A Database, and a Building, to Watch

The US Department of Energy High-Performance Buildings Database is an intriguing source of information for green design. On the positive side, it presents design intents (including architectural vision) and performance strategies for 125 progressive buildings, as well as links to sources and contacts. On the other hand, many entries dwell on the acquisition of LEED points, present only modeled performance “data”, gloss over any interesting problems that may have arisen, and show no evidence of post-occupancy investigation.

At least two of the buildings in the list, however, present honest, detailed, useful accounts of their experiences: the Adam Joseph Lewis Center for Environmental Studies at Oberlin College and the Environmental Technology Center at Sonoma State University. Interestingly, these are both university buildings dedicated to environmental studies; with luck, more submissions will follow their examples! Here is a bit about the first one.

Adam Joseph Lewis Center for Environmental Studies at Oberlin College. With a design team led by passionately idealistic professor, David Orr, equally visionary architects at McDonough + Partners, and outstandingly generous donors in the Lewis family, this building was privileged from the beginning. Few of its contemporaries would be able to support a living machine, for example! But many of its features are widely relevant:

  • elongated form to maximize permeability to light and air
  • east-west orientation to simplify shading
  • passive solar heating design incorporating thermal mass (not a simple choice in a cold winter climate)
  • radiant heating in large open areas, such as the atrium, that have high infiltration
  • a geothermal heat pump
  • daylighting with photosensors to dim electric lights
  • an outstanding low lighting power density of 0.9 W/sf
  • automated operable windows for passive ventilation and cooling
  • demand (CO2)-controlled ventilation to save fan energy
  • very expensive highly-insulating glass
  • a stunningly large 4,000sf photovoltaic array (an imperfect answer to the ideal of “sustainability”)

All of these features have been debated in projects I’ve worked on in the last year: they are gradually entering the mainstream, and owners and architects are in need of solid precedent studies.

The exceptional part of the Lewis Center effort is the commitment by both Oberlin staff and students and NREL scientists to evaluate the performance of systems in their contexts, to make both big changes (e.g. replacement of the original electric boiler with a ground-source heat pump) and small ones (controls adjustments) and to track the effects of changes. An excellent real-time display is provided on the Oberlin website, and field work results by Paul Torcellini and colleagues at NREL are also now a public document.

So: how well does this building perform? The answer is: quite well! It has an EUI of about 32 kBtu/sf-yr, met entirely (on an annual basis) by the PV array. This is about 1/3 of the EUI of other Oberlin buildings. Is it “really” net-zero? Some purists argue persuasively that the point of a net-zero building is not to offset energy use with enormous arrays of silicon wafers, glass, and metal. We’re working on a net-zero K-12 school now that’s seeking the same vast-PV solution, understandably – getting heating loads down, after a point, is just really, really hard. In any case, Oberlin has taken a fantastic step forward for all of us – not only by creating this progressive building, but by sharing their experiences, good and otherwise, with all of us.


What does building performance “research” mean, exactly? (part 2)

The Scandinavian hospital study illustrates several contributions that building research offers, that energy modeling, commissioning, and system monitoring cannot:

1. Understanding of design intent

The vision of the design team – owners, stakeholders, architects, and engineers – has a permanent impact on the energy performance of a building. Siting, geometry, orientation, envelope materials, organization of the program, spatial configuration, and environmental control system choices all reflect this vision, tempered by constraints of time, money, and building codes, and all have substantial impacts on the energy needed to heat, cool, light, and ventilate the building.

The effects of these key performance-affecting decisions are so decisive, and so long-lasting, that we must understand the forces that shape them in the design process. This element is therefore a cornerstone of building performance research.

2. Systems-level analysis

Buildings operate as complex systems of people, machines, structures, and climates. Since these elements affect building performance as an interacting set, a clear understanding of building performance necessarily requires investigation into the operation of the system as a whole. Energy models do simulate buildings as systems, but their best role is to inform decisions among a limited set of design alternatives; they are definitely not diagnostic tools. To understand the operation, and especially the malfunction, of a real building requires field investigation.

Revealing unexpected activities of occupants, and understanding the origins of these activities, is an especially important aspect of field research. Unprogrammed use of spaces, manual overriding of system controls, and propping openings open or closed are just a few ways that occupants can unwittingly diminish the energy performance of their building; tracing these to their motivations, and then to concrete aspects of the building design or operation, is an essential component of systems-level building research.

3. Pattern discernment through comparison of multiple buildings of a type

Each building is substantially unique: despite common elements, a particular assembly of spaces, materials, climate, program, and occupants is rarely duplicated. This complicates efforts to determine which performance strategies work well, and which don’t, in a particular building type. Yet, as the Scandinavian hospital study shows, patterns do emerge when enough examples of a type are compared.

The seeking of patterns among multiple examples of complex systems has excellent precedents in field ecology (see work by E. Odum and J. Lovelock) and in architecture (C. Alexander, A Pattern Language). Their field techniques and analytical tools are directly applicable to building systems, as well, and should inspire us toward a “comparative building ecology” that illustrates performance patterns, in their contexts, robustly.

4. Controlled experimentation

Although every building is an uncontrolled experiment, some controlled experiments can be conducted within a building, nonetheless. Energy use of analogous spaces that differ only in occupancy or equipment can be compared; conditions in individual spaces can be tracked through varying seasons; passive airflow paths can be obstructed or cleared; light shelf sizes and angles can be varied; setpoints and schedules of mechanical systems can be adjusted, for example. While such adjustments are often undertaken by facilities managers, rarely are the results of individual changes tracked over time to yield meaningful information. Such experiments might also be simulated with models; an intriguing document by the New Buildings Institute presents such analysis for large buildings. Given the limitations of models, however, the realm of controlled experimentation remains an important one for teasing apart relationships among spaces, people, and environmental control systems in real buildings.

5. Publication

The ultimate goal of pure research is to publicize the results, so that a wide audience can learn from them (where “publication” includes meetings, talks, discussion forums, and websites as well as design and science journals). In contrast, commissioning reports and building energy models are private documents; indeed, building designers and owners are understandably reluctant to publish evidence of performance below expectations.

At the same time, the disclosure of design decision pathways, model predictions, operational realities, and performance outcomes would help future design teams immensely. This, therefore, is the most important of the unique contributions that building performance research has to offer: the provision of reliable information, obtained through rigorous investigation and experimentation, unbiased by financial or legal interests, in straightforward, accessible forms with the strength to change common practice.

What does performance “research” mean, exactly? (part 1)

If “research” means original research, and if buildings are created from principles of spatial design, climate responsiveness, structural stability, heat transfer, fluid mechanics, and hydrophobicity, among others, that are already known, then what, exactly, does “research” mean in the context of the built environment? What else is new, in these constructions, that we have yet to learn? More specifically, what can we learn that is generally true, and generally applicable – that can inform the design of future buildings?

To an architect, there is still quite a lot! To a social scientist, there is an equal abundance. But in terms of energy performance…what can original research really offer, that energy modeling or commissioning or just system monitoring cannot?

One study that’s influencing our current work is still in progress by members of the University of Washington Integrated Design Lab in collaboration with BetterBricks and I-Sustain. Several additional people, including NBBJ architects and my boss, have also been closely involved. In this work, the group investigated a number of different Scandinavian hospitals, analyzed their geometry, orientation, layout, loads patterns, envelopes, hvac system design, and energy usage, and distilled a series of truths from the results that are evidently unknown to the US hospital design community. This will be a fantastic resource when it’s published!

One striking result, for example, was the superior effectiveness of decentralized ventilation systems in combination with radiant heating and cooling. While the inherent wastefulness of VAV reheat systems might seem obvious, with the strategy of cooling air to the lowest temperature needed by any of the zones served, and then reheating it locally for all the rest, the majority of American hospitals are designed with this exact system type. The next question, of course, is why? Are they cheaper? Easier to install? More reliable? Or just “the way it’s always been done”? There must be some obstacle out there to changing common practice.

(to be continued…)

Building performance investigation doesn’t pay – yet

Today I was working away on trendlogs for the under-performing LEED building, trying to figure out why air-handling units were misbehaving in various ways, when my boss startled me. “Stop working on that!”, he exclaimed. “We have to watch the budget REALLY closely on this one!” Considering that we were only 8 hours into the project, this was rather unexpected. We clearly have an exceptionally small budget to bring this complicated 150,000sf building into line with its energy model – and this is a building that has gotten the best that money can buy from the very beginning.

So if THIS privileged building isn’t worth more study, even to save thousands of dollars a year and to gain its last (necessary) M&V point, where do the thousands of other green building hopefuls find incentives for maintaining their performance goals? That is, before a dramatic energy crisis solves the problem for us?

Since we’re already discussing LEED issues, a closer look at the LEED v.3 Minimum Requirements is worthwhile, and is mildly encouraging. For new construction and major renovations, Item 6 reads:

6. Must Commit to Sharing Whole-Building Energy and Water Usage Data
All certified projects must commit to sharing with USGBC and/or GBCI all available actual whole-project energy and water usage data for a period of at least 5 years. This period starts on the date that the LEED project begins typical physical occupancy if certifying under New Construction, Core & Shell, Schools, or Commercial Interiors, or the date that the building is awarded certification if certifying under Existing Buildings: Operations & Maintenance. Sharing this data includes supplying information on a regular basis in a free, accessible, and secure online tool or, if necessary, taking any action to authorize the collection of information directly from service or utility providers. This commitment must carry forward if the building or space changes ownership or lessee.

While this commitment is still technically voluntary, certification can be withdrawn for failure to comply. Still, the performance is not required to reach or, apparently, to even approach the performance originally predicted. Even to gain Monitoring & Verification points, it is only necessary to “Provide a process for corrective action if the results of the M&V plan indicate that energy savings are not being achieved.”

The issue of confidentiality was also broached, and real estate lawyers are already taking note. Apparently the ugly jungle of building design lawsuits has just added a fertile patch of soil. Regrettably, this has potential to delay further the all-important transfer of information from real buildings back to the design community.

Meanwhile, I was immediately reassigned to a new project: construction of a DOE2.2 energy model for a large ambitious new building, not so very different in program from the troubled LEED building. It’s in schematic design, of course, which means that a large number of important details are completely absent. Energy models are valuable tools, and this model-building exercise will surely be a useful guide for the architects. Unfortunately, it will surely not predict the actual energy performance. But it sure does have a nice big budget.

Nature calls

On September 10, the prestigious journal Nature published a commentary entitled Overrated Ratings”, in which it criticizes the LEED green building award system for falling short in promoting design of low-energy buildings.

“…as is well known in the building research community but not outside it,” the editors write, “neither LEED nor any other such rating is a reliable guide to energy performance. Labelled buildings often perform no better in energy terms than the general building stock, and sometimes worse.”

Sometimes WORSE. How is that possible? True, the original LEED system weighted energy performance rather lightly, giving nearly equal weight to site design, water and waste management, and green materials. But these should not be causing the energy use to be occasionally WORSE than average, especially since the energy performance of every LEED building must be modeled if it is to gain energy points.

Here, unfortunately, is the problem, as the editors continue: “most ratings assess a building’s energy performance using theoretical projections from engineers’ models, but don’t measure its real, post-occupancy performance, which often can be much poorer.”


My colleagues and I have just begun work on a project with this exact problem. The building was awarded LEED Gold two years ago, in part through energy points gained with a careful, thorough DOE2.2 model that used the best information available at the time (i.e., pre-occupancy). Two years later, this beautiful building is using about 50% more energy per year than modeled. Fifty percent. This is not the kind of error that results from occupants squeezing a few more people into a space than projected, or cranking the thermostats up a few degrees. This is a sign of pervasive use of building spaces and systems in ways very, very different than intended by the owner, designed by the architects and engineers, and yes, simulated by the modelers.

Why are we doing this? Because, thank goodness, LEED 2.2 had a “Monitoring and Verification” point that this building needed to keep its Gold label. The teeth are sharper in LEED v.3.

We don’t know what the problems are, yet, though we have some preliminary ideas. We believe they will be relatively easy to identify, though we fear that some of them result from conflict between design intents and occupants’ reality. Most of all, though, we hope that this effort will become part of a new, larger, and sustained national effort to evaluate the performance of not just LEED buildings, but all new green buildings, and vast numbers of green-ing existing buildings as well.

Every building is an experiment

Scientists always replicate their experiments. No journal would publish their work if they didn’t.  Field ecologists agonize over finding comparable niches, hoping that statistical analysis will tease truth from the daunting variability of the natural environment. Chefs try a recipe repeatedly, adjusting ingredients minutely, to find the perfect taste. Athletes and musicians practice moves and passages over and over, striving for the subtle nuances that will make all the difference in the final performance.

But in architecture, every building is its own experiment. With rare exception, it is an experiment of one. No controls, no replicates, just one glorious culmination of years of hopes and ideas and struggles to get the thing built. And what of its performance?? Spatial performance is rewarded with glossy features in Architectural Record or Dwell. But “Star-chitects” like Peter Eisenman notwithstanding, energy performance is also a very real concern among designers! Buildings use enormous amounts of energy, and as a result, enormous effort has gone into programs to promote green building design: LEED, SEED, Architecture 2030, International Living Building Challenge, to name just a few. Buildings routinely win awards, gain certifications, and advertise themselves as “green” on the basis of design alone. Think of how astonishing this is: would a recipe, a music composition, a dance, a space, win awards without ever being experienced?

To be sure, building performance can be predicted to some extent from digital models. I build DOE2 and EnergyPlus models for a living, for example. They are valuable, even essential, tools. But they are not the whole story. Each model is only as good as the information that goes into it, and information is always incomplete in schematic design, which is when most building energy models are built. Elements get changed in design development without re-analysis. Decisions are re-visited during construction, materials are substituted, mistakes are made, budgets are cut, timelines are quickened, and occupants use the building in unexpected ways.

But what about that performance? Isn’t the final performance important enough that we actually go check, to see if the experiment worked? Actually – no! Tragically, most buildings never receive any greater performance assessment than the owner’s monthly glance at the utility bill. This is tragic because vast resources are invested in buildings, because thousands of green buildings have been built in the last decade and thousands more are on the boards, and because the lessons available from those we’ve built are not even being listened to, with a few laudable exceptions (such as those published here and here), so that new ones can benefit.

This blog exists for the green building agents who are out there listening, for public agencies that need persuasion to fund green building research, for new green buildings with valuable messages, and for designers who urgently need these messages.