Intelligent drainage tube central monitoring method, system and drainage monitoring sensor

Abstract

The application discloses a kind of intelligent drainage tube central monitoring method, system and drainage monitoring sensor, it is related to medical monitoring field, including: through the monitoring sensor of external in the wall of drainage tube, the multimodal sensing data of drainage fluid is collected;Respectively calculate drainage flow rate real-time value and drainage cumulative total amount, carry out hierarchical determination to drainage fluid blood component;According to the preset three-level monitoring rule, execute local alarm determination and pinch control decision;According to the real-time monitoring data of patient, periodically execute risk assessment, automatically adjust the monitoring level group of patient;Data is uploaded to central monitoring platform, the monitoring data of all access sensor is centralized visual display, abnormal alarm and data storage management.The application has the advantages that: through external non-contact multimodal sensor and dynamic risk grading strategy, according to the real-time state of patient, the monitoring density is self-adaptively adjusted, realizes accurate, efficient and gives consideration to medical resource optimization drainage whole process intelligent management.

Description

Technical Field

[0001] This invention relates to the field of medical monitoring, specifically to a method, system, and drainage monitoring sensor for central monitoring of intelligent drainage tubes. Background Technology

[0002] Postoperative drainage tubes are commonly required for clinical surgical procedures and ICU patients to drain oozing fluids such as blood, effusion, and pus. Medical staff need to regularly check the drainage bags at the bedside, record the total drainage volume, observe the fluid color and flow rate, and assess for complications such as active bleeding or abnormal effusion. Standard clinical drainage devices typically consist only of a drainage tube, drainage bag, and graduated scale, lacking an automated monitoring module. Existing simple drainage monitoring accessories are mostly independent timing and measuring devices, usable only by a single bed, without a centralized management platform.

[0003] Currently, there are two types of drainage monitoring equipment on the market. The first is a single-bed independent drainage metering sensor, which can only be placed on a single patient's drainage tube and can only collect drainage fluid volume data. It lacks a centralized monitoring terminal, requiring medical staff to check the device bed-by-bed, making unified monitoring across the entire department impossible. This device relies solely on gravity measurement to calculate flow rate and cannot identify blood components in the drainage fluid; nurses can only visually judge the fluid color, resulting in a lag in identification. The second type is a ward-centralized vital signs monitor, which only monitors basic vital signs such as blood pressure, heart rate, and blood oxygenation. It is incompatible with drainage tube monitoring modules and cannot collect information on drainage speed, total drainage volume, or drainage fluid characteristics, nor does it have an alarm function for abnormal drainage.

[0004] Currently, clinical drainage devices and general monitoring equipment lack integrated capabilities for measuring drainage fluid flow rate, identifying blood components, and centrally managing them. This forces medical staff to rely on manual inspections to assess drainage status, which is not only inefficient but also makes it difficult to promptly detect abnormal changes such as occult bleeding. Furthermore, the lack of a tiered alarm mechanism based on multi-dimensional monitoring data and automatic tube clamping mechanisms makes it easy for the handling of exceeding drainage limits to be delayed or forgotten. Drainage data also cannot be centrally stored or dynamically analyzed, resulting in deficiencies in the comprehensiveness of perception, the timeliness of alarms, and the closed-loop management of central monitoring of drainage tubes. Summary of the Invention

[0005] To address the aforementioned technical problems, this paper provides a central monitoring method, system, and drainage monitoring sensor for intelligent drainage tubes. This technical solution solves the problems mentioned in the background art, such as the lack of integrated capabilities for drainage fluid flow rate measurement, blood component identification, and centralized control, the reliance on manual inspection to determine the drainage status, low efficiency, difficulty in timely detection of abnormal changes such as occult bleeding, lack of a hierarchical alarm mechanism based on multi-dimensional monitoring data and automatic tube clamping execution means, and the easy delay or forgetting of the handling after the total drainage volume exceeds the limit.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for central monitoring of intelligent drainage tubes, comprising: Multimodal sensing data of the drainage fluid is collected by a monitoring sensor placed on the wall of the drainage tube. The multimodal sensing data includes at least fluid flow rate data and liquid spectral data. Based on multimodal sensor data, the real-time value of drainage flow rate and the cumulative total drainage volume are calculated respectively, and the blood components of the drainage fluid are classified and determined, and blood grade labels are output. Based on the real-time value of drainage flow rate, the cumulative drainage volume, and the blood grade label, local alarm judgment and tube clamping control decisions are executed in accordance with the preset three-level monitoring rules; Based on the postoperative monitoring status of patients, patients are divided into different monitoring level groups. Different sensor data acquisition frequencies and data upload strategies are set for different monitoring level groups. Risk assessments are periodically performed based on the real-time monitoring data of patients, and the monitoring level group of patients is automatically adjusted based on the risk assessment results. Multimodal sensor data, real-time drainage flow rate, cumulative drainage volume, and blood grade labels are uploaded to the central monitoring platform via wireless communication. The central monitoring platform then provides centralized visualization, alarm functions, and data storage management for all monitoring data from the connected sensors.

[0007] Preferably, the multimodal sensing data of the drainage fluid collected by the monitoring sensor externally placed on the wall of the drainage tube includes: An infrared non-contact flow detection unit is used to collect liquid flow rate signals by penetrating the wall of the drainage tube with infrared light. After signal amplification and analog-to-digital conversion, the raw flow rate data is output. A multi-band optical detection unit is used, with multi-band LED light sources and photodetectors set on both sides of the drainage tube to collect the intensity of transmitted light signals in the green, red, and near-infrared bands, respectively, as raw spectral data.

[0008] Preferably, the step of calculating the real-time value of drainage flow rate and the cumulative drainage volume based on multimodal sensor data, and classifying the blood components of the drainage fluid to output blood grade labels includes: The raw flow velocity data is subjected to sliding window filtering to remove abnormal pulse signals, resulting in a filtered flow velocity signal. Combined with the preset inner diameter parameters of the drainage tube, the real-time value of the drainage flow velocity is calculated. The cumulative total of drainage is calculated by integrating the real-time value of the drainage flow rate over time. The raw spectral data is preprocessed to output the preprocessed transmitted light signal intensity of each band. The preprocessing includes ambient light subtraction, dynamic baseline calibration, bubble identification, and abnormal signal removal. Based on the normalized absorbance values ​​of each band, a multi-band spectral feature vector is calculated. The multi-band spectral feature vector includes the normalized absorbance of each band, the band absorbance ratio, and the signal change rate. Multiple standard drainage fluid samples with known blood concentrations were collected. Under standard tubing conditions, multi-band optical detection units were used to acquire multi-band spectral data of each sample, and multi-band spectral feature vectors of each sample were extracted. Using multi-band spectral feature vectors as input and known blood concentration grading labels of standard drainage fluid samples as output targets, a classification algorithm is used for training to obtain a drainage fluid blood grading model; The multi-band spectral feature vector is input into the drainage fluid blood grading model, and the blood grade label is output. The blood grade label includes at least low blood component grade, light blood fluid grade, and high blood component grade.

[0009] Preferably, the method for constructing the multi-band spectral feature vector further includes: Using the reference spectral data collected by the standard blank solution under the same pipeline conditions as the baseline, the normalized absorbance values ​​for each band are calculated based on the preprocessed transmitted light signal intensity and the reference spectral data:

[0010] In the formula, For band The normalized absorbance value, For band The intensity of the transmitted light signal after preprocessing For band The intensity of the reference optical signal; The band absorbance ratio is the ratio of the normalized absorbance values ​​of any two bands.

[0011] In the formula, For the first Band and the The ratio of absorbance in different wavelength bands, and The first Bands and the Normalized absorbance value for the band. This is a preset regularization constant used to prevent the denominator from being zero; Under a fixed sampling interval, the absolute value of the difference between the normalized absorbance value at the current sampling time and the previous sampling time is calculated, and the ratio of the absolute value of the difference to the sampling interval is defined as the signal change rate. The normalized absorbance values, absorbance ratios of each band, and signal change rate are combined into a multi-dimensional vector, which serves as the multi-band spectral feature vector.

[0012] Preferably, the step of executing local alarm judgment and tube clamping control decisions based on the real-time value of drainage flow rate, the cumulative drainage volume, and the blood grade label, according to the preset three-level monitoring rules, includes: Preset flow rate alarm threshold and total flow limit threshold; When the real-time value of the drainage flow rate exceeds the flow rate alarm threshold, an abnormal flow rate alarm is triggered. When the total amount of drainage reaches the upper limit threshold, a total amount compliance alarm is triggered, and a clamping control command is generated at the same time. When the blood grade label reaches the preset blood alarm level, a blood abnormality alarm is triggered; When any of the abnormal flow rate alarm, total volume target alarm, and abnormal blood flow alarm is triggered, an audible and visual alarm signal is issued, and the corresponding tube clamping control action is executed according to the alarm priority. When the cumulative drainage volume reaches a preset percentage of the total capacity of the drainage bag, or when the estimated time for the drainage bag to fill is less than a preset time threshold, an "Immediate Clamping Reminder" label is attached to the patient. For patients with the "About to Clamp Tube" tag, multiple total volume thresholds are set. When the total drainage volume reaches each threshold, a progressive reminder of the corresponding level is pushed. When the total drainage volume reaches the highest threshold, an audible and visual alarm and a servo motor clamp tube command are automatically triggered.

[0013] Preferably, the step of periodically performing risk assessments based on the patient's real-time monitoring data and automatically adjusting the patient's monitoring level group based on the risk assessment results includes: The monitoring level groups should include at least: a high-risk observation group and a routine group; Patients are automatically placed into the high-risk observation group after surgery and are continuously monitored for a preset initial duration. After the initial monitoring period ends, a risk assessment is performed every preset assessment cycle. Based on multi-band spectral blood data, drainage fluid flow rate and flow rate fluctuation coefficient, a comprehensive risk assessment model is constructed to conduct a comprehensive assessment and output a comprehensive risk score. Based on the comparison between the comprehensive risk score and the preset risk threshold, patients are assigned to either the high-risk observation group or the routine group. The risk assessment also includes an immediate triggering mechanism, including: When the flow rate fluctuation coefficient of a patient in the standard group exceeds the fluctuation threshold in a series of preset monitoring sessions, an immediate risk assessment is triggered. When any patient's blood grade label reaches or exceeds the preset instant assessment trigger level, an instant risk assessment is triggered. Once the immediate risk assessment is completed, if the overall risk score exceeds the preset escalation threshold, the patient will be immediately escalated to the high-risk observation group.

[0014] Preferably, the centralized visualization, anomaly alarm, and data storage management of monitoring data from all connected sensors by the central monitoring platform includes: The monitoring data from the risk observation team is uploaded to the central monitoring platform in real time at the highest upload frequency. The monitoring data of the regular group is uploaded to the central monitoring platform at a second upload frequency, which is lower than the first upload frequency. The data uploaded by the regular group is an aggregated data packet within a preset time window. The aggregated data packet includes at least the mean, maximum and minimum values ​​of the monitoring data within the time window. The central monitoring platform receives real-time monitoring data uploaded by each monitoring sensor through a wireless communication gateway, and displays the patient drainage status information corresponding to each sensor on a visual dashboard by bed. When monitoring data uploaded by any sensor triggers abnormal conditions, the central monitoring platform simultaneously displays a pop-up notification and an audible and visual alarm, and highlights the abnormal bed. The central monitoring platform automatically saves all monitoring data according to the preset storage period, supports querying and exporting historical data by patient ID and time period, and automatically generates drainage flow rate change curves and cumulative drainage volume growth curves. When the total drainage volume reaches the preset clamping reminder threshold, a clamping reminder window will pop up on the visual dashboard to prompt medical staff to clamp the drainage tube. The clamping reminder threshold is less than or equal to the upper limit threshold of the total drainage volume.

[0015] Furthermore, this solution proposes an intelligent central monitoring system for drainage tubes to implement the intelligent central monitoring method for drainage tubes as described above, including: An external drainage monitoring sensor is used to install on the outer wall of the drainage tube to collect multimodal sensing data of the drainage fluid, and to perform local calculations, alarm judgments and tube clamping control based on the multimodal sensing data. A wireless communication gateway is used to receive monitoring data uploaded by one or more external drainage monitoring sensors, and after completing data protocol conversion and data caching, transmit the data to the central monitoring station. The central monitoring station is used for centralized visualization, alarm management, parameter configuration, tube clamping reminder, data storage, and equipment status management of monitoring data from all connected sensors. It includes: a multi-patient visualization dashboard module, an alarm management module, a parameter configuration module, a tube clamping reminder module, a data storage and statistics module, and an equipment status management module.

[0016] Furthermore, a drainage monitoring sensor, adapted to a central monitoring system for intelligent drainage tubes, includes: a housing, an infrared non-contact flow detection device 5, a multi-band optical detection device, a tube clamping actuator 8, and a circuit board module 9. The outer shell includes a front shell 1 and a rear shell 2. The bottom of the front shell 1 and the rear shell 2 are connected by a hinge 4. The front shell 1 includes a front shell plate 101. The front shell plate 101 has a fixing groove plate 102 symmetrically located on both sides. A pin 3 is installed on the groove plate 102. The pin 3 passes through the groove plate 102 and is slidably connected to the groove plate 102. Rotating buckles 103 are provided on the left and right sides of the lower edge of the top of the front shell plate 101. Buckle holes 203 are provided on the left and right sides of the lower edge of the top of the rear shell 2. Pipe through holes 201 and pin grooves 202 are coaxially opened on the left and right side walls of the rear shell 2. The infrared non-contact flow detection device 5 is fixed in the left pipe perforation 201, and the main body is placed on the outside of the rear housing 2. The multi-band optical detection device includes a photodetector 6 and a background platform 7. The photodetector 6 is fixed at the top of the inner cavity of the rear housing 2, and the background platform 7 is fixed at the bottom of the inner cavity of the rear housing 2 and faces the photodetector 6. The drainage tube 10 is located between the photodetector 6 and the background platform 7. The tube clamping device 8 is fixed to the bottom of the inner cavity of the rear housing 2, on the right side of the background platform 7, directly below the drainage tube 10. The tube clamping device 8 includes a control motor 801, a main rotary transmission gear 802, a driven rotary transmission gear 804, an active elliptical tube clamping head 803, and a driven elliptical tube clamping head 805. The control motor 801 is fixed to the bottom of the inner cavity of the rear housing 2. The circuit board module 9 is fixed to the upper inner wall of the rear housing 2, located to the right of the photodetector 6 and directly above the clamping actuator 8.

[0017] Preferably, the circuit board module further includes: The flow metering unit is used to perform sliding window filtering on the raw flow velocity data output by the infrared non-contact flow detection device, calculate the real-time value of the drainage flow velocity in combination with the preset inner diameter parameters of the drainage tube, and calculate the cumulative total drainage by integrating the real-time value of the drainage flow velocity over time. The local alarm unit is used to control the buzzer and LED warning lights to emit audible and visual alarm signals when the abnormal flow rate alarm, total volume target alarm, or abnormal blood alarm is triggered, based on the judgment results of the three-level monitoring rules. The wireless transmission unit is used to upload multimodal sensor data, real-time drainage flow rate, cumulative drainage volume, and blood grade label to the wireless communication gateway via wireless communication, and to receive parameter configuration instructions issued by the central monitoring platform. The parameter storage and adjustment unit is used to store preset flow rate alarm thresholds, total drainage volume upper limit thresholds, blood alarm levels, drainage tube inner diameter parameters, and model parameters of the drainage fluid blood grading model. It also supports medical staff to adjust the stored parameters on-site or remotely via physical buttons or wireless communication.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: Integrated sensors externally mounted on the drainage tube wall synchronously collect flow velocity and multi-band spectral data. At the edge, local alarms and tube clamping are executed based on flow velocity integration, intelligent blood grade classification, and three-level monitoring rules. This enables real-time quantitative identification of active bleeding and drainage abnormalities, while ensuring independent closed-loop control even when the central platform is offline. It can quickly and accurately identify early postoperative bleeding tendencies, predict bleeding risks early, significantly shorten emergency response time, and buy critical time for clinical rescue. Simultaneously, through dynamic grouping of monitoring levels and variable frequency data acquisition and uploading strategies, it effectively solves the data throughput and monitoring density issues under centralized management of multiple patients. The challenge lies in balancing the need for precise monitoring of high-risk patients with the need to reduce system resource consumption. Furthermore, by leveraging the visualization dashboards and data storage of the central monitoring platform, medical staff can remotely monitor the real-time status of all drainage lines in the ward from the nurses' station or physician's office. This allows for comprehensive control of drainage abnormalities, improving the overall safety level of postoperative monitoring. Simultaneously, it reduces the frequency of bedside rounds, optimizes departmental nursing staff allocation, effectively alleviates the workload of clinical medical staff, and enables centralized monitoring, trend analysis, and historical tracing of drainage status across the entire department. This significantly enhances the intelligence level of postoperative drainage management, the efficiency of abnormal response, and the safety of clinical operations. Attached Figure Description

[0019] Figure 1 This is a flowchart of the intelligent drainage tube central monitoring method of the present invention; Figure 2 The present invention uses multimodal sensor data to calculate the real-time value of drainage flow rate and the cumulative total drainage, and classifies the blood components of the drainage fluid to output a blood grade label flowchart. Figure 3 This is a diagram of the intelligent drainage tube central monitoring system architecture of the present invention; Figure 4 This is a schematic diagram of the drainage monitoring sensor structure of the present invention; Figure 5 This is a cross-sectional schematic diagram of the pin on the front housing of the drainage monitoring sensor of the present invention; The numbers on the map are: 1-Front housing, 101-Front housing plate, 102-Groove plate, 103-Rotating buckle; 2-Rear housing, 201-Pipe perforation, 202-Pin slot, 203-Snap-in hole; 3-Pin, 301-Sliding steel ball, 302-Spring structure; 4-Flip hinge; 5-Infrared non-contact flow detection device; 6-Photodetector; 7-Background platform; 8-Pipe clamping actuator, 801-Control motor, 802-Main rotary transmission gear, 803-Active elliptical pipe clamping head, 804-Driven rotary transmission gear, 805-Driven elliptical pipe clamping head; 9-Circuit board module; 10-Drainage pipe. Detailed Implementation

[0020] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0021] Reference Figure 1 As shown, a method for central monitoring of intelligent drainage tubes includes: Multimodal sensing data of the drainage fluid is collected by a monitoring sensor placed on the wall of the drainage tube. The multimodal sensing data includes at least fluid flow rate data and liquid spectral data. Based on multimodal sensor data, the real-time value of drainage flow rate and the cumulative total drainage volume are calculated respectively, and the blood components of the drainage fluid are classified and determined, and blood grade labels are output. Based on the real-time value of drainage flow rate, the cumulative drainage volume, and the blood grade label, local alarm judgment and tube clamping control decisions are executed in accordance with the preset three-level monitoring rules; Based on the postoperative monitoring status of patients, patients are divided into different monitoring level groups. Different sensor data acquisition frequencies and data upload strategies are set for different monitoring level groups. Risk assessments are periodically performed based on the real-time monitoring data of patients, and the monitoring level group of patients is automatically adjusted based on the risk assessment results. Multimodal sensor data, real-time drainage flow rate, cumulative drainage volume, and blood grade labels are uploaded to the central monitoring platform via wireless communication. The central monitoring platform then provides centralized visualization, alarm functions, and data storage management for all monitoring data from the connected sensors.

[0022] Understandably, to achieve intelligent central monitoring of drainage tubes, it is first necessary to acquire multi-dimensional physical and optical information of the drainage fluid without contact. Therefore, this solution employs a monitoring sensor externally placed on the drainage tube wall, using both infrared non-contact flow detection and multi-band transmission spectroscopy to collect flow velocity and spectral data of the drainage fluid in parallel. This avoids the risk of cross-infection associated with invasive sensors and provides a reliable data foundation for subsequent flow velocity calculation, total volume accumulation, and blood component identification. Secondly, after acquiring multimodal sensor data, preliminary data processing and calculation need to be performed locally on the sensor to reduce the amount of data transmitted wirelessly and improve alarm response speed. Therefore, this solution calculates the real-time drainage flow velocity at the sensor end, integrates the flow velocity over time to obtain the total cumulative drainage volume, and inputs the spectral data into a pre-constructed drainage fluid blood grading model, outputting a blood grade label. Building upon this foundation, this solution establishes three levels of monitoring rules locally on the sensors. Based on real-time monitoring results across three dimensions—flow rate, total volume, and blood grade—local alarms and tube clamping controls are triggered independently or in conjunction with each other. This ensures that even if the central monitoring platform is disconnected or experiences communication delays, the sensors can still independently perform basic safety monitoring functions. Finally, all monitoring data is uploaded to the central monitoring platform via wireless communication. The central monitoring platform then aggregates data from all connected sensors, enabling centralized, visual monitoring of patients across the entire department, unified alarm management, and historical data storage and analysis.

[0023] Furthermore, considering the varying severity of patients' conditions and their differing needs for monitoring frequency and data processing methods, this solution introduces a monitoring level classification and dynamic grouping mechanism. Patients are divided into different monitoring level groups based on their real-time monitoring status, and differentiated data collection frequencies and data upload strategies are set for each group. This ensures both the precise monitoring needs of high-risk patients and reduces system data storage pressure by lowering the collection frequency for routine patients, thus achieving a rational allocation of monitoring resources.

[0024] Multimodal sensing data of the drainage fluid are collected by monitoring sensors placed externally on the wall of the drainage tube, including: An infrared non-contact flow detection unit is used to collect liquid flow rate signals by penetrating the wall of the drainage tube with infrared light. After signal amplification and analog-to-digital conversion, the raw flow rate data is output. A multi-band optical detection unit is used, with multi-band LED light sources and photodetectors set on both sides of the drainage tube to collect the intensity of transmitted light signals in the green, red, and near-infrared bands, respectively, as raw spectral data.

[0025] It can be explained that the flow rate of the drainage fluid and the blood composition are two core indicators for determining whether a patient has postoperative complications such as active bleeding or abnormal exudation. Abnormal flow rate can indicate the risk of bleeding or drainage tube blockage, while changes in blood composition are a direct basis for determining active bleeding. This solution uses an infrared non-contact flow detection unit and a multi-band optical detection unit integrated in the same sensor to simultaneously acquire these two types of data, providing raw input for subsequent flow rate calculation, total volume accumulation, and blood grading determination. The infrared non-contact flow detection unit works as follows: the infrared LED emitter emits infrared light into the drainage tube, which penetrates the tube wall and the drainage fluid inside before reaching the infrared photoelectric receiver. As the drainage fluid flows within the tube, the volume of the liquid flowing through the detection area constantly changes, causing the intensity of the transmitted light signal to fluctuate regularly with respect to the flow rate. By measuring the frequency and amplitude of these fluctuations, combined with a pre-calibrated flow rate conversion coefficient and the inner diameter of the drainage tube, the real-time flow rate can be calculated in subsequent steps. The core advantage of this detection method is that the detection unit is completely external to the drainage tube wall, without compromising the airtightness of the drainage system, and does not come into contact with the drainage fluid, thus avoiding the risk of cross-infection associated with invasive sensors. The multi-band optical detection unit employs three-band transmission detection technology—green light (525nm), red light (660nm), and near-infrared light (940nm)—to identify blood components based on Lambert-Beer's law and multi-band scattering characteristics. The physical principle is as follows: different wavelengths of light are absorbed to varying degrees by hemoglobin when penetrating drainage fluid. According to Lambert-Beer's law, absorbance is linearly related to the concentration of the absorbing substance. By measuring the attenuation of transmitted light intensity relative to incident light intensity, the hemoglobin concentration in the drainage fluid can be calculated. Among them, the green band has high absorption sensitivity for low concentrations of hemoglobin, making it suitable for detecting trace amounts of blood components in drainage fluid and a key channel for early detection of occult bleeding; the red band maintains good linear response even at high hemoglobin concentrations, making it suitable for quantifying medium to high concentrations of blood components; the near-infrared band has good penetration and stability for differences in tube wall thickness, changes in drainage fluid turbidity, and changes in the light source path, and can be used as an auxiliary correction channel to compensate for systematic errors introduced by individual differences in the tube and environmental changes, thereby improving the accuracy and robustness of blood grading determination; The integrated sensor clip is installed on the outer wall of the drainage tube, and is compatible with different specifications of drainage tubes commonly used in clinical practice, such as abdominal drainage tubes, closed thoracic drainage tubes, wound drainage tubes and ventricular drainage tubes. Its external, non-contact and universal design not only ensures the standard of clinical aseptic operation, but also enables the system to be deployed quickly without replacing the dedicated drainage tube.

[0026] Reference Figure 2As shown, based on multimodal sensor data, the real-time value of drainage flow rate and the cumulative drainage volume are calculated respectively, and the blood components of the drainage fluid are classified and determined, outputting blood grade labels including: The raw flow velocity data is subjected to sliding window filtering to remove abnormal pulse signals, resulting in a filtered flow velocity signal. Combined with the preset inner diameter parameters of the drainage tube, the real-time value of the drainage flow velocity is calculated. The cumulative total of drainage is calculated by integrating the real-time value of the drainage flow rate over time. The raw spectral data is preprocessed to output the preprocessed transmitted light signal intensity of each band. The preprocessing includes ambient light subtraction, dynamic baseline calibration, bubble identification, and abnormal signal removal. Based on the normalized absorbance values ​​of each band, a multi-band spectral feature vector is calculated. The multi-band spectral feature vector includes the normalized absorbance of each band, the band absorbance ratio, and the signal change rate. Multiple standard drainage fluid samples with known blood concentrations were collected. Under standard tubing conditions, multi-band optical detection units were used to acquire multi-band spectral data of each sample, and multi-band spectral feature vectors of each sample were extracted. Using multi-band spectral feature vectors as input and known blood concentration grading labels of standard drainage fluid samples as output targets, a classification algorithm is used for training to obtain a drainage fluid blood grading model; The multi-band spectral feature vector is input into the drainage fluid blood grading model, and the blood grade label is output. The blood grade label includes at least low blood component grade, light blood fluid grade, and high blood component grade.

[0027] It can be explained that postoperative drainage tube placement is a routine clinical procedure, and the color, flow rate, and total volume of the drainage fluid are core indicators for determining whether a patient has active bleeding, abnormal exudation, or other complications. Traditional methods rely entirely on nurses' regular bedside checks, which may lead to the failure to detect minor or occult bleeding in the early stages, delaying timely intervention. This solution uses a monitoring sensor externally placed on the drainage tube wall to continuously collect flow rate and spectral data of the drainage fluid. The control module automatically calculates the flow rate, accumulates the total volume, and extracts spectral features locally. Combined with a pre-built drainage fluid blood grading model, it outputs a blood grade label, transforming drainage monitoring from manual visual judgment to automatic quantitative judgment by sensors. Based on the blood grade label, it can quickly and accurately identify the patient's early postoperative bleeding tendency, predict bleeding risks as early as possible, significantly shorten emergency response time, and gain critical time for clinical rescue. To further illustrate this plan, the following details the implementation of obtaining real-time drainage flow rate, cumulative drainage volume, and blood grade labels: After infrared light penetrates the wall of the drainage tube, the intensity of the transmitted light signal received by the photodetector fluctuates regularly with the flow of liquid inside the tube. The fluctuation frequency is proportional to the liquid flow velocity. After filtering out abnormal pulse signals through a sliding window, the fluctuation frequency of the filtered flow velocity signal is statistically analyzed and then multiplied by a pre-calibrated flow velocity conversion coefficient to obtain the preliminary flow velocity value of the drainage fluid. The calibration method of the flow velocity conversion coefficient is as follows: on the calibration test bench, a standard liquid with a known constant flow velocity flows through a standard calibration tube, and the fluctuation frequency of the signal output by the photodetector is recorded. The ratio of the known flow velocity value to the signal fluctuation frequency is used as the flow velocity conversion coefficient. After calibration, the flow velocity conversion coefficient is stored in the control module for real-time flow velocity calculation. Because different drainage tubes have different diameters, the volume of liquid flowing through the tube per unit time at the same flow rate varies. Therefore, it is necessary to adjust the flow rate based on the inner diameter parameter of the drainage tube. The specific adjustment method is as follows: multiply the initial flow rate value by the inner diameter correction factor, which is the ratio of the square of the currently used drainage tube's inner diameter to the square of the standard calibration tube's inner diameter. This converts the initial flow rate value into the actual real-time drainage flow rate under the current drainage tube specification. The inner diameter parameter of the drainage tube is preset by medical staff according to the actual drainage tube specifications used clinically. The cumulative drainage volume is calculated by integrating the real-time drainage flow rate value over time. Preprocessing takes the raw spectral data as input and includes: Ambient light subtraction: Acquire ambient light signals with the LED light source off and subtract the ambient light component from the actual transmitted light signal intensity; Dynamic baseline calibration: Using the reference light signal intensity collected by the standard blank solution under the same pipeline conditions as the baseline, and taking into account the light source attenuation and detector sensitivity drift, the baseline calibration is automatically performed once every preset time period to update the reference light signal intensity value; Bubble detection: Set a signal change rate threshold. When the signal change rate of any band exceeds the signal change rate threshold, it is determined that there is bubble interference at the current sampling time, and the current sampled data is marked as invalid data. Abnormal signal rejection: Set a physical reasonable range and a normal range for signal change rate, and reject sampling points whose signal strength exceeds the physical reasonable range or whose signal change rate exceeds the normal range; The drainage fluid blood grading model employs a support vector machine (SVM) as the classification algorithm, using a radial basis function (RBF) kernel. The model's input layer consists of multi-band spectral feature vectors, and the output layer provides three levels of classification labels: low blood component level, lightly bloody fluid level, and high blood component level. During training, the multi-band spectral feature vectors of standard drainage fluid samples are used as input features, and the corresponding known blood concentration grading labels are used as supervision labels. Optimal penalty and kernel function parameters are determined through grid search and cross-validation, with the goal of maximizing the classification margin to obtain the classification decision function. After training, the model parameters are stored in the monitoring sensor's storage unit. During online inference, the real-time extracted multi-band spectral feature vectors are input into the classification decision function, outputting the corresponding blood grade labels. This upgrades manual visual identification to spectral quantification grading, reducing subjective judgment differences and improving the consistency and timeliness of identifying trace blood components.

[0028] Methods for constructing multi-band spectral eigenvectors also include: Using the reference spectral data collected by the standard blank solution under the same pipeline conditions as the baseline, the normalized absorbance values ​​for each band are calculated based on the preprocessed transmitted light signal intensity and the reference spectral data:

[0029] In the formula, For band The normalized absorbance value, For band The intensity of the transmitted light signal after preprocessing For band The intensity of the reference optical signal; The band absorbance ratio is the ratio of the normalized absorbance values ​​of any two bands.

[0030] In the formula, For the first Band and the The ratio of absorbance in different wavelength bands, and The first Bands and the Normalized absorbance value for the band. This is a preset regularization constant used to prevent the denominator from being zero; Under a fixed sampling interval, the absolute value of the difference between the normalized absorbance value at the current sampling time and the previous sampling time is calculated, and the ratio of the absolute value of the difference to the sampling interval is defined as the signal change rate. The normalized absorbance values, absorbance ratios of each band, and signal change rate are combined into a multi-dimensional vector, which serves as the multi-band spectral feature vector.

[0031] This can be explained by converting the original light signal from the transmission intensity domain to the absorbance domain, allowing subsequent calculations to directly reflect the concentration of blood components in the drainage fluid. The band absorbance ratio eliminates common systematic errors such as fluctuations in light source intensity and differences in tube wall thickness, making the subsequent feature vector more robust to individual tube variations and environmental changes. Furthermore, the signal change rate captures the dynamic changes in the drainage fluid's properties over time, enabling a rapid response to sudden changes in the blood composition. Thus, a multi-band spectral feature vector is formed by combining the normalized absorbance values ​​of the three bands, the three band absorbance ratios, and the signal change rates of each band. This multi-band spectral feature vector contains both static information about the current blood concentration in the drainage fluid and dynamic information about its changes over time, providing complete input features for subsequent blood grading models.

[0032] Based on the real-time drainage rate, cumulative drainage volume, and blood grade label, local alarm judgment and tube clamping control decisions are executed according to the preset three-level monitoring rules, including: Preset flow rate alarm threshold and total flow limit threshold; When the real-time value of the drainage flow rate exceeds the flow rate alarm threshold, an abnormal flow rate alarm is triggered. When the total amount of drainage reaches the upper limit threshold, a total amount compliance alarm is triggered, and a clamping control command is generated at the same time. When the blood grade label reaches the preset blood alarm level, a blood abnormality alarm is triggered; When any of the abnormal flow rate alarm, total volume target alarm, and abnormal blood flow alarm is triggered, an audible and visual alarm signal is issued, and the corresponding tube clamping control action is executed according to the alarm priority. When the cumulative drainage volume reaches a preset percentage of the total capacity of the drainage bag, or when the estimated time for the drainage bag to fill is less than a preset time threshold, an "Immediate Clamping Reminder" label is attached to the patient. For patients with the "About to Clamp Tube" tag, multiple total volume thresholds are set. When the total drainage volume reaches each threshold, a progressive reminder of the corresponding level is pushed. When the total drainage volume reaches the highest threshold, an audible and visual alarm and a servo motor clamp tube command are automatically triggered.

[0033] This can be explained by the fact that the three-level monitoring rules correspond to three different clinical abnormality scenarios: excessively rapid drainage flow indicates potential active bleeding or drainage tube blockage risk; the cumulative drainage volume reaching the safety limit indicates the need to terminate drainage; and abnormal blood grade indicates the presence of abnormal blood components in the drainage fluid. When any abnormality is triggered, the sensor immediately issues an audible and visual alarm, ensuring that bedside medical staff can detect the abnormality immediately and drive the servo motor to perform a tube clamping action according to the preset alarm priority. The "About to Clamp" alert group is a further refined monitoring strategy in terms of total volume. Before reaching the final upper limit threshold, this plan introduces an intermediate threshold early warning mechanism based on the drainage bag capacity: when the cumulative drainage volume reaches a preset percentage of the total drainage bag capacity, or when the remaining filling time of the drainage bag estimated based on the current flow rate is less than a preset time threshold, it is determined that the drainage is about to reach the upper limit, the patient is labeled with the "About to Clamp" alert group, and the sensor data acquisition frequency is increased accordingly to obtain more accurate real-time data; After adding tags, within the framework of multi-level total volume thresholds, the central monitoring platform pushes progressive reminders of different levels based on the threshold range of the cumulative drainage volume, providing nurses with progressive operation prompts and allowing them to prepare for tube clamping in advance. When the cumulative drainage volume reaches the highest threshold, a local audible and visual alarm and a servo motor tube clamping command are automatically triggered, ensuring that drainage is terminated in a timely manner when the safe limit is reached. Through early warning and progressive reminders, complications such as fluid reflux and infection caused by insufficient preparation time for nurses due to a one-time alarm or forgetting to clamp the tube are effectively avoided. The estimated time to fill the drainage bag is calculated based on the relationship between the total capacity of the drainage bag, the cumulative amount of drainage already drained, and the current real-time drainage rate. Let the total capacity of the drainage bag be... , is a preset constant, configured by medical staff in the system according to the specifications of drainage bags actually used in clinical practice. The current cumulative drainage volume is . The real-time value of the drainage flow rate is Then the remaining capacity of the drainage bag is Expected filling time The theoretical calculation formula is:

[0034] In the formula, ,when When the time is less than the preset threshold, the system determines that the drainage is about to reach its limit and adds a "reminder group" tag to the patient indicating that the tube will be clamped soon. The formula calculation continues on the platform as the drainage process progresses. and Continuously updated The time is then dynamically refreshed to ensure the accuracy of the estimated time; It should be noted that each threshold is determined based on a combination of clinical guidelines and statistical analysis of historical clinical data: the flow rate alarm threshold is initially set at the 95th percentile of the normal drainage flow rate range at different postoperative stages (e.g., 0-6 hours, 6-24 hours, and after 24 hours). Exceeding this value indicates a potential risk of active bleeding or drainage tube blockage. The upper limit threshold for total drainage volume is directly specified by the attending physician in the medical order based on the patient's specific condition, and the system synchronously reads the medical order value from the hospital information system as the basis for judgment. The preset percentage of total drainage bag capacity and the preset threshold for remaining filling time are determined based on a combination of the nurse's average response time and shift handover buffer time, ensuring that nurses have sufficient time to prepare for clamping after receiving progressive reminders. The threshold grading scheme (e.g., 60%, 80%, and 90% of the total drainage bag capacity) corresponds to three levels of operational lead time: "Prepare for monitoring," "Prepare for clamping instruments," and "Approaching the upper limit." The interval between each level is configured based on the minimum preparation time required by the standard clinical nursing procedure. The thresholds of each level of the multi-level total threshold are divided proportionally based on the upper limit of the total drainage volume. The above initial thresholds are initially set through retrospective clinical data analysis before system deployment. After deployment, statistical dynamic optimization based on percentile method is adopted after sufficient clinical data is accumulated. Clinicians fine-tune and confirm the thresholds in the human-computer interaction interface of the central monitoring platform according to individual patient differences to ensure that the threshold settings take into account both safety and clinical operability.

[0035] Risk assessments are periodically performed based on real-time patient monitoring data, and the patient's monitoring level group is automatically adjusted based on the risk assessment results. The monitoring level groups should include at least: a high-risk observation group and a routine group; Patients are automatically placed into the high-risk observation group after surgery and are continuously monitored for a preset initial duration. After the initial monitoring period ends, a risk assessment is performed every preset assessment cycle. Based on multi-band spectral blood data, drainage fluid flow rate and flow rate fluctuation coefficient, a comprehensive risk assessment model is constructed to conduct a comprehensive assessment and output a comprehensive risk score. Based on the comparison between the comprehensive risk score and the preset risk threshold, patients are assigned to either the high-risk observation group or the routine group. The risk assessment also includes an immediate triggering mechanism, including: When the flow rate fluctuation coefficient of a patient in the standard group exceeds the fluctuation threshold in a series of preset monitoring sessions, an immediate risk assessment is triggered. When any patient's blood grade label reaches or exceeds the preset instant assessment trigger level, an instant risk assessment is triggered. Once the immediate risk assessment is completed, if the overall risk score exceeds the preset escalation threshold, the patient will be immediately escalated to the high-risk observation group.

[0036] This can be explained by the fact that the early postoperative period is a high-risk period for complications such as bleeding and exudation, requiring the highest intensive monitoring. Therefore, this protocol adopts a mechanism that mandates patients to enter the high-risk observation group in the initial postoperative stage, and then dynamically adjusts the monitoring level through periodic assessments: patients are automatically placed in the high-risk observation group after surgery, during which flow rate and spectral data are collected at the highest frequency to ensure no omissions in monitoring during the critical postoperative window period; After the postoperative high-risk window period, patients' conditions tend to stabilize, but individual differences and changes in condition still exist. Adopting a uniform high-frequency monitoring strategy would lead to data redundancy and wasted system resources, while uniformly reducing the frequency might miss abnormal changes in some patients. Therefore, this protocol introduces a periodic risk assessment mechanism, performing a risk assessment every preset assessment cycle after the initial monitoring period ends. The comprehensive risk assessment model outputs a comprehensive risk score, which is composed of three weighted assessment dimensions: First, the multi-band spectral blood data dimension, using the current blood grade label value as input; a higher blood grade reflects a greater risk. Second, the drainage fluid flow rate dimension, using the deviation of the current real-time drainage flow rate value from the preset median of the normal flow rate range as input; a larger deviation reflects a greater risk of drainage abnormalities. Third, the flow rate fluctuation coefficient dimension, using the ratio of the standard deviation to the average value of the real-time drainage flow rate within a preset time window as input; this reflects the stability of the drainage process; a larger fluctuation coefficient indicates a higher risk of abnormalities such as bleeding, blockage, or positional changes. The expression for the comprehensive risk assessment model is:

[0037] In the formula, To calculate the overall risk score, These are the blood quality label values, with 0 for low blood component levels, 1 for lightly bloody fluid levels, and 2 for high blood component levels. This is the real-time value of the drainage flow rate. The median value of the preset normal flow rate range. The velocity fluctuation coefficient is... The weighting coefficients are pre-calibrated using an expert scoring method. Data for each dimension is normalized before being entered into the formula to eliminate the influence between different dimensions. Among these, the flow velocity fluctuation coefficient... The specific calculation method is as follows: Calculate the standard deviation of the real-time values ​​of the drainage flow rate within the preset time window. Compared with the average ,calculate Patients were assigned to either a high-risk observation group or a routine group based on the comparison between their risk scores and preset risk thresholds. Since patients in the control group are not exempt from risk assessment, this protocol includes an immediate triggering mechanism as a supplementary safeguard to periodic assessments: when the flow rate fluctuation coefficient of patients in the control group exceeds the fluctuation threshold in a consecutive preset number of monitoring sessions, it indicates persistent instability in the drainage status, potentially indicating a risk of occult bleeding; when any patient's blood grade label reaches or exceeds the preset immediate assessment trigger level, it indicates a real-time change in the drainage fluid characteristics. When any of the above conditions are met, the system immediately triggers an immediate risk assessment. After the immediate risk assessment, if the risk score exceeds the preset escalation threshold, the patient is immediately escalated to the high-risk observation group, resuming high-frequency monitoring. This immediate triggering mechanism ensures that sudden changes in the patient's condition within the periodic assessment interval can be detected promptly and the monitoring level upgraded, forming a dynamic monitoring closed loop of routine monitoring, abnormal triggering, immediate assessment, and level escalation. It should be noted that the median normal flow rate range is determined based on the statistical median of historical normal drainage flow rates at different postoperative stages (e.g., 0-6 hours, 6-24 hours, and after 24 hours), and is switched in stages according to the postoperative recovery process; the flow rate fluctuation coefficient threshold is set based on the 95th percentile of the flow rate fluctuation coefficient in historical data of normal patients, and exceeding this value indicates persistent instability in the drainage status; the preset number of consecutive measurements is determined according to the clinical definition of persistent abnormality, usually set to 2-3 times, that is, real-time assessment is only triggered when multiple consecutive monitoring exceed the threshold, in order to avoid false triggering caused by noise from a single measurement; real-time assessment The trigger level is set to either a lightly bloody fluid level or a high blood component level. That is, when the blood grade label reaches the lightly bloody fluid level or above, an immediate assessment is triggered to ensure that any clinically significant color changes in the drainage fluid are promptly detected. The escalation threshold is set to a risk score threshold higher than that required for transfer from the standard group to the high-risk observation group. The specific value is determined by ROC curve analysis of the risk score distribution of patients diagnosed with clinically necessary escalation of monitoring levels in historical cases, using the score corresponding to the maximum Youden index as the escalation threshold to balance escalation sensitivity and specificity. The weighting coefficients in the risk comprehensive assessment model... The risk threshold was determined using an expert scoring method. Clinical experts compared the relative importance of the three assessment dimensions pairwise and calculated the weight coefficient of each dimension using a judgment matrix. The risk threshold was determined based on the difference in the comprehensive risk score distribution between patients diagnosed with complications and normal patients in historical cases. ROC curve analysis was used to determine the optimal cutoff value, and patients were divided into a high-risk observation group and a routine group. Based on the above settings, the specific rules for classifying guardianship levels are illustrated with an example: Post-operatively, patients were mandatory into a 24-hour high-risk observation group, followed by adaptive grouping. The adaptive grouping rules were as follows: risk assessment was performed every 8 hours, based on multi-band spectral blood data, drainage fluid flow rate, and flow rate fluctuation coefficient, to output a risk score. Patients were assigned to either the high-risk observation group or the standard group based on the comparison of the risk score with a preset risk threshold. An instant trigger mechanism was also set: if a patient in the standard group had two consecutive flow rate fluctuation coefficients exceeding the normal range, or if spectral scanning detected a lightly bloody fluid level or higher, an instant risk assessment was automatically triggered. If the risk score exceeded the escalation threshold, the patient was immediately escalated to the high-risk observation group. When the patient's cumulative drainage volume reaches 80% of the total drainage bag capacity, or when the estimated filling time of the drainage bag, calculated based on the cumulative drainage volume and real-time drainage flow rate, is less than 30 minutes, an "Imminent Tube Clamping Reminder" tag is added to the patient. For patients with added tags, three total volume thresholds are set: 90%, 95%, and 100% of the total drainage bag capacity. When the cumulative drainage volume reaches each threshold, the central monitoring platform pushes a progressive reminder of the corresponding level. When the cumulative drainage volume reaches 100%, a local audible and visual alarm and a servo motor tube clamping command are automatically triggered.

[0038] The central monitoring platform provides centralized visualization, anomaly alarm, and data storage management for monitoring data from all connected sensors, including: The monitoring data from the risk observation team is uploaded to the central monitoring platform in real time at the highest upload frequency. The monitoring data of the regular group is uploaded to the central monitoring platform at a second upload frequency, which is lower than the first upload frequency. The data uploaded by the regular group is an aggregated data packet within a preset time window. The aggregated data packet includes at least the mean, maximum and minimum values ​​of the monitoring data within the time window. The central monitoring platform receives real-time monitoring data uploaded by each monitoring sensor through a wireless communication gateway, and displays the patient drainage status information corresponding to each sensor on a visual dashboard by bed. When monitoring data uploaded by any sensor triggers abnormal conditions, the central monitoring platform simultaneously displays a pop-up notification and an audible and visual alarm, and highlights the abnormal bed. The central monitoring platform automatically saves all monitoring data according to the preset storage period, supports querying and exporting historical data by patient ID and time period, and automatically generates drainage flow rate change curves and cumulative drainage volume growth curves. When the total drainage volume reaches the preset clamping reminder threshold, a clamping reminder window will pop up on the visual dashboard to prompt medical staff to clamp the drainage tube. The clamping reminder threshold is less than or equal to the upper limit threshold of the total drainage volume.

[0039] This can be explained by the fact that the central monitoring platform, as a centralized management terminal that gathers monitoring data from various sensors, continuously receives drainage status information from each bed through a wireless communication gateway. It integrates the data that originally needed to be checked at each bedside into a unified visual dashboard, realizing centralized management of drainage monitoring throughout the department. Medical staff can remotely view the real-time status of all drainage tubes in the ward from the nurses' station or the doctor's office. On the one hand, it can control the risk of drainage abnormalities throughout the process and improve the overall safety level of postoperative monitoring. On the other hand, it reduces the frequency of bedside rounds, optimizes the allocation of nursing staff in the department, and effectively alleviates the workload of clinical medical staff. It should be noted that the sensor data acquisition frequency and the monitoring data upload frequency are set independently, and together they determine the timeliness of the data on the central monitoring platform. The acquisition frequency determines the time interval for the sensor to sample various types of data, while the upload frequency determines the strategy for transmitting data from the sensor to the central monitoring platform after sampling. After data acquisition is completed, it is sent to the platform and refreshed for display according to the corresponding upload strategy. Therefore, the timeliness of the platform display is jointly determined by the acquisition frequency and the upload strategy. The acquisition frequency and upload frequency configurations for different monitoring levels and additional tag states are shown in Table 1: Table 1. Data collection and upload frequency configurations under different monitoring levels and additional tag statuses.

[0040] As shown in Table 1, the high-risk observation group sampled data at the first acquisition frequency (flow rate every 5 seconds, spectrum every 10 seconds, total volume every 30 seconds) and simultaneously uploaded the data to the platform at the first upload frequency (real-time, one-by-one upload), ensuring that any transient abnormalities during the critical postoperative window period for high-risk patients could be detected promptly. The routine group sampled data at the second acquisition frequency (flow rate every 30 seconds, spectrum every 2 minutes, total volume every 5 minutes) and uploaded the data to the platform at the second upload frequency (aggregate upload every 5 minutes), effectively reducing the data reception and storage pressure on the central monitoring platform while meeting the basic monitoring needs of stable patients. After triggering the tag indicating imminent tube clamping, the sensor data acquisition frequency is adjusted differently based on the patient's original monitoring level. For patients originally in the routine group, the acquisition frequency is increased from the second to the third frequency after tagging. Specifically, the infrared flow data acquisition interval is adjusted to every 10 seconds, the multi-band spectral scan trigger interval is adjusted to every 30 seconds, and the cumulative drainage volume update interval is adjusted to every 10 seconds. For patients originally in the high-risk observation group, the first acquisition frequency remains unchanged after tagging, i.e., infrared flow data is acquired every 5 seconds, multi-band spectral scan is triggered every 10 seconds, and the cumulative drainage volume update interval is adjusted to every 10 seconds. The upload frequency is also adjusted synchronously: the cumulative volume is uploaded in real time every 10 seconds. The uploaded data includes three elements: flow rate, cumulative volume percentage, and blood grade. The central monitoring platform updates the total volume comparison result every 10 seconds based on this data and pushes corresponding progressive reminders when the cumulative volume reaches the three thresholds of 90%, 95%, and 100% of the total drainage bag capacity. When the cumulative volume reaches 100%, a local audible and visual alarm and a servo motor tube clamping command are automatically triggered. The above-mentioned collection frequency and upload frequency settings correspond to the data collection density and platform display timeliness requirements under different monitoring levels. The collection and upload strategies for each level are independent of each other and are automatically switched and executed by the system according to the monitoring level classification results and the status of the attached tags.

[0041] Furthermore, based on the same inventive concept as the aforementioned intelligent drainage tube central monitoring method, this solution proposes an intelligent drainage tube central monitoring system, such as... Figure 3 As shown: An external drainage monitoring sensor is used to install on the outer wall of the drainage tube to collect multimodal sensing data of the drainage fluid, and to perform local calculations, alarm judgments and tube clamping control based on the multimodal sensing data. A wireless communication gateway is used to receive monitoring data uploaded by one or more external drainage monitoring sensors, and after completing data protocol conversion and data caching, transmit the data to the central monitoring station. The central monitoring station is used for centralized visualization, alarm management, parameter configuration, tube clamping reminder, data storage, and equipment status management of monitoring data from all connected sensors. It includes: a multi-patient visualization dashboard module, an alarm management module, a parameter configuration module, a tube clamping reminder module, a data storage and statistics module, and an equipment status management module.

[0042] Furthermore, this solution proposes an intelligent drainage tube central monitoring and drainage sensor, including: a housing, an infrared non-contact flow detection device 5, a multi-band optical detection device, a tube clamping actuator 8, and a circuit board module 9, such as... Figure 4 As shown: The outer casing includes a front shell 1 and a rear shell 2. The bottoms of the two shells are connected by a hinge 4. When the front shell 1 and the rear shell 2 are fastened together, they form a receiving cavity. The interior of the cavity maintains a light-proof environment. The front shell 1 includes a front shell plate 101. The front shell plate 101 has symmetrically positioned fixing groove plates 102 on both sides. Pins 3 are installed on the groove plates 102, passing through the groove plates 102 and slidingly connected to them. Rotating buckles 103 are provided on the left and right sides of the bottom edge of the top of the front shell plate 101. Buckle holes are provided on the left and right sides of the bottom edge of the top of the rear shell 2. 203, the position of the buckle hole 203 corresponds to the position of the rotating buckle 103 of the front shell plate 101. When the front shell 1 and the rear shell 2 are fastened together, the rotating buckle 103 is inserted into the buckle hole 203. By rotating the rotating buckle 103, the front shell 1 and the rear shell 2 are stably fastened together. The left and right side walls of the rear shell 2 are coaxially provided with pipe through holes 201 and pin grooves 202. Before the front shell 1 and the rear shell 2 are fastened together, the pin 3 is kept in an uninserted state. After the front shell 1 and the rear shell 2 are fastened together, the pin 3 is inserted into the pin groove 202 through the groove plate 102, so that the pin 3 clamps the drainage tube 10. The infrared non-contact flow detection device 5 is fixed in the left pipe perforation 201, and the main body is placed on the outside of the rear housing 2. When the drainage tube 10 is placed, the infrared non-contact flow detection device 5 adopts an opening and closing snap-fit ​​structure, which can be opened and snapped into the outer wall of the drainage tube 10 to complete the fixation of the detection position. The multi-band optical detection device includes a photodetector 6 and a background platform 7. The photodetector 6 is fixed at the top of the inner cavity of the rear housing 2. The background platform 7 is equipped with a multi-band LED light source, which is fixed at the bottom of the inner cavity of the rear housing 2 and faces the photodetector 6. The drainage tube 10 is located between the photodetector 6 and the background platform 7. The detection light emitted by the background platform 7 penetrates vertically from bottom to top through the lower wall of the drainage tube 10, the drainage fluid inside the tube, and the upper wall of the drainage tube 10, and is then received by the photodetector 6. The detection element does not come into contact with the drainage fluid throughout the entire process. The tube clamping actuator 8 is fixed to the bottom of the inner cavity of the rear housing 2, on the right side of the background platform 7, directly below the corresponding drainage tube 10. The tube clamping actuator 8 includes a control motor 801, a main rotary transmission gear 802, a driven rotary transmission gear 804, an active elliptical tube clamping head 803, and a driven elliptical tube clamping head 805. The control motor 801 is fixed to the bottom of the inner cavity of the rear housing 2. The power output end of the control motor 801 is connected to the main rotary transmission gear 802. The main rotary transmission gear 802 is connected to the driven rotary transmission gear 804. The upper end of the main rotary transmission gear 802... An active elliptical tube clamping head 803 is fixedly installed, and a driven elliptical tube clamping head 805 is fixedly installed on the upper end of a rotary transmission gear 804. When a tube clamping control command is received, the control motor 801 drives the active elliptical tube clamping head 803 to rotate through the main rotary transmission gear 802. At the same time, the main rotary transmission gear 802 drives the driven rotary transmission gear 804 to rotate, thereby driving the driven elliptical tube clamping head 805 to rotate. The rotation of the active elliptical tube clamping head 803 and the driven elliptical tube clamping head 805 clamps the drainage tube, thereby closing the cavity of the drainage tube. The circuit board module 9 is fixed to the upper inner wall of the rear housing 2, located to the right of the photodetector 6 and directly above the pipe clamping actuator 8, and is electrically connected to the infrared non-contact flow detection device 5, the multi-band optical detection device and the pipe clamping actuator 8 respectively.

[0043] Furthermore, the rear end of the latch 3 is also equipped with a locking mechanism, the cross-section of which is as follows: Figure 5 As shown: The locking mechanism consists of a sliding steel ball 301 and a spring structure 302. When the pin 3 is inserted into the grooved plate 102, the first set of sliding steel balls 301 and spring structure 302 touch the outer wall of the grooved plate 102, squeezing the sliding steel ball 301 to move towards the spring structure 302, so that the spring structure 302 is in a compressed state. After the sliding steel ball 301 crosses the outer wall of the grooved plate 102, the spring structure 302 pushes the sliding steel ball 301 into the cavity in the grooved plate 102 for locking. The locking mechanism is provided in multiple sets, and the positions of the multiple sets of locking mechanisms are adapted to the standard specifications of the drainage tube. By controlling the number of sets of sliding steel balls 301 and spring structure 302 entering the grooved plate, the insertion depth of the pin 3 is controlled to clamp drainage tubes of different sizes.

[0044] The steps for using the drainage monitoring sensor include: First step (with the drainage monitoring sensor not turned on): pull out the pins 3 on both sides of the groove plate 102 of the front housing 1, rotate the rotating buckle 103, and open the front housing 1 of the drainage monitoring sensor; The second step is to open the buckle structure of the infrared non-contact flow detection device 5, pass the drainage tube 10 through the infrared non-contact flow detection device 5, and through the pipe through holes 201 on both sides of the rear housing 2. Place the drainage tube 10 between the active elliptical clamping head 803 and the driven elliptical clamping head 805, adjust the active elliptical clamping head 803 and the driven elliptical clamping head 805 to keep the drainage tube 10 in a non-compression state, straighten the drainage tube 10, and fasten the infrared non-contact flow detection device 5. The third step is to close the front housing 1 and the rear housing 2, insert the pin 3 into the groove plate 102, rotate the rotating buckle 103, and complete the installation of the drainage tube 10. The fourth step involves activating the infrared non-contact flow detection device 5 and the multi-band optical detection device based on the control signals from the circuit board module, and dynamically adjusting the pipe clamping actuator 8.

[0045] The circuit board module also includes: The flow metering unit is used to perform sliding window filtering on the raw flow velocity data output by the infrared non-contact flow detection device, calculate the real-time value of the drainage flow velocity in combination with the preset inner diameter parameters of the drainage tube, and calculate the cumulative total drainage by integrating the real-time value of the drainage flow velocity over time. The local alarm unit is used to control the buzzer and LED warning lights to emit audible and visual alarm signals when the abnormal flow rate alarm, total volume target alarm, or abnormal blood alarm is triggered, based on the judgment results of the three-level monitoring rules. The wireless transmission unit is used to upload multimodal sensor data, real-time drainage flow rate, cumulative drainage volume, and blood grade label to the wireless communication gateway via wireless communication, and to receive parameter configuration instructions issued by the central monitoring platform. The parameter storage and adjustment unit is used to store preset flow rate alarm thresholds, total drainage volume upper limit thresholds, blood alarm levels, drainage tube inner diameter parameters, and model parameters of the drainage fluid blood grading model. It also supports medical staff to adjust the stored parameters on-site or remotely via physical buttons or wireless communication.

[0046] In summary, the advantages of this invention are: by using an external non-contact multimodal sensor and a dynamic risk grading strategy, the monitoring density can be adaptively adjusted according to the patient's real-time status, thereby achieving precise, efficient, and intelligent management of the entire patient flow management process that also optimizes medical resources.

[0047] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for central monitoring of intelligent drainage tubes, characterized in that, include: Multimodal sensing data of the drainage fluid is collected by a monitoring sensor placed on the wall of the drainage tube. The multimodal sensing data includes at least fluid flow rate data and liquid spectral data. Based on multimodal sensor data, the real-time value of drainage flow rate and the cumulative total drainage volume are calculated respectively, and the blood components of the drainage fluid are classified and determined, and blood grade labels are output. Based on the real-time value of drainage flow rate, the cumulative drainage volume, and the blood grade label, local alarm judgment and tube clamping control decisions are executed in accordance with the preset three-level monitoring rules; Based on the postoperative monitoring status of patients, patients are divided into different monitoring level groups. Different sensor data acquisition frequencies and data upload strategies are set for different monitoring level groups. Risk assessments are periodically performed based on the real-time monitoring data of patients, and the monitoring level group of patients is automatically adjusted based on the risk assessment results. Multimodal sensor data, real-time drainage flow rate, cumulative drainage volume, and blood grade labels are uploaded to the central monitoring platform via wireless communication. The central monitoring platform then provides centralized visualization, alarm functions, and data storage management for all monitoring data from the connected sensors.

2. The intelligent drainage tube central monitoring method according to claim 1, characterized in that, The multimodal sensing data of the drainage fluid collected by the monitoring sensor externally placed on the wall of the drainage tube includes: An infrared non-contact flow detection unit is used to collect liquid flow rate signals by penetrating the wall of the drainage tube with infrared light. After signal amplification and analog-to-digital conversion, the raw flow rate data is output. A multi-band optical detection unit is used, with multi-band LED light sources and photodetectors set on both sides of the drainage tube to collect the intensity of transmitted light signals in the green, red, and near-infrared bands, respectively, as raw spectral data.

3. The intelligent drainage tube central monitoring method according to claim 2, characterized in that, Based on multimodal sensor data, the real-time value of drainage flow rate and the cumulative drainage volume are calculated respectively, and the blood components of the drainage fluid are classified and determined, outputting blood grade labels including: The raw flow velocity data is subjected to sliding window filtering to remove abnormal pulse signals, resulting in a filtered flow velocity signal. Combined with the preset inner diameter parameters of the drainage tube, the real-time value of the drainage flow velocity is calculated. The cumulative total of drainage is calculated by integrating the real-time value of the drainage flow rate over time. The raw spectral data is preprocessed to output the preprocessed transmitted light signal intensity of each band. The preprocessing includes ambient light subtraction, dynamic baseline calibration, bubble identification, and abnormal signal removal. Based on the normalized absorbance values ​​of each band, a multi-band spectral feature vector is calculated. The multi-band spectral feature vector includes the normalized absorbance of each band, the band absorbance ratio, and the signal change rate. Multiple standard drainage fluid samples with known blood concentrations were collected. Under standard tubing conditions, multi-band optical detection units were used to acquire multi-band spectral data of each sample, and multi-band spectral feature vectors of each sample were extracted. Using multi-band spectral feature vectors as input and known blood concentration grading labels of standard drainage fluid samples as output targets, a classification algorithm is used for training to obtain a drainage fluid blood grading model; The multi-band spectral feature vector is input into the drainage fluid blood grading model, and the blood grade label is output. The blood grade label includes at least low blood component grade, light blood fluid grade, and high blood component grade.

4. The intelligent drainage tube central monitoring method according to claim 3, characterized in that, The method for constructing the multi-band spectral feature vector also includes: Using the reference spectral data collected by the standard blank solution under the same pipeline conditions as the baseline, the normalized absorbance values ​​for each band are calculated based on the preprocessed transmitted light signal intensity and the reference spectral data: In the formula, For band The normalized absorbance value, For band The intensity of the transmitted light signal after preprocessing For band The intensity of the reference optical signal; The band absorbance ratio is the ratio of the normalized absorbance values ​​of any two bands. In the formula, For the first Band and the The ratio of absorbance in different wavelength bands, and The first Bands and the Normalized absorbance value for the band. This is a preset regularization constant used to prevent the denominator from being zero; Under a fixed sampling interval, the absolute value of the difference between the normalized absorbance value at the current sampling time and the previous sampling time is calculated, and the ratio of the absolute value of the difference to the sampling interval is defined as the signal change rate. The normalized absorbance values, absorbance ratios of each band, and signal change rate are combined into a multi-dimensional vector, which serves as the multi-band spectral feature vector.

5. The intelligent drainage tube central monitoring method according to claim 4, characterized in that, The process of executing local alarm judgment and tube clamping control decisions based on real-time drainage flow rate, cumulative drainage volume, and blood grade label, according to preset three-level monitoring rules, includes: Preset flow rate alarm threshold and total flow limit threshold; When the real-time value of the drainage flow rate exceeds the flow rate alarm threshold, an abnormal flow rate alarm is triggered. When the total amount of drainage reaches the upper limit threshold, a total amount compliance alarm is triggered, and a clamping control command is generated at the same time. When the blood grade label reaches the preset blood alarm level, a blood abnormality alarm is triggered; When any of the abnormal flow rate alarm, total volume target alarm, and abnormal blood flow alarm is triggered, an audible and visual alarm signal is issued, and the corresponding tube clamping control action is executed according to the alarm priority. When the cumulative drainage volume reaches a preset percentage of the total capacity of the drainage bag, or when the estimated time for the drainage bag to fill is less than a preset time threshold, an "Immediate Clamping Reminder" label is attached to the patient. For patients with the "About to Clamp Tube" tag, multiple total volume thresholds are set. When the total drainage volume reaches each threshold, a progressive reminder of the corresponding level is pushed. When the total drainage volume reaches the highest threshold, an audible and visual alarm and a servo motor clamp tube command are automatically triggered.

6. The intelligent drainage tube central monitoring method according to claim 5, characterized in that, The process of periodically performing risk assessments based on real-time patient monitoring data and automatically adjusting the patient's monitoring level group based on the risk assessment results includes: The monitoring level groups should include at least: a high-risk observation group and a routine group; Patients are automatically placed into the high-risk observation group after surgery and are continuously monitored for a preset initial duration. After the initial monitoring period ends, a risk assessment is performed every preset assessment cycle. Based on multi-band spectral blood data, drainage fluid flow rate and flow rate fluctuation coefficient, a comprehensive risk assessment model is constructed to conduct a comprehensive assessment and output a comprehensive risk score. Based on the comparison between the comprehensive risk score and the preset risk threshold, patients are assigned to either the high-risk observation group or the routine group. The risk assessment also includes an immediate triggering mechanism, including: When the flow rate fluctuation coefficient of a patient in the standard group exceeds the fluctuation threshold in a series of preset monitoring sessions, an immediate risk assessment is triggered. When any patient's blood grade label reaches or exceeds the preset instant assessment trigger level, an instant risk assessment is triggered. Once the immediate risk assessment is completed, if the overall risk score exceeds the preset escalation threshold, the patient will be immediately escalated to the high-risk observation group.

7. The intelligent drainage tube central monitoring method according to claim 6, characterized in that, The centralized monitoring platform provides centralized visualization, alarm functions, and data storage management for all connected sensors' monitoring data, including: The monitoring data from the risk observation team is uploaded to the central monitoring platform in real time at the highest upload frequency. The monitoring data of the regular group is uploaded to the central monitoring platform at a second upload frequency, which is lower than the first upload frequency. The data uploaded by the regular group is an aggregated data packet within a preset time window. The aggregated data packet includes at least the mean, maximum and minimum values ​​of the monitoring data within the time window. The central monitoring platform receives real-time monitoring data uploaded by each monitoring sensor through a wireless communication gateway, and displays the patient drainage status information corresponding to each sensor on a visual dashboard by bed. When monitoring data uploaded by any sensor triggers abnormal conditions, the central monitoring platform simultaneously displays a pop-up notification and an audible and visual alarm, and highlights the abnormal bed. The central monitoring platform automatically saves all monitoring data according to the preset storage period, supports querying and exporting historical data by patient ID and time period, and automatically generates drainage flow rate change curves and cumulative drainage volume growth curves. When the total drainage volume reaches the preset clamping reminder threshold, a clamping reminder window will pop up on the visual dashboard to prompt medical staff to clamp the drainage tube. The clamping reminder threshold is less than or equal to the upper limit threshold of the total drainage volume.

8. A central monitoring system for intelligent drainage tubes, characterized in that, A method for implementing the central monitoring of an intelligent drainage tube as described in any one of claims 1-7, comprising: An external drainage monitoring sensor is used to install on the outer wall of the drainage tube to collect multimodal sensing data of the drainage fluid, and to perform local calculations, alarm judgments and tube clamping control based on the multimodal sensing data. A wireless communication gateway is used to receive monitoring data uploaded by one or more external drainage monitoring sensors, and after completing data protocol conversion and data caching, transmit the data to the central monitoring station. The central monitoring station is used for centralized visualization, alarm management, parameter configuration, tube clamping reminder, data storage, and equipment status management of monitoring data from all connected sensors. It includes: a multi-patient visualization dashboard module, an alarm management module, a parameter configuration module, a tube clamping reminder module, a data storage and statistics module, and an equipment status management module.

9. A drainage monitoring sensor, adapted to the intelligent drainage tube central monitoring system as described in claim 8, comprising: The housing, the infrared non-contact flow detection device (5), the multi-band optical detection device, the pipe clamping actuator (8), and the circuit board module (9) are characterized by: The outer shell includes a front shell (1) and a rear shell (2). The bottom of the front shell (1) and the rear shell (2) are connected by a flap hinge (4). The front shell (1) includes a front shell plate (101). The front shell plate (101) has a fixed groove plate (102) symmetrically positioned on both sides. A pin (3) is installed on the groove plate (102). Rotary buckles (103) are provided on the left and right sides of the lower edge of the top of the front shell plate (101). Buckle holes (203) are provided on the left and right sides of the lower edge of the top of the rear shell (2). Pipe through holes (201) and pin grooves (202) are coaxially opened on the left and right side walls of the rear shell (2). The infrared non-contact flow detection device (5) is fixed in the left pipe perforation (201), and the main body is placed on the outside of the rear housing (2); The multi-band optical detection device includes a photodetector (6) and a background platform (7). The photodetector (6) is fixed at the top of the inner cavity of the rear housing (2), and the background platform (7) is fixed at the bottom of the inner cavity of the rear housing (2) and faces the photodetector (6). The drainage tube (10) is located between the photodetector (6) and the background platform (7). The tube clamping device (8) is fixed at the bottom of the inner cavity of the rear housing (2), on the right side of the background platform (7), directly below the drainage tube (10). The tube clamping device (8) includes a control motor (801), a main rotary transmission gear (802), a driven rotary transmission gear (804), an active elliptical tube clamping head (803), and a driven elliptical tube clamping head (805). The control motor (801) is fixed at the bottom of the inner cavity of the rear housing (2). The circuit board module (9) is fixed to the upper inner wall of the rear housing (2), located to the right of the photodetector (6) and directly above the clamping actuator (8).

10. A drainage monitoring sensor according to claim 9, characterized in that, The circuit board module also includes: The flow metering unit is used to perform sliding window filtering on the raw flow velocity data output by the infrared non-contact flow detection device, calculate the real-time value of the drainage flow velocity in combination with the preset inner diameter parameters of the drainage tube, and calculate the cumulative total drainage by integrating the real-time value of the drainage flow velocity over time. The local alarm unit is used to control the buzzer and LED warning lights to emit audible and visual alarm signals when the abnormal flow rate alarm, total volume target alarm, or abnormal blood alarm is triggered, based on the judgment results of the three-level monitoring rules. The wireless transmission unit is used to upload multimodal sensor data, real-time drainage flow rate, cumulative drainage volume, and blood grade label to the wireless communication gateway via wireless communication, and to receive parameter configuration instructions issued by the central monitoring platform. The parameter storage and adjustment unit is used to store preset flow rate alarm thresholds, total drainage volume upper limit thresholds, blood alarm levels, drainage tube inner diameter parameters, and model parameters of the drainage fluid blood grading model. It also supports medical staff to adjust the stored parameters on-site or remotely via physical buttons or wireless communication.

Images