Plateau oxygen production and health monitoring system management platform

Abstract

A plateau oxygen production and health monitoring system management platform provided by the present invention comprises a data management module, an equipment management module, a user management module and a database, the data management module is used for storing, managing and analyzing system data, and the data management module carries out multi-source fusion on health data of a user and environmental data, real-time analysis is realized through big data analysis and an artificial intelligence algorithm to obtain a comprehensive index of the healthy environment, and whether a result is transmitted to the equipment management module or not is determined according to the index; the equipment management module is used for managing the operation state and parameter setting of the equipment, realizing data interaction and cooperative work between the oxygen generator and the health monitoring equipment according to the analysis result of the data management module, and automatically adjusting the operation parameters of the oxygen generator according to the real-time health state of the user; the user management module is used for managing user permission and login authentication of the system; the database comprises various information tables and alarm record tables.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental control, and particularly relates to a management platform for a plateau oxygen generation and health monitoring system. Background Art

[0002] Due to the high altitude, thin air and low oxygen content in the plateau area, long-term exposure to a hypoxic environment will have an adverse impact on human health, especially on the student group in the growth and development stage. The hypoxic environment may cause students to have symptoms such as dizziness, fatigue, and memory decline. In severe cases, it may even affect the physical development and learning efficiency of students. With the increase in plateau tourism, engineering construction, and scientific research activities, the demand for plateau oxygen generation technology and health monitoring is becoming increasingly urgent. Existing plateau oxygen generation equipment usually operates independently, lacks systematic management, and has insufficient integration with health monitoring data, making it impossible to achieve personalized oxygen supply and real-time health intervention. Currently, schools in plateau areas usually adopt traditional ventilation equipment or simple oxygen cylinder oxygen supply methods, but these methods have many deficiencies. First, traditional ventilation equipment cannot effectively increase the oxygen concentration in the classroom. Especially in the plateau area, the oxygen content in the air itself is low, and it is difficult to meet the oxygen demand of students simply by ventilation. Second, although the oxygen cylinder oxygen supply method can provide oxygen, there are problems such as unstable oxygen supply, inconvenient management, and great potential safety hazards. In addition, the existing oxygen supply system lacks the function of real-time monitoring of students' health conditions and cannot dynamically adjust the oxygen supply amount according to the physiological state of students. Summary of the Invention

[0003] The technical problem solved by the present invention is to provide a management platform for a plateau oxygen generation and health monitoring system, which solves the problem that the oxygen generation equipment and health monitoring equipment in the prior art are not deeply integrated by integrating Internet of Things, big data analysis, and artificial intelligence technologies.

[0004] The technical solution adopted by the present invention is as follows: A management platform for a plateau oxygen generation and health monitoring system. The plateau oxygen generation and health monitoring system includes an oxygen generator, sensors, solenoid valves, a blood oxygen concentration detector, a data transmission system, a control system, and a management platform. The oxygen generator provides oxygen for school classrooms, the sensors continuously monitor various parameters in the classrooms, the data transmission system transmits the sensor data to the management platform, the management platform adjusts the control system according to the sensor data, and the blood oxygen concentration detector continuously monitors the blood oxygen concentration of students and transmits the data to the management platform to judge the physical condition of students through big data analysis; the management platform includes a data management module, an equipment management module, a user management module, and a database. The data management module is used to store, manage, and analyze the sensor data, blood oxygen concentration data, and oxygen generator data of the system. The data management module performs multi-source fusion of the user's health data and environmental data, and realizes real-time analysis through big data analysis and artificial intelligence algorithms to obtain a comprehensive health environment index. Whether to transfer the result to the equipment management module is determined by comparing the value of this index with a preset value; the equipment management module is used to manage the operating status and parameter settings of the oxygen generator, sensors, and solenoid valves, and realizes data interaction and collaborative work between the oxygen generator and health monitoring devices through Internet of Things technology according to the analysis result of the data management module, and automatically adjusts the operating parameters of the oxygen generator according to the user's real-time health status; the user management module is used to manage the user permissions and login authentication of the system; the database includes at least a student information table, a classroom information table, an oxygen generator information table, a sensor information table, a health information table, and an alarm record table; the system management platform supports centralized management of multiple users and multiple devices, visualizes health data, and provides corresponding oxygen supply strategy suggestions.

[0005] Preferably, the database is a relational database, and the student information is associated with the classroom information, the classroom information is associated with the oxygen generator information, the classroom information is associated with the sensor information, the student information is associated with the health information, and the alarm record is associated with the classroom information and student information.

[0006] Preferably, it further has a prediction module, which uses machine learning algorithms to perform trend analysis on user health data, predicts potential health risks, and adjusts the oxygen supply strategy in advance.

[0007] Preferably, the data processing method of the management platform includes the following steps: Step 1, data collection: The collected data includes sensor data and student health data, which are respectively stored in the corresponding sensor information table and health information table in the database; Step 2, data fusion: Use the HEI algorithm to fuse the sensor data and student health data to generate a comprehensive health environment index; Step 3, dynamic adjustment: Dynamically adjust the output of the oxygen generator according to the comprehensive health environment index, update the oxygen generator information table in the database, and generate an oxygen supply strategy; Step 4, Anomaly Detection: Use the TSAD algorithm to detect students' health anomalies and generate alarm records in the alarm record table of the database. Step 5, Fault Prediction: Use the MLFP model to predict the faults of the oxygen generator and give early warnings.

[0008] Preferably, the HEI algorithm described in Step 2 is a dynamic priority adjustment algorithm based on the comprehensive health environment index. According to the HEI value obtained by the algorithm, that is, the comprehensive health environment index, the oxygen supply strategy is divided into an emergency mode, an optimization mode, and an energy-saving mode; in different modes, the output of the oxygen generator is dynamically adjusted. HEI = α⋅(ω1⋅O2 + ω2⋅(1−CO2)) + β⋅(ω3⋅SpO2 + ω4⋅(1−HR norm )) Where α is the initial value of the environmental data weight coefficient, which is 0.6; β is the initial value of the physiological data weight coefficient, which is 0.4; ω1, ω2, ω3, ω4 are sub-weight coefficients optimized through historical data; HR norm is the heart rate normalization value dynamically adjusted based on the student's age and resting heart rate.

[0009] Preferably, the TSAD algorithm described in Step 4 is an anomaly detection algorithm based on time series. First, perform time series modeling on students' health data; then use a sliding window to calculate statistical features such as mean and variance; then detect anomaly points through a dynamic threshold; if an anomaly is found, start the alarm.

[0010] Preferably, the MLFP model described in Step 5 is a fault prediction model based on machine learning. Input the operating parameters of the oxygen generator into the model. After the model performs modeling and analysis through a random forest or a long short-term memory network, the model outputs the fault probability.

[0011] The beneficial effects of the present invention are as follows: The present invention can realize the remote control of oxygen generation equipment and the real-time analysis of health data, dynamically adjust the oxygen supply plan according to the user's health status and environmental conditions, support the unified management of multiple users and multiple devices, and has a higher integration degree and intelligence level; realize the deep integration of oxygen generation equipment and health monitoring equipment, break through the limitations of a single data source; propose the TSAD algorithm to detect students' health anomalies in real time, with high sensitivity and low false alarm rate; propose the MLFP model to predict equipment faults based on machine learning, and improve the reliability of the system. Detailed Embodiments

[0012] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the following further describes the present invention in detail in conjunction with embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0013] A management platform for a plateau oxygen generation and health monitoring system, which is used to manage the plateau oxygen generation and health monitoring system. The plateau oxygen generation and health monitoring system includes an oxygen generator, sensors, solenoid valves, blood oxygen concentration detectors, a data transmission system, a control system and a management platform. The oxygen generator provides oxygen for school classrooms, the sensors real-time monitor various parameters in the classroom, the data transmission system transmits the sensor data to the management platform, the management platform adjusts the control system according to the sensor data, the blood oxygen concentration detector real-time monitors the blood oxygen concentration of students and transmits the data to the management platform, and judges the physical state of students through big data analysis.

[0014] The management platform includes a data management module, an equipment management module, a user management module and a database. The data management module is used to store, manage and analyze the sensor data, blood oxygen concentration data and oxygen generator data of the system. The data management module performs multi-source fusion on the user's health data and environmental data, and realizes real-time analysis through big data analysis and artificial intelligence algorithms to obtain a comprehensive health environment index. According to the comparison between the value of this index and the preset value, it is decided whether to transmit the result to the equipment management module; the equipment management module is used to manage the operating status and parameter settings of the oxygen generator, sensors and solenoid valves, and realizes data interaction and collaborative work between the oxygen generator and health monitoring devices through the Internet of Things technology according to the analysis result of the data management module, and automatically adjusts the operating parameters of the oxygen generator according to the user's real-time health status; the user management module is used to manage the user permissions and login authentication of the system; the database at least includes student information tables, classroom information tables, oxygen generator information tables, sensor information tables, health information tables, and alarm record tables; the system also includes a prediction module, which uses machine learning algorithms to perform trend analysis on user health data, predicts potential health risks, and adjusts the oxygen supply strategy in advance. The system management platform supports centralized management of multiple users and multiple devices, visualizes health data and provides corresponding oxygen supply strategy suggestions.

[0015] The database is a relational database, such as MySQL, Oracle, etc., which serves as the system's database. According to the system's functional requirements and data characteristics, the database structure is designed and association relationships are established. Student information is associated with classroom information, where one student belongs to one classroom; classroom information is associated with oxygen generator information, where one classroom corresponds to one oxygen generator; classroom information is associated with sensor information, where one classroom has multiple sensor data records corresponding to that classroom; student information is associated with health information, where one student has multiple health data records; and alarm records are associated with classroom information and student information. A data backup and recovery mechanism is established to regularly back up the database to ensure data security and availability. When the database fails, data can be restored in a timely manner to ensure the normal operation of the system.

[0016] The data processing flow of the management platform is as follows: Step 1, data collection: The collected data includes sensor data and student health data, which are respectively stored in the corresponding sensor information table and health information table in the database; Step 2, data fusion: The HEI algorithm is used to fuse sensor data and student health data to generate a comprehensive health environment index; The HEI algorithm is a dynamic priority adjustment algorithm based on the comprehensive health environment index. According to the HEI value obtained by the algorithm, that is, the comprehensive health environment index, the oxygen supply strategy is divided into an emergency mode, an optimization mode, and an energy-saving mode. When HEI < 0.4, it is the emergency mode: maximize the oxygen supply and trigger an audible and visual alarm; When 0.4 ≤ HEI < 0.7, it is the optimization mode: adjust the oxygen supply according to the scenario rules; When HEI ≥ 0.7, it is the energy-saving mode: maintain the minimum oxygen supply and give priority to reducing energy consumption. Under different modes, the output of the oxygen generator is dynamically adjusted; HEI = α⋅(ω1⋅O2 + ω2⋅(1−CO2)) + β⋅(ω3⋅SpO2 + ω4⋅(1−HR norm )) Among them, α is the initial value of the environmental data weight coefficient, which is 0.6; β is the initial value of the physiological data weight coefficient, which is 0.4; ω1, ω2, ω3, ω4 are sub-weight coefficients optimized through historical data; HR norm is the heart rate normalization value dynamically adjusted based on the student's age and resting heart rate. This value is used to quantify the student's heart rate status and is one of the important parameters in the comprehensive health environment index for dynamically adjusting the oxygen supply strategy. The calculation formula is: ; Among them, HR current is the student's current heart rate value; HR min is the lower limit of the resting heart rate of the student's age group; HR max is the upper limit of the resting heart rate of the student's age group.

[0017] Step 3, Dynamic adjustment: Dynamically adjust the output of the oxygen generator according to the comprehensive health environment index, update the oxygen generator information table in the database, and generate an oxygen supply strategy; Step 4, Abnormality detection: Use the TSAD algorithm to detect students' health abnormalities and generate alarm records in the alarm record table of the database. By analyzing students' health data (blood oxygen, heart rate, respiratory rate), health abnormalities are detected in real time; The TSAD algorithm is an anomaly detection algorithm based on time series. First, perform time series modeling on students' health data; Then use a sliding window to calculate statistical features such as mean and variance; Then detect anomaly points through dynamic thresholds; If anomalies such as sudden drops in blood oxygen or abnormal increases in heart rate are found, alarms are triggered.

[0018] The sliding window is used to extract local features from time series data. The window size is set according to the characteristics of the data and the application scenario. The specific methods include: 1. Based on data frequency: If the data collection frequency is high (such as once per second), select a smaller window, for example, 60 data points, that is, 1 minute.

[0019] If the data collection frequency is low (such as once per minute), select a larger window, for example, 60 data points, that is, 1 hour.

[0020] 2. Based on data periodicity: If the data has obvious periodicity (such as regular changes every day or week), set the window size to an integer multiple of the period.

[0021] For example, if the data shows periodic changes every day, the window size can be set to the number of data points for 24 hours.

[0022] 3. Based on empirical values: According to the analysis results of historical data, select an empirical value as the window size.

[0023] For example, in the scenario of plateau schools, the window size can be initially set to the number of data points for 30 minutes.

[0024] Adopt a mechanism for dynamically adjusting the window size, specifically: 1. Based on data volatility: When the data fluctuates greatly, reduce the window size to improve sensitivity.

[0025] When the data fluctuates little, increase the window size to reduce the false alarm rate.

[0026] 2. Based on anomaly detection results: If consecutive anomaly points are detected, the window size can be reduced to analyze the data more finely.

[0027] If no abnormal points are detected for a long time, the window size can be increased to reduce the computational complexity.

[0028] The dynamic threshold is used to determine whether a data point is abnormal. Compared with the fixed threshold, the dynamic threshold can adapt to the changes in the data, reducing the false alarm rate and the missed alarm rate. The dynamic threshold is set based on statistical features or according to historical data or machine learning models. When a data point exceeds the threshold range, it is marked as an abnormal point.

[0029] To adapt to the changes in the data, a mechanism for dynamically adjusting the threshold is adopted: 1. Based on the data distribution: When the data distribution changes (such as seasonal or sudden changes), recalculate the threshold.

[0030] For example, in the scenario of a plateau school, different thresholds are set according to different time periods (such as during class, between classes, at night).

[0031] 2. Based on the anomaly detection results: If consecutive abnormal points are detected, appropriately relax the threshold to reduce the false alarm rate.

[0032] If no abnormal points are detected for a long time, appropriately tighten the threshold to improve the sensitivity.

[0033] Step 5, Fault prediction: Use the MLFP model to predict the faults of the oxygen generator and give early warnings. By analyzing the operation data of the oxygen generator, predict the equipment faults. The MLFP model is a machine learning-based fault prediction model. Input the operation parameters of the oxygen generator such as oxygen output, pressure, and temperature into the model. After the model performs modeling and analysis through algorithms such as RandomForest or Long Short-Term Memory (LSTM) algorithms, the model outputs the fault probability.

[0034] The RandomForest algorithm is suitable for structured data, medium and small-scale data sets, and scenarios with high requirements for interpretability.

[0035] The LSTM algorithm is suitable for time series data, large-scale data sets, and scenarios with low requirements for real-time performance.

[0036] The accuracy and efficiency of fault prediction can be improved by designing a model selector to automatically select the optimal model according to the data characteristics.

[0037] The selection of the model algorithm is based on the data characteristics and application scenarios, as follows: The selection basis according to the data characteristics is as follows: 1. Data type: If the data is structured data (such as tabular data, including the operating parameters of oxygen generators, fault records, etc.), the random forest algorithm model is preferably selected.

[0038] If the data is time series data (such as continuously collected sensor data), the LSTM algorithm model is preferably selected.

[0039] 2. Data scale: If the data scale is small (such as several thousand records), the random forest algorithm model is preferably selected.

[0040] If the data scale is large (such as several hundred thousand records), the LSTM algorithm model is preferably selected.

[0041] 3. Data dimension: If the data dimension is low (such as dozens of features), the random forest algorithm model is preferably selected.

[0042] If the data dimension is high (such as several hundred features), the LSTM algorithm model is preferably selected.

[0043] The selection basis according to the application scenario is as follows: 1. Real-time requirement: If the real-time requirement is high (such as rapid fault prediction is needed), the random forest algorithm model is preferably selected.

[0044] If the real-time requirement is low (such as a long training time can be accepted), the LSTM algorithm model is preferably selected.

[0045] 2. Interpretability requirement: If the prediction results of the model need to be explained (such as analyzing which features cause faults), the random forest algorithm model is preferably selected.

[0046] If the interpretability requirement is low, the LSTM algorithm model is preferably selected.

[0047] To automatically select a model according to the data characteristics, a model selector (Model Selector) is designed, and its working process is as follows: (1) Data characteristic analysis; Data type detection: Determine whether the data is structured data or time series data.

[0048] Data scale detection: Calculate the number of records and features of the data.

[0049] Data distribution detection: Analyze the statistical characteristics of the data (such as mean, variance, correlation).

[0050] (2) Model selection rules; Based on the results of data characteristic analysis, formulate model selection rules: 1. If the data is structured data and of small scale, select the random forest algorithm model.

[0051] 2. If the data is time series data and of large scale, select the LSTM algorithm model.

[0052] 3. If the data characteristics are between the two, train both models simultaneously and select the model with better performance.

[0053] (3)Performance evaluation; Use cross-validation or the holdout method to evaluate the performance of the model (such as accuracy, recall, F1 score). Select the model with the optimal performance as the final model.

[0054] The management platform can perform data statistics, trend analysis, and correlation analysis: conduct a comparative test experiment, select some classrooms as the control group, without installing the intelligent oxygen supply system, and only perform conventional ventilation and air purification treatments. Compare the changes in environmental indicators and students' physiological indicators between the control group and the experimental group classrooms, conduct statistical analysis on the obtained data, calculate parameters such as the average value and standard deviation of students' physiological indicators, and the average value and standard deviation of classroom environmental indicators. Evaluate the effect of the intelligent oxygen supply system by comparing the data differences between the experimental group and the control group.

[0055] Then conduct trend analysis: conduct trend analysis on the test data to observe the changing trends of students' physiological indicators and classroom environmental indicators over time. Understand the long-term effect and stability of the intelligent oxygen supply system by analyzing the trend changes. And correlation analysis: conduct correlation analysis on students' physiological indicators and classroom environmental indicators to understand the relationship between the two. Further optimize the control strategy of the intelligent oxygen supply system by analyzing the correlation to improve the effect and efficiency of the system.

[0056] The above are the specific embodiments of the present invention and the technical principles applied. Any modification or equivalent transformation based on the technical solution of the present invention shall be included within the protection scope of the present invention.

Claims

1. A plateau oxygen production and health monitoring system management platform, characterized by: The plateau oxygen production and health monitoring system includes an oxygen generator, a sensor, a solenoid valve, a blood oxygen concentration detector, a data transmission system, a control system, and a management platform. The oxygen generator provides oxygen to school classrooms. The sensor monitors various parameters in the classroom in real time. The data transmission system transmits the sensor data to the management platform. The management platform adjusts the control system according to the sensor data. The blood oxygen concentration detector monitors the blood oxygen concentration of students in real time and transmits the data to the management platform. The physical condition of the students is judged through big data analysis. The management platform includes a data management module, an equipment management module, a user management module, and a database. The data management module is used to store, manage, and analyze the sensor data, blood oxygen concentration data, and oxygen generator data of the system. The data management module integrates the user's health data and environmental data from multiple sources and analyzes the data through big data. The data management module uses the data analysis and artificial intelligence algorithm to realize real-time analysis, obtain the comprehensive index of the health environment, and decide whether to pass the result to the equipment management module based on the comparison between the value of the index and the preset value; the equipment management module is used to manage the operating status and parameter settings of the oxygen generator, sensor, and solenoid valve, and realizes data interaction and collaborative work between the oxygen generator and the health monitoring equipment through the Internet of Things technology according to the analysis results of the data management module, and automatically adjusts the operating parameters of the oxygen generator according to the real-time health status of the user; the user management module is used to manage the user rights and login authentication of the system; the database includes at least student information table, classroom information table, oxygen generator information table, sensor information table, health information table, and alarm record table; the system management platform supports centralized management of multiple users and multiple devices, visualizes health data and provides corresponding oxygen supply strategy recommendations.

2. A plateau oxygen production and health monitoring system management platform according to claim 1, characterized in that: The database is a relational database, in which student information is associated with classroom information, classroom information is associated with oxygen concentrator information, classroom information is associated with sensor information, student information is associated with health information, and alarm records are associated with classroom information and student information.

3. A plateau oxygen production and health monitoring system management platform according to claim 1, characterized in that: It also has a prediction module that uses machine learning algorithms to perform trend analysis on user health data, predict potential health risks, and adjust oxygen supply strategies in advance.

4. A plateau oxygen production and health monitoring system management platform according to claim 1, characterized in that: The data processing method of the management platform comprises the following steps: Step 1, data collection: The collected data includes sensor data and student health data and is stored in the corresponding sensor information table and health information table of the database; Step 2: Data fusion: Use the HEI algorithm to fuse sensor data and student health data to generate a comprehensive health environment index; Step 3, dynamic adjustment: dynamically adjust the output of the oxygen generator according to the comprehensive health environment index, update the oxygen generator information table in the database, and generate an oxygen supply strategy; Step 4, anomaly detection: use TSAD algorithm to detect student health anomalies and generate alarm records to the alarm record table of the database; Step 5: Fault prediction: Use the MLFP model to predict oxygen concentrator failure and provide early warning.

5. A plateau oxygen production and health monitoring system management platform according to claim 4, characterized in that: The HEI algorithm described in step 2 is a dynamic priority adjustment algorithm based on the health environment comprehensive index. The oxygen supply strategy is divided into emergency mode, optimization mode and energy-saving mode according to the HEI value obtained by the algorithm, that is, the health environment comprehensive index; in different modes, the output of the oxygen concentrator is dynamically adjusted; HEI=α⋅(ω1⋅O2+ω2⋅(1−CO2))+β⋅(ω3⋅SpO2+ω4⋅(1−HR) norm )) Among them, α is the environmental data weight coefficient, and its initial value is 0.6; β is the physiological data weight coefficient, and its initial value is 0.4; ω1, ω2, ω3, and ω4 are sub-weight coefficients obtained through optimization of historical data; HR norm A heart rate normalized value that is dynamically adjusted based on the student's age and resting heart rate.

6. A plateau oxygen production and health monitoring system management platform according to claim 4, characterized in that: The TSAD algorithm described in step 4 is an anomaly detection algorithm based on time series. First, the student health data is modeled as a time series; then the sliding window is used to calculate statistical features such as mean and variance; Then, abnormal points are detected through dynamic thresholds; if abnormalities are found, an alarm is activated.

7. A plateau oxygen production and health monitoring system management platform according to claim 1, characterized in that: The MLFP model described in step 5 is a fault prediction model based on machine learning. The operating parameters of the oxygen generator are input into the model. After the model is modeled and analyzed through a random forest or a long short-term memory network, the model outputs the fault probability.