Page 77 - Handbook of Electrical Engineering

56 HANDBOOK OF ELECTRICAL ENGINEERING

2.6.1.4 Power amplifier

Power amplification is necessary in order to develop sufficient power to drive the fuel value open or
closed. The amplifier incorporates,

• The droop constant K d1 .
• The lag term time constants T g1 and T g2 which are inherently present in the electronic circuits.
• The derivative damping gain K g2 which is often made adjustable.

2.6.1.5 Governor compensation
In order to improve the speed of response a lag-lead compensation circuit is employed in some
governor control systems. It contains a gain term K g3 , a lag time constant T g4 and a lead time constant
T g3 . If data are not available for these they may be assumed to be K g3 = 1.0and T g3 = T g4 = 0.

2.6.1.6 Fuel valve mechanism lag
The fuel valve actuator and its mechanism may have sufficient inductance or inertia to introduce a
perceptible lag in the valve stem response to its input signal. The equivalent time constant is T f 1 .

2.6.1.7 Fuel valve limits
The fuel valve naturally has an upper and lower physical limit of the ‘hard’ type, i.e. a limit that is
suddenly reached by the moving part. (A ‘soft’ limit is one in which the moving part reaches a region
of increasing resistance before it eventually comes to rest. An electrical analogy would be magnetic
saturation in an exciter, see sub-section 4.2.) The two hard limits are f min and f max where f min is
usually set at zero. Occasionally f min has a negative value to artificially account for the no-load
turbine power needed to drive the compressor. Hence at no load on the gas-turbine coupling the
valve would be represented as having its position set to zero, whereas in practice it would open to
about 15% of its travel.
Some fuel valves are driven by constant speed servomechanisms such as stepper motors.
When they move the stem from one position to another the initial acceleration to constant speed is
rapid, and likewise when the final position is reached. Feedback is applied in the valve controller to
accurately relate the stem position to the magnitude of the control signal. Often this type of device is
not modelled in computer programs, and so some form of approximation should be used to account
for the lag in time between the receipt of the signal and the valve stem reaching its correct position.
The constant speed motion of the valve actuator is also called ‘slewing’ and the ‘slewing rate’ is the
measure of the rate of change of position during the constant speed motion.
An exponential approximation of slewing is now considered. Assume that the valve can move
from its zero position to its 100% position in T 100 seconds, at a constant rate, when a step input signal
is applied at t = 0 seconds. Assume that an equivalent exponential lag term responds to the same
step input over the same period of T 100 seconds. Figure 2.18 shows the two responses referred to a
common base of time. A good ‘measure of fit’ can be made by choosing the time constant T fa such
that the area represented by the lower part (A) equals that represented by the upper area (B). This
is determined by equating these two areas. The areas are found by integration. Area (A) is found by