30 November 1998
F) temperature would have occurred with proportional control, at point A. Similarly, when the outside air
temperature falls to 8 degrees C (47 degrees F), the integral mode component prevents the valve from
opening until the outside air temperature falls below 7 degrees C (45 degrees F). The proportional only
mode controller would have begun to open the valve at 8 degrees C (47 degrees F) outside air temperature
at point B. From the time the valve begins to open at the outside air temperature of 7 degrees C (45
degrees F) until the fifth hour when the outside air temperature is -4 degrees C (25 degrees F) the integral
mode component of the controller output signal becomes smaller until at the fifth hour the integral mode
component becomes zero. This additional pressure on the normally-open valve allows the coil air
discharge temperature to remain at the 7 degrees C (45 degrees F) setpoint instead of controlling between
8 and 7 degrees C (47 and 45 degrees F) when the outside air temperature is between 7 and -4 degrees C
(45 and 25 degrees F) as would have occurred with proportional only control. Similarly, when the outside
air temperature falls below -4 degrees C (25 degrees F), the integral mode component is subtracted from
the proportional mode component of the controller output signal to open the valve enough to keep the
discharge air temperature at 7 degrees C (45 degrees F) rather than at 6 degrees C (43 degrees F) as
would have occurred with proportional mode control.
Figure 2-14. Proportional plus integral control mode.
e. Effects of rapid load changes. The rate of change of load imposed on the HVAC system by the
process affects how well the controller will perform its task of controlling at setpoint. The temperature of
the outside air changes relatively slowly, and the temperature conditions inside also require some time to
change. Inside conditions change as a function of air temperature changes made by the HVAC system,
which warm up or cool down masses of material within the building. Inside conditions also change as
functions of lighting load and occupancy. Because of these relatively slow rates of change, most of the
HVAC processes that require gradual controller output changes can be controlled quite well with
proportional plus integral (PI) control modes. Except for lighting loads on the HVAC system, these variable
changes are relatively slow compared to the rates of change of variables that affect some non-HVAC
processes. Lighting loads are sometimes imposed on the HVAC system quickly. This is an example of a
step change in the process variable. The combined actions of proportional and integral modes are not
always adequate to control rapidly changing variables.
f. Proportional-integral-derivative (PID) mode.
(1) Some processes require a controller that can respond to rapidly changing process variables.
One answer to control of rapidly changing processes has been the addition of another control mode called
derivative mode. When this control mode is added to proportional-integral control, the combination is
known as proportional-integral-derivative (PID) control mode. The PID control mode adds a component
algebraically to the output signal; this component is proportional to the rate of change of the error between
the control point and setpoint. This automatic adjustment also affects the proportional and integral output
components in a manner analogous to the way in which the integral component affects the proportional
component. As the valve actuator stroke position changes, the temperature changes as a result of
changing flow through the valve, and the rate of error signal change between the control point and setpoint
varies as the control point comes closer to setpoint.
(2) There are a few HVAC control applications that are difficult for either P-mode or PI-mode
control because of the fast rates of change of the process variables. One such application is the control of
I-mode component of PI-mode takes care of the varying range of offsets due to loads that occur in
domestic hot water heating applications. For example, high-rise residential buildings have morning and
evening peak periods of demand for hot water use. These peak demand periods drive the domestic hot
water temperature to the low end of the offset range. Periods of relative nonoccupancy, such as late
morning and early afternoon, drive the temperature to the high end of the offset range due to the minimal
demand for domestic hot water use. Periods of relative inactivity during occupancy, such as late evening