TY  - JOUR
T1  - Calibration Using Artificial Neural Networks
AU - , H. Vasquez AU - , D.J. Fonseca 
JO  - Journal of Engineering and Applied Sciences
VL  - 1
IS  - 1
SP  - 1
EP  - 6
PY  - 2006
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2006.1.6
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2006.1.6
KW  - Artificial neural networks
KW  -load cell calibration
KW  -wheatstone bridge
KW  -levenberg-marquardt optimization
AB  - This study discusses the design and development of an Artificial Neural network (ANN) model to monitor the force applied to a strain-gage load cell.  The reference voltage applied to a Wheatstone bridge formed by the strain gages, the amplification of the Wheatstone bridge’s output voltage and the digitized value of the amplifier’s output voltage acquired by a microprocessor represented the input to the ANN model.  The output of the ANN was defined as the estimated value of the load acting on the load cell.  In this study, a 5-3-1 neural network architecture proved to yield the best results, being the backpropagation Levenberg-Marquardt optimization algorithm the selected training paradigm.  Based on the results obtained, it was concluded that neural networks offer a good option to calibrate an instrument, equipment, or system that operates under variable input conditions.
ER  -