Вы находитесь на странице: 1из 4

ASHRAE Journal

Fundamentals of DDC
By Rob Moult VDC) or current (typically 0–20 mA).
• Digital inputs. An example of a digital input is a flow switch.

T
Direct
his article provides an introduction to direct digital con-
trol (DDC) by discussing “direct,” “digital” and “con-
trol” in their own sections.
Digital inputs sense a discrete input (on/off). Digital in-
puts are isolated so that induced voltages do not damage
the DDC controller. Some digital inputs may be able to
accumulate pulse streams from devices such as power
meters or flow meters.
Until about 20 years ago, almost all HVAC controls installed • Digital outputs. An example of a digital output is a relay
in large commercial buildings in North America were pneumatic. used to start or stop a piece of equipment.
In pneumatic controls, a remote-sensing device, such as a tem-
perature transmitter using a bimetallic strip and a leak port, Microprocessor
varied the air pressure sent to a pneumatic receiver-controller. The microprocessor is the “brain” of the DDC controller, co-
The receiver-controller then sent a pneumatic signal to an out- ordinating the activities. The microprocessor has a very simple
put device such as a valve or damper actuator. These three operating system, which either loops through the following
devices (sensing device, receiver-controller and output device) functions at a periodic rate (typically about once per second) or
formed a simple, stand-alone control loop. A dual input receiver- upon the occurrence of input/output changes (called event-
controller was available that allowed the setpoint of the control driven):
loop to be varied based on a pneumatic signal. • Processing inputs and outputs.
The first generation building automation system used this fea- • Passing data to and from the communications circuitry.
ture to save energy through setpoint reset. As shown in Figure • Updating user interface (if present).
1, the central minicomputer of the building automation system • Executing control sequences.
would decide that the setpoint of a particular control loop should
Processing inputs and outputs involves the following func-
be increased or decreased. The minicomputer would then in-
tions:
struct a field panel to adjust a voltage applied to an electric-
• Ranging analog inputs to the correct engineering units.
pressure transducer, which in turn would change the pressure
For example, the output from the analog to digital con-
signal applied to the setpoint input of the receiver-controller.
verter, which is proportional to the resistance of the tem-
Figure 1 shows that the building automation system was
perature sensor, is converted into degrees Fahrenheit or
unable to directly control the output device. The building auto-
degrees Celsius. As another example, the pulses from an
mation system only was able to control the output device indi-
accumulator are converted into a flow reading using a
rectly, by changing the setpoint of the receiver-controller. The
scaling constant. Ranging includes linearizing standard
DDC controller, introduced about 20 years ago, replaces the
sensor types based on pre-defined sensor curves and
function of the field panel, electric pressure transducer and
converting velocity pressure into an airflow reading.
receiver-controller. A DDC controller allows the building auto-
• Filtering of analog input signals. Selected analog inputs
mation system to control the output device directly. This is the
will be smoothed1 so that they are less sensitive to instan-
origin of the word “direct” in “direct digital control.”
taneous fluctuations. This is important if the analog input
is to be used in a control loop. Filtering is typically used
Digital
on inputs subject to instantaneous fluctuations, such as
Figure 2 shows the internal architecture of a typical DDC
pressure.
controller.
An engineer uses a personal computer to generate the point
database for the DDC controller, including ranging, filtering
Inputs/Outputs
and automatic interlock parameters. The PC software also is
The inputs/outputs portion of a DDC controller is respon-
used to create the control sequence to be executed. The three
sible for interfacing to the following types of signals:
techniques used to create control sequences are:
• Analog inputs. An example of an analog input is a tem-
• Question and Answer. The engineer enters responses to a
perature sensor. The DDC controller converts the raw ana-
log signal of voltage, current or resistance into a digital
value for processing.
• Analog outputs. An example of an analog output is a sig- About the Author
nal sent to control a valve actuator. The DDC controller Rob Moult is director-Asia Pacific operations with Johnson
converts the digital value into a voltage (typically 0–10 Controls.
November 2000 ASHRAE Journal 19
ASHRAE Journal
series of questions and an expert system creates a detailed
sequence of operation. This approach has the least flexibil-
ity but delivers the most reliable sequence of operation.
• Graphical Programming. The engineer “draws the con-
trol flow chart” of the desired sequence of operation on a
PC. This approach can be quite intuitive, but can become
unwieldy when a complex sequence of operation is re-
quired. The graphical editor used to create the sequence
of operation often comes equipped with a simulator for
troubleshooting.
• Text Programming. The engineer develops the sequence
from scratch or using library routines. This approach pro-
vides the most flexibility, but is also the most prone to
errors. This editor often comes with a debugger for trouble-
shooting.

Memory
A typical DDC controller uses FLASH memory for the oper-
ating system and configuration parameters and uses RAM for Figure 1: Indirect control of a pneumatic control loop.
its working “scratchpad.” If the DDC controller has an on-
board battery, configuration parameters may be stored in RAM User Interface
rather than in non-volatile FLASH memory. RAM also is used The features and type of user interface available (LEDs, LED
typically to store historical information such as trending and display or LCD display) depend on the intended application of
temporary schedules. the DDC controller. DDC controllers that are not routinely ac-

Advertisement for the print edition


formerly in this space.

20 ASHRAE Journal w w w. a s h r a e j o u r n a l . o r g November 2000


Control Equipment

Figure 3: Matching the valve to the coil means a linear


system.

Figure 2: Internal architecture of a typical DDC controller. Traditionally, the hardware used in the communications por-
tion of the DDC controller has been EIA485 circuitry. In recent
cessed, such as those mounted on VAV boxes, do not have on- years, there has been some demand for alternative hardware
board user interfaces. This type of DDC controller typically options including traditional LAN technologies. As the cost of
provides a connection point for a hand-held device or a laptop Ethernet circuitry continues to drop, it is possible that DDC
computer. In most cases, the user interface allows access to controllers of the future, like PCs, will standardize on Ethernet
local information and a limited amount of “global” information. communications.
The communications protocol used by DDC controllers has
Communications been evolving from vendor-specific protocols to standard pro-
The role of the communications portion of the DDC control- tocols such as BACnet® and LonMark. As this trend contin-
ler is to allow the microprocessor to send data and receive ues, it will become easier to have DDC controllers from different
commands to and from the communications bus. vendors sharing data on the same communications bus.

Power Supply
Most DDC controllers are powered from 24VAC transformers.
The power supply in the DDC controller supports the on-board
functions and provides power to drive binary outputs. Some
DDC controllers are equipped with on-board batteries that pro-
tect memory in the event of a loss of power. The on-board bat-
tery does not maintain control. A universal power supply (UPS)
must be provided if the DDC controller is to continue to control
during a power failure. DDC controllers that are not routinely
accessed, such as those mounted on VAV boxes, typically do
not have batteries because of the maintenance issue.

Control
Advertisement for the print edition Today, most DDC controllers use a simple proportional plus
integral (PI) control algorithm. When controls were mechanical,
formerly in this space.
pneumatic or electronic, this was the only choice as it was
difficult to implement a more advanced algorithm. Since their
inception, DDC controllers have been using a PI algorithm be-
cause it was proven, computationally efficient and required a
minimal amount of memory. Some DDC controllers also have a
derivative term to implement a proportional plus integral plus
derivative (PID) control algorithm, but PID is rarely used for
HVAC applications because the derivative term makes the con-
trol algorithm overly sensitive to instantaneous fluctuations.
A PI control algorithm assumes a linear relationship between
the controlled variable and the output. This is rarely true for
HVAC applications. Let us consider the control of a heating or
cooling coil. The relationship between water flow and the heat-
ing or cooling capacity of a coil is non-linear. As shown in Figure
3, the solution is to use a valve plug with an equal percentage
characteristic so that the combined system (valve plus coil) has
November 2000 ASHRAE Journal 21
ASHRAE Journal
an overall linear characteristic. ergy inefficient), or a single PI loop serv-
In practice, it is not possible to perfectly ing multiple outputs through spring range
match the valve with the coil. As a result, selection (further complicating the linear-
the characteristic of the combined system ity issue). DDC is very good at computa-
is not perfectly linear. This effect is not tional logic. The way that we think of con-
significant over a limited operating range. trol through DDC should be (and is
However, the optimal tuning parameters gradually becoming) different than how
when operating at one end of the range we thought of control when we used
can be significantly different from the op- pneumatics. This opens the door, for ex-
timal tuning parameters when operating at ample, to finite state machine models and
the other end of the range. This is why hybrid control (Seem, 1999).
many temperature control loops have to
be re-tuned during season changes. Final Words
Other applications of control loops in In the 1980s, the main question sur-
HVAC are also non-linear. These include: rounding DDC was, “can the extra cost
be justified?” In the 1990s, the focus
• Damper position vs. airflow for VAV
changed to interoperability through stan-
boxes.
dard protocols. Slowly, this issue is start-
• Damper position vs. mixed air tem-
ing to be resolved. It is the author’s opin-
perature for air-handling units.
ion that the next decade will see us revis-
• Fan speed vs. duct static pressure
iting control algorithms to achieve break-
for VAV air-handling units.
through performance improvements. In
• Inlet guide vane position vs. duct
the past decade, the cost of processing
static pressure for VAV air-handling Advertisement for the print edition
power and the cost of memory have
units.
dropped dramatically. The sophistication formerly in this space.
The non-linearity of HVAC processes of HVAC simulation tools has increased
is well known. The industry has grown to significantly, making it possible to de-
accept the limitations of PI control algo- velop more advanced control algorithms.
rithms—sluggish response and frequent It is therefore timely and appropriate for
re-tuning. Recently published papers and the ASHRAE Journal to feature this se-
books (see bibliography) are looking at ries of articles on applications of controls.
new approaches to address this issue.
The VAV box damper actuator control Notes
loop is a very common control loop in the 1. A first-order exponential filter is typi-
HVAC industry. Virtually all VAV boxes use cally used of the form where n is a smooth-
an incremental (also known as position ing constant between zero and one.
adjust) actuator in this application. It is 2. The desired airflow setpoint is deter-
common practice to apply a PI control al- mined based on the actual space tempera-
gorithm to control the damper to achieve ture and temperature setpoint.
a desired airflow setpoint.2 This is unfor-
tunate, as the input signal (calculated air- Bibliography
flow) is prone to rapid changes. Each time 1. Åström, K.J. and T. Hägglund. 1995.
any VAV box changes damper position, PID Controllers: Theory, Design and
all other VAV boxes connected to the Tuning, 2nd edition, Instrument Society of
same AHU immediately feel the impact. America, Research Triangle Park, NC.
To ensure stability, the flow control loop 2. Seem, J.E. 1998. “A new pattern rec-
requires a relatively large deadband of 50 ognition adaptive controller with appli-
cfm (24 L/s) or greater. A large deadband cation to HVAC systems.” Automatica,
will decrease the accuracy of control and 34(8):969–982.
slow the responsiveness of the loop. 3. Seem, J.E., C. Park, and J. M. House.
Again, the industry is looking at practical 1999. “A new sequencing control strat-
ways to address this issue. egy for air-handling units.” International
Pneumatic controls had a difficult time Journal of Heating, Ventilating, Air-Con-
effectively doing complex logic, so most ditioning and Refrigerating Research,
sequences of operation were limited to 5(1):35–58.
having multiple PI loops controlling a 4. Tan, K.K., et al. 1999. Advances in
piece of equipment independently (en- PID Control, Springer-Verlag, NY. (Circle No. 28 on Reader Service Card)

November 2000 ASHRAE Journal 22

Вам также может понравиться