CARVALHO, E. A. N.; http://lattes.cnpq.br/9031123002753606; CARVALHO, Elyson Adan Nunes.
Resumen:
In this work a system to perform sensory positioning for automatic inspection
of the fillet welds found on fuel tanks is implemented and tested. To be able to
do this, the system need to have features as a high speed of operation,
accuracy and precision, with low cost. Moreover, the processes of manufacture
and maintenance of the system should be simple.
The fillet weld identification is done with the information from a linear array of
15 optical-reflexive distance sensors. The sensor type chosen was
characterized and it was developed a circuit for data acquisition using a
microcontroller and serial communication. To process the data acquisition
system signal output, the Sinc Interpolation and the Information and α-β filters
were used, aiming to a greater reliability, accuracy and precision of the system.
A proportional control system was used to correct the linear and angular
positions of the sensory platform. A mechanical platform was also developed,
joining characteristics like simple maintenance and manufacture, as well as high
accuracy, speed of operation and durability.
Simulation and experimental tests are presented for the proposed system.
Also, preliminary development of an inductive distance sensor and the use of
image processing to perform fillet weld identification are proposed as future
work.