ANN Based Flow Velocity Measurement Scheme Using Constant Temperature Anemometer (CTA) With Temperature Correction

Debangshu Dey and Sugata Munshi[1]

Department of Electrical Engineering. , Jadavpur University Kolkata-700032, India.
[1] Corresponding author;
The transfer curve of Constant Temperature Anemometer (CTA) is highly nonlinear and the output voltage is also cross sensitive to fluid temperature. In this work a signal processing arrangement employing neural computation has been devised that performs linearization as well as compensation of temperature error, thereby predicting the fluid flow velocity from the measured value of bridge error voltage. During training, the Artificial Neural Network (ANN) is initialized with Differential Evolution (DE) algorithm and then trained with conventional Levenberg-Marquadt algorithm. Simulation results for the velocity measurement of air show very good performance. The scheme is even superior to earlier schemes, which used ANNs only for linearization, and the problem of temperature variation of fluid was avoided. In that respect this scheme is able to perform more complex jobs than previous schemes and still it yields better results, so far as the full scale RMS error of the fluid velocity measurement is concerned. An attractive feature of the present scheme is the use of a low-cost sensor (thermistor) for sensing the fluid temperature.

To read the full article,
please log in:
If you haven't registered already,
you can do so for free:
Copyright 2000-2010 Professor F.R. Hall & Dr I. Oraifige, University of Wolverhampton.
close this window