Online monitoring method and system for resin saturation of separation column

Abstract

The invention discloses a separation column resin saturation online monitoring method and system, and belongs to the technical field of medical instruments, and the method comprises the following steps: obtaining real-time acoustic impedance data of an inlet end and an outlet end of a separation column; calculating an instantaneous difference value of the real-time acoustic impedance data of the inlet end and the outlet end as a real-time acoustic impedance instantaneous difference value; measuring radial blood flow velocity distribution of the middle section of the separation column, and generating a flow field disturbance factor according to the radial blood flow velocity distribution; applying local pulse thrust to a resin bed layer in the separation column, and obtaining an instantaneous acoustic impedance difference value after disturbance; generating a relaxation recovery curve according to the change of the instantaneous acoustic impedance difference value along with time after disturbance; according to the method, the problem that the monitoring precision is insufficient due to the fact that an existing monitoring method cannot adapt to the non-Newtonian fluid characteristics of blood and dynamic fluctuation of a flow field interferes with a monitoring signal can be solved, and then the effects of eliminating non-adsorptive signal deviation caused by the fluctuation of the flow field and further improving the resin saturation monitoring precision are achieved.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, and more specifically, to a method and system for online monitoring of resin saturation in separation columns. Background Technology

[0002] In bedside blood purification therapy, the adsorption resin in the separation column is the component that removes harmful substances such as toxins and inflammatory mediators from the patient's blood. Its adsorption saturation directly determines the treatment effect and medical safety. If perfusion is not stopped in time after the resin reaches saturation, the adsorbed toxins will seep back into the patient's blood, causing serious complications. If it is stopped prematurely, the resin adsorption capacity will be wasted, increasing treatment costs and the burden on the patient. Therefore, accurate and real-time online monitoring of the resin saturation in the separation column is extremely important. Currently, most online monitoring methods for resin saturation are based on calibration models constructed with fixed flow rates and standard matrices. The principle is to indirectly determine the adsorption saturation state of the resin by detecting changes in the solute concentration at the outlet of the separation column. Such methods have certain applicability in laboratory simulation scenarios or standardized industrial separation processes and can meet basic monitoring needs. However, when applied to bedside blood purification scenarios, they have certain technical limitations and are difficult to adapt to the complex conditions of clinical treatment.

[0003] The characteristics of bedside blood purification include significant individual patient differences and real-time fluctuations in hemodynamics during treatment. On the one hand, different patients have different blood flow velocities and blood components, such as blood lipids, platelets, and plasma protein concentrations, resulting in varying physicochemical properties of the blood. On the other hand, blood is a typical non-Newtonian fluid, and its viscosity changes with shear rate. Affected by heartbeats, the flow field is constantly in dynamic fluctuation. Existing calibration models are built based on fixed flow rates and standard matrices, failing to consider the complexity of the aforementioned clinical scenarios. They cannot effectively distinguish between resin adsorption saturation and the detection signal shift caused by changes in the blood matrix and hemodynamic fluctuations, easily leading to saturation misjudgments. Misjudgments not only waste resin resources but may also cause toxin backflow due to resin oversaturation, posing serious medical safety risks to patients and contradicting the original clinical treatment purpose of bedside blood purification.

[0004] Furthermore, commonly used online monitoring methods in existing technologies, such as ultraviolet-visible spectroscopy and near-infrared spectroscopy, all require optical windows to be installed on blood vessels to detect light transmission and reflection. However, during the clinical use of medical devices, components such as fibrinogen and platelets in the blood are easily adsorbed and adhered to the surface of the optical window, gradually forming a dense biofilm. In severe cases, microthrombi may even form. Contaminants on the surface of the optical window can seriously interfere with the normal transmission and reflection of light, causing spectral signal distortion and rendering the resin saturation judgment based on spectral analysis completely ineffective. At the same time, due to the continuous requirements of bedside treatment, it is impossible to clean the optical window offline in a timely manner, while online cleaning has problems such as complex operation and poor cleaning effect. Summary of the Invention

[0005] To address the problems existing in the prior art, the present invention aims to provide an online monitoring method and system for the saturation of separation column resin, which can overcome the technical problem that existing monitoring methods do not consider common-mode interference caused by individual patient differences and hemodynamic fluctuations in bedside blood purification, and are prone to saturation misjudgment.

[0006] To solve the above problems, the present invention adopts the following technical solution:

[0007] The first aspect is a method and system for online monitoring of resin saturation in separation columns, including:

[0008] Obtain real-time acoustic impedance data at the inlet and outlet of the separation column;

[0009] Calculate the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet ends, and use it as the real-time acoustic impedance instantaneous difference value.

[0010] The radial blood flow velocity distribution in the middle section of the separation column is measured, and the flow field disturbance factor is generated based on the radial blood flow velocity distribution;

[0011] A local pulse thrust is applied to the resin bed in the separation column, and the instantaneous acoustic impedance difference after the disturbance is obtained;

[0012] A relaxation recovery curve is generated based on the change of the instantaneous acoustic impedance difference after disturbance over time, and relaxation characteristic values ​​are extracted from the relaxation recovery curve.

[0013] The instantaneous difference of real-time acoustic impedance is corrected by the flow field disturbance factor to obtain the corrected signal. The corrected signal is integrated and the integration result is compared with the relaxation characteristic value. Based on the comparison result, a saturation alarm signal is output.

[0014] Furthermore, real-time acoustic impedance data at the inlet and outlet ends of the separation column are obtained, including:

[0015] Broadband acoustic excitation is applied at the inlet and outlet of the separation column, the echo is received, the echo of the blood segment is extracted by gating through a time window, and the amplitude attenuation spectrum and phase shift spectrum of the blood in the broadband are obtained as the blood spectral fingerprint.

[0016] Based on the principal component resonance peak identified by blood spectral fingerprinting, the acoustic impedance at the inlet and outlet ends is measured using the frequency of this resonance peak to obtain the purification acoustic impedance change curve.

[0017] Furthermore, the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet is calculated, including:

[0018] The purification acoustic impedance change curve at the inlet end is time-delayed, with the delay time being the average passage time of blood flowing from the inlet end to the outlet end, generating a delayed waveform that includes rheological common-mode noise.

[0019] The purification acoustic impedance change curve at the outlet end is superimposed with the delayed waveform by acoustic phase cancellation, and the superimposed signal is extracted as the difference in pure adsorption contribution.

[0020] Furthermore, the radial blood flow velocity distribution in the middle section of the separation column is measured, and a flow field disturbance factor is generated based on the radial blood flow velocity distribution, including:

[0021] Multi-angle acoustic wave incidentness was applied to the middle section of the separation column to obtain Doppler frequency shift clusters of blood flow at different radial depths;

[0022] The Doppler frequency shift clusters are converted into equivalent acoustic impedance modulation coefficient sequences for each radial layer;

[0023] Extract the near-wall velocity gradient exponent from the equivalent acoustic impedance modulation coefficient sequence;

[0024] The near-wall velocity gradient exponent is modulated with the delay waveform to generate the flow field disturbance factor.

[0025] Furthermore, a localized pulse thrust is applied to the resin bed within the separation column, including:

[0026] Based on the difference signal of pure adsorption contribution and the flow field disturbance factor, a physical disturbance trigger pulse is generated;

[0027] Based on the physical disturbance trigger pulse, a multi-channel delayed trigger signal is generated using a geometric acoustic model of the separation column;

[0028] Local pulse thrust is applied to the resin bed based on multiple delayed trigger signals.

[0029] Furthermore, the instantaneous acoustic impedance difference after the disturbance is obtained, including:

[0030] After applying a local pulse thrust, acoustic impedance data at different time points after the disturbance are collected to obtain the acoustic impedance time series after the disturbance.

[0031] The difference between the time series of acoustic impedance after disturbance and the pure adsorption contribution before disturbance is calculated to obtain the net disturbance response curve.

[0032] The relaxation characteristic value is extracted from the net disturbance response curve and used as the characteristic value of the instantaneous acoustic impedance difference after the disturbance.

[0033] Furthermore, a relaxation recovery curve is generated based on the change of the instantaneous acoustic impedance difference over time after the disturbance, including:

[0034] Generate a set of time-off copies of the net disturbance response curve with different time delays;

[0035] The original net disturbance response curve is multiplied with each time offset replica to obtain the autocorrelation intensity signal corresponding to each delay time.

[0036] The autocorrelation intensity signals corresponding to each delay time are sorted according to the delay time and reconstructed into autocorrelation function curves.

[0037] Furthermore, relaxation eigenvalues ​​are extracted from the relaxation recovery curve, including:

[0038] Identify and extract the dominant relaxation time constant from the autocorrelation function curve;

[0039] The relaxation time constant and the initial amplitude of the net disturbance response curve are combined and encoded to generate relaxation eigenvalues.

[0040] Furthermore, the corrected signal is integrated, and the integration result is compared with the relaxation eigenvalue. Based on the comparison result, a saturation alarm signal is output, including:

[0041] The difference between the flow field disturbance factor and the pure adsorption contribution is acoustically amplitude modulated to generate a flow field normalized adsorption signal.

[0042] The normalized adsorption signal of the flow field is acoustically integrated to generate a characterization value of the total adsorption.

[0043] The total adsorption characterization value and the relaxation characteristic value are compared using acoustic dual thresholds, and a saturation alarm signal is output based on the comparison result.

[0044] Secondly, the present invention also provides an online monitoring system for the resin saturation of a separation column, comprising:

[0045] The data acquisition module is used to acquire real-time acoustic impedance data at the inlet and outlet ends of the separation column;

[0046] The calculation module is used to calculate the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet ends, which is used as the real-time acoustic impedance instantaneous difference value.

[0047] The disturbance generation module is used to measure the radial blood flow velocity distribution in the middle section of the separation column and generate a flow field disturbance factor based on the radial blood flow velocity distribution.

[0048] The disturbance acquisition module is used to apply a local pulse thrust to the resin bed in the separation column and acquire the instantaneous acoustic impedance difference after the disturbance.

[0049] The feature extraction module is used to generate a relaxation recovery curve based on the change of the instantaneous acoustic impedance difference after disturbance over time, and to extract relaxation feature values ​​from the relaxation recovery curve.

[0050] The saturation alarm module corrects the instantaneous difference of real-time acoustic impedance using the flow field disturbance factor to obtain a corrected signal. It then integrates the corrected signal and compares the integration result with the relaxation eigenvalue. Based on the comparison result, it outputs a saturation alarm signal.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0052] (1) This scheme uses the real-time acoustic impedance data of the inlet and outlet of the separation column and extracts the difference in pure adsorption contribution by means of time delay and phase cancellation superposition. This overcomes the technical problem that the existing monitoring method does not take into account the common mode interference caused by individual differences of patients and hemodynamic fluctuations in bedside blood purification, and is prone to saturation misjudgment. In this way, it achieves the effect of accurately extracting the acoustic impedance change caused only by resin adsorption and improving the accuracy of the basic data of resin saturation monitoring.

[0053] (2) This scheme adopts the technical means of applying multi-angle acoustic wave incident in the middle section of the separation column to obtain Doppler frequency shift clusters, thereby generating flow field disturbance factors and using them to correct the instantaneous difference of acoustic impedance. This overcomes the technical problem that the existing monitoring methods cannot adapt to the non-Newtonian fluid characteristics of blood and the interference of flow field dynamic fluctuations on the monitoring signal, resulting in insufficient monitoring accuracy. In this way, it achieves the effect of eliminating the non-adsorption signal deviation caused by flow field fluctuations and further improving the accuracy of resin saturation monitoring.

[0054] (3) This scheme adopts the technical means of applying local pulse thrust to the resin bed, obtaining the time series of acoustic impedance after disturbance and extracting relaxation characteristic values, which overcomes the technical problem that the acoustic impedance signal under steady state is easily affected by random noise and is difficult to directly characterize the adsorption saturation state of the resin itself, thereby achieving accurate reflection of the adsorption saturation degree from the dimension of the physical properties of resin particles.

[0055] (4) This scheme uses the technique of integrating the normalized adsorption signal to generate the total adsorption characterization value and comparing it with the relaxation characteristic value using acoustic double threshold to output an alarm signal. This overcomes the technical problem that the single signal judgment of the existing monitoring method is prone to misjudgment, which may lead to resin waste or toxin backflow. In this way, it achieves the technical effect of realizing the two-dimensional cross-verification of resin saturation, reducing the probability of misjudgment, and ensuring the safety and economy of bedside blood purification treatment. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0057] Figure 1 This is a flowchart of the online monitoring method for resin saturation in the separation column according to the present invention;

[0058] Figure 2 This is a data flow diagram between various modules in the online monitoring system for resin saturation of the separation column of the present invention. Detailed Implementation

[0059] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0060] Example 1

[0061] Please see Figure 1 A method for online monitoring of resin saturation in a separation column, comprising the following steps:

[0062] Step 1: Obtain real-time acoustic impedance data at the inlet and outlet of the separation column. The specific operation is as follows:

[0063] In bedside blood purification treatment, the inlet end of the separation column is the tubing section before the blood enters the resin adsorption bed, and the outlet end is the tubing section after the blood flows out of the resin adsorption bed. The acoustic impedance of the blood medium is determined by the density of the medium and the propagation speed of sound waves in the medium. Changes in the concentration of solutes such as toxins and inflammatory mediators to be adsorbed in the blood will directly change the density of the blood medium and the propagation speed of sound waves, thus causing corresponding changes in acoustic impedance. The acoustic impedance of the blood at the inlet end, which has not undergone resin adsorption treatment, can reflect the initial concentration of solutes to be adsorbed in the patient's blood. The acoustic impedance of the blood at the outlet end, which has undergone adsorption treatment in the resin bed, can reflect the concentration of remaining solutes in the blood after adsorption. The acquisition of real-time acoustic impedance data is carried out synchronously with the bedside blood purification treatment process. The time resolution of the acquisition is matched with the flow velocity of blood in the separation column tubing, ensuring that the acquired data can continuously and uninterruptedly reflect the acoustic state of the blood at the inlet and outlet ends at the corresponding moment. The acoustic impedance data acquisition at the inlet and outlet ends is synchronously triggered by the same clock source to ensure that the data acquired at both ends have a unified time reference.

[0064] Step 1 also includes the following steps:

[0065] Step 11: Apply broadband acoustic excitation to the inlet and outlet of the separation column, receive the echo, extract the echo of the blood segment through time window gating, and obtain the amplitude attenuation spectrum and phase shift spectrum of the blood in the broadband as the blood spectral fingerprint. The specific operation is as follows:

[0066] Broadband acoustic excitation is applied through ultrasonic transducers installed on the outer walls of the tubing at both the inlet and outlet ends of the separation column. An acoustic coupling medium is filled between the ultrasonic transducers and the tubing outer wall to reduce energy loss during sound wave propagation. The frequency coverage of the broadband acoustic excitation matches the acoustic response frequency bands of the solute to be adsorbed and the inherent components of the blood, ensuring effective acoustic interaction between the excitation sound waves and various components in the blood. Simultaneously, the ultrasonic transducers receive echo signals reflected by the tubing medium. These echo signals include reflections from the outer wall, inner wall, blood medium section, and other interfaces within the tubing. The reflection signals from different interfaces correspond to different propagation times due to the varying propagation path lengths of the sound waves. The time window gating operation is based on the propagation speed of the sound waves in the tubing material and the blood medium, combined with… The pipe wall thickness and the inner diameter of the blood segment are used to calculate the time interval corresponding to the propagation of sound waves in the blood medium segment. The start and end positions of the time window are set accordingly. Only the echo signal corresponding to the blood medium segment within the time window is extracted, excluding interference signals from pipe wall and other interface reflections. The extracted blood segment echo signal is subjected to Fourier transform to convert the time domain echo signal into a frequency domain signal, obtaining the amplitude and phase parameters of the echo signal at each frequency point in the broadband range. The initial spectrum of the broadband excitation signal is compared point by point with the spectrum of the blood segment echo signal to calculate the amplitude attenuation and phase shift corresponding to each frequency point in the broadband range. This forms the amplitude attenuation spectrum and phase shift spectrum of blood in the broadband range. This amplitude attenuation spectrum and phase shift spectrum can uniquely characterize the acoustic characteristics of the blood medium at the corresponding moment, serving as the blood spectral fingerprint.

[0067] Step 12: Based on the blood spectral fingerprint, identify the principal component resonance peak, and measure the acoustic impedance at the inlet and outlet ends using the resonance peak frequency to obtain the purification acoustic impedance change curve. The specific operation is as follows:

[0068] The blood spectral fingerprint, containing amplitude attenuation and phase shift spectra, reflects the acoustic response characteristics of different blood components at various frequencies across a wide bandwidth. The acoustic responses of different components exhibit superposition and enhancement at specific frequencies, forming resonance peaks. The identification of principal component resonance peaks is achieved by traversing the frequency domain data of the blood spectral fingerprint. This involves traversing all frequency points within the wide bandwidth, calculating the correlation coefficient between amplitude attenuation and phase shift at each frequency point, and selecting the frequency with the highest correlation coefficient and signal amplitude meeting a preset signal-to-noise ratio threshold as the frequency corresponding to the principal component resonance peak. At this frequency, the acoustic response sensitivity of the solute to be adsorbed in the blood is highest, and the common-mode interference from the inherent components of the blood is minimized. Based on the identified principal component resonance peak frequencies, the acoustic impedance of the blood at the inlet and outlet of the separation column is continuously measured. The acoustic impedance measurement is based on the reflection characteristics of ultrasound at the interface between the media. When ultrasound is incident perpendicularly onto the interface of two acoustic media, there is a definite mathematical relationship between the reflection coefficient and the characteristic acoustic impedance of the two media. Based on this relationship, the acoustic impedance of the blood medium can be calculated using the measured reflection coefficient.

[0069] This mathematical relationship is derived from the plane wave perpendicular incident boundary condition of the acoustic wave equation. When a plane ultrasound wave is perpendicularly incident on the interface of two homogeneous media, part of the incident wave's energy passes through the interface into the second medium, while the other part is reflected by the interface to form a reflected wave. Based on the continuous boundary condition of sound pressure and particle velocity at the interface, a quantitative relationship between the sound pressure reflection coefficient and the characteristic acoustic impedance of the two media can be derived. Combined with the pre-calibrated characteristic acoustic impedance of the pipeline material, the characteristic acoustic impedance of the blood medium can be calculated from the measured reflection coefficient. The corresponding formula is: ;

[0070] In the formula, The characteristic acoustic impedance of blood medium is expressed in Rayleigh (Ra). The characteristic acoustic impedance of the pipe wall material in the separation column is expressed in Ra, which is a pre-calibrated known constant. The acoustic pressure reflection coefficient at the principal component resonance frequency is calculated using the amplitude ratio and phase difference between the incident and reflected acoustic pressures at the same frequency. After measuring the acoustic impedance at each time point at the principal component resonance frequency, the acoustic impedance values ​​collected at multiple time points are sequentially arranged according to the order of collection time to form a purified acoustic impedance variation curve that continuously changes with treatment time. This curve can effectively filter out interference signals from non-target frequencies and stably reflect the changes in the acoustic impedance of the blood at the corresponding tubing end as the treatment progresses.

[0071] In a preferred embodiment of the present invention, step 2 is further included: calculating the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet ends, which is used as the real-time acoustic impedance instantaneous difference value. The specific operation is as follows:

[0072] During bedside blood purification treatment, fluctuations in the physicochemical properties of the patient's blood synchronously alter the acoustic impedance of the blood at the inlet and outlet of the separation column. These changes are unrelated to resin adsorption and constitute common-mode interference signals unrelated to adsorption behavior. The adsorption of toxins and inflammatory mediators in the blood by the resin only changes the acoustic impedance of the blood at the outlet and does not affect the acoustic impedance of the blood at the inlet. The difference in acoustic impedance between the inlet and outlet is solely due to resin adsorption and can be used to characterize the resin adsorption process. The calculation of the instantaneous difference in real-time acoustic impedance is based on the synchronously acquired real-time acoustic impedance data at the inlet and outlet. Through corresponding calculations, common-mode interference caused by changes in the blood's own rheology is eliminated, and the amount of acoustic impedance change caused solely by resin adsorption is accurately extracted, providing an effective signal basis after removing interference for subsequent determination of resin saturation.

[0073] Step 2 also includes the following steps:

[0074] Step 21: Time-delay the purification acoustic impedance change curve at the inlet end. The delay time is the average transit time of blood flowing from the inlet end to the outlet end, generating a delayed waveform that includes rheological common-mode noise. The specific operation is as follows:

[0075] When blood flows within the flow channel of the separation column, it takes a certain amount of time to travel from the inlet acoustic impedance measurement point to the outlet acoustic impedance measurement point. After the same blood clump completes acoustic impedance data acquisition at the inlet, it needs this time to reach the measurement point at the outlet. There is a time offset between the purified acoustic impedance change curves at the inlet and outlet, which is related to blood flow. The average transit time of blood from the inlet to the outlet is calculated based on the total flow channel volume between the inlet and outlet measurement points of the separation column and the real-time average blood flow rate. The total flow channel volume is an inherent geometric parameter of the separation column, which can be obtained through pre-calibration. It includes the pipe volume from the inlet measurement point to the resin bed inlet, the effective flow volume of the resin bed, and the pipe volume from the resin bed outlet to the outlet measurement point. The real-time average blood flow rate is obtained from the real-time feedback data of the blood pump in the bedside blood purification equipment. The average transit time is the sum of the total flow channel volume and the real-time average flow rate. The ratio; the time delay operation shifts the purified acoustic impedance change curve at the inlet end backward along the time axis. The shift duration is consistent with the calculated average throughput time. The shifted curve and the purified acoustic impedance change curve at the outlet end are on the same time reference. The two sets of curve data corresponding to the same time point come from the measurement results of the same blood clot at the inlet and outlet ends respectively. The delayed waveform obtained after the shift completely preserves the acoustic impedance fluctuations caused by the rheological changes of the patient's blood. These fluctuations are rheological common-mode noise, which has a high degree of consistency in the measurement results at the inlet and outlet ends of the same blood clot.

[0076] Step 22: The purification acoustic impedance change curve at the outlet end is superimposed with the delayed waveform by acoustic phase cancellation, and the superimposed signal is extracted as the difference in purification adsorption contribution. The specific operation is as follows:

[0077] The purified acoustic impedance change curve at the outlet is composed of two superimposed signals: one is the rheological common-mode noise signal consistent with the delayed waveform, and the other is the acoustic impedance change signal caused by resin adsorption. This latter signal exists only in the measurement results at the outlet and will not appear in the delayed waveform. The acoustic wave phase cancellation superposition operation is based on the phase and amplitude consistency characteristics of the common-mode signal. The purified acoustic impedance change curve at the outlet and the delayed waveform are superimposed point by point in opposite phase. In the two sets of waveforms corresponding to the same time point, the rheological common-mode noise signals with consistent phase and amplitude will cancel each other out and exist only in the waveform at the outlet. The adsorption contribution signal is completely preserved; the purified acoustic impedance measured at the inlet of the same blood clot contains only the acoustic impedance component caused by the rheological properties of the patient's blood itself, which is the rheological common-mode noise. After the blood clot flows through the resin bed, the purified acoustic impedance measured at the outlet contains the same rheological common-mode noise component as at the inlet, as well as the acoustic impedance change component caused by resin adsorption. After aligning the time references of the inlet curve and the outlet curve through a time delay operation, the difference between the two can cancel out the common-mode noise component, retaining only the adsorption contribution component, thus obtaining the quantitative calculation as follows: ;

[0078] In the formula, The difference in pure adsorption at the corresponding time point is expressed in Rayles. The values ​​are the output acoustic impedance change curves at the same time point, in Rayles. The value of the inlet purification acoustic impedance change curve after time delay at the same time point is expressed in Rayles. The average transit time of blood from the inlet to the outlet is expressed in seconds. The pure adsorption contribution difference obtained after superposition calculation completely removes the common-mode interference caused by the rheological fluctuations and hemodynamic changes of the patient's blood, and only retains the difference in acoustic impedance between the inlet and outlet caused by the adsorption of solutes in the blood by the resin. It can directly characterize the adsorption intensity of the resin at the corresponding time point.

[0079] In a preferred embodiment of the present invention, step 3 is further included: measuring the radial blood flow velocity distribution in the middle section of the separation column, and generating a flow field disturbance factor based on the radial blood flow velocity distribution. The specific operation is as follows:

[0080] During bedside blood purification, the blood flow field within the separation column is constantly in a dynamic fluctuation state due to the patient's heartbeat, changes in vascular compliance, and the non-Newtonian fluid properties of blood. Changes in the radial velocity distribution of the flow field alter the contact state between blood and resin particles, and simultaneously modulate the sound wave propagation path and energy attenuation during acoustic impedance measurement, thereby causing a non-adsorption shift in the acoustic impedance measurement signal. The middle section of the separation column is the main section of the resin bed and the area where sufficient contact between blood and the resin adsorption medium occurs, and where solute adsorption behavior mainly takes place. The radial blood flow velocity distribution in this region can directly reflect the real-time state of the flow field within the separation column. The measurement of the radial blood flow velocity distribution covers the entire radial depth of the separation column tube cross-section from the central axis to the tube wall, and can completely characterize the differences in blood flow velocity at different radial positions. The flow field disturbance factor is generated based on the characteristic parameters of the radial blood flow velocity distribution, and can quantitatively characterize the degree of influence of the dynamic fluctuation of the flow field on the acoustic impedance measurement signal and resin adsorption behavior.

[0081] Step 3 also includes the following steps:

[0082] Step 31: Apply multi-angle acoustic wave incidentness to the middle section of the separation column to obtain Doppler frequency shift clusters of blood flow at different radial depths. The specific operation is as follows:

[0083] Multiple sets of ultrasonic transducers are arranged circumferentially in the middle section of the separation column. The installation position and sound wave emission direction of each set of ultrasonic transducers are pre-calibrated, allowing sound wave excitation at different incident angles to be applied to the blood flow medium inside the separation column. The multi-angle incident sound wave angle range covers the entire radial depth of the separation column cross-section, and sound waves at different incident angles can reach blood flow regions at different radial positions within the cross-section, achieving full radial coverage of the blood flow state. The flowing blood contains a large number of blood cells with acoustic scattering properties; the interaction between the incident sound waves and the flowing blood cells generates… The frequency of the scattered echo differs from that of the incident sound wave; this frequency difference is known as the Doppler frequency shift. Blood flow at different radial depths has different flow velocities, corresponding to different Doppler frequency shift values. Sound wave excitation at different incident angles corresponds to the acquisition of Doppler frequency shift signals at different radial depths. All Doppler frequency shift signals corresponding to the incident angles are arranged and combined in order of radial depth to form a Doppler frequency shift cluster covering the entire radial range. This frequency shift cluster can completely characterize the blood flow motion state at different radial positions within the cross-section of the middle section of the separation column.

[0084] The quantitative relationship between Doppler frequency shift and blood flow velocity is derived from the basic principle of the acoustic Doppler effect. When an incident sound wave irradiates flowing blood cells at a fixed angle, the blood cells, as moving scatterers, first receive the incident sound wave, resulting in a first frequency shift, and then scatter the sound wave back to the transducer, resulting in a second frequency shift. The superposition of the two frequency shifts is the finally detected Doppler frequency shift. Based on the basic laws of sound wave propagation and the Doppler effect of moving objects, a quantitative expression for the relationship between Doppler frequency shift and blood flow velocity can be derived, specifically: ;

[0085] In the formula, The measured Doppler frequency shift is expressed in Hertz. denoted as the center frequency of the incident sound wave, in Hertz; v is the blood flow velocity at the corresponding radial position, in meters per second; θ is the angle between the incident direction of the sound wave and the direction of blood flow, in degrees; c is the speed of sound in human blood, in meters per second, which is a pre-calibrated fixed constant.

[0086] Step 32: Convert the Doppler frequency shift clusters into equivalent acoustic impedance modulation coefficient sequences for each radial layer. The specific operations are as follows:

[0087] There is a definite acoustic correspondence between the Doppler frequency shift value and the blood flow velocity at the corresponding radial position. Based on this correspondence, each frequency shift value within the Doppler frequency shift cluster can be converted into the blood flow velocity at the corresponding radial depth, thus obtaining the blood flow velocity distribution of each radial layer within the cross-section of the middle section of the separation column. Changes in blood flow velocity alter the propagation path and energy attenuation characteristics of sound waves within the blood medium, thereby modulating the sound wave reflection and transmission behavior during the acoustic impedance measurement process. The intensity of this modulation is quantitatively correlated with the blood flow velocity value. Based on pre-completed acoustic calibration experiments, a correspondence between blood flow velocity and acoustic impedance measurement modulation intensity can be established. This correspondence can convert the blood flow velocity values ​​of each radial layer into the equivalent acoustic impedance modulation coefficient of the corresponding layer. The equivalent acoustic impedance modulation coefficient can quantitatively characterize the modulation amplitude of the acoustic impedance measurement signal caused by the change in blood flow velocity at the corresponding radial layer. Arranging the equivalent acoustic impedance modulation coefficients of each radial layer in order from the center of the pipe to the pipe wall according to the radial depth forms an equivalent acoustic impedance modulation coefficient sequence. This sequence can completely characterize the distribution of the modulation effect of flow field changes on acoustic impedance measurement within the entire radial range.

[0088] Step 33: Extract the near-wall velocity gradient exponent from the equivalent acoustic impedance modulation coefficient sequence. The specific operation is as follows:

[0089] Blood flow within the separation column is affected by the viscous resistance of the tube wall. The velocity variation near the wall is much greater than that in the central region of the tube, and the velocity variation characteristics near the wall directly reflect the overall disturbance of the flow field. The equivalent acoustic impedance modulation coefficient sequence corresponds one-to-one with the blood velocity distribution of each radial layer. The rate of change of blood velocity at the corresponding radial depth with radial position can be calculated from this sequence, and this rate of change is the velocity gradient. The range of the near-wall region is predetermined based on the inner diameter of the separation column tube, covering an annular area from the inner wall of the tube to a preset radial depth. The radial range of this region can be set by a fixed proportion of the inner diameter of the tube. The sequence segment corresponding to the near-wall region is extracted from the equivalent acoustic impedance modulation coefficient sequence, and the velocity gradient at each radial position within the near-wall region is calculated based on this sequence segment. The velocity gradient within this region is statistically calculated to obtain the near-wall velocity gradient index. The near-wall velocity gradient index can quantitatively characterize the intensity of blood velocity variation in the near-wall region, thereby reflecting the overall disturbance intensity of the flow field within the separation column. The magnitude of the index value is positively correlated with the degree of flow field disturbance.

[0090] Step 34: Modulate the near-wall velocity gradient exponent with the delay waveform to generate the flow field disturbance factor. The specific operation is as follows:

[0091] The delayed waveform is obtained by time-delaying the purified acoustic impedance change curve at the inlet end. This waveform changes continuously with the treatment time, fully reflecting the real-time fluctuations of the patient's blood rheological properties and exhibiting temporal synchronization with the dynamic changes of the flow field within the separation column. The near-wall velocity gradient index is a quantitative parameter characterizing the intensity of flow field disturbance. It can be used for acoustic amplitude modulation calculation with the delayed waveform. During the modulation calculation, the near-wall velocity gradient index serves as a modulation depth parameter to normalize the real-time amplitude of the delayed waveform. The modulation calculation is performed using a unified time reference. The near-wall velocity gradient index and the amplitude of the delayed waveform at the same time point participate in the modulation calculation at the corresponding moment, resulting in a modulated output signal that changes continuously with the treatment time. This modulated output signal can simultaneously reflect the dynamic disturbance intensity of the flow field and the real-time changes in blood rheological properties. After normalization, a flow field disturbance factor is generated. The flow field disturbance factor is continuously updated with the treatment time, quantitatively characterizing the combined influence of flow field fluctuations and blood rheological changes on the acoustic impedance measurement signal at the corresponding moment, providing a quantitative correction parameter for subsequent correction of the instantaneous difference in acoustic impedance.

[0092] The quantitative relationship of acoustic amplitude modulation is derived from the basic principles of signal modulation. The core logic of amplitude modulation is to synchronously adjust the output amplitude of the carrier signal by the real-time changes in the modulating signal, so that the output signal simultaneously carries all the information of both the carrier and modulating signals. Here, the delayed waveform serves as the carrier signal, carrying information on the continuous changes in blood rheological properties, while the near-wall velocity gradient exponent serves as the modulating signal, carrying information on the real-time intensity of flow field disturbances. Amplitude modulation is achieved through the multiplication of these two signals, and the corresponding quantitative expression is: ;

[0093] In the formula, M(t) is the modulated output signal at the corresponding time, in Rayleigh; k(t) is the near-wall velocity gradient exponent at the corresponding time, which is a dimensionless parameter; and W(t) is the amplitude of the delayed waveform at the corresponding time, in Rayleigh.

[0094] In a preferred embodiment of the present invention, step 4 is further included: applying a local pulse thrust to the resin bed in the separation column and obtaining the instantaneous acoustic impedance difference after disturbance. The specific operation is as follows:

[0095] The adsorption saturation state of the resin bed alters the physical properties of the resin particles, including density, pore filling state, and surface mechanical properties. The steady-state acoustic impedance difference signal can reflect the resin adsorption process, but it is susceptible to residual interference from blood rheological properties and flow field fluctuations. By applying a local pulse thrust to the resin bed, the resin particles can generate a transient mechanical response. The relaxation and recovery process of this response is only related to the adsorption saturation state of the resin particles themselves and is not affected by steady-state changes in the blood matrix. The transient acoustic impedance difference after disturbance can completely record the transient acoustic changes of the resin bed after disturbance. The characteristic parameters extracted from this can directly characterize the degree of resin adsorption saturation, complementing the steady-state acoustic impedance signal and further eliminating signal interference caused by non-adsorption factors, providing direct characteristic basis for accurate determination of resin saturation.

[0096] Step 4 also includes the following steps:

[0097] Step 41: Based on the difference signal of pure adsorption contribution and the flow field disturbance factor, generate a physical disturbance trigger pulse. The specific operation is as follows:

[0098] The pure adsorption contribution difference signal can continuously characterize the real-time adsorption process of the resin bed, and the flow field disturbance factor can continuously characterize the real-time stability of the flow field within the separation column. The two signals have a unified time reference and can be used synchronously to generate physical disturbance trigger pulses. During the generation process, the two continuous signals are first time-aligned to ensure that the two sets of data at the same time point correspond to the same working state of the separation column. Then, the trigger judgment logic is set. When the cumulative change of the pure adsorption contribution difference signal reaches the preset trigger threshold and the real-time value of the flow field disturbance factor is within the preset stable range, a physical disturbance trigger pulse is generated. The timing parameters of the trigger pulse include the rise time, pulse width, and fall time, all of which are preset based on the flow channel geometry parameters of the separation column and the rheological characteristics of the blood. This ensures that the trigger pulse can accurately trigger the subsequent generation of multiple signals and the application of thrust, while not having a continuous impact on the structure of the resin bed and the flow state of the blood.

[0099] Step 42: Based on the physical disturbance trigger pulse, generate multiple delayed trigger signals based on the geometric acoustic model of the separation column. The specific operation is as follows:

[0100] The geometric acoustic model of the separation column is pre-established based on the inherent geometry of the separation column, the spatial distribution parameters of the resin bed, and the installation positions and acoustic characteristics of the circumferentially arranged ultrasonic transducers. It can accurately describe the propagation path, propagation time, and coherent superposition law of sound waves in the medium inside the separation column. The physical disturbance trigger pulse is used as a synchronous start signal input to the geometric acoustic model. The model calculates the sound wave propagation delay corresponding to each ultrasonic transducer according to the spatial position of the preset target disturbance area in the resin bed. This delay is the time difference required for the sound waves emitted by each transducer to reach the target disturbance area. Based on the calculated delay values, multiple delay trigger signals corresponding to the number of transducers are generated. The trigger time difference of each trigger signal is perfectly matched with the delay value calculated by the model, ensuring that the sound waves emitted by each transducer can achieve in-phase coherent superposition within the target disturbance area, forming a concentrated local force, while avoiding the generation of unnecessary acoustic disturbances in non-target areas.

[0101] Step 43: Apply local pulse thrust to the resin bed based on the multi-channel delayed trigger signal. The specific operation is as follows:

[0102] Multiple ultrasonic transducers arranged circumferentially in the separation column correspond one-to-one with multiple delayed trigger signals. Upon receiving the corresponding delayed trigger signal, each transducer immediately outputs a high-energy pulsed sound wave with a preset frequency and amplitude. The pulsed sound waves output by the multiple transducers are emitted sequentially according to the preset delay sequence, and undergo acoustic coherent superposition in the target disturbance area of ​​the resin bed within the separation column. The superimposed sound wave energy is concentrated in the target local area, forming a directional acoustic radiation force, i.e., local pulse thrust. The duration of the local pulse thrust is consistent with the pulse width of the pulsed sound wave, and its range of action is limited to the preset local area of ​​the resin bed. It will not cause large-scale disturbance to the entire resin bed, nor will it affect the normal flow of blood within the separation column. The local pulse thrust acts on the resin particles in the target area, causing the resin particles to produce tiny transient displacements, changing the gaps between the resin particles and the contact state between the resin and the surrounding blood medium. This leads to transient changes in the acoustic impedance characteristics of the local area of ​​the resin bed. These transient changes are directly related to the adsorption saturation state of the resin particles, which can provide a basis for subsequent acoustic impedance data acquisition and feature extraction.

[0103] Step 44: After applying the local pulse thrust, collect acoustic impedance data at different time points after the disturbance to obtain the acoustic impedance time series after the disturbance. The specific operation is as follows:

[0104] The acquisition of acoustic impedance data is strictly synchronized with the application of local pulse thrust. The moment the local pulse thrust is applied is taken as the zero point of the acquisition time sequence. Acoustic impedance data at the inlet and outlet of the separation column are continuously acquired at preset fixed time intervals. The acquisition frequency and total duration are preset based on the recovery characteristics of the resin bed after disturbance, ensuring complete coverage of the entire process from disturbance to recovery to the initial steady state of the resin bed. During the acquisition process, the acoustic impedance data acquisition at the inlet and outlet ends are synchronously triggered by the same clock source to ensure that the data acquired at both ends have a unified time reference. The acoustic impedance data at the inlet and outlet ends corresponding to each time point are acquired synchronously. The acoustic impedance data at the inlet and outlet ends acquired at different time points are arranged in chronological order of acquisition time to form a time series of acoustic impedance after disturbance that changes continuously with time. This series completely records the entire process of transient changes in the acoustic impedance of the blood at the inlet and outlet ends after the resin bed is disturbed by local pulse thrust.

[0105] Step 45: Difference the time series of acoustic impedance after perturbation with the difference between the pure adsorption contribution before perturbation to obtain the net perturbation response curve. The specific operation is as follows:

[0106] The difference in pure adsorption contribution before disturbance is the steady-state pure adsorption contribution difference calculated through the aforementioned steps before the application of local pulse thrust. This value characterizes the steady-state adsorption state of the resin bed when undisturbed, excluding the transient change component caused by disturbance. Each time point in the acoustic impedance time series after disturbance corresponds to a set of synchronously acquired inlet and outlet acoustic impedance data. Following the same calculation logic as in the aforementioned steps, the instantaneous acoustic impedance difference after disturbance corresponding to each time point is calculated, forming a time series of instantaneous acoustic impedance differences after disturbance. The value corresponding to each time point in this time series is differentially analyzed point-by-point with the steady-state pure adsorption contribution difference before disturbance to obtain the net disturbance response value corresponding to each time point. The net disturbance response values ​​corresponding to all time points are arranged in chronological order to form a net disturbance response curve that changes continuously with time. This curve completely removes the acoustic impedance difference component caused by the steady-state adsorption state of the resin bed, retaining only the transient acoustic impedance change component caused by the local pulse thrust disturbance, and can completely characterize the transient response and recovery process of the resin bed after disturbance.

[0107] The net disturbance response is the separation of acoustic impedance changes caused solely by the pulse thrust disturbance. It requires eliminating the steady-state adsorption contribution difference basis. The instantaneous acoustic impedance difference after disturbance consists of two parts: the steady-state basis and the transient disturbance component. Through differential operations on these two parts, the steady-state basis can be directly canceled, yielding the net response caused solely by the disturbance. Based on the principle of linear superposition of time-series signals and the rules of differential operations, a corresponding quantitative formula can be established, specifically: ;

[0108] In the formula, The value represents the net disturbance response at the corresponding time point, in Rayleigh units. The instantaneous acoustic impedance difference after the disturbance at the corresponding moment is expressed in Rayles. The difference in contribution to the steady-state pure adsorption before the disturbance is expressed in Rayles; t is the time variable with the completion time of the local pulse thrust application as zero, expressed in seconds.

[0109] Step 46: Extract the relaxation characteristic value from the net disturbance response curve as the characteristic value of the instantaneous acoustic impedance difference after the disturbance. The specific operation is as follows:

[0110] The net disturbance response curve characterizes the relaxation process of the acoustic impedance difference in the resin bed after disturbance, from transient change to steady state. The characteristics of this relaxation process are directly related to the adsorption saturation state of the resin particles. The more solute adsorbed by the resin particles and the higher the saturation, the changes in the density of the resin particles, the medium characteristics within the pores, and the interaction forces between particles will occur accordingly, resulting in corresponding differences in relaxation characteristics such as the recovery rate and the amplitude of the transient response after disturbance. The process of extracting relaxation feature values ​​from the net disturbance response curve involves first smoothing the curve to remove high-frequency noise introduced during the acquisition process, and then identifying key feature points of the curve, including the initial peak amplitude after disturbance, the time corresponding to the peak, and the time corresponding to different proportions of curve decay to the steady-state value. Based on these key feature points, feature parameters that can fully characterize the relaxation process are extracted. These feature parameters are combined and encoded to generate relaxation feature values. These relaxation feature values ​​can quantitatively characterize the relaxation recovery characteristics of the resin bed after disturbance and directly reflect the adsorption saturation state of the resin.

[0111] In a preferred embodiment of the present invention, step 5 is further included: generating a relaxation recovery curve based on the change of the instantaneous acoustic impedance difference after disturbance over time, and extracting relaxation feature values ​​from the relaxation recovery curve. The specific operation is as follows:

[0112] The transient acoustic impedance change generated by the resin bed under local pulse thrust will gradually recover to the steady state before the disturbance over time. The pattern of this recovery process is directly related to the adsorption saturation state of the resin particles. The higher the solute content adsorbed by the resin particles, the more obvious the change in the coupling state between the particles' physical properties and the surrounding blood medium, and the more prominent the difference in the recovery process after disturbance. The net disturbance response curve fully records the entire process of transient acoustic impedance change after disturbance, but the curve inevitably contains blood flow micro-disturbances and random noise introduced by signal acquisition. Directly extracting features from the original curve is prone to deviation. By processing the net disturbance response curve through autocorrelation operation, a relaxation recovery curve that characterizes the inherent recovery law of the signal can be generated. This curve can effectively suppress random noise and enhance the relaxation characteristics directly related to the resin adsorption state. The relaxation feature values ​​extracted from the relaxation recovery curve can stably and accurately characterize the adsorption saturation degree of the resin bed.

[0113] Step 5 also includes the following steps:

[0114] Step 51: Generate a set of time-off copies of the net disturbance response curve with different time delays. The specific operations are as follows:

[0115] The net disturbance response curve is a discrete time-series signal sampled at fixed time intervals. The signal duration covers the complete process of the resin bed from disturbance to recovery to steady state. Each sampling point corresponds to a unique time coordinate and acoustic impedance difference value. The generation of time-offset replicas is based on the original net disturbance response curve. First, a set of delayed time sequences covering the complete recovery process is set. The interval between adjacent delayed times in the sequence is consistent with the sampling time interval of the original curve, ensuring that all replicas are completely aligned with the sampling points of the original curve. For each delayed time in the sequence, the original net disturbance response curve is shifted backward along the time axis by a duration equal to that delayed time. The portion exceeding the time range of the original curve after shifting and the portion before the start of the time axis are filled with zero values, so that the shifted curve maintains the same time length and number of sampling points as the original curve. All the shifted curves are classified according to their corresponding delayed times, forming a set of time-offset replicas with different time delays.

[0116] Step 52: Multiply the original net disturbance response curve with each time offset replica to obtain the autocorrelation intensity signal corresponding to each delay time. The specific operation is as follows:

[0117] The autocorrelation operation of time-series signals can characterize the similarity of the signals themselves at different time offsets. The inherent relaxation recovery characteristics of the resin bed will show a stable and regular similarity at different time offsets, while the random noise introduced during the acquisition process does not have this kind of regularity. Through corresponding operations, the effective signal of relaxation correlation can be effectively enhanced and the irregular random noise can be suppressed. In the operation process, the original net disturbance response curve and each time offset replica are multiplied point by point along a unified time axis. The original curve value and the replica value corresponding to the same time coordinate are multiplied to obtain the product result of that coordinate. Then, the product results corresponding to all time coordinates are summed to obtain the autocorrelation intensity signal of the time offset replica corresponding to the delay time. Each time offset replica corresponds to a unique delay time and also corresponds to a unique autocorrelation intensity signal. The autocorrelation intensity signals corresponding to all delay times completely characterize the self-similarity of the net disturbance response curve at different time offsets.

[0118] This computation is derived from the basic definition of autocorrelation in discrete-time signals. The autocorrelation function of a discrete-time series describes the similarity of the sequences at different time delays. Essentially, it is the summation of the product of the original sequence and the delayed sequence over the entire time interval. For the net disturbance response curve in this scheme, this computation can quantify the result of point-by-point multiplication into the autocorrelation strength corresponding to the delay time, effectively eliminating the influence of random noise, enhancing the inherent relaxation characteristics of the signal, and ultimately forming a quantitative expression suitable for this application scenario. Specifically: ;

[0119] In the formula, R is the autocorrelation intensity signal corresponding to the delay time, and the unit is Rayleigh squared; The time offset replica used in the calculation corresponds to the delay time in seconds; x is the time series value of the original net disturbance response curve in Rayleigh. is the time coordinate variable, in seconds; N is the total number of sampling points of the original net disturbance response curve, a dimensionless integer constant.

[0120] Step 53: Sort the autocorrelation intensity signals corresponding to each delay time according to the delay time, and reconstruct them into autocorrelation function curves. The specific operation is as follows:

[0121] Each autocorrelation intensity signal corresponds one-to-one with a unique delay time value. All autocorrelation intensity signals are arranged in ascending order of their corresponding delay times, forming a discrete data point sequence with delay time as the horizontal axis and autocorrelation intensity as the vertical axis. This discrete data point sequence is then subjected to smooth interpolation, with the interpolation interval consistent with the sampling interval of the original net disturbance response curve. This ensures that the reconstructed curve has a time resolution that matches the original signal. After interpolation, a continuous autocorrelation function curve is formed. This curve fully characterizes the inherent relaxation characteristics of the net disturbance response curve and corresponds to the entire relaxation recovery process of the resin bed after disturbance. Therefore, it is used as a relaxation recovery curve. The variation of this curve is only related to the adsorption saturation state of the resin bed and is not affected by random noise or blood flow micro-disturbances during the acquisition process.

[0122] Step 54: Identify and extract the dominant relaxation time constant from the autocorrelation function curve. The specific operation is as follows:

[0123] The decay process of the autocorrelation function curve corresponds to the energy decay characteristics of the relaxation recovery process of the resin bed. This decay process can be composed of multiple exponential decay terms superimposed. The time constant corresponding to the exponential decay term with the largest weight can determine the main rate of change in the relaxation recovery process. This time constant is the dominant relaxation time constant. The dominant relaxation time constant is directly related to the adsorption saturation state of the resin particles. The higher the solute content adsorbed by the resin, the higher the saturation, the stronger the mechanical coupling characteristics between the resin particles and the surrounding blood medium, the slower the recovery rate after disturbance, and the larger the corresponding dominant relaxation time constant value. In the identification and extraction process, the autocorrelation function curve is first subjected to multi-exponential fitting operation. The fitting interval covers the complete range of the autocorrelation function curve from the peak decay to near zero. The fitting parameters are optimized by least squares method to minimize the sum of squared residuals between the fitted curve and the measured autocorrelation function curve. From the multiple exponential decay terms obtained by fitting, the term with the largest weight is extracted. The time constant corresponding to this term is the final dominant relaxation time constant. This constant can quantitatively characterize the main rate of the resin bed recovering to steady state after disturbance and directly reflect the degree of resin adsorption saturation.

[0124] Step 55: Encode the combination of the dominant relaxation time constant and the initial amplitude of the net disturbance response curve to generate relaxation eigenvalues. The specific operation is as follows:

[0125] The initial amplitude of the net disturbance response curve is the value corresponding to the first peak reached by the curve after the application of the local pulse thrust. This amplitude directly characterizes the transient response intensity of the resin bed after being disturbed by the local pulse thrust. The higher the resin saturation, the more significant the change in particle density and mechanical properties, and the higher the transient response amplitude after disturbance. Therefore, the initial amplitude can directly reflect the relevant information of the resin adsorption saturation state. In the combined encoding process, the dominant relaxation time constant and the initial amplitude of the net disturbance response curve are first normalized. The normalization reference value is pre-set based on the calibration data of the initial unloaded state and the fully saturated state of the resin bed in the separation column, so that the numerical range of the two parameters is unified within the same interval. Then, the two normalized parameters are linearly combined according to the preset encoding rules to generate a unique corresponding relaxation characteristic value. This relaxation characteristic value simultaneously includes the rate characteristics of the resin bed relaxation recovery and the intensity characteristics of the transient response, which can comprehensively and uniquely characterize the adsorption saturation state of the resin bed.

[0126] In a preferred embodiment of the present invention, step 6 is further included: correcting the instantaneous difference of real-time acoustic impedance using a flow field disturbance factor to obtain a corrected signal; integrating the corrected signal; comparing the integration result with the relaxation characteristic value; and outputting a saturation alarm signal based on the comparison result. The specific operation is as follows:

[0127] The real-time acoustic impedance instantaneous difference reflects the difference in acoustic impedance between the blood at the inlet and outlet of the separation column caused by resin adsorption. This signal has eliminated common-mode interference caused by changes in the rheological properties of the blood itself, but it is still affected by residual amplitude modulation caused by dynamic fluctuations in the flow field within the separation column. Directly using it to determine resin saturation will result in a deviation from the actual adsorption state. The flow field disturbance factor is generated based on the radial blood flow velocity distribution in the middle section of the separation column. It can quantitatively characterize the degree of modulation of the acoustic impedance measurement signal by flow field fluctuations. It can be used to normalize and correct the real-time acoustic impedance instantaneous difference, eliminate non-adsorption signal deviations caused by flow field fluctuations, and obtain a corrected signal that is only directly related to the resin adsorption behavior. The integral operation of the corrected signal can accumulate the time-varying instantaneous adsorption intensity into the total adsorption of the resin bed. The system employs a quantitative characterization value, which reflects the resin saturation level from the perspective of steady-state adsorption process. The relaxation characteristic value, generated based on the relaxation recovery characteristics of the resin bed after disturbance, reflects the resin's adsorption saturation state from the perspective of changes in the physical properties of the resin particles themselves, unaffected by steady-state changes in the blood matrix and flow field. By comparing the total adsorption characterization value and the relaxation characteristic value using a dual-threshold comparison, the resin saturation can be cross-validated from both steady-state adsorption accumulation and transient material properties dimensions. This avoids misjudgments caused by interference from a single signal, preventing waste of resin adsorption capacity due to premature treatment termination and avoiding the risk of toxin backflow caused by resin oversaturation. Finally, based on the comparison results, a corresponding saturation alarm signal is output, providing a reliable basis for the process control of bedside blood purification treatment.

[0128] Step 6 also includes the following steps:

[0129] Step 61: The difference between the flow field disturbance factor and the pure adsorption contribution is acoustically amplitude modulated to generate a normalized adsorption signal. The specific operation is as follows:

[0130] The pure adsorption contribution difference is the instantaneous adsorption intensity signal after eliminating blood rheological common-mode noise. The amplitude of this signal is linearly related to the total amount of solute adsorbed by the resin per unit time. However, dynamic fluctuations in the flow field within the separation column alter the contact efficiency between blood and resin particles, and simultaneously modulate the acoustic impedance measurement signal, leading to a deviation between the measured amplitude and the actual adsorption intensity of the resin. The flow field disturbance factor quantitatively characterizes the modulation amplitude of the acoustic impedance signal by flow field fluctuations and shares a unified time reference with the pure adsorption contribution difference. The acoustic amplitude modulation calculation of both can realize the pure adsorption contribution difference. The normalization correction eliminates the amplitude deviation caused by flow field fluctuations. The measured difference in pure adsorption contribution is a linear superposition of the true adsorption contribution component of the resin and the deviation component caused by flow field modulation. The flow field disturbance factor can quantitatively characterize the modulation ratio of the deviation component. To eliminate the influence of flow field modulation, the measured difference in pure adsorption contribution needs to be inversely modulated through amplitude modulation calculation to separate the true adsorption contribution signal. Based on the basic principle of linear modulation, the measured signal and the true signal satisfy the amplitude ratio relationship. The modulation effect caused by the flow field can be canceled through inverse modulation calculation. The corresponding formula is: ;

[0131] In the formula, S is the normalized adsorption signal of the flow field at the corresponding time, and the unit of the corresponding value is Rayleigh; ΔZ is the difference in pure adsorption contribution at the corresponding time, and the unit of the corresponding value is Rayleigh; M is the flow field disturbance factor at the corresponding time, which is a dimensionless parameter. The treatment time variable corresponds to a value in seconds. During the calculation, the difference between the flow field disturbance factor value and the pure adsorption contribution value at the same time point participates in the modulation calculation at the corresponding time, resulting in a signal that is only related to the actual adsorption intensity of the resin. This signal changes continuously with the treatment time, forming a complete flow field normalized adsorption signal sequence.

[0132] Step 62: Perform acoustic integration on the normalized adsorption signal of the flow field to generate a characterization value of the total adsorption amount. The specific operation is as follows:

[0133] The instantaneous value of the flow field normalized adsorption signal characterizes the instantaneous adsorption rate of solutes in the blood by the resin bed at the corresponding moment, which is the change in acoustic impedance corresponding to the amount of solute adsorbed by the resin per unit time. The total adsorption capacity of the resin bed is a fixed value. As the treatment time progresses, the cumulative total of the instantaneous adsorption rate is positively correlated with the adsorption saturation of the resin. The integral of the instantaneous adsorption rate over time is equal to the total adsorption amount within that time period. This is the basic physical meaning of integral calculation. The instantaneous value of the flow field normalized adsorption signal has a linear correspondence with the amount of solute adsorbed by the resin per unit time. Therefore, the definite integral of this signal over the treatment time interval can directly characterize the total amount of solute cumulatively adsorbed by the resin bed. Based on the mathematical definition of definite integral and the cumulative characteristics of the adsorption process, the corresponding formula is: ;

[0134] In the formula, Q is the total adsorption characterization value, and the corresponding value is in Rayleigh multiplied by seconds; S is the normalized adsorption signal of the flow field, and the corresponding value is in Rayleigh. The treatment time variable is represented by a value in seconds; 0 represents the start time of treatment, and T represents the current calculation time, both in seconds. The integral operation is updated in real time as the treatment progresses, and a corresponding total adsorption value is generated at each calculation time, providing a real-time updated quantitative basis for subsequent threshold comparisons.

[0135] Step 63: Compare the total adsorption characterization value with the relaxation characteristic value using acoustic dual thresholds, and output a saturation alarm signal based on the comparison result. The specific operation is as follows:

[0136] The total adsorption characterization value reflects the resin's adsorption process from the perspective of steady-state adsorption accumulation, while the relaxation characteristic value reflects the resin's adsorption saturation state from the perspective of changes in the physical properties of the resin particles themselves. These two values ​​are generated based on completely different measurement principles, eliminating common-mode interference factors. Dual-threshold comparison enables cross-validation of both dimensions, significantly reducing the probability of misjudgment of resin saturation. The values ​​of the dual thresholds are pre-determined through calibration experiments using an unloaded resin bed and different saturation gradients. The calibration process employs blood matrix and flow conditions consistent with clinical treatment to ensure the adaptability of the thresholds to the actual clinical scenario. In the acoustic dual-threshold comparison process, the actual... The updated total adsorption value is compared with a preset adsorption threshold, and the synchronously generated relaxation characteristic value is compared with a preset relaxation characteristic threshold. The two sets of comparison results jointly determine the output logic of the saturation alarm signal. Only when the total adsorption value and the relaxation characteristic value simultaneously reach or exceed the corresponding set threshold will the corresponding level of saturation alarm signal be output, avoiding false alarms caused by interference from a single parameter. The output saturation alarm signal can be directly connected to the control system of the bedside blood purification equipment, providing a trigger basis for manual prompts or automatic control of the treatment process, ensuring the safety and economy of the treatment.

[0137] Example 2

[0138] Please see Figure 2 Based on Example 1, this embodiment provides an online monitoring system for the resin saturation of a separation column, including:

[0139] The data acquisition module is used to acquire real-time acoustic impedance data at the inlet and outlet ends of the separation column;

[0140] The calculation module is used to calculate the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet ends, which is used as the real-time acoustic impedance instantaneous difference value.

[0141] The disturbance generation module is used to measure the radial blood flow velocity distribution in the middle section of the separation column and generate a flow field disturbance factor based on the radial blood flow velocity distribution.

[0142] The disturbance acquisition module is used to apply a local pulse thrust to the resin bed in the separation column and acquire the instantaneous acoustic impedance difference after the disturbance.

[0143] The feature extraction module is used to generate a relaxation recovery curve based on the change of the instantaneous acoustic impedance difference after disturbance over time, and to extract relaxation feature values ​​from the relaxation recovery curve.

[0144] The saturation alarm module corrects the instantaneous difference of real-time acoustic impedance using the flow field disturbance factor to obtain a corrected signal. It then integrates the corrected signal and compares the integration result with the relaxation eigenvalue. Based on the comparison result, it outputs a saturation alarm signal.

[0145] The above description is merely a preferred embodiment of the present invention; however, 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 its improved concepts, should be covered within the scope of protection of the present invention.

Claims

1. A method for online monitoring of resin saturation in a separation column, characterized in that, include: Obtain real-time acoustic impedance data at the inlet and outlet of the separation column; Calculate the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet ends, and use it as the real-time acoustic impedance instantaneous difference value. The radial blood flow velocity distribution in the middle section of the separation column is measured, and the flow field disturbance factor is generated based on the radial blood flow velocity distribution; A local pulse thrust is applied to the resin bed in the separation column, and the instantaneous acoustic impedance difference after the disturbance is obtained; A relaxation recovery curve is generated based on the change of the instantaneous acoustic impedance difference after disturbance over time, and relaxation characteristic values ​​are extracted from the relaxation recovery curve. The instantaneous difference of real-time acoustic impedance is corrected by the flow field disturbance factor to obtain the corrected signal. The corrected signal is integrated and the integration result is compared with the relaxation characteristic value. Based on the comparison result, a saturation alarm signal is output.

2. The method for online monitoring of resin saturation in a separation column according to claim 1, characterized in that, Acquire real-time acoustic impedance data at the inlet and outlet of the separation column, including: Broadband acoustic excitation is applied at the inlet and outlet of the separation column, the echo is received, the echo of the blood segment is extracted by gating through a time window, and the amplitude attenuation spectrum and phase shift spectrum of the blood in the broadband are obtained as the blood spectral fingerprint. Based on the principal component resonance peak identified by blood spectral fingerprinting, the acoustic impedance at the inlet and outlet ends is measured using the frequency of this resonance peak to obtain the purification acoustic impedance change curve.

3. The method for online monitoring of resin saturation in a separation column according to claim 2, characterized in that, Calculate the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet ends, including: The purification acoustic impedance change curve at the inlet end is time-delayed, with the delay time being the average passage time of blood flowing from the inlet end to the outlet end, generating a delayed waveform that includes rheological common-mode noise. The purification acoustic impedance change curve at the outlet end is superimposed with the delayed waveform by acoustic phase cancellation, and the superimposed signal is extracted as the difference in pure adsorption contribution.

4. The method for online monitoring of resin saturation in a separation column according to claim 3, characterized in that, The radial blood flow velocity distribution in the middle section of the separation column is measured, and a flow field disturbance factor is generated based on the radial blood flow velocity distribution, including: Multi-angle acoustic wave incidentness was applied to the middle section of the separation column to obtain Doppler frequency shift clusters of blood flow at different radial depths; The Doppler frequency shift clusters are converted into equivalent acoustic impedance modulation coefficient sequences for each radial layer; Extract the near-wall velocity gradient exponent from the equivalent acoustic impedance modulation coefficient sequence; The near-wall velocity gradient exponent is modulated with the delay waveform to generate the flow field disturbance factor.

5. The method for online monitoring of resin saturation in a separation column according to claim 4, characterized in that, Applying localized pulse thrust to the resin bed within the separation column includes: Based on the difference signal of pure adsorption contribution and the flow field disturbance factor, a physical disturbance trigger pulse is generated; Based on the physical disturbance trigger pulse, a multi-channel delayed trigger signal is generated using a geometric acoustic model of the separation column; Local pulse thrust is applied to the resin bed based on multiple delayed trigger signals.

6. The method for online monitoring of resin saturation in a separation column according to claim 5, characterized in that, Obtain the instantaneous acoustic impedance difference after the disturbance, including: After applying a local pulse thrust, acoustic impedance data at different time points after the disturbance are collected to obtain the acoustic impedance time series after the disturbance. The difference between the time series of acoustic impedance after disturbance and the pure adsorption contribution before disturbance is calculated to obtain the net disturbance response curve. The relaxation characteristic value is extracted from the net disturbance response curve and used as the characteristic value of the instantaneous acoustic impedance difference after the disturbance.

7. The method for online monitoring of resin saturation in a separation column according to claim 6, characterized in that, The relaxation recovery curve is generated based on the change of the instantaneous acoustic impedance difference after the disturbance over time, including: Generate a set of time-off copies of the net disturbance response curve with different time delays; The original net disturbance response curve is multiplied with each time offset replica to obtain the autocorrelation intensity signal corresponding to each delay time. The autocorrelation intensity signals corresponding to each delay time are sorted according to the delay time and reconstructed into autocorrelation function curves.

8. The method for online monitoring of resin saturation in a separation column according to claim 7, characterized in that, Extracting relaxation eigenvalues ​​from the relaxation recovery curve includes: Identify and extract the dominant relaxation time constant from the autocorrelation function curve; The relaxation time constant and the initial amplitude of the net disturbance response curve are combined and encoded to generate relaxation eigenvalues.

9. The method for online monitoring of resin saturation in a separation column according to claim 8, characterized in that, The corrected signal is integrated, and the integration result is compared with the relaxation eigenvalue. Based on the comparison result, a saturation alarm signal is output, including: The difference between the flow field disturbance factor and the pure adsorption contribution is acoustically amplitude modulated to generate a flow field normalized adsorption signal. The normalized adsorption signal of the flow field is acoustically integrated to generate a characterization value of the total adsorption. The total adsorption characterization value and the relaxation characteristic value are compared using acoustic dual thresholds, and a saturation alarm signal is output based on the comparison result.

10. An online monitoring system for the resin saturation of a separation column, applied to the online monitoring method for resin saturation of a separation column according to any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire real-time acoustic impedance data at the inlet and outlet ends of the separation column; The calculation module is used to calculate the instantaneous difference between the real-time acoustic impedance data at the inlet and outlet ends, which is used as the real-time acoustic impedance instantaneous difference value. The disturbance generation module is used to measure the radial blood flow velocity distribution in the middle section of the separation column and generate a flow field disturbance factor based on the radial blood flow velocity distribution. The disturbance acquisition module is used to apply a local pulse thrust to the resin bed in the separation column and acquire the instantaneous acoustic impedance difference after the disturbance. The feature extraction module is used to generate a relaxation recovery curve based on the change of the instantaneous acoustic impedance difference after disturbance over time, and to extract relaxation feature values ​​from the relaxation recovery curve. The saturation alarm module corrects the instantaneous difference of real-time acoustic impedance using the flow field disturbance factor to obtain a corrected signal. It then integrates the corrected signal and compares the integration result with the relaxation eigenvalue. Based on the comparison result, it outputs a saturation alarm signal.

Images