A pressure compensation method for a non-invasive ventilator and a ventilator

Abstract

The application provides a pressure compensation method and a breathing machine for a non-invasive breathing machine, and the method comprises the following steps: collecting a plurality of flow setting values and corresponding actual measured values of a machine pressure sensor under the condition that a breathing pipeline is open; determining a pressure compensation prediction value by fitting and calculating a model parameter through a least square matrix method; using a goodness of fit R 2 to perform model accuracy verification; if the goodness of fit R 2 meets the requirements, the pressure compensation prediction value is the pressure compensation value; otherwise, the pressure compensation prediction value is recalculated; and the pressure compensation value is superimposed on a user pressure setting value to realize real-time pipeline pressure compensation. The advantage of the application is that the method can estimate the pipeline pressure compensation parameter in the machine self-checking process, greatly reduces the calculation amount, and can quickly realize real-time pressure compensation. Moreover, the application increases a model verification mechanism, and avoids the harm caused by high compensation pressure to the user due to the wrong parameter estimation.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of respirators, and particularly relates to a pressure compensation method for a non-invasive respirator and the respirator. BACKGROUND

[0002] In recent years, non-invasive respirators provide users with good application experience due to advantages such as non-invasive operation, light injury and good curative effect. However, in the pressure control process, the airflow provided by the turbine needs to pass through components such as humidification and breathing pipeline, resulting in a certain pressure loss, and the user end pressure cannot reach the preset value. In addition, in the inhalation process, due to the large lung inflow, the pressure drops significantly, and the gas supply effect is weakened. Therefore, how to compensate for the pressure loss has become a problem to be solved for non-invasive ventilation equipment.

[0003] The existing pressure compensation method is to input the collected pressure and flow values into a preset BP neural network model, obtain the pressure compensation value at the current time through the training of the BP neural network model, superimpose the pressure compensation value on the set pressure, and then use PID for pressure control. That is: target pressure value = user set value + compensation pressure value. 2 The defect of this compensation scheme is that the calculation amount is large, the chip requirement is high, the machine cost is increased, and there is no verification process, and the possible error compensation cannot be estimated. SUMMARY

[0005] The purpose of the present application is to overcome the defects.

[0006] In order to achieve the above purpose, the present application provides a pressure compensation method for a non-invasive respirator, which comprises:

[0007] Step 1: under the condition that the breathing pipeline is open, a plurality of flow setting values and corresponding actual measurement values of the machine pressure sensor are collected;

[0008] Step 2: the pressure compensation prediction value is determined by fitting and calculating the model parameters through the least square matrix method;

[0009] Step 3: the goodness of fit R 2 of the model is verified; if the goodness of fit R 2 meets the requirements, the pressure compensation prediction value is the pressure compensation value; otherwise, step 2 is repeated to recalculate;

[0010] Step 4: the pressure compensation value is superimposed on the user pressure set value to realize real-time compensation of the pipeline pressure.

[0011] As an improvement of the above method, the step 1 is specifically:

[0012] With the breathing tubing open, collect N flow setting values ​​{x1, x2, ..., x k ,…,x N The actual measured values ​​of the machine pressure sensor corresponding to {y1,y2,…,y} are as follows: k ,…,y N}

[0013] As an improvement to the above method, step 2 specifically includes:

[0014] Calculate the parameter estimates based on the following system of equations. and

[0015]

[0016] Calculate the predicted pressure compensation value:

[0017]

[0018] in, Q represents the predicted value of pressure compensation. seonsor This indicates that the flow rate value is set.

[0019] As an improvement to the above method, the goodness-of-fit coefficient R 2 The system of equations to be solved is as follows:

[0020]

[0021] in, SS is the average value of the pressure sensor. tot For the total sum of squares, SS reg This is the sum of squares of the regression.

[0022] As an improvement to the above method, the goodness of fit R 2 Meeting the requirements means R 2 >0.9.

[0023] The present invention also provides a ventilator that achieves pressure compensation based on the above method.

[0024] Compared with the prior art, the advantages of the present invention are:

[0025] Existing solutions utilize neural networks for parameter solving, resulting in high computational demands, high requirements for the MCU, and increased machine costs. The method of this invention can estimate pipeline pressure compensation parameters during machine self-testing, significantly reducing computational load and enabling rapid real-time pressure compensation. Furthermore, this invention incorporates a model verification mechanism to prevent harm to users caused by high compensation pressure resulting from erroneous parameter estimation. Attached Figure Description

[0026] Figure 1 A non-invasive ventilator system block diagram is shown;

[0027] Figure 2 A pipe lumped parameter model diagram is shown;

[0028] Figure 3 A pressure compensation method calculation flow chart is shown. DETAILED DESCRIPTION

[0029] The present application proposes a pressure compensation method for non-invasive ventilators based on the least square method, increases a model verification module, ensures that the user obtains better pressure support during the entire use process, and improves the comfort of use.

[0030] The technical solutions of the present application will be described in detail below with reference to the accompanying drawings.

[0031] As shown in Figure 1 , the ventilator is mainly composed of a main machine, a humidification tank and a series of pipes, and the final gas is input to the mask to assist the user to breathe.

[0032] As shown in Figure 2 , the present application introduces a dynamic gas-electric model of a breathing pipe, including an electronic simulator that reacts to the breathing pipe. In the model, C T represents the compliance of the breathing pipe, Q T represents the pipe gas compression flow, which can generally be considered as C T = 0. R pump represents the resistance encountered by the gas flow in the patient's ventilation pipe, Q Sensor represents the flow setting value, P Sensor represents the pressure sensor acquisition value, P aw represents the user's proximal pressure value.

[0033] For the problem of pressure loss at the patient end, the present application proposes a pressure compensation method based on the least square method, which can realize automatic compensation of the pressure of the ventilation equipment and realize accurate control of the pressure during the entire breathing cycle of the user.

[0034] As shown in Figure 3 , the pressure compensation method of the present application includes the following steps: under the condition that the breathing pipe is open, N flow setting values {x1, x2, …, x N} are collected, and the corresponding machine pressure sensor actual measurement values {y1, y2, … y N} are collected, wherein N is a positive integer; according to {x1, x2, …, x N} and {y1, y2, … y N}, the pressure compensation value is determined by fitting the calculation model parameters in the least square matrix mode. The pressure compensation value is superimposed on the user pressure set value to realize real-time compensation of pipeline pressure and real-time control.

[0035] As shown in Figure 2 , R pump in the model is also called parabolic gas flow constant, which satisfies the following relationship:

[0036]

[0037] Since the pipeline is empty, P aw = 0 cmH2O (2)

[0038] In the formula, Q Sensor is the flow set value, P Sensor is the pressure sensor acquisition value, and P aw is the user proximal pressure value.

[0039] From formulas (1) and (2), the flow rate and resistance relationship is:

[0040]

[0041] At this time, P sensor is the pressure compensation value.

[0042] To improve the system fault tolerance and simplify the humidification in the model, the gas flow may generate turbulence, so a nonlinear model needs to be used:

[0043]

[0044] In the formula , a and b are two undetermined parameters, representing the constant value of the pipeline resistance, and the parameter value needs to be determined in the system self-checking process using the least square method for the pressure compensation coefficient.

[0045] For N groups of data (x k , y k ) obtained by the flow set value and the pressure sensor measurement value, k = 1, 2, …, N, the least square fitting is performed according to the nonlinear model for the fluid parabola, that is, to find such that:

[0046]

[0047] is the minimum.

[0048] According to the least square principle, the partial derivative of formula (5) is taken.

[0049]

[0050]

[0051] The equation set can be arranged:

[0052]

[0053] The best estimate of the parameters [k1 k2] can be obtained by solving the equation set

[0054] The goodness of fit R is used for the fitting result 2 The model accuracy is verified, and Substitute equation (4) to get a pressure compensation prediction value based on the model parameters, that is:

[0055]

[0056] The goodness of fit coefficient R 2 The equation set is solved as:

[0057]

[0058] In the formula, The mean value of the pressure sensor is SS tot The total sum of squares is SS reg The regression sum of squares is SS res The residual sum of squares is SS

[0059] The value of R 2 The closer the value of R 2 The smaller the value of R

[0060] If the coefficient R 2 > 0.9, it is considered that the pressure compensation estimation is successful, otherwise it needs to be recalibrated.

[0061] In an actual application process, set the flow value, and collect the pressure sensor value corresponding to it. The experimental sampling record is as follows:

[0062]

[0063] According to the above steps, the pressure compensation fitting calculation of the pipeline resistance is obtained

[0064] The actual use result is as follows:

[0065]

[0066] Among them, the user end pressure is measured by a third party device.

[0067] Finally, it should be noted that the above examples are merely used to illustrate the technical solutions of the present application but not to limit. Although the present application is explained in detail with reference to the examples, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A breathing machine, characterized by, The ventilator uses a pressure compensation method for non-invasive ventilators to compensate for the pipeline pressure, and the method comprises: Step 1: Collect multiple flow setting values and corresponding actual measurement values of the machine pressure sensor under the condition that the breathing pipeline is open; Step 2: Determine the pressure compensation prediction value by fitting and calculating the model parameters through the least square matrix method; Step 3: Use goodness of fit R for pressure compensation prediction value 2 Model accuracy verification is performed; if the goodness of fit R 2 meets the requirements, the pressure compensation prediction value is the pressure compensation value; otherwise, step 2 is repeated for recalculation. Step 4: Superimpose the pressure compensation value on the user pressure setting value to realize real-time compensation of the pipeline pressure; The step 2 is specifically: The parameter estimates are calculated according to the following system of equations and : ； Calculate the pressure compensation prediction value: ； wherein, represents a pressure-compensated prediction value; represents a set flow value; The goodness of fit R 2 The system of equations to be solved is: ； wherein, is the pressure sensor mean, is the total sum of squares, is the regression sum of squares; The step 1 is specifically: N number of flow set values are collected with the breathing circuit open The corresponding machine pressure sensor actual measurement values .

2. The ventilator of claim 1, wherein, The goodness of fit R 2 Fits the requirements mean that R 2 > 0.9.

Images