An abnormal blood sampling test data evaluation processing method and system of a blood sampling system

Abstract

The application provides an abnormal blood sampling test data evaluation processing method and system of a blood sampling system. Wherein, the diffusion trajectory dynamics of the contact surface between the anticoagulant and blood in the blood collection tube is tracked in real time, the distortion characteristics of the dielectric response waveforms on both sides are synchronously collected by a capacitive sensor array to generate diffusion form parameters; the sound frequency modulated by the sound flow device triggered by the sudden change of curvature is analyzed to form a composite disturbance field; the capacitive axial scanning is performed on the disturbed mixed fluid to extract the phase angle offset and gradient mutation point coordinates to construct a three-dimensional dielectric anomaly map; the standard dielectric model is matched to locate the concentration gradient imbalance area and vortex decay area coordinate set to generate an anticoagulant anomaly index and a mixed defect matrix combined with the phase angle offset difference; the data anomaly degree evaluation function is established based on the anomaly index and the defect matrix, the clinical error threshold is dynamically associated, and the test evaluation result is output. The application can monitor the anticoagulant diffusion in real time and accurately evaluate the mixed quality error.

Description

Technical Field

[0001] This application relates to the field of data evaluation and processing technology, and in particular to a method and system for evaluating and processing abnormal blood collection test data in a blood collection system. Background Technology

[0002] In clinical blood testing, the homogeneous mixing of anticoagulants and blood is crucial for ensuring accurate test results. Uneven mixing can lead to imbalances in the local anticoagulant concentration gradient or eddy current attenuation, causing coagulation abnormalities, blood cell morphology damage, or deviations in biochemical indicators, directly impacting the reliability of key tests such as complete blood count and coagulation function. Therefore, real-time monitoring of anticoagulant diffusion dynamics and precise control of the mixing process are necessary to mitigate the risk of testing errors caused by microscopic mixing defects.

[0003] In current technology, a mainstream approach is a mixing process monitoring system based on optical imaging. This system uses high-speed microscopic imaging and image analysis algorithms to capture the macroscopic morphological changes in the diffusion trajectory of anticoagulants, and then combines this with a preset threshold to trigger a mechanical stirring device to adjust the mixing intensity. This system dynamically captures the deformation characteristics of the mixing interface and optimizes mixing uniformity using an empirical model of stirring rate and time.

[0004] However, optical imaging schemes rely on visible light transmission, which cannot penetrate highly turbid blood media, resulting in insufficient sensitivity for monitoring the diffusion trajectory of deep mixing interfaces. At the same time, their image analysis algorithms can only identify macroscopic morphological anomalies, which limits the accuracy of locating mixing defects and makes it impossible to effectively identify local eddy current attenuation areas. Summary of the Invention

[0005] This application provides a method and system for evaluating and processing abnormal blood sampling test data in a blood sampling system, in order to solve the problem of detection errors caused by uneven mixing of anticoagulant and blood in the prior art.

[0006] In a first aspect, embodiments of this application provide a method for evaluating and processing abnormal blood collection test data in a blood collection system, comprising: real-time tracking of the dynamic changes in the diffusion trajectory of the anticoagulant and the blood contact surface within the blood collection tube; synchronously capturing the waveform distortion characteristics of the dielectric response on both sides of the blood contact surface using a capacitance sensor array to generate diffusion trajectory morphology parameters; analyzing the curvature abrupt change characteristics of the diffusion trajectory morphology parameters; triggering an acoustic-fluid mixing device to modulate the acoustic wave frequency and fluid flow velocity to form a composite disturbance field within the mixing cavity of the acoustic-fluid mixing device; performing a capacitance gradient axial scan on the mixed fluid after the composite disturbance field is applied, and extracting the phase angle offset and gradient abrupt change points of the capacitance change curve. The spatial coordinates are used to construct a three-dimensional distribution map of the dielectric abnormality region. Based on the spatial topological features of the phase angle offset in the three-dimensional distribution map, a corresponding pre-stored standard dielectric model is matched to locate the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region. An anticoagulant distribution abnormality index is generated by calculating the spatial topological feature deviation. A mixed defect assessment matrix is ​​established based on the difference in phase angle offset between the spatial coordinate set and the standard dielectric model. A data anomaly assessment function is constructed based on the anticoagulant distribution abnormality index and the mixed defect assessment matrix. The mapping relationship between the assessment function output value and the clinical test error threshold is dynamically linked to generate the test data assessment result.

[0007] Secondly, embodiments of this application provide an abnormal blood collection test data evaluation and processing system for a blood collection system, comprising: a tracking module, used to track the dynamic changes in the diffusion trajectory of the anticoagulant and blood contact surface in the blood collection tube in real time, and synchronously capture the waveform distortion characteristics of the dielectric response on both sides of the blood contact surface through a capacitance sensor array to generate diffusion trajectory morphology parameters; an analysis module, used to analyze the curvature abrupt change characteristics of the diffusion trajectory morphology parameters, trigger an acoustic-fluid mixing device to modulate the acoustic wave frequency and fluid flow velocity, and form a composite disturbance field in the mixing cavity of the acoustic-fluid mixing device; and a construction module, used to perform an axial scan of the capacitance gradient on the mixed fluid after the composite disturbance field, and extract the phase angle offset and gradient of the capacitance change curve. The system uses the spatial coordinates of mutation points to construct a three-dimensional distribution map of dielectric anomaly regions. A matching module matches pre-stored standard dielectric models based on the spatial topological features of phase angle offsets in the three-dimensional distribution map to locate the spatial coordinate sets of anticoagulant concentration gradient imbalance regions and mixed eddy current attenuation regions. It also generates an anticoagulant distribution anomaly index by calculating the spatial topological feature deviation and establishes a mixed defect assessment matrix based on the difference in phase angle offsets between the spatial coordinate set and the standard dielectric model. An evaluation module constructs a data anomaly assessment function based on the anticoagulant distribution anomaly index and the mixed defect assessment matrix. It dynamically correlates the output value of the evaluation function with the mapping relationship of clinical test error thresholds to generate test data evaluation results.

[0008] The technical solution of this application has the following beneficial effects: This application achieves high-precision dynamic monitoring and parameter extraction of the mixing process by real-time tracking of the dynamic changes in the diffusion trajectory of the anticoagulant and blood contact surface in blood collection tubes and synchronously capturing the waveform distortion characteristics of the dielectric response on both sides of the interface using a capacitive sensor array; by analyzing the curvature change characteristics of the diffusion trajectory to trigger the control of the composite disturbance field of the acoustic fluid device, the mixing conditions are dynamically optimized to suppress concentration gradient imbalance and eddy current attenuation; based on the phase angle offset and gradient change point coordinates extracted by the axial scanning of the capacitive gradient, a three-dimensional dielectric anomaly map is constructed, and combined with the spatial topological matching of the standard dielectric model, the mixing defect region is accurately located and the anticoagulant distribution anomaly index and mixing defect assessment matrix are quantified; finally, through the dynamic mapping correlation between the anomaly assessment function and the clinical error threshold, the entire process of closed-loop feedback control of mixing quality is realized. This solution significantly improves the monitoring sensitivity and assessment efficiency of anticoagulant mixing uniformity, reduces the risk of testing errors caused by mixing defects, and provides dynamic assurance for the reliability of clinical test data.

[0009] Furthermore, by integrating the curvature distribution characteristics of diffusion trajectory morphology parameters, acoustic-fluid parameters of the composite perturbation field, and the spatial coordinate set of dielectric anomaly regions, combined with dynamic feedback from test data evaluation results, multi-dimensional collaborative compensation parameters are generated to achieve precise correction of anticoagulant concentration gradient imbalance and fluid shear defects within the mixing chamber. Based on the collaborative compensation parameters, the acoustic frequency, phase, and velocity distribution of the acoustic-fluidic device are dynamically adjusted to optimize the energy distribution and eddy current generation mode of the composite perturbation field, suppressing local concentration polarization and enhancing shear uniformity. Simultaneously, through real-time updating of the dielectric anomaly three-dimensional map and topological matching of the standard model, the weight coefficients of the mixing defect evaluation matrix are iteratively corrected to improve the adaptability of anomaly region location and compensation parameters. Ultimately, a fully closed-loop control mechanism of "dynamic monitoring - parameter analysis - perturbation optimization - defect correction" is formed, significantly improving the mixing uniformity and stability of anticoagulant and blood, effectively reducing test errors caused by concentration shifts or shear anomalies, and providing high-precision, adaptive quality control assurance for clinical sample preprocessing. These or other aspects of this application will become more apparent in the following description of embodiments. Attached Figure Description

[0010] Figure 1 A flowchart of a method for evaluating and processing abnormal blood collection test data from a blood collection system provided in this application is shown;

[0011] Figure 2 A schematic diagram of the structure of an abnormal blood collection test data evaluation and processing system provided in this application is shown. Detailed Implementation

[0012] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0013] This application aims to overcome the problems of insufficient dynamic monitoring accuracy, delayed location of mixing defects, and lack of traceability mechanisms for testing errors in traditional blood collection systems. This technology constructs a collaborative mechanism encompassing "dynamic tracking, parameter feedback, disturbance control, and closed-loop evaluation" to achieve high-precision real-time monitoring and proactive intervention of the mixing interface. Based on the fusion analysis of capacitance sensing and dielectric response distortion characteristics, it accurately captures abnormal diffusion trajectory morphology; through dynamic modulation of the acoustic-fluid composite disturbance field, it suppresses concentration gradient imbalance and eddy current attenuation; combining three-dimensional topological modeling of dielectric anomaly maps with spatial matching of standard parameters, it quantifies mixing defects and constructs an evaluation system. Finally, it dynamically correlates the anomaly index with clinical error thresholds, forming a real-time closed-loop feedback control of mixing quality, thereby improving the uniformity of anticoagulant distribution and the reliability of test data, providing an intelligent quality control solution for clinical blood sample pretreatment.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0015] Figure 1 This application provides a flowchart of a method for evaluating and processing abnormal blood sampling test data in a blood collection system, as shown in the embodiments of this application. Figure 1 As shown, the method includes: 101. Real-time tracking of the dynamic changes in the diffusion trajectory of the anticoagulant and blood contact surface within the blood collection tube, and synchronously capturing the waveform distortion characteristics of the dielectric response on both sides of the blood contact surface through a capacitive sensor array to generate diffusion trajectory morphology parameters; in this step, the dynamic change in diffusion trajectory refers to the real-time deformation process of the diffusion path at a time resolution of 0.1 seconds between the anticoagulant and blood contact surface. The capacitive sensor array refers to a ring detection device composed of 24 high-frequency capacitive sensors. The dielectric response refers to the change in capacitance-frequency characteristics at the blood-anticoagulant mixing interface due to polarization effect. The waveform distortion characteristics refer to the asymmetric waveform peak shift that occurs during the diffusion of the capacitance signal at the mixing interface. The diffusion trajectory morphology parameters refer to a quantized dataset containing the radius of curvature and diffusion rate (0.5-3 mm / s).

[0016] In this embodiment, a 24-channel capacitive sensor array (0.5 mm spacing) distributed in a ring continuously detects the change in dielectric constant at the interface between blood and anticoagulant in a blood collection tube using a 10 MHz high-frequency signal. Each sensor node, based on the principle of edge electric field perturbation, captures the differences in polarization response between different phases of the substance (plasma / anticoagulant) during diffusion at the contact surface. After eliminating environmental electromagnetic interference through adaptive Kalman filtering, the waveform rising edge slope (dV / dt ≥ 5 μV / ms) and valley oscillation frequency (8-12 kHz) are extracted as dynamic feature parameters. An improved Canny edge detection algorithm is used to perform sub-pixel-level localization of the contact surface contour. Combined with cubic spline interpolation, a sequence of curvature radii (R = 0.2-1.5 mm) and contact angle change rate (Δθ = 3° / s) of the diffusion trajectory is generated, ultimately forming a diffusion trajectory morphological parameter matrix with a spatiotemporal resolution of 0.1 mm / 10 ms.

[0017] In a fully automated blood collection tube anticoagulant mixing quality monitoring system, when EDTA anticoagulant is injected into a blood collection tube (4mm inner diameter), a high-density flexible capacitive sensor array (128 channels, 0.2mm spacing) is used to track the diffusion trajectory of the blood-anticoagulant interface in real time. The sensors detect the difference in dielectric constant between the two sides of the interface (blood ε=58±3, anticoagulant ε=27±2), capturing the waveform distortion characteristics of the diffusion interface: the rise time increases from the baseline value of 0.8ms to 1.5ms, and the waveform amplitude attenuates by 25%. Simultaneously, the system extracts the curvature extrema of the diffusion trajectory (curvature κ≥0.6mm⁻¹) and the diffusion rate gradient (Δv=0.1-0.3mm / s). A modified Savitzky-Golay filtering algorithm is used to eliminate high-frequency noise, generating a morphological parameter dataset containing curvature distribution, diffusion direction, and interface fluctuation frequency (2-8Hz), providing high-precision input for subsequent control.

[0018] 102. Analyze the curvature abrupt change characteristics of the diffusion trajectory morphological parameters, trigger the acoustic-fluid mixing device to modulate the acoustic wave frequency and fluid flow velocity, forming a composite disturbance field within the mixing cavity of the acoustic-fluid mixing device; in this step, the curvature abrupt change characteristic refers to a local abrupt change point in the diffusion trajectory where the first derivative of curvature exceeds a threshold. The acoustic-fluid mixing device refers to a synergistic device integrating a piezoelectric transducer and a microfluidic channel (flow velocity 0.2-5 mL / s). The composite disturbance field refers to a three-dimensional vortex field formed by the superposition of acoustic radiation force and fluid shear stress (τ=0.1-0.8 Pa).

[0019] In this embodiment, based on the first derivative mutation detection of the diffusion trajectory curvature (threshold δR / δt≥1.2mm / ms), a sliding window variance analysis method is used to identify the instability critical point of the mixing interface (window width 50ms). A piezoelectric ceramic acoustic wave generator (frequency adjustable from 1-5MHz) and a micro-peristaltic pump (flow rate 0.1-5mL / s) are coordinated for control. PID closed-loop regulation is used to parametrically couple the sound pressure level (SPL=145dB) and Reynolds number (Re=200-500). Within the 3D-printed spiral flow channel mixing cavity, the nonlinear superposition of the acoustic flow effect (acoustic radiation force F=0.3μN) and laminar shear force generates a composite disturbance field of axial vortices (vorticity ω=15s⁻¹) and radial secondary flow. Particle image velocimetry (PIV) verifies a uniform mixing state with a flow rate standard deviation ≤5%.

[0020] When morphological parameter analysis shows that the frequency of curvature abrupt changes in the Z=5-8mm axial segment exceeds the threshold (≥5Hz) and the density of gradient abrupt change points exceeds the limit (Δκ / Δx≥0.4mm⁻²), the intelligent control module of the acoustic-fluid mixing device is triggered. The piezoelectric array (frequency 11-15MHz, sound pressure level 140dB) inside the device switches to dual-frequency standing wave mode (13MHz+15MHz orthogonal superposition), and simultaneously adjusts the peristaltic pump flow rate from 0.8mL / s to 1.5mL / s, forming a composite disturbance field with vortex intensity ω≥25s⁻¹. Through high-speed particle imaging velocimetry verification, the synergistic effect of acoustic radiation force (F=1.8μN) and fluid shear stress (τ=0.6Pa) reduces the curvature abrupt change region of the anticoagulant diffusion interface by 60%, and stabilizes the interface fluctuation frequency to within ±1Hz.

[0021] 103. Perform an axial capacitance gradient scan on the mixed fluid after the combined disturbance field, extract the phase angle offset and spatial coordinates of the gradient abrupt change points of the capacitance change curve, and construct a three-dimensional distribution map of the dielectric anomaly region. In this step, the axial capacitance gradient scan refers to a capacitance tomography method performed along the axis of the mixing tube with a step resolution of 10 μm. The phase angle offset refers to the phase lag of the capacitance signal relative to the reference signal caused by the dielectric loss of the mixed fluid. The gradient abrupt change point refers to the spatial coordinate point where the capacitance gradient amplitude exceeds a set threshold (▽C≥0.6pF / mm). The dielectric anomaly region refers to a three-dimensional spatial distribution region where the dielectric constant deviates from the standard value by ±20%.

[0022] In this embodiment, a capacitance gradient axial scanner (axial resolution 10 μm) is used to perform 32-slice tomographic scanning along the length of the hybrid tube, acquiring 256 capacitance measurement points per slice. The phase angle shift of each scanned slice in the 2-8 MHz frequency band is extracted using Fast Fourier Transform, and a modified Difference of Gaussians algorithm is used to detect abrupt changes in capacitance gradient. An improved back-projection algorithm is used to reconstruct the three-dimensional dielectric constant distribution, mapping the phase shift to voxel grayscale values. The Marching Cubes algorithm is then used to generate a three-dimensional dielectric anomaly map containing the vortex core region (grayscale value ≥ 180) and the concentration boundary layer.

[0023] After the perturbation field was applied, a capacitive gradient axial scanner (step accuracy 0.05 mm) was used to perform a full axial scan of the mixed fluid (Z=0-12 mm). A phase angle shift (Δθ=+9°) and a gradient abrupt change point (ΔC / Δx=0.45 pF / mm) were detected in the capacitance curve at Z=6.2 mm. A dielectric anomaly region map was constructed using a three-dimensional inverse distance weighted interpolation algorithm (IDW, search radius 0.3 mm). The map showed that the high-concentration imbalance region (Δε≥12) was located in the X=3-5 mm / Y=2-4 mm range, and the mixed eddy current attenuation region (vortex radius ≤0.08 mm) was concentrated in the Z=5-7 mm layer. The polygonal mesh boundaries of the anomaly region (vertex spacing ≤0.1 mm) were extracted using the Marching Cubes algorithm to generate a three-dimensional coordinate set.

[0024] 104. Based on the spatial topological features of the phase angle offset in the three-dimensional distribution map, match the corresponding pre-stored standard dielectric model to locate the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixing eddy attenuation region. Generate an anticoagulant distribution anomaly index by calculating the spatial topological feature deviation. Establish a mixing defect evaluation matrix based on the difference in phase angle offset between the spatial coordinate set and the standard dielectric model. In this step, the spatial topological features refer to the geometric relationship between the vortex core center spacing (2-5 mm) and the concentration layer thickness (0.1-0.5 mm) in the three-dimensional dielectric distribution. The standard dielectric model refers to the spatial distribution template of dielectric parameters under ideal mixing conditions pre-stored through CFD simulation. The anticoagulant concentration gradient imbalance region refers to the spatial range where the absolute value of the concentration gradient is below the critical value.

[0025] The mixed eddy current attenuation region refers to the fluid region where the eddy current value attenuates to less than 30% of the initial value. Spatial topological feature deviation refers to the difference between the actual dielectric distribution and the spatial structure of the standard model. The mixed defect evaluation matrix refers to a quantitative evaluation table containing eddy current attenuation level and concentration imbalance intensity. In this embodiment, a point cloud registration algorithm is used to perform non-rigid matching between the three-dimensional dielectric spectrum and the standard dielectric model (NURBS surface modeling), and an improved ICP algorithm is used to optimize the spatial registration error (RMSE≤0.05mm). Topological features of the concentration gradient imbalance region (triangular patch curvature κ≥0.8mm⁻¹) and the eddy current attenuation region (velocity attenuation rate≥20% / mm) are extracted through Delaunay triangulation. The Mahalanobis distance (threshold D²≥2.8) between the actual phase angle offset and the standard model is calculated, and combined with the vortex core center coordinate offset (ΔXYZ≥0.3mm), an anticoagulant distribution anomaly index (0-1 normalized) is generated. A mixed defect evaluation matrix containing the spatial coordinate-phase offset correlation matrix is ​​constructed through tensor decomposition. The phase angle offset topological features of the 3D map (such as the annular phase hysteresis band θ=+7°±1°) were registered with a pre-stored standard dielectric model (healthy sample θ=±2°) using NURBS surface registration to locate the core area of ​​anticoagulant concentration imbalance (coordinates X=4.2mm, Z=6.5mm) and the eddy current attenuation boundary fracture zone (fracture width 0.12mm). The spatial topological deviation was calculated using Hausdorff distance (D=0.73, threshold D<0.5), generating an anticoagulant distribution anomaly index AEI=85%. A hybrid defect evaluation matrix (defect score S=4.2 / 5.0) was constructed based on the phase angle difference (Δθ=5°) and coordinate offset (Δd=0.25mm), with high-weight defect regions (weight ≥0.8) accounting for 35% of the matrix.

[0026] 105. Based on the anticoagulant distribution anomaly index and the mixed defect assessment matrix, a data anomaly assessment function is constructed. The mapping relationship between the output value of the assessment function and the clinical test error threshold is dynamically linked to generate the test data assessment result. In this step, the data anomaly assessment function refers to a linear model that integrates the anticoagulant distribution anomaly index (weight 0.7) and the entropy value of the mixed defect matrix (weight 0.3). The clinical test error threshold refers to the maximum allowable detection deviation (total error ≤ 15%) set based on the CLIA'88 standard. The test data assessment result refers to a comprehensive judgment report that includes the anomaly probability (0-100%), defect heatmap, and confidence interval (95% CI). In this embodiment, a linear weighted assessment function is constructed: anomaly = 0.6 × anticoagulant distribution anomaly index + 0.4 × mixed defect matrix Frobenius norm. The test error threshold mapping curve (R² = 0.93) is obtained by fitting a clinical database (n = 1200 cases). The Dynamic Time Warping (DTW) algorithm is used to align the output value of the evaluation function with historical failure modes in real time. When the abnormality exceeds the preset clinical threshold (0.78), a three-level early warning mechanism is triggered, and finally, a multi-dimensional test data evaluation result including probability confidence (95% CI) and defect space heatmap is generated.

[0027] Based on an AEI of 85% and a defect matrix (S=4.2), a data anomaly assessment function f=0.7AEI+0.3S (threshold f≤60) was constructed, dynamically correlated with clinical laboratory error standards (CLIA allows AEI≤70%). When f=78.6 exceeds the threshold, the system generates the test result "mixing quality substandard" and triggers a closed-loop optimization strategy: adjusting the acoustic frequency to 14MHz (Δf=+1MHz) and the flow rate to 1.8mL / s (Δv=+0.3mL / s), while compensating for phase lag (Δθ=-3°). After secondary mixing, the AEI decreased to 58% (f=52.4), and the eddy current attenuation zone area decreased by 80%, verifying the effectiveness of the dynamic assessment-control system in clinical blood collection anticoagulation processes.

[0028] In summary, steps 101 to 105 achieve high-precision dynamic monitoring and parameter extraction of the mixing interface by real-time tracking of the dynamic changes in the diffusion trajectory of the anticoagulant at the blood contact surface within the blood collection tube and capturing dielectric response waveform distortion characteristics using a capacitive sensor array. Based on curvature mutation characteristics-triggered acoustic-fluid composite disturbance field regulation, the acoustic frequency and fluid velocity can be dynamically optimized, suppressing local concentration gradient imbalances and enhancing eddy current stability. Through spatial topological matching between the three-dimensional dielectric anomaly map constructed by capacitive gradient scanning and the standard model, the mixing defect region is accurately located and the anticoagulant distribution anomaly index is quantified. A multi-dimensional defect assessment system is established by combining the mixing defect assessment matrix. Finally, through the dynamic correlation between the anomaly assessment function and the clinical error threshold, a closed-loop control mechanism of "monitoring-regulation-assessment-feedback" is formed, significantly improving the uniformity of anticoagulant mixing and the reliability of test data, reducing the risk of coagulation abnormalities caused by concentration shifts or eddy current attenuation, and providing intelligent quality control assurance for clinical blood sample preprocessing.

[0029] In some embodiments, steps 101 to 105 further include: 201, fusing the curvature distribution of the diffusion trajectory morphology parameters, the parameters of the composite disturbance field, and the correlation of the spatial coordinate set, and generating collaborative compensation parameters based on the evaluation results of the test data; in step 201, the curvature distribution of the diffusion trajectory morphology parameters refers to the variation characteristics of the curvature value in the diffusion path with time and space. The parameters of the composite disturbance field refer to the control parameters of the acoustic-fluid mixing device, such as acoustic frequency, flow velocity, and disturbance mode. The spatial coordinate set refers to the specific location coordinates of the dielectric anomaly region in three-dimensional space. The evaluation results of the test data refer to a comprehensive judgment report based on the anomaly probability and defect distribution of the mixing quality analysis. The collaborative compensation parameters refer to the set of acoustic frequency compensation amount, flow velocity adjustment gradient, and disturbance mode optimization coefficients used to optimize the mixing process.

[0030] In this embodiment, multi-source data fusion technology is used to perform multi-dimensional correlation analysis on the curvature distribution characteristics of the diffusion trajectory, the acoustic pressure-velocity coupling parameters of the composite disturbance field, and the spatial coordinates of the dielectric anomaly region. An improved graph convolutional network (GCN) is used to construct a parameter correlation model. A node feature extraction module encodes curvature abrupt change points (node ​​attributes include curvature change rate and timestamp), disturbance field intensity (node ​​attributes include sound pressure level and Reynolds number), and spatial coordinate offset (node ​​attributes include XYZ coordinates and dielectric gradient value). An edge weight calculation module establishes the spatial-temporal correlation of cross-domain parameters. Combining the evaluation results of inspection data (anomaly probability, defect heatmap), an attention mechanism is used to dynamically weight the influence of different parameters on the mixing quality. Finally, a fully connected layer outputs a set of collaborative compensation parameters, including acoustic frequency compensation (Δf), velocity adjustment gradient (Δv), and disturbance mode optimization coefficients (k_p), to achieve the generation of a multi-dimensional dynamic compensation strategy for the mixing process.

[0031] 202. Correct the anticoagulant distribution imbalance and fluid shear defects within the mixing chamber based on the aforementioned synergistic supplementation parameters. In step 202, the synergistic supplementation parameters refer to the optimized parameters of acoustic frequency, flow velocity, and disturbance mode used to correct mixing defects. The mixing chamber refers to the microfluidic device used to achieve mixing of the anticoagulant and blood. Anticoagulant distribution imbalance refers to the phenomenon of uneven concentration distribution of the anticoagulant within the mixing chamber. Fluid shear defects refer to eddy current attenuation caused by insufficient shear force during fluid mixing.

[0032] In this embodiment, based on the collaborative compensation parameters, an embedded control system performs real-time closed-loop control of the acoustic-fluid mixing device. A fuzzy PID control algorithm is employed to convert the acoustic frequency compensation (Δf = ±0.5MHz) in the compensation parameters into a driving voltage modulation signal for the piezoelectric transducer (accuracy ±0.1V). Simultaneously, the pulse frequency of the stepper motor of the micro-peristaltic pump (resolution 1Hz) is dynamically adjusted according to the flow rate adjustment gradient (Δv = 0.1-2mL / s). For the disturbance mode optimization coefficient (k_p = 0.8-1.2), the tilt angle of the guide vanes within the mixing cavity is adjusted (adjustable from 0° to 15°) using flow channel topology reconstruction technology to enhance the coupling strength between the acoustic-fluid effect and the shear flow. The system verifies the compensation effect through high-speed particle imaging velocimetry (PIV). When the standard deviation of the anticoagulant concentration gradient is detected to drop to the threshold (σ<0.05mg / mL) and the vortex recovery rate (η_vortex≥85%) is detected, the mixing defect correction is determined to be complete. Ultimately, the system achieves the technical indicators of 40% improvement in the spatial distribution uniformity of anticoagulants and 25% improvement in mixing efficiency, ensuring that blood sample pretreatment meets the accuracy requirements of clinical testing.

[0033] Here is a specific example: In the dynamic optimization scenario of the anticoagulant injection system of a hemodialysis machine, when the equipment detects an abnormal local concentration gradient at the blood-anticoagulant mixing interface within the dialysis tubing, the system activates a real-time compensation mechanism. A miniature capacitive sensor array (0.1 pF resolution) continuously collects data on curvature abrupt changes in the diffusion trajectory, simultaneously acquiring the current operating parameters of the acoustic-fluid mixing device, combined with the coordinate location results of the dielectric anomaly region. Based on an improved graph convolutional network, coupled analysis is performed on the aforementioned multidimensional parameters to identify the region of insufficient shear stress and acoustic-fluid coupling failure caused by dialysate flow rate fluctuations. The system generates collaborative compensation parameters, instructing the piezoelectric transducer to increase its frequency to 3.8 MHz to enhance acoustic radiation, while simultaneously adjusting the microfluidic pump pulse mode to achieve a flow rate gradient Δv = +0.5 mL / s, and applying a 12° tilt correction to the mixing chamber guide vanes. This increases the anticoagulant diffusion rate from 1.2 mm / s to 2.4 mm / s, and reduces the standard deviation of the concentration gradient from 0.08 mg / mL to 0.03 mg / mL. Verification using particle imaging velocimetry showed that the eddy recovery rate increased from 70% to 92%, achieving closed-loop optimization of anticoagulant distribution throughout the dialysis process.

[0034] In summary, steps 201 to 202 achieve intelligent closed-loop control of the mixing process between the anticoagulant and blood in the blood collection tube. By integrating diffusion trajectory morphology, acoustic-fluid disturbance field parameters, and dielectric anomaly spatial characteristics, a dynamic compensation model is constructed to accurately correct the imbalance of the anticoagulant distribution gradient and the attenuation defects of fluid eddies within the mixing chamber. Based on multi-dimensional biophysical characteristic collaborative analysis, the system autonomously optimizes acoustic frequency, flow velocity, and disturbance mode, significantly improving the uniformity of anticoagulant diffusion and mixing efficiency, while simultaneously reducing the coagulation risk caused by uneven mixing, ensuring the stability of blood sample quality, and providing a highly reliable pretreatment guarantee for clinical testing.

[0035] In some embodiments, step 104, which involves matching the spatial topological features of the phase angle offset in the three-dimensional distribution map with a corresponding pre-stored standard dielectric model to locate the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixing eddy current attenuation region, includes: 301. Obtaining the spatial topological features of the phase angle offset in the three-dimensional distribution map and constructing a spatial distribution set of the phase angle offset including spatial frequency and amplitude gradient; in step 301, the three-dimensional distribution map refers to the three-dimensional spatial distribution model of the dielectric constant reconstructed by axial scanning of the capacitance gradient. The spatial topological features of the phase angle offset refer to the geometric connection relationship and morphological distribution characteristics of the phase angle in three-dimensional space.

[0036] Spatial frequency refers to the periodic variation of phase angle offset in spatial dimensions. Amplitude gradient refers to the rate of change of the amplitude of phase angle offset in space. The spatial distribution set of phase angle offset refers to a dataset of phase angle features containing spatial frequency and amplitude gradient. In this embodiment, spatial frequency decomposition of phase angle offset in dielectric anomaly spectrum is performed using three-dimensional fast Fourier transform (3D-FFT) to extract amplitude gradient features in the main frequency band (2-8MHz). An improved Morlet wavelet analysis is then used to construct the spatial distribution set of phase angle offset. A spatial topology mapping algorithm is employed to convert the phase angle offset data into a three-dimensional voxel model containing the joint frequency-amplitude distribution. The amplitude gradient vector field is calculated using a gradient operator (a variant of the Sobel operator) and fused with the spatial frequency spectrum using tensors. Finally, a spatial distribution set of phase angle offset containing spatial frequency parameters, amplitude gradient direction, and local energy density is generated, providing a multi-dimensional feature basis for defect region identification.

[0037] 302. The reference phase angle distribution parameters of the pre-stored standard dielectric model are invoked, and the spatial frequency and amplitude gradient of the spatial distribution set of the phase angle offset are matched layer by layer. Spatial nodes whose deviation from the reference parameters exceeds the dynamic fluctuation threshold are marked. In step 302, the pre-stored standard dielectric model refers to the dielectric parameter distribution template under an ideal mixed state, pre-stored through simulation or experiment. The reference phase angle distribution parameters refer to the ideal spatial distribution characteristics of the phase angle in the standard dielectric model. The dynamic fluctuation threshold refers to the maximum allowable deviation between the phase angle offset and the reference parameters. Spatial nodes refer to the specific coordinates of the phase angle offset in three-dimensional space that exceed the dynamic fluctuation threshold. In this embodiment, based on the reference phase angle distribution of the pre-stored standard dielectric model (NURBS surface parametric modeling), the dynamic time warping (DTW) algorithm is used to match the spatial frequency spectrum of the actual phase angle offset layer by layer. By calculating the local frequency difference (Δf≤±0.2cycle / mm) and amplitude gradient deviation angle (θ≤15°), and combining adaptive threshold segmentation (an improved version of the Otsu algorithm), spatial nodes exceeding the dynamic fluctuation threshold (ΔA≥0.3dB / mm) are marked. Point cloud density clustering (DBSCAN algorithm) is used to remove isolated noise points, retaining the coordinates of continuous out-of-range regions (cluster diameter≥0.5mm), forming a set of spatial nodes that significantly deviate from the standard model, thus completing the preliminary localization of the dielectric anomaly region.

[0038] 303. Load the real-time energy decay trajectory of the hybrid eddy current decay process, associate it with the spatial nodes, and filter nodes whose energy decay rate and the magnitude gradient trend of the phase angle offset are opposite to those of the phase angle offset to generate a primary spatial cluster. In step 303, the hybrid eddy current decay process refers to the physical phenomenon that the eddy energy gradually weakens over time during fluid mixing. The real-time energy decay trajectory refers to the dynamic decay curve of eddy current energy changing over time. The energy decay rate refers to the decay speed of eddy current energy changing over time. The primary spatial cluster refers to the set of nodes whose energy decay rate and the magnitude gradient trend of the phase angle offset are opposite to those of the phase angle offset. In this embodiment, the real-time energy decay trajectory of the hybrid eddy current decay process (sampling rate 1kHz) is loaded, and the mutual information entropy algorithm is used to establish a correlation model between the spatial node energy decay rate (dE / dt = -0.5 to -2mJ / ms) and the phase angle magnitude gradient trend. Conflicting nodes with accelerated energy decay (d²E / dt²≤-0.1mJ / ms²) and reverse-increasing phase angle magnitude gradient (▽A≥0.4dB / mm) were screened using Pearson correlation coefficient (threshold r≤-0.7). An improved spectral clustering algorithm (k=3-5) was applied to generate a primary spatial cluster containing a vortex core offset region (offset ≥0.3mm) and a concentration boundary fracture zone, thereby achieving the co-location of fluid energy dissipation and dielectric anomaly.

[0039] 304. Within the area covered by the primary spatial cluster, the spatial frequency distribution of the phase angle offset and the fluctuation characteristics of the eddy current energy attenuation trajectory are fused to iteratively correct the overlapping boundary of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region. In step 304, the overlapping boundary refers to the spatial intersection of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region. Iterative correction refers to the algorithmic process of adjusting the boundary position through multiple optimizations. In this embodiment, within the area covered by the primary spatial cluster, the spatial frequency distribution of the phase angle (dominant frequency fluctuation ±10%) and the temporal fluctuation characteristics of the eddy current energy attenuation trajectory (fluctuation amplitude ≥15%) are fused using an improved level set method. The overlapping boundary is iteratively optimized using the alternating direction multiplier method (ADMM) to dynamically adjust the boundary weight coefficients (α=0.6-0.8) of the anticoagulant concentration gradient imbalance region (concentration difference ≥0.2mg / mL) and the mixed eddy current attenuation region. After 3-5 iterations, the boundary positioning error converges to within 0.05mm, generating a high-precision defect region overlapping boundary model, solving the boundary ambiguity and misjudgment problems in traditional methods.

[0040] 305. Integrate the dynamic fluctuation threshold of the spatial nodes, the temporal fluctuation characteristics of the eddy energy decay trajectory, and the iterative correction weight of the overlapping boundaries to output the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixed eddy attenuation region. In step 305, the temporal fluctuation characteristics refer to the change law of the eddy energy decay trajectory in the time dimension. The iterative correction weight refers to the contribution ratio of different parameters in the boundary correction process. The spatial coordinate set refers to the specific location coordinates of the anticoagulant concentration gradient imbalance region and the mixed eddy attenuation region. In this embodiment, the dynamic fluctuation threshold (ΔA=0.3dB / mm), the temporal fluctuation index (β=0.2-0.5) of the eddy energy decay trajectory, and the iterative correction weight of the overlapping boundaries (α=0.7) of the spatial nodes are integrated, and a multi-parameter fusion decision system is constructed using a random forest regression model. By assigning a 40% weight to temporal fluctuation features, a 35% weight to spatial thresholds, and a 25% weight to boundary features through feature importance analysis (Gini index ranking), the final output is a set of spatial coordinates (positioning accuracy ±0.1mm) containing the anticoagulant concentration gradient imbalance region (confidence ≥90%) and the mixing eddy attenuation region (confidence ≥85%), providing a quantitative defect distribution map for mixing process optimization.

[0041] Here is a specific example: In a blood center's platelet abnormality detection scenario using intelligent apheresis, an uneven mixing of EDTA anticoagulant led to abnormal platelet aggregation. The system constructed a three-dimensional phase angle distribution map of the blood collection tubes using a ring array eddy current sensor, detecting a phase angle shift of 0.18-0.35 rad in the 120-240 kHz frequency band, forming an abnormal gradient set containing 12 spatial frequency nodes. A pre-stored EDTA-K2 dielectric model was used for comparison, marking an abnormal node with a phase gradient deviation of 42% at a longitudinal distance of 28 mm in blood collection tube No. 3. Simultaneously loading the mixed eddy current energy attenuation trajectory revealed an abnormal decrease in the energy attenuation rate to -0.15 dB / ms in this region, forming a primary spatial cluster with a radius of 3.2 mm centered on this node. By iteratively fusing the phase frequency distribution and eddy current fluctuation characteristics, the overlapping boundary between the identified anticoagulant concentration imbalance zone (EDTA deviation ±18%) and the eddy current attenuation blind zone was corrected, ultimately outputting an abnormal core region with a diameter of 4.7 mm in the longitudinal 28-45 mm segment. Upon review, the platelet aggregation rate in the area was found to be 85% (normal <15%). The system then adjusted the ACD infusion rate to the standard value of 120% and started tubing preheating, which increased the platelet viability from 68% to 92%, effectively solving the data distortion problem caused by abnormal anticoagulation gradient.

[0042] In summary, steps 301 to 305 achieve intelligent collaborative localization of dielectric anomalies and fluid energy attenuation during anticoagulant mixing. By fusing the spatial frequency characteristics of phase angle offset and the dynamic fluctuation characteristics of eddy current energy attenuation trajectory, a multi-dimensional data-driven defect region identification model is constructed, accurately delineating the spatial boundaries between the anticoagulant concentration gradient imbalance zone and the mixing eddy current attenuation zone. Based on dynamic fluctuation thresholds and iterative correction algorithms, the system optimizes the acoustic-fluid coupling parameters and fluid shear modes within the mixing chamber in real time, significantly improving the uniformity of anticoagulant diffusion and effectively suppressing abnormal eddy current energy dissipation, ensuring the stability and repeatability of the mixing process, and providing high-precision quality control assurance for clinical blood sample processing.

[0043] In some embodiments, step 304, within the area covered by the primary spatial cluster, integrates the spatial frequency distribution of the phase angle offset and the fluctuation characteristics of the eddy current energy decay trajectory to iteratively correct the overlapping boundary of the anticoagulant concentration gradient imbalance region and the mixing eddy current decay region, including: 401, within the boundary range of the primary spatial cluster, performing global interpolation of the spatial frequency distribution of the phase angle offset and the fluctuation characteristics of the eddy current energy decay trajectory to generate a spatial interpolation grid covering the entire mixed fluid region; in step 401, the boundary range of the primary spatial cluster refers to the mixed fluid spatial region initially located by defect features. The spatial frequency distribution of the phase angle offset refers to the distribution characteristics of the phase angle in space as a function of frequency. The fluctuation characteristics of the eddy current energy decay trajectory refer to the dynamic change law of eddy current energy decay over time. Global interpolation refers to a method of processing discrete data continuously within a spatial range. The spatial interpolation grid refers to a regularized three-dimensional data grid covering the entire mixed fluid region.

[0044] In this embodiment, an improved radial basis function (RBF) interpolation algorithm is used to perform global interpolation on the spatial frequency distribution of phase angle offset (main frequency band 3-8 cycles / mm) and the fluctuation characteristics of eddy current energy decay trajectory (time-domain fluctuation index β=0.2-0.6) within the geometric boundary of the primary spatial cluster. By constructing a non-uniform sampling point adaptive weight model, discrete phase angle frequency data (sampling interval 0.5mm) and eddy current energy time series (sampling rate 1kHz) are coupled in a spatial-temporal dimension. Thin plate spline functions are used to eliminate boundary effects, generating a 0.1mm resolution interpolation grid covering the entire hybrid cavity. The spatial frequency amplitude gradient (▽A=0.1-0.8dB / mm) is calculated using the Sobel operator, and the eddy current fluctuation amplitude is extracted by Hilbert transform to obtain the envelope slope (dE / dt=-0.3~-1.5mJ / ms). Finally, a three-dimensional interpolation grid containing the joint frequency-energy distribution characteristics is formed, providing a continuous data basis for global defect analysis.

[0045] 402. Traverse each grid cell in the spatial interpolation grid, quantify the difference between the spatial frequency amplitude change rate of the phase angle offset and the fluctuation amplitude of the eddy current energy attenuation trajectory, and generate a dynamic fluctuation threshold mask; in step 402, a grid cell refers to the smallest data unit in the spatial interpolation grid. The spatial frequency amplitude change rate of the phase angle offset refers to the rate of change of the phase angle amplitude in space. The fluctuation amplitude difference of the eddy current energy attenuation trajectory refers to the deviation of the eddy current energy fluctuation amplitude from the reference value. The dynamic fluctuation threshold mask refers to the binarized matrix used to mark grid cells that exceed the fluctuation threshold. In this embodiment, based on the spatial interpolation grid cells, the sliding window variance analysis method is used to quantify the difference index between the spatial frequency amplitude change rate of the phase angle and the fluctuation amplitude of the eddy current energy. The two types of parameters are converted into dimensionless statistics through Z-score standardization, and a dynamic fluctuation threshold mask is constructed using improved fuzzy logic rules: when the phase angle change rate Z value is ≥2.5 and the eddy current difference Z value is ≤-1.8, it is marked as an abnormal cell. Morphological closing operations are used to eliminate discrete noise, retain continuous out-of-range regions (minimum cluster volume 0.2 mm³), and generate a binarized dynamic threshold mask to achieve multi-dimensional quantitative screening of mixed defect features.

[0046] 403. Superimpose the dynamic fluctuation threshold mask with the original boundary contour of the primary spatial cluster, extract the grid cells that synchronously exceed the limits of the phase angle offset spatial frequency amplitude change rate and eddy current energy attenuation fluctuation amplitude, and construct a multimodal feature superposition field; in step 403, the dynamic fluctuation threshold mask refers to the screening matrix that marks the grid cells that exceed the limits. The original boundary contour of the primary spatial cluster refers to the geometric boundary of the initially located defect region. The synchronously exceeding limit grid cells refer to the grid cells that simultaneously satisfy the limits of both the phase angle amplitude change rate and the eddy current fluctuation amplitude. The multimodal feature superposition field refers to the composite data field that fuses the phase angle and eddy current energy features.

[0047] In this embodiment, a spatial logical AND operation is performed between the dynamic threshold mask and the original boundary of the primary spatial cluster (a polygonal mesh extracted by the Marching Cubes algorithm). An improved region growing algorithm is used to extract mesh elements that simultaneously satisfy the requirements of phase angle frequency amplitude exceeding the limit (ΔA≥0.5dB / mm) and eddy current energy fluctuation exceeding the limit in the opposite direction (dE / dt≤-1.0mJ / ms). Tensor fusion technology is used to map the two types of feature parameters into multimodal feature vectors (dimensionality=32). After dimensionality reduction using principal component analysis (PCA), density peak clustering (DPC) algorithm is used to generate a multimodal feature superposition field containing the anticoagulant concentration boundary fracture zone (fracture width≥0.15mm) and the vortex core offset zone (offset≥0.25mm). Its spatial resolution is improved to 0.05mm, achieving accurate correlation mapping of defect regions.

[0048] 404. Using the multimodal feature superposition field as a benchmark, iteratively adjust the boundary weight parameters of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region to correct the overlapping boundary. In step 404, the multimodal feature superposition field refers to a multidimensional data model that includes phase angle and eddy current energy characteristics. The boundary weight parameters refer to the optimization coefficients used to adjust the overlap of the defect region boundaries. The overlapping boundary refers to the intersection of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region.

[0049] In this embodiment, a boundary weight parameter optimization model is constructed using a multimodal feature superposition field as input: the boundary overlap is defined as the objective function, where C_i is the concentration gradient imbalance intensity (0-1 normalized), V_j is the eddy current decay rate (0-1 normalized), and the initial weights are α=0.6 and β=0.4. A genetic algorithm (GA) is used for iterative optimization. Each iteration verifies the boundary correction effect through finite element simulation. The iteration terminates when the standard deviation of the concentration gradient (σ≤0.03mg / mL) and the eddy current recovery rate (η≥90%) simultaneously meet the targets. After 3-5 generations of evolution, the optimal weight parameters α=0.72 and β=0.28 are obtained, and the corrected overlapping boundary model is output, achieving synergistic optimization of anticoagulant distribution and hydrodynamic defects.

[0050] Here is a specific example: In a smart apheresis platelet quality monitoring scenario, a blood bank discovered that uneven mixing of EDTA anticoagulant resulted in microclots with a diameter of 2.3 mm within the collection bag. The system, using a three-dimensional phase angle distribution map constructed from a ring array eddy current sensor, showed a phase angle shift of 0.28-0.41 rad in the 180-220 kHz frequency band (normal threshold 0.18 rad), forming a primary spatial cluster containing 8 abnormal nodes. During global interpolation, an improved Kriging algorithm was used to generate a spatial grid with 0.5 mm precision. The system detected a phase frequency amplitude change rate of 0.12 rad / mm² in region 3, while the eddy current energy attenuation fluctuation amplitude deviation in this region reached -0.18 dB / ms, exceeding the dynamic threshold by 1.7 times. The superimposed feature field identified a multimodal superposition region centered at coordinates (35 mm, 22 mm) with a radius of 4.1 mm. In this region, the anticoagulant concentration gradient deviation reached ±23%, and the overlap rate with the eddy current attenuation blind zone reached 81%. By adjusting the adaptive boundary weights, the boundary accuracy of the anticoagulation imbalance zone was improved to ±0.8 mm, triggering the system to automatically initiate a secondary compensation procedure: increasing the anticoagulant infusion rate to 130% of the baseline value, while simultaneously activating the centrifugal mixing device for dynamic homogenization. A follow-up examination after 30 seconds of intervention showed that the EDTA concentration fluctuation in this area narrowed to ±6%, and platelet aggregation decreased from 78% to 12%, successfully eliminating the data distortion caused by the abnormal anticoagulation gradient.

[0051] In summary, steps 401 to 404 achieve intelligent global detection and dynamic boundary optimization of multimodal defect features in mixed fluids. By fusing the spatial frequency distribution of phase angles and the fluctuation characteristics of eddy current energy decay trajectories, a global interpolation grid and a dynamic threshold mask are constructed to accurately extract synchronous anomaly regions within the superimposed field of multimodal features. Based on an iterative weight adjustment algorithm, the system adaptively corrects the boundary overlap between the anticoagulant concentration gradient imbalance and the eddy current decay region, overcoming the limitations of traditional single-parameter detection. This significantly improves the positioning accuracy and boundary resolution of mixed defects, ensuring the uniformity and stability of the mixing process, and providing a highly reliable real-time quality monitoring solution for complex fluid industrial processes.

[0052] In some embodiments, step 104, which involves generating an anticoagulant distribution anomaly index by calculating spatial topological feature deviation, and establishing a hybrid defect evaluation matrix based on the phase angle offset difference between the spatial coordinate set and the standard dielectric model, includes: 501. Obtaining the topological feature density distribution of the spatial coordinate set and the dielectric constant distribution of the standard dielectric model, and calculating the gradient difference between the topological feature density distribution and the dielectric constant distribution to generate spatial topological feature deviation; In step 501, the spatial coordinate set refers to the set of three-dimensional spatial coordinate points of the region to be analyzed. The topological feature density distribution refers to the distribution of density features in the spatial coordinate set. The dielectric constant distribution of the standard dielectric model refers to the dielectric constant distribution at each point in the standard model. The gradient difference refers to the difference in the spatial rate of change between the two distributions. The spatial topological feature deviation is a comprehensive quantitative index of the difference between the topological feature and dielectric constant distributions.

[0053] In this embodiment, the nonlinear correlation between the topological feature density distribution (sampling interval 0.2 mm) of the spatial coordinate set and the dielectric constant distribution (resolution 0.1 mm³) of the standard dielectric model is calculated using the mutual information entropy algorithm. An improved Gaussian kernel density estimation method is used to extract local density extrema (density difference ≥ 15%). The Sobel operator is used to calculate the spatial gradient fields of the two distributions (gradient magnitude range 0.1-2.0 dB / mm). Regions with gradient direction difference angles exceeding 30° are marked using dynamic threshold segmentation (an improved version of the Otsu algorithm). A spatial interpolation algorithm is then used to generate a topological feature deviation heatmap covering the entire region (deviation on a 0-1 scale). Finally, a deviation matrix containing curvature abrupt change points (κ ≥ 0.5 mm⁻¹) and dielectric anomaly core regions (εr ≥ 3.5) is output, achieving preliminary quantitative localization of mixed defects. 502. Dynamically correlate and match the spatial topological feature deviation with the phase angle offset of the standard dielectric model to generate an anticoagulant distribution anomaly index; in step 502, the phase angle offset of the standard dielectric model refers to the change in phase angle at each point in the standard model. Dynamic correlation and matching refers to nonlinearly aligning and comparing two feature sequences. The anticoagulant distribution anomaly index refers to a quantitative indicator of the anticoagulant distribution deviating from the normal state. In this embodiment, based on the NURBS parametric surface of the phase angle offset of the standard dielectric model, the dynamic time warping (DTW) algorithm is used to nonlinearly align the spatial topological feature deviation sequence (sampling rate 500Hz) with the standard phase angle sequence, and calculate the phase lag (Δθ=0.1-0.8rad) within a local window (5×5 voxels). Principal component analysis (PCA) was used to extract the main features of deviation (contribution rate ≥85%). The deviations of the axial layer (weight 0.5), radial layer (0.3) and tangential layer (0.2) were weighted and fused using an improved entropy weighting method to generate an anticoagulant distribution anomaly index (0-100% dynamic range). Regions with an anomaly index exceeding the threshold (≥80%) were identified as the core area of ​​concentration gradient imbalance, thus completing the correlation mapping between defects and dielectric anomalies.

[0054] 503. Based on the anticoagulant distribution anomaly index, divide the spatial coordinate set into anomaly regions, extract the co-variation parameters of the anticoagulant concentration gradient and phase angle offset difference in the anomaly regions, and generate the initial weight coefficients of the hybrid defect assessment matrix; in step 503, the anomaly region refers to the spatial region where the anticoagulant distribution anomaly index exceeds a threshold. The anticoagulant concentration gradient refers to the rate of change of anticoagulant concentration in space. The phase angle offset difference refers to the deviation of the actual phase angle from the standard value. The co-variation parameter refers to the degree of correlation between the concentration gradient and the phase angle offset difference. The initial weight coefficients of the hybrid defect assessment matrix refer to the initial weight values ​​of each dimension in the defect assessment matrix.

[0055] In this embodiment, within the abnormal region, an improved DBSCAN clustering algorithm (eps=0.3mm, minPts=8) is used to delineate the anticoagulant concentration gradient imbalance region (concentration difference ≥0.3mg / mL). A sliding window covariance analysis (window size 3×3×3 voxels) is used to extract the co-variance parameters (Pearson correlation coefficient r≤-0.7) between the phase angle offset difference (Δφ≥0.6°) and the concentration gradient. Fuzzy logic rules (membership function μ=0.65) are used to convert the co-variance parameters into initial weight coefficients (0-1 scale). The weights are adaptively adjusted based on the spatial frequency distribution (dominant frequency 3-7 cycles / mm) to generate an initial weight coefficient matrix containing defect density (≥85%), phase hysteresis intensity (≥0.5rad), and spatial correlation, providing benchmark parameters for the multi-dimensional evaluation of mixed defects.

[0056] 504. Dynamically optimize the initial weight coefficients of the hybrid defect assessment matrix based on the aforementioned collaborative change parameters. This dynamic optimization process adjusts the matrix dimension correlation strength by matching the cumulative threshold of the phase angle offset difference with the distribution trend of the anticoagulant concentration gradient, generating a multi-dimensional hybrid defect assessment matrix that integrates spatial topological feature deviation and phase angle offset difference. In step 504, dynamic optimization refers to the process of adjusting the weight coefficients based on new data. The cumulative threshold of the phase angle offset difference refers to the total threshold of the phase angle deviation. The distribution trend of the anticoagulant concentration gradient refers to the spatial variation law of the concentration gradient. The matrix dimension correlation strength refers to the degree of correlation between the dimensions in the matrix. The multi-dimensional hybrid defect assessment matrix refers to a defect assessment matrix that integrates multiple feature dimensions.

[0057] In this embodiment, an improved genetic algorithm (GA, population size 50, iterations 5) is used to optimize the initial weight coefficients. The objective function is defined as the matching degree between the cumulative threshold of phase angle offset difference (ΣΔφ≥1.2rad) and the distribution trend of anticoagulant concentration gradient (slope ≤-0.5mg / mL / mm). The matrix dimensional correlation strength (R²≥0.9) is verified by finite element simulation. Tensor decomposition (CP decomposition, rank=3) is used to fuse spatial topological feature deviation (dimensional 32) and phase angle difference (dimensional 16). Principal component analysis (preserving variance ≥95%) is then used to reduce the dimensionality to an 8-dimensional feature space. Finally, a multi-dimensional hybrid defect evaluation matrix is ​​generated, which includes defect type (concentration imbalance / eddy current decay), spatial hierarchy weight (0.1-0.9), and confidence level (≥90%), realizing a holographic quantitative characterization of defects in complex fluid systems.

[0058] Here is a specific example: In the anomaly detection of an intelligent platelet apheresis system, a blood center found abnormal platelet aggregation (aggregation degree 82%) in a longitudinal 18-35mm segment of blood collection tubes using heparin sodium anticoagulant. The system acquired a set of spatial coordinates through a three-dimensional dielectric sensor array (accuracy 0.2mm) and detected a gradient difference (Δ dielectric constant 0.78) between the topological feature density distribution and the standard dielectric model at coordinates (42mm, 15mm), generating a spatial topological feature deviation of 0.35 / mm². After dynamic correlation with the phase angle offset (0.45rad, normal threshold 0.2rad), the anticoagulant distribution anomaly index reached 4.2 (threshold 1.8). When dividing the abnormal region, the synergistic parameter (correlation coefficient 0.89) of the anticoagulant concentration gradient deviation ±28% and the phase angle offset in that region was extracted to construct the initial weight coefficients [0.38, 0.42, 0.2] of the hybrid defect assessment matrix. By dynamically optimizing the phase accumulation threshold (2.3 times) and concentration distribution trend, and adjusting the matrix correlation strength to improve compensation efficiency by 35%, a multi-dimensional assessment matrix was generated to trigger level 3 compensation: increasing the anticoagulant pump speed to 140% of the baseline value, and simultaneously activating the centrifugal mixing device (2200 rpm) and the pipeline temperature control module (37±0.5℃). Post-intervention follow-up showed that the heparin sodium distribution fluctuation narrowed to ±7%, platelet aggregation decreased to 9%, and the phase shift returned to 0.18 rad, validating the effectiveness of multi-field coupling compensation.

[0059] In summary, steps 501 to 504 achieve intelligent quantitative assessment of dynamic distortion and spatial defects caused by dielectric anomalies during anticoagulant mixing. By fusing geometric invariance identification of phase angle offset with dynamic distortion pattern analysis, a multi-level difference fusion model is constructed to accurately locate the coupling region between abnormal anticoagulant concentration gradient and hydrodynamic defects within the mixing cavity. Based on the spatial topological correlation between dynamic distortion feature set and phase angle difference vector, the system overcomes the limitations of traditional static parameter detection, significantly improves the quantitative accuracy and traceability of mixing defects, and enables real-time visual diagnosis of mixing process quality. This provides a complete chain solution for defect location, assessment, and optimization in complex fluid industrial processes.

[0060] In some embodiments, step 105, which involves constructing a data anomaly assessment function based on the anticoagulant distribution anomaly index and the mixed defect assessment matrix, and dynamically associating the output value of the assessment function with the mapping relationship of the clinical test error threshold to generate test data assessment results, includes: 601, decomposing the spatial density gradient of the anticoagulant distribution anomaly index and the vector magnitude of the mixed defect assessment matrix, marking the spatial coordinate offset of the abnormal region, and generating a matrix feature set of the anomaly index and defects; in step 601, the spatial density gradient of the anticoagulant distribution anomaly index refers to the rate of change of the anomaly index in space. The vector magnitude of the mixed defect assessment matrix refers to the L2 norm of each row vector in the matrix. The spatial coordinate offset of the abnormal region refers to the deviation of the geometric center of the abnormal region from the standard position. The matrix feature set of the anomaly index and defects refers to a feature dataset containing the anomaly index, vector magnitude, and spatial offset.

[0061] In this embodiment, the Gaussian kernel density estimation (KDE) algorithm is used to nonparametrically decompose the spatial density gradient (resolution 0.1 mm) of the anticoagulant distribution anomaly index. Simultaneously, the L2 norm (magnitude range 0.1-2.5) of each row vector in the hybrid defect assessment matrix is ​​calculated as a defect intensity index. An improved K-means clustering method (k=5-8) is used to jointly cluster the spatial density gradient (bandwidth h=0.5 mm) and vector magnitude, identifying regions with both high density and high magnitude (density ≥ 85%, magnitude ≥ 1.8). Delaunay triangulation is used to mark the geometric center coordinate offset (ΔXYZ ≥ 0.3 mm). Finally, a three-dimensional matrix feature set (data dimension = 256 × 256 × 3) containing the anomaly index amplitude, defect magnitude, and spatial offset is generated, providing structured input for multi-parameter fusion analysis. 602. Construct a weight function for the product of the spatial density of the matrix features and the vector magnitude, associate it with the segmented interval weights of the clinical test error threshold, and generate a data anomaly assessment function; In step 602, the spatial density of the matrix features refers to the distribution density of the anomaly index in space. The weight function for the product of the vector magnitude refers to the product function of the spatial density and the vector magnitude. The segmented interval weights of the clinical test error threshold refer to the weight values ​​corresponding to different error intervals. In this embodiment, based on the matrix feature set, a weight function for the product of spatial density (ρ) and vector magnitude (‖V‖) is constructed: W=αρ+(1-α)‖V‖ (α=0.6), and the weight coefficient α is dynamically adjusted using the entropy weight method (step size Δα=0.05). The segmented interval weights of the clinical test error threshold (CLIA standard) are associated (error ≤5% weight 0.2, 5-10% weight 0.5, ≥10% weight 0.8), and the threshold intervals are mapped to weight adjustment factors β (0.3-1.0) through fuzzy logic rules (membership function μ=0.7). The gradient descent algorithm is used to optimize the function parameters (learning rate η=0.01), generating a set of dynamic parameters for the data anomaly assessment function that includes dynamic weights (W'=βW), error sensitivity (γ=0.05-0.2), and spatial constraints (λ=0.1), thereby realizing the quantitative correlation between clinical standards and mixed defects.

[0062] 603. Quantify the segmented interval fluctuation amplitude of the data anomaly assessment function, map it to the probability distribution boundary of the clinical test error threshold, and generate a dynamic mapping relationship set between the assessment function output and the threshold interval; In step 603, the segmented interval fluctuation amplitude of the data anomaly assessment function refers to the fluctuation intensity of the function within each segmented interval. The probability distribution boundary of the clinical test error threshold refers to the boundary range of the error threshold in the probability distribution. The dynamic mapping relationship set between the assessment function output and the threshold interval refers to the matching relationship data between the function output and the error threshold. In this embodiment, wavelet packet transform (db4 wavelet, decomposition level = 5) is used to decompose the data anomaly assessment function in the time-frequency domain to extract the fluctuation amplitude (E = 0.1-1.2mV²) of each segmented interval (bandwidth 0.5Hz). The probability distribution boundary of the clinical test error threshold is fitted by kernel density estimation (KDE, bandwidth = 0.2) (confidence level 95%), and the dynamic time warping (DTW) algorithm is used to nonlinearly align the assessment function output sequence with the threshold boundary, and the mapping offset (ΔT ≤ ±10ms) is calculated. Generate a dynamic mapping set containing fluctuation amplitude-probability boundary matching degree (R²≥0.85), time-domain phase difference (Δφ=0-π / 2), and confidence interval (CI=90-99%) to achieve a probabilistic association between the evaluation results and clinical standards. 604. Extract abnormal boundary regions exceeding the clinical test error threshold from the dynamic mapping set, extract the phase offset of the data anomaly assessment function, and generate a set of abnormal test data events;

[0063] In step 604, the dynamic mapping relationship set refers to the mapping data between the evaluation function output and the error threshold. The clinical test error threshold refers to the clinically acceptable deviation threshold of the test result. The abnormal boundary region refers to the spatial range exceeding the error threshold. The phase offset of the data anomaly evaluation function refers to the phase delay of the function output relative to the reference signal. The test data anomaly event set refers to the dataset containing the anomaly event type, spatial coordinates, and phase offset. In this embodiment, based on the dynamic mapping relationship set, an improved level set method (iterations = 5) is used to extract the abnormal boundary region (area ≥ 0.2 mm²) exceeding the clinical test error threshold (Δ ≥ 8%). The intrinsic mode functions (IMFs) within the boundary region are extracted using the Hilbert-Huang transform, and their instantaneous phase offset (Δθ = 0.1-0.5 rad) is calculated. An improved DBSCAN clustering method (eps=0.3mm, minPts=5) was applied to spatially aggregate the phase offset, generating a set of abnormal events in the inspection data that includes the abnormal event type (concentration imbalance / eddy current decay), spatial coordinates, and phase offset (0.2-1.0rad), thus achieving refined classification of defect events.

[0064] 605. Aggregate the phase offset of the set of abnormal test data events with the boundary parameters of the dynamic mapping relationship set, arrange them according to the spatial coordinate hierarchy of the anticoagulant distribution anomaly index, and generate the test data evaluation result. In step 605, the phase offset of the set of abnormal test data events refers to the phase delay of the abnormal event relative to the reference signal. The boundary parameters of the dynamic mapping relationship set refer to the boundary constraints in the mapping relationship. The spatial coordinate hierarchy of the anticoagulant distribution anomaly index refers to the hierarchical structure of the anomaly index in space. The test data evaluation result refers to a comprehensive judgment report including defect severity and clinical risk level.

[0065] In this embodiment, tensor decomposition (CP decomposition, rank=3) is used to fuse the phase offset (dimension=32) of the abnormal event set of the test data with the boundary parameters (dimension=16) of the dynamic mapping relationship. Principal component analysis (PCA, retaining variance ≥95%) is then used to reduce the dimensionality to an 8-dimensional feature space. Based on the spatial coordinate hierarchy of the anticoagulant distribution anomaly index (axial layer thickness 0.5mm), spatial coding technology (Z-order curve) is used to arrange the fused features in hierarchical order, constructing a structured test data evaluation result that includes defect severity (0-5 levels), clinical risk level (AD level), and spatial positioning accuracy (±0.05mm). The output format is compatible with the HL7 standard and can be directly connected to the hospital LIS system.

[0066] Here is a specific example: In a smart apheresis platelet quality monitoring scenario, a blood center discovered abnormal platelet aggregation in blood collection tubes using EDTA-K2 anticoagulant. The system constructed a three-dimensional anticoagulant distribution map using a ring array eddy current sensor, detecting an abnormal spatial density gradient in the longitudinal 28-45mm segment, where the vector magnitude of the node at coordinates (35mm, 22mm) reached 4.7 (normal threshold is 2.8). When constructing the product weight function, combined with the clinical test error threshold (platelet count allowable deviation ±15%), a dynamic parameter set was generated showing that the data anomaly assessment value for this region was 3.8 (threshold 1.5). Quantitative analysis revealed a phase offset of 0.35 rad (normal 0.15 rad), and when mapped to the error probability distribution boundary, this region exceeded the clinical threshold range by 2.6 times. The system extracted the abnormal boundary to generate an event set, identifying the abnormal core area of ​​the anticoagulant distribution index at Z-axis levels 3-5 (corresponding to the middle layer of the blood collection tube), with an ACD infusion rate deviation of 32% (standard value 18:1). The final aggregated data triggered a three-level compensation mechanism: automatically increasing the anticoagulant pump speed to 135% of the baseline value, simultaneously activating the centrifugal mixing device. After a 30-second intervention, a follow-up examination showed that the EDTA concentration fluctuation narrowed to ±6%, platelet aggregation decreased to 12%, and the dynamic mapping parameters returned to the threshold range. This solution was successfully applied to a blood center in Shanghai, improving the efficiency of handling abnormal blood collection events by 40%.

[0067] In summary, steps 601 to 605 achieve multi-dimensional dynamic assessment of anticoagulant mixing defects and intelligent correlation analysis of clinical laboratory errors. By integrating spatial density gradient, vector magnitude features, and clinical laboratory thresholds, a dynamic parameter-driven anomaly assessment system is constructed. Based on product weighting functions and probability boundary mapping technology, the system accurately identifies abnormal boundary regions exceeding clinical error thresholds and automatically generates a set of abnormal laboratory data events including phase offset and spatial coordinate levels. This overcomes the limitations of traditional single-indicator judgment, significantly improves the sensitivity and clinical relevance of mixing defect detection, and achieves closed-loop management across the entire chain from data anomaly detection to clinical decision support, providing real-time and accurate assessment basis for medical quality control.

[0068] In some embodiments, step 102, analyzing the curvature abrupt change characteristics of the diffusion trajectory morphological parameters, triggering the acoustic-fluid mixing device to modulate the acoustic wave frequency and fluid flow velocity, and forming a composite disturbance field within the mixing cavity of the acoustic-fluid mixing device, includes: 701, decomposing the local curvature abrupt change characteristics of the diffusion trajectory morphological parameters, calculating the rate of curvature change at the curvature abrupt change point and the curvature gradient direction of the adjacent region in the local curvature abrupt change characteristics, and generating a curvature abrupt change parameter set; in step 701, the local curvature abrupt change characteristics of the diffusion trajectory morphological parameters refer to the local region characteristics where the curvature changes significantly in the diffusion trajectory. The rate of curvature change at the curvature abrupt change point refers to the rate of change of curvature over time at the curvature abrupt change point. The curvature gradient direction of the adjacent region refers to the direction of curvature change in the region surrounding the curvature abrupt change point. The curvature abrupt change parameter set refers to a feature dataset containing the coordinates, rate of change, and gradient direction of the curvature abrupt change point.

[0069] In this embodiment, an improved Canny edge detection algorithm is used to identify abrupt changes in the local curvature of the diffusion trajectory. The rate of change of curvature (Δκ / Δt ≥ 0.8 mm⁻² / ms) is determined by calculating the second derivative of curvature (d²κ / ds²), and the curvature gradient direction (θ = 0-360°) of adjacent regions is extracted using the Gaussian gradient operator. False abrupt changes are eliminated using non-maximum suppression (NMS) technology, retaining valid points where the rate of change of curvature exceeds a threshold (threshold = 1.2 mm⁻²) and the gradient direction is consistent (standard deviation of direction angle ≤ 15°). A parameter set (data dimension = 256 × 3) containing the coordinates of curvature abrupt changes, the rate of change, and the gradient vector is generated through cubic spline interpolation, providing high-precision input for acoustic fluid parameter matching.

[0070] 702. Based on the curvature change rate and gradient direction of the curvature mutation parameter set, match the acoustic frequency band range of the acoustic fluid mixing device with the linear response range of the fluid flow velocity to generate acoustic frequency modulation parameters; in step 702, the acoustic frequency band range of the acoustic fluid mixing device refers to the operating frequency range of the acoustic transmitter. The linear response range of the fluid flow velocity refers to the range of the linear relationship between the flow velocity and the acoustic frequency. The acoustic frequency modulation parameters refer to the set of parameters used to optimize the matching of acoustic frequency and flow velocity. In this embodiment, based on the curvature mutation parameter set, the Fast Fourier Transform (FFT) is used to analyze the linear response characteristics of the acoustic frequency band (1-5MHz) and the fluid flow velocity (0.5-5mL / s) to construct a sound pressure-flow velocity transfer function model (H(f)=A·e^(-jωτ)). By fitting the correlation coefficient matrix (R²≥0.85) between the rate of curvature change (Δκ) and the acoustic frequency (f) using the least squares method, and combining the gradient direction constraint (flow velocity direction matching within θ±30°), an acoustic frequency modulation parameter table containing the optimal frequency modulation range (Δf=±0.3MHz) and the flow velocity adjustment gradient (Δv=0.2-1.2mL / s) is generated, thereby achieving quantitative matching of acoustic-flow coupling parameters.

[0071] 703. Based on the linear response range between the acoustic frequency modulation parameters and the fluid flow velocity, adjust the phase synchronization sequence of the acoustic transmitter and the flow velocity gradient parameters of the fluid pump within the mixing chamber to generate a set of coupling parameters between the acoustic wave and the fluid. In step 703, the linear response range of the fluid flow velocity refers to the range of the linear relationship between the flow velocity and the acoustic frequency. The phase synchronization sequence of the acoustic transmitter refers to the phase adjustment sequence of the acoustic transmitter. The flow velocity gradient parameters of the fluid pump refer to the parameters of the flow velocity change over time or space. The set of coupling parameters between the acoustic wave and the fluid refers to the parameter set containing the acoustic phase sequence and the flow velocity gradient. In this embodiment, a phase synchronization control algorithm (PLL) is used to convert the acoustic frequency modulation parameters into a phase sequence of the acoustic transmitter (phase resolution 0.1°), and frequency tracking is achieved by adjusting the piezoelectric ceramic driving voltage (accuracy ±0.05V) through PID closed-loop control. The pulse frequency (resolution 0.1Hz) of the stepper motor of the micro peristaltic pump is dynamically adjusted synchronously based on the velocity gradient parameter. Computational fluid dynamics (CFD) simulation is used to verify the acoustic-velocity coupling effect (vorticity ω ≥ 20s⁻¹). A set of coupling parameters is generated, including phase sequence (32-bit encoding), velocity gradient (Δv = 0.5mL / s), and sound pressure level (SPL = 140dB), ensuring the spatiotemporal consistency of the acoustic-fluid synergy. 704. The dynamic matching rules between the phase synchronization sequence and the velocity gradient parameter are optimized by traversing the area of ​​the acoustic transmitter and the fluid pump covered by the coupling parameter set, generating a set of anti-diffusion disturbance rules.

[0072] In step 704, the interaction area between the acoustic wave transmitter and the fluid pump refers to the spatial range in which the acoustic wave and the fluid interact. The phase synchronization sequence refers to the phase adjustment sequence of the acoustic wave transmitter. The velocity gradient parameter refers to the parameter that changes the velocity over time or space. The dynamic matching rule refers to the set of rules used to optimize the matching between the acoustic wave and the velocity. The anti-diffusion disturbance rule set refers to the dataset of optimization rules used to eliminate diffusion resistance. In this embodiment, based on the set of coupling parameters, a genetic algorithm (GA) is used to perform multi-objective optimization on the acoustic wave transmitter array (8×8 layout) and the fluid pump interaction area (axial segment length 2mm): the objective functions are defined as vorticity uniformity (σ≤0.1) and energy efficiency (η≥80%). The optimal combination of phase sequence (fitness ≥ 0.9) and flow rate gradient parameter (Δv = 0.8 mL / s) is iteratively screened through crossover and mutation operations (crossover rate 0.7, mutation rate 0.05). Particle swarm optimization (PSO) is used to dynamically adjust the matching rule weights (sound weight 0.6, flow rate weight 0.4), and finally a rule set containing 32 sets of anti-diffusion disturbance rules (such as frequency-flow rate matching pairs and phase delay compensation) is generated to achieve adaptive control of the disturbance field.

[0073] 705. Map the dynamic matching threshold of the anti-diffusion disturbance rule set to the phase synchronization distribution of the coupling parameter set to form a composite disturbance field that eliminates diffusion resistance at the contact surface. In step 705, the dynamic matching threshold refers to the critical value used to determine the matching effect between sound waves and flow velocity. The phase synchronization distribution of the coupling parameter set refers to the spatial distribution characteristics of the sound wave phase. The composite disturbance field refers to the multidimensional disturbance field formed by the interaction of sound waves and fluid. In this embodiment, the anti-diffusion disturbance rule set is mapped to the three-dimensional space of the hybrid cavity using the field superposition principle. The coupling strength (τ≥0.5Pa) between the sound wave phase distribution (phase difference ≤5°) and the flow velocity gradient field is calculated using finite element analysis (FEA). The sound pressure cloud map (resolution 0.1mm) and the flow velocity vector field are fused using the inverse distance weighted interpolation (IDW) algorithm to generate a shear stress peak region (τ_max=0.8Pa) at the contact surface (interface thickness ≤0.05mm), forming a composite disturbance field that eliminates diffusion resistance. Verification using a laser Doppler velocimeter (LDV) showed that the uniformity of vorticity within the disturbed field was increased to 95%, and the diffusion rate increased from 1.2 mm / s to 2.5 mm / s, achieving dynamic suppression of contact surface resistance.

[0074] Here is a specific example: In a smart apheresis platelet mixed monitoring scenario, a blood center discovered that uneven distribution of EDTA anticoagulant caused platelet aggregation (aggregation rate up to 85%). The system captures the diffusion trajectory of the mixed fluid through an eddy current sensor array, identifies curvature abrupt change points in the longitudinal 28-45mm segment, and generates a curvature abrupt change parameter set containing 12 abnormal nodes. The system matches the 400-600kHz linear response frequency band of the acoustic fluid device to generate acoustic modulation parameters. The phase sequence of the acoustic transmitter (phase difference ±15°) and the fluid pump pulsation gradient (0.05mL / s²) are adjusted to form a coupled parameter set covering an area with a radius of 4.7mm. When optimizing the dynamic matching rules, it was found that when the flow velocity gradient > 0.1mL / s² and the acoustic frequency offset < 2%, the anti-diffusion efficiency improves by 40%, ultimately generating a composite perturbation field to eliminate contact surface resistance. After implementation, the EDTA concentration fluctuation decreased from ±23% to ±6%, and the platelet aggregation returned to the normal range, simultaneously triggering a synergistic intervention mechanism of tubing preheating (37℃) and centrifugation mixing.

[0075] In summary, steps 701 to 705 achieve intelligent synergistic control of diffusion resistance and fluid disturbance during the anticoagulant-blood mixing process. By integrating curvature change feature analysis and dynamic matching of acoustic wave-flow velocity parameters, a multi-modal parameter synergistic composite disturbance field is constructed. Based on adaptive identification of curvature change rate and gradient direction, the system optimizes the coupling strength between acoustic wave phase sequence and fluid shear force in real time, overcoming the limitations of traditional static parameter control. This significantly improves the diffusion efficiency and distribution uniformity of the anticoagulant within the mixing chamber, effectively eliminates diffusion resistance at the contact surface, and ensures the high efficiency and stability of the mixing process. It provides a closed-loop solution for dynamic disturbance optimization in complex fluid industrial mixing processes. In some embodiments, step 704, which involves traversing the operational regions of the acoustic transmitter and fluid pump covered by the set of coupling parameters, optimizing the dynamic matching rules of the phase synchronization sequence and the velocity gradient parameters, and generating a set of anti-diffusion disturbance rules, includes: 801, analyzing the spatial distribution characteristics of the acoustic transmitter phase synchronization sequence and the fluid pump velocity gradient parameters, extracting the temporal fluctuation amplitude of the acoustic wave propagation constant and the fluid velocity, and generating a set of dynamic response parameters; in step 801, the acoustic transmitter phase synchronization sequence refers to the set of phase adjustment order and time intervals between elements in a multi-element acoustic transmitter. The fluid pump velocity gradient parameter refers to the rate of change of the fluid pump output velocity in the spatial dimension. The acoustic wave propagation constant refers to the composite parameter describing the amplitude attenuation coefficient and the phase change rate when the acoustic wave propagates in the medium.

[0076] The temporal fluctuation amplitude of fluid velocity refers to the maximum fluctuation amplitude of the velocity signal in the time dimension. The dynamic response parameter set refers to a multi-dimensional feature dataset that integrates the acoustic wave propagation characteristics and the dynamic fluctuation of fluid velocity. In this embodiment, a multi-physics coupling analysis method is used to analyze the spatial distribution characteristics of the acoustic wave transmitter phase synchronization sequence (sampling rate 1MHz) through time-frequency domain decomposition technology (such as wavelet packet transform). Combined with laser Doppler velocimetry data of fluid pump velocity gradient parameters (axial gradient 0.2-1.5mL / s²), the acoustic wave propagation constant (attenuation coefficient α=0.1-0.5dB / mm) and the temporal fluctuation amplitude of fluid velocity (peak-to-peak value 0.3-1.2m / s) are extracted. Improved principal component analysis (PCA) is used to reduce the dimensionality of the acoustic-flow parameters (preserving variance ≥95%), generating a dynamic response parameter set that includes the acoustic wave propagation direction angle (θ=0-360°), velocity gradient amplitude, and fluctuation frequency, providing a multi-dimensional data basis for collision detection.

[0077] 802. Traverse the sound wave propagation direction and fluid velocity gradient direction in the dynamic response parameter set, calculate the phase shift of the sound wave amplitude attenuation rate and the dynamic response of the fluid velocity, and generate a dynamic response difference matrix; in step 802, the dynamic response parameter set refers to the sound wave-fluid coupling dynamic feature dataset generated in step 801. The sound wave propagation direction refers to the angle of the main vector direction of sound wave energy transmission in space (based on the coordinate system). The fluid velocity gradient direction refers to the direction vector of the maximum increase of the velocity change rate in space. The sound wave amplitude attenuation rate refers to the amount of attenuation of the sound wave amplitude per unit propagation distance or time. The phase shift of the fluid velocity dynamic response refers to the phase angle (in radians) corresponding to the time delay of the velocity change relative to the sound wave excitation signal. The dynamic response difference matrix refers to a multi-dimensional matrix data structure constructed by quantifying the sound wave-fluid response difference. In this embodiment, based on the dynamic response parameter set, a cross-correlation algorithm is used to calculate the cosine similarity (threshold ≥ 0.8) of the angle between the sound wave propagation direction and the flow velocity gradient direction. Hilbert transform is used to extract the sound wave amplitude attenuation rate (ΔA / Δt = 0.05-0.3 dB / ms) and the instantaneous phase shift of the flow velocity response (Δφ = 0.1-0.8 rad). Tensor operations are used to construct a three-dimensional dynamic response difference matrix (dimension = 32 × 32 × 3), where the matrix elements are weighted and fused from the sound-flow direction matching degree, amplitude attenuation rate, and phase shift (weighting coefficients α = 0.6, β = 0.4), thus achieving a quantitative characterization of the sound-flow coupling effect.

[0078] 803. Based on the amplitude attenuation rate and phase offset of the dynamic response difference matrix, detect the conflict region between the sound wave propagation path and the fluid velocity gradient, and construct a dynamic response conflict node set; in step 803, the dynamic response difference matrix refers to the set of quantified data on the difference between the sound wave and the fluid response generated in step 802. The conflict region between the sound wave propagation path and the fluid velocity gradient refers to the spatial region where the angle between the sound wave energy transfer direction and the velocity gradient direction exceeds a threshold (e.g., >90°) or where the energy attenuation rate does not match the velocity response. The dynamic response conflict node set refers to a structured dataset that marks the spatial coordinates of the conflict region, the conflict type (directional conflict / energy conflict), and the conflict severity level. In this embodiment, a density clustering algorithm (DBSCAN, eps=0.3mm, minPts=5) is applied to spatially scan the dynamic response difference matrix to identify conflict regions where the amplitude attenuation rate exceeds the limit (ΔA / Δt≥0.2dB / ms) and the phase offset is reversed (Δφ≥π / 4). Finite element analysis (FEA) was used to verify the shear stress conflict (τ≥0.6Pa) between the sound wave propagation path and the velocity gradient field. A dynamic response conflict node set was constructed, including coordinates (accuracy ±0.05mm), conflict type (sound pressure-velocity direction conflict / energy attenuation conflict), and severity level (1-5), to locate 93% of the diffusion resistance sources within the mixing cavity. 804. Within the area covered by the dynamic response conflict node set, the phase synchronization frequency band of the sound wave transmitter and the velocity gradient threshold of the fluid pump were adjusted to generate a dynamic matching rule optimization framework.

[0079] In step 804, the dynamic response conflict node set refers to the set of characteristic data of the acoustic-fluid conflict region detected in step 803. The phase synchronization frequency band of the acoustic transmitter refers to the effective frequency band within the acoustic frequency range where the phase relationship can be dynamically adjusted. The velocity gradient threshold of the fluid pump refers to the maximum critical value of the allowable velocity change rate (e.g., Δv_max = 1.0 mL / s²). The dynamic matching rule optimization framework refers to the optimization rule structure generated by adjusting the acoustic-velocity parameter matching relationship. In this embodiment, a genetic algorithm (GA) is used to optimize the acoustic phase synchronization frequency band (frequency tuning step size 0.05 MHz) and the velocity gradient threshold (Δv = 0.1-0.9 mL / s²) within the conflict node set coverage area. The fitness function is defined as the weighted sum of vorticity uniformity (σ ≤ 0.1) and energy efficiency (η ≥ 75%), and the optimal parameter combination is iteratively screened through crossover mutation (crossover rate 0.7, mutation rate 0.03). Particle swarm optimization (PSO) is used to dynamically adjust the acoustic-flow weights (sound wave 0.65, flow velocity 0.35) to generate a dynamic matching rule optimization framework that includes frequency-flow velocity matching rules (e.g., 2.2MHz corresponds to 0.6mL / s), phase delay compensation (Δφ=±0.1rad), and spatial constraints. 805. The dynamic matching rule optimization framework is dynamically matched with the phase offset parameters of the conflict node set to generate and output an anti-diffusion disturbance rule set. In step 805, the phase offset parameters of the conflict node set refer to the phase delay angle data of the flow velocity response relative to the acoustic excitation within the conflict region.

[0080] The anti-diffusion disturbance rule set refers to a set of control parameters generated through dynamic matching to suppress diffusion resistance. In this embodiment, a fuzzy logic inference engine is used to dynamically adapt the phase offset parameters of the dynamic matching rule optimization framework and the conflict node set, and a rule triggering mechanism (confidence ≥ 85%) is used to screen effective anti-diffusion rules. The rules are mapped to the three-dimensional space of the hybrid cavity using the field superposition principle to generate an anti-diffusion disturbance rule set that includes a sound pressure cloud map (resolution 0.1 mm), a flow velocity vector field, and a shear stress peak region (τ_max = 0.8 Pa). By real-time control of the piezoelectric array (phase error ≤ 0.5°) and the micro-pump (flow velocity error ± 0.02 mL / s), the diffusion resistance is reduced by 62%, verifying the engineering effectiveness of the rule set in complex flow fields.

[0081] Here is a specific example: In a smart apheresis platelet anticoagulant mixing monitoring scenario, a blood center discovered an abnormal EDTA-K2 distribution leading to platelet aggregation of 85%. The system constructed a three-dimensional phase distribution map using a ring array eddy current sensor, detecting anomalies in the acoustic propagation constant and fluid velocity gradient deviation (±22%) in the longitudinal 28-45mm segment, generating a parameter set containing 12 dynamic response nodes. Calculations revealed a conflict between the acoustic amplitude attenuation rate and the velocity phase offset (32°) in region 3, constructing a dynamic response conflict node set covering a 4.7mm radius region. The acoustic transmitter phase synchronization frequency band was adjusted to 420-580kHz, simultaneously increasing the fluid pump velocity gradient threshold to 0.12mL / s², generating a dynamic matching rule that improved anticoagulant mixing efficiency by 38%. After implementation, EDTA concentration fluctuations narrowed from ±23% to ±6%, and platelet aggregation decreased to 12%, triggering a collaborative intervention mechanism of centrifugal mixing device and pipeline preheating (37℃), verifying the effectiveness of the anti-diffusion disturbance rule.

[0082] In summary, steps 801 to 805 achieve intelligent optimization of dynamic coordinated control of acoustic wave and fluid parameters. By analyzing the spatiotemporal distribution characteristics of acoustic wave propagation and fluid velocity, a dynamic response difference-driven anti-diffusion disturbance system is constructed. Based on the conflict detection of acoustic wave amplitude attenuation rate and flow velocity phase offset, the system accurately locates the acoustic-flow coupling conflict region, optimizes the phase synchronization frequency band and flow velocity gradient threshold in real time, overcomes the limitations of traditional static parameter matching, significantly improves acoustic-flow coordination efficiency and disturbance field uniformity, effectively suppresses diffusion resistance and enhances mixing stability, and provides a dynamic parameter adaptive anti-diffusion solution for precision industrial processes such as microfluidic reactions and drug synthesis. In some embodiments, step 201, which involves fusing the curvature distribution of the diffusion trajectory morphology parameters, the parameters of the composite perturbation field, and the correlation of the spatial coordinate set, and combining the evaluation results of the test data to generate collaborative compensation parameters, includes: 901, decomposing the curvature density gradient of the high-frequency abrupt region and the low-frequency smooth region in the curvature distribution of the diffusion trajectory morphology parameters, quantifying the spatial weight ratio of different curvature intervals, and generating a weight mapping table for the anticoagulant diffusion path; in step 901, the curvature distribution of the diffusion trajectory morphology parameters refers to the spatial variation characteristics of the curvature of the anticoagulant diffusion path.

[0083] The high-frequency abrupt change region refers to a spatial region with drastic curvature changes and a frequency higher than 5 cycles / mm. The low-frequency smooth region refers to a spatial region with gentle curvature changes and a frequency lower than 1 cycle / mm. The curvature density gradient refers to the difference in the amplitude of curvature changes per unit area. The curvature interval refers to different ranges divided according to curvature values ​​(such as low, medium, and high curvature bands). The spatial weight proportion refers to the quantified proportion of the influence of different curvature intervals on the overall diffusion path. The weight mapping table refers to an optimized path reference table that marks the correspondence between curvature weights and spatial coordinates. In this embodiment, this step uses wavelet packet decomposition technology to divide the curvature distribution signal of the diffusion trajectory into a high-frequency abrupt change region (frequency ≥ 5 cycles / mm) and a low-frequency smooth region (frequency ≤ 1 cycle / mm) in the frequency domain, and calculates the curvature density gradient of each frequency band using the Morlet wavelet function. An improved K-means clustering algorithm (cluster number k = 3-5) is used to perform density statistics on the spatial distribution of different curvature intervals, and the ratio of the high-frequency region weight (accounting for 35-60%) to the low-frequency region weight (accounting for 40-65%) is calculated. Based on spatial coding technology, the weight ratio and coordinate position (0.1mm resolution) are mapped to a three-dimensional mesh to generate an anticoagulant diffusion path weight mapping table containing curvature type labels (high frequency / low frequency), weight coefficients (0-1 normalized), and gradient vectors, providing a quantitative basis for multi-field coupling compensation. 902. The acoustic frequency spectrum distribution and fluid velocity vector direction in the composite perturbation field parameters are extracted, and the energy proportion of the main acoustic frequency band is fused with the spatial tilt angle of the velocity vector to generate a strongly coupled acoustic-fluid field tensor. In step 902, the composite perturbation field parameters refer to a set of multi-physics field parameters that simultaneously contain acoustic and fluid velocity characteristics.

[0084] The frequency spectrum distribution of sound waves refers to the intensity proportion of sound wave energy in different frequency bands (e.g., 2-8MHz). The direction of the fluid velocity vector refers to the direction vector of the velocity movement in three-dimensional space. The energy proportion of the dominant frequency band of sound waves refers to the intensity proportion of the frequency band where the sound wave energy is concentrated (e.g., energy proportion ≥70%). The spatial tilt angle of the velocity vector refers to the angle between the velocity direction and the reference coordinate axis (e.g., the horizontal axis). The acoustic-flow field strong coupling tensor refers to the three-dimensional matrix data structure that quantifies the synergistic effect of sound wave energy and velocity direction. In the embodiments of this application, the dominant frequency band of the sound wave frequency spectrum in the composite perturbation field is extracted by Fast Fourier Transform (FFT), and the direction of the fluid velocity vector (tilt angle θ = 0-360°, modulus 0.5-3m / s) is analyzed by combining particle imaging velocimetry (PIV) data. Tensor product operations are used to perform high-order fusion of the acoustic wave dominant frequency energy distribution matrix (32×32) and the flow velocity vector direction matrix (32×32×2), generating a third-order acoustic-flow field strongly coupled tensor (dimension = 32×32×5) containing sound pressure energy (140-160dB), flow velocity inclination angle matching degree (cosine similarity 0.6-0.95), and energy-direction coupling coefficient (0.3-0.8), quantifying the spatial intensity distribution characteristics of the acoustic-flow synergistic effect. 903. Traverse the coordinate points of the anticoagulant concentration gradient imbalance region in the spatial coordinate set, and superimpose the weight mapping table and the acoustic-flow field strongly coupled tensor according to the coordinate hierarchy to construct a local compensation intensity matrix for multi-field coupling; In step 903, the spatial coordinate set refers to the three-dimensional coordinate point set of the anticoagulant diffusion path. The anticoagulant concentration gradient imbalance region refers to the spatial region where the concentration change rate exceeds a threshold. The coordinate hierarchy refers to the grid division method of layering along the axial direction (e.g., each layer is 0.1mm along the Z-axis). The local compensation intensity matrix refers to a multi-field collaborative compensation quantification model that combines curvature weights and acoustic-flow coupling intensity. In this embodiment, based on the coordinate set (accuracy ±0.05mm) of the anticoagulant concentration gradient imbalance region, a layered interpolation algorithm is used to superimpose the weight mapping table (layer thickness 0.1mm) and the acoustic-flow field intensity tensor according to axial layers (Z=0-50mm). The combined effect coefficient of curvature weights (0-1) and acoustic-flow coupling intensity (0-1) at each coordinate point is calculated by the improved Hadamard product. The matching degree between the local compensation intensity (compensation amount 0.5-2.0Pa) and the shear stress distribution is verified by finite element simulation (FEA). A 128×128 multi-field coupling local compensation intensity matrix containing curvature-acoustic-flow combined effect intensity, spatial location, and compensation priority is constructed to realize multi-modal compensation quantification modeling of the defect region.

[0085] 904. Inject the anticoagulant distribution anomaly index from the test data evaluation results, correct the dynamic attenuation coefficients of curvature density gradient and acoustic flow field intensity in the multi-field coupling compensation intensity matrix, and generate an optimized compensation parameter kernel function. In step 904, the test data evaluation results refer to the anticoagulant distribution quality detection data based on experiments or simulations (such as CFD). The anticoagulant distribution anomaly index refers to the quantitative index of concentration uniformity deviating from the standard value (such as 0-100%). The dynamic attenuation coefficient refers to the energy loss coefficient of acoustic flow coupling effect changing with time or space. The compensation parameter kernel function refers to the optimized acoustic flow-curvature co-compensation mathematical model. In this embodiment, the anticoagulant distribution anomaly index (AEI=0-100%) based on clinical tests is injected, and a backpropagation neural network (BPNN) is used to establish a prediction model of the dynamic attenuation coefficients (α=0.1-0.9) of curvature density gradient (input layer node=32) and acoustic flow field intensity (input layer node=32) in the compensation intensity matrix. The attenuation coefficient is iteratively optimized using a gradient descent algorithm (learning rate η=0.01) until the mean square error (MSE) between the compensation intensity output value and the measured concentration gradient standard deviation (σ≤0.05mg / mL) converges to a threshold (MSE≤0.001). Finally, an optimized compensation parameter kernel function is generated, containing the spatial attenuation coefficient (α(x,y,z)), dynamic response delay (τ=0.1-0.5s), and energy compensation gain (β=1.2-2.0), achieving dynamic time-space adaptation for defect compensation. 905. Based on the phase modulation characteristics of the compensation parameter kernel function in the spatial coordinate set, the acoustic frequency offset and velocity vector compensation amounts at different coordinate levels are aggregated to generate a multidimensional control sequence containing collaborative compensation parameters. In step 905, the phase modulation characteristics refer to the ability of the compensation parameter kernel function to control the acoustic phase delay (e.g., ±0.2rad).

[0086] The acoustic frequency offset refers to the active adjustment of the acoustic frequency (e.g., ±0.3MHz). The velocity vector compensation refers to the correction value of the velocity direction or magnitude (e.g., Δv=0.1-0.6m / s). The multidimensional control sequence refers to a set of coordinated control commands including acoustic frequency, velocity vector, and phase parameters. In this embodiment, based on the phase modulation characteristics (phase difference Δφ=±0.2rad) of the compensation parameter kernel function, principal component analysis (PCA) is used to reduce the acoustic frequency offset (Δf=±0.3MHz) and velocity vector compensation (Δv=0.1-0.6m / s) at different coordinate levels (layer thickness 0.2mm) to a 3-dimensional feature space. The feature vectors are arranged in axial hierarchical order (Z-axis priority) using spatiotemporal coding (STC) technology to generate a 128-dimensional coordinated compensation control sequence including the frequency modulation sequence (32-bit encoding), the velocity compensation gradient (0.5m / s² step size), and the phase synchronization trigger condition (phase tolerance ≤0.1rad). This sequence can directly drive the acoustic wave emission array (response time ≤ 1ms) and the fluid pump (flow rate accuracy ±0.02m / s), achieving full-domain dynamic optimization of the anticoagulant diffusion path.

[0087] Here is a specific example: In a smart apheresis platelet mixing monitoring scenario, a blood bank found that uneven mixing of EDTA anticoagulant led to platelet aggregation (aggregation degree 85%) in the longitudinal 28-45mm section of the blood collection tube. The system captured the diffusion trajectory through an eddy current sensor array, decomposing it into a high-frequency abrupt change region (curvature density gradient 0.18 rad / mm²) and a low-frequency smooth region (0.06 rad / mm²), generating a weighted mapping table showing that the high-frequency region accounted for 68%. Simultaneously, the 420-580kHz main frequency band (energy percentage 72%) of the acoustic-fluidic device and the velocity vector tilt angle of 32° were extracted to construct a strongly coupled acoustic-fluidic field tensor (coupling coefficient 0.87). Traversing 183 coordinate points in the anticoagulation imbalance region, after superimposing the weights and acoustic-fluidic tensor, a high-compensation region with [the specified value] as its core was identified, generating a local compensation matrix. After injecting the anticoagulant distribution anomaly index (deviation ±23%), the dynamic attenuation coefficient was corrected, improving the curvature gradient compensation efficiency by 40%. This ultimately generated a compensation kernel function to regulate the acoustic frequency offset (±15kHz) and flow rate compensation. After implementation, EDTA concentration fluctuations narrowed to ±6%, triggering a synergistic mechanism of centrifugal mixing and pipeline preheating (37℃), validating the effectiveness of the multidimensional control sequence.

[0088] In summary, steps 901 to 905 achieve intelligent dynamic compensation for the multi-physics coupling effect during anticoagulant mixing. By integrating the curvature distribution characteristics of the diffusion trajectory, the strong coupling parameters of the acoustic-flow field, and the anomaly assessment results of the test data, a compensation kernel function with spatial weight and dynamic attenuation synergistic optimization is constructed. Based on the hierarchical superposition analysis of curvature density gradient and acoustic-flow field strong tensor, the system breaks through the limitations of traditional static parameter control, accurately identifies the high-frequency resistance zone and low-frequency imbalance band in the anticoagulant diffusion path, and generates a multi-dimensional control sequence for acoustic frequency shift-flow velocity compensation in real time. This significantly improves the uniformity of anticoagulant distribution and diffusion efficiency within the mixing chamber, providing a multi-modal data fusion-driven adaptive anticoagulation solution for clinical blood processing equipment.

[0089] Figure 2 This application provides a schematic diagram of the structure of an abnormal blood collection test data evaluation and processing system, as shown in the embodiment of the present application. Figure 2 As shown, the system includes: a tracking module 21, used to track the dynamic changes in the diffusion trajectory of the anticoagulant and blood contact surface in the blood collection tube in real time, and synchronously capture the waveform distortion characteristics of the dielectric response on both sides of the blood contact surface through a capacitance sensor array to generate diffusion trajectory morphology parameters; an analysis module 22, used to analyze the curvature abrupt change characteristics of the diffusion trajectory morphology parameters, trigger the acoustic-fluid mixing device to modulate the acoustic wave frequency and fluid flow velocity, and form a composite disturbance field in the mixing cavity of the acoustic-fluid mixing device; and a construction module 23, used to perform a capacitance gradient axial scan on the mixed fluid after the composite disturbance field, extract the phase angle offset of the capacitance change curve and the spatial coordinates of the gradient abrupt change point, and construct a dielectric anomaly. A three-dimensional distribution map of the normal region; a matching module 24, used to match the corresponding pre-stored standard dielectric model according to the spatial topological features of the phase angle offset in the three-dimensional distribution map, so as to locate the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region, and generate an anticoagulant distribution anomaly index by calculating the spatial topological feature deviation, and establish a mixed defect assessment matrix according to the difference between the spatial coordinate set and the phase angle offset of the standard dielectric model; an assessment module 25, used to construct a data anomaly assessment function based on the anticoagulant distribution anomaly index and the mixed defect assessment matrix, dynamically associate the mapping relationship between the assessment function output value and the clinical test error threshold, and generate test data assessment results. Figure 2 The aforementioned abnormal blood collection test data evaluation and processing system of a blood collection system can perform... Figure 1 The implementation principle and technical effects of the abnormal blood sampling test data evaluation and processing method of the blood sampling system described in the above embodiment will not be repeated here. The specific operation methods of each module and unit in the abnormal blood sampling test data evaluation and processing system of the blood sampling system in the above embodiment have been described in detail in the embodiments of the relevant method, and will not be elaborated upon here.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for evaluating and processing abnormal blood collection test data in a blood collection system, characterized in that, include: The diffusion trajectory of the anticoagulant in the blood collection tube and the blood contact surface is dynamically tracked in real time. The waveform distortion characteristics of the dielectric response on both sides of the blood contact surface are captured synchronously by a capacitive sensor array to generate diffusion trajectory morphology parameters. Analyzing the curvature abrupt change characteristics of the diffusion trajectory morphology parameters triggers the acoustic-fluid mixing device to modulate the acoustic wave frequency and fluid flow velocity, thereby forming a composite disturbance field within the mixing cavity of the acoustic-fluid mixing device; A capacitance gradient axial scan is performed on the mixed fluid after the combined disturbance field, and the phase angle offset of the capacitance change curve and the spatial coordinates of the gradient abrupt change point are extracted to construct a three-dimensional distribution map of the dielectric anomaly region. Based on the spatial topological features of the phase angle offset in the three-dimensional distribution map, a corresponding pre-stored standard dielectric model is matched to locate the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region. An anticoagulant distribution anomaly index is generated by calculating the spatial topological feature deviation. A mixed defect evaluation matrix is ​​established based on the difference between the spatial coordinate set and the phase angle offset of the standard dielectric model. Based on the anticoagulant distribution anomaly index and the mixed defect assessment matrix, a data anomaly assessment function is constructed. The mapping relationship between the output value of the assessment function and the clinical test error threshold is dynamically associated to generate test data assessment results. Based on the anticoagulant distribution anomaly index and the mixed defect assessment matrix, a data anomaly assessment function is constructed. The mapping relationship between the assessment function output value and the clinical laboratory error threshold is dynamically linked to generate laboratory data assessment results, including: Decompose the spatial density gradient of the anticoagulant distribution anomaly index and the vector magnitude of the mixed defect evaluation matrix, mark the spatial coordinate offset of the anomaly region, and generate a matrix feature set of anomaly index and defects. Construct a weight function that is the product of the spatial density of the matrix features and the vector magnitude, associate it with the segmented interval weights of the clinical test error threshold, and generate a data anomaly assessment function; The amplitude of the segmented interval fluctuation of the data anomaly assessment function is quantified, and the probability distribution boundary of the clinical test error threshold is mapped to generate a set of dynamic mapping relationships between the assessment function output and the threshold interval. Extract the abnormal boundary regions that exceed the clinical test error threshold from the dynamic mapping relationship set, extract the phase offset of the data anomaly assessment function, and generate a set of abnormal test data events; The boundary parameters of the phase offset and dynamic mapping relationship set of the abnormal event set of the test data are aggregated and arranged according to the spatial coordinate hierarchy of the anticoagulant distribution anomaly index to generate the test data evaluation result.

2. The method according to claim 1, characterized in that, Also includes: By integrating the curvature distribution of the diffusion trajectory morphology parameters, the parameters of the composite perturbation field, and the correlation of the spatial coordinate set, and combining the evaluation results of the test data, collaborative compensation parameters are generated. The imbalance in anticoagulant distribution and fluid shear defects within the mixing cavity are corrected based on the aforementioned synergistic supplementary parameters.

3. The method according to claim 1, characterized in that, Based on the spatial topological features of the phase angle offset in the three-dimensional distribution map, a corresponding pre-stored standard dielectric model is matched to locate the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixing eddy attenuation region, including: Obtain the spatial topological features of the phase angle offset in the three-dimensional distribution map, and construct a spatial distribution set of phase angle offset including spatial frequency and amplitude gradient; The reference phase angle distribution parameters of the pre-stored standard dielectric model are called, and the spatial frequency and amplitude gradient of the spatial distribution set of phase angle offset are matched layer by layer. Spatial nodes whose deviation from the reference parameters exceeds the dynamic fluctuation threshold are marked. Load the real-time energy decay trajectory of the hybrid eddy current decay process, associate the spatial nodes, filter the nodes whose energy decay rate is opposite to the magnitude gradient trend of the phase angle offset, and generate a primary spatial cluster. Within the area covered by the primary spatial cluster, the spatial frequency distribution of the phase angle offset and the fluctuation characteristics of the eddy current energy attenuation trajectory are fused to iteratively correct the overlapping boundary of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region. By integrating the dynamic fluctuation threshold of the spatial nodes, the temporal fluctuation characteristics of the eddy energy decay trajectory, and the iterative correction weight of the overlapping boundaries, the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixed eddy decay region is output.

4. The method according to claim 3, characterized in that, Within the region covered by the primary spatial cluster, the spatial frequency distribution of the phase angle offset and the fluctuation characteristics of the eddy current energy attenuation trajectory are fused to iteratively correct the overlapping boundary of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region, including: Within the boundary range of the primary spatial cluster, global interpolation of the spatial frequency distribution of the phase angle offset and the fluctuation characteristics of the eddy current energy decay trajectory is performed to generate a spatial interpolation grid covering the entire domain of the mixed fluid. By traversing each grid cell in the spatial interpolation grid, the difference between the spatial frequency amplitude change rate of the phase angle offset and the fluctuation amplitude of the eddy current energy decay trajectory is quantified to generate a dynamic fluctuation threshold mask. By superimposing the dynamic fluctuation threshold mask and the original boundary contour of the primary spatial cluster, the grid cells that synchronously exceed the limits of the phase angle offset spatial frequency amplitude change rate and eddy current energy attenuation fluctuation amplitude are extracted to construct a multimodal feature superposition field. Based on the superimposed field of multimodal features, the boundary weight parameters of the anticoagulant concentration gradient imbalance region and the mixed eddy attenuation region are iteratively adjusted to correct the overlapping boundary.

5. The method according to claim 1, characterized in that, An anticoagulant distribution anomaly index is generated by calculating the spatial topological feature deviation. A hybrid defect evaluation matrix is ​​established based on the difference in phase angle offset between the spatial coordinate set and the standard dielectric model, including: Obtain the topological feature density distribution of the spatial coordinate set and the dielectric constant distribution of the standard dielectric model, and calculate the gradient difference between the topological feature density distribution and the dielectric constant distribution to generate the spatial topological feature deviation. The spatial topological feature deviation is dynamically correlated and matched with the phase angle offset of the standard dielectric model to generate an anticoagulant distribution anomaly index. Based on the anticoagulant distribution anomaly index, the abnormal regions of the spatial coordinate set are divided, and the co-variation parameters of the anticoagulant concentration gradient and phase angle offset difference in the abnormal regions are extracted to generate the initial weight coefficients of the mixed defect evaluation matrix. The initial weight coefficients of the hybrid defect assessment matrix are dynamically optimized based on the collaborative change parameters. The dynamic optimization process adjusts the matrix dimension correlation strength by matching the cumulative threshold of the phase angle offset difference with the distribution trend of the anticoagulant concentration gradient, thereby generating a multi-dimensional hybrid defect assessment matrix that integrates spatial topological feature deviation and phase angle offset difference.

6. The method according to claim 1, characterized in that, Analyzing the curvature abrupt change characteristics of the diffusion trajectory morphology parameters triggers the acoustic-fluid mixing device to modulate the acoustic wave frequency and fluid flow velocity, forming a composite perturbation field within the mixing cavity of the acoustic-fluid mixing device, including: The local curvature abrupt change features of the diffusion trajectory morphology parameters are decomposed, and the curvature change rate of the curvature abrupt change points and the curvature gradient direction of the adjacent regions are calculated to generate a set of curvature abrupt change parameters. Based on the curvature change rate and gradient direction of the curvature mutation parameter set, the acoustic frequency band range of the acoustic fluid mixing device is matched with the linear response range of the fluid flow velocity to generate acoustic frequency modulation parameters. Based on the linear response range between the acoustic frequency modulation parameters and the fluid flow velocity, the phase synchronization sequence of the acoustic transmitter in the mixing cavity and the flow velocity gradient parameters of the fluid pump are adjusted to generate a set of coupling parameters between the acoustic wave and the fluid. Traverse the active regions of the acoustic transmitter and fluid pump covered by the set of coupling parameters, optimize the dynamic matching rules of the phase synchronization sequence and the velocity gradient parameters, and generate a set of rules to resist diffusion disturbances. The dynamic matching threshold of the anti-diffusion disturbance rule set is mapped to the phase synchronization distribution of the coupling parameter set to form a composite disturbance field that eliminates the diffusion resistance of the contact surface.

7. The method according to claim 6, characterized in that, Traversing the operational regions of the acoustic transmitter and fluid pump covered by the set of coupling parameters, the dynamic matching rules between the phase synchronization sequence and the velocity gradient parameters are optimized to generate a set of anti-diffusion disturbance rules, including: The spatial distribution characteristics of the phase synchronization sequence of the acoustic transmitter and the velocity gradient parameters of the fluid pump are analyzed, and the temporal fluctuation amplitude of the acoustic propagation constant and the fluid velocity is extracted to generate a set of dynamic response parameters. By traversing the sound wave propagation direction and fluid velocity gradient direction in the set of dynamic response parameters, the phase offset of the sound wave amplitude attenuation rate and the dynamic response of fluid velocity is calculated, and a dynamic response difference matrix is ​​generated. Based on the amplitude attenuation rate and phase offset of the dynamic response difference matrix, the conflict region between the sound wave propagation path and the fluid velocity gradient is detected, and a set of dynamic response conflict nodes is constructed. Within the area covered by the set of dynamic response conflict nodes, the phase synchronization frequency band of the acoustic transmitter and the velocity gradient threshold of the fluid pump are adjusted to generate a dynamic matching rule optimization framework. The dynamic matching rule optimization framework is dynamically matched with the phase offset parameters of the conflict node set to generate and output an anti-diffusion disturbance rule set.

8. The method according to claim 2, characterized in that, By integrating the curvature distribution of the diffusion trajectory morphology parameters, the parameters of the composite perturbation field, and the correlation of the spatial coordinate set, and combining the evaluation results of the test data, collaborative compensation parameters are generated, including: The curvature density gradient of the high-frequency abrupt region and the low-frequency smooth region in the curvature distribution of the diffusion trajectory morphology parameters is decomposed, the spatial weight ratio of different curvature intervals is quantified, and a weight mapping table of the anticoagulant diffusion path is generated. The acoustic frequency spectrum distribution and fluid velocity vector direction in the composite disturbance field parameters are extracted, and the energy proportion of the main frequency band of the acoustic wave is fused with the spatial tilt angle of the velocity vector to generate a strongly coupled acoustic-flow field tensor. Traverse the coordinate points of the anticoagulant concentration gradient imbalance region in the spatial coordinate set, and superimpose the weight mapping table and the acoustic-flow field strong coupling tensor according to the coordinate level to construct a local compensation intensity matrix for multi-field coupling. The abnormal distribution index of the anticoagulant in the evaluation results of the test data is injected to correct the dynamic attenuation coefficient of curvature density gradient and acoustic flow field intensity in the multi-field coupling compensation intensity matrix, and generate the optimized compensation parameter kernel function. Based on the phase modulation characteristics of the compensation parameter kernel function in the spatial coordinate set, the acoustic frequency offset and flow velocity vector compensation of different coordinate levels are aggregated to generate a multidimensional control sequence containing collaborative compensation parameters.

9. A system for evaluating and processing abnormal blood sampling test data in a blood sampling system, used in accordance with the method for evaluating and processing abnormal blood sampling test data in a blood sampling system as described in any one of claims 1 to 8, characterized in that, include: The tracking module is used to track the dynamic changes of the diffusion trajectory between the anticoagulant and the blood contact surface in the blood collection tube in real time. It synchronously captures the waveform distortion characteristics of the dielectric response on both sides of the blood contact surface through a capacitive sensor array to generate diffusion trajectory morphology parameters. The analysis module is used to analyze the curvature abrupt change characteristics of the diffusion trajectory morphology parameters, trigger the acoustic fluid mixing device to modulate the sound wave frequency and fluid flow velocity, and form a composite disturbance field in the mixing cavity of the acoustic fluid mixing device. The construction module is used to perform axial scanning of the capacitance gradient of the mixed fluid after the combined disturbance field, extract the phase angle offset of the capacitance change curve and the spatial coordinates of the gradient abrupt change point, and construct a three-dimensional distribution map of the dielectric anomaly region. The matching module is used to match the corresponding pre-stored standard dielectric model according to the spatial topological features of the phase angle offset in the three-dimensional distribution map, so as to locate the spatial coordinate set of the anticoagulant concentration gradient imbalance region and the mixed eddy current attenuation region, and generate the anticoagulant distribution anomaly index by calculating the spatial topological feature deviation, and establish a mixed defect evaluation matrix according to the difference between the spatial coordinate set and the phase angle offset of the standard dielectric model. The evaluation module is used to construct a data anomaly evaluation function based on the anticoagulant distribution anomaly index and the mixed defect evaluation matrix, dynamically associate the output value of the evaluation function with the mapping relationship of the clinical test error threshold, and generate test data evaluation results.

Images