Prototype Fast Breeder Reactor (PFBR) has 8-Steam Generators(SG) to produce superheated
steam using the heat produced by the reactor core to run the turbines for generating the electrical
power. Unlike the other conventional heat exchangers, PFBR SGs use Sodium in the primary circuit
and water in the secondary circuit for transferring heat to produce the steam. Any breach in the SG
tubes will lead to a violent exothermic reaction and local explosion of SG. To maintain safety of
plant, proper inspection campaigns are conducted to ascertain the healthiness of the tubes. SG
tubes inspections are carried out using Remote Field Eddy Current (RFEC) probe, which needs to be
positioned in the exact centre of each tube and pushed into the entire length (23 meters) of steam
tube to collect the probe data to detect any flaw or degradation of tubes.
In order to automate thisinspection process, in this research work a PFBR Steam Generation
Inspection System (PSGIS) has been developed that consists a 2-DoF robot with the necessary
inspection probe cable mechanism attached to its end effector. The system also has imaging system,
which capture’s SG tube sheet image, processes it to identify the individual tubes and guides the
robot end effector (inspection probe) to the centre of each tube using vision servoing technique for
pushing the probe into the tube to detect the tube flaws. This seminar talk covers the development
of 2-DoF robotic arm, and its kinematic studies carried out to develop the robotic manipulator
control algorithm to reach tubes centre for inspection with accuracy. The robot arm manipulability
is studied to arrive the optimal damping factor, which is used in the Damped Least Squares Inverse
kinematics (DLSIK) for accurate positioning of the inspection probe. Also, this optimal damping
factor helps in reducing tracking error and maintain good numerical stability in the convergence of
IK algorithm. A rigid body transformation parameter estimation is also done to calibrate the robotic
manipulator for the precision reach of the tubes in the task space. An optimal inspection sequence
for the random inspection of tubes through Genetic Algorithm (GA) is also carried out.