Respirator valve assembly modeling method and system based on neural network fitting, computer readable storage medium and respirator

The invention relates to the technical field of breathing machines, in particular to a breathing machine valve assembly modeling method and system based on neural network fitting, a computer readable storage medium and a breathing machine. The method comprises the following steps: step 1, obtaining static experiment data of the valve assembly through testing; 2, dynamic experiment data of the valve assembly are obtained through testing; 3, on the basis of the static experiment data, a static model is constructed through a neural network fitting method so as to reflect the nonlinear relation among the input control voltage, the input pressure and the output flow; 4, on the basis of the dynamic experiment data, assuming an ordinary differential equation structure representing a dynamic model, and obtaining the dynamic model reflecting dynamic response data of flow between different adjacent test points by utilizing a linear system identification algorithm; and 5, limiting the working range of the static model b