Identification of a nonlinear inertial measuring pressure channel

Abstract

The method of nonlinear inertial measuring systems identification is considered in the article on the example of a measuring pressure channel, which consists of a measuring line and a pressure sensor. The latter, as a rule, is a linear system, and the measuring line, which for the pressure channel is a pipe filled with liquid or gas, is projected as a linear system. During the operation, the measuring line gradually becomes contaminated, the liquid partially freezes or air enters it. In this case, the loss of the pressure channel linearity is possible, which in practice, is usually not measured. The proposed method allows to obtain mathematical channel model on the basis of measured input and output signals. It is based on the well-known model, which divides the channel into two virtual blocks, one of which is a nonlinear inertial, and the other is linear inertial. The conversion function of the first block is described by a polynomial of finite degree, and the transforming properties of the second block are described by the convolution integral. This allows to record the signal at the output of the channel in the form of mathematical dependence on the unknown coefficients of the polynomial and parameters of the pulse characteristic of the linear block. Then the distance in the functional space with the quadratic metric between this mathematical dependence and the output signal is minimized. As a result of minimization by the method of the global random search, the polynomial coefficients and the parameters of the impulse response are calculated, which means that the solution of the system identification problem is determined. The received channel model was used for the theoretical determination of the output signal under the already known input signal. The verification of the model based on Fisher’s criterion with a significance level of 0.95 confirmed its adequacy.