Synthesis of smart sensors based on the introduction of minimal structural redundancy

Abstract

The problems of operation of measuring systems and changes in their parameters due the influence of various kinds of destabilizing factors, including time, are considered, and the primary elements of the measuring path – sensors and converting equipment – are most susceptible to such an influence. The deviation of the values of the parameters of the measurers from the nominal values leads to significant errors in the assessment of the unknown input signal, which necessitates the current (during operation) identification of the measurer. The problem of the current identification of the measurer parameters under unknown input influences is solved. The identification procedure is carried out by introducing an additional channel for transforming the measured value in the spatial domain. In this case, the transformation of the measured value in the additional channel of the structurally redundant measurer has the form of a preliminary functional (nonlinear) transformation of the m-th power over the input signal. The solution to the current identification problem is considered for the linear static characteristic of the main channel of the measuring transducer. An additional equation in solving the problem of current identification and conducting metrological self-control in an intelligent sensor is presented in the form of a regression relationship between the output signals of the additional and main channels of the structurally redundant measurer. The unknown parameters of the regression equation are determined from the results of processing the time samples of the output signals of the additional and main channels using the least squares method. The dependences of the rms measurement error of the input quantity on the rms value of the measurement noise in the output signals of the structurally redundant measurer, the dynamics of the input signal change and the power of nonlinearity m of the preliminary functional transformation in the additional measuring channel are presented. In the course of research, it was revealed that the shape and spectrum of the input signal do not significantly affect the measurement accuracy. It is shown that the highest measurement accuracy is provided by the preliminary quadratic conversion of the input signal in the additional channel of the minimum-redundant sensor. The research results can be used for metrological selfcontrol in smart sensors or smart measuring systems.