Controllable filling method and system for helium mixture for dilution refrigerator

Abstract

The application discloses a controllable filling method and system for a helium mixture of a dilution refrigerator, and relates to the technical field of image processing and intelligent control, and comprises the following steps: collecting an image of an outer wall of a target thin tube section to generate a frost band enhanced profile; extracting a half-height width and an edge gradient to calculate a frost band shape anomaly ratio, and generating a throttling influence coefficient according to the frost band shape anomaly ratio; collecting pressure and temperature of a metering cavity to calculate a metering release molar quantity, and correcting the metering release molar quantity by using the throttling influence coefficient to obtain an effective inlet molar quantity; and generating a valve control instruction based on the effective inlet molar quantity. The application can improve the metering accuracy and safety of the filling process.

Description

Technical Field

[0001] This invention relates to the field of image processing and intelligent control technology, and in particular to a controllable method and system for adding helium mixtures for dilution refrigerators. Background Technology

[0002] The operation of a dilution refrigerator relies on a helium mixture forming a stable cycle within the loop and establishing the required cryogenic environment. In actual engineering, during initial installation, resetting after major maintenance, recharging the mixture after recovery, and refilling after replacing valve assemblies or pipelines, it is usually necessary to quantitatively release the mixture in the metering chamber and send it into the dilution refrigerator loop through the filling valve to ensure that the working fluid storage in the loop reaches the target level and meets the requirements for subsequent condensation and circulation establishment. Due to the presence of capillary flow segments, narrow pipe sections, and cryogenic components in the loop, there is a risk of localized narrowing, distributed throttling, or flow blockage caused by contaminants. This can cause the release amount displayed on the metering side to deviate from the actual effective inflow into the loop, further leading to problems such as undercharging causing difficulty in establishing the loop or overcharging causing increased pressure risks.

[0003] Current technologies generally employ calculation methods based on the pressure and temperature of the metering chamber to indirectly estimate the refill amount. This involves converting the pressure drop within the metering chamber into the released molar quantity according to the gas state equation, and using this as the criterion for valve opening / closing and refill termination. This method only focuses on the release data at the metering end, ignoring the weakening effect of flow resistance changes in the downstream loop on the actual flow rate. When there is throttling or blockage in the loop, a significant systematic deviation will appear between the amount of mixed gas released as displayed in the metering chamber and the actual effective amount of gas entering the dilution loop and participating in the circulation, leading to severe distortion of the refill amount. This deviation can easily cause undercharging, preventing the refrigerator from starting, or overcharging, resulting in excessive loop pressure and safety hazards. Furthermore, the throttling expansion of the gas inside the pipe usually causes vivid frost or frost ring morphology changes on the outer wall of the thin pipe, but traditional refill strategies fail to utilize this visual information to correct the flow calculation, and cannot promptly detect and compensate for refill errors caused by abnormal flow resistance. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies that rely on metering cavity parameters to calculate the release amount, which cannot detect changes in throttling or flow resistance within the circuit, resulting in a mapping distortion between the metered release amount and the actual effective inlet amount. Furthermore, the invention addresses the shortcomings of neglecting the visual indication of the operating conditions by the frost pattern on the outer wall, leading to inaccurate filling control. Therefore, this invention proposes a controllable filling method and system for diluting helium mixtures for refrigeration machines.

[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution:

[0006] A controlled method for adding a helium mixture for a dilution refrigerator includes:

[0007] S1. Acquire an image of the outer wall of the target thin tube section of the dilution refrigeration unit, and process the image to generate a frost-enhanced profile along the tube axis.

[0008] S2. Extract the full width at half maximum (FWHM) and edge gradient from the frost-enhanced profile, and calculate the anomalous ratio of the frost-band morphology based on the FWHM and edge gradient.

[0009] S3. Generate the throttling influence coefficient based on the anomalous ratio of the frost zone morphology;

[0010] S4. Collect the real-time pressure and temperature in the metering chamber, calculate the metering release molar amount, and correct the metering release molar amount through the throttling influence coefficient to obtain the effective inlet molar amount.

[0011] S5. Generate valve control commands based on the effective input molar quantity.

[0012] Preferably, generating a frost-reinforced profile along the tube axis includes:

[0013] Select the target region containing the target thin tube segment from the acquired image and perform grayscale processing on the target region;

[0014] The average value of the horizontal pixels corresponding to each axial coordinate is calculated along the tube axis direction of the target area to obtain the initial axial average profile. The axial coordinates are calibrated by the known geometric scale of the target thin tube segment to convert the axial pixel coordinates into axial length coordinates.

[0015] The median of the initial axial average profile is used as the background baseline. Based on the difference between the initial axial average profile and the background baseline, the frost-enhanced profile is obtained.

[0016] Preferably, the full width at half maximum (FWHM) and edge gradient are extracted from the frost-enhanced profile, and the anomalous ratio of the frost-band morphology is calculated based on the FWHM and edge gradient, including:

[0017] Locate the main peak on the frost-enhanced profile and determine the half-height value corresponding to half the peak height of the main peak.

[0018] Find the intersection of the frost-enhanced profile and the half-height value on both sides of the main peak, and take the axial distance between the two intersection points as the half-height width;

[0019] Calculate the rate of change of profile values ​​near the intersection point to obtain the edge gradient;

[0020] The anomalous ratio of the frost band morphology is obtained by using the ratio of half-width at half-maximum to edge gradient.

[0021] Preferably, the throttling influence coefficient is generated based on the anomalous ratio of the frost zone morphology, including:

[0022] During the reference time period before the valve is opened, the minimum value of the anomalous ratio of the frost band pattern is used as the reference anomalous ratio, and the median absolute deviation of the anomalous ratio of the frost band pattern is calculated as the scale term.

[0023] During the injection process, the difference between the real-time acquired frost band morphology anomaly ratio and the baseline anomaly ratio is calculated, and the difference is divided by the scale term to obtain the normalized anomaly quantity.

[0024] The throttling effect coefficient is obtained by mapping the normalized outliers using a monotonically decreasing function.

[0025] Preferably, the metered release molar amount is corrected by a throttling effect coefficient to obtain the effective inlet molar amount, including:

[0026] Based on the known volume, real-time pressure, and real-time temperature within the metering chamber, the instantaneous molar quantity of the metering chamber at the initial and current moments is calculated using the gas state equation.

[0027] The difference between the instantaneous molar quantity at the start time and the instantaneous molar quantity at the current time is used as the measured molar quantity released.

[0028] The effective inlet molar amount is obtained by multiplying the metered release molar amount by the throttling effect coefficient.

[0029] Preferably, the valve control command is generated based on the effective input molar quantity, including:

[0030] Calculate the difference between the preset injection target value and the current effective inlet molar quantity to obtain the remaining effective quantity;

[0031] The release rate is obtained by calculating the rate of change of the amount of molar released over time.

[0032] The control increment is calculated based on the ratio of the remaining effective amount to the release rate;

[0033] The control increment is superimposed on the valve control command from the previous moment, and the superposition result is limited to the range that the valve can control, to obtain the current valve control command.

[0034] Preferably, the method further includes a safety termination step: real-time monitoring of the circuit pressure of the dilution refrigeration unit circuit; and immediately closing the filling valve when the effective inlet molar amount is greater than or equal to the preset filling target value, or when the circuit pressure is greater than or equal to the preset rated pressure boundary.

[0035] To address the aforementioned problems, the present invention also provides a controllable dosing system for a helium mixture used in a dilution refrigerator, the system comprising:

[0036] The image acquisition and processing module is used to acquire images of the outer wall of the target thin tube segment of the dilution refrigeration machine, and process the images to generate a frost-enhanced profile along the tube axis.

[0037] The morphological anomaly ratio module is used to extract the full width at half maximum (FWHM) and edge gradient on the frost-enhanced profile, and to calculate the anomaly ratio of the frost morphology based on the FWHM and edge gradient.

[0038] The throttling coefficient module is used to generate a throttling influence coefficient based on the anomalous ratio of the frost zone morphology;

[0039] The correction module is used to collect the real-time pressure and temperature in the metering chamber, calculate the metering release molar amount, and correct the metering release molar amount through the throttling influence coefficient to obtain the effective inlet molar amount.

[0040] The valve control module is used to generate valve control commands based on the effective input molar quantity.

[0041] Compared with the prior art, the beneficial effects of the present invention are:

[0042] 1. This invention constructs a throttling influence coefficient characterizing the internal flow resistance state of the pipeline by acquiring images of the outer wall of the target thin tube section and extracting the anomalous ratio of the frost pattern. This coefficient is used to correct the metered release molar amount calculated based on pressure and temperature in real time, thereby eliminating the flow attenuation caused by local diameter reduction, distributed throttling, or trace contamination inside the loop. It solves the systematic deviation between the release amount displayed on the metering chamber side and the actual effective inflow amount into the dilution loop, ensuring that the filling process can accurately reflect the actual working fluid storage changes inside the loop. It effectively prevents the risk of undercharging and failure to start or overcharging and pressure exceeding the limit due to mapping distortion, and significantly improves the metering accuracy and safety reliability of the dilution refrigeration mixture filling process.

[0043] 2. This invention establishes a quantitative mapping mechanism between the visual perception of the outer wall frost band and the internal flow resistance conditions. By analyzing the half-width at half-maximum and edge gradient of the frost band enhancement profile, it can sensitively identify the abnormal trend of the frost band being wide and having blunt edges, giving the filling system a non-contact perception capability for invisible flow resistance changes. Based on the visually corrected effective inlet molar quantity, valve control commands are generated, realizing the adaptive matching of filling rate and remaining effective quantity. This overcomes the limitation of simply relying on pressure parameters to identify distributed throttling, ensuring the consistency and stability of filling control under complex flow resistance conditions. Attached Figure Description

[0044] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0045] Figure 1 This is a schematic flowchart of a controlled filling method for a helium mixture used in a dilution refrigerator according to the present invention.

[0046] Figure 2 This is a functional block diagram of a controllable dosing system for a helium mixture used in a dilution refrigerator according to the present invention.

[0047] Figure 3 This is a schematic diagram of the half-width at half-maximum and edge gradient of the frost-reinforced cross section of the present invention. Detailed Implementation

[0048] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0049] Example: This example provides a controlled method for adding helium mixtures to a dilution refrigerator. See [link to example]. Figure 1 Specifically, including:

[0050] S1. Acquire an image of the outer wall of the target thin tube section of the dilution refrigeration unit, and process the image to generate a frost-enhanced profile along the tube axis.

[0051] In an embodiment of the present invention, an image of the outer wall of the target thin tube segment of the dilution refrigerator is acquired, and the image is processed to generate a frost-enhanced profile along the tube axis, including:

[0052] Acquire images of the outer wall of the target thin tube segment of the dilution chiller;

[0053] Select the target region containing the target thin tube segment from the acquired image and perform grayscale processing on the target region;

[0054] Specifically, acquiring an image of the outer wall of the target capillary tube segment of the dilution chiller refers to using a camera to obtain the intensity distribution of reflection and scattering on the outer surface of the capillary tube segment under visible light illumination, in order to record the difference in brightness between the frosted and non-frosted areas on the outer wall. The target capillary tube segment refers to the outer surface of the slender tube that is most prone to throttling and cooling during the filling process and forms a frost band on its surface. The target area refers to the image sub-region that covers the outer wall of the capillary tube segment in the acquired image to reduce the interference of irrelevant background on the measurement.

[0055] In detail, during the preparation stage of controlled helium mixture filling in the dilution refrigeration unit, the target thin tube segment most prone to developing an outer wall frost zone near the filling circuit is selected as the imaging object. An industrial camera or high-definition camera is fixedly installed to the side of the outer wall of the target thin tube segment, with the lens optical axis approximately perpendicular to the axis of the thin tube segment to reduce the influence of perspective compression on the axial dimension measurement of the frost zone. Preferably, the working distance from the camera to the outer wall of the thin tube is 20 to 60 centimeters to balance field of view coverage and spatial resolution. Preferably, an imaging mode with a resolution of not less than 1920 x 1080 is used to ensure that the transition zone at the edge of the frost zone has sufficient pixel width in the image. Preferably, the frame rate is 10 to 30 frames per second to cover the critical process of frost zone formation and change before and after valve opening without generating excessive storage burden. After the camera is installed, the outer wall of the target thin tube segment is illuminated. Preferably, a constant-brightness light source with a color temperature of 4000 to 6000 K and stable illuminance is used to obliquely illuminate the target thin tube segment from a different direction than the camera. The outer wall is designed to reduce the interference of specular reflection highlights on the brightness of the frost band and improve the brightness contrast caused by frost scattering. At the same time, a calibration object with known geometric dimensions is fixed next to the target thin tube segment, or the outer diameter of the thin tube is used as a scale reference so that the pixel coordinates can be converted into axial length coordinates laterally. During the formal acquisition, a reference image sequence is continuously acquired before the valve is opened to cover the frost-free or weak frost state. Then, during the filling process, the outer wall image of the target thin tube segment is continuously acquired, and each frame image is timestamped for traceability. In each frame image acquired, the bounding box of the target thin tube segment is manually initialized once, and the bounding box position is updated in subsequent frames using a line detection method based on edge strength or a tracking method based on template matching, so as to obtain a target area that always contains the target thin tube segment. The target area preferably covers the entire possible length of the frost band along the thin tube axis and the visible width of the thin tube outer wall in the transverse direction, with an additional edge margin of no more than half the outer diameter of the thin tube to avoid excessive noise from the background area.

[0056] After obtaining the target area, the target area is processed into grayscale. Specifically, the red, green and blue channels of the target area are weighted and the single-channel brightness image is calculated and output as a grayscale image. The preferred weighting method is 0.299 multiplied by the red channel, 0.587 multiplied by the green channel and 0.114 multiplied by the blue channel to match the brightness response of the human eye and improve the contrast of the frosting scattering area in the grayscale domain. If necessary, the grayscale image is normalized to eliminate the influence of small fluctuations in the light source, thereby obtaining the stable grayscale input required for the subsequent generation of the frost band enhanced profile.

[0057] Calculate the average horizontal pixel value corresponding to each axial coordinate along the tube axis of the target area to obtain the initial axial average profile.

[0058] The median of the initial axial average profile is used as the background baseline. Based on the difference between the initial axial average profile and the background baseline, the frost-enhanced profile is obtained.

[0059] Specifically, the tube axis direction refers to the direction along the length of the thin tube, corresponding to the expansion direction of the frost band along its length. The axial coordinates refer to numbering each axial position in the target area based on the tube axis direction and mapping the pixel positions to actual length positions using known geometric scales. The initial axial average profile is a one-dimensional brightness distribution curve formed by arranging the lateral pixel average values ​​corresponding to each axial coordinate in axial order, used to characterize the brightness fluctuations of the frost band along its length. The frost band enhancement profile is a one-dimensional enhancement curve obtained in the form of differences, whose peak value and peak width correspond to the strength and coverage of the frost band on the outer wall. The frost band refers to the area along the outer wall of the target thin tube section of the dilution refrigeration unit. The formation mechanism of the frost-covered area, which is distributed in a band along the axial direction, is that during the filling process, the gas inside the pipe is cooled by local throttling or expansion. When the pipe wall temperature is lower than the condensation temperature of water vapor in the surrounding air and further lower than the freezing temperature, the water vapor in the air first condenses on the surface of the pipe wall and then freezes into fine ice crystals as the temperature continues to drop. The ice crystal layer produces stronger scattering and reflection of incident light, which is manifested as a banded texture with increased brightness in the image. The axial extension range and brightness fluctuation of this banded area reflect the spatial distribution of the cooling intensity of the outer wall. Therefore, the frost band can be used as an observable carrier for identifying the degree of local throttling distribution and filling mapping distortion.

[0060] In detail, after obtaining the grayscale image of the target area containing the target thin tube segment, the range of axial pixel coordinates in the tube axis direction of the target area is first determined, and a calibration relationship from axial pixel coordinates to axial length coordinates is established. Preferably, the known outer diameter of the target thin tube segment or the known length of the lateral calibration object is used as the scale reference. The pixel length of the calibration object in the image is converted into the length corresponding to a unit pixel, so that each axial pixel coordinate corresponds to an axial length coordinate. Then, the target area is scanned column by column or row by row. For each axial pixel coordinate, the corresponding horizontal pixel set is extracted, and the grayscale arithmetic mean of the horizontal pixel set is calculated. The horizontal pixel set preferably covers the visible width of the outer wall of the target thin tube segment and retains a small margin on both sides to absorb slight jitter at the edge of the tube wall. Preferably, the horizontal width is 1.2 to 1.8 times the number of pixels corresponding to the outer diameter of the target thin tube to reduce local texture noise without excessively introducing background. This results in a one-dimensional sequence that varies with the axial length coordinate, and the one-dimensional sequence is arranged in axial order to form an initial axial average profile. The initial axial average profile reflects the average reflection and scattering intensity fluctuations of the outer wall of the target thin tube segment along the length direction.

[0061] The median of the initial axial average profile is calculated and used as the background baseline. The reason for using the median as the background baseline is that frost bands usually only occupy part of the axial position, while most axial positions are in a non-frost or weakly frost state. The median can represent the typical gray level of the non-frost region and is not sensitive to local frost band peaks, thus avoiding the background being raised by the frost band. Then, the background baseline is subtracted from the value at each axial coordinate of the initial axial average profile to obtain the difference sequence, and this difference sequence is used as the frost band enhancement profile. In the frost band enhancement profile, the region with a positive difference value indicates that the gray level of the outer wall at that axial position is higher than that of the non-frost background and corresponds to the frost band region with enhanced scattering. The region with a difference value close to zero indicates that the gray level of the outer wall at that axial position is close to that of the background and corresponds to the non-frost outer wall region. Thus, a frost band enhancement profile that can be directly used for subsequent main peak localization and half-width extraction is obtained.

[0062] S2. Extract the full width at half maximum (FWHM) and edge gradient from the frost-enhanced profile, and calculate the anomalous ratio of the frost-band morphology based on the FWHM and edge gradient.

[0063] In an embodiment of the present invention, the full width at half maximum (FWHM) and edge gradient are extracted from the frost-enhanced profile, and the anomalous ratio of the frost-band morphology is calculated based on the FWHM and edge gradient, including:

[0064] Locate the main peak on the frost-enhanced profile and determine the half-height value corresponding to half the peak height of the main peak.

[0065] Find the intersection of the frost-enhanced profile and the half-height value on both sides of the main peak, and take the axial distance between the two intersection points as the half-height width;

[0066] In detail, after obtaining the frost-enhanced profile, the frost-enhanced profile is regarded as a one-dimensional sequence of axial length coordinate changes along the target thin tube segment. First, the axial position with the largest amplitude is searched in this one-dimensional sequence and the axial position is determined as the main peak position. At the same time, the profile value corresponding to this position is determined as the main peak value to characterize the strongest scattering enhancement point generated by the frost on the outer wall. Then, half of the main peak value is defined as the half-height value, so that the half-height value corresponds to the intermediate intensity level of the frost from background enhancement to peak enhancement. Next, with the main peak position as the center, the sequence is scanned point by point to both sides of the axis to find the nearest transition segment where the profile value changes from higher than the half-height value to lower than the half-height value. In the transition segment on each side, the axial length coordinate when the profile value is exactly equal to the half-height value is calculated by linear interpolation between two adjacent sampling points, thereby obtaining the left intersection point and the right intersection point of the main peak. Finally, the axial length coordinate of the right intersection point and the axial length coordinate of the left intersection point are subtracted to obtain the half-height width. The half-height width is used to characterize the effective coverage range of the frost in the axial direction and to provide the width component for subsequent anomalous ratio calculation.

[0067] Calculate the rate of change of profile values ​​near the intersection point to obtain the edge gradient;

[0068] The anomalous ratio of the frost band morphology is obtained based on the ratio of half-width at half-maximum to edge gradient.

[0069] It should be noted that the anomalous ratio of frost morphology is an indicator used to quantify the degree of morphological anomaly of the frost band on the outer wall of the target slender tube segment in the tube axis direction. It is composed of the proportional relationship between the axial coverage width corresponding to the main peak of the frost band at half-height in the frost band enhancement profile and the steepness of the frost band edge transition. The half-height width reflects the effective extension range of the frost band in the axial direction to achieve moderate reinforcement strength, while the edge gradient reflects the maximum rate of change of the profile value with axial length when the frost band transitions from the background area to the frost band area, and corresponds to the concentration of cooling distribution on the outer wall. The ratio of these two values ​​can... When the width and bluntness of the frost band are aligned to the same scale, a large anomaly in the frost band shape indicates that the frost band is wider in the axial direction and the edge transition is gentler. This corresponds to a distributed cooling or throttling effect and is more likely to cause a deviation in the mapping between the metered release amount and the effective inlet amount. When the anomaly in the frost band shape is smaller, it indicates that the frost band is narrower in the axial direction and the edge transition is steeper. This corresponds to a more concentrated cooling or throttling effect and a relatively lower degree of mapping deviation. Thus, this anomaly ratio can serve as a key input for subsequently generating the throttling influence coefficient and correcting the effective inlet molar amount.

[0070] In detail, after obtaining the intersection points on the left and right sides of the main peak, local profile segments near each intersection point are extracted, centered on the axial length coordinate of each intersection point. The rate of change of profile values ​​within these local profile segments is calculated to characterize the steepness of the transition from the background to the frost zone edge. The axial range of these local profile segments is preferably one to three axial sampling intervals to ensure that the rate of change estimation mainly falls within the edge transition zone near the intersection points and is not affected by the flat top of the main peak or distant noise. Specifically, the difference in frost zone enhanced profile values ​​between two adjacent axial sampling points near the intersection points is divided by the corresponding axial length difference to obtain a differential derivative sequence, which is then calculated on the left side... The maximum absolute value of the differential derivative sequence near the intersection point and the right intersection point is taken as the edge gradient component on that side. Then, the arithmetic mean of the edge gradient components on the left and right sides is calculated to obtain the edge gradient. The larger the edge gradient, the steeper the edge of the frost zone and the more concentrated the throttling cold point. The smaller the edge gradient, the gentler the edge of the frost zone and the more dispersed the cooling distribution. Then, the ratio of the obtained half-width at half-maximum to the edge gradient is calculated to obtain the anomalous ratio of the frost zone shape. The larger the anomalous ratio of the frost zone shape, the wider and blunter the frost zone shape tends to be. The smaller the anomalous ratio of the frost zone shape, the narrower and steeper the frost zone shape tends to be. Thus, a shape quantification result that can be used to generate the subsequent throttling influence coefficient is formed.

[0071] S3. Generate the throttling influence coefficient based on the anomalous ratio of the frost zone morphology;

[0072] In an embodiment of the present invention, the throttling influence coefficient is generated based on the anomalous ratio of the frost zone morphology, including:

[0073] During the reference time period before the valve is opened, the minimum value of the anomalous ratio of the frost band pattern is used as the reference anomalous ratio, and the median absolute deviation of the anomalous ratio of the frost band pattern is calculated as the scale term.

[0074] During the injection process, the difference between the real-time acquired frost band morphology anomaly ratio and the baseline anomaly ratio is calculated, and the difference is divided by the scale term to obtain the normalized anomaly quantity.

[0075] The throttling effect coefficient is obtained by mapping the normalized outliers using a monotonically decreasing function.

[0076] Specifically, the normalized anomaly is a dimensionless quantity used to quantify the degree of deviation of the frost band morphology from the baseline state during the filling process. It is obtained by scaling the difference between the real-time acquired anomaly ratio of the frost band morphology and the baseline anomaly ratio using a scaling term. The baseline anomaly ratio characterizes the lower reference limit of the frost band morphology on the outer wall before valve opening, and the scaling term characterizes the typical fluctuation amplitude of the anomaly ratio under this reference state. Therefore, the magnitude of the normalized anomaly directly reflects whether the current frost band exhibits a wider and blunter edge trend, and the significance of this trend relative to natural fluctuations. A larger normalized anomaly indicates a more significant deviation of the frost band morphology from the baseline and suggests a more dispersed cooling distribution on the outer wall, thus more likely accompanied by distributed throttling within the loop, affecting effective filling. The weakening effect; the throttling effect coefficient is a dimensionless correction coefficient obtained by monotonically decreasing mapping of the normalized anomaly. Its value is used to characterize the degree of weakening of the conversion of the metered release molar amount into the effective inlet molar amount by throttling or equivalent diameter reduction in the loop. When the normalized anomaly increases, the throttling effect coefficient decreases accordingly to indicate that the proportion of the same metered release amount that can effectively enter the dilution loop and form a usable charging state is reduced. When the normalized anomaly is close to zero, the throttling effect coefficient is close to one to indicate that the mapping deviation between the metered release amount and the effective inlet amount is small. Thus, the throttling effect coefficient can serve as a key multiplicative correction factor in the subsequent calculation of the effective inlet molar amount and make the filling control command reflect the strength of the throttling effect revealed by the frost band morphology on the outer wall.

[0077] In detail, while the filling valve remains closed and no helium mixture is released into the dilution refrigeration circuit, the anomalous ratios of the frost band morphology obtained from the continuously acquired outer wall images during this stage, calculated using the aforementioned method, form a baseline sequence. The time range corresponding to this baseline sequence is defined as the baseline time period. Preferably, the baseline time period covers several seconds after the camera and illumination have stabilized to include natural fluctuations in non-frost or weak frost conditions and avoid changes in outer wall temperature caused by valve operation. Subsequently, the minimum value of the anomalous ratio baseline sequence of the frost band morphology is taken as the baseline anomalous ratio within the baseline time period. The baseline anomalous ratio is used to characterize the lower limit of the morphology degree of wide and blunt frost bands under the baseline state and serves as a zero-point reference for subsequent anomalies. At the same time, the median of the absolute deviation of the anomalous ratio baseline sequence of the frost band morphology relative to its median is calculated within the baseline time period and used as a scale term. The reason for using the median absolute deviation as a scale term is that it is insensitive to outliers caused by occasional bright spot noise or brief local condensation and can stably characterize the typical fluctuation amplitude of the anomalous ratio under the baseline state, thereby making the normalization result robust.

[0078] During the filling process after the valve is opened, the anomaly ratio of the real-time acquired frost band morphology is continuously calculated frame by frame and the difference is obtained by subtracting it from the reference anomaly ratio. The difference is then divided by the scale term to obtain the normalized anomaly. When the normalized anomaly is positive, it indicates that the current frost band morphology is wider and has blunter edges than the reference. When the normalized anomaly is close to zero, it indicates that the current frost band morphology is close to the reference state. Subsequently, the normalized anomaly is input into a monotonically decreasing function to obtain the throttling influence coefficient. The selection of the monotonically decreasing function is based on the principle that when the normalized anomaly increases, the throttling influence coefficient should decrease to reflect the weakening effect of distributed throttling on the effective inlet molar amount. Preferably, the monotonically decreasing function adopts the form of one plus the reciprocal of the square of the normalized anomaly to ensure that the output falls between zero and one and has smooth saturation characteristics, so that the throttling influence coefficient can continuously characterize the strength of the throttling influence and be used for subsequent correction of the metered release molar amount.

[0079] S4. Collect the real-time pressure and temperature in the metering chamber, calculate the metering release molar amount, and correct the metering release molar amount through the throttling influence coefficient to obtain the effective inlet molar amount.

[0080] In an embodiment of the present invention, the metered release molar amount is calculated, and the metered release molar amount is corrected by a throttling influence coefficient to obtain the effective inlet molar amount, including:

[0081] Collect real-time pressure and temperature data within the metering chamber;

[0082] Based on the known volume, real-time pressure, and real-time temperature within the metering chamber, the instantaneous molar quantity of the metering chamber at the initial and current moments is calculated using the gas state equation.

[0083] The difference between the instantaneous molar quantity at the start time and the instantaneous molar quantity at the current time is used as the measured molar quantity released.

[0084] Specifically, the metered release molar quantity refers to the amount of gas actually released from the metering chamber to the dilution refrigeration circuit during the helium mixture filling process. It is obtained by comparing the change in the amount of gas contained in the metering chamber at the initial moment and the current moment. The initial moment corresponds to the gas storage in the metering chamber when it is in the ready-to-release state before the valve is opened, and the current moment corresponds to the remaining gas storage in the metering chamber at a certain sampling moment during filling. The difference between the two represents the total amount of gas flowing out of the metering chamber and into the downstream pipeline during that time period. This quantity directly reflects the degree of reduction in the metering chamber inventory and has quantitative traceability. It is the basic input for converting the release quantity at the metering level into the effective inlet molar quantity at the dilution refrigeration circuit level and generating valve control commands.

[0085] In detail, during the controlled refueling process of helium mixtures, the metering chamber is used as a measuring container to quantitatively characterize the amount of gas released. The known volume of the metering chamber is pre-obtained and stored as a fixed parameter. This known volume is preferably obtained from the geometric dimensions of the metering chamber or provided by factory calibration data to ensure traceability of molar quantity conversion. Subsequently, a pressure sensor and a temperature sensor are installed at the gas connection point of the metering chamber, ensuring sufficient heat exchange between both and the gas inside the chamber. The pressure sensor is preferably a model with a range covering the upper limit of refueling and an accuracy not less than 0.25% of full scale to reduce the amplification effect of pressure reading errors on molar quantity conversion. The temperature sensor is preferably... Select a model with a measurement range covering ambient temperature variations and an accuracy of not less than 0.5 Kelvin to avoid conversion errors caused by temperature drift. Before the valve is opened, record the sampling time as the starting time and simultaneously read the metering chamber pressure and temperature at that starting time as the starting pressure and temperature. During the filling process, continuously collect the real-time pressure and temperature in the metering chamber at a fixed sampling period. The sampling period is preferably 0.1 to 1 second to balance control response and noise suppression. For each sampling time, substitute the collected real-time pressure and temperature into the gas state equation to calculate the instantaneous molar quantity in the metering chamber at that sampling time. The formula for calculating the instantaneous molar quantity is:

[0086] In the formula, Let be the instantaneous molar quantity of the measuring cavity at time t. The real-time pressure of the metering chamber at time t. Given the known volume of the measuring cavity, Let be the real-time temperature of the metering cavity at time t. The gas constant is set to 8.314. The basis of this formula is that the gas in the metering chamber is in a macroscopically uniform state and its pressure and temperature are measurable during the filling process. Through the gas state relationship, the measurable pressure and temperature and the known volume can be converted into the amount of gaseous substance, thereby realizing the quantitative measurement of the release amount. Furthermore, the difference between the instantaneous molar amount at the beginning and the instantaneous molar amount at the current moment is used to obtain the metered release molar amount. The metered release molar amount is used to characterize the amount of gas released from the metering chamber into the dilution refrigeration circuit.

[0087] The effective inlet molar amount is obtained by multiplying the metered release molar amount by the throttling effect coefficient.

[0088] In detail, after obtaining the metered release molar quantity and the throttling effect coefficient, the two are aligned at the same sampling time and multiplicatively corrected to obtain the effective inlet molar quantity. Specifically, the metered release molar quantity is read at the current sampling time to characterize the total amount of gas released from the metering chamber to the downstream loop, while the throttling effect coefficient is read to characterize the degree to which throttling or equivalent diameter reduction in the loop weakens the effective proportion of the released gas that can enter the dilution refrigerator loop. Then, the metered release molar quantity and the throttling effect coefficient are multiplied to obtain the effective inlet molar quantity. The effective inlet molar quantity represents the amount of gas that can effectively enter the downstream loop during the release process corresponding to the metered release molar quantity. The amount of gaseous substance diluted in the refrigeration circuit to form a usable charge state is based on the multiplicative correction of the throttling effect coefficient, which is a proportional quantity between zero and one that describes the effective proportional change. When the throttling effect coefficient is close to one, it means that the mapping deviation between the metered release amount and the effective amount in the inlet is small. At this time, the effective inlet molar amount is close to the metered release molar amount. When the throttling effect coefficient decreases, it means that the distributed throttling has a stronger weakening effect on the inlet. At this time, the effective inlet molar amount is correspondingly lower than the metered release molar amount, so that the effective inlet molar amount can adaptively reflect the charging mapping distortion with the change of frost morphology and be used for the generation of subsequent valve control commands.

[0089] S5. Generate valve control commands based on the effective input molar quantity;

[0090] In an embodiment of the present invention, valve control commands are generated based on the effective input molar quantity, including:

[0091] Calculate the difference between the preset injection target value and the current effective inlet molar quantity to obtain the remaining effective quantity;

[0092] The release rate is obtained by calculating the rate of change of the amount of molar released over time.

[0093] In detail, during the filling control phase, the filling target value is used as the target effective inlet molar quantity to be achieved and written into the controller before filling begins. At each control sampling moment, the current effective inlet molar quantity is read as the currently achieved effective filling quantity, and the filling target value is subtracted from the current effective inlet molar quantity to obtain the remaining effective quantity. The remaining effective quantity is used to characterize the amount of effective gaseous substance that still needs to be replenished before filling is completed and serves as the main driving quantity for subsequent control increment calculations. At the same sampling moment, the rate of change of the metered release molar quantity over time is calculated using a discrete sequence of metered release molar quantities to obtain the release rate. Specifically, the difference between the metered release molar quantity at the current sampling moment and the metered release molar quantity at the previous sampling moment is divided by the sampling period between the two sampling moments to obtain the release rate. The release rate is used to characterize the speed at which the metering chamber releases gas under the current valve control action and provides a scale reference for subsequently converting the remaining effective quantity into valve control increments, thereby enabling the controller to generate coherent valve control commands based on the target difference and the actual release capacity.

[0094] The control increment is calculated based on the ratio of the remaining effective amount to the release rate;

[0095] The control increment is superimposed on the valve control command of the previous moment, and the superposition result is limited to the range that the valve can control, to obtain the current valve control command;

[0096] In detail, after obtaining the remaining effective quantity and release rate, these two are used to construct the incremental update quantity for valve control to achieve a closed-loop approximation of the effective inlet molar quantity. Specifically, the remaining effective quantity is divided by the release rate to obtain the base quantity of the control increment. This base quantity corresponds to the equivalent adjustment range required to complete the remaining effective refueling under the current release capacity level. Its meaning is to normalize the amount of effective gaseous substance that needs to be replenished according to the amount of substance that can be released per unit time, thereby enabling the control update to have an adaptive scale for different flow stages. Considering that the release rate may reach a minimum value, leading to ratio divergence, it is preferable to add a micro-term determined by the measurement resolution and sampling period to the release rate to ensure that the denominator is always non-zero and does not change the dominant trend of the release rate. Subsequently, the control increment is superimposed... The unconstrained current command is obtained by adding the valve control command from the previous moment. This superposition is used to form a smooth recursive control process and avoid discontinuous jumps in the valve control quantity between adjacent sampling moments. Furthermore, the unconstrained current command is restricted to the range that the valve can control to obtain the current valve control command. The range that the valve can control is determined by the minimum controllable opening and the maximum controllable opening of the valve actuator. It is preferably normalized to zero corresponding to full closure and one corresponding to full opening. If the unconstrained current command is less than the minimum controllable opening, it is set to the minimum controllable opening. If the unconstrained current command is greater than the maximum controllable opening, it is set to the maximum controllable opening. This ensures that the output valve control command always falls within the action range that the actuator can achieve and can be reliably issued and executed.

[0097] The circuit pressure of the dilution refrigeration unit is monitored in real time. When the effective inlet molar amount is greater than or equal to the preset charging target value, or when the circuit pressure is greater than or equal to the preset rated pressure boundary, the charging valve is immediately closed.

[0098] In detail, throughout the entire process of controlled helium mixture refueling, the dilution refrigerator circuit pressure is used as a safety constraint and continuously collected by a circuit pressure sensor. The circuit pressure sensor is preferably located downstream of the refueling valve and near the inlet of the dilution refrigerator circuit, so that its readings represent the actual pressure state of the circuit and can respond promptly to abnormal pressure increases. The controller reads the circuit pressure in real time at a fixed sampling period and compares it with the rated pressure boundary. Simultaneously, it reads or updates the effective inlet molar quantity calculated from the metered release molar quantity and the throttling effect coefficient, and compares it with the refueling target value. The refueling target value defines the effective inlet molar quantity termination point to be achieved in this refueling, and the rated pressure boundary defines the maximum safe pressure the circuit can withstand, preferably based on… The pressure rating or safety valve setting of the dilution refrigeration unit circuit is determined to ensure consistency with equipment safety specifications. When the effective inlet molar quantity is determined to be greater than or equal to the charging target value, the controller immediately generates a valve closing command and drives the charging valve to a fully closed state to terminate the charging process, thereby avoiding the risk of instability or overpressure caused by exceeding the target charging amount. When the circuit pressure is determined to be greater than or equal to the rated pressure boundary, the controller also immediately generates a valve closing command and drives the charging valve to a fully closed state to eliminate the source of continuous pressure rise. At the same time, the circuit pressure value and effective inlet molar quantity at the trigger time are recorded as a safety event log for subsequent review, thereby realizing a safety termination mechanism constrained by both circuit pressure and effective inlet molar quantity, and ensuring that the charging process is controllable in terms of both target achievement and pressure safety.

[0099] like Figure 2 The diagram shown is a functional block diagram of a controllable filling system for a diluted helium mixture used in a refrigeration unit, provided by an embodiment of the present invention.

[0100] In this embodiment, the functions of each module / unit are as follows:

[0101] The image acquisition and processing module is used to acquire images of the outer wall of the target thin tube segment of the dilution refrigeration machine, and process the images to generate a frost-enhanced profile along the tube axis.

[0102] The morphological anomaly ratio module is used to extract the full width at half maximum (FWHM) and edge gradient on the frost-enhanced profile, and to calculate the anomaly ratio of the frost morphology based on the FWHM and edge gradient.

[0103] The throttling coefficient module is used to generate a throttling influence coefficient based on the anomalous ratio of the frost zone morphology;

[0104] The correction module is used to collect the real-time pressure and temperature in the metering chamber, calculate the metering release molar amount, and correct the metering release molar amount through the throttling influence coefficient to obtain the effective inlet molar amount.

[0105] The valve control module is used to generate valve control commands based on the effective input molar quantity.

[0106] like Figure 3 As shown in the figure, this diagram illustrates the one-dimensional distribution of frost-enhanced profiles along the tube axis and the measurement elements used for morphological quantification. The horizontal axis represents the axial length of the target thin tube segment, and the vertical axis represents the frost-enhanced profile value. The frost-enhanced profile value characterizes the enhancement magnitude of the frost-covered area on the outer wall relative to the background baseline. The peak of the curve corresponds to the main peak position of the frost-enhanced profile, reflecting the strongest enhancement point of the frost band in the axial direction. The horizontal dashed line in the figure represents the half-height value, which is half the peak value of the main peak. At the half-height value, the curve and the horizontal dashed line intersect on both sides of the main peak. The axial distance between the two intersection points is defined as the half-height width. The half-height width characterizes the effective coverage area of ​​the frost band in the axial direction to achieve a moderate enhancement level. The steepness of the rising and falling edges of the curve near the intersection points corresponds to the edge gradient. The edge gradient characterizes the rate of change of the profile value with axial length when the frost band transitions from the background area to the frost band area. Thus, this figure intuitively illustrates the key quantities required for frost band morphological quantification and their value positions on the profile curve through the geometric relationship between the half-height value, half-height width, and edge gradient.

[0107] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A controlled method for adding a helium mixture to a dilution refrigerator, characterized in that, include: S1. Acquire an image of the outer wall of the target thin tube section of the dilution refrigeration unit, and process the image to generate a frost-enhanced profile along the tube axis. S2. Extract the full width at half maximum (FWHM) and edge gradient from the frost-enhanced profile, and calculate the anomalous ratio of the frost-band morphology based on the FWHM and edge gradient. S3. Generate the throttling influence coefficient based on the anomalous ratio of the frost zone morphology; S4. Collect the real-time pressure and temperature in the metering chamber, calculate the metering release molar amount, and correct the metering release molar amount through the throttling influence coefficient to obtain the effective inlet molar amount. S5. Generate valve control commands based on the effective input molar quantity.

2. The method for controlled refueling of a helium mixture for a dilution refrigerator according to claim 1, characterized in that, Generating a frost-reinforced profile along the tube axis includes: Select the target region containing the target thin tube segment from the acquired image and perform grayscale processing on the target region; The average value of the horizontal pixels corresponding to each axial coordinate is calculated along the tube axis direction of the target area to obtain the initial axial average profile. The axial coordinates are calibrated by the known geometric scale of the target thin tube segment to convert the axial pixel coordinates into axial length coordinates. The median of the initial axial average profile is used as the background baseline. Based on the difference between the initial axial average profile and the background baseline, the frost-enhanced profile is obtained.

3. The controlled refueling method for a helium mixture used in a dilution refrigerator according to claim 1, characterized in that, The full width at half maximum (FWHM) and edge gradient are extracted from the frost-enhanced profile, and the anomalous ratio of the frost-band morphology is calculated based on the FWHM and edge gradient, including: Locate the main peak on the frost-enhanced profile and determine the half-height value corresponding to half the peak height of the main peak. Find the intersection of the frost-enhanced profile and the half-height value on both sides of the main peak, and take the axial distance between the two intersection points as the half-height width; Calculate the rate of change of profile values ​​near the intersection point to obtain the edge gradient; The anomalous ratio of the frost band morphology is obtained by using the ratio of half-width at half-maximum to edge gradient.

4. The controlled refueling method for a helium mixture used in a dilution refrigerator according to claim 1, characterized in that, Throttling influence coefficients are generated based on the anomalous ratio of frost zone morphology, including: During the reference time period before the valve is opened, the minimum value of the anomalous ratio of the frost band pattern is used as the reference anomalous ratio, and the median absolute deviation of the anomalous ratio of the frost band pattern is calculated as the scale term. During the injection process, the difference between the real-time acquired frost band morphology anomaly ratio and the baseline anomaly ratio is calculated, and the difference is divided by the scale term to obtain the normalized anomaly quantity. The throttling effect coefficient is obtained by mapping the normalized outliers using a monotonically decreasing function.

5. The controlled refueling method for a helium mixture used in a dilution refrigerator according to claim 1, characterized in that, The effective inlet molar quantity is obtained by correcting the metered release molar quantity using a throttling effect coefficient, including: Based on the known volume, real-time pressure, and real-time temperature within the metering chamber, the instantaneous molar quantity of the metering chamber at the initial and current moments is calculated using the gas state equation. The difference between the instantaneous molar quantity at the start time and the instantaneous molar quantity at the current time is used as the measured molar quantity released. The effective inlet molar amount is obtained by multiplying the metered release molar amount by the throttling effect coefficient.

6. The method for controlled refueling of a helium mixture for a dilution refrigerator according to claim 1, characterized in that, Valve control commands are generated based on the effective input molar quantity, including: Calculate the difference between the preset injection target value and the current effective inlet molar quantity to obtain the remaining effective quantity; The release rate is obtained by calculating the rate of change of the amount of molar released over time. The control increment is calculated based on the ratio of the remaining effective amount to the release rate; The control increment is superimposed on the valve control command from the previous moment, and the superposition result is limited to the range that the valve can control, to obtain the current valve control command.

7. The method for controlled refueling of a helium mixture for a dilution refrigerator according to claim 1, characterized in that, The method also includes a safe termination step: real-time monitoring of the circuit pressure of the dilution refrigeration unit circuit; when the effective inlet molar amount is greater than or equal to the preset charging target value, or when the circuit pressure is greater than or equal to the preset rated pressure boundary, the charging valve is immediately closed.

8. A controllable dosing system for a helium mixture used in a dilution refrigerator, characterized in that, The system includes: The image acquisition and processing module is used to acquire images of the outer wall of the target thin tube segment of the dilution refrigeration machine, and process the images to generate a frost-enhanced profile along the tube axis. The morphological anomaly ratio module is used to extract the full width at half maximum (FWHM) and edge gradient on the frost-enhanced profile, and to calculate the anomaly ratio of the frost morphology based on the FWHM and edge gradient. The throttling coefficient module is used to generate a throttling influence coefficient based on the anomalous ratio of the frost zone morphology; The correction module is used to collect the real-time pressure and temperature in the metering chamber, calculate the metering release molar amount, and correct the metering release molar amount through the throttling influence coefficient to obtain the effective inlet molar amount. The valve control module is used to generate valve control commands based on the effective input molar quantity.

Images