An abnormal blockage real-time monitoring and early warning system based on clinical infusion data characteristics

Abstract

The application relates to the technical field of medical care information science and clinical big data processing, and discloses an abnormal blockage real-time monitoring and early warning system based on clinical infusion data characteristics, which comprises a controlled path characteristic signal capturing module, a driving load deviation compensation module, a flow channel pressure information inversion module and a blockage situation early warning discrimination module. The electric response signal generated by the controlled infusion path driving end in the commutation excitation transient state is captured, the response slope is extracted and is mapped into a virtual probe quantity representing the physical characteristics of the consumable interface, the virtual probe quantity is used to offset the nonlinear adhesion deviation component, and the pressure distribution characteristics are inverted based on a quasi-rolling dynamics mapping model. The application decouples the consumable interface resistance interference, improves the signal recognition sensitivity of the system under the condition of extremely low flow rate, and realizes high-reliability real-time early warning of the clinical infusion situation.

Description

A Real-time Monitoring and Early Warning System for Abnormal Obstruction Based on Clinical Infusion Data Characteristics Technical Field

[0001] This invention relates to a real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics, belonging to the fields of healthcare informatics and clinical big data processing technology. Background Technology

[0002] Current precision drug delivery in clinical settings typically uses infusion pumps to control the flow rate of the drug solution and determines the status of the tubing by integrating pressure sensors or monitoring the bus current of the drive motor. This provides basic protection against physical blockages. However, from an informatics perspective, the physical boundary conditions of the controlled pathway evolve dynamically over time, and the resulting nonlinear noise from the materials causes a systematic shift in the preset monitoring characteristic benchmarks.

[0003] Existing technologies primarily avoid interference by extending the judgment window or relaxing the warning threshold. Essentially, this sacrifices the timeliness of warnings for a lower false alarm rate. It is difficult to isolate system noise caused by the evolution of consumable physical characteristics from mixed load data without adding external physical sensors by constructing advanced informatics processing algorithms. For example, Chinese invention patent CN104606737A discloses an infusion pump and an infusion pump blockage detection method. This method involves setting pressure sensors attached to the pipeline upstream and downstream of the pump body, comparing the signal difference between the two with a preset threshold to eliminate the influence of changes in the elasticity of the infusion tubing during use on the detection results. However, this approach relies on the mechanical coupling between external physical sensors and the pipeline. Under extremely low flow rate conditions, linear differential logic struggles to identify non-linearity generated at the consumable interface. Linear adhesion resistance, lack of analysis on drug viscosity drift caused by environmental temperature fluctuations and stress relaxation characteristics under long-distance compression of flexible consumables, prevents the system from extracting pure fluid excitation features from underlying data in complex clinical environments. This limits the sensitivity of the early warning system in micro-infusion scenarios, leading to non-pathological false alarms or missed alarms. To address these challenges, the industry has attempted to introduce software filtering algorithms based on empirical data. However, due to the lack of characterization of the physical nature of fluid dynamics under flexible constraints, the cognitive dissonance caused by the evolution of dynamic boundary conditions remains unresolved. This technical bottleneck, caused by the instability of interface physical properties and the evolution of material properties, restricts the safety assurance dimension of clinical drug administration. There is an urgent need to establish a digital mapping model with physical property perception capabilities through healthcare informatics.

[0004] Therefore, the technical problem to be solved by this invention is how to establish an adaptive monitoring logic that can automatically resist physical drift of consumables and interface interference, and use the microscopic response characteristics of the existing electrical signal at the drive end to invert the internal pressure state of the pipeline through information acquisition and data modeling. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics, the system comprising:

[0006] The controlled pathway characteristic signal acquisition module is used to acquire electrophysical characteristic signals that characterize the driving pulse jump behavior of the controlled infusion pathway, and to start synchronous acquisition of the driving end electrical response parameters at the start of the phase excitation transient, generating a mixed load response data sequence containing interface adhesion noise.

[0007] The drive load deviation compensation module is connected to the controlled path characteristic signal capture module. It is used to extract the current change rate slope after the commutation excitation front in the mixed load response data sequence. According to the preset linear mapping rule, the current change rate slope is converted into a virtual probe quantity that characterizes the physical constraint characteristics of the interface of the controlled infusion path. The virtual probe quantity is used to perform deviation field cancellation operation on the mixed load response data sequence to filter out nonlinear noise caused by the adhesion resistance of the infusion consumable interface and generate a pure fluid excitation feature vector.

[0008] The flow channel pressure information inversion module is connected to the drive load deviation compensation module. It is used to input the pure fluid excitation feature vector into the preset pseudo-rolling dynamics mapping model. The pseudo-rolling dynamics mapping model is configured to analyze the energy conservation equation between the drive energy consumption distribution and the fluid load, and calculate the situation distribution feature mapping map reflecting the internal pressure of the controlled fluid delivery channel.

[0009] The blockage situation early warning and discrimination module is connected to the flow channel pressure information inversion module. It is used to calculate the statistical offset of the situation distribution feature map relative to the preset reference map, and output an abnormal blockage early warning signal when the statistical offset continues to exceed the preset blockage judgment threshold.

[0010] Preferably, the observation window period for synchronous acquisition performed by the controlled path characteristic signal acquisition module is limited to [a specific timeframe]. to Within the range, the controlled path characteristic signal acquisition module extracts the characteristic amplitude within the observation window period to generate a non-stationary pulsating signal characterizing the instability of the interface lubrication state, and inputs the non-stationary pulsating signal as an additional constraint variable into the blocking situation early warning and discrimination module.

[0011] Preferably, the system further includes a stress relaxation correction module, which is used to perform dynamic calibration on the elastic compensation coefficient in the simulated rolling dynamics mapping model based on the creep characteristics of the flexible consumables in the controlled infusion pathway by acquiring the characteristic residual of the non-extrusion phase period of the controlled infusion pathway within the feature space generated by the driving load deviation compensation module.

[0012] Preferably, the system also includes an environmental temperature drift sensing module, which is used to extract the winding resistance value of the drive end, analyze the physical mapping relationship between the winding resistance value and the ambient temperature to generate environmental thermal balance parameters, and perform gain compensation for the viscosity load fluctuation of the liquid medicine based on the environmental thermal balance parameters.

[0013] Preferably, when performing gain compensation, the environmental temperature drift sensing module calculates the calibrated judgment threshold according to the following logical rules. : ,in, The calibrated decision threshold; The preset benchmark threshold; This is the preset viscosity-temperature correlation coefficient of the drug solution; The environmental thermal balance parameters are generated in real time by the environmental temperature drift sensing module, in °C. This is the preset standard reference temperature value, in °C.

[0014] Preferably, the system also includes a liquid physical quality identification module, which is used to perform frequency domain analysis of the high-frequency ripple component on the mixed load response data sequence and extract the characteristic energy distribution value that characterizes the heterogeneity of the liquid.

[0015] Preferably, the blocking situation early warning and discrimination module is configured to establish a multi-dimensional judgment rule based on the characteristic energy distribution value and statistical offset, and to analyze the abnormal state into a classification evaluation result including local pulse interference, global resistance increase and medium physical property variation.

[0016] Preferably, the obstruction status early warning and judgment module is pre-set with gradient response logic, which is used to output clinical maintenance guidance signals or emergency intervention commands in stages according to the rate of change of statistical offset.

[0017] Preferably, the system also includes a sensitivity adaptive adjustment module, which is used to dynamically reconstruct the judgment weights of the blocking situation early warning and discrimination module under different clinical medication tasks based on the historical evolution characteristics of the situation distribution feature map.

[0018] Preferably, the system is used for controlled infusion pathways with flow rates below [previous value]. Under certain operating conditions, the output of the drive load deviation compensation module is corrected using virtual probes to maintain the linear sensitivity of abnormal blockage monitoring.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. In clinical infusion data features, physical decoupling of obstruction features and consumable interface interference under low flow rate conditions driven by a medical information computing model is achieved. The current response slope at the moment of commutation of the drive unit is extracted as a virtual probe using a feature signal capture algorithm. The instantaneous adhesion damping parameter of the interface between the tubing and the pump head is extracted at the information level, eliminating the nonlinear torque fluctuation caused by the stick-slip effect during low-speed compression of flexible tubing. This solves the problem of useful signals being masked by background noise of consumables in clinical micro-infusion scenarios, extends the lower limit of sensitivity of obstruction monitoring to the micro-infusion range below 5 mL / h, improves the stability of the medical and healthcare information system in sensing underlying physiological parameters, and ensures real-time early warning during high-precision drug administration.

[0021] 2. Establish a data-driven monitoring benchmark dynamic compensation mechanism that adapts to the evolution of pipeline physical properties. By acquiring the characteristic response of the drive circuit during the non-extrusion phase cycle of the pump head rotation, the system can perceive the creep and stress relaxation phenomena of the flexible pipeline caused by long-term pressure in real time through feature extraction logic. Based on this, the elastic compensation coefficient in the simulated rolling dynamic model is dynamically corrected, eliminating the zero-point drift of monitoring caused by the fatigue of consumable materials. This ensures that the monitoring system maintains a constant judgment threshold during continuous infusion for more than 24 hours, avoiding the risk of false alarms or missed alarms caused by pipeline softening.

[0022] 3. Achieving adaptive adjustment of ambient temperature drift based on software-defined sensing: This solution utilizes the physical characteristics of the resistance of the stepper motor phase windings changing with temperature. Through an algorithm model, the drive motor is reused as a distributed temperature sensing element. By analyzing the instantaneous proportional relationship between winding current and voltage and inverting the ambient thermal equilibrium temperature in the digital medical information space, the gain compensation is performed for load fluctuations caused by drug viscosity. This solves the pressure judgment error caused by direct air conditioning in the operating room or ambient temperature difference, ensuring that the system has consistent logic sensitivity under different climatic conditions and improving the operational reliability of medical equipment in complex clinical environments. Attached Figure Description

[0023] Figure 1 is a block diagram illustrating the data processing logic and algorithm flow principle of the abnormal blockage monitoring of the present invention.

[0024] Figure 2 is a schematic diagram of the overall hardware architecture of the system of the present invention and the interaction of multi-level early warning signals. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the present invention and are not intended to limit the scope of protection of the present invention.

[0026] This invention provides a real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics, including a controlled pathway characteristic signal acquisition module, a drive load deviation compensation module, a flow channel pressure information inversion module, and an obstruction status early warning discrimination module. The controlled pathway characteristic signal acquisition module is electrically connected to the drive load deviation compensation module and is used to acquire the phase current signal and electrical response parameters of the infusion pump drive end. The drive load deviation compensation module is electrically connected to the flow channel pressure information inversion module and is used to extract transient slope features from the electrical response parameters and calculate the virtual probe quantity to offset the nonlinear noise at the consumable interface. The information inversion module is electrically connected to the obstruction situation early warning and discrimination module. It is used to convert the fluid excitation feature vector into a pressure situation distribution map based on the pseudo-rolling dynamics mapping model. During clinical precision infusion, when the flow rate is below 5.0 mL / h, the physical boundary conditions of the controlled infusion pathway dynamically evolve over time. The adhesive damping generated at the contact interface between the flexible tubing and the pump head rollers can mask the true pressure load signal. To handle signal anomalies in this application scenario, the controlled pathway feature signal capture module acquires the electrophysical feature signals characterizing the driving pulse jump behavior of the controlled infusion pathway, and detects the transient at the start of the phase excitation. Initiate synchronous acquisition of the electrical response parameters of the drive end; the observation window period for the acquisition process is limited to 10. s to 50 Within a range of s, a mixed load response data sequence including interface adhesion noise is generated; the drive load deviation compensation module receives the mixed load response data sequence and extracts the slope of the current change rate after the commutation excitation leading edge. ; slope of current change rate The calculation procedure is as follows: the least squares method is used to sample the phase current sequence within the observation window. Perform linear fitting, slope of current change rate The calculation formula is as follows: ;in, The slope of the rate of change of current, in A / s; This is the change in current, expressed in amperes (A). The change in current is measured in seconds (s); the module maps the slope of the current change rate according to a preset linear mapping rule. It is converted into a virtual probe quantity that characterizes the physical constraint properties of the interface of the controlled infusion pathway.

[0027] The virtual probe measurement reflects the adhesion damping parameter at the instant of contact between the pump head roller and the outer wall of the pipeline. ; Slope of current change rate in drive load deviation compensation module With virtual probe quantity The mapping rule calibration process is as follows: Standard medical silicone tubing with a hardness of Shore A 50 to 70 degrees is installed in the controlled infusion pathway. During the pre-charge stage with zero fluid pressure, a stepper motor is driven to rotate at a gradient angular velocity range of 0.1 rad / s to 10.0 rad / s. The current change rate slope benchmark value at each angular velocity node is recorded. The least squares method is used to regress the angular velocity, current change rate slope benchmark value, and empirical constant of interfacial adhesion damping to determine the proportional mapping factor, which is then stored in a feature lookup table. The monitoring process is based on the current stepper motor angular velocity... The index corresponds to the mapping factor, which will collect the slope of the current change rate in real time. Converted into virtual probe quantity of physical constraint characteristics of controlled infusion pathway interface The module uses virtual probes to perform deviation field cancellation operations on the mixed load response data sequence, filtering out nonlinear noise caused by adhesion damping at the infusion consumable interface, and generating a pure fluid excitation feature vector. The flow channel pressure information inversion module inputs the pure fluid excitation feature vector into the pseudo-rolling kinetic mapping model, which is used to analyze the energy conservation equation between the driving energy consumption distribution and the fluid load. The pseudo-rolling kinetic mapping model calculates the situational distribution feature mapping of the internal pressure of the controlled infusion passage according to the following formula: ;in, This represents the total input drive power, expressed in watts (W). The effective power consumption to overcome fluid output resistance, measured in W; The internal mechanical losses of the system are expressed in W. This model converts the motor's phase current and output frequency into virtual rolling pressure loads. The generated situational distribution feature map reflects the spatial distribution of pressure within the flow channel and its evolution with phase. The internal mechanical losses of the pseudo-rolling dynamics mapping model are... In the analysis, the calibration weight coefficients are determined. The method for determining this is as follows: under no-load conditions, the drive unit is started and operated within its full speed range, and the total input power of the drive motor is collected through a current sensor. After deducting the heat loss from the motor windings and the amount measured by the virtual probe, Mapping mechanical friction power consumption, and relating residual power consumption to motor drive angular velocity. Correlation analysis was used to determine the calibration weighting coefficients reflecting the efficiency of the transmission mechanism and the bearing damping characteristics. Calibration weight coefficients The torque loss component at unit angular velocity, in N·m, is written into the non-volatile memory during system initialization and used as a fixed input to the model's energy conservation equation. This allows the pressure distribution feature map output by the flow channel pressure information inversion module to remove power fluctuations caused by non-fluid loads.

[0028] The preset baseline mapping map is composed of the baseline power consumption matrix collected by the system during the pre-charging phase; the processor drives the stepper motor to rotate one revolution under no-load or constant flow speed conditions, and the controlled path characteristic signal acquisition module moves in preset angular steps. collection The driving electrical response parameters at discrete phase points, and the interface adhesion damping parameters corresponding to each phase point. With driving angular velocity The baseline characteristic values ​​of the corresponding addresses are recorded to generate a spatial topology matrix reflecting the distribution of pipeline compression deformation. This topology matrix is ​​stored in a local buffer as a preset benchmark mapping map, which is used in subsequent monitoring stages to perform point-by-point residual comparison with the real-time situation distribution mapping map and calculate the statistical offset. The calculation formula is as follows: ;in, The offset is calculated in kPa. This represents the total number of discrete phase samples in a single cycle. For the first Real-time pressure inversion values ​​at each phase point, in kPa; The system is designed to detect the background pressure value at the corresponding phase point in the preset reference mapping diagram, in kPa. To address the stress relaxation and creep phenomena caused by continuous extrusion of flexible consumables, the system includes a stress relaxation correction module. During the intervals when the drive unit performs non-extrusion actions, i.e., within a specific phase angle when the pump head roller disengages from the pipeline, a standardized micro-perturbation command is injected into the motor, and the characteristic residual at the drive end is collected. A hysteresis component is used to characterize the pipeline's rebound performance. The module dynamically updates the elastic compensation coefficient in the simulated rolling dynamics mapping model based on the hysteresis component to compensate for the monitoring zero-point drift caused by continuous pressure on the pipeline.

[0029] To address the viscosity changes in the medication caused by fluctuations in ambient temperature, the ambient temperature drift sensing module acquires the real-time resistance characteristics of the infusion pump drive unit windings. This is defined as a temperature characterization parameter; based on the physical characteristics of the copper wire resistance changing with temperature, the module deduces the environmental thermal balance parameters. The calculation formula is as follows: ;in, This is the real-time resistance value, in units of... ; The resistance value at the standard reference temperature, in units of ; Temperature coefficient of resistance, unit: ; These are environmental heat balance parameters, expressed in °C. This is the standard reference temperature value, in °C; the module is based on the environmental thermal balance parameters. For the judgment threshold Perform gain compensation; the compensation logic is as follows: ;in, The calibrated threshold value is expressed in kPa. The preset benchmark threshold is expressed in kPa. This is the viscosity-temperature correlation coefficient of the drug solution, in units of... The liquid physical quality identification module performs frequency domain analysis of the high-frequency ripple component on the mixed load response data sequence. Specifically, the system is configured with a current sensor sampling frequency of 100 kHz, continuously capturing 512 sampling points and storing them in a circular buffer. A rectangular window function is applied to the buffer data for fast Fourier transform, with a 50% data overlap rate, meaning that the operation is performed once every 256 points. The frequency domain analysis focuses on extracting the high-frequency energy distribution values ​​between 10 kHz and 20 kHz. When the sum of the amplitudes in this frequency band exceeds 1.3 times the reference value of the pre-charging stage, it is logically determined that micro-crystallization has occurred in the liquid. The high-frequency ripple component in the driving current is extracted through fast Fourier transform, and the harmonic distortion parameters of the response signal are calculated. When the harmonic distortion parameters deviate from the homogeneous fluid reference value, the blockage situation warning and discrimination module outputs an identification signal characterizing micro-crystallization or heterogeneous aggregation of the liquid. The discrimination module calculates the statistical offset of the situation distribution feature map relative to the preset reference map, and outputs an abnormal blockage warning signal when the statistical offset continuously exceeds the judgment threshold.

[0030] Example 1: In a micro-infusion pump application scenario in a pediatric intensive care unit, the system performs a drug infusion task at a flow rate of 1.0 mL / h. The controlled infusion pathway uses medical-grade silicone tubing. Due to the long time interval between each step of the stepper motor inside the drive unit, a stick-slip effect occurs at the interface between the pump head roller and the tubing at a low shear rate. This causes nonlinear torque fluctuations in the current signal detected by the drive end. These fluctuations overlap with the pressure load signal caused by early blockage of the flow channel. The current threshold-based method cannot distinguish between the background noise of the consumables and the pressure load caused by the blockage. To address the monitoring challenges posed by the instability of the interface physical properties, the system initiates a processing flow based on feature signal capture and deviation compensation. The controlled pathway feature signal capture module detects the transient start of the phase excitation at the drive end. Enable synchronous sampling to capture the 30° front of the commutation excitation. The electrophysical characteristic signal within s; the current change rate slope is extracted from the mixed load response data sequence by the drive load deviation compensation module. ; slope of current change rate The calculation formula is as follows: ;in, The slope of the rate of change of current, in A / s; This is the change in current, expressed in amperes (A). The change in current is measured in seconds (s); the module maps the slope of the rate of change of current according to a linear mapping rule. This is converted into a virtual probe quantity characterizing the physical constraint properties of the interface of the controlled infusion pathway, namely the interfacial adhesion damping parameter. The extraction of virtual probe quantities provides dynamic reference field correction parameters for flow channel pressure inversion.

[0031] By using interfacial adhesion damping parameters By introducing a pseudo-rolling kinetic mapping model, the system achieves the stripping of the internal resistance of the consumable interface; the pressure situation distribution map output by the flow channel pressure information inversion module reflects the actual fluid load inside the flow channel; when the system faces a blockage caused by drug crystallization, the resistance increment generated by the fluid domain is converted into a statistical offset in the pressure situation distribution map through energy conservation analysis of the pseudo-rolling kinetic mapping model; since the system has filtered out the interface adhesion damping component, the blockage situation early warning and discrimination module identifies the evolution of the pressure trend within 200s; before the flow channel pressure reaches the preset safety limit, the system outputs an abnormal blockage early warning signal; through the information processing closed loop from the electrical response characteristics of the drive end to the physical properties inversion, the system eliminates the contradiction between monitoring sensitivity and false alarm rate in clinical micro-infusion while maintaining the mechanical structure of the infusion pump unchanged.

[0032] Example 2: The experimental verification stage employed a physical simulation platform, including a precision infusion pump power unit, a medical-grade silicone infusion tubing with preset damping characteristics, and an electrical signal recorder with a sampling frequency of 500 kSps. Experimental data was derived from the phase current sequence of the motor drive end collected by this physical experimental platform. An adjustable fluid resistance valve was installed at the end of the infusion tubing to simulate obstruction conditions. To verify the system's stability in complex clinical electromagnetic environments, Gaussian white noise with a signal-to-noise ratio of 20 dB was actively injected into the signal source. The experimental design included considerations for the sampling period. Execution logic calibration; sampling period The setting is constrained by the steepness of the rising edge of the drive pulse and the computational load of the controller; sampling period The selection of the sampling period is used to balance the fidelity of transient feature extraction with the data processing load; when the commutation excitation period is shortened, the sampling period is adjusted to avoid signal aliasing. It should tend towards 10 The lower limit of the value of s; under the typical operating condition of the stepper motor running in 16 microstepping mode, the sampling period will be... It was determined to be 20. At this point, 50 effective sampling points are acquired within the single commutation observation window, meeting the statistical accuracy requirements of least squares fitting; the experiment is conducted within a flow rate range of 0.1 mL / h to 10.0 mL / h; the controlled pathway characteristic signal acquisition module acquires the mixed load response data sequence in real time; the drive load deviation compensation module calculates the slope of the current change rate within the observation window. And convert it into virtual probe quantity The control group used a current threshold determination method without deviation compensation, while the sample group of this invention used an interface adhesion decoupling treatment method.

[0033] Table 1: Comparison of data between the present invention sample group and the control group under different flow velocities and blockage conditions.

[0034]

[0035] Based on the experimental data in Table 1, when the flow rate is below 0.5 mL / h, the control group identifies nonlinear torque fluctuations as blockage signals and generates false alarms. This is mitigated by introducing a virtual probe. After performing deviation field cancellation calculations, the deviation rate between the inverted pressure value and the actual fluid load inside the pipeline is less than 5%; when the flow rate exceeds 5.0 mL / h, the virtual probe quantity... The value of decreases and tends to level off as the shear rate increases. This phenomenon indicates that the adhesion damping at the consumable interface enters the physical saturation region at high speeds, and further increases in flow velocity cannot obtain compensation gain through the slope characteristic. During the evolution of blockage, the flow channel pressure information inversion module uses the energy conservation equation... Computational fluid effective power consumption ;in, The total input drive power, To overcome the effective power consumption of fluid output resistance, This refers to internal mechanical losses within the system; as the flow channel resistance increases, the total driving electric power... After deducting the virtual probe quantity Mapped mechanical losses After that, the effective power consumption of the fluid. It exhibits a monotonically increasing trend; even under conditions of superimposed power frequency interference and random noise, the system maintains its ability to identify pressure fluctuations above 15 kPa.

[0036] Example 3: This example, in conjunction with Figures 1 and 2, describes a real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics. As shown in Figure 1, the electrical response signal from the infusion pump drive end is input to the controlled pathway characteristic signal acquisition module. This module is configured to synchronously acquire transient response parameters and generate a mixed load response data sequence, which is then transmitted to the drive load deviation compensation module. This module extracts the slope of the current change rate and maps it to a virtual probe quantity. Then, it generates a pure fluid excitation feature vector by canceling nonlinear adhesion noise and inputs it into the flow channel pressure information inversion module. The flow channel pressure information inversion module analyzes the energy conservation equation based on the pseudo-rolling kinetic mapping model and generates a situation distribution feature mapping map. During this process, the stress relaxation correction module acquires the non-extrusion phase feature residual and dynamically calibrates the elastic compensation coefficient to correct the pseudo-rolling kinetic mapping model. Finally, the situation distribution feature mapping map enters the obstruction situation early warning judgment module. Combined with the environmental temperature drift sensing module, it calculates and statistically analyzes the offset by extracting the environmental thermal balance parameters generated by the winding resistance and performing judgment threshold gain compensation, and compares it with the preset benchmark mapping map to perform judgment threshold logic analysis. When the conditions are met, an abnormal obstruction early warning signal is output.

[0037] As shown in Figure 2, the embedded monitoring system of the intelligent infusion pump terminal device mainly consists of an infusion pump motor drive unit, a high-precision current sensing module, an ambient temperature monitoring probe, and an embedded computing and control unit. The infusion pump motor drive unit operates under the physical load resistance generated by the physical infusion pipeline system containing medication and consumables. Its drive current loop collects the drive response through the high-precision current sensing module and converts it into analog and digital signals, which are then input to the embedded computing and control unit. The ambient temperature monitoring probe is used to collect environmental parameter data to assist in temperature drift compensation. The embedded computing and control unit integrates an abnormal blockage monitoring and early warning firmware that includes a feature signal capture algorithm, a dynamic model inversion algorithm, and a blockage status discrimination logic. On the one hand, this unit outputs real-time status data streams to the human-machine interaction display screen to display the status and waveforms. On the other hand, it sends alarm trigger commands to the local audible and visual alarm. Simultaneously, it encapsulates the early warning information packet and transmits it remotely via the hospital's internal LAN / Wi-Fi through wireless and wired network communication modules to achieve synchronous early warning. Finally, it transmits the data to the centralized monitoring center of the nurse station, which is equipped with central monitoring and management software.

[0038] Example 4: In clinical continuous precision drug delivery applications, when replacing medical flexible tubing consumables with different physical specifications, due to microscopic differences in consumable wall thickness and polymer molecular weight distribution, a characteristic offset occurs between the system's original drive load compensation benchmark and the physical pathway. To address the technical challenge posed by the evolution of consumable properties to monitoring accuracy, the system initiates on-site calibration during the initial operating cycle after consumable loading. At the interval when the pump head rotates into the non-extrusion phase cycle, i.e., under the initial physical condition of the pump head roller being detached from the outer wall of the tubing, the controlled pathway characteristic signal capture module injects a set of micro-current disturbance sequences with standardized amplitudes into the drive winding. The drive load deviation compensation module extracts the current slope benchmark value after commutation transient by obtaining the electrical response feedback of the drive circuit under no-load conditions. The module uses the least squares method to determine the current slope reference value. Perform statistical regression calculations to dynamically update the analytical parameters in the simulated rolling kinetics mapping model used to characterize the internal mechanical losses of the system; internal mechanical losses of the system The calculation formula is as follows: ;in, This refers to the internal mechanical losses of the system, expressed in W. This represents the angular velocity of the stepper motor drive, measured in rad / s. These are the calibration weighting coefficients determined through on-site disturbance feedback, in N·m. The virtual probe quantity is used to characterize the physical constraint properties of the current interface and is a dimensionless parameter.

[0039] Based on this, the blockage situation early warning and discrimination module performs online correction of the blockage judgment threshold based on the viscosity-temperature characteristics of the drug solution according to the environmental thermal balance parameters obtained by the environmental temperature drift sensing module; when the drug solution inside the pipeline is switched from room temperature saline to a high-viscosity fat emulsion preparation, the module calculates the gain compensation amount of the judgment threshold according to the physical law of the exponential migration of drug solution viscosity with temperature, and adds it to the benchmark judgment threshold; the early warning and discrimination module calculates the pressure situation distribution mapping map in continuous statistical steps. The statistical offset within each step period, where the number of statistical steps is... The value is determined by the product of the flow rate setpoint and the sampling frequency to ensure consistent time resolution at different flow rates. When the statistical offset continuously exceeds the judgment threshold after online correction, the system triggers an early warning mechanism and outputs a quantified blockage risk indicator. Through dynamic calibration procedures for batch differences and adaptive correction logic for media characteristics, the system eliminates the risk of misjudgment caused by fluctuations in consumable consistency and changes in the physical properties of the drug solution, and achieves reliable inversion of the blockage situation under complex clinical variable environments.

[0040] Example 5: In an offline laboratory testing environment, the system constructs a blocking determination mapping matrix by acquiring the drive motor response sequence. This testing environment is equipped with a current sensor with a sampling rate of 1.0MHz and a pressure reference instrument with a measurement accuracy of 0.1%. The system records the phase current response sequence within the speed step range of the drive motor and calculates the corresponding drive power consumption distribution. The energy transfer coefficient in the simulated rolling dynamics mapping model is determined through a fitting algorithm, and the drive angular velocity is generated. Virtual probe quantity With internal mechanical losses of the system A three-parameter mapping table is used; this table is stored in a lookup table and covers discrete phase points within the clinical dosing flow rate range to provide deterministic output of energy loss predictions at different flow rate levels.

[0041] In the initial physical phase of system deployment at the infusion pump power terminal, the system addresses resistance offsets caused by hardware assembly through on-site baseline calibration. With tubing assembled but before medication filling, the system drives a stepper motor to perform standardized phase-jump actions, and a controlled path characteristic signal acquisition module collects the static background current generated by the pump head mechanical assembly. The system calculates the zero-point compensation factor according to the following calibration formula. : ;in, This is the zero-point compensation factor, in kPa. The current-to-pressure conversion factor is determined by the mapping table of the aforementioned three parameters, and the unit is kPa / A; The collected static background current value is in amperes (A); this is the zero-point compensation factor. The data is written into non-volatile memory to perform zero-point correction on the pressure distribution map output by the flow channel pressure information inversion module, in order to suppress measurement offsets caused by bearing wear or pump head pressure fluctuations. Before performing zero-point compensation, the system calibrates the current-to-pressure conversion factor: at standard ambient temperature, the stepper motor is driven to push physiological saline at a flow rate of 10 ml per hour, and the average steady-state phase current is recorded as 100 mA, with the corresponding internal reference pressure defined as 5 kPa. Thus, the conversion factor is determined to be 0.08 kPa pressure increase for every 1 mA increase in current. This calibration process is performed when the device is first powered on or when the pump head assembly is replaced, to ensure that the calculated kPa value has a clear physical reference.

[0042] Example 6: In an offline test scenario for calibrating the internal mechanical loss characteristics of the system, the system executes a standardized characterization procedure for the energy conversion relationship; it starts up in a controlled environment with a temperature of 25°C and a relative humidity of 50%, and the system drives the stepper motor to perform gradient scanning within an angular velocity range of 0.1 rad / s to 10.0 rad / s. The controlled path characteristic signal acquisition module synchronously records the commutation transient current slope at each speed node. The system determines the calibration weight coefficients in the proposed rolling kinetics mapping model by performing linear regression analysis on the acquired sampling sequences. Calibration weight coefficients The formula for determining it is as follows: ;in, The calibration weighting coefficients are in N·m. This represents the total input drive power, expressed in watts (W). The effective power consumption to overcome fluid output resistance, measured in W; This represents the angular velocity of the stepper motor drive, measured in rad / s. The virtual probe quantity is a dimensionless parameter. This procedure establishes a predictive model for internal system losses by decomposing the energy flow direction under different power output modes.

[0043] In clinical applications involving changing medication solutions of different viscosities, the system utilizes real-time sensed environmental characteristic parameters to perform online correction of judgment thresholds. Once the infusion pathway is fully loaded and the system is in the pre-charge operation phase, the environmental temperature drift sensing module acquires the real-time resistance characteristics of the stepper motor's phase windings. And calculate the environmental heat balance parameters. The congestion situation early warning and judgment module determines the congestion situation based on environmental thermal balance parameters. For the judgment threshold Perform compensation calculations and calibrated decision thresholds. The calculation process is as follows: ;in, The calibrated threshold value is expressed in kPa. The threshold value is used as the benchmark for judgment, and the unit is kPa. This is the viscosity-temperature correlation coefficient of the drug solution, in units of... ; The current environmental thermal balance parameter, in units of ; The standard reference temperature value is in °C. This procedure suppresses the shift in monitoring sensitivity caused by changes in ambient temperature, enabling the system to maintain consistent early warning accuracy when facing media with different physical properties.

[0044] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics, characterized in that, The system includes: The controlled pathway characteristic signal acquisition module is used to acquire electrophysical characteristic signals characterizing the driving pulse jump behavior of the controlled infusion pathway, and to initiate synchronous acquisition of the driving end electrical response parameters at the start of the phase excitation transient, generating a mixed load response data sequence containing interface adhesion noise. The driving load deviation compensation module, connected to the controlled pathway characteristic signal acquisition module, is used to extract the current change rate slope after the commutation excitation front in the mixed load response data sequence, convert the current change rate slope into a virtual probe quantity characterizing the physical constraint characteristics of the interface of the controlled infusion pathway according to a preset linear mapping rule, and use the virtual probe quantity to perform deviation field cancellation operation on the mixed load response data sequence to filter out nonlinear noise caused by the interface adhesion resistance of the infusion consumables, generating a pure fluid excitation characteristic vector. The flow channel pressure information inversion module, connected to the driving load deviation compensation module, is used to input the pure fluid excitation characteristic vector into a preset pseudo-rolling kinetic mapping model. The pseudo-rolling kinetic mapping model is configured to analyze the energy conservation equation between the driving energy consumption distribution and the fluid load, and calculate the situation distribution characteristic mapping map reflecting the internal pressure of the controlled infusion pathway. The blockage situation early warning and discrimination module is connected to the flow channel pressure information inversion module. It is used to calculate the statistical offset of the situation distribution feature map relative to the preset reference map, and output an abnormal blockage early warning signal when the statistical offset continues to exceed the preset blockage judgment threshold.

2. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 1, characterized in that, The observation window period for synchronous acquisition performed by the controlled path characteristic signal acquisition module is limited to [a certain period]. to Within the range, the controlled path characteristic signal acquisition module extracts the characteristic amplitude within the observation window period to generate a non-stationary pulsating signal characterizing the instability of the interface lubrication state, and inputs the non-stationary pulsating signal as an additional constraint variable into the blocking situation early warning and discrimination module.

3. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 1, characterized in that, The system also includes a stress relaxation correction module, which is used to perform dynamic calibration on the elastic compensation coefficients in the simulated rolling dynamics mapping model based on the creep characteristics of the flexible consumables in the controlled fluid delivery path by acquiring the characteristic residuals of the non-extrusion phase period of the controlled fluid delivery path, using the pure fluid excitation feature vector generated by the drive load deviation compensation module.

4. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 1, characterized in that, The system also includes an environmental temperature drift sensing module, which is used to extract the winding resistance value of the drive end, analyze the physical mapping relationship between the winding resistance value and the ambient temperature to generate environmental thermal balance parameters, and perform gain compensation for the viscosity load fluctuation of the liquid drug based on the environmental thermal balance parameters.

5. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 4, characterized in that, When performing gain compensation, the environmental temperature drift sensing module calculates the calibrated judgment threshold according to the following logical rules. ： ,in, The calibrated threshold value is expressed in kPa. The preset blocking threshold is expressed in kPa. This is the preset viscosity-temperature correlation coefficient of the drug solution, in units of... ； The environmental thermal balance parameters are generated in real time by the environmental temperature drift sensing module, in °C. This is the preset standard reference temperature value, in °C.

6. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 1, characterized in that, The system also includes a liquid medicine physical quality identification module, which is used to perform frequency domain analysis of the high-frequency ripple component on the mixed load response data sequence and extract the characteristic energy distribution value that characterizes the heterogeneity of the liquid medicine.

7. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 6, characterized in that, The blocking situation early warning and discrimination module is configured to establish multi-dimensional judgment rules based on characteristic energy distribution values ​​and statistical offsets, and to analyze abnormal states into classification and evaluation results including local pulse interference, global resistance increase and changes in medium physical properties.

8. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 1, characterized in that, The obstruction situation early warning and judgment module is pre-set with gradient response logic, which is used to output clinical maintenance guidance signals or emergency intervention commands in stages according to the rate of change of statistical offset.

9. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 1, characterized in that, The system also includes a sensitivity adaptive adjustment module, which dynamically reconstructs the judgment weights of the blocking situation early warning and discrimination module under different clinical medication tasks based on the historical evolution characteristics of the situation distribution feature map.

10. The real-time monitoring and early warning system for abnormal obstruction based on clinical infusion data characteristics according to claim 1, characterized in that, The system is used for controlled infusion pathways with flow rates below [missing information]. Under the operating conditions, the output of the drive load deviation compensation module is corrected by using virtual probe measurements.

Images