Hospital medical gas monitoring and regulating system and method based on big data intelligent judgment

Abstract

The application provides a hospital medical gas monitoring and regulating system and method based on big data intelligent judgment, belongs to the technical field of medical informatization, and realizes real-time collection of pressure, flow rate and flow data of various medical gases in areas such as wards and operating rooms through a pipeline monitoring module, generates a pipeline operation index, performs time synchronization and quality evaluation on the index through a data integration module, forms integrated data, performs adaptive classification on the data through a background analysis module by using a topological space mapping and a nonlinear probability matrix, generates four operation level marks of A, B, C and D, and intelligently controls adjusting devices of various gas pipelines according to the level marks in combination with chaos prediction and dynamic optimization, so that automatic monitoring, analysis and regulation of the medical gas pipeline are realized, safe and efficient operation of the system is ensured, a technical leap from passive response to active prediction and from static threshold to dynamic classification is realized, the control precision of the gas concentration is improved, and the abnormal response time is shortened to seconds.

Description

Technical Field

[0001] This invention relates to the field of medical information technology, and in particular to a hospital medical gas monitoring and control system and method based on big data intelligent judgment. Background Technology

[0002] Medical gas pipeline systems are a crucial infrastructure component of modern hospitals, responsible for delivering medical oxygen, medical air, medical nitrous oxide, and medical vacuum gases to various departments. According to the "Technical Specification for Medical Gas Engineering" (GB50751-2012) and "Medical Gas Systems Part 1: Piping Systems for Compressed Medical Gases and Vacuum" (YY / T0186.1-2016), medical gas pipeline systems must ensure a stable and reliable gas supply. Pipeline pressure, flow rate, and flow parameters must be strictly controlled within specified ranges to ensure medical safety.

[0003] Traditional medical gas pipeline monitoring systems primarily employ fixed-point pressure monitoring and simple threshold alarm mechanisms, triggering an alarm when the pipeline pressure exceeds a preset range. This monitoring method has the following shortcomings: First, it only monitors pressure parameters, lacking comprehensive monitoring of flow rate and velocity, thus failing to fully assess the pipeline's operational status; second, it uses fixed thresholds, making it unable to adapt to changes in gas demand across different departments and at different times; third, it lacks predictive capabilities, only able to respond passively after problems occur, unable to prevent pipeline failures in advance; and fourth, it does not consider the mutual influence and coordinated control requirements between multiple gas pipeline systems.

[0004] As hospitals expand and medical needs increase, medical gas pipeline systems are becoming increasingly complex. Wards primarily require medical oxygen, medical air, and negative pressure suction, while operating rooms require medical oxygen, medical air, carbon dioxide, and medical nitrous oxide for anesthesia and anesthetic waste gas emission systems. ICUs and special departments may also require medical nitrogen and other gases. Existing monitoring systems struggle to effectively manage such a complex network of pipelines across multiple locations, gases, and parameters. Therefore, a new medical gas pipeline management system capable of intelligent monitoring, predictive control, and multi-objective optimization is urgently needed. Summary of the Invention

[0005] The purpose of this invention is to provide a monitoring and control system and method for hospital medical gas pipelines based on big data intelligent judgment. By introducing advanced technologies such as multi-dimensional topological space mapping, nonlinear probability matrix and chaotic prediction, it realizes multi-parameter comprehensive monitoring and intelligent control of medical gas pipelines, solves the problems of single monitoring parameters, lag response, coarse control and unpredictability in traditional systems, and improves the safety, stability and resource utilization efficiency of medical gas pipeline systems.

[0006] This invention proposes a hospital medical gas pipeline monitoring and control system based on big data intelligent judgment, comprising a pipeline monitoring module, a data integration module, a background analysis module, a control module, an early warning module, and a knowledge base module. The pipeline monitoring module monitors pressure, velocity, and flow rate data of medical gas pipelines in different locations. In the ward area, it monitors medical oxygen, medical air, and medical vacuum pipelines; in the operating room area, it monitors medical oxygen, medical air, medical nitrous oxide, medical vacuum, and anesthetic waste gas emission pipelines. The data integration module receives pipeline operation indices and integrates them after time synchronization and quality assessment. The background analysis module maps the integrated data to a multi-dimensional pipeline state space based on a topological space mapping mechanism, and performs adaptive hierarchical judgment using a nonlinear probability matrix to generate pipeline operation level labels. The control module controls the regulating devices of each medical gas pipeline based on these level labels through chaotic prediction and dynamic optimization.

[0007] The present invention has the following beneficial effects:

[0008] First, by using multi-dimensional pipeline topology state space representation, multi-parameter comprehensive monitoring of medical gas pipeline systems is realized. It considers three key parameters: pressure, velocity, and flow rate, breaking through the limitations of traditional single pressure monitoring and enabling a comprehensive assessment of pipeline operating status.

[0009] Second, the adaptive classification and determination mechanism using a nonlinear probability matrix can comprehensively consider the absolute value, rate of change, historical trend, and environmental factors of pipeline parameters. It can dynamically adjust the classification standards according to the gas consumption patterns of different locations and time periods, thereby achieving more accurate pipeline status classification and avoiding the limitations of fixed threshold judgment.

[0010] Third, the introduction of a collaborative control and decision-making system combining chaotic prediction and dynamic optimization can predict the changing trends of pipeline parameters in advance, carry out preventive control before a fault occurs, reduce pipeline fluctuations, improve system stability, and transform from a passive response to an active predictive management mode.

[0011] Fourth, differentiated monitoring and control are implemented for the gas usage characteristics of different locations. The ward area focuses on ensuring basic gas supply and negative pressure suction, the operating room area focuses on ensuring anesthetic gas supply and exhaust gas emission, and special departments are configured with corresponding gas pipelines according to actual needs. This achieves refined management, improves the safety and reliability of the medical gas pipeline system, increases pipeline pressure control accuracy by more than 60%, shortens abnormal response time to the second level, and reduces system failure rate by more than 75%. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the overall structure of the hospital medical gas pipeline monitoring and control system based on big data intelligent judgment, which is the subject of this invention.

[0013] Figure 2This is a detailed structural diagram of the pipeline monitoring module of the present invention.

[0014] Figure 3 This is a detailed structural diagram of the background analysis module of the present invention.

[0015] Figure 4 This is a detailed structural schematic diagram of the state characterization unit of the present invention.

[0016] Figure 5 This is a detailed structural diagram of the hierarchical determination unit of the present invention.

[0017] Figure 6 This is a detailed structural diagram of the control and decision-making unit of the present invention.

[0018] Figure 7 This is a detailed structural diagram of the control module of the present invention.

[0019] Figure 8 This is a flowchart illustrating the hospital medical gas pipeline monitoring and control method based on big data intelligent judgment, as described in this invention. Detailed Implementation

[0020] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments, so as to better understand the advantages and features of the present invention. It should be noted that the following embodiments are for illustrative purposes only and should not be construed as limiting the scope of protection of the present invention.

[0021] like Figure 1 As shown, the hospital medical gas pipeline monitoring and control system based on big data intelligent judgment provided by the present invention includes a pipeline monitoring module 1, a data integration module 2, a background analysis module 3, a control module 4, an early warning module 5, and a knowledge base module 6.

[0022] Pipeline monitoring module 1 is electrically connected to data integration module 2, data integration module 2 is electrically connected to backend analysis module 3, backend analysis module 3 is electrically connected to control module 4, early warning module 5 is electrically connected to both backend analysis module 3 and control module 4, and knowledge base module 6 is electrically connected to backend analysis module 3. The entire system forms a closed-loop control architecture, enabling real-time monitoring, intelligent analysis, predictive control, and continuous optimization of medical gas pipelines.

[0023] The pipeline monitoring module 1 is used to monitor pressure, velocity, and flow rate data of medical gas pipelines in different locations and generate a pipeline operation index. Preferably, the pipeline monitoring module 1 performs routine sampling every 5 minutes. When the rate of change of pipeline parameters exceeds a preset threshold, the sampling frequency is automatically increased to 10 seconds per sampling to ensure timely detection of abnormal situations. For example, when the pressure change rate of the oxygen pipeline in the operating room exceeds 0.02 MPa / min, the system determines that there may be abnormal gas demand or pipeline leakage, and immediately increases the sampling frequency for intensive monitoring.

[0024] In one embodiment of the present invention, the pipeline pressure is measured by a pressure sensor installed at a key node of the pipeline, with a measurement range of 0-1.0 MPa, an accuracy of ±0.5%, and a response time of less than 1 s. The pipeline flow velocity is measured by a vortex flow meter or an ultrasonic flow meter, with a measurement range of 0-100 L / min, an accuracy of ±1%, and a response time of less than 2 s. The pipeline flow rate is calculated by integrating the flow velocity and the pipeline cross-sectional area, with units of m³ / h. The temperature parameter is measured by a thermocouple temperature sensor, with a measurement range of -20℃ to 80℃ and an accuracy of ±0.5℃. All these sensors comply with the requirements of "Medical Gas Systems Part 1: Piping Systems for Compressed Medical Gases and Vacuum" (YY / T 0186.1-2016).

[0025] Data integration module 2 receives pipeline operation indices from pipeline monitoring module 1 and integrates these indices into unified data after time synchronization and quality assessment. Since there may be slight differences in the acquisition times of different monitoring points, data integration module 2 uses Network Time Protocol (NTP) for time synchronization to ensure the consistency of data timestamps, achieving millisecond-level synchronization accuracy. Furthermore, data integration module 2 performs quality assessment on the received data, identifies and marks abnormal data points, and uses a Kalman filter algorithm for data smoothing to improve data reliability. For example, when the measured value of a pressure sensor suddenly deviates from the historical mean by more than three standard deviations, the system marks this data as suspicious and simultaneously calls data from adjacent monitoring points for cross-validation.

[0026] The background analysis module 3 is the core of this system. It receives and integrates data, maps the integrated data to a multi-dimensional pipeline state space based on a topology space mapping mechanism, and performs adaptive classification using a nonlinear probability matrix to generate pipeline operation level labels. The background analysis module 3 employs innovative technologies such as multi-dimensional topology space representation, nonlinear probability classification, and chaotic prediction, achieving a technological leap from passive response to active prediction, and from static thresholds to dynamic classification. In one embodiment of this invention, the background analysis module 3 runs on a high-performance server configured with an Intel Xeon processor and 128GB of memory, capable of simultaneously processing real-time data streams from more than 300 monitoring points with a data processing latency of less than 100ms, meeting the requirements for real-time monitoring and rapid response.

[0027] The control module 4 receives level markers and, based on these markers, uses chaotic prediction and dynamic optimization to control the regulating devices of each medical gas pipeline. According to the control strategy generated by the background analysis module 3, the control module 4 coordinates the control of electric regulating valves, pressure regulators, vacuum pumps, and other equipment in each gas pipeline to ensure the safety and stability of the pipeline system. In one embodiment of the invention, the control module 4 adopts a distributed control architecture, setting up area controllers in each department and communicating with the central control system via a fieldbus, which ensures both real-time control and improved system reliability.

[0028] The early warning module 5 generates tiered early warning signals when abnormal pipeline pressure, flow rate, or flow rate, equipment malfunction, or deviation in control effectiveness is detected. The module has three warning levels: alert (yellow), warning (orange), and emergency (red). For example, an alert warning is issued when the oxygen pipeline pressure in the ward falls below 0.35 MPa, a warning warning below 0.3 MPa, and an emergency warning below 0.25 MPa. The warning signals are sent to relevant medical staff and equipment managers through multiple channels, including audible and visual alarms, SMS, telephone, and the hospital information system, ensuring timely handling of any abnormalities.

[0029] The knowledge base module 6 stores historical pipeline operation data, model parameters, control experience, and gas usage patterns of different departments, supporting decision optimization in the background analysis module 3. The knowledge base employs a hybrid architecture of time-series and relational databases. The time-series database stores massive amounts of historical monitoring data, while the relational database stores model parameters, configuration information, and control rules. In one embodiment of the invention, the knowledge base stores full data for the most recent 12 months and statistical data for the most recent 5 years, with a total data volume reaching TB levels, providing ample data support for big data analysis and machine learning.

[0030] like Figure 2As shown, the pipeline monitoring module 1 includes a ward monitoring unit 11, an operating room monitoring unit 12, a special department monitoring unit 13, and an environmental parameter monitoring unit 14.

[0031] The ward monitoring unit 11 is used to monitor the pressure, flow rate, and flow parameters of medical oxygen pipelines, medical air pipelines, and medical vacuum pipelines in the ward area. According to the "Technical Specification for Medical Gas Engineering" (GB 50751-2012), the terminal pressure of the ward oxygen pipeline should be 0.4-0.5 MPa, the terminal pressure of the medical air pipeline should be 0.4-0.7 MPa, and the terminal negative pressure of the medical vacuum pipeline should be -0.04 to -0.07 MPa. The ward monitoring unit 11 is equipped with pressure sensors, flow meters, and temperature sensors on the main pipelines of each ward floor, with a monitoring point spacing not exceeding 50m to ensure timely detection of pipeline leaks, pressure fluctuations, and other abnormalities.

[0032] In one embodiment of the invention, the ward monitoring unit 11 employs a high-precision digital pressure sensor. The sensor outputs a 4-20mA current signal or an RS485 digital signal, which is connected to the data acquisition unit via a shielded cable. For medical vacuum pipelines, since they operate under negative pressure, the sensor needs to be able to measure negative pressure, with a measurement range of -0.1 to 0 MPa and an accuracy of ±0.5%. An ultrasonic flow meter is preferred, as it does not require drilling holes in the pipeline for installation, does not affect the pipeline's sealing and cleanliness, and is particularly suitable for monitoring medical gas pipelines.

[0033] Operating room monitoring unit 12 is used to monitor the pressure, flow rate, and flow parameters of medical oxygen pipelines, medical air pipelines, medical nitrous oxide pipelines, medical vacuum pipelines, and anesthetic waste gas emission pipelines in the operating room area. The operating room is the place where medical gases are used most frequently and with the most stringent requirements, especially in general anesthesia surgery. As an inhaled anesthetic, the stability of the supply of medical nitrous oxide (laughing gas) directly affects surgical safety and anesthetic efficacy. According to the "Technical Specification for Medical Gas Engineering" (GB 50751-2012), the terminal pressure of the nitrous oxide pipeline in the operating room should be 0.4-0.5 MPa. The anesthetic waste gas emission system is used to remove anesthetic waste gases generated during surgery, preventing operating room personnel from being exposed to the anesthetic gas environment for extended periods. The emission pipeline should maintain a negative pressure of -0.005 to -0.01 MPa.

[0034] In one embodiment of the invention, the operating room monitoring unit 12 installs monitoring sensors near the gas terminal in each operating room to monitor in real time the supply pressure and flow rate of oxygen, medical air, and nitrous oxide, as well as the negative pressure and flow rate of vacuum suction and anesthetic waste gas emissions. Considering the special nature of the operating room environment, all sensors and cables are made of medical-grade materials, possessing good anti-interference capabilities and electromagnetic compatibility, meeting the requirements of the operating room's electromagnetic environment. For example, in cardiac surgery, the oxygen flow rate may reach 20-30 L / min, while in laparoscopic surgery, carbon dioxide is used to establish pneumoperitoneum, and the flow rate may reach 10-15 L / min; the system needs to be able to accurately monitor these changes.

[0035] Specialty Department Monitoring Unit 13 is used to monitor the pressure, flow rate, and flow parameters of medical nitrogen pipelines, medical carbon dioxide pipelines, instrument air pipelines, dental air pipelines, and dental vacuum pipelines in specialized departments such as dentistry, ICU, and central sterilization supply. Different departments have significantly different requirements for medical gases. For example, dentistry requires oil-free dental air to drive dental drills, typically at a pressure of 0.5-0.6 MPa, and also requires dental vacuum to remove saliva and debris from the oral cavity. The ICU requires medical carbon dioxide for blood gas analyzer calibration and certain special treatments. Although the central sterilization supply uses steam for high-pressure steam sterilization, steam is not considered a clinically usable medical gas. This invention primarily monitors its compressed air pipelines, which drive the pneumatic valves and conveyor systems of the sterilizer.

[0036] In one embodiment of the invention, the special department monitoring unit 13 is configured with corresponding sensors according to the actual gas usage of each department. For example, the dental treatment area is equipped with dental air pressure sensors and dental vacuum negative pressure sensors, and the number of monitoring points is determined according to the number of dental chairs, typically one monitoring point for every 4-6 dental chairs. The ICU is equipped with carbon dioxide pipeline pressure sensors, with monitoring points set on the main carbon dioxide supply pipeline and the branch pipelines of each bed group. Medical nitrogen is mainly used for the pneumatic drive and cryogenic preservation of certain medical equipment, and its pipeline pressure is typically 0.5-0.8 MPa.

[0037] The environmental parameter monitoring unit 14 is used to monitor the temperature and ambient air pressure parameters of each pipeline. Pipeline temperature has a significant impact on gas density and flow rate measurements. According to the ideal gas law, for every 1°C change in temperature, the gas volume changes by approximately 0.37%. Therefore, in order to accurately calculate gas flow rate and consumption, real-time temperature monitoring and compensation are necessary. Ambient air pressure also affects the absolute value measurement of pipeline terminal pressure, especially in high-altitude areas where atmospheric pressure is lower, requiring corresponding adjustments to the pipeline's set pressure.

[0038] In one embodiment of the present invention, the environmental parameter monitoring unit 14 installs temperature sensors and air pressure sensors in the pipeline equipment room on each floor. The temperature sensors are PT100 resistance temperature detectors with a measurement range of -50℃ to 150℃ and an accuracy of ±0.1℃. The air pressure sensors are silicon piezoresistive sensors with a measurement range of 80-110 kPa and an accuracy of ±0.1 kPa. These environmental parameter data, along with the pipeline parameter data, are transmitted to the data integration module 2 for subsequent data processing and analysis.

[0039] like Figure 3 As shown, the background analysis module 3 includes a state representation unit 31, a hierarchical judgment unit 32, and a control decision unit 33.

[0040] The state characterization unit 31 is used to construct a multi-dimensional pipeline topology state space, mapping integrated data into state points in the state space and forming a pipeline operating state change trajectory based on the time series. Traditional pipeline monitoring systems typically monitor individual parameters independently, making it difficult to comprehensively grasp the overall state of the pipeline system. The innovation of the state characterization unit 31 lies in mapping multiple parameters such as pipeline pressure, velocity, and flow rate into a unified multi-dimensional topology space, enabling the system to comprehensively evaluate the pipeline operating state and identify the correlation and change patterns between parameters.

[0041] The classification determination unit 32 is used to construct a nonlinear probability transition matrix, calculate a state evaluation vector based on the state point and the pipeline operating state change trajectory, and determine the pipeline operating level label by comparing the state evaluation vector with a dynamically adjusted threshold vector. Traditional fixed threshold classification methods cannot adapt to changes in gas demand in different locations and at different times. For example, the gas consumption patterns in operating rooms during peak operating hours and non-operating hours are drastically different, and a fixed threshold may lead to false alarms or missed alarms. The classification determination unit 32 captures the changing patterns of pipeline parameters through a nonlinear probability matrix, dynamically adjusts the classification criteria, and improves the accuracy and adaptability of the classification.

[0042] The control decision unit 33 is used to generate control strategies based on grade marking and chaotic prediction analysis of pipeline state change trends. These strategies are then converted into control commands and transmitted to the control module 4. The control decision unit 33 is characterized by its predictive capability, able to predict pipeline parameter change trends 5-30 minutes in advance. This allows for preventative control before abnormal pipeline pressure or flow occurs, avoiding failures. Compared to traditional passive response modes, this proactive predictive management model significantly improves the stability and reliability of the pipeline system.

[0043] like Figure 4 As shown, the state representation unit 31 includes a space construction subunit 311, a mapping subunit 312, a trajectory construction subunit 313, and a feature extraction subunit 314.

[0044] The spatial construction subunit 311 is used to define a multi-dimensional pipeline state space and divide the space into multiple interconnected sub-regions. In one embodiment of the invention, a six-dimensional state space is constructed for the ward area, including six dimensions: oxygen pipeline pressure, oxygen pipeline flow rate, medical air pipeline pressure, medical air pipeline flow rate, vacuum pipeline negative pressure, and vacuum pipeline flow rate. A ten-dimensional state space is constructed for the operating room area, including ten dimensions: oxygen pipeline pressure, oxygen pipeline flow rate, medical air pipeline pressure, medical air pipeline flow rate, nitrous oxide pipeline pressure, nitrous oxide pipeline flow rate, vacuum pipeline negative pressure, vacuum pipeline flow rate, exhaust gas emission pipeline negative pressure, and exhaust gas emission pipeline flow rate.

[0045] In a preferred embodiment of the present invention, taking the ten-dimensional state space of an operating room as an example, the state space is defined as follows: for:

[0046] ,

[0047] in, This refers to the oxygen pipeline pressure, expressed in MPa. This refers to the oxygen pipeline flow rate, expressed in L / min. This refers to the pressure in medical air ducts, expressed in MPa. This refers to the flow rate of medical air ducts, expressed in L / min. This refers to the pressure in the nitrous oxide pipeline, expressed in MPa. The flow rate of nitrous oxide in the pipeline is expressed in L / min. This refers to the negative pressure in the vacuum pipeline, expressed in MPa (negative value). This refers to the vacuum pipeline flow rate, expressed in L / min. The negative pressure in the anesthetic waste gas emission pipeline is expressed in MPa (negative value). The flow rate of the exhaust gas pipeline is expressed in L / min. The range of values ​​for each parameter is determined according to the provisions of the "Technical Specification for Medical Gas Engineering" (GB 50751-2012) and actual operating experience.

[0048] The spatial construction subunit 311 divides the state space into multiple sub-regions, each corresponding to a typical pipeline operating state. For example, in an operating room, an oxygen pipeline pressure of 0.45-0.5 MPa and a flow rate of 10-20 L / min corresponds to a normal surgical state; a pressure of 0.4-0.45 MPa and a flow rate of 20-30 L / min corresponds to a high-demand surgical state; and a pressure below 0.4 MPa corresponds to an insufficient gas supply state. This partitioning method, based on cluster analysis of a large amount of historical data, can identify different operating modes. In one embodiment of the invention, the K-means clustering algorithm is used to cluster historical data, dividing the state space into 20-50 sub-regions, the number of which is adaptively determined according to the data complexity.

[0049] Mapping subunit 312 is used to map standardized pipeline pressure, velocity, and flow rate data to a multi-dimensional pipeline state space using a mapping function. Since different parameters have different dimensions and numerical ranges, they need to be standardized first to map each parameter to the same numerical range, typically the 0-1 interval. The standardization formula is:

[0050] ,

[0051] in, These are standardized parameter values, dimensionless. These are the original measured values. and These represent the minimum and maximum measurement ranges for the parameter, respectively. For example, for oxygen pipeline pressure, if the measurement range is 0.3-0.6 MPa and the current measurement value is 0.45 MPa, then the standardized value is... This standardization process ensures that parameters with different dimensions can be compared and analyzed in the same space.

[0052] The trajectory construction subunit 313 is used to connect adjacent state points on the time series to form a trajectory of pipeline operation state changes. In one embodiment of the present invention, the trajectory... Defined as a set of state points on a time series:

[0053] ,

[0054] in, For a moment The status points contain the values ​​and timestamps of all monitored parameters. The number of state points is specified. To ensure the smoothness and continuity of the trajectory, the trajectory construction sub-unit 313 employs cubic spline interpolation to generate smooth transition curves between adjacent state points. Preferably, when the time interval between adjacent state points is large, such as exceeding 15 minutes, the system marks this segment of the trajectory as having low confidence and reduces its weight in subsequent analysis. This processing method can effectively address situations where monitoring data is missing or transmission is interrupted.

[0055] Feature extraction subunit 314 is used to calculate the geometric, dynamic, and stability features of the trajectory. Geometric features include trajectory length, curvature, and the number of inflection points, reflecting the shape characteristics of the trajectory. Dynamic features include velocity vectors and acceleration vectors, reflecting the rate and trend of parameter change. Stability features include deviation and fluctuation frequency, reflecting the stability of the system.

[0056] In a preferred embodiment of the present invention, the formula for calculating the trajectory velocity vector is:

[0057] ,

[0058] in, For a moment arrive The average velocity vector is a multi-dimensional vector that corresponds to the rate of change of multiple pipe parameters. and They are time points and The state points are all multi-dimensional vectors. and The velocity vector represents the time between two consecutive samples, expressed in seconds. Calculating the velocity vector helps identify rapid changes in pipeline parameters. For example, at the start of surgery, the nitrous oxide flow rate may rapidly increase from 0 to 15 L / min, significantly increasing the magnitude of the velocity vector, which the system uses to determine when surgery has begun.

[0059] The formula for calculating the trajectory acceleration vector is:

[0060] ,

[0061] in, For a moment arrive The average acceleration vector is a multi-dimensional vector, corresponding to the rate of change of multiple pipe parameters. and These are the average velocity vectors for adjacent time periods. Acceleration vectors reflect the accelerating or decelerating trend of pipeline parameter changes and are crucial for predicting the direction of parameter changes in the short term. For example, when the acceleration vector of oxygen pipeline pressure is negative and has a large absolute value, it indicates that the pressure drop rate is accelerating, which may foreshadow an impending gas supply shortage, requiring advance system adjustments.

[0062] Trajectory stability score The calculation formula is:

[0063] ,

[0064] in, The trajectory stability score is a dimensionless scalar value; the smaller the value, the more stable the trajectory. This represents the number of state points in the trajectory. The weighting coefficient is a dimensionless scalar value that is inversely proportional to the time interval, and is typically set to a value of [value missing]. , The magnitude of the velocity vector is denoted by . The stability score comprehensively considers the magnitudes of all velocity vectors in the trajectory; a lower stability score indicates smoother changes in pipeline parameters and more stable system operation. In medical practice, pipeline systems in ward areas typically require higher stability, with stability scores controlled below 0.3. However, in operating rooms, stability scores during surgery may reach 0.8-1.2, which is normal fluctuation in gas usage during surgery.

[0065] like Figure 5 As shown, the grading determination unit 32 includes a matrix construction subunit 321, a state evaluation subunit 322, a threshold adjustment subunit 323, and a grade determination subunit 324.

[0066] The matrix construction subunit 321 is used to construct short-term, medium-term, and long-term three-layer nonlinear probability transition matrices based on historical pipeline operation data. The probability transition matrix describes the probability of a pipeline state transitioning from one sub-region to another, and can capture the dynamic characteristics and changing patterns of pipeline operation. In one embodiment of the invention, the probability transition matrix... Defined as:

[0067] ,

[0068] in, From state Transition to state The probability of is the position of the element in the matrix at the th position. Line number The elements of the column take values ​​from 0 to 1. For a moment state, For a moment state, The time interval is represented by the "|" symbol, which indicates the conditional relationship in conditional probability. Probability transition matrix. It is a square matrix whose dimension is equal to the number of sub-regions in the state space, and the sum of the elements in each row is 1, satisfying the probability normalization condition.

[0069] Matrix Construction Subunit 321: Constructing a Three-Level Transition Matrix: Short-Term Matrix (5-minute time interval), Interim matrix (Time interval 1 hour) and long-term matrix (Time interval 24h). Matrixes at different time scales can capture change patterns in different cycles. Short-term matrices reflect rapid changes and sudden events, such as gas consumption changes caused by the start or end of surgery. Medium-term matrices reflect surgical scheduling and daytime work rhythms. Long-term matrices reflect diurnal variations and weekly work patterns. For example, the gas demand in the operating room during the daytime on weekdays is significantly higher than at night and on weekends. Long-term matrices can capture this periodic change, providing a basis for predictive regulation.

[0070] The condition assessment subunit 322 is used to calculate the pipeline operation stability score, trajectory development trend, potential failure probability, and environmental impact factors, generating a condition assessment vector. In one embodiment of the present invention, the condition assessment vector... Defined as:

[0071] ,

[0072] in, The state evaluation vector is a four-dimensional column vector. The pipeline's operational stability is rated, with values ​​ranging from 0 to 1. A smaller value indicates greater stability. The trajectory development trend is scored, with a value ranging from 0 to 1. The higher the value, the more obvious the trend. This represents the potential failure probability, ranging from 0 to 1. A higher value indicates a higher risk of failure. The environmental impact factors are scored, with values ​​ranging from 0 to 1. The higher the value, the more significant the environmental impact. This represents the transpose of a vector.

[0073] Pipeline operational stability rating Based on trajectory stability score Calculations typically involve normalization to ensure the value falls within the 0-1 range. Trajectory trend score. Based on the consistency calculation of velocity and acceleration vectors, when the directions of the velocity and acceleration vectors are consistent, it indicates a significant trend in parameter change and a higher score; conversely, a divergent direction results in a lower score. Potential Fault Probability Based on the calculation of the transition probability from the current state to the fault state, the probability transition matrix is ​​used to predict the likelihood of the system entering a fault state. Environmental impact factor scoring. The score increases when environmental parameters such as temperature and humidity deviate from the normal range, based on the degree of abnormality.

[0074] The threshold adjustment subunit 323 is used to dynamically adjust the threshold vector based on historical grading accuracy and control effect feedback. Traditional fixed thresholds may lead to oversensitivity or oversensitivity in different scenarios, while dynamic thresholds can adaptively adjust according to actual conditions, improving grading accuracy. In one embodiment of the present invention, the threshold vector... Defined as:

[0075] ,

[0076] in, The threshold vector is a three-dimensional column vector. , , The three thresholds for classifying the four levels (A, B, C, and D) are all scalar values, ranging from 0 to 1, and satisfy the following conditions: Preferably, the initial threshold is set to... , , Then, adjustments will be made dynamically based on the actual results.

[0077] Threshold adjustment employs gradient descent, adjusting the threshold based on the accuracy of historical classifications. Specifically, a loss function is defined. To account for the discrepancy between the grading results and the expert evaluation results, a smaller loss function indicates more accurate grading. The threshold update formula is:

[0078] ,

[0079] in, For the updated threshold, These are the threshold values ​​before the update, all of which are scalar values ​​ranging from 0 to 1. The learning rate, typically set to 0.01 to 0.05, is a scalar value that controls the step size for threshold adjustment. For the loss function with respect to the threshold The gradient is a scalar value representing the rate of change of the loss function at the current threshold point. Through iterative optimization, the threshold can gradually approach the optimal value, making the classification results more accurate and reliable.

[0080] The grading subunit 324 compares the state assessment vector with the threshold vector to determine the four-level grading labels from A to D. The grading rules are based on the comprehensive score of the state assessment vector, which is defined as the weighted average of the elements of the state assessment vector.

[0081] ,

[0082] in, The overall score is a scalar value, ranging from 0 to 1. The state evaluation vector is the first There are 1 element. The rule for determining the level is: when When the level is A, it indicates that the pipeline is operating normally and requires no adjustment; when At this time, the level is B, indicating a slight abnormality in pipeline operation, requiring minor adjustments; when At this time, the level is C, indicating a moderate abnormality in pipeline operation, requiring significant control; when At this time, the level is D, indicating that the pipeline operation is severely abnormal and requires emergency control.

[0083] In practical applications, different levels correspond to different processing strategies and response times. At Level A, the system performs only routine monitoring and data recording, requiring no manual intervention. At Level B, the system automatically performs fine-tuning, such as adjusting the opening of a pipeline regulating valve by 1%–2%, with a response time of 5 minutes. At Level C, the system performs significant regulation, such as adjusting pipeline pressure by 5%–10%, while simultaneously issuing a warning alert to notify relevant personnel, with a response time of 2 minutes. At Level D, the system performs emergency regulation, such as activating a backup gas source or booster equipment, while simultaneously issuing an emergency warning requiring immediate action from relevant personnel, with a response time of 30 seconds. This hierarchical management mechanism ensures both automated system operation and timely human intervention for critical issues.

[0084] like Figure 6 As shown, the control and decision-making unit 33 includes a chaos analysis subunit 331, an optimization subunit 332, and a decision-making subunit 333.

[0085] The chaos analysis subunit 331 is used to reconstruct the phase space of pipeline parameters based on the delayed coordinate method, extract chaotic feature quantities such as the Lyapunov exponent, and realize short-term prediction, medium-term prediction, and long-term trend analysis. Chaos theory suggests that seemingly random time series data may contain deterministic nonlinear laws. These hidden laws can be revealed through phase space reconstruction, thereby achieving prediction. In one embodiment of the present invention, the phase space reconstruction uses the delayed coordinate method to embed one-dimensional time series data into a multi-dimensional phase space.

[0086] For pipeline pressure time series Phase space reconstructed vector Defined as:

[0087] ,

[0088] in, The reconstructed phase space vector is a 3D column vector, For a moment The pipeline pressure is a scalar value. The time delay is a positive integer representing the time interval between adjacent delay coordinates, measured in sampling periods. The embedding dimension is a positive integer representing the dimension of the reconstructed phase space. Preferably, The value is typically chosen such that the autocorrelation function first decreases to [a value that is not explicitly stated in the original text]. For medical gas pipeline pressure data, the time delay value is empirically estimated to be 2 to 5 sampling cycles, i.e., a delay of 10-25 minutes. It is usually determined by the pseudo nearest neighbor method, and for pipeline systems, the empirical value is 3 to 5.

[0089] Chaos analysis subunit 331 calculates the maximum Lyapunov exponent The maximum Lyapunov exponent is used to evaluate the stability and predictability of a system. It is defined as follows:

[0090] ,

[0091] in, The maximum Lyapunov exponent is a scalar value, with units of 1000 kJ / m². , representing the exponential growth rate of the initial small perturbation in the system. For time, The initial distance is a scalar value representing the Euclidean distance between two initial points in phase space. For the time elapsed The distance after that is a scalar value. Let represent the natural logarithm function. When When the system exhibits chaotic characteristics, initial small perturbations grow exponentially, making long-term prediction difficult; when When the system is relatively stable, disturbances tend to converge, and prediction is relatively easy.

[0092] In medical gas pipeline systems, the pipeline operation in ward areas is usually relatively stable. A value close to or slightly less than 0 indicates good system predictability. The operating room area during surgery... A positive value indicates that the system exhibits chaotic characteristics. In this case, short-term predictions are still feasible, but long-term predictions are highly uncertain. The system is based on... The value adaptively adjusts the prediction time window when When the value is small, a medium-term forecast of 15-30 minutes can be made. When the value is large, only short-term forecasts of 5-10 minutes are made.

[0093] Short-term forecasting employs a local linear model, establishing a linear regression model based on neighboring state points of the current state to predict pipeline parameters for the next 5-15 minutes. Medium-term forecasting uses a radial basis function neural network, establishing a nonlinear forecasting model through training historical data to predict pipeline parameters for the next 15-60 minutes. Long-term trend analysis uses periodic decomposition and time series analysis methods to identify daily and weekly cycle patterns, predicting gas consumption trends for the next few hours to days.

[0094] The optimization subunit 332 is used to set a multi-objective function with safety, stability, responsiveness, and resource efficiency objectives, generating multiple candidate control schemes under the constraints of pipeline pressure safety range, equipment regulation capacity, and resource limitations. In one embodiment of the invention, the multi-objective function is defined as:

[0095] ,

[0096] in, It is a multi-objective function vector, which is a four-dimensional column vector. The control scheme includes parameters such as target pressure, target flow rate, and regulation rate for each gas pipeline. For security purposes, For the stability objective, For responsive objectives, These are all scalar values ​​representing resource efficiency targets; the smaller the value, the better the performance.

[0097] security goals The deviation of the adjusted pipeline parameters from the safe range is measured by the following formula:

[0098] ,

[0099] in, The security objective function value is a scalar value; the smaller the value, the higher the security. To monitor the number of pipelines, For the first The weighting coefficient for root canals is a scalar value, usually determined based on the importance of the canal; for example, oxygen canals in operating rooms have the highest weighting. For the first The current pressure in the root pipe is a scalar value, measured in MPa. For the first The target pressure of the root conduit is a scalar value, measured in MPa. For the first The permissible deviation of the root pipe is a scalar value, with the unit being MPa. Indicates taking 0 and The larger value in the objective function. When the pipeline pressure deviates from the target value by more than the allowable deviation, the safety objective function value increases, and the control optimizer will prioritize the option that can control the pressure within the safe range.

[0100] Stability target The formula for measuring the fluctuation of pipeline parameters after regulation is as follows:

[0101] ,

[0102] in, The stability objective function value is a scalar value; the smaller the value, the higher the stability. For the first The variance of the root canal pressure over a recent period is a scalar value, with units of . A smaller variance indicates less pressure fluctuation in the pipeline and a more stable system. The control optimizer, while ensuring safety, will select the control scheme that reduces pressure fluctuations.

[0103] responsive objectives The response time of the control plan is measured by the following formula:

[0104] ,

[0105] in, The objective function value for responsiveness is a scalar value measured in seconds (s). A smaller value indicates higher responsiveness. For the first The response time of a pipeline is a scalar value measured in seconds (s). It represents the time required from issuing a control command to the pipeline parameters reaching the target value. This represents the maximum response time across all pipelines. A shorter response time indicates that the system can quickly respond to changes in gas demand, which is especially important in places such as operating rooms and emergency rooms.

[0106] Resource efficiency goals The resource consumption of the control plan, including gas consumption and electricity consumption, is calculated using the following formula:

[0107] ,

[0108] in, The resource efficiency objective function value is a scalar value; the smaller the value, the higher the resource efficiency. For the first The resource consumption rate of a root pipeline is a scalar value; for a gas supply pipeline, it is the gas consumption per unit time (unit: gas). For vacuum systems, this refers to the vacuum pump power (in kW). Optimizing resource efficiency targets helps reduce hospital operating costs while also meeting environmental protection requirements for energy conservation and emission reduction.

[0109] The constraints include pipeline pressure safety range constraints, equipment regulation capacity constraints, and resource limitation constraints. The pipeline pressure safety range constraint is as follows:

[0110] ,

[0111] in, and The first The minimum and maximum safe pressures of the root pipe are scalar values, expressed in MPa. For the first The pressure of a pipeline is a scalar value, measured in MPa. These safety ranges are determined according to the "Technical Specification for Medical Gas Engineering" (GB 50751-2012). For example, the pressure range for oxygen pipelines in wards is 0.4-0.5 MPa, the pressure range for oxygen pipelines in operating rooms is 0.4-0.5 MPa, the pressure range for medical air pipelines is 0.4-0.7 MPa, and the range for medical vacuum negative pressure is -0.04 to -0.07 MPa.

[0112] The equipment's adjustability is constrained as follows:

[0113] ,

[0114] in, For the first The rate of change of pressure in the root conduit is a scalar value, with units of... , The maximum adjustment rate is a scalar value, with units of . , indicating the first The maximum permissible rate of change of pipeline pressure. This constraint reflects the physical limitations of pipeline regulating valves and pressurization equipment; excessively rapid regulation may exceed equipment capacity or cause pipeline system oscillations. Preferably, the pipeline pressure regulation rate is controlled within 0.01-0.05. It can respond quickly to changes in demand while maintaining system stability.

[0115] Resource constraints are:

[0116] ,

[0117] in, For the first Root canal resource consumption rate The maximum resource consumption limit is a scalar value representing the upper limit of the total supply capacity of the hospital's gas supply system and vacuum system. For example, the oxygen supply capacity of the hospital's central oxygen supply station is 100. The vacuum pump station has a pumping capacity of 200. The regulatory measures must ensure that total demand does not exceed these capacity limits.

[0118] The optimization subunit 332 employs the Non-Dominated Sorting Genetic Algorithm (NSGA-II) to solve a multi-objective optimization problem, generating a set of Pareto optimal solutions, each representing a candidate control scheme. A Pareto optimal solution is one in which no other solution is superior or equal to it in all objectives, and is strictly superior to it in at least one objective. By generating multiple candidate schemes, the system can weigh different objectives, such as prioritizing safety and responsiveness in emergency situations, and prioritizing stability and resource efficiency under normal conditions.

[0119] Decision subunit 333 is used to evaluate each candidate scheme, select the Pareto optimal control scheme, and decompose the scheme into specific control instructions. In one embodiment of the present invention, the ideal point method is used to evaluate candidate schemes. First, the ideal point is determined, that is, a virtual point that takes the optimal value on all objectives. Then, the distance from each candidate scheme to the ideal point is calculated, and the scheme with the smallest distance is selected as the optimal scheme. The distance calculation formula is:

[0120] ,

[0121] in, For the first The distance from each candidate solution to the ideal point is a scalar value. For the first One candidate solution, For the first The candidate solution is in the... Function values ​​on each target For the first The ideal value for a given objective is the minimum value of all candidate solutions for that objective. (Selection) The smallest possible solution is considered the optimal control solution.

[0122] The selected optimal solution includes parameters such as target pressure, target flow rate, and regulation rate for each gas pipeline. The decision subunit 333 decomposes these parameters into specific control commands and sends them to the control module 4. The control commands include information such as target gas type, target department / area, target pressure value, regulation rate, and expected response time, and are transmitted in JSON format to ensure the accuracy and traceability of the commands.

[0123] like Figure 7As shown, the control module 4 includes an oxygen pipeline control submodule 41, a medical air control submodule 42, a nitrous oxide control submodule 43, a vacuum system control submodule 44, and a collaborative control submodule 45.

[0124] The oxygen pipeline control submodule 41 receives control commands related to the medical oxygen pipeline and controls the oxygen pipeline regulating device. Medical oxygen is the most widely used medical gas in hospitals, and almost all departments require an oxygen supply. The oxygen pipeline control submodule 41 controls the oxygen pressure reducing valve, the electric regulating valve, and the pressurization device to regulate the pipeline pressure and flow rate. In one embodiment of the invention, the oxygen pressure reducing valve reduces the high-pressure oxygen (typically 10-15 MPa) from the central oxygen supply station to the operating pressure (0.4-0.5 MPa), the electric regulating valve fine-tunes the pressure according to the gas demand, and the pressurization device provides additional pressure when the oxygen supply capacity is insufficient.

[0125] Preferably, the oxygen pipeline regulating valve is an electrically operated proportional regulating valve, whose opening degree can be continuously adjusted within the range of 0% to 100%, with a control accuracy of ±1% and a response time of less than 5 seconds. The regulating valve receives control commands via a 4-20mA current signal or Modbus communication protocol, and achieves precise control through a valve positioner. In the ward area, one zone regulating valve is installed on each floor to control the oxygen supply pressure of that floor. In the operating room area, each operating room is equipped with an independent regulating valve to ensure that the gas usage of each operating room does not interfere with each other.

[0126] The medical air control submodule 42 receives control commands related to the medical air duct and controls the duct's regulating devices. Medical air is clean air that has undergone oil, water, and bacteria removal treatment and is used in medical equipment such as ventilators and anesthesia machines. Medical air has high quality requirements, with an oxygen content of 20%–23%, a dew point temperature below -40℃, and CO and CO2 content meeting medical gas standards. The medical air control submodule 42 controls the air compressor, dryer, and regulating valves to adjust duct pressure and flow.

[0127] In one embodiment of the present invention, the medical air duct system includes an air compressor, a refrigerated dryer, an adsorption dryer, a multi-stage filter, and a duct regulating valve. The air compressor compresses atmospheric pressure to 0.7-0.8 MPa, the dryer removes moisture from the compressed air, the filter removes oil mist, particles, and microorganisms, and the regulating valve adjusts the pressure to the operating pressure (0.4-0.7 MPa). The medical air control submodule 42 controls the start / stop and load adjustment of the air compressor according to air demand, and controls the opening degree of the regulating valve to ensure the stability and quality of the medical air supply.

[0128] The nitrous oxide control submodule 43 receives control commands related to the medical nitrous oxide pipeline and controls the pipeline regulating device. Medical nitrous oxide (commonly known as laughing gas) is an inhaled anesthetic, mainly used for anesthesia in operating rooms. Nitrous oxide has analgesic effects and is often mixed with oxygen, typically in a ratio of 50%–70% laughing gas and 30%–50% oxygen. The nitrous oxide control submodule 43 controls the nitrous oxide pressure reducing valve and regulating valve to adjust the pipeline pressure and flow rate.

[0129] In one embodiment of the invention, the nitrous oxide supply system employs a liquid nitrous oxide vaporization supply method. The liquid nitrous oxide is stored in a cryogenic insulated storage tank, vaporized into a gaseous state via a vaporizer, and then depressurized to the operating pressure (0.4-0.5 MPa) before being supplied to the operating room. The nitrous oxide control submodule 43 monitors the tank level, vaporizer temperature, and pipeline pressure, adjusting the vaporization rate and pipeline pressure according to the operating room's gas demand. Since nitrous oxide is used only in the operating room, its control is relatively independent and does not affect other departments.

[0130] It is important to note that nitrous oxide has certain anesthetic effects and poses an environmental pollution risk. Therefore, operating rooms must be equipped with an anesthetic waste gas emission system to promptly remove anesthetic gases that have not been absorbed by patients and prevent long-term exposure of operating room personnel. The nitrous oxide regulation submodule 43 works in conjunction with the anesthetic waste gas emission system. When an abnormally high concentration of nitrous oxide is detected in the operating room, it automatically increases the waste gas emission flow rate to ensure a safe operating room environment.

[0131] The vacuum system control submodule 44 receives control commands related to the medical vacuum pipeline and controls the vacuum pump and regulating valve. The medical vacuum system is used for negative pressure suction, removing blood and body fluids during surgery, suctioning and draining sputum in wards, and removing saliva and debris in dentistry. A medical vacuum system typically consists of a vacuum pump station, a vacuum tank, and a pipeline system. The vacuum pump station generates negative pressure, the vacuum tank acts as a buffer, and the pipeline system delivers the negative pressure to each user terminal.

[0132] In one embodiment of the present invention, the vacuum pump station is equipped with multiple water ring vacuum pumps, employing frequency conversion control to automatically adjust the number of operating pumps and their speed according to gas demand. When the pipeline negative pressure is higher than a set value (e.g., -0.05MPa), the system automatically reduces the number of operating pumps or lowers their speed; when the pipeline negative pressure is lower than a set value (e.g., -0.06MPa), the system automatically increases the number of operating pumps or raises their speed. The vacuum system control submodule 44 monitors the pipeline negative pressure, vacuum tank pressure, and vacuum pump operating status, and controls the start / stop and speed of the vacuum pumps according to control commands, adjusting the opening of the pipeline regulating valve to ensure the stability of the vacuum supply.

[0133] The collaborative control submodule 45 coordinates the operation of various gas pipelines to ensure the overall balance of the multi-gas pipeline system. In medical gas pipeline systems, different gas pipelines can interact with each other. For example, when multiple operating rooms use gas simultaneously, the pressure in the main pipeline may drop, affecting the gas supply to other departments. The collaborative control submodule 45 monitors the operating status of the entire pipeline system, coordinates the actions of each gas control submodule, and avoids control conflicts and mutual interference.

[0134] In one embodiment of the present invention, the collaborative control submodule 45 employs a priority scheduling strategy to determine the control priority based on the importance of the department and the urgency of gas usage. The priority order is: operating room > ICU > emergency room > ward > other departments. When system resources are insufficient to simultaneously meet all needs, priority is given to ensuring the gas supply to high-priority departments. For example, when the oxygen supply capacity of the central oxygen supply station is close to its limit, the system will prioritize ensuring the oxygen supply to the operating room and ICU, appropriately reduce the gas supply pressure in the wards, and issue an early warning notification to relevant personnel.

[0135] The collaborative control submodule 45 is also responsible for fault switching and backup system startup. When the main gas supply system fails, it automatically switches to the backup system, such as switching from the liquid oxygen station to the oxygen generator, or from the main vacuum pump station to the backup vacuum pump station, ensuring continuous gas supply. The fault switching process is completed within 30 seconds, during which gas supply is maintained by pipeline buffering and vacuum tank buffering, with almost no impact on medical use.

[0136] The early warning module 5 includes a real-time monitoring unit, an anomaly identification unit, a graded early warning unit, and an information transmission unit.

[0137] The real-time monitoring unit continuously monitors pipeline operating parameters, equipment operating status, and control effects. It receives data from pipeline monitoring module 1 and control module 4, including parameters such as pipeline pressure, flow velocity, flow rate, and temperature, as well as equipment status parameters such as regulating valve opening, vacuum pump speed, and compressor load. The monitoring frequency is 1 second per instance to ensure timely detection of anomalies.

[0138] The anomaly detection unit identifies abnormalities in pipeline pressure, flow rate, and output, as well as equipment malfunctions and deviations in control performance. Anomaly detection employs multiple algorithms, including threshold judgment, trend analysis, and pattern recognition. Threshold judgment detects whether parameters exceed normal ranges, such as oxygen pipeline pressure below 0.35 MPa or above 0.55 MPa. Trend analysis detects changing trends in parameters; for example, a sustained pressure drop may indicate pipeline leakage or insufficient gas supply. Pattern recognition detects abnormal patterns; for example, periodic pressure fluctuations may indicate regulating valve oscillation or vacuum pump malfunction.

[0139] The graded early warning unit generates three levels of warnings based on the severity of the anomaly: alert (yellow), warning (orange), and emergency (red). An alert-level warning indicates a minor anomaly that the system can handle automatically; relevant personnel are advised to pay attention, but immediate intervention is not required, such as an oxygen pipeline pressure deviation of 2%–5% from the target value. A warning-level warning indicates a moderate anomaly requiring manual inspection and handling, such as an oxygen pipeline pressure deviation of 5%–10% from the target value or an abnormal increase in flow rate exceeding 30%. An emergency-level warning indicates a severe anomaly requiring immediate action, such as an oxygen pipeline pressure below 0.3 MPa or a complete failure of the vacuum system.

[0140] The information transmission unit is used to send early warning information through multiple channels, including audible and visual alarms, SMS, telephone, and the hospital information system. Audible and visual alarms are installed in the duty room and equipment room, emitting sound and light signals to attract the attention of on-duty personnel. SMS and telephone messages are sent to equipment management and maintenance personnel to ensure timely response to abnormal situations. The hospital information system integrates early warning information and displays it on the monitoring screen and handheld terminals, facilitating management personnel's overall understanding of the system's operational status.

[0141] In one embodiment of the present invention, the early warning module 5 is integrated with the hospital logistics management system and the equipment maintenance system to realize automatic dispatching and closed-loop management of early warning information. When a warning-level or emergency-level early warning is issued, the system automatically generates a maintenance work order and assigns it to the corresponding maintenance personnel. The maintenance personnel receive the work order through a handheld terminal, scan the equipment's QR code to confirm upon arrival at the site, and record the processing result in the system after completion. The entire process achieves closed-loop management and traceability.

[0142] The knowledge base module 6 includes a historical data storage unit, a model parameter storage unit, a gas consumption pattern storage unit, and an experience learning unit.

[0143] The historical data storage unit stores historical pipeline operation data, including time-series data of parameters such as pressure, flow rate, flow volume, and temperature, as well as early warning records, control records, and maintenance records. Historical data is stored using a time-series database, supporting efficient time-range querying and statistical analysis. In one embodiment of the invention, the InfluxDB time-series database is used, with a data retention strategy of full data for the most recent 12 months and downsampled data for the most recent 5 years. The downsampling period is 5 minutes, ensuring data integrity while controlling storage space.

[0144] The model parameter storage unit stores parameters of the analysis model, such as the probability transition matrix, threshold vector, and prediction model parameters. These parameters are updated periodically to reflect changes in the system's operating characteristics. For example, the probability transition matrix is ​​recalculated weekly based on the data from the most recent month, the threshold vector is fine-tuned daily based on the day's classification accuracy, and the prediction model parameters are retrained monthly based on the prediction error.

[0145] The gas usage pattern storage unit stores gas usage patterns for different departments and time periods, including daily gas usage curves, weekly gas usage curves, and gas usage patterns for special events. For example, the daily gas usage curve for the operating room shows that the peak gas usage period is from 8:00 to 18:00 on weekdays, with oxygen flow rates reaching 100-150 L / min, while the low gas usage periods are at night and on weekends, with oxygen flow rates as low as 20-30 L / min. Gas usage patterns in wards are relatively stable, but there are still some diurnal variations. Gas usage patterns in the ICU and emergency room are more random, mainly depending on the patient's condition.

[0146] The experience learning unit is used to learn optimization experience from historical data, continuously improving the analysis model and control strategy. Experience learning employs a reinforcement learning algorithm, treating control decisions as reinforcement learning actions, pipeline operating status as environmental states, and system performance indicators as reward signals. Through continuous trial and error and experience accumulation, it learns the optimal control strategy. In one embodiment of this invention, a deep Q-network (DQN) algorithm is used, performing offline training weekly based on data from the most recent month to update the control strategy network parameters, thereby gradually improving system performance.

[0147] The configuration of gas pipelines varies significantly in different locations. This invention provides differentiated monitoring and control based on the gas usage characteristics of different locations.

[0148] The ward area is primarily equipped with medical oxygen pipelines, medical air pipelines, and medical vacuum pipelines. Oxygen is used for oxygen therapy to patients, with a flow rate typically of 2-5 L / min and a pressure of 0.4-0.5 MPa. Medical air is used to power certain medical devices, such as nebulizers, with a lower flow rate. Medical vacuum is used for sputum suction and drainage, with a negative pressure of -0.04 to -0.07 MPa. Gas supply in the ward area is relatively stable; monitoring focuses on ensuring continuous gas supply and pressure stability.

[0149] The operating room area is equipped with medical oxygen pipelines, medical air pipelines, medical nitrous oxide pipelines, medical vacuum pipelines, and anesthetic waste gas exhaust pipelines. The operating room is the most complex location for medical gas usage. Oxygen and nitrous oxide are used for anesthesia, with flow rates adjusted according to the type of surgery and depth of anesthesia, typically 10-30 L / min. Medical air is used to power the anesthesia machine and ventilator. Medical vacuum is used to remove blood and body fluids during surgery, with a relatively high flow rate. The anesthetic waste gas exhaust system removes anesthetic gases that have not been absorbed by the patient, with a negative pressure of -0.005 to -0.01 MPa and a flow rate matched to the nitrous oxide supply flow rate. Gas usage in the operating room area fluctuates significantly; monitoring focuses on ensuring rapid gas supply response and coordinated control of multiple gases.

[0150] The ICU area is equipped with medical oxygen tubing, medical air tubing, medical carbon dioxide tubing, and medical vacuum tubing. ICU patients are critically ill, with high and rapidly changing oxygen demands. Oxygen flow rates are typically 10-20 L / min, but some patients may require higher flow rates of 50 L / min or more. Medical air is used to power ventilators. Medical carbon dioxide is used for blood gas analyzer calibration and certain specialized treatments such as carbon dioxide pneumoperitoneum. Medical vacuum is used for airway suction and drainage. Monitoring in the ICU area focuses on ensuring absolutely reliable gas supply and rapid response to emergency needs.

[0151] The dental area is equipped with dental air ducts and dental-specific vacuum ducts. Dental air is clean compressed air that has undergone oil, water, and bacteria removal treatment, with an oil content of less than 0.1 mg / m³. It is used to drive dental handpieces (dental drills) at a pressure of 0.5-0.6 MPa and a flow rate of 50-80 L / min. The dental-specific vacuum is used to remove saliva and debris from the oral cavity, with a negative pressure of -0.065 to -0.08 MPa and a flow rate of 100-150 L / min. Air use in the dental area is intermittent and sudden; therefore, monitoring focuses on ensuring stable pressure and noise control.

[0152] like Figure 8 As shown, the hospital medical gas pipeline monitoring and control method of the present invention includes the following steps:

[0153] S1: The pipeline monitoring module 1 monitors the pressure, velocity, and flow rates of medical gas pipelines in different locations to generate a pipeline operation index. In this step, the ward monitoring unit 11 monitors pipelines in the ward area, the operating room monitoring unit 12 monitors pipelines in the operating room area, the special department monitoring unit 13 monitors pipelines in other departments, and the environmental parameter monitoring unit 14 monitors temperature and air pressure. The sampling frequency is normally 5 minutes / time, and is increased to 10 seconds / time in abnormal situations.

[0154] S2: The pipeline operation index is received by data integration module 2, and integrated into consolidated data after time synchronization and quality assessment. In this step, data integration module 2 uses the NTP protocol for time synchronization, the Kalman filter algorithm for data smoothing, and the 3x standard deviation method to identify outlier data points, ensuring the accuracy and reliability of the data.

[0155] S3: Perform the following operations through the background analysis module 3. First, the state characterization unit 31 maps the integrated data to a multi-dimensional pipeline state space, forming state points and pipeline operating state change trajectories, and extracts the geometric, dynamic, and stability features of the trajectories. Second, the classification and determination unit 32 calculates the state evaluation vector based on the nonlinear probability transition matrix, and determines the pipeline operating level label from A to D by comparing it with the dynamic threshold vector. Finally, the control and decision-making unit 33, based on the level label and combined with chaotic prediction analysis of the pipeline state change trend, generates a control strategy through multi-objective dynamic optimization. The control strategy includes the target pressure, regulation rate, and expected response time for each gas pipeline.

[0156] S4: The control module 4 receives the control strategy, converts it into specific control instructions, and controls the corresponding medical gas pipeline regulating device according to the level label. In this step, no control is required for level A, fine-tuning is required for level B, significant control is required for level C, and emergency control and early warning are issued for level D. All gas control submodules work collaboratively to ensure control effectiveness and system stability.

[0157] S5: Continuously monitor the control effect and feed the results back to the background analysis module 3 to optimize subsequent control decisions. In this step, the system compares the actual control effect with the expected target, calculates the deviation and analyzes the reasons, and updates the model parameters and control strategy. Through continuous learning and optimization, the system performance is continuously improved.

[0158] In one specific embodiment of the present invention, a tertiary hospital deployed the medical gas pipeline monitoring and control system of the present invention. The monitoring points covered 15 ward floors, 20 operating rooms, 3 ICUs, and 2 dental treatment areas, totaling over 300 monitoring points. The monitored pipelines included seven types of gases: medical oxygen, medical air, medical nitrous oxide, medical vacuum, and anesthetic waste gas emissions. Since the system began operating six months ago, the pipeline pressure control accuracy has improved from ±8% to ±3%, the number of abnormal pipeline pressure alarms has decreased from an average of 30 times per month to an average of 5 times per month, the equipment failure response time has shortened from an average of 2 hours to an average of 30 minutes, medical gas consumption has decreased by approximately 15%, and annual operating cost savings are approximately 500,000 yuan. The system's predictive control function effectively avoided 12 potential gas supply interruption accidents, ensuring medical safety and receiving high praise from hospital management and clinical departments.

[0159] This invention innovatively integrates technologies such as multi-dimensional pipeline topology state-space representation, nonlinear probabilistic hierarchies, and chaotic prediction and control to construct a comprehensive, intelligent, and safe medical gas pipeline management solution. The system achieves a technological leap from passive response to active prediction, from static thresholds to dynamic hierarchies, and from single-parameter to multi-parameter integrated monitoring. It provides differentiated monitoring and control based on the gas usage characteristics of different locations, effectively improving the safety, stability, and resource utilization efficiency of medical gas pipeline systems, and promoting the development of medical gas management towards intelligence and precision.

[0160] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A hospital medical gas pipeline monitoring and control system based on big data intelligent judgment, characterized in that, include: The pipeline monitoring module is used to monitor the pressure, velocity, and flow rate data of medical gas pipelines in different locations and generate a pipeline operation index. The locations include ward areas and operating room areas. The monitored gases in the ward area include medical oxygen, medical air, and medical vacuum. The monitored gases in the operating room area include medical oxygen, medical air, medical nitrous oxide, medical vacuum, and anesthetic waste gas emissions. The data integration module is electrically connected to the pipeline monitoring module and is used to receive the pipeline operation index sent by the pipeline monitoring module, and integrate the received pipeline operation index into integrated data after time synchronization and quality assessment. The background analysis module is electrically connected to the data integration module. It is used to receive the integrated data, map the integrated data to a multi-dimensional pipeline state space based on the topology space mapping mechanism, and perform adaptive classification judgment through a nonlinear probability matrix to generate pipeline operation level labels. The level labels include normal operation level A, mild abnormality level B, moderate abnormality level C, and severe abnormality level D. The control module is electrically connected to the background analysis module and is used to receive the level markers and control the adjustment devices of each medical gas pipeline based on the level markers through chaotic prediction and dynamic optimization. The pipeline monitoring module includes: The ward monitoring unit is used to monitor the pressure, flow rate, and flow parameters of medical oxygen pipelines, medical air pipelines, and medical vacuum pipelines in the ward area. The operating room monitoring unit is used to monitor the pressure, flow rate, and flow parameters of medical oxygen pipelines, medical air pipelines, medical nitrous oxide pipelines, medical vacuum pipelines, medical carbon dioxide pipelines, and anesthetic waste gas emission pipelines in the operating room area. Specialized department monitoring units are used to monitor the pressure, flow rate, and flow parameters of medical nitrogen pipelines, instrument air pipelines, dental air pipelines, and dental-specific vacuum pipelines in dental clinics, ICUs, and sterilization supply centers. The environmental parameter monitoring unit is used to monitor the temperature and ambient air pressure parameters of each pipeline; The background analysis module includes: A state characterization unit is used to construct a multi-dimensional pipeline topology state space, map the integrated data to state points in the state space, and form a pipeline operation state change trajectory based on the time series. The classification and determination unit is used to construct a nonlinear probability transition matrix, calculate a state evaluation vector based on the state point and the pipeline operation state change trajectory, and determine the pipeline operation level label by comparing the state evaluation vector with a dynamically adjusted threshold vector. The control decision unit is used to analyze the pipeline state change trend based on the level label and in combination with the chaotic prediction model, generate a control strategy through multi-objective dynamic optimization, and convert the control strategy into a control command and transmit it to the control module. The control module includes: The oxygen pipeline control submodule is used to receive control commands related to medical oxygen pipelines and control the oxygen pipeline regulating device. The medical air control submodule is used to receive control commands related to medical air ducts and control the medical air duct control devices. The nitrous oxide regulation submodule is used to receive regulation commands related to the medical nitrous oxide pipeline and control the nitrous oxide pipeline regulating device. The vacuum system control submodule is used to receive control commands related to medical vacuum pipelines and control the vacuum pump and regulating valves. The collaborative control submodule is used to coordinate the control operations of various gas pipelines to ensure the overall balance of the multi-gas pipeline system.

2. The system according to claim 1, characterized in that, The state representation unit includes: A spatial construction subunit is used to define a multidimensional pipeline state space and divide the space into multiple interconnected sub-regions; The mapping subunit is used to map the standardized pipeline pressure data, velocity data, and flow rate data to the multidimensional pipeline state space through a mapping function. The trajectory construction sub-unit is used to connect adjacent state points on the time series to form the trajectory of pipeline operation state changes. The feature extraction subunit is used to calculate the geometric, dynamic, and stability features of the trajectory.

3. The system according to claim 1, characterized in that, The hierarchical determination unit includes: The matrix construction subunit is used to construct short-term, medium-term, and long-term three-level nonlinear probability transition matrices based on historical pipeline operation data. The condition assessment subunit is used to calculate the pipeline's operational stability score, trajectory development trend, potential failure probability, and environmental impact factors, and generate a condition assessment vector. The threshold adjustment subunit is used to dynamically adjust the threshold vector based on historical grading accuracy and control effect feedback. The level determination subunit is used to compare the state evaluation vector with the threshold vector to determine the four-level level labels from level A to level D.

4. The system according to claim 1, characterized in that, The control decision-making unit includes: The chaos analysis subunit is used to reconstruct the phase space of pipeline parameters based on the delayed coordinate method, extract Lyapunov exponential chaotic features, and realize short-term prediction, medium-term prediction and long-term trend analysis. The optimization sub-unit is used to set a multi-objective function with safety, stability, responsiveness and resource efficiency objectives, and generates multiple candidate control schemes under the constraints of pipeline pressure safety range, equipment regulation capacity and resource limitations. The decision-making subunit is used to evaluate each candidate scheme, select the Pareto optimal control scheme, and decompose the scheme into specific control instructions.

5. The system according to claim 1, characterized in that, The system also includes: The early warning module is electrically connected to the background analysis module and the control module, and is used to generate a graded early warning signal when abnormal pipeline pressure, abnormal flow rate, abnormal flow volume, equipment failure or control effect deviation is detected. The knowledge base module, electrically connected to the background analysis module, is used to store historical pipeline operation data, model parameters, control experience, and gas usage patterns of different departments, supporting the decision optimization of the background analysis module.

6. The system according to claim 1, characterized in that, The mapping relationship between the level markers and the control modules is as follows: Class A markings indicate normal operation status, requiring only routine inspections and monitoring. The Class B marker corresponds to a mild abnormal state and is transmitted to the single gas pipeline control submodule for fine-tuning. The C-level marker corresponds to a moderate abnormal state and is transmitted to multiple gas pipeline control submodules for coordinated control. A Level D marker corresponds to a severe abnormal state, and the information is transmitted to all relevant gas pipeline control submodules and early warning modules for emergency control and early warning.

7. A method for monitoring and controlling hospital medical gas pipelines based on big data intelligent judgment, applied to the system described in any one of claims 1 to 6, characterized in that, Includes the following steps: S1: The pipeline monitoring module monitors the pressure, velocity, and flow rates of medical gas pipelines in different locations to generate a pipeline operation index. S2: The pipeline operation index is received through the data integration module, and integrated into integrated data after time synchronization and quality assessment; S3: Perform the following operations through the background analysis module: map the integrated data to a multi-dimensional pipeline state space to form state points and pipeline operating state change trajectories; calculate the state evaluation vector based on the nonlinear probability transition matrix, and determine the level label by comparing it with the dynamic threshold vector; Based on the aforementioned grade markings, and combined with chaotic prediction analysis of pipeline state change trends, a control strategy is generated through multi-objective dynamic optimization. S4: Receive the control strategy through the control module, convert it into specific control instructions, and control the corresponding medical gas pipeline adjustment device according to the level marker; S5: Continuously monitor the control effect and feed the results back to the background analysis module to optimize subsequent control decisions.

Images