June 2013

BlueRithm - Improve Quality, Accelerate Deliverables, Save Time and Money
Cloud-Based Commissioning Software

(Click Message to Learn More)

A Path to Multi-Agent Operation of Buildings

Binding oBIX through BIM creates a simple path to multi-agent transactive operation of buildings.

Toby ConsidineToby Considine
TC9 Inc

The New Daedalus

Contributing Editor

New Products
Site Search
Past Issues
Distech Controls
Reliable Controls

Recently, I had a long conversation about why oBIX 2.0 will include a BIM interface with Ludo Bertsch. Ludo represents CABA (the Continental Building Automation Association) on the oBIX Technical Committee. Ludo pushed me until this BIM framework coalesced into expectations for agent-based transactional negotiations in buildings that a customer already had me thinking about.

Earlier that same week, an energy manager asked me to begin considering how agents could support the special requirements for each type of space. Each agent would be given a budget and be tasked with defending its mission. The agents would mask complexity, and they would compete in a transactive market to implement policy.

Talking to Ludo, I realized how well these projects fit together. Before I explain that, however, I need to provide some background.

One view of oBIX is that it offers ways to build a collection of points, and then to interact with those points as a collection. These interactions can read values, set values, or request ongoing telemetry. In the core oBIX specification (1.x), the means of establishing these collections are simple and literal. I write 1.x. because this family of specifications will continue to be the core of even the 2.x implementations and there may be further refinements to after 1.1.

The enterprise interactions for oBIX (2.x) define ways to expand the means to establish these collections, i.e., to make the discovery means more abstract. (oBIX 2.x has other goals as well.) An abstract semantic framework adds new ways to build the oBIX collections. oBIX 2.x anticipate bindings using both tagging frameworks and BIM. oBIX 2.0 features are considered standard extensions to the core oBIX 1.x features; a server that conforms to 2.0 must support 1.x as well as at least one of the enterprise extensions.

oBIX 2.0 will define how to interact with an oBIX server in conjunction with a tagging standard. For example, the Project Haystack is growing in popularity as a tagging framework. If a client knows the server is tagged with Haystack, then it can make a request such as “Return a collection of all points that are tagged as follows.” Using similar semantics, Haystack tags could be used to define security. “Tenants may not modify points with the following tags”. We must also support the symmetric query, i.e., “What Tags are applied to this point?” oBIX 2.0 will not require the use of Haystack or any other tagging standard; it will specify how to use tagging where it exists.

Current best practices create a BIM (Building Information Models) for use in design and construction. oBIX 2.0 defines how to take advantage of a BIM if available. BIM can name the spaces in a building, and link BAS equipment to those spaces. Because people occupy space, and perform business activities in space, space provides a means to view BAS activities from an enterprise perspective, i.e., based on how they support business activities. “Return a collection of points that service Suite 207.” BIM-based interactions can also support secure access to server information, as described above for tagging.

Now, I will turn to the energy agents.

During times of power shortage, whether stemming from market conditions or distribution faults, the portfolio of buildings must adjust to a new power budget. The capability of a building to adjust its energy use based upon temporary signals from a grid or microgrid is known as Demand Response. Demand Response is usually described as temporary non-destructive reductions in service provision, sometimes accompanied by pre-consumption. For example, an office might turn up the thermostat to be slightly less comfortable in summer. The office building might also shift the cooling process to overnight storing [chilled water] to enhance its ability to respond when asked.

Reliable ControlsThe energy manager’s problem was the different spaces that had markedly different abilities to tolerate service degradation, including difference in the type of degradation. Some spaces could accept less outside air so long as a precise temperature was maintained. Animal Quarters could become as much as 10 degrees warmer, but must maintain rigorous standards for air exchange. Biosafety Level 3 (BL3) laboratories can accept both changes in temperature and changes in ventilation, but must maintain negative air pressure. In this environment, changing the wrong parameter slightly is not service degradation, but a service failure, and potentially one with long term effects.

The energy manager’s portfolio has won national awards for energy management; there is very little margin for error in these decisions. The building portfolio consists of a mix of brands and technologies, installed over time on government low bids.

The energy manager proposed that each space be represented by a software agent; that agent would understand the unique needs of each space. During times of shortage, the agents would compete to acquire power to support the needs of their space. The general model for competition leading to optimum control would follow the pattern outlined by Huberman and Clearwater (“Multi-Agent control of Building Environments”, 1995). Unlike that work, the agents would compete over multiple dimensions of building services.

As we talked, I began to tie these two projects, the BIM-based services of oBIX 2.0 and software agents for spaces, together.

Through BIM-based queries, one can create collections of points that are each, in effect, a virtual BAS. Each virtual BAS can be represented by an agent. A traditional general purpose BAS can be treated as many special purpose BAS, the interests of each supported by an agent. From the perspective of a high-level architecture, it makes no difference whether an agent is “near” a system or in the cloud controlling a system. Binding oBIX through BIM creates a simple path to multi-agent transactive operation of buildings.

If you want to find out more about high level architectures for transactive operations, drop me a line, or look for the paper “Understanding Microgrids as the essential Smart Energy Architecture”, (Considine, Cazalet, & Cox, 2012).


Haystack Connect 2015
[Click Banner To Learn More]

[Home Page]  [The Automator]  [About]  [Subscribe ]  [Contact Us]


Want Ads

Our Sponsors