Intelligent water level monitoring, regulating and controlling system and method based on breathing humidifier

Abstract

The invention discloses an intelligent water level monitoring, regulating and controlling system and method based on a breathing humidifier. A data acquisition module monitors the real-time temperature and humidity in a humidifying water tank, the temperature and humidity of the external environment and the flow of an air source and transmits the acquired data to a processing and identifying module; the processing identification module receives the temperature, humidity and flow data of the data acquisition module, carries out calculation to obtain a key difference value and a change rate index, and carries out function training on the monitored data by constructing a water level monitoring data model; the main control module executes a corresponding humidifying water infusion instruction to the flow control module, and a flow adjusting valve is controlled to adjust the humidifying water infusion water amount; and the alarm feedback module receives an alarm instruction of the main control module and sends out an alarm prompt tone when the processing identification module monitors abnormity. Internal and external data of the breathing humidifier can be monitored in real time, the water level of the humidifying water tank is intelligently regulated and controlled through the data model so as to reach the target temperature and humidity range of the patient end, and therefore the most reasonable airway humidifying scheme is provided for a patient.

Description

Technical Field

[0001] The present invention relates to a water level monitoring and control system and method, and in particular to a water level intelligent monitoring and control system and method based on a respiratory humidifier. Background Art

[0002] A respiratory humidifier is a medical device that helps maintain the normal physiological function of the patient's airway system by heating and humidifying the respiratory gas. Its main purpose is to prevent airway dryness and inflammation and improve the mucus clearance function on the airway surface. It is usually used in conjunction with a ventilator to provide patients with heated and humidified respiratory gas to maintain the normal clearance function of the patient's airway surface mucociliary system and the normal contraction and diffusion characteristics of the alveolar epithelium, helping patients maintain normal airway function and avoid airway dryness and inflammatory symptoms.

[0003] A respiratory humidifier generally includes a heating plate, a humidification water tank, a temperature and humidity monitoring device, and a control display panel. During operation, sterile water for injection enters the humidification water tank through a dedicated pipeline, and the control display panel adjusts the power of the heating plate according to preset parameters, heats the humidification water in the humidification water tank to generate water vapor, and carries the heated and humidified gas to the patient through the respiratory tube.

[0004] During this process, the water level in the humidification tank is crucial to the humidification effect. However, in actual applications, medical staff are busy at work and may cause abnormal water levels in the humidification tank due to negligence or improper operation. When the power of the heating plate is fixed, a water level that is too low may cause the humidification tank to dry out or insufficient water evaporation, which will not only damage the equipment, but also reduce the gas humidification level at the patient end and fail to meet treatment needs. On the contrary, a water level that is too high may cause water overflow in the humidification tank or excessive gas humidification, affecting the normal operation of the equipment and the gas humidification effect, and bringing unnecessary risks to patients.

[0005] The existing technology usually determines the water level in the humidification water tank by setting a liquid level monitor at a fixed position, or by obtaining data such as the outlet temperature of the humidification water tank and the temperature of the heating plate, and then inferring the water level based on the power of the heating plate. However, this method has poor accuracy and cannot accurately report changes in the water level in real time, resulting in the inability to accurately identify the water level, increasing the potential risks in patient treatment. In addition, the existing technology relies on adjusting the heating rate of the heating plate to prompt the gas to quickly reach the temperature and humidity targets at the patient end. However, this approach not only has safety issues that need to be improved, but is also likely to cause the humidification water to be consumed too quickly, making it difficult to replenish the humidification water in a timely manner.

[0006] Therefore, the existing technology has great deficiencies in the monitoring and control of the water level in the humidification water tank. There is an urgent need for a more accurate, real-time and intelligent water level monitoring and control system to ensure that patients obtain ideal airway humidification effects while improving the safety and reliability of the equipment. Summary of the invention

[0007] The purpose of the present invention is to overcome the defects of the prior art and provide a water level intelligent monitoring and control system and method based on a respiratory humidifier, which monitors the water level changes in real time and intelligently adjusts the infusion amount of humidified water according to the temperature and humidity feedback from the patient end; based on historical data and real-time data, the constructed water level monitoring data model is used to predict the water level change trend in the next 24 hours.

[0008] The purpose of the present invention is achieved in one aspect as follows: a water level intelligent monitoring and control system based on a respiratory humidifier, comprising a data acquisition module, a processing and identification module, a main control module, a flow control module, an alarm feedback module, a wireless module and a user-end display module;

[0009] The data acquisition module is used to monitor the temperature and humidity data in the humidification water tank, the ambient temperature and humidity data, the air source flow data, the humidification water infusion flow data and the user-end temperature and humidity data in real time; and transmit the collected data to the processing and identification module;

[0010] The processing and identification module is used to calculate the collected data to obtain key difference and change rate indicators, and use the constructed water level monitoring data model to predict the water level change trend in the next 24 hours based on historical data and real-time data, and send real-time signals to the main control module according to the water level status in the tank;

[0011] The main control module receives the signal from the processing module, executes the corresponding humidification water flow command to the flow control module, intelligently controls the flow regulating valve, and automatically adds humidification water to the humidification water tank to a safe water level range; at the same time, the main control module also combines the real-time temperature and humidity feedback of the patient end to perform coordinated adjustment to ensure that the humidification water volume is sufficient to meet the water volume demand of the patient end to reach the target temperature and humidity after heating; when an abnormal signal is received, an alarm command is sent to the alarm feedback module;

[0012] The alarm feedback module is used to receive instructions from the main control module and send out an alarm prompt sound when the data is abnormal, indicating that the water level is abnormal or that spare humidification water needs to be added in time;

[0013] The wireless module is used to upload the data of the main control module and the alarm feedback module to the user-side display module;

[0014] The flow control module is connected to the flow regulating valve arranged on the liquid inlet of the humidification water tank, and is used to receive the signal of the main control module, adjust the amount of humidification water infusion, and keep the water level in the humidification water tank within a safe range;

[0015] The user-side display module is used to display the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning and sound and light alarm information to prompt medical staff to handle;

[0016] The data acquisition module, processing and identification module, flow control module, alarm feedback module, wireless module and main control module are electrically connected;

[0017] The data acquisition module is electrically connected to the internal temperature and humidity sensor, the environmental temperature and humidity sensor, the air source flow sensor, the humidified water infusion flow sensor and the patient end temperature and humidity sensor.

[0018] As a further limitation of the present invention, the air source flow sensor is arranged on the inner wall of the air inlet of the humidification water tank, and the internal temperature and humidity sensor is arranged in the humidification water tank; the ambient humidity sensor is arranged on the base of the humidification water tank; the humidification water infusion flow sensor is on the inner wall of the liquid inlet of the humidification water tank; and the patient-end temperature and humidity sensor is located at the patient-end of the ventilator pipeline.

[0019] Another aspect of the present invention is to achieve the purpose of: a method for intelligent monitoring and controlling water level based on a respiratory humidifier, comprising the following steps:

[0020] 1) The data acquisition module monitors the real-time temperature and humidity in the humidification water tank (unit: Celsius (℃) and relative humidity (%RH)), external environmental temperature and humidity (unit: Celsius (℃) and relative humidity (%RH)), air source flow (unit: liter per minute (L / min)), humidification water infusion flow (unit: milliliter per minute (ml / min)) and patient-side temperature and humidity (unit: Celsius (℃) and relative humidity (%RH)) through the internal temperature and humidity sensor, the environmental temperature and humidity sensor, the air source flow sensor, the humidification water infusion flow sensor and the patient-side temperature and humidity sensor in real time, and transmits the collected data to the processing and identification module;

[0021] 2) The processing and identification module receives the data from the data acquisition module, calculates the key difference and change rate indicators, and performs function training on the monitored and calculated data through the constructed water level monitoring data model; generates the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve and abnormal mode warning signal, and transmits the corresponding signal to the main control module;

[0022] 3) The main control module receives the signal from the processing module, executes the corresponding humidification water flow command to the flow control module, intelligently controls the flow regulating valve, and automatically adds humidification water to the humidification water tank to a safe water level range; at the same time, the main control module also combines the real-time temperature and humidity feedback from the patient end to coordinate and adjust to ensure that the humidification water volume is sufficient to meet the water volume demand of the patient end to reach the target temperature and humidity after heating; when an abnormal signal is received, an alarm command is sent to the alarm feedback module;

[0023] 4) When the processing module detects abnormal data, the alarm feedback module receives instructions from the main control module and issues an alarm prompt to indicate that the water level is abnormal or that spare humidification water needs to be added in time; the main control module uploads the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning signal, and sound and light alarm information to the user-end display module through the wireless module to prompt medical staff to handle it.

[0024] As a further limitation of the present invention, the processing and identification module in step 2) receives the data from the data acquisition module and calculates to obtain the key difference and change rate indicators, specifically including:

[0025] The temperature difference between the humidification water tank and the environment ΔT: ΔT = T 罐 -T 环 , ΔT reflects the difference between the temperature in the humidification water tank and the ambient temperature;

[0026] The humidity difference between the humidification water tank and the ambient humidity ΔRH: ΔRH=RH 罐 -RH 环 , the ΔRH difference represents the difference between the humidity in the humidification water tank and the ambient humidity;

[0027] 10s temperature difference change rate ΔT' / Δt: First, record the temperature difference ΔT1 and ΔT2 at two adjacent 10-second time points, which correspond to the temperature difference between the humidification water tank and the environment at t1 and t2 respectively. Then, calculate the change between the two temperature differences ΔT'=ΔT2-ΔT1; calculate the 10s temperature difference change rate by the formula: ΔT' / Δt=ΔT' / 10s. The ΔT' / Δt change rate reflects the speed of change of the temperature difference between the humidification water tank and the environment within 10 seconds.

[0028] 10s humidity difference change rate ΔRH' / Δt: record the humidity difference ΔRH1 and ΔRH2 at two adjacent 10-second time points, then calculate the change between the two humidity differences ΔRH'=ΔRH2-ΔRH1, and finally get the 10s humidity difference change rate by the formula: ΔRH' / Δt=ΔRH' / 10s, the ΔRH' / Δt change rate indicates how fast the humidity difference between the humidification water tank and the environment changes within 10 seconds;

[0029] 10s gas source flow rate change rate ΔQ / Δt: record the gas source flow rates Q1 and Q2 at two adjacent 10-second time points, and calculate the change between the two flow rates ΔQ=Q2-Q1; calculate the 10s gas source flow rate change rate by the formula: ΔQ / Δt=ΔQ / 10s. The ΔQ / Δt change rate reveals the change of the gas source flow rate within 10 seconds.

[0030] As a further limitation of the present invention, the function training of the data through the constructed water level monitoring data model in step 2) specifically includes: using 1D-CNN to extract the temperature and humidity inside the humidification water tank, the temperature and humidity of the external environment, the temperature and humidity difference between the inside and outside of the humidification water tank, the air source flow rate and its rate of change characteristics, and then learning the time series information through the LSTM network to complete the prediction of water level, water consumption rate, remaining available time, etc.

[0031] As a further limitation of the present invention, preprocessing is required before using 1D-CNN to extract the temperature and humidity in the humidification water tank, the temperature and humidity of the external environment, the temperature and humidity difference between the inside and outside of the humidification water tank, the air source flow rate and its change rate characteristics, specifically including:

[0032] The input data includes the temperature data T in the humidification water tank 罐 , humidity data RH in the humidification water tank 罐 , Ambient temperature data T 环 , Ambient temperature data RH 环 , air source flow rate Q, temperature difference between the humidification water tank and the environment ΔT, humidity difference between the humidification water tank and the environment ΔRH, 10s temperature difference change rate ΔT' / Δt, 10s humidity difference change rate ΔRH' / Δt, 10s air source flow rate change rate ΔQ / Δt; the data of each time step t is represented as a one-dimensional vector containing multiple features;

[0033] Before the data is input into the 1D-CNN-LSTM model, it needs to be weighted. Different data items are assigned different weight coefficients according to their importance:

[0034] The primary weight w1 (0.7-0.9) is applicable to the original collected data (T 罐 (t), RH 罐 (t), T 环 (t), RH 环 (t), Q(t))

[0035] The intermediate weight w2 (0.9 to 1.1) is suitable for difference data (ΔT(t), ΔRH(t));

[0036] Advanced weight w3 (1.1-1.3) is suitable for rate of change data

[0037] The features of each time step are weighted to form a weighted data matrix, expressed as:

[0038]

[0039] The weighted data matrix X weighted, whose shape is (n,m), where n is the number of time steps and m is the number of features at each time step; this data matrix will be used as the input of the subsequent 1D-CNN layer and LSTM layer network for further analysis and prediction.

[0040] As a further limitation of the present invention, the step 2) of generating a 24-hour water level change prediction curve specifically includes:

[0041] The 1D-CNN-LSTM model is used to predict the changing trend of the water level in the humidification tank within the next 24 hours, and the water level change curve is output to help medical staff analyze the water level dynamics in advance;

[0042] The 1D-CNN-LSTM model outputs the water level forecast for the next 24 hours:

[0043]

[0044] in, is the predicted water level value at time t+1, N is the number of time steps in the next 24 hours, and t is the time step in minutes (set to 5 minutes);

[0045] The predicted water level Draw a water level change curve, output a water level change trend chart for the next 24 hours, upload it to the user-side display module for real-time update and embed it into the interactive interface of medical staff.

[0046] As a further limitation of the present invention, the abnormal mode warning signal in step 2) specifically includes:

[0047] Real-time detection of water level changes in the humidification water tank and abnormal modes of humidification water infusion, triggering early warning so that medical staff can promptly handle equipment failures or replace spare humidification water;

[0048] The water level change rate calculation formula is:

[0049]

[0050] Where Δt represents the time interval, is the water level value at the current time t output by the 1D-CNN-LSTM model, The water level value at the previous time t-1 output by the 1D-CNN-LSTM model;

[0051] The abnormal mode determination criteria are as follows:

[0052] Abnormal water level drop rate Abnormal water level rise rate Abnormal humidified water infusion add =0;

[0053] Among them, r 正常上限 is the normal upper limit of the rate of water level drop, r 正常下限 Q is the normal lower limit of the rate of water level rise add Add flow rate of humidification water, unit: ml / s;

[0054] When an abnormal pattern is detected, an alarm signal is generated and transmitted to the main control module.

[0055] As a further limitation of the present invention, the main control module in step 3) performs coordinated adjustment in combination with the real-time temperature and humidity feedback from the patient, specifically including:

[0056] Conversion of water level in humidification tank:

[0057]

[0058] Where: W level The relative position of the current water level of the humidification tank within the safe water level range, ranging from 0 to 1; H is the current water level of the humidification tank (unit: mm); max is the maximum water level height of the humidification water tank (unit: mm, the upper limit water level of the humidification water tank);

[0059] W level =0 means the water level in the humidification water tank has dropped to the minimum safety threshold; W level =1 means the water level in the humidification water tank has reached the upper limit of the maximum safe water level; When the value is between 0 and 1, it means that the water level is within the safe range, and the proportional value reflects the relative relationship between the current water level and the safe water level.

[0060] Patient end temperature difference ΔT: represents the difference between the patient end temperature and the target temperature ΔT = T p -T target , when ΔT>0, it means the temperature is too high; when ΔT=0, it means the temperature is appropriate; when ΔT<0, it means the temperature is too low;

[0061] Patient end humidity difference ΔH: indicates the difference between the patient end humidity and the target humidity ΔH = H p -H target , when ΔH>0, it means the humidity is too high; when ΔH=0, it means the humidity is appropriate; when ΔH<0, it means the humidity is too low;

[0062] Add humidification water flow calculation formula

[0063] Q add =k×(ΔT·w T +ΔH·w H )×(1-W level )

[0064] Among them, Q add Add flow rate of humidification water (unit: ml / s); ΔT = T p -T target is the temperature difference; ΔH=H p -H target is the humidity difference; w T (0.6~0.7) and w H (0.4~0.5) are the weight coefficients of temperature and humidity, respectively, which are used to adjust the relative influence of temperature and humidity on the humidification water flow; W level is the current water level ratio of the humidification water tank (between 0 and 1); k (0.8 to 1.2) is the adjustment coefficient.

[0065] When the present invention is working, the data acquisition module collects the temperature and humidity data in the humidification water tank, the ambient temperature and humidity data, the air source flow data, the humidification water infusion flow and the temperature and humidity at the patient end, and the processing and identification module calculates the key difference and change rate index in the humidification water tank, and the water level in the humidification water tank is obtained according to the preset algorithm, and the humidification water consumption rate, the remaining available time, the 24-hour water level change prediction curve and the abnormal mode warning signal are predicted. The main control module receives the signal of the processing module, executes the corresponding humidification water flow instruction to the flow control module, intelligently controls the flow regulating valve, and automatically adds humidification water to the humidification water tank to the safe water level range; at the same time, the main control module also combines the real-time temperature and humidity feedback of the patient end to coordinate and adjust to ensure that the humidification water volume is sufficient to meet the water volume demand of the patient end to reach the target temperature and humidity after heating; when an abnormal signal is received, an alarm instruction is sent to the alarm feedback module. Finally, the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning and sound and light alarm information are displayed in real time on the computer screen of medical staff through the wireless module.

[0066] The present invention adopts the above technical scheme, and compared with the prior art, the beneficial effects are as follows: the present invention applies a 1D-CNN-LSTM structure, which can effectively extract the characteristics such as the temperature and humidity difference inside and outside the humidification water tank, the air source flow rate and its change rate, and significantly reduce the network parameter amount and calculation complexity, so as to provide guarantee for real-time water level monitoring; then the time series information is learned through the LSTM network to complete the prediction of water level, water consumption rate and remaining available time, and an alarm prompt is issued in case of abnormality, so as to reserve accurate and safe operation time for medical staff, thereby improving the safety and reliability of respiratory humidification equipment; finally, combined with the real-time temperature and humidity feedback of the patient end, the flow regulating valve is intelligently controlled to automatically add humidification water to the humidification water tank to a safe water level range, and at the same time, the water volume demand of the patient end after heating can be met to reach the required target temperature and humidity, so that the temperature and humidity of the patient end reach the target range, thereby providing the patient with the most reasonable airway humidification plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a structural schematic diagram of the system of the present invention.

[0068] Figure 2 is a flow chart of the method of the present invention.

[0069] Figure 3 It is a schematic diagram of the input data of the water level monitoring data model of the present invention (taking the temperature data in the humidification water tank as an example).

[0070] Figure 4 It is a schematic diagram of the structure of the 1DD-CNN-LSTM model of the present invention.

[0071] Figure 5 It is a schematic diagram of the 24-hour water level change prediction curve output by the present invention. DETAILED DESCRIPTION

[0072] like Figure 1 The water level intelligent monitoring and control system based on the respiratory humidifier shown includes a data acquisition module, a processing and identification module, a main control module, a flow control module, an alarm feedback module, a wireless module and a user-end display module;

[0073] The data acquisition module is used to monitor the temperature and humidity data in the humidification water tank, the ambient temperature and humidity data, the air source flow data, the humidification water infusion flow data, and the patient end temperature and humidity data in real time; and transmit the collected data to the processing and identification module;

[0074] The processing and identification module receives the temperature, humidity and flow data from the data acquisition module, calculates the key difference and change rate indicators, and performs function training on the monitored temperature and humidity data through the constructed water level monitoring data model; generates the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, and abnormal mode warning signal, and transmits the corresponding signal to the main control module;

[0075] The main control module receives the signal from the processing identification module and executes the corresponding humidification water flow instruction to the flow control module; when an alarm instruction is received, it is sent to the alarm feedback module;

[0076] When the processing module detects abnormal data, the alarm feedback module receives instructions from the main control module and issues an alarm prompt to indicate that the water level is abnormal or that spare humidification water needs to be added in time;

[0077] The wireless module is used to upload the data of the main control module and the alarm feedback module to the user-side display module;

[0078] The flow control module is connected to the flow regulating valve arranged on the liquid inlet of the humidification water tank, and is used to receive the signal of the main control module, adjust the rate of infusing humidification water, and keep the water level of humidification water within a safe range;

[0079] The user-side display module is used to display the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning and sound and light alarm information to prompt medical staff to handle it;

[0080] The data acquisition module, the processing and identification module, the flow control module, the alarm feedback module, the wireless module are electrically connected to the main control module; the data acquisition module is electrically connected to the internal temperature and humidity sensor, the environmental temperature and humidity sensor, the air source flow sensor, the humidification water infusion flow sensor and the patient-end temperature and humidity sensor; the internal temperature and humidity sensor is arranged in the humidification water tank; the environmental humidity sensor is arranged on the base of the humidification water tank; the air source flow sensor is arranged on the inner wall of the air inlet of the humidification water tank; the humidification water infusion flow sensor is on the inner wall of the liquid inlet of the humidification water tank; the patient-end temperature and humidity sensor is located at the patient end of the ventilator pipeline.

[0081] like Figure 2 As shown, a method for intelligent monitoring and controlling water level based on a respiratory humidifier comprises the following steps:

[0082] 1) The data acquisition module monitors the real-time temperature and humidity in the humidification water tank (unit: Celsius (℃) and relative humidity (%RH)), external environmental temperature and humidity (unit: Celsius (℃) and relative humidity (%RH)), air source flow (unit: liter per minute (L / min)), humidification water infusion flow (unit: milliliter per minute (ml / min)) and patient-side temperature and humidity (unit: Celsius (℃) and relative humidity (%RH)) through internal temperature and humidity sensors, environmental temperature and humidity sensors, air source flow sensors, humidification water infusion flow sensors and patient-side temperature and humidity sensors, and collects the data (such as Figure 3 As shown, retain two decimal places) is transmitted to the processing and identification module;

[0083] 2) The processing and identification module receives the temperature, humidity and flow data from the data acquisition module, calculates the key difference and change rate indicators, and performs function training on the monitored temperature and humidity data through the constructed water level monitoring data model; generates the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, and abnormal mode warning signal, and transmits the corresponding signal to the main control module;

[0084] The processing and identification module receives the temperature, humidity and flow data from the data acquisition module and calculates the key difference and change rate indicators, including:

[0085] The temperature difference between the humidification water tank and the environment ΔT: ΔT = T 罐 -T 环 , ΔT reflects the difference between the temperature in the humidification water tank and the ambient temperature;

[0086] The humidity difference between the humidification water tank and the ambient humidity ΔRH: ΔRH=RH 罐 -RH 环 , the ΔRH difference represents the difference between the humidity in the humidification water tank and the ambient humidity;

[0087] 10s temperature difference change rate ΔT' / Δt: First, record the temperature differences ΔT1 and ΔT2 at two adjacent 10-second time points (t1 and t2, and t2 = t1 + 10s), which correspond to the temperature differences between the humidification water tank and the environment at t1 and t2, respectively. Then, calculate the change between the two temperature differences ΔT' = ΔT2-ΔT1; calculate the 10s temperature difference change rate by the formula: ΔT' / Δt = ΔT' / 10s. The ΔT' / Δt change rate reflects the speed of change of the temperature difference between the humidification water tank and the environment within 10 seconds.

[0088] 10s humidity difference change rate ΔRH' / Δt: record the humidity difference ΔRH1 and ΔRH2 at two adjacent 10-second time points, then calculate the change between the two humidity differences ΔRH'=ΔRH2-ΔRH1, and finally get the 10s humidity difference change rate by the formula: ΔRH' / Δt=ΔRH' / 10s, the ΔRH' / Δt change rate indicates how fast the humidity difference between the humidification water tank and the environment changes within 10 seconds;

[0089] 10s gas source flow rate change rate ΔQ / Δt: record the gas source flow rates Q1 and Q2 at two adjacent 10-second time points, and calculate the change between the two flow rates ΔQ=Q2-Q1; calculate the 10s gas source flow rate change rate by the formula: ΔQ / Δt=ΔQ / 10s. The ΔQ / Δt change rate reveals the change of the gas source flow rate within 10 seconds.

[0090] In order to further optimize the computational efficiency of the network and ensure the high accuracy of water level monitoring, 1D-CNN is used to extract the temperature and humidity inside the humidification water tank, the temperature and humidity of the external environment, the temperature and humidity difference between the inside and outside of the humidification water tank, the air source flow rate and its rate of change. Then, the time series information is learned through the LSTM network to complete the prediction of water level, humidification water consumption rate, remaining available time, etc.

[0091] The input data includes the temperature data T in the humidification water tank 罐 , humidity data RH in the humidification water tank 罐 , Ambient temperature data T 环 , Ambient temperature data RH 环, air source flow rate Q, temperature difference between the humidification water tank and the environment ΔT, humidity difference between the humidification water tank and the environment ΔRH, 10s temperature difference change rate ΔT' / Δt, 10s humidity difference change rate ΔRH' / Δt, 10s air source flow rate change rate ΔQ / Δt. The data of each time step t (set to every 5 minutes) can be represented as a one-dimensional vector containing multiple features.

[0092] Before the data is input into the 1D-CNN-LSTM model, it needs to be weighted. Different data items are assigned different weight coefficients according to their importance:

[0093] The primary weight w1 (0.7-0.9) is applicable to the original collected data (T 罐 (t), RH 罐 (t), T 环 (t), RH 环 (t), Q(t))

[0094] The intermediate weight w2 (0.9 to 1.1) is suitable for difference data (ΔT(t), ΔRH(t));

[0095] Advanced weight w3 (1.1-1.3) is suitable for rate of change data

[0096] The features of each time step are weighted to form a weighted data matrix, expressed as:

[0097]

[0098] The weighted data matrix X weighted , whose shape is (n,m), where n is the number of time steps and m is the number of features at each time step.

[0099] First, the weighted data matrix is ​​input into the 1D-CNN layer to extract local features from the multidimensional features of each time step. The convolution kernel slides on the time series to extract local features. Assuming the size of the convolution kernel is k, that is, each convolution operation covers k time steps of data, the size of the output feature map is n-k+1. The output of the convolution layer is a multidimensional feature map with a shape of (n-k+1, filters), where filters is the number of convolution kernels.

[0100] Convolutional layer formula:

[0101] Z1=Conv1D(X weighted ,filters=64,kernel_size=3,activation=′relu′)

[0102] Among them, filters=64 means the number of convolution kernels is 64, kernel_size=3 means each convolution kernel covers 3 time steps of data, and activation=relu means the activation function uses relu.

[0103] The feature map Z1 extracted by the 1D-CNN layer has a shape of (n-k+1, filters) and is passed to the LSTM layer. The LSTM layer captures the temporal dependencies in the sequence by cyclically processing the time series data. The LSTM layer contains "cell state" and "hidden state", and its update process is regulated by the forget gate, input gate, and output gate:

[0104] Forget Gate:

[0105] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0106] Input Gate:

[0107] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0108]

[0109] Cell State Update:

[0110]

[0111] Output Gate:

[0112] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0113] h t =o t tanh(C t )

[0114] Among them, W f ,W i ,W c ,W o and b f ,b i ,bc ,b o are the weight matrix and bias vector respectively, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function.

[0115] The output of the LSTM layer is a new sequence with shape (n-k+1, units), where units is the number of neural units in the LSTM.

[0116] LSTM layer formula:

[0117] LSTM out =LSTM(Z1,units=128,return_sequences=False)

[0118] Among them, units = 128 means that the number of neural units of LSTM is 128, and return_sequences = False means that LSTM only outputs the final hidden state instead of the hidden state of each time step.

[0119] The output of the LSTM layer is connected to the fully connected layer (Dense layer) to generate the final prediction result. The fully connected layer maps the output of the LSTM layer to one or more target variables, including water level prediction value, humidification water consumption rate, and remaining available time, through an activation function.

[0120] The prediction formula of the fully connected layer is:

[0121]

[0122] Among them, units = 1 means outputting a predicted value, and activation = 'linear' means the output is a continuous value.

[0123] Based on the output results of the 1D-CNN-LSTM network, advanced output functions are added to further meet the needs of medical scenarios, including 24-hour water level change prediction curve and abnormal pattern warning.

[0124] a) Through the 1D-CNN-LSTM model (such as Figure 4 As shown in the figure, the changing trend of the water level in the humidification tank within the next 24 hours is predicted, and the water level change curve is output to help medical staff analyze the water level dynamics in advance;

[0125] The 1D-CNN-LSTM model outputs the water level forecast for the next 24 hours:

[0126]

[0127] in, is the predicted water level value at time t+1, N is the number of time steps in the next 24 hours, t is the time step in minutes (set to 5 minutes, which can be adjusted according to specific circumstances);

[0128] The predicted water level Draw a water level change curve and output a water level change trend chart for the next 24 hours (such as Figure 5 As shown), it is uploaded to the user-side display module for real-time update and embedded in the interactive interface of medical staff.

[0129] b) Real-time detection of water level changes in the humidification water tank and abnormal modes of humidification water infusion, triggering early warning so that medical staff can promptly handle equipment failures or replace spare humidification water;

[0130] The water level change rate calculation formula is:

[0131]

[0132] Where Δt represents the time interval, is the water level value at the current time t output by the 1D-CNN-LSTM model, The water level value at the previous time t-1 output by the 1D-CNN-LSTM model;

[0133] The abnormal mode determination criteria are as follows:

[0134] Abnormal water level drop rate Abnormal water level rise rate Abnormal humidified water infusion add =0

[0135] Among them, r 正常上限 The normal upper limit of the rate when the water level drops is set to 3 mm / 5 minutes, r 正常下限 It is the normal lower limit of the rate when the water level rises, which is set to 1.5mm / 5 minutes. The specific data can be adjusted according to the actual situation. add Add flow rate for humidification water (unit: ml / s);

[0136] When an abnormal pattern is detected, an alarm signal is generated and transmitted to the main control module.

[0137] 3) The main control module executes the corresponding humidification water flow instruction to the flow control module according to the signal received from the processing module. The module intelligently controls the flow regulating valve and automatically adds humidification water to the humidification water tank to a safe water level range; at the same time, the main control module also combines the real-time temperature and humidity feedback of the patient end to coordinate and adjust to ensure that the humidification water volume is sufficient to meet the water volume demand of the patient end to reach the target temperature and humidity after heating; when an abnormal signal is received, an alarm instruction is sent to the alarm feedback module;

[0138] Combined with the real-time temperature and humidity feedback from the patient, coordinated adjustments are made to intelligently control the flow control valve every 10 minutes (adjustable according to specific circumstances), including:

[0139] Conversion of water level in humidification tank:

[0140]

[0141] Where: W level The relative position of the current water level of the humidification tank within the safe water level range, ranging from 0 to 1; H is the current water level of the humidification tank (unit: mm); max is the maximum water level height of the humidification water tank (unit: mm, the upper limit water level of the humidification water tank);

[0142] W level =0 means the water level in the humidification water tank has dropped to the minimum safety threshold; W level =1 means the water level in the humidification water tank has reached the upper limit of the maximum safe water level; When the value is between 0 and 1, it means that the water level is within the safe range, and the proportional value reflects the relative relationship between the current water level and the safe water level.

[0143] Patient end temperature difference: represents the difference between the patient end temperature and the target temperature ΔT = T p -T target , when ΔT>0, it means the temperature is too high; when ΔT=0, it means the temperature is appropriate; when ΔT<0, it means the temperature is too low.

[0144] Patient humidity difference: indicates the difference between the patient humidity and the target humidity ΔH = H p -H target , when ΔH>0, it means the humidity is too high; when ΔH=0, it means the humidity is appropriate; when ΔH<0, it means the humidity is too low.

[0145] Add humidification water flow calculation formula

[0146] Q add =k×(ΔT·w T +ΔH·w H )×(1-W level )

[0147] Among them, Q add Add flow rate of humidification water (unit: ml / s); ΔT = T p -T target is the temperature difference; ΔH=H p -H target is the humidity difference; w T (0.6~0.7) and w H(0.4~0.5) are the weight coefficients of temperature and humidity, respectively, which are used to adjust the relative influence of temperature and humidity on the humidification water flow; W level is the current water level ratio of the humidification water tank (between 0 and 1); k (0.8 to 1.2) is the adjustment coefficient.

[0148] 4) When the processing module detects abnormal data, the alarm feedback module receives instructions from the main control module and issues an alarm prompt to indicate that the water level is abnormal or that spare humidification water needs to be added in time; the main control module uploads the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning signal, and sound and light alarm information to the user-end display module through the wireless module to prompt medical staff to handle it.

[0149] When the present invention is working, the data acquisition module collects the temperature and humidity data in the humidification water tank, the ambient temperature and humidity data, the air source flow data, the humidification water infusion flow data and the patient end temperature and humidity data to the processing and identification module;

[0150] The processing and identification module receives the temperature, humidity and flow data from the data acquisition module, calculates the key difference and change rate indicators, and performs function training on the monitored temperature and humidity data through the constructed water level monitoring data model; generates the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, and abnormal mode warning signal, and transmits the corresponding signal to the main control module;

[0151] The main control module receives the signal of the processing and identification module, executes the corresponding humidification water flow command to the flow control module (flow regulating valve), and automatically adds humidification water to the humidification water tank to the safe water level range. When the processing and identification module identifies that the temperature at the patient end is too high and the humidity is appropriate, the system will reduce the humidification water flow; when the temperature at the patient end is too high and the humidity is too low, the system will increase the humidification water flow to increase the humidity and maintain the temperature; when the temperature at the patient end is appropriate and the humidity is too high, the system will increase the humidification water flow to replenish water to avoid the water level being too low; when the temperature at the patient end is appropriate and the humidity is too low, the system will increase the humidification water flow to increase the humidity and meet the heating requirements; when the temperature at the patient end is too low and the humidity is too high, the system will increase the humidification water flow to increase the humidity and replenish the water level; when the temperature at the patient end is too low and the humidity is too low, the system will increase the humidification water flow to meet the dual requirements of temperature and humidity; when the temperature at the patient end is appropriate and the humidity is appropriate, the system will maintain the humidification water flow to stabilize the water level. The intelligent control flow regulating valve ensures that the supply of humidification water can meet the temperature and humidity requirements of the patient and ensure that the water level of the humidification tank is always kept within a safe range. When the processing module detects that the humidification water has stopped infusion, the alarm feedback module sends out an audible and visual alarm message and prompts the remaining available time in real time; when the processing module detects abnormal data, it receives instructions from the main control module and sends out an alarm prompt sound, indicating that the water level is abnormal or that spare humidification water needs to be added in time.

[0152] The user-end display module is located on the medical staff's computer. The information displayed on the user page includes but is not limited to the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning and sound and light alarm information, etc., prompting medical staff to take action and improve the safety and reliability of medical equipment.

[0153] The present invention provides a water level intelligent monitoring and control system and method based on a respiratory humidifier. A 1D-CNN-LSTM structure is adopted to effectively extract the characteristics such as the temperature and humidity difference inside and outside a humidification water tank, the air source flow rate and its change rate, and the like, while significantly reducing the network parameter amount and the calculation complexity; providing guarantee for real-time water level monitoring; then, the time series information is learned through an LSTM network to complete the prediction of the water level, the water consumption rate and the remaining available time, and an alarm is issued in case of abnormality, so as to reserve accurate and safe operation time for medical staff, thereby improving the safety and reliability of the respiratory humidification equipment; finally, in combination with the real-time temperature and humidity feedback of the patient end, a flow regulating valve is intelligently controlled to automatically add humidification water to the humidification water tank to a safe water level range, while satisfying the water volume demand for the patient end to reach the required target temperature and humidity after heating, so that the temperature and humidity of the patient end reach the target range, thereby providing the patient with the most reasonable airway humidification plan.

[0154] The present invention is not limited to the above-mentioned embodiments. On the basis of the technical solution disclosed in the present invention, technicians in this field can make some substitutions and deformations to some technical features therein according to the disclosed technical content without creative labor, and these substitutions and deformations are all within the protection scope of the present invention.

Claims

1. A water level intelligent monitoring and control system based on a respiratory humidifier, characterized in that: It includes data acquisition module, processing and identification module, main control module, flow control module, alarm feedback module, wireless module and user-end display module; The data acquisition module is used to monitor the temperature and humidity data in the humidification water tank, the ambient temperature and humidity data, the air source flow data, the humidification water infusion flow data, and the patient end temperature and humidity data in real time; and transmitting the collected data to the processing and identification module; The processing and identification module is used to process and calculate the collected data to obtain key difference and change rate indicators, and use the constructed water level monitoring data model to predict the water level change trend in the next 24 hours based on historical data and real-time data, and send real-time signals to the main control module according to the water level status in the tank and the temperature and humidity conditions at the user end; The main control module receives the signal from the processing identification module and executes the corresponding humidification water flow instruction to the flow control module; when an abnormal signal is received, an alarm instruction is sent to the alarm feedback module; The alarm feedback module is used to receive instructions from the main control module and send out an alarm prompt sound when the data is abnormal, indicating that the water level is abnormal or that spare humidification water needs to be added in time; The wireless module is used to upload the data of the main control module and the alarm feedback module to the user-side display module; The flow control module is connected to the flow regulating valve arranged on the liquid inlet of the humidification water tank, and is used to receive the instruction of the main control module, adjust the infusion rate of humidification water, keep the water level of humidification water within a safe range and the water volume can be used for the patient to reach the target temperature and humidity required by the patient; The user-side display module is used to display the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning and sound and light alarm information to prompt medical staff to handle; The data acquisition module, processing and identification module, flow control module, alarm feedback module, wireless module and main control module are electrically connected; The data acquisition module is electrically connected to the internal temperature and humidity sensor, the environmental temperature and humidity sensor, the air source flow sensor, the humidified water infusion flow sensor and the patient end temperature and humidity sensor.

2. The water level intelligent monitoring and control system based on a respiratory humidifier according to claim 1 is characterized in that: The internal temperature and humidity sensor is arranged in the humidification water tank; the environmental humidity sensor is arranged on the base of the humidification water tank; the air source flow sensor is arranged on the inner wall of the air inlet of the humidification water tank; the humidification water infusion flow sensor is on the inner wall of the liquid inlet of the humidification water tank; the patient-end temperature and humidity sensor is located at the patient-end of the ventilator pipeline.

3. A water level intelligent monitoring and control method based on a respiratory humidifier, characterized in that: The following steps are involved: 1) The data acquisition module monitors the real-time temperature and humidity in the humidification water tank in real time through the internal temperature and humidity sensor, the environmental temperature and humidity sensor, the air source flow sensor, the humidification water infusion flow sensor and the patient-end temperature and humidity sensor, the external environmental temperature and humidity, the unit is Celsius °C and the relative humidity %RH; the air source flow, the unit is liter per minute L / min; the humidification water infusion flow, the unit is milliliter per minute ml / min and the patient-end temperature and humidity, the unit is Celsius °C and the relative humidity %RH, and transmits the collected data to the processing and identification module; 2) The processing and identification module receives the temperature, humidity and flow data from the data acquisition module, calculates the key difference and change rate indicators, and performs function training on the monitored temperature and humidity data through the constructed water level monitoring data model; generates the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve and abnormal mode warning signal, and transmits the corresponding signal to the main control module; 3) The main control module executes the corresponding humidification water flow instruction to the flow control module according to the signal received from the processing module. The flow control module intelligently controls the flow regulating valve to automatically add humidification water to the humidification water tank to a safe water level range; at the same time, the main control module combines the real-time temperature and humidity feedback from the patient end to perform coordinated adjustment; when an abnormal signal is received, an alarm instruction is sent to the alarm feedback module; 4) When the processing module detects an abnormal signal, the alarm feedback module receives the alarm command from the main control module and sends out an alarm prompt sound to indicate that the water level is abnormal or that spare humidification water needs to be added in time; the main control module uploads the current water level, water consumption rate, remaining available time, 24-hour water level change prediction curve, abnormal mode warning signal, and sound and light alarm information to the user-end display module through the wireless module to prompt medical staff to handle it.

4. The method for intelligent monitoring and controlling water level based on a respiratory humidifier according to claim 3 is characterized in that: The processing and identification module in step 2) receives the data from the data acquisition module and calculates and obtains the key difference and change rate indicators, which specifically include: The temperature difference between the humidification water tank and the environment ΔT: ΔT = T 罐 -T 环 , ΔT reflects the difference between the temperature in the humidification water tank and the ambient temperature; The humidity difference between the humidification water tank and the ambient humidity ΔRH: ΔRH=RH 罐 -RH 环 , the ΔRH difference represents the difference between the humidity in the humidification water tank and the ambient humidity; 10s temperature difference change rate ΔT' / Δt: First, record the temperature difference ΔT1 and ΔT2 at two adjacent 10-second time points, which correspond to the temperature difference between the humidification water tank and the environment at t1 and t2 respectively. Then, calculate the change between the two temperature differences ΔT'=ΔT2-ΔT1; calculate the 10s temperature difference change rate by the formula: ΔT' / Δt=ΔT' / 10s. The ΔT' / Δt change rate reflects the speed of change of the temperature difference between the humidification water tank and the environment within 10 seconds. 10s humidity difference change rate ΔRH' / Δt: record the humidity difference ΔRH1 and ΔRH2 at two adjacent 10-second time points, then calculate the change between the two humidity differences ΔRH'=ΔRH2-ΔRH1, and finally get the 10s humidity difference change rate by the formula: ΔRH' / Δt=ΔRH' / 10s, the ΔRH' / Δt change rate indicates how fast the humidity difference between the humidification water tank and the environment changes within 10 seconds; 10s gas source flow rate change rate ΔQ / Δt: record the gas source flow rates Q1 and Q2 at two adjacent 10-second time points, and calculate the change between the two flow rates ΔQ=Q2-Q1; calculate the 10s gas source flow rate change rate by the formula: ΔQ / Δt=ΔQ / 10s. The ΔQ / Δt change rate reveals the change of the gas source flow rate within 10 seconds.

5. The method for intelligent monitoring and controlling water level based on a respiratory humidifier according to claim 3, characterized in that: The function training of the monitored temperature and humidity data through the constructed water level monitoring data model described in step 2) specifically includes: using one-dimensional convolution 1D-CNN to extract the temperature and humidity in the humidification water tank, the temperature and humidity of the external environment, the temperature and humidity difference between the inside and outside of the humidification water tank, the air source flow rate and its rate of change characteristics, and then learning the time series information through the LSTM network to complete the prediction of water level, humidification water consumption rate, remaining available time, etc.

6. The method for intelligent monitoring and controlling water level based on a respiratory humidifier according to claim 5, characterized in that: Before using the one-dimensional convolution 1D-CNN to extract the temperature and humidity in the humidification water tank, the temperature and humidity of the external environment, the temperature and humidity difference between the inside and outside of the humidification water tank, the air source flow rate and its change rate characteristics, preprocessing is required, specifically including: The input data includes the temperature data T in the humidification water tank 罐 , humidity data RH in the humidification water tank 罐 , Ambient temperature data T 环 , Ambient temperature data RH 环 , air source flow rate Q, temperature difference between the humidification water tank and the environment ΔT, humidity difference between the humidification water tank and the environment ΔRH, 10s temperature difference change rate ΔT' / Δt, 10s humidity difference change rate ΔRH' / Δt, 10s air source flow rate change rate ΔQ / Δt; the data of each time step t is represented as a one-dimensional vector containing multiple features; Before the data is input into the 1D-CNN-LSTM model, it needs to be weighted. Different data items are assigned different weight coefficients according to their importance: Primary weight w1, 0.7~0.9 is applicable to the original collected data T 罐 (t), RH 罐 (t), T 环 (t), RH 环 (t), Q(t); The intermediate weight w2, 0.9 to 1.1, is applicable to the difference data ΔT(t) and ΔRH(t); Advanced weight w3, 1.1 to 1.3 is suitable for rate of change data The features of each time step are weighted to form a weighted data matrix, expressed as: The weighted data matrix X weighted , whose shape is (n,m), where n is the number of time steps and m is the number of features at each time step; the weighted data matrix will be used as the input of the subsequent 1D-CNN layer and LSTM layer network for further analysis and prediction.

7. The method for intelligent monitoring and controlling water level based on a respiratory humidifier according to claim 3, characterized in that: Generating a 24-hour water level change prediction curve in step 2) specifically includes: The 1D-CNN-LSTM model is used to predict the water level of the humidification tank in the next 24 hours and output the water level change curve; The 1D-CNN-LSTM model outputs the water level forecast for the next 24 hours: in, is the predicted water level value at time t+1, N is the number of time steps in the next 24 hours, and t is the time step in minutes; The predicted water level Draw a water level change curve, output a water level change trend chart for the next 24 hours, upload it to the user-side display module for real-time update and embed it into the interactive interface of medical staff.

8. The method for intelligent monitoring and controlling water level based on a respiratory humidifier according to claim 3, characterized in that: The abnormal mode warning signal in step 2) specifically includes: Real-time detection of water level changes in the humidification water tank and abnormal modes of humidification water infusion, triggering early warning so that medical staff can promptly handle equipment failures or replace spare humidification water; The water level change rate calculation formula is: Where Δt represents the time interval, is the water level value at the current time t output by the 1D-CNN-LSTM model, The water level value at the previous time t-1 output by the 1D-CNN-LSTM model; The abnormal mode determination criteria are as follows: Abnormal humidified water infusion add =0; Among them, r 正常上限 is the normal upper limit of the rate of water level drop, r 正常下限 is the normal lower limit of the rate of water level rise, Q add Add flow rate of humidification water, unit: ml / s; When an abnormal pattern is detected, an alarm signal is generated and transmitted to the main control module.

9. The method for intelligent monitoring and controlling water level based on a respiratory humidifier according to claim 3, characterized in that: The main control module in step 3) combines the real-time temperature and humidity feedback from the patient to perform coordinated adjustment, specifically including: Conversion of water level in humidification tank: Where: W level The relative position of the current water level of the humidification tank within the safe water level range, ranging from 0 to 1; is the current water level of the humidification tank, unit: mm; H max The maximum water level of the humidification water tank, unit: mm, the upper limit water level of the humidification water tank; W level =0 means the water level in the humidification water tank has dropped to the minimum safety threshold; W level =1 means the water level in the humidification water tank has reached the upper limit of the maximum safe water level; When the value is between 0 and 1, it means that the water level is within the safe range, and the proportional value reflects the relative relationship between the current water level and the safe water level; Patient end temperature difference: represents the difference between the patient end temperature and the target temperature ΔT = T p -T target , when ΔT>0, it means the temperature is too high; when ΔT=0, it means the temperature is appropriate; when ΔT<0, it means the temperature is too low; Patient humidity difference: indicates the difference between the patient humidity and the target humidity ΔH = H p -H target , when ΔH>0, it means the humidity is too high; when ΔH=0, it means the humidity is appropriate; when ΔH<0, it means the humidity is too low; Add humidification water flow calculation formula: Q add =k×(ΔT·w T +ΔH·w H )×(1-W level ) Among them, Q add Add flow rate of humidification water, unit: ml / s; ΔT = T p -T target is the temperature difference; ΔH=H p -H target is the humidity difference; w T =0.6~0.7 and w H = 0.4 ~ 0.5 are the weight coefficients of temperature and humidity, respectively, which are used to adjust the relative influence of temperature and humidity on the humidification water flow; W level is the current water level ratio of the humidification water tank, between 0 and 1; k=0.8~1.2 is the adjustment coefficient.

Images