Investigation of Insulation Performance of High Voltage Rotating Machines Using Intelligent Techniques
Investigation of Insulation Performance of High Voltage Rotating Machines Using Intelligent Techniques
1.
The condition of insulation performance of a high voltage rotating machine has been
analyzed using Back Propagation Neural Network (BPN), Fuzzy logic and BPN network
with Proportional Integral and Derivative (PID) controller concept along with slope
parameter. The economic aspects of its periodic maintenance, the loss in production due
to breakdown, the investment cost due to frequent replacement with a new machine is
huge. The insulation condition of the machine can be assessed by measuring the various
parameters like capacitance, leakage current, dissipation factor, polarization index, surge
voltage withstanding strength and partial discharge magnitude. To assess these
parameters of the insulations used in high voltage rotating machines, a number of
measurements have been made on actual stator coils of machines. For various test
voltages, the capacitance, leakage current and dissipation factor are measured and
correlated as a function of test voltages. It is observed from the simulation results that
BPN with PID techniques gives accurate results, predicting the performance of stator
winding insulation.
2.
This paper presents a low cost, digitally controlled, closed loop BLDC motor drive.
Normally encoder feedback is required to get a speed feedback between two hall
sensor status changes, for robust controller designing. With a view to reduce the
cost, in this design an encoder feedback is avoided and moving average speed
estimation with variable Kp-Ki PI controller is realised. The digital controller Cortex
M3 ARM microcontroller LM3S2616 is implemented with only hall-sensors feedback
and variable Kp-Ki PI controller algorithm. The experimental setup is verified by
loading motor from no load to full load as well as with continuous and impact load.
Serially, motor parameters like speed, current are gathered on PC and analysed,
they confirmed the simulated results.
type relationships for alternating stress versus cycles to failure have been developed.
Based on these relationships, the surface treatment modification factors in terms of
power law relationships have been evolved.