HPLC sample pretreatment injection speed of viscosity feedforward control method

Abstract

This invention relates to the field of automatic control technology, specifically to a viscosity feedforward control method for injection rate in HPLC sample pretreatment. The method includes: correcting the transmitted light intensity of each group of experimental samples using blank samples and particle-scattered samples; determining the color-corrected relative light intensity of each group of transmitted light intensity; determining the anomaly index of each group of transmitted light intensity; filtering out several retained transmitted light intensities for each group of experimental samples; and, combined with the particle concentration, determining the true transmitted light intensity of each group of experimental samples. This is used to construct a mapping model of true transmitted light intensity versus viscosity, outputting the viscosity of the current sample to determine the actual aspirated volume and actual injection rate. This invention enables precise and adaptive injection rate control in HPLC sample pretreatment through viscosity feedforward.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, and specifically to a viscosity feedforward control method for the injection rate in HPLC sample pretreatment. Background Technology

[0002] In automated HPLC (High Performance Liquid Chromatography) sample pretreatment, the accuracy and stability of the injection (i.e., sample aspiration and injection) process directly determine the accuracy and repeatability of the final analytical results. Currently, most mainstream automated samplers employ fixed aspiration volume settings and constant injection rate control logic.

[0003] Existing problems: The physical properties of actual test samples (such as serum, plasma, tissue homogenates, and traditional Chinese medicine extracts) vary significantly, especially in viscosity. A constant injection rate easily leads to fluctuations in tubing resistance, sample volume deviation, peak distortion, and reduced repeatability. For example, high-viscosity samples generate air bubbles during aspiration due to high resistance, and during injection, the sample adheres to the needle wall due to poor fluidity, resulting in significant deviations in the actual injection volume. Conversely, low-viscosity samples are prone to turbulent air bubbles if aspiration is too fast, while injection speeds that are too slow cause sample evaporation, leading to large concentration deviations. Furthermore, current mainstream equipment relies on post-injection pressure feedback correction or manual speed adjustment, which suffers from control lag, passive response, and dependence on human experience, making it unsuitable for high-throughput, multi-matrix sample batch testing scenarios. Simultaneously, existing technologies lack a pre-injection viscosity identification mechanism and have not established a dynamic matching model between viscosity and injection speed, making it difficult to compensate for fluid resistance disturbances in advance. This results in the injection accuracy and stability of high-viscosity samples failing to meet the requirements of high-precision testing. Summary of the Invention

[0004] This invention provides a viscosity feedforward control method for injection rate in HPLC sample pretreatment to solve existing problems.

[0005] The viscosity feedforward control method for injection rate in HPLC sample pretreatment of the present invention adopts the following technical solution: One embodiment of the present invention provides a viscosity feedforward control method for injection rate in HPLC sample pretreatment, the method comprising the following steps: Obtain the transmitted light intensity and corresponding pressure value for each group of blank samples and each group of experimental samples; the viscosity, hemoglobin concentration, number of microbubbles and particulate matter concentration of the experimental samples are all known. Particle scattering samples were selected from the experimental samples. The transmitted light intensity of each group of experimental samples was corrected by comparing the transmitted light intensity of blank samples with that of particle scattering samples, and the relative light intensity of each group of transmitted light intensity after color correction was determined. Based on the relative light intensity after color correction of all groups of transmitted light intensity in each experimental sample, and the fluctuation of the corresponding pressure value, the abnormality index of the transmitted light intensity of each group of experimental samples is determined, which is used to screen out several groups of retained transmitted light intensity in each experimental sample; based on the relative light intensity after color correction of each group of retained transmitted light intensity and the abnormality index, combined with the magnitude of particulate matter concentration, the true transmitted light intensity of each group of experimental samples is determined. Based on the viscosity and true transmitted light intensity of all experimental samples, a mapping model of true transmitted light intensity-viscosity is constructed; the true transmitted light intensity of the current sample is obtained, input into the mapping model of true transmitted light intensity-viscosity, and the viscosity of the current sample is output; based on the viscosity of the current sample, the actual aspirate volume and actual injection rate of the current sample are determined.

[0006] Furthermore, the specific steps involved in selecting the particle scattering samples from the experimental samples are as follows: All experimental samples with a hemoglobin concentration of 0, a microbubble count of 0, but a particulate matter concentration of 0 were recorded as particulate scattering samples.

[0007] Furthermore, the specific steps for determining the color-corrected relative light intensity of each group of transmitted light intensity for each group of experimental samples are as follows: Each group of transmitted light intensity includes: transmitted light intensity under green light and transmitted light intensity under near-infrared light; For blank samples, the mean value of the transmitted light intensity of all groups under green light is obtained and recorded as the baseline green light transmitted light intensity. The mean value of the transmitted light intensity of all groups under near-infrared light is obtained and recorded as the baseline near-infrared light transmitted light intensity. For each group of transmitted light intensities, the ratio of the transmitted light intensity of each experimental sample under green light to the transmitted light intensity of the reference green light is obtained, and recorded as the calibrated transmitted light intensity under green light; the ratio of the transmitted light intensity of each experimental sample under near-infrared light to the transmitted light intensity of the reference near-infrared light is obtained, and recorded as the calibrated transmitted light intensity under near-infrared light. Based on the calibration transmitted light intensity of the particle scattering sample under green light and near-infrared light, determine the particle scattering wavelength reference correction coefficient. The relative light intensity after color correction is determined based on the calibrated transmitted light intensity of each experimental sample under green light and near-infrared light, as well as the particle scattering wavelength reference correction coefficient.

[0008] Furthermore, the specific steps for determining the particle scattering wavelength reference correction coefficient are as follows: For each group of transmitted light intensities, the ratio of the calibrated transmitted light intensity under near-infrared light to the calibrated transmitted light intensity under green light for each group of particle scattering samples is obtained and denoted as the particle scattering wavelength correction coefficient. The mean value of the particle scattering wavelength correction coefficients corresponding to the transmitted light intensity of all groups of particle scattering samples is denoted as the particle scattering wavelength reference correction coefficient.

[0009] Furthermore, the specific steps for determining the color-corrected relative light intensity based on the calibrated transmitted light intensity of each experimental sample under green light and near-infrared light, and the particle scattering wavelength reference correction coefficient, are as follows: For each group of transmitted light intensities, the product of the ratio of the calibrated transmitted light intensity under green light to the calibrated transmitted light intensity under near-infrared light for each experimental sample and the particle scattering wavelength reference correction coefficient is obtained and recorded as the color-corrected relative light intensity.

[0010] Furthermore, the specific steps for determining the anomalous index of the transmitted light intensity for each group of experimental samples are as follows: For each experimental sample, a three-standard-deviation interval is constructed based on the mean and standard deviation of the color-corrected relative light intensity corresponding to all groups of transmitted light intensity. The deviation cutoff coefficient of the color-corrected relative light intensity corresponding to each group of transmitted light intensity based on the three-standard-deviation interval is denoted as the relative light intensity anomaly index in each group of transmitted light intensity. The upper limit cutoff coefficient of the absolute value of the relative deviation of the pressure values ​​corresponding to two adjacent groups of transmitted light intensity is calculated and denoted as the pipeline pressure anomaly index in each group of transmitted light intensity. Based on the preset relative light intensity anomaly weight and the preset pipeline pressure anomaly weight, the relative light intensity anomaly index and the pipeline pressure anomaly index in each group of transmitted light intensity are weighted and summed to obtain the anomaly index of each group of transmitted light intensity.

[0011] Furthermore, the specific steps for selecting several sets of retained transmitted light intensities for each group of experimental samples are as follows: If the abnormal index of the transmitted light intensity of each group of experimental samples is less than the preset abnormal threshold, it is recorded as the retained transmitted light intensity.

[0012] Furthermore, the specific steps for determining the true transmitted light intensity of each group of experimental samples are as follows: The window length for each experimental sample group is determined based on the particulate matter concentration; the particulate matter concentration is positively correlated with the window length. For each experimental sample, obtain the color-corrected relative light intensity sequence corresponding to the retained transmitted light intensity of all groups. Using the complement of the anomaly index of the retained transmitted light intensity of each group as the weight, perform a weighted moving average based on the window length to obtain the true transmitted light intensity sequence. The mean of the true transmitted light intensity sequence is recorded as the true transmitted light intensity of each experimental sample.

[0013] Furthermore, the specific steps for determining the actual aspirated volume and actual injection rate of the current sample are as follows: Based on the difference between the current sample viscosity and the preset standard viscosity, and in conjunction with the preset target injection volume, determine the actual amount of the current sample absorbed; The product of the ratio of the preset standard viscosity to the viscosity of the current sample and the preset standard injection rate is denoted as the actual injection rate of the current sample.

[0014] Furthermore, the specific steps for determining the actual amount of sample drawn from the current sample based on the difference between the current sample's viscosity and the preset standard viscosity, combined with the preset target injection volume, are as follows: The product of the difference between the current sample viscosity and the preset standard viscosity and the preset viscosity correction coefficient is recorded as the initial injection volume correction coefficient. The product of the sum of the initial injection volume correction coefficient and the preset reference constant and the preset target injection volume is recorded as the actual amount of sample absorbed.

[0015] The beneficial effects of the technical solution of the present invention are: In this embodiment of the invention, the transmitted light intensity of each group of experimental samples is corrected by using the transmitted light intensity of each group of blank samples and particle scattering samples. The relative light intensity of each group of transmitted light intensity after color correction is determined, and then the anomaly index of each group of transmitted light intensity is determined to screen out several groups of retained transmitted light intensities for each group of experimental samples. Combined with the particle concentration, the true transmitted light intensity of each group of experimental samples is determined, which is used to construct a mapping model of true transmitted light intensity-viscosity. The true transmitted light intensity of the current sample is obtained, input into the mapping model, and the viscosity of the current sample is output. The viscosity is determined by the actual aspirated volume and injection rate of the current sample. Through multi-interference factor fusion correction, the influence of sample color, bubbles, and particulate matter is effectively overcome, reducing viscosity detection errors and improving the fitting accuracy of the mapping model. This allows for real-time and accurate perception of sample viscosity (indirectly characterized by transmitted light intensity), dynamic adjustment of aspirated volume and injection rate, and completion of pre-injection viscosity identification. A dynamic matching model between viscosity and injection rate is established, achieving viscosity feedforward control of injection rate. This solves the problems of wall adhesion and bubbles in high-viscosity samples (such as lipemia and hyperglobulinemia) and volatilization and turbulence in low-viscosity samples. Therefore, this invention enables precise and adaptive injection rate control in HPLC sample pretreatment through viscosity feedforward. Attached Figure Description

[0016] Figure 1 This is a flowchart of the viscosity feedforward control method for injection rate in HPLC sample pretreatment according to the present invention. Figure 2 Sample needle cross-section; Figure 3 A schematic diagram of the transmitted light intensity collected by the sample needle. Detailed Implementation

[0017] The following description, in conjunction with the accompanying drawings, details the specific scheme of the viscosity feedforward control method for the injection rate in HPLC sample pretreatment provided by this invention.

[0018] Please see Figure 1 The diagram illustrates a flow chart of a viscosity feedforward control method for injection rate in HPLC sample pretreatment according to an embodiment of the present invention. The method includes the following steps: Step S001: Obtain the transmitted light intensity and corresponding pressure value for each group of blank samples and each group of experimental samples; the viscosity, hemoglobin concentration, number of microbubbles and particulate matter concentration of the experimental samples are known.

[0019] In this embodiment, the modules used for HPLC sample pretreatment include: (1) Sample needle assembly: integrated dynamic optical coupler detection module, with the needle tip adopting a multi-porous anti-clogging design (pore diameter of 0.1 mm). (2) Dynamic optical coupler detection module: integrated inside the sample needle cleaning swab, consisting of dual-wavelength LED light source (550 nm green light and 850 nm near-infrared light) and a broadband photoelectric sensor on the opposite side of the optical path. During operation, the two light sources emit light alternately in time, and the sensor collects light signals passing through the sampling needle area at different wavelengths, and monitors the transmitted light intensity of the liquid at the needle tip in real time (negatively correlated with viscosity). (3) Micro-dispensing drive module: includes a high-precision plunger pump (resolution 0.1 μL) and a stepper motor, supporting dynamic adjustment of the aspiration or injection speed (range 10 to 500 μL / s). (4) Core control module: built-in microprocessor, performing transmitted light intensity-viscosity mapping and parameter calculation.

[0020] First, transmitted light intensity analysis was performed on a large number of experimental samples with known viscosity to inform the subsequent construction of the transmitted light intensity-viscosity mapping model. Specifically: (1) Determine the experimental sample group.

[0021] (1.1) Prepare standard liquids covering the viscosity range of common HPLC samples.

[0022] Eleven sets of standard samples with viscosities of 1, 5, 10, 15, 20, 25, 30, 35, 40, 45, and 50 mPa·s were prepared using standard simulated solutions with synchronously varying viscosity and turbidity (e.g., by proportionally blending standard lipid emulsions of gradient concentrations into glycerol-water solutions, or by adding standard formalinium scattering agent). These formulated simulated solutions, while exhibiting the set kinetic viscosity, introduce suspended particles strongly correlated with their viscosity levels into the system, thereby generating a stable, progressively increasing scattering attenuation along the detection optical path. This ensures effective gradient data support for back-calculating the predicted viscosity from the actual transmitted light intensity.

[0023] (1.2) To simulate the complexity of real samples, additional "subsamples" containing different interfering factors were prepared for each viscosity grade of the standard solution: (1.2.1) Color interference: Simulate hemolysis (red) and jaundice (yellow). Add hemoglobin standards to make hemoglobin concentrations of 0, 2.5, 5, 7.5, and 10 (unit: g / L).

[0024] (1.2.2) Bubble interference: Subsamples containing approximately 0, 1, 2, or 3 tiny bubbles were prepared by slight oscillation or micro-injection of air.

[0025] (1.2.3) Particulate matter interference: Simulate cell debris or undissolved matter. Add micron-sized silica or standard particulate matter to achieve particulate matter concentrations of 0, 1, 2.5, and 5 (unit: mg / mL).

[0026] This resulted in 880 subsamples (11 (viscosity) × 5 (color) × 4 (bubbles) × 4 (particulate matter)) as experimental samples, ensuring the model's broad representativeness.

[0027] (2) Determine the blank sample group.

[0028] Set a pure water sample as a blank sample.

[0029] (3) Determine the transmitted light intensity of each sample.

[0030] For any sample, the sample needle is lowered to 2 mm below the liquid surface in the sampling tube. A pre-set detection volume (5 μL) of sample is aspirated at a low speed (50 μL / s) using the micro-dispensing drive module, avoiding turbulence and air bubbles. The sample needle is then raised to rest at the origin. The transmitted light intensity under 550 nm green light and 850 nm near-infrared light (unit: ADC counts) are alternately collected using the dynamic optical coupler detection module. This constitutes one set of transmitted light intensity data. A set of transmitted light intensity data is collected every 1 second, for a total of 10 sets. Simultaneously, a miniature pressure sensor is used to collect the pressure value within the sample needle tubing corresponding to each set of transmitted light intensity data (the average pressure during the time period of each transmitted light intensity data collection is taken, unit: bar).

[0031] Therefore, for each group of transmitted light intensities, the transmitted light intensity under 550nm green light and the transmitted light intensity under 850nm near-infrared light, as well as the pressure value, are obtained for each sample group. Since the transmitted light intensity is always greater than 0, the denominator in the subsequent calculation formula based on the transmitted light intensity must not be 0.

[0032] In this embodiment, the sample needle cross-sectional view is shown as follows: Figure 2 As shown, Figure 2 This includes: plunger pump connection hole, needle inner hole, sampling hole, and needle tip. A schematic diagram of the transmitted light intensity collected by the sample needle is shown below. Figure 3As shown, Figure 3 It includes: a sampling needle, a cleaning swab, a sample, a dynamic optocoupler shown in the magnified image A of the needle tip (magnification ratio of 3:1), and a sample drop hanging on the outer wall of the needle.

[0033] Step S002: Select particle scattering samples from the experimental samples, and correct the transmitted light intensity of each group of experimental samples by comparing the transmitted light intensity of the blank samples with that of the particle scattering samples, and determine the relative light intensity of each group of transmitted light intensity after color correction for each group of experimental samples.

[0034] In this embodiment, firstly, an optical module integrated on the sample needle is used to pre-absorb a small amount of sample and collect its transmitted light intensity signal. Then, through a multi-interference factor fusion correction model, interference from color, bubbles, and particles is removed from the original light intensity signal to obtain the "true transmitted light intensity" that is only related to viscosity. Next, the viscosity of the current sample is quickly calculated using a pre-established "true transmitted light intensity-viscosity" mapping model. Finally, based on the calculated viscosity, the sample aspiration volume and injection speed are adjusted by feedforward using a quantization formula to ultimately perform a precise and appropriate injection operation.

[0035] The multi-interference factor fusion correction model works as follows: First, by comparing the "detection wavelength" (550nm green light), which is affected by color, and the "reference wavelength" (850nm near-infrared light), which is not affected by color, the ratio of calibrated transmitted light intensity at the two wavelengths after calibration with blank samples is calculated, which can offset the intensity change caused by color absorption. Then, by using particle scattering samples with different particle concentrations, a particle scattering wavelength reference correction coefficient is determined to correct the ratio of calibrated transmitted light intensity at the two wavelengths, eliminating the offset caused by the inconsistent scattering attenuation of particles at the two wavelengths, and thus obtaining the color... The corrected relative light intensity was analyzed. Further analysis revealed that tiny bubbles within the sample needle strongly scatter light, causing an abnormally high spike in transmitted light intensity within a very short time. Additionally, the passage of bubbles caused minor pressure fluctuations in the pipeline. From the 10 transmitted light intensity sets for each experimental sample, multiple sets of transmitted light intensities were selected to retain them, while sets with bubble interference were removed. Finally, the color-corrected relative light intensity sequences corresponding to all retained transmitted light intensities were subjected to adaptive window-length weighted filtering to eliminate irregular high-frequency spikes in the transmitted light intensity signal caused by random scattering of light from particulate matter in the sample. This process completed the fusion and correction of multiple interference factors in the transmitted light intensity, allowing the model to output the true transmitted light intensity for each experimental sample.

[0036] It should be noted that the transmitted light intensity method is based on the principle that "the more viscous the sample, the worse the light transmittance may be (e.g., lipemia turbidity)," but there are interferences: (1) Sample color: Hemolysis (red) or jaundice (yellow) will absorb light of specific wavelengths, affecting the transmitted light intensity, but it is not directly related to viscosity. (2) Bubble interference: Tiny bubbles in the needle tip or tubing will severely scatter light, leading to an abnormal increase in transmitted light intensity. (3) Particulate matter: Cell debris and other particulate matter in samples that have not been fully centrifuged will affect scattering.

[0037] Color interference correction: It should be noted that sample color (e.g., hemoglobin) selectively absorbs light of specific wavelengths, causing an abnormal decrease in the intensity of that wavelength, but this decrease is unrelated to viscosity. This embodiment uses a dual-wavelength ratio method to eliminate this effect. A "detection wavelength" (550nm green light) affected by color and a "reference wavelength" (850nm near-infrared light) unaffected by color are selected. Color has minimal impact on the intensity of the reference wavelength, primarily affecting the intensity of the detection wavelength. Calculating the ratio of the two can offset the intensity change caused by color absorption. Using only the original dual-wavelength ratio can only eliminate the interference of pigment selective absorption; however, inherent optical path losses and the synchronous scattering attenuation of particles at both wavelengths introduce a fixed system offset, which cannot be eliminated by the original ratio alone.

[0038] Therefore, particle scattering samples were first selected from the experimental samples. Then, the transmitted light intensity of each group of experimental samples was corrected by comparing the transmitted light intensity of the blank samples with that of the particle scattering samples, and the relative light intensity of each group of transmitted light intensity after color correction was determined.

[0039] Step S003: Based on the relative light intensity after color correction of all groups of transmitted light intensity in each experimental sample and the fluctuation of the corresponding pressure value, determine the abnormality index of each group of transmitted light intensity in each experimental sample, so as to screen out several groups of retained transmitted light intensity in each experimental sample; based on the relative light intensity after color correction of each group of retained transmitted light intensity and the abnormality index, combined with the magnitude of particulate matter concentration, determine the true transmitted light intensity of each experimental sample.

[0040] Bubble interference correction.

[0041] It should be noted that tiny air bubbles within the sample needle strongly scatter light, causing an abnormally high "peak" in transmitted light intensity within a very short time. This peak is unrelated to viscosity and severely distorts the average value. Therefore, outlier (peak) removal is necessary, along with auxiliary verification using data from a micro-pressure sensor (air bubbles cause minute pressure changes in the tubing as they pass through). This is because the sample needle and thin tubing are microchannels; liquid viscosity, the weight of the liquid column, and tubing wall friction create continuous flow resistance. Even when the sample needle is raised and stationary, without suction or drainage, a static pressure gradient still exists within the tubing. Minor temperature fluctuations, tubing deformation, and sample evaporation will slowly alter the pressure baseline.

[0042] Therefore, based on the relative light intensity after color correction of all groups of transmitted light intensity in each experimental sample, and the fluctuation of the corresponding pressure value, the abnormal index of each group of transmitted light intensity in each experimental sample is determined, in order to screen out several groups of retained transmitted light intensity in each experimental sample.

[0043] Particulate matter interference correction.

[0044] It should be noted that particulate matter in the sample randomly scatters light, causing irregular high-frequency "glitch" fluctuations in the transmitted light intensity signal. These fluctuations do not reflect viscosity and require smoothing using digital filtering techniques.

[0045] Therefore, based on the color-corrected relative light intensity and anomaly index of the retained transmitted light intensity of each experimental sample, combined with the particle concentration, the true transmitted light intensity of each experimental sample was determined.

[0046] Step S004: Based on the viscosity and true transmitted light intensity of all experimental samples, construct a mapping model of true transmitted light intensity-viscosity; obtain the true transmitted light intensity of the current sample, input it into the mapping model of true transmitted light intensity-viscosity, and output the viscosity of the current sample; determine the actual aspirate volume and actual injection rate of the current sample based on the viscosity of the current sample.

[0047] Construct a mapping model of real transmitted light intensity and viscosity.

[0048] Based on the known viscosity and true transmitted light intensity of all experimental samples, a quadratic polynomial fitting using the least squares method is employed to construct a mapping model between true transmitted light intensity and viscosity. Specifically: in The model predicts the first The viscosity of the experimental samples , as well as All coefficients are obtained by least squares fitting, which is a well-known technique. The specific method will not be introduced here. For the first The true transmitted light intensity of the experimental samples.

[0049] Determine the current sample aspiration volume and injection rate.

[0050] Following the method described above for obtaining the true transmitted light intensity of the experimental samples, obtain the true transmitted light intensity of the current sample. .

[0051] The true transmitted light intensity of the current sample The input is fed into the mapping model of real transmitted light intensity-viscosity, and the output is the viscosity of the current sample. Among them, viscosity It must be greater than 0.

[0052] Therefore, based on the viscosity of the current sample The size determines the actual amount of the current sample drawn. Compared with the actual injection rate .

[0053] For the current sample, complete the preprocessing workflow: (1) Empty the pre-absorbed sample: Move the sample needle to the waste liquid tank and drain the 5 μL sample previously used for testing. This is to ensure that the test sample does not contaminate subsequent analyses.

[0054] (2) Formal aspiration: The sample needle is re-inserted into the sampling tube, and the calculated actual aspiration volume is drawn. Collect samples.

[0055] (3) Precise injection: The sample needle is moved to the injection port of the HPLC injection valve and the actual injection rate is calculated. The sample is injected into the quantitative loop of the injection valve. For highly viscous samples, Slower, but the propulsion is smoother; for low-viscosity samples, It may be faster, but it is still within the range where bubbles are unlikely to form.

[0056] (4) After the injection is completed, the sample needle is subjected to "reverse pulse air blowing" (air pressure is 0.2MPa, lasting 0.5s) to remove residual droplets at the needle tip, thereby completing the sample needle cleaning and preparing for the next injection.

[0057] This completes the full and precise HPLC sample pretreatment injection process based on viscosity feedforward control.

[0058] Preferably, in some possible implementations of the embodiments of the present invention, all experimental samples (a total of 11 (viscosity) × 3 (particles) = 33 sub-samples) with a hemoglobin concentration of 0, a microbubble count of 0, but a particulate matter concentration of not 0 are recorded as particulate scattering samples.

[0059] Preferably, in some possible implementations of the embodiments of the present invention, for a blank sample (pure water sample), the average value of the transmitted light intensity of all groups under 550nm green light is obtained and recorded as the reference green light transmitted light intensity. The mean value of the transmitted light intensity of all groups under 850nm near-infrared light was obtained and recorded as the reference near-infrared light transmitted light intensity. .

[0060] For the Group transmitted light intensity, obtain the first The transmitted light intensity of the experimental samples under 550nm green light Compared with the reference green light transmission intensity ratio , denoted as the The calibration transmitted light intensity of the experimental samples under 550nm green light ; Get the The transmitted light intensity of the experimental samples under 850nm near-infrared light Compared with the reference near-infrared light transmission intensity ratio , denoted as the The calibration transmitted light intensity of the experimental samples under 850nm near-infrared light . No. Group of transmitted light intensity This represents a total of 10 groups of transmitted light intensities. Group experimental samples This represents a total of 880 experimental samples.

[0061] This eliminates optical path loss.

[0062] Preferably, in some possible implementations of the embodiments of the present invention, for the first Group transmitted light intensity, obtain the first Calibrated transmitted light intensity of the particle scattering sample under 850nm near-infrared light With the Calibrated transmitted light intensity of the particle scattering sample under 550nm green light ratio , denoted as the particle scattering wavelength correction coefficient. Group of particle scattering samples This represents a total of 33 groups of particle scattering samples.

[0063] The mean value of the particle scattering wavelength correction coefficients corresponding to the transmitted light intensity of all groups of particle scattering samples is denoted as the particle scattering wavelength reference correction coefficient. .

[0064] It should be noted that for particle-scattered samples, the transmitted light intensity is only affected by viscosity and particle scattering. Therefore, the transmitted light intensity is equivalent to the product of viscosity and particle scattering attenuation coefficient. If the viscosity of each group of particle-scattered samples is fixed, then... Equivalent to the first The first group of particle scattering samples The ratio of the scattering attenuation coefficient of 850nm infrared particles to that of 550nm green particles under the transmitted light intensity represents the particle scattering wavelength correction coefficient. Since the correction is made by comparing the ratio of the calibrated transmitted light intensity under 550nm green light to the calibrated transmitted light intensity under 850nm near-infrared light, the reciprocal of the above ratio is used for compensation to eliminate the offset caused by the inconsistent scattering attenuation of particles under the two wavelengths.

[0065] Preferably, in some possible implementations of the embodiments of the present invention, for the first Group transmitted light intensity, obtain the first The calibration transmitted light intensity of the experimental samples under 550nm green light With the The calibration transmitted light intensity of the experimental samples under 850nm near-infrared light The ratio of this ratio to the particle scattering wavelength reference correction factor is used. product , denoted as the relative light intensity after color correction. .

[0066] It should be noted that: This is achieved by calibrating the ratio of transmitted light intensity to green light and near-infrared light to offset the selective absorption of hemoglobin, then multiplying by a particle scattering wavelength reference correction factor to compensate for the difference in dual-wavelength scattering by particles. The final result is a relative light intensity that is unaffected by pigments, optical path loss, or differences in particle scattering wavelengths, and is only related to the sample viscosity. If the... If the particulate matter concentration in the experimental group is 0 and there is no interference from dual-wavelength scattering differences, then... , denoted as the relative light intensity after color correction. .

[0067] Preferably, in some possible implementations of the embodiments of the present invention, for the first For each experimental sample, obtain the color-corrected relative light intensity sequence corresponding to the transmitted light intensity of all groups. Calculate the mean and standard deviation Construct a three-standard-deviation interval , obtain Based on the deviation cutoff factor of three standard deviation intervals , denoted as the The first group of experimental samples Relative intensity anomaly index in the group of transmitted light intensity .

[0068] It should be noted that: To find the minimum value function, For absolute value functions, in this embodiment, when Exceeding hour, When it is 1, Not exceeding hour, A value between 0 and 1 is used to characterize the relative light intensity anomaly index. If... A value of 0 indicates a sequence If it is stable, then directly order It is 0.

[0069] For the Group 1 experimental samples, calculate the first Pressure value corresponding to the group of transmitted light intensity Compared to the first Pressure value corresponding to the group of transmitted light intensity Upper limit cutoff coefficient of the absolute value of the relative deviation , denoted as the The first group of experimental samples Abnormal pipeline pressure index in group transmitted light intensity .

[0070] Among them, the pipeline pressure anomaly index in the first group of transmitted light intensity is equal to the pipeline pressure anomaly index in the second group of transmitted light intensity.

[0071] The weighting for relative light intensity anomalies is preset to 0.7, and the weighting for pipeline pressure anomalies is preset to 0.3. This will be used as an example for the following description.

[0072] Based on the preset relative light intensity anomaly weight and the preset pipeline pressure anomaly weight, the first... The relative light intensity anomaly index and the pipeline pressure anomaly index in the group of transmitted light intensity are weighted and summed to obtain the first... Abnormal index of transmitted light intensity .

[0073] Preferably, in some possible implementations of the embodiments of the present invention, the preset abnormal threshold is 0.6, which will be used as an example for description. If If it is greater than or equal to 0.6, then the first... The first group of experimental samples The intensity of transmitted light is recorded as the intensity of transmitted light to be removed, which is then identified as bubble interference.

[0074] like If it is less than 0.6, then the first The first group of experimental samples The intensity of the transmitted light is denoted as the retained transmitted light intensity.

[0075] It should be noted that: if the first If the number of groups with the highest transmitted light intensity that are excluded from the experimental samples is greater than 5, then the group is considered to be the first group. The sample collection for the experimental group was invalid and needed to be repeated. This further eliminated bubble interference.

[0076] Preferably, in some possible implementations of the embodiments of the present invention, for the first For each experimental sample, obtain the color-corrected relative light intensity sequence corresponding to the transmitted light intensity for all groups. ,in For the first The first group of experimental samples The group retains the color-corrected relative light intensity corresponding to the transmitted light intensity. For the first The number of groups that retain transmitted light intensity corresponding to the experimental samples. .

[0077] In this embodiment, based on the adaptive filtering window of particle concentration, if the first... If the particulate matter concentration of the first group of experimental samples is 0 or 1 (unit: mg / mL), then the second group is set. Window length of experimental samples It is 3; if the first If the particulate matter concentration of the first group of experimental samples is 2.5 (unit: mg / mL), then the second group is set up. Window length of experimental samples It is 5; if the first If the particulate matter concentration of the first group of experimental samples is 5 (unit: mg / mL), then the second group is set up. Window length of experimental samples The value is 7. This means that the higher the particulate matter concentration, the larger the window required for smoothing.

[0078] In sequence In the middle, with the first For each group of experimental samples, the complement of the anomalous index of the transmitted light intensity is retained as the weight, based on the window length. Perform a weighted moving average to obtain the first... True transmitted light intensity sequence of experimental samples ,in For the first The first group of experimental samples The group retains the color correction and filtered relative light intensity corresponding to the transmitted light intensity, i.e., the first group. The first group of experimental samples The group retains the true transmitted light intensity.

[0079] It should be noted that for the current sample with unknown particulate matter concentration, a weighted moving average filtering operation is performed using a preset fixed window length (preferably, the window length of the current sample is set to 5).

[0080] Among them, the weighted moving average is a well-known technique, specifically: In sequence In China, with Centered on, forming a length of window ,but The calculation formula is: in, For window The number of elements present in the memory. For window The memory of the first The data value of each element For window The memory of the first The anomaly index corresponding to each element.

[0081] This further eliminates particulate matter interference.

[0082] The first True transmitted light intensity sequence of experimental samples The mean of is denoted as the th. The true transmitted light intensity of the experimental samples .

[0083] Preferably, in some possible implementations of the embodiments of the present invention, a target injection volume is preset. 30 μL, preset standard viscosity The value is 1.5 mPa·s, and the preset reference constant is 1. This will be used as an example for the description.

[0084] Actual amount of the current sample drawn The calculation formula is: in, This is the preset viscosity correction factor. This is used to eliminate the influence of dimensions and to prevent the adjusted actual aspirate volume from exceeding the minimum and maximum injection volume range set by the sample needle. If Greater than Then increase the actual absorption to compensate for the loss due to wall adhesion. If Less than If so, reduce the actual amount absorbed to avoid overdose. This is the correction factor for the initial injection volume. The viscosity is the output of the true transmitted light intensity of the current sample, which is input into the true transmitted light intensity-viscosity mapping model.

[0085] Preferably, in some possible implementations of the embodiments of the present invention, a preset standard injection rate is used. The flow rate is 100 μL / s, and this will be used as an example for explanation.

[0086] Actual injection rate of the current sample The calculation formula is: Among them, for Greater than For highly viscous samples, reduce the injection rate and decrease needle wall adhesion. Less than Low-viscosity samples increase injection speed and prevent evaporation.

[0087] It should be noted that the minimum and maximum injection rates set for the sample needle are obtained. If... If the injection rate is less than the minimum injection rate, then let Equal to the minimum injection rate. If If the injection rate is greater than the maximum injection rate, then let It equals the maximum injection speed.

[0088] This invention is now complete.

[0089] In summary, in this embodiment of the invention, the transmitted light intensity of each group of experimental samples is corrected by using the transmitted light intensity of each group of blank samples and particle scattering samples. The relative light intensity of each group of transmitted light intensity after color correction is determined, and then the anomaly index of each group of transmitted light intensity is determined to screen out several groups of retained transmitted light intensities for each group of experimental samples. Combined with the particle concentration, the true transmitted light intensity of each group of experimental samples is determined, which is used to construct a true transmitted light intensity-viscosity mapping model. The true transmitted light intensity of the current sample is obtained, input into the true transmitted light intensity-viscosity mapping model, and the viscosity of the current sample is output to determine the actual aspirated volume and actual injection rate of the current sample. This invention enables precise and adaptive injection rate control in HPLC sample pretreatment through viscosity feedforward.

Claims

1. A viscosity feedforward control method for injection rate in HPLC sample pretreatment, characterized in that, The method includes the following steps: Obtain the transmitted light intensity and corresponding pressure value for each group of blank samples and each group of experimental samples; the viscosity, hemoglobin concentration, number of microbubbles and particulate matter concentration of the experimental samples are all known. Particle scattering samples were selected from the experimental samples. The transmitted light intensity of each group of experimental samples was corrected by comparing the transmitted light intensity of blank samples with that of particle scattering samples, and the relative light intensity of each group of transmitted light intensity after color correction was determined. Based on the relative light intensity after color correction of all groups of transmitted light intensity in each experimental sample, and the fluctuation of the corresponding pressure value, the abnormality index of the transmitted light intensity of each group of experimental samples is determined, which is used to screen out several groups of retained transmitted light intensity in each experimental sample; based on the relative light intensity after color correction of each group of retained transmitted light intensity and the abnormality index, combined with the magnitude of particulate matter concentration, the true transmitted light intensity of each group of experimental samples is determined. Based on the viscosity and true transmitted light intensity of all experimental samples, a mapping model of true transmitted light intensity-viscosity is constructed; the true transmitted light intensity of the current sample is obtained, input into the mapping model of true transmitted light intensity-viscosity, and the viscosity of the current sample is output; based on the viscosity of the current sample, the actual aspirate volume and actual injection rate of the current sample are determined.

2. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 1, characterized in that, The specific steps involved in selecting the particle scattering samples from the experimental samples are as follows: All experimental samples with a hemoglobin concentration of 0, a microbubble count of 0, but a particulate matter concentration of 0 were recorded as particulate scattering samples.

3. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 1, characterized in that, The specific steps for determining the color-corrected relative light intensity of each group of transmitted light intensity for each group of experimental samples are as follows: Each group of transmitted light intensity includes: transmitted light intensity under green light and transmitted light intensity under near-infrared light; For blank samples, the mean value of the transmitted light intensity of all groups under green light is obtained and recorded as the baseline green light transmitted light intensity. The mean value of the transmitted light intensity of all groups under near-infrared light is obtained and recorded as the baseline near-infrared light transmitted light intensity. For each group of transmitted light intensities, the ratio of the transmitted light intensity of each experimental sample under green light to the transmitted light intensity of the reference green light is obtained, and recorded as the calibrated transmitted light intensity under green light; the ratio of the transmitted light intensity of each experimental sample under near-infrared light to the transmitted light intensity of the reference near-infrared light is obtained, and recorded as the calibrated transmitted light intensity under near-infrared light. Based on the calibration transmitted light intensity of the particle scattering sample under green light and near-infrared light, determine the particle scattering wavelength reference correction coefficient. The relative light intensity after color correction is determined based on the calibrated transmitted light intensity of each experimental sample under green light and near-infrared light, as well as the particle scattering wavelength reference correction coefficient.

4. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 3, characterized in that, The specific steps for determining the particle scattering wavelength reference correction coefficient are as follows: For each group of transmitted light intensities, the ratio of the calibrated transmitted light intensity under near-infrared light to the calibrated transmitted light intensity under green light for each group of particle scattering samples is obtained and denoted as the particle scattering wavelength correction coefficient. The mean value of the particle scattering wavelength correction coefficients corresponding to the transmitted light intensity of all groups of particle scattering samples is denoted as the particle scattering wavelength reference correction coefficient.

5. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 3, characterized in that, The specific steps for determining the color-corrected relative light intensity based on the calibrated transmitted light intensity of each experimental sample under green light and near-infrared light, and the particle scattering wavelength reference correction coefficient, are as follows: For each group of transmitted light intensities, the product of the ratio of the calibrated transmitted light intensity under green light to the calibrated transmitted light intensity under near-infrared light for each experimental sample and the particle scattering wavelength reference correction coefficient is obtained and recorded as the color-corrected relative light intensity.

6. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 1, characterized in that, The specific steps for determining the anomaly index of the transmitted light intensity for each group of experimental samples are as follows: For each experimental sample, a three-standard-deviation interval is constructed based on the mean and standard deviation of the color-corrected relative light intensity corresponding to all groups of transmitted light intensity. The deviation cutoff coefficient of the color-corrected relative light intensity corresponding to each group of transmitted light intensity based on the three-standard-deviation interval is denoted as the relative light intensity anomaly index in each group of transmitted light intensity. The upper limit cutoff coefficient of the absolute value of the relative deviation of the pressure values ​​corresponding to two adjacent groups of transmitted light intensity is calculated and denoted as the pipeline pressure anomaly index in each group of transmitted light intensity. Based on the preset relative light intensity anomaly weight and the preset pipeline pressure anomaly weight, the relative light intensity anomaly index and the pipeline pressure anomaly index in each group of transmitted light intensity are weighted and summed to obtain the anomaly index of each group of transmitted light intensity.

7. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 1, characterized in that, The specific steps involved in selecting several sets of retained transmitted light intensities for each group of experimental samples are as follows: If the abnormal index of the transmitted light intensity of each group of experimental samples is less than the preset abnormal threshold, it is recorded as the retained transmitted light intensity.

8. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 1, characterized in that, The specific steps for determining the true transmitted light intensity of each experimental sample are as follows: The window length for each experimental sample group is determined based on the particulate matter concentration; the particulate matter concentration is positively correlated with the window length. For each experimental sample, obtain the color-corrected relative light intensity sequence corresponding to the retained transmitted light intensity of all groups. Using the complement of the anomaly index of the retained transmitted light intensity of each group as the weight, perform a weighted moving average based on the window length to obtain the true transmitted light intensity sequence. The mean of the true transmitted light intensity sequence is recorded as the true transmitted light intensity of each experimental sample.

9. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 1, characterized in that, The specific steps involved in determining the actual aspirated volume and actual injection rate of the current sample are as follows: Based on the difference between the current sample viscosity and the preset standard viscosity, and in conjunction with the preset target injection volume, determine the actual amount of the current sample absorbed; The product of the ratio of the preset standard viscosity to the viscosity of the current sample and the preset standard injection rate is denoted as the actual injection rate of the current sample.

10. The viscosity feedforward control method for injection rate in HPLC sample pretreatment according to claim 9, characterized in that, The specific steps for determining the actual amount of sample drawn from the current sample based on the difference between the current sample viscosity and the preset standard viscosity, combined with the preset target injection volume, are as follows: The product of the difference between the current sample viscosity and the preset standard viscosity and the preset viscosity correction coefficient is recorded as the initial injection volume correction coefficient. The product of the sum of the initial injection volume correction coefficient and the preset reference constant and the preset target injection volume is recorded as the actual amount of sample absorbed.

Images