of 5-15 minutes.

[0007] Optionally, in the embodiments of the present application, the gas filling system for filling alkali metal vapor and buffer gas into the atomic cell comprises: an alkali metal dosing device for filling the alkali metal vapor into the atomic cell through a connecting pipeline made of inert titanium alloy material, wherein the alkali metal vapor is filtered through a molecular sieve filter module before being filled into the atomic cell to adsorb and remove impurity gases carried in the alkali metal vapor; a mass sensor coupled to the alkali metal dosing device for outputting an alkali metal filling amount signal in real time, wherein the mass sensor has an accuracy of ±0.1 mg, the alkali metal dosing device is provided with a heating device for heating solid alkali metal to above the melting point to form the alkali metal vapor; a micro-flow sensor for monitoring the instantaneous flow of neon gas and carbon monoxide in real time; a distributed optical fiber temperature sensor arranged inside the atomic cell for detecting the temperature of the atomic cell in real time; a pressure sensor for detecting the pressure inside the atomic cell in real time; a high-speed digital I / O module comprising an FPGA for parallel outputting multiple control signals and performing priority management on the control signals; a PID control loop comprising the pressure sensor, a PID controller, a gas filling valve, and a gas cell pressure feedback channel for adjusting the pressure of the atomic cell; and a proportional control loop comprising the distributed optical fiber temperature sensor, the mass sensor, a proportional controller, a neon gas / carbon monoxide adjusting valve, and a spectral feedback channel for adjusting the flow ratio of neon gas and carbon monoxide.

[0008] Optionally, in the embodiment of the present application, the gas filling system for filling alkali metal vapor and buffer gas into the atomic chamber further comprises: a feedforward compensation module receiving a predicted value of temperature change and a predicted value of the alkali metal filling amount, inputting the predicted values into the PID controller, and compensating for potential gas leakage to reduce hysteresis of the gas filling system, wherein the predicted values are generated by a prediction model deployed in the FPGA, the prediction model uses historical data of the gas filling cycle as a training set, uses an LSTM network for modeling, and updates the parameters of the prediction model in an incremental manner after each gas filling is completed; a dynamic weight distribution module adjusts the integral term weight in the PID parameters according to the real-time temperature value, so that the integral term weight is increased under high temperature conditions to suppress the volatilization accumulation effect of the alkali metal vapor, wherein the integral term weight refers to the integral gain Ki; and a parameter self-tuning module uses a genetic algorithm, takes a weighted sum of the atomic chamber pressure error, the alkali metal filling amount error, the neon / carbon monoxide ratio error, and the temperature fluctuation range as a fitness function, and loads the optimal PID parameters and proportional control parameters calculated offline to the PID control loop and the proportional control loop to adapt to the coupling requirements under different environmental conditions.

[0009] Optionally, in the embodiment of the present application, the gas filling system for filling alkali metal vapor and buffer gas into the atomic chamber further comprises: a multivariate disturbance test module separating the leakage effect and the temperature effect through a step response experiment, and establishing a dynamic relationship matrix between the leakage effect and the temperature effect; embedding the dynamic relationship matrix into the feedforward compensation module, in the case that the distributed optical fiber temperature sensor detects a change in the temperature of the atomic chamber, the gas filling system corrects the gas leakage compensation amount according to the dynamic relationship matrix to avoid the gas leakage error caused by thermal expansion; a TEC refrigeration module comprising a ring-shaped TEC refrigeration sheet and the distributed optical fiber temperature sensor form a closed-loop temperature control channel, the FPGA adjusts the TEC drive current in real time according to the feedback of the distributed optical fiber temperature sensor, so that the temperature of the atomic chamber is stabilized within ±0.01K; after the gas filling is completed, the atomic chamber is sealed by a chamber sealing device.

[0010] Optionally, in the embodiment of the present application, the FPGA further comprises: preferentially enabling a temperature-based flow ratio adjustment mode after the gas filling system is started; automatically switching to a flow ratio adjustment mode based on the alkali metal filling amount in the case that the temperature of the atomic chamber is stabilized in the ±5℃ interval and exceeds 10 minutes; immediately switching back to the temperature-based flow ratio adjustment mode when the temperature fluctuation of the atomic chamber exceeds a set threshold.

[0011] Optionally, in the embodiment of the present application, the proportional controller comprises: comparing the flow ratio of neon and carbon monoxide in real time; and outputting a closing signal to the alkali metal quantitative dispenser when the flow ratio reaches a set ratio.

[0012] Optionally, in the embodiment of the present application, the feedforward compensation module continuously updates the predicted value of the temperature change and the predicted value of the alkali metal filling amount in a sliding window manner, and superimposes the updated predicted values on the output end of the PID controller in real time.

[0013] Optionally, in the embodiment of the present application, when performing the step response experiment, the multivariable disturbance test module simultaneously applies a step input to the temperature channel and the pressure channel through a programmable disturbance generator, and quantifies the coupling coefficient of the leakage effect and the temperature effect based on the spectral feedback signal, thereby updating the dynamic relationship matrix.

[0014] Optionally, in the embodiment of the present application, after receiving the sealing trigger signal sent by the FPGA, the gas chamber sealing device performs the following sealing procedure: closing the alkali metal quantitative dispenser; synchronously closing the neon / carbon monoxide regulating valve; and driving the laser sealing or mechanical pressure sealing mechanism to complete the sealing of the atomic gas chamber in a preset time sequence.

[0015] Optionally, in the embodiment of the present application, in the case of ultra-low temperature demand scenarios, the gas filling system for filling alkali metal vapor and buffer gas into the atomic gas chamber further comprises: lowering the temperature of the atomic gas chamber to 10K by using a mechanical refrigeration device; after the mechanical refrigeration, finely adjusting the temperature of the atomic gas chamber with an accuracy of ±0.01K by the TEC refrigeration module; and the FPGA coordinates the start-stop and power distribution of the mechanical refrigeration device and the TEC refrigeration module to realize seamless switching and joint control of two-stage refrigeration.

[0016] In the embodiment of the present application, cleaning the atomic gas chamber with acetone solution, hydrochloric acid solution and supercritical CO2 in turn significantly improves the cleanliness of the atomic gas chamber and reduces impurities in the atomic gas chamber. In addition, the formation of the passivation layer on the inner wall of the atomic gas chamber avoids the release of impurities and the consumption of alkali metal, further improving the filling accuracy of the atomic gas chamber. Moreover, the flow ratio of neon and carbon monoxide is adjusted in real time based on the two variables of alkali metal filling amount and atomic gas chamber temperature during the gas filling operation, ensuring the filling accuracy of the atomic gas chamber, thereby solving the problem of low filling accuracy of the atomic gas chamber caused by the simple and extensive atomic gas chamber filling process in the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

[0018] Figure 1 is an optional atomic clock atomic gas chamber charging method schematic diagram provided according to an embodiment of the application;

[0019] Figure 2 is another optional atomic clock atomic gas chamber charging method schematic diagram according to an embodiment of the application. DETAILED DESCRIPTION

[0020] In order to enable persons skilled in the art to better understand the scheme of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative labor should fall within the scope of protection of the present application.

[0021] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] In one embodiment, an atomic clock atomic gas chamber charging method is provided, as shown in Figure 1 The above-mentioned atomic clock atomic gas chamber charging method comprises:

[0023] S102, sequentially using acetone solution, hydrochloric acid solution, supercritical CO2 to clean the atomic gas chamber;

[0024] S104, baking the cleaned atomic cell in an oven with ultraviolet irradiation, and introducing a pulsed vacuum-inert gas exchange environment into the atomic cell through the interface of the atomic cell, so that a passivation layer is formed on the inner wall of the atomic cell, wherein the wavelength of the ultraviolet irradiation is 254 mm, and the pulsed vacuum-inert gas exchange environment refers to circulating vacuum and inert gas filling in the atomic cell;

[0025] S106, placing the cooled atomic cell into a gas filling system, the gas filling system being used to fill alkali metal vapor and buffer gas into the atomic cell, the buffer gas being neon and carbon monoxide, and the proportion between the neon and the carbon monoxide filled into the atomic cell is adjusted based on the amount of alkali metal filled into the atomic cell and the temperature of the atomic cell during the filling process of the atomic cell;

[0026] S108, sealing the atomic cell when the alkali metal vapor and the buffer gas in the atomic cell reach a set value;

[0027] S110, taking out the atomic cell from the gas filling system, and completing the filling process.

[0028] Optionally, in the embodiment of the present application, the oily substances on the inner wall of the atomic cell are cleaned using an acetone solution first, then the acetone remaining on the inner wall of the atomic cell is cleaned using distilled water, then other soluble substances on the inner wall of the atomic cell are cleaned using a hydrochloric acid solution, then the hydrochloric acid remaining on the inner wall of the atomic cell is cleaned using distilled water, and finally the nanoscale silicates and metal oxides on the inner wall of the atomic cell are cleaned using supercritical CO2.

[0029] The interface of the atomic cell can be a glass tail pipe or a metal pipe, and the supercritical CO2 refers to carbon dioxide in a supercritical fluid state which is neither liquid nor gas under the condition of a temperature ≥ 31.1 ℃ and a pressure ≥ 7.38 MPa, and has high diffusivity and high solubility.

[0030] The pulsed vacuum-inert gas exchange environment refers to circulating vacuum and inert gas filling in the atomic cell, and gradually using the gas flow to carry away the small particles, water vapor molecules and other volatile pollutants adsorbed on the inner wall of the atomic cell. The pulsed operation includes a vacuum pumping stage, an inert gas filling stage and a circulation repetition, wherein the vacuum pumping stage gradually reduces the impurity gas in the atomic cell by multiple times of pumping by a vacuum pump, the inert gas filling stage dilutes the residual impurities by filling high-purity inert gas such as argon or nitrogen, and the atomic cell impurity concentration is reduced to a very low level by multiple times of vacuum-pumping and gas-filling circulation. The last step of the pulsed vacuum-inert gas exchange is to pump to a high vacuum to remove the inert gas and impurities together.

[0031] The core of an atomic clock is to use the hyperfine level transition frequency of alkali metal atoms as the frequency reference, and the alkali metal atoms are usually cesium (Cs) or rubidium (Rb). The alkali metal atoms are placed in an atomic cell, and a buffer gas is filled in the atomic cell in addition to the alkali metal atoms. The purpose is to limit the movement range of the alkali metal atoms by frequent collisions between the buffer gas and the alkali metal atoms, reduce the Doppler broadening and interatomic collision broadening, thereby compressing the linewidth of the resonance signal from hundreds of kHz to the order of hundreds of Hz, and significantly improving the frequency stability of the atomic clock. The buffer gas can also prolong the interaction time of the alkali metal atoms with the electromagnetic field by reducing the collision between the alkali metal atoms and the wall of the atomic cell (the wall collision of the atomic cell will cause the atoms to decohere), so as to prolong the coherence time. The alkali metal vapor is a single, free alkali metal atom in a gaseous state, and the solid alkali metal is a metal crystal formed by a large number of alkali metal atoms combined by metal bonds.

[0032] Further, the atomic cell is sequentially cleaned by using acetone solution, hydrochloric acid solution and supercritical CO2; then the cleaned atomic cell is baked, and a pulsed vacuum-inert gas exchange environment is introduced during the baking process. At the same time, a wavelength of 254 mm ultraviolet light is arranged in the oven to irradiate, so that a passivation layer is formed on the inner wall of the atomic cell. After the baking of the atomic cell, the atomic cell is cooled; then the cooled atomic cell is placed in a gas filling system to fill the alkali metal vapor and the buffer gas into the atomic cell. During the gas filling operation, the alkali metal vapor is first filled into the atomic cell, and then the flow ratio of neon gas and carbon monoxide filled into the atomic cell is adjusted based on the temperature of the atomic cell. When the temperature of the atomic cell is less than 250℃, the flow ratio of neon gas to carbon monoxide is adjusted to 8:2; when the temperature of the atomic cell is greater than or equal to 250℃ and less than 300℃, the flow ratio of neon gas to carbon monoxide is adjusted to 6:4; when the temperature of the atomic cell is greater than or equal to 300℃ and less than 350℃, the flow ratio of neon gas to carbon monoxide is adjusted to 5:5; when the temperature of the atomic cell is greater than or equal to 350℃ and less than 460℃, the flow ratio of neon gas to carbon monoxide is adjusted to 4:6; when the temperature of the atomic cell is greater than or equal to 460℃, the gas filling operation is stopped. When the temperature of the atomic cell is stable in the interval of ±5℃ and exceeds 10 minutes, the flow ratio of neon gas to carbon monoxide filled into the atomic cell is adjusted based on the amount of alkali metal filled. When the amount of alkali metal filled is between 3mg and 10mg, the flow ratio of neon gas to carbon monoxide is adjusted to 8:2; when the amount of alkali metal filled is between 10mg and 20mg, the flow ratio of neon gas to carbon monoxide is adjusted to 6:4. When the flow of neon gas and carbon monoxide reaches the set value, the gas filling operation is stopped and the sealing is performed; the sealed atomic cell is taken out, thereby completing the preparation of the atomic cell.

[0033] By the embodiment of the present application, the buffer gas of neon and carbon monoxide can significantly prolong the service life of the atomic chamber while maintaining the stability of the atomic chamber; sequentially using acetone solution, hydrochloric acid solution, and supercritical CO2 to clean the atomic chamber significantly improves the cleanliness of the atomic chamber and reduces the impurities in the atomic chamber; in addition, the formation of the passivation layer on the inner wall of the atomic chamber avoids the release of impurities and the consumption of alkali metals, further improving the charging accuracy of the atomic chamber; and when performing the charging operation, the flow ratio of neon and carbon monoxide is adjusted in real time based on the two variables of the alkali metal charging amount and the atomic chamber temperature, ensuring the charging accuracy of the atomic chamber.

[0034] As an optional solution, the supercritical CO2 cleaning process includes:

[0035] S202, real-time acquisition of the above-mentioned atomic chamber inner wall image by the optical fiber endoscope, with an acquisition interval less than or equal to 3 seconds;

[0036] S204, using a convolutional neural network to identify the residue distribution characteristics of the collected above-mentioned atomic chamber inner wall image;

[0037] S206, dynamically adjusting the cleaning parameters of the above-mentioned supercritical CO2 according to the identification result;

[0038] S208, cleaning based on the above-mentioned cleaning parameters to remove the nanoscale silicate and metal oxide residues on the inner wall of the above-mentioned atomic chamber, wherein the pressure of the supercritical CO2 in the above-mentioned cleaning parameters is adjusted in the range of 10-30 MPa, the temperature is adjusted in the range of 40-60℃, and the cleaning time is adjusted in the range of 5-15 minutes.

[0039] Optionally, in the embodiment of the present application, the optical fiber endoscope is a miniature imaging device based on the principle of optical fiber image transmission and light transmission, which can enter a closed space through a narrow channel (such as the charging port or observation window of the atomic clock atomic chamber) to realize real-time acquisition of high-resolution atomic chamber inner wall images.

[0040] According to the identification result, the cleaning parameters of supercritical CO2 are dynamically adjusted, which means that the type of the residual on the inner wall of the atomization chamber is determined first. If it is metal oxide, the pressure is adjusted first. If it is silicate, the temperature is adjusted first. Then, the distribution of the residual on the inner wall of the atomization chamber is determined. If it is locally accumulated, the gradient pressure is used, that is, the medium pressure (such as 20 MPa) is used for global cleaning, and the high pressure (such as 30 MPa) is switched to the stubborn deposition area. If it is uniformly distributed, the stable pressure and temperature are used, such as the pressure of 20 MPa and the temperature of 50℃. Finally, according to the cleaning effect after adjusting the pressure and temperature parameters, it is determined whether the cleaning time needs to be extended. For example, when the residual on the inner wall of the atomization chamber is dense metal oxide, the cleaning parameters of high pressure (28-30 MPa), medium temperature (45-50℃) and long time (12-15 min) are used. When the residual on the inner wall of the atomization chamber is silicate, the cleaning parameters of medium pressure (20-25 MPa), high temperature (55-60℃) and medium time (10-12 min) are used. When the residual on the inner wall of the atomization chamber is thin-layer uniform pollution, the cleaning parameters of low pressure (10-15 MPa), low temperature (40-45℃) and short time (5-8 min) are used. When the residual on the inner wall of the atomization chamber is local stubborn deposition, the cleaning parameters of pulse high pressure (30-35 MPa), constant temperature (50℃) and segmented cleaning (10 min) are used.

[0041] Through the embodiments of the present application, the above-mentioned atomization chamber inner wall image is collected in real time by an optical fiber endoscope, and the collection interval is less than or equal to 3 seconds. The convolutional neural network is used to identify the residual distribution characteristics of the collected above-mentioned atomization chamber inner wall image. According to the identification result, the cleaning parameters of the above-mentioned supercritical CO2 are dynamically adjusted. Based on the above-mentioned cleaning parameters, the cleaning is performed to remove the nanoscale silicate and metal oxide residual on the above-mentioned atomization chamber inner wall. The pressure in the above-mentioned cleaning parameters is adjusted in the range of 10-30 MPa, the temperature is adjusted in the range of 40-60℃, and the cleaning time is adjusted in the range of 5-15 minutes. That is, the supercritical CO2 cleaning adjusts the cleaning parameters by introducing image recognition, thereby significantly improving the atomization chamber cleaning efficiency and the atomization chamber cleanliness.

[0042] As an optional solution, the above-mentioned inflation system is used to inflate the alkali metal vapor and the buffer gas into the above-mentioned atomization chamber, which includes:

[0043] ​S1, the alkali metal vapor is filled into the above-mentioned atomic gas chamber by the alkali metal quantitative dispenser through the connecting pipeline, wherein the connecting pipeline is made of titanium alloy material which has been inertized, and the alkali metal vapor is filtered by a molecular sieve filtering module before being filled into the atomic gas chamber to adsorb and remove the impurity gas carried by the alkali metal vapor; the alkali metal filling amount signal is output in real time by the mass sensor coupled to the alkali metal quantitative dispenser, wherein the accuracy of the mass sensor is ±0.1 mg, and the alkali metal quantitative dispenser is provided with a heating device, and the heating device is used to heat the solid alkali metal to above the melting point;

[0044] S2, the instantaneous flow rates of the neon gas and the carbon monoxide are monitored in real time by a micro-flow sensor; the temperature of the atomic gas chamber is detected in real time by a distributed optical fiber temperature sensor arranged inside the atomic gas chamber; the pressure inside the atomic gas chamber is detected in real time by a pressure sensor; a plurality of control signals are output in parallel by a high-speed digital I / O module comprising an FPGA, and priority management is performed on the control signals; the pressure of the atomic gas chamber is adjusted by a PID control loop composed of the pressure sensor, a PID controller, a gas filling valve and an atomic gas chamber pressure feedback channel; the flow rate ratio of the neon gas to the carbon monoxide is adjusted by a proportional control loop composed of the distributed optical fiber temperature sensor, the mass sensor, a proportional controller, a neon gas / carbon monoxide adjusting valve and a spectral feedback channel.

[0045] Optionally, in the embodiment of the present application, the alkali metal quantitative dispenser is a micro device for quantitatively releasing alkali metal atoms or vapor in a vacuum or controlled environment. The connecting pipeline can be made of titanium alloy material which has been inertized to prevent the alkali metal vapor from reacting with or contaminating the buffer gas. The molecular sieve filtering module comprises a titanium alloy shell, a cavity inside the titanium alloy shell and a molecular sieve adsorption medium filled in the cavity, wherein the titanium alloy shell is provided with sealing interfaces for docking with the connecting pipeline at both ends, the cavity is used to form a gas passage for the flow of alkali metal vapor, and the molecular sieve adsorption medium is selected from 3A, 4A, 5A, 13X or Ag + Modified type, for selectively adsorbing impurity gas, in addition, it also comprises a sintered metal filter screen or metal wire mesh support which is arranged at the upstream end and the downstream end of the molecular sieve adsorption medium to fix the molecular sieve adsorption medium and prevent particle migration.

[0046] The micro-flow sensor is a miniaturized sensor for measuring extremely small volume flow. The distributed optical fiber temperature sensor is a distributed temperature measurement system that uses an entire optical fiber as both a transmission line and a thermometer. The high-speed digital I / O module is a high-precision and high-response-speed digital signal input / output device that can detect feedback signals such as pressure sensors and temperature sensors to ensure the stability of the atomic gas chamber charging process. The FPGA (field programmable gate array) is the core processing unit of the high-speed digital I / O module, responsible for implementing high-precision timing control, real-time signal processing, and multi-device synchronization. The PID controller (proportional-integral-derivative controller) is the core control algorithm for ensuring the stability of the charging process parameters (such as pressure, temperature, gas flow, etc.). It adjusts in real time through feedback to make the inflation system quickly converge to the set value and suppress external disturbances. The heating device can be a heating furnace, and the mass sensor can be a quartz crystal microbalance sensor.

[0047] After the heating device heats the solid alkali metal to the melting point, the solid alkali metal becomes alkali metal vapor. Then the alkali metal quantitative dispenser fills the alkali metal vapor into the atomic gas chamber through the connecting pipeline, while the temperature control is above 5-10°C above the melting point of the alkali metal, the sampling frequency of the mass sensor is ≥100Hz, and the inner wall of the nozzle of the alkali metal quantitative dispenser has an alkali metal non-wetting coating. In addition, the alkali metal quantitative dispenser will pass through the molecular sieve filtering module before flushing the alkali metal vapor into the atomic gas chamber to adsorb and remove impurity gases carried in the alkali metal vapor.

[0048] For further example, the pressure sensor detects the total pressure inside the atomic gas chamber in real time and continuously, and converts the physical pressure signal into an electrical signal that can be recognized by the PID controller as the measurement value input of the PID control loop. The PID controller receives the target pressure and the current pressure feedback from the pressure sensor, calculates the error between the two, and outputs a control signal according to the proportional-integral-derivative algorithm to adjust the opening of the inflation valve, achieving fast and zero-error pressure control. The inflation valve, as an actuator, converts the control signal output by the PID controller into the actual gas flow, and the increase in the opening of the inflation valve indicates that more gas is filled into the atomic gas chamber, and the decrease in the opening of the inflation valve indicates that the inflation is reduced or stopped, and if necessary, it can be reversed to release pressure. The inflation valve directly affects the rising or falling rate of the pressure of the atomic gas chamber. The atomic gas chamber pressure feedback channel is composed of the gas space inside the atomic gas chamber, the pressure guide pipeline and the signal transmission link, responsible for transmitting the instantaneous true pressure of the atomic gas chamber to the pressure sensor without loss and low delay. Based on the above steps, the closed-loop feedback of detection-control-execution-detection is completed to achieve the regulation of the total pressure of the atomic gas chamber through the PID control loop.

[0049] In addition, the proportional control loop realizes dynamic adjustment of the flow ratio of neon and carbon monoxide through multi-sensor cooperation. The mass sensor monitors the actual flow of neon and carbon monoxide in real time. The distributed fiber temperature sensor monitors the temperature distribution of the atomic gas chamber to indirectly reflect the mixing uniformity of neon and carbon monoxide. The spectral feedback channel directly verifies the ratio of neon and carbon monoxide through laser absorption spectroscopy. The proportional controller continuously compares the target flow ratio of neon and carbon monoxide with the above feedback data, calculates the adjustment amount, and drives the neon / carbon monoxide adjustment valve to adjust the opening degree. For example, when the spectrum shows that the CO concentration is too high, the controller will reduce the CO valve opening degree or increase the Ne valve opening degree, so that the flow ratio of neon and carbon monoxide in the atomic gas chamber is adjusted to the specified target flow ratio.

[0050] According to the embodiment of the present application, the alkali metal vapor is filled into the atomic gas chamber through the alkali metal quantitative dispenser; the alkali metal filling amount signal is output in real time through the mass sensor coupled to the alkali metal quantitative dispenser; the pressure of the atomic gas chamber is adjusted through the PID control loop composed of the pressure sensor, the PID controller, the filling valve, and the atomic gas chamber pressure feedback channel; and the flow ratio of the neon and the carbon monoxide is adjusted through the proportional control loop composed of the distributed fiber temperature sensor, the mass sensor, the proportional controller, the neon / carbon monoxide adjustment valve, and the spectral feedback channel. That is, through multi-sensor closed-loop control, the adjustment accuracy of the alkali metal amount, the buffer gas flow, the temperature, and the pressure is improved, and the filling accuracy of the atomic clock atomic gas chamber is further improved.

[0051] As an optional solution, the gas filling system for filling the alkali metal vapor and the buffer gas into the atomic gas chamber further comprises:

[0052] S1, receiving the predicted value of the temperature change and the predicted value of the alkali metal filling amount through the feedforward compensation module, inputting the predicted value into the PID controller, compensating for potential gas leakage to reduce the hysteresis of the gas filling system, wherein the predicted value is generated by a prediction model deployed in the FPGA, the prediction model uses historical data of the gas filling cycle as a training set, uses an LSTM network for modeling, and updates the parameters of the prediction model in an incremental manner after each gas filling is completed;

[0053] S2, the dynamic weight distribution module adjusts the integral term weight in the PID parameter according to the real-time temperature value, so that the integral term weight is increased under high temperature conditions to suppress the volatilization accumulation effect of the alkali metal vapor, wherein the integral term weight refers to the integral gain Ki;

[0054] S3, the parameter self-tuning module adopts a genetic algorithm, taking the weighted sum of the atomic cell pressure error, the alkali metal charge error, the neon / carbon monoxide ratio error, and the temperature fluctuation range as the fitness function, and loading the optimal PID parameters and proportional control parameters calculated offline to the PID control loop and the proportional control loop to adapt to the coupling requirements under different environmental conditions.

[0055] Optionally, in the embodiments of the present application, the LSTM network refers to a long short-term memory network, the input of which includes atomic cell temperature, atomic cell pressure, neon instantaneous flow, carbon monoxide instantaneous flow, mass sensor reading, valve opening instruction, and spectrum feedback error, and the output of the LSTM network is temperature change prediction value and alkali metal charge prediction value, and the PID parameters include proportional gain Kp, integral gain Ki, and differential gain Kd.

[0056] The fitness function is the only evaluation standard of the genetic algorithm. For example, the fitness function = 0.4 x atomic cell pressure error + 0.3 x neon / carbon monoxide ratio error + 0.2 x temperature fluctuation range + 0.1 x alkali metal charge error. The genetic algorithm takes the fitness function as the evaluation standard to calculate the optimal PID parameters and proportional control parameters (such as the proportional coefficient of the neon valve), thereby adapting to the coupling requirements under different environmental conditions.

[0057] Further, the prediction model deployed in the FPGA generates temperature change prediction value and alkali metal charge prediction value for the current charging cycle based on historical charging cycle data; these prediction values are input to the PID controller and superimposed with real-time feedback signals from the distributed optical fiber temperature sensor and the mass sensor to adjust the opening of the neon / carbon monoxide regulating valve in advance. For example, when the prediction model identifies that the temperature at the end of the atomic cell will rise by 0.5°C due to excessive charging of alkali metal, the feedforward compensation module will increase the buffer gas flow in advance to offset the thermal expansion effect. After each charging is completed, the charging system automatically records the deviation between the actual gas leakage and the estimated gas leakage based on the temperature change prediction value and the alkali metal charge prediction value, and updates the LSTM network model parameters in an incremental learning manner to gradually improve the prediction accuracy.

[0058] By the embodiment of the present application, the predicted value of temperature change and the predicted value of the alkali metal filling amount are received by the feedforward compensation module, the predicted values are input into the PID controller, potential gas leakage is compensated for, and the hysteresis of the gas filling system is reduced; the dynamic weight distribution module adjusts the integral term weight in the PID parameter according to the real-time temperature value, so that the integral term weight is increased under high temperature conditions to suppress the volatilization accumulation effect of the alkali metal vapor; the parameter self-tuning module adopts a genetic algorithm, takes the weighted sum of the atomic cell pressure error, the alkali metal filling amount error, the neon / carbon monoxide ratio error and the temperature fluctuation range as the fitness function, and loads the optimal PID parameter and proportional control parameter calculated by an offline method into the PID control loop and the proportional control loop to adapt to the coupling requirements under different environmental conditions. That is, the response hysteresis of the gas filling system is reduced by feedforward compensation, the stability of the atomic cell filling is improved by dynamic weight distribution, the flexibility of the atomic cell filling is improved by parameter self-tuning, and the accuracy of the atomic cell filling is further improved.

[0059] As an optional solution, the gas filling system for filling the alkali metal vapor and the buffer gas into the atomic cell further comprises:

[0060] S1, the multivariable disturbance test module separates the leakage effect and the temperature effect by step response experiment, and establishes a dynamic relationship matrix between the leakage effect and the temperature effect;

[0061] S2, the dynamic relationship matrix is embedded into the feedforward compensation module, in the case that the distributed optical fiber temperature sensor detects that the temperature of the atomic cell changes, the gas filling system corrects the gas leakage compensation amount according to the dynamic relationship matrix to avoid the gas leakage error caused by thermal expansion;

[0062] S3, a closed-loop temperature control channel is formed by the TEC refrigeration module containing the annular TEC refrigeration sheet and the distributed optical fiber temperature sensor, the FPGA adjusts the TEC driving current in real time according to the feedback of the distributed optical fiber temperature sensor, so that the temperature of the atomic cell is stabilized within ±0.01K; S4, after the gas filling is completed, the atomic cell is sealed by the atomic cell sealing device.

[0063] Optionally, in the embodiment of the present application, the TEC refers to a thermoelectric cooler, the cold face of the annular TEC refrigeration sheet is attached to the outer wall of the atomic cell, and the hot face is connected with the heat sink and the fan through the heat conduction interface material.

[0064] To further illustrate, step response experiments were conducted under controlled conditions. A step-change in neon gas flow was introduced at a constant temperature to stimulate a pure leakage effect. Subsequently, a step-change in temperature was applied under sealed conditions to extract the pure temperature effect. By comparing the pressure response curves of the atomic gas chamber from the two sets of step response experiments, the least squares method was used to fit a dynamic relationship matrix between the leakage effect and the temperature effect. Its off-diagonal elements represent the impact of temperature fluctuations on the leakage rate (e.g., for every 1°C increase in the atomic gas chamber temperature, the leak detection system will mistakenly interpret the pressure change caused by this 1°C as a 0.2% gas leak). This dynamic relationship matrix is ​​embedded in a feedforward compensation module. When the distributed fiber optic temperature sensor detects a sudden temperature change at the end of the atomic gas chamber, the inflation system automatically adjusts the leakage compensation according to the dynamic relationship matrix to avoid misidentifying pressure changes caused by thermal expansion as actual leaks.

[0065] Through the embodiments of the present application, the closed-loop temperature control channel composed of the TEC refrigeration module and the distributed optical fiber temperature sensor keeps the temperature of the atomic gas chamber stable, and the multivariable disturbance test makes the feedforward compensation more accurate, further improving the pressure stability of the atomic gas chamber.

[0066] Optionally, in an embodiment of the present application, the FPGA further includes:

[0067] After the above-mentioned inflation system is started, the temperature-based flow ratio adjustment mode is preferentially enabled; when the temperature of the above-mentioned atomic gas chamber is stable in the range of ±5°C for more than 10 minutes, it automatically switches to the flow ratio adjustment mode based on the above-mentioned alkali metal filling amount; when the temperature fluctuation of the above-mentioned atomic gas chamber exceeds the set threshold, it immediately switches back to the temperature-based flow ratio adjustment mode.

[0068] Optionally, in an embodiment of the present application, a temperature-based flow ratio adjustment mode is preferentially executed. When the temperature of the atomic gas chamber is less than 250°C, the flow ratio of neon and carbon monoxide is adjusted to 8:2; when the temperature of the atomic gas chamber is greater than or equal to 250°C and less than 300°C, the flow ratio of neon and carbon monoxide is adjusted to 6:4; when the temperature of the atomic gas chamber is greater than or equal to 300°C and less than 350°C, the flow ratio of neon and carbon monoxide is adjusted to 5:5; when the temperature of the atomic gas chamber is greater than or equal to 350°C and less than 460°C, the flow ratio of neon and carbon monoxide is adjusted to 4:6; when the temperature of the atomic gas chamber is greater than or equal to 460°C, the inflation operation is stopped. When the temperature of the atomic gas chamber is stable in the range of ±5°C for more than 10 minutes, the flow ratio adjustment mode based on the alkali metal charging amount is as follows: when the alkali metal charging amount is between 3mg and 10mg, the flow ratio of neon and carbon monoxide is adjusted to 8:2; when the alkali metal charging amount is between 10mg and 20mg, the flow ratio of neon and carbon monoxide is adjusted to 6:4.

[0069] Through the embodiment of the present application, after the inflation system is started, the temperature-based flow ratio regulation mode is preferentially enabled; in the case that the temperature of the atomic chamber is stabilized in the interval of ±5℃ and exceeds 10 minutes, the flow ratio regulation mode based on the alkali metal filling amount is automatically switched to; when the temperature fluctuation of the atomic chamber exceeds the set threshold, the flow ratio regulation mode based on the temperature is immediately switched back. That is, the accuracy of the atomic chamber filling is improved through the two modes of the temperature-based flow ratio regulation mode and the flow ratio regulation mode based on the alkali metal filling amount.

[0070] As an optional solution, the above-mentioned proportional controller comprises:

[0071] The flow ratio of the neon gas and the carbon monoxide is compared in real time; in the case that the flow ratio of the neon gas and the carbon monoxide reaches the set value, a closing signal is output to the alkali metal quantitative dispenser.

[0072] Optionally, in the embodiment of the present application, the micro-flow sensor monitors the instantaneous flow of the neon gas and the carbon monoxide in real time, the proportional controller calculates the monitored instantaneous flow, and when the flow calculated for five consecutive times reaches the set value, a closing signal is output to the alkali metal quantitative dispenser.

[0073] Through the embodiment of the present application, the flow ratio of the neon gas and the carbon monoxide is compared in real time; in the case that the flow ratio of the neon gas and the carbon monoxide reaches the set value, a closing signal is output to the alkali metal quantitative dispenser, thereby avoiding manual closing and shortening the filling period of the atomic chamber.

[0074] As an optional solution, the feedforward compensation module continuously updates the predicted value of the temperature change and the predicted value of the alkali metal filling amount in a sliding window manner, and superimposes the updated predicted values on the output end of the PID controller in real time.

[0075] Optionally, in the embodiment of the present application, the prediction model deployed in the FPGA outputs the predicted value of the temperature change and the predicted value of the alkali metal filling amount for the next 5 seconds every 10 seconds, and these predicted values are superimposed on the output end of the PID controller to correct the opening degree of the neon / carbon monoxide regulating valve. For example, when the prediction shows that the alkali metal filling will cause the temperature at the end of the atomic chamber to rise by 0.8℃, the feedforward compensation module will increase the buffer gas flow in advance according to the dynamic relationship matrix to offset the pressure fluctuation caused by thermal expansion.

[0076] Through the embodiment of the present application, the feedforward compensation module continuously updates the predicted value of the temperature change and the predicted value of the alkali metal filling amount in a sliding window manner, and superimposes the updated predicted values on the output end of the PID controller in real time, thereby further improving the accuracy of the atomic chamber filling.

[0077] As an optional solution, the multivariable disturbance testing module, when performing the above-mentioned step response experiment, applies step input to the temperature channel and the pressure channel simultaneously through the programmable disturbance generator, and quantifies the coupling coefficient of the leakage effect and the temperature effect based on the spectral feedback signal, so as to update the dynamic relationship matrix, and make the feedforward compensation module always use the latest dynamic relationship matrix data to offset the pseudo-leakage signal caused by thermal expansion.

[0078] Optionally, in the embodiment of the present application, the programmable disturbance generator is a signal source or an actuator which can be programmed by software or hardware in real time to set the amplitude, frequency, waveform and start and end time, and is used to charge a known disturbance as needed, repeatedly and programmably in a control system or an experimental device. The coupling coefficient is used to measure how much gas leakage is equivalent to a temperature change of 1K. The programmable disturbance generator applies step input to the temperature channel, such as pulling the atomic cell temperature from 70℃ to 75℃ in 0.5s. The response of the gas filling system to temperature disturbance and pressure disturbance is read out in real time through the spectral feedback signal.

[0079] Through the embodiment of the present application, the multivariable disturbance testing module, when performing the above-mentioned step response experiment, applies step input to the temperature channel and the pressure channel simultaneously through the programmable disturbance generator, and quantifies the coupling coefficient of the leakage effect and the temperature effect based on the spectral feedback signal, so as to update the dynamic relationship matrix, and further improve the control accuracy of the atomic cell during the filling process.

[0080] As an optional solution, the atomic cell sealing device executes the following sealing procedure after receiving the sealing trigger signal sent by the FPGA:

[0081] The alkali metal quantitative dispenser is closed; the neon / carbon monoxide regulating valve is closed synchronously; and the laser sealing or mechanical sealing mechanism is driven by the preset time sequence to complete the sealing of the atomic cell.

[0082] Optionally, in the embodiment of the present application, during the sealing of the atomic cell, the preset time sequence refers to a computer program or an electronic instruction set which is pre-set and accurately controls the trigger time and duration of each operation step. The metal atomic cell triggers laser sealing, the glass atomic cell calls mechanical sealing, the laser sealing is detected by real-time X-ray imaging, and the mechanical sealing adopts helium mass spectrometry leak detection.

[0083] Through the embodiment of the present application, the alkali metal quantitative dispenser is closed; the neon / carbon monoxide regulating valve is closed synchronously; and the laser sealing or mechanical sealing mechanism is driven by the preset time sequence to complete the sealing of the atomic cell. That is, the preset time sequence makes the laser sealing or mechanical sealing action start when the filling target of the atomic cell just reaches the standard, avoiding the delay of time and the overshoot of the atomic cell, and further improving the filling accuracy of the atomic cell.

[0084] Optionally, in the case of ultra-low temperature demand scenarios, the above-mentioned inflation system for inflating alkali metal vapor and buffer gas into the above-mentioned atomic chamber further comprises:

[0085] The above-mentioned atomic chamber temperature is reduced to 10K by a mechanical refrigeration device; after the above-mentioned mechanical refrigeration, the above-mentioned atomic chamber temperature is finely regulated by the above-mentioned TEC refrigeration module with an accuracy of ±0.01K; the above-mentioned FPGA coordinates the start-stop and power distribution of the above-mentioned mechanical refrigeration device and the above-mentioned TEC refrigeration module, realizing seamless switching and joint control of two-stage refrigeration.

[0086] Optionally, in the embodiment of the present application, the FPGA first starts the mechanical refrigeration device, and rapidly reduces the atomic chamber from room temperature 300K to 10K within 8min; then the FPGA closes the mechanical refrigeration and cuts into the TEC refrigeration module according to the real-time reading of the optical fiber temperature sensor at the end of the atomic chamber, and locks the temperature difference of the atomic chamber temperature within ±0.01K.

[0087] By the embodiment of the present application, the above-mentioned atomic chamber temperature is reduced to 10K by a mechanical refrigeration device; after the above-mentioned mechanical refrigeration, the above-mentioned atomic chamber temperature is finely regulated by the above-mentioned TEC refrigeration module with an accuracy of ±0.01K; the above-mentioned FPGA coordinates the start-stop and power distribution of the above-mentioned mechanical refrigeration device and the above-mentioned TEC refrigeration module, realizing seamless switching and joint control of two-stage refrigeration. That is, the two-stage refrigeration mode realizes extremely fast cooling and improves the control accuracy of the atomic chamber temperature.

[0088] Optionally, as an optional implementation, the above-mentioned method is explained and described as a whole by the following steps:

[0089] In the beginning stage of the filling of the atomic clock atomic cell, the atomic cell is sequentially cleaned by using an acetone solution, a hydrochloric acid solution and supercritical CO2, an image of the inner wall of the atomic cell is collected by the optical fiber endoscope every 3 seconds during the cleaning by supercritical CO2, and the image of the inner wall of the atomic cell is input into a convolutional neural network for residual distribution feature recognition, and according to the recognition results such as the type of the residual and the distribution of the residual, the cleaning parameters (such as pressure, temperature and cleaning time) of the supercritical CO2 are dynamically adjusted to completely remove the nanoscale silicate and metal oxide residues on the inner wall of the atomic cell. Subsequently, the cleaned atomic cell is placed in an oven under the irradiation of 254 nm ultraviolet light to form a passivation layer, and in the baking process, a pulsed vacuum-inert gas exchange environment is introduced, wherein the purity of the inert gas (such as neon or argon) needs to be ≥99.999% to avoid impurity pollution of the atomic cell. In order to form a uniform and thickness-controllable passivation layer on the inner wall of the atomic cell, the average power density E of the 254 nm ultraviolet light on the crystal surface, the continuous irradiation time t in a single pulse and the number N of pulsed vacuum-inert gas exchange cycles need to satisfy the following relationship: D = E × t × N^0.5 ≤ 150 mJ·cm -2 , the value of E is generally 0.5-2.0 mW·cm -2 , the value of N is generally 3-8 times, and D represents the cumulative energy measurement, which is generally 100-150 mJ·cm -2 , and when D exceeds 150 mJ·cm -2 , the passivation layer thickens and cracks appear, and when D is less than 80 mJ·cm -2 , the coverage is incomplete. The baked atomic cell is cooled for subsequent gas filling operation.

[0090] During the inflation phase, the FPGA control module is activated and the optimal PID parameters and proportional control parameters calculated offline are loaded. These optimal PID parameters and proportional control parameters are calculated using a genetic algorithm in the parameter self-tuning module, with the weighted sum of the atomic chamber pressure error, alkali metal charge error, neon / carbon monoxide ratio error, and temperature fluctuation as the fitness function. Various sensors, including distributed fiber optic temperature sensors, pressure sensors, mass sensors, and microflow sensors, are initialized to complete calibration. The TEC refrigeration module starts preheating and enters standby mode. The FPGA sends a start signal to the alkali metal quantitative dispenser, activating its built-in heating device, heating the solid alkali metal to above its melting point to form alkali metal vapor. The alkali metal quantitative dispenser then fills the atomic chamber with alkali metal vapor through a connecting pipeline. The mass sensor outputs a real-time alkali metal charge signal (accuracy ±0.1mg), which is synchronously transmitted to the FPGA. The FPGA records the charge data in real time as a basis for subsequent control. The neon and carbon monoxide supply systems are activated, and a microflow sensor monitors the instantaneous flow rates of both gases in real time. A PID control loop begins operating, with a pressure sensor detecting the pressure in the atomic chamber. The PID controller adjusts the gas flow rate via the inflation valve, forming a closed-loop feedback control loop for the atomic chamber pressure. The proportional control loop begins operating, with the FPGA prioritizing temperature-based flow proportional regulation (primarily based on data from a distributed fiber-optic temperature sensor). Once the atomic chamber temperature stabilizes within the ±5°C range for 10 minutes, it automatically switches to alkali metal charge proportional regulation (primarily based on data from a mass sensor). If the temperature fluctuates beyond a set threshold, it immediately switches back to temperature-based flow proportional regulation. The proportional controller compares the neon and carbon monoxide flow ratio in real time and, when it reaches the set value, outputs a signal to shut down the alkali metal dispenser. A predictive model in the FPGA (trained using historical inflation data and incrementally updating parameters after each inflation) generates predicted values ​​for temperature change and alkali metal charge. These values ​​are then added to the PID controller output via a feedforward compensation module to compensate for potential gas leaks and reduce system hysteresis. Adjust the integral gain Ki in the PID parameters based on the real-time temperature value (increasing Ki at high temperatures suppresses the cumulative effect of alkali metal volatilization). Through step response experiments, step inputs are applied to the temperature and pressure channels. The coupling coefficient between the leakage effect and the temperature effect is quantified using the spectral feedback signal as a reference, and the dynamic relationship matrix is ​​updated. When the alkali metal charge, neon gas, and carbon monoxide charge in the atomic gas chamber all reach the set values ​​and stabilize, the charging is considered complete.

[0091] After the inflation is completed, all actuators are reset, the sensor stops data collection, the alkali metal quantitative dispenser and the neon gas / carbon monoxide adjusting valve are synchronously closed by the inflation system, the FPGA sends a sealing trigger signal, and the atomic cell sealing device starts laser or mechanical sealing (1080nm laser scanning welding is used for a metal atomic cell, and 80N pressure fitting ultrasonic vibration sealing is used for a glass atomic cell) according to a preset time sequence, and the sealing of the atomic cell is completed. In an ultralow-temperature scenario, two-stage refrigeration is realized by the FPGA in cooperation with a mechanical refrigeration module and a TEC refrigeration module. The mechanical refrigeration module reduces the temperature of the atomic cell to 10K, and the TEC refrigeration module ensures that the temperature fluctuation in the whole process is controlled within ±0.01K. The FPGA records the data of the whole inflation process, which is used to update the parameters of the prediction model.

[0092] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited to the action sequence described, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0093] The sequence numbers of the embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0094] In the above embodiments of the present application, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0095] In the several embodiments provided by the present application, the above-described device embodiments are only schematic, and the division of the above-mentioned units is only a logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed components can be indirect coupling or communication connection through some interface, and can be electrical or other forms.

[0096] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0097] In addition, each of the functional units in the various embodiments of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0098] The above merely describes the preferred embodiments of the present application, and it should be pointed out that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A method for filling an atomic gas chamber of an atomic clock, characterized in that: include: The atomic gas chamber was cleaned with acetone solution, hydrochloric acid solution and supercritical CO2 in sequence; The cleaned atomic gas chamber is placed in an oven equipped with ultraviolet light irradiation for baking, and a pulsed vacuum-inert gas exchange environment is introduced into the atomic gas chamber through the interface of the atomic gas chamber to form a passivation layer on the inner wall of the atomic gas chamber, wherein the wavelength of the ultraviolet light irradiation is 254 mm, and the pulsed vacuum-inert gas exchange environment refers to cyclically evacuating the atomic gas chamber and filling it with inert gas; placing the cooled atomic gas chamber into a gas filling system, wherein the gas filling system is used to fill the atomic gas chamber with alkali metal vapor and a buffer gas, wherein the buffer gas is neon and carbon monoxide; during the filling process of the atomic gas chamber, adjusting the ratio of the neon and carbon monoxide filled into the atomic gas chamber based on the amount of alkali metal filled in the atomic gas chamber and the temperature of the atomic gas chamber; and sealing the atomic gas chamber when the amount of the alkali metal vapor and the buffer gas in the atomic gas chamber reaches a set value; The atomic gas chamber is taken out from the inflation system to complete the filling.

2. The method according to claim 1, wherein The supercritical CO2 cleaning process includes: The image of the inner wall of the atomic gas chamber is collected in real time by a fiber optic endoscope, with an acquisition interval of less than or equal to 3 seconds; Using a convolutional neural network to identify the distribution characteristics of residues on the collected image of the inner wall of the atomic gas chamber; Dynamically adjusting the cleaning parameters of the supercritical CO2 according to the identification results; Cleaning is performed based on the cleaning parameters to remove nano-sized silicate and metal oxide residues on the inner wall of the atomic gas chamber, wherein the pressure of supercritical CO2 in the cleaning parameters is adjusted within the range of 10-30 MPa, the temperature is adjusted within the range of 40-60°C, and the cleaning time is adjusted within 5-15 minutes.

3. The method according to claim 1, wherein The gas filling system is used to fill the atomic gas chamber with alkali metal vapor and buffer gas, and includes: The alkali metal vapor is filled into the atomic gas chamber by an alkali metal quantitative distributor through a connecting pipeline, wherein the connecting pipeline is made of a titanium alloy material that has undergone inertization treatment. Before being filled into the atomic gas chamber, the alkali metal vapor passes through a molecular sieve filter module to adsorb and remove impurity gases carried by the alkali metal vapor; The alkali metal charging amount signal is output in real time by a mass sensor coupled to the alkali metal quantitative dispenser, wherein the accuracy of the mass sensor is ±0.1 mg, and the alkali metal quantitative dispenser is provided with a heating device, wherein the heating device is used to heat the solid alkali metal to above the melting point to form the alkali metal vapor; The instantaneous flow rates of the neon gas and the carbon monoxide are monitored in real time by a micro flow sensor; Detecting the temperature of the atomic gas chamber in real time by using a distributed optical fiber temperature sensor arranged inside the atomic gas chamber; Detecting the internal pressure of the atomic gas chamber in real time by a pressure sensor; Outputting multiple control signals in parallel through a high-speed digital I / O module including an FPGA, and performing priority management on the control signals; Regulating the pressure of the atomic gas chamber through a PID control loop consisting of the pressure sensor, PID controller, inflation valve and gas chamber pressure feedback channel; The flow ratio of the neon gas and the carbon monoxide is adjusted by a proportional control loop composed of the distributed optical fiber temperature sensor, the mass sensor, a proportional controller, a neon gas / carbon monoxide regulating valve and a spectral feedback channel.

4. The method according to claim 1, wherein The gas filling system for filling the atomic gas chamber with alkali metal vapor and buffer gas also includes: receiving, through a feedforward compensation module, a predicted value of temperature change and a predicted value of the alkali metal charge, inputting the predicted values ​​into the PID controller to compensate for potential gas leakage and thereby reduce the hysteresis of the inflation system, wherein the predicted values ​​are generated by a prediction model deployed in the FPGA, the prediction model being trained using historical data of inflation cycles and modeled using an LSTM network, and the parameters of the prediction model being incrementally updated after each inflation is completed; The dynamic weight allocation module adjusts the integral term weight in the PID parameter according to the real-time temperature value, so that the integral term weight is increased under high temperature conditions to suppress the volatilization cumulative effect of the alkali metal vapor, wherein the integral term weight refers to the integral gain Ki; The parameter self-tuning module adopts a genetic algorithm, with the weighted sum of the atomic gas chamber pressure error, the alkali metal filling amount error, the neon / carbon monoxide ratio error and the temperature fluctuation range as the fitness function, and loads the optimal PID parameters and proportional control parameters calculated in an offline manner into the PID control loop and the proportional control loop to adapt to the coupling requirements under different environmental conditions.

5. The method according to claim 1, wherein The gas filling system for filling the atomic gas chamber with alkali metal vapor and buffer gas also includes: The multivariable disturbance test module separates leakage effect and temperature effect through step response experiment and establishes a dynamic relationship matrix between the leakage effect and the temperature effect; The dynamic relationship matrix is ​​embedded in the feedforward compensation module. When the distributed optical fiber temperature sensor detects a change in the temperature of the atomic gas chamber, the inflation system corrects the gas leakage compensation amount according to the dynamic relationship matrix to avoid gas leakage errors caused by thermal expansion. A closed-loop temperature control channel is formed by a TEC cooling module including an annular TEC cooling plate and the distributed optical fiber temperature sensor, and the FPGA adjusts the TEC drive current in real time according to feedback from the distributed optical fiber temperature sensor to stabilize the temperature of the atomic gas chamber within ±0.01K; After the inflation is completed, the atomic gas chamber is sealed by a gas chamber sealing device.

6. The method according to claim 3, characterized in that The FPGA further comprises: After the inflation system is started, the temperature-based flow ratio regulation mode is enabled; when the temperature of the atomic gas chamber is stable in the range of ±5°C for more than 10 minutes, it is automatically switched to the flow ratio regulation mode based on the alkali metal filling amount; when the temperature fluctuation of the atomic gas chamber exceeds the set threshold, it is immediately switched back to the temperature-based flow ratio regulation mode.

7. The method according to claim 3, characterized in that The proportional controller comprises: The flow ratio of the neon gas and the carbon monoxide is compared in real time; when the flow rates of the neon gas and the carbon monoxide reach the set value, a closing signal is output to the alkali metal quantitative distributor.

8. The method according to claim 4, characterized in that The feedforward compensation module continuously updates the predicted value of the temperature change and the predicted value of the alkali metal charging amount in a sliding window manner, and superimposes the updated predicted values ​​on the output end of the PID controller in real time.

9. The method according to claim 5, characterized in that When performing the step response experiment, the multivariable disturbance test module applies step input to the temperature channel and the pressure channel simultaneously through a programmable disturbance generator, and quantifies the coupling coefficient between the leakage effect and the temperature effect based on the spectral feedback signal, thereby updating the dynamic relationship matrix.

10. The method according to claim 5, characterized in that After receiving the sealing trigger signal from the FPGA, the air chamber sealing device executes the following sealing procedure: The alkali metal quantitative distributor is closed; the neon / carbon monoxide regulating valve is closed synchronously; and the laser sealing or mechanical pressure sealing mechanism is driven in a preset time sequence to complete the sealing of the atomic gas chamber.

11. The method according to any one of claims 3 to 10, characterized in that In the case of ultra-low temperature demand scenarios, the gas filling system for filling the atomic gas chamber with alkali metal vapor and buffer gas further includes: A mechanical refrigeration device is used to reduce the temperature of the atomic gas chamber to 10K. After the mechanical refrigeration, the TEC refrigeration module finely controls the temperature of the atomic gas chamber with an accuracy of ±0.01K. The FPGA coordinates the start and stop and power distribution of the mechanical refrigeration device and the TEC refrigeration module to achieve seamless switching and joint control of two-stage refrigeration.

Images