Page 389 - Boiler_Operators_Handbook,_Second_Edition

374 Boiler Operator’s Handbook

set bellows and the controller will be satisfied as long as error times gain. The output changes when the error
the setpoint and process variable pressures are the same. changes. Integral is the reset adjustment and it repeats
Unlike proportional control we don’t need a deviation in the error multiplied by the value of the integral. Notice
the process variable to get the required output. that the reset effect is the error repeated; an integral
It’s reset control that keeps the boiler level right adjustment is normally marked to indicate repeats per
at the center of the glass while changing the feedwater minute, meaning that is how many times the error will
control valve position from closed at low loads to al- be repeated in one minute. That doesn’t mean the con-
most wide open at high loads. Its also reset control that troller only repeats the error for a minute either; it con-
makes it possible to keep the steam header pressure at tinues repeating the error every minute. It also doesn’t
120 psig whether we’re at low fire or high fire and even repeat it at the end of a minute. If the integral is set at 60
when we’re running five boilers instead of one. It’s reset repeats per minute it will increase or decrease the out-
control that allows us to run air/fuel ratios so tight that put by value equal to the error every second.
oxygen in the flue gas can be held at one half percent. A proper combination of gain (proportional con-
Tuning a reset controller is nowhere near as easy trol) and integral (reset control) will make the process
as tuning a proportional controller but the additional return to the setpoint quickly and smoothly. Now that
feature of the controller (you have P + I, proportional you understand the way the controller operates you
plus integral) allows you more flexibility in matching should have a better idea of which adjustment to use
the process. Aw shucks, another fancy word! Integral and which direction to turn it, a big step in tuning a con-
is a mathematical term that sort of means accumulat- troller.
ing the average value. It’s not important to understand Adjusting the gain of a ratio totalizer was a lot
mathematical integrals, only that it’s another name for more complicated than adjusting it on a modern con-
reset. troller. You had to release two set screws that held the
Tuning consists of changing the gain (proportional pivot spring to the base and the beam then slide the
control) and reset (integral control) until the combina- pivot spring assembly to a new location and tighten
tion provides a response to an upset in process condi- the screws. While you did that the output was always
tions where the process variable returns to setpoint jumping all over as you handled the parts and turned
within a short period of time and with only a little the set screws so you didn’t have any idea of what the
overshoot in response to the initial error. You’ve prob- results would be until you got your hands off it. Gain
ably seen the curve in another book, where the error is on modern controllers can be adjusted without affecting
plotted versus time, it starts as a big error with a rapid the output except for the difference in the gain (times the
change in process variable quickly approaching the set- error). Increase the gain and the output changes more
point, overshooting it a bit, then turning back toward for a given error. Just to make sure you understand it,
the setpoint and falling in line with it. It’s a pretty pic- the error is the difference between setpoint and process
ture but making it do that in the real world can be damn variable, what you want and what you’ve got.
difficult at times. Adjusting the integral of a ratio totalizer didn’t up-
On several occasions I’ve run into a reset controller set the operation so much because adjusting the needle
that had all reset blocked out (like completely closing valve didn’t have the effect that grabbing the beam to
the needle valve) because an operator didn’t understand adjust gain did. It was more like a modern controller.
reset control adjustments. Keep in mind that a simple If you opened up the needle valve you increased the
proportional controller requires an error to do its job and repeats per minute because the air could flow through
you’ll find that attempts to minimize that error result in (adding air to or bleeding air from the volume chamber
some pretty wild swings in the output of the controller; and reset bellows) faster. Closing down on the valve de-
what we call instability. Those swings are primarily as- creased the repeats per minute. Closing the needle valve
sociated with the fact that the process doesn’t respond entirely made for a pretty sloppy controller because the
instantly to changes in the controller output. It can take proportional part had to compensate for changing the
a few hundredths of a second to several seconds before compression of the air in the reset bellows and volume
the full effect of a change in controller output is appar- chamber as well.
ent by looking at the process variable. As for tuning the controller, you adjust the gain or
You tune a reset controller to deal with those de- reset to balance the system response. If a change in con-
lays. The controller will have two adjustments, gain and troller output produces an almost instantaneous change
integral. Gain is the proportional part, the output is the in process variable then most of the control function can