Dynamic Neural Network Calibration of Quartz Transducers

Abstract

Abstract Quartz pressure transducers are plagued by degraded performance when subjected to thermal transients. In this paper, two new methods of calibrating quartz transducers in which a neural network is used to accomplish a dynamic calibration of a transducer pair for a range of temperatures and pressures will be described. When a quartz pressure transducer is exposed to an environmental temperature change, thermal stress is induced in the sensing element. This thermal stress causes erroneous pressure readings until thermal equilibrium is achieved within the transducer. In order to compensate for the undesirable transient thermal effects on the pressure transducer performance, multiple past and future temperature- and pressure-sensor outputs are used as inputs to a calibrating neural network. These new methods allow the time-dependent qualities of the quartz-oscillator transducer to be included in the calibration scheme. Generally, these methods can be applied to other types of transducers subject to transient environmental effects and are not limited to quartz-oscillator transducers. A theoretical discussion of this technology is presented, followed by experimental results obtained by applying both methods to quartz pressure-transducer calibration.