wilamowski-b-m-irwin-j-d-industrial-communication-systems-2011

Building and Home Automation 26-3
26.2.2 Distributed Functions
Functional buildings often reach considerable spatial dimensions. At the same time, in each building
and probably in each building part different requirements are made on building services. New technologies
have also enabled new services. The original, centralized approach of DDC stations has reached its
limits. The I/O points of a DDC station typically controlled a single zone of a specific building service.
The higher number of devices today also requires a considerably higher level of networking between
sensors and actuators.
The building control network can be modeled as a directional graph of data points. A data point is
a node in that graph and represents a single, physical value of the system, for example, a temperature.
The data point is an object that contains this physical value as a present value and a number of metadata
describing the context of the value, such as engineering unit, resolution, and minimum and maximum
value. The present value is communicated along the links of the graph according to its direction implied
by input and output data points. In this model, sensors are data sources, controllers are data processors,
and actuators are data sinks. The metadata can be used by engineering and commissioning tools
to identify data points in the building and to establish the correct communication links. This process is
referred to as binding.
The basic purpose of data points is to represent typed values that may be communicated to implement
distributed control. Thus, automation functions in BAS operate on data points. Frequently used
functions are trending, alarming, and scheduling. Trending refers to the historical logging of data point
values over time. Current implementations embrace interval-based change of value or trigger-based
logging and record present values or aggregated values such as minimum, average, and maximum.
Trend logs can be temporarily instantiated to monitor the effect of specific changes in the BAS. They
can also be used to generate reports on energy usage or provide data to detect aging components in
order to assist early replacement. Alarming is a function in BAS, which monitors data point values
according to predefined rules. A typical alarm generator monitors one or a set of data points and generates
an alarm if one of these values exceeds the specified normal range longer than a given time delay.
Most alarm systems require an operator to acknowledge the alarm condition for recording that the
condition was taken care of. This can be done on an operator panel, via E-mail, or Web-based services.
Other alarm systems simply collect and log alarm conditions as they occur and go away for later processing.
Scheduling in BAS allows to set and withdraw data point values at given times of a day. This is,
for example, commonly used to issue set points of HVAC systems, turn on lights before opening hours,
or schedule maintenance tasks at night. The simplest schedule is week based and defines times and values
for each day of the week. To cover exceptional days such as holidays, a calendar-based approach is
used. In this case, a (central) calendar defines the fixed or recurring dates of special days. The schedule
refers to such a calendar and defines exception days, which override the regular weekday schedule.
The standard approach to apply the control network model to a building was to split the system into
three layers [8]. This architecture is depicted in Figure 26.1. On the lowest layer, the field layer, control
over actuators, such as fans or valves for hot water and coolant is provided, and an adaption of them in
response to data point values for sensors, actuators, and set points (which may be provided from a central
location or via local operating panels) is performed. The automation level communicates vertically
with the field level by exchanging values with the field devices as well as horizontally to peer controllers
on the same level. Typical functions in this level are controlling functions, alarm generation, and
scheduling. The management layer is the top level and the most abstract layer in the BAS. Here, global
parameters are maintained for the automation layer. Typical functions are management, data retrieval
and storage with trending, global schedules, global alarm collection and acknowledgement, visualization,
and interfaces to other systems.
In today’s BAS, the different layers are comprised of certain types of devices. The field layer is typically
populated by I/O devices. These can be accessed through direct I/Os linked to a DDC or by devices con-
© 2011 by Taylor and Francis Group, LLC