Neural network‐based target‐type flowmeter for high‐velocity liquid flow measurement

Abstract

Target flowmeter is commonly used for measuring flow rates, but there are some shortcomings of the conventional target flowmeter such as complex structure, low sensitivity, non-linearity. Here, a reluctance-type target flowmeter has been developed which converts the flow rates into an electrical signal using a modified Maxwell–Wien bridge. Two ferromagnetic cores are used in which one is attached to the target disc and in other core, coils are winded. These two cores along with coils convert the displacement of the target disc into a change in inductance. The change in inductance is converted into an electrical signal using a modified bridge. The electrical output is non-linear with respect to the flow rate. The electrical output is linearised with the help of a neural network technique. This minimises the non-linearity error up to 0.27% of full scale output (FSO). This research study includes a theoretical analysis of the proposed flow transmitter. This sensor has been designed and tested for a 0–2000 liters per hour (LPH) range in the laboratory. Experimental studies revealed that the transmitter under investigation possessed linear characteristics and obeyed the theoretical equations.