Defining the Importance of Physical Architecture in the Design Process
Why is physical architecture important to the design, and what technologies should you choose?
My last two columns dealt with point selection and sequence writing. These determine the majority of a BAS project’s scope of work; however, there is one other issue necessary to the core of the design: the system architecture. In this case, I am referring to the physical architecture that deals mainly with the layout/quantity of the controllers of various types along with the wiring topology and transport technology used to connect them. (In my June 2018 column, I covered BAS application architecture, which is a different issue typically only important to integration challenges.) Why is it important to define the physical architecture (or simply “architecture” in this column) in the design?
IP vs. MS/TP
The two major communications wiring and transport technologies used in BAS are IP and various forms of EIA-485 (mainly BACnet’s MS/TP). EIA-485 is a very old and slow technology that, because of its use in MS/TP, is still the dominant choice for lower-end controllers (e.g., VAV boxes). IP is the dominant choice for higher-level controllers (i.e., BACnet “B-BCs”) and is necessary for connecting the system to the operator interface and internet remote access. IP is much faster and more robust than MS/TP, allows the controller to communicate directly with the operator interface (i.e., its communications are not routed through a B-BC controller), and can also allow for BAS networking via the building’s IP infrastructure (a cost-effective approach if allowed by the building owner).
Many believe that critical facilities should maximize the use of IP over MS/TP (even at the VAV box level), though how much this actually improves BAS reliability is still a subject of debate. For most current BAS choices, the IP vs. MS/TP choice mainly applies to middle-level controllers (e.g., BACnet “B-AACs”), and this is where I’d recommend IP communications be used in all but low-criticality projects. However, there’s more to assuring reliable communications and system operations than the use of IP.
Controller Quantities and Segregation
To minimize the effect from a loss of a controller, the following are some guidelines to follow in further developing the architecture:
Use only one (and no more than one) mid-level controller per controlled equipment (e.g., AHU). The controller shall control that equipment only;
Use the above mid-level controller to manage the MS/TP network for its (and only its) associated VAV boxes and other terminal devices;
AHUs that serve a large number of VAV boxes will involve multiple MS/TP networks. Each network should serve a separate floor (or contiguous floors); and
If there are more MS/TP networks than can be managed by the single AHU controller, then additional controllers with MS/TP-to-IP routing (or dedicated routers) shall be used and reside on the same IP subnet as the AHU controller.
Fault Immunity and Redundancy
Mission-critical facilities should consider further steps like:
When multiple-redundant equipment serves an area (e.g., two or more AHUs), then each unit should have its own sensors for critical portions of the sequence. These might include space temperature, supply air sensors, duct and space static pressure sensors, etc.
If multiple-redundant AHUs serve VAV boxes, then each AHU controller should have its VAV controller network divided not by area. Instead, and simply put as possible, adjacent VAV boxes (especially if they serve the same area/room) shall not be connected to the same AHU controller.
If the highest-level controllers perform building-wide control (direct or supervisory), they should be on separate IP sub-networks and set up in a primary/standby manner such that a failure in one enables a redundant sequence in the other and vice versa. Loss of IP communications between the controllers would be one way to detect a possible failure; however, BI/BO point pairs wired between the controllers would improve the chance that an actual controller failure is being detected. Each controller’s BO would be wired to a BI in the other controller and kept energized during normal operation. A failure would be likely when a change in the BO’s state is detected by the other controller.
There are more architecture issues to consider than are covered in a typical building project. The need to consider these issues depends upon the criticality of the building’s operations. Either way, the architecture details can be (mostly) documented in the project design using a detailed architecture drawing.