SELECTION OF THE SENSOR CONVERSION CHARACTERISTIC MODEL FOR CONTROLLING THE ERROR IN THE MEASUREMENT OF PHYSICAL QUANTITIES

Abstract

On the example of a pressure sensor, the problem of selecting a model and parameters of
the conversion function of a microprocessor sensor is considered. The conversion function is
based on a mathematical model that associates the electrical signal coming from the sensor's
measuring transducer with the value of a physical quantity. The model of the conversion function
of a microprocessor sensor must repeat the real spatial dependence of the electrical signal on the
measured value and take into account the influence of external factors, such as temperature. Microprocessor
sensors are used to measure the parameters of an object with a given accuracy. The
main contribution to the measurement error is made by the inaccuracy of the approximation of the
real transformation function by its model. The need to achieve the optimal level of parameter
measurement error in the system, taking into account the complexity and cost of measurements,
requires the control of the sensor error. For this purpose, various models and methods of approximation
are presented. For efficient error control, a method of multi-segment spatial approximation
based on models of linear or non-linear spatial elements is proposed. The error control procedure
is formulated. The procedure for using the model of multi-segment spatial approximation
of the transformation characteristic for pressure calculations taking into account the influence of
temperature is based on the combined use of linear and non-linear spatial elements within the
same model. The segment type selection procedure should begin with an assessment of the possibility
of using a linear spatial element first, and if it is impossible to meet the accuracy requirements,
an analysis of the use of a non-linear element. The method allows you to change the types
and configuration of spatial elements and in this way influence the measurement error. The advantages
of this approach are confirmed by the simulation results.