Selected Types of Neural Networks for Magnetoelastic Sensor Error Suppression

The main advantage of magnetoelastic sensors in comparison with the classic wire ones or semiconductor ones is their marked sensitivity and higher resistance to environmental moisture. These properties predestinate them for use in civil engineering and geo-technological applications [2]. Magnetoelastic sensors of pressure force are currently based on the change of core permeability caused by mechanical stress, while altering its magnetization. The sensor works as a transformer with a variable coefficient of transformation. Under the action of an external force the output voltage changes. Some shortcomings of magnetoelastic sensors are higher power consumption, noticeable sensor errors, eghysteresis and nonlinearity [6]. At present, the requirements for accuracy and reliability of sensor measuring systems are getting higher. The total accuracy of standard measuring system can be significantly improved by adding a data conditioning block. The aim of the paper is to present selected types of neural networks as an effective tool for errors suppression.