A Model of Sustainable Design

The $354 million David L. Lawrence Convention Center, located on the Allegheny River in Pittsburgh, is the newest green building in what is perhaps the greenest city in the United States. The Center, which is one of just a handful of facilities to earn a Gold rating under the U.S. Green Building Council's LEED system, incorporates a number of energy efficiency and sustainability features that h...

By Dr. Malcolm Lewis, P.E., and Tom Lunneberg, P.E., CTG Energetics, Inc., Irvine, Calif. June 1, 2004

The $354 million David L. Lawrence Convention Center, located on the Allegheny River in Pittsburgh, is the newest green building in what is perhaps the greenest city in the United States. The Center, which is one of just a handful of facilities to earn a Gold rating under the U.S. Green Building Council’s LEED system, incorporates a number of energy efficiency and sustainability features that have not previously been included in a convention center. Its impressive list of sustainability features includes ample daylighting and natural ventilation for 330,000 sq. ft. of exhibit space, enabling the center to meet most of its lighting, ventilation and space cooling needs with minimal energy input for most of the year. These features required a significant analysis effort on the part of the design team.

Analysis and results

A computer-based energy model was developed to predict energy use for the center, and this model served multiple purposes during different stages of the project. Initially, the model was used to validate the peak cooling load that the facility would experience. Establishing a reasonable estimate of the peak cooling load was important be-cause the convention center was to be connected to a district cooling plant from which it would purchase chilled water and steam. The size of the interconnecting pipes, pumps and heat exchangers—along with the monthly capacity charges that are levied to recover the costs for this equipment on an amortized basis—would be determined based on the size of the worst-case demand for chilled water. There had been differences of opinion regarding the size of the cooling load, largely due to different underlying assumptions about how the convention center would be used. The number of people in the building, use of exhibit lighting and the projected schedule for major events all influence the building loads. Fortunately, the mechanical engineer for the project had first-hand experience with another recently completed convention center project and was able to obtain reliable data from that project based on actual operation. This real-world data was included in the DOE-2 energy model in an effort to make the cooling load calculations as accurate as possible.

During the design development phase of the project, the energy model was used to evaluate a variety of energy-efficiency concepts that were conceived by the design team. Some of these ideas, such as high-efficiency lighting, improved fan efficiency and high-performance glazing, were easy to model. Other, less common measures required creative analysis approaches from the energy modeling team. For example, it was proposed that the chiller plant could reject its heat into the Allegheny River instead of using traditional evaporative cooling towers. The team felt that this approach would have energy as well as aesthetic advantages because no cooling towers would be visible—or take up precious space—and there would be no plumes of water vapor emanating from the towers. Further, there would be no energy use associated with cooling tower fans. Since DOE-2, the building simulation program used for modeling the convention center, does not have the capability to directly model such a heat rejection approach, the modeling team, in conjunction with the HVAC designer, conceived a way to mimic the performance of such a system. This method involved specifying cooling towers with no associated fan horsepower and control settings that provided a condenser water temperature profile that mimicked what would be available from the river. Though not a perfect solution, this method gave decision makers an order-of-magnitude answer to the question of how much energy could be saved by using the river as a heat sink.

Towards the end of construction, when the various energy efficiency measures had been incorporated into the final design, the DOE-2 model was used to document compliance with ASHRAE Standard 90.1-1999. This standard is the basis by which LEED credits are awarded under Credit 1 of the LEED Energy & Atmosphere category. Using the Energy Cost Budget method, which excludes process and other unregulated loads from its calculations, the Lawrence Convention exceeded minimum ASHRAE 90.1-1999 requirements by 35.6% (see Table), which qualified the project for five LEED points—a significant contribution toward the project’s Gold rating.

The DOE-2 energy model was able to serve multiple beneficial purposes because the design team was constantly looking for new uses for the model. The substantial effort associated with generating a model leads many projects to put it on the shelf instead of employing it to answer relevant questions.

Lessons learned

The convention center pushed the DOE-2 simulation software to its limit with respect to its ability to model the building shape and energy efficiency features accurately. One of the building’s most prominent features is its sail-shaped roof, a large curved element that spans the main exhibition hall. Because most building simulation programs can only handle rectangular wall and roof elements, it would have been difficult to accurately model the curved roof and the accuracy of the energy calculations would have suffered.

In order to better model its shape, team members from Lawrence Berkeley National Laboratory initially proposed using (at the time) a very new version of the DOE-2.1E simulation program (bug release 107) that could model polygon-shaped wall and roof components. While improved accuracy is always appreciated, using the newer version of the program had some undesirable ramifications. Most significantly, for certain model inputs, the program used different syntax that had to be learned by the modeling team in the course of a very aggressive completion schedule. In addition, the limited documentation for the new program resulted in wasted time that could have been better employed on energy analysis instead of software troubleshooting. In retrospect, it would have been preferable to stick with a tried-and-true, albeit slightly less capable, version of DOE-2.1E. When it comes to achieving accurate results in a timely manner, our experience shows that expertise with a particular simulation tool trumps greater capabilities of a less familiar tool every time.

Despite the fact that DOE-2.1E is quite capable of modeling a variety of energy efficiency measures, its limitations were noticeable for certain aspects of the convention center analysis. For example, DOE-2 calculates heating and cooling loads by tallying the quantity of heat generated within a space and the amount of heat gain or loss between that space and adjoining spaces (indoor or outdoor). If there is a net increase in heat, then DOE-2 considers this to be a cooling load. A net loss represents a heating load. DOE-2 does not account for air movement, whether it is driven by convection or outdoor wind speed and direction. In essence, the DOE-2 model is not aware that hot air rises, which is of great importance when modeling high ceiling spaces such as exhibit halls. In addition, DOE-2 has very limited ability to model natural ventilation strategies because it does not consider airflow patterns. For these reasons, the design team commissioned a natural ventilation study using computational fluid dynamics (CFD) software to predict the thermal comfort delivered by natural ventilation in the exhibit halls.

Another shortcoming of the DOE-2 program is its limited daylighting modeling capabilities. While the program does a reasonable job of estimating the energy benefits of different lighting control strategies—dimming, stepped or simple on/off control—in response to incoming daylight, it cannot be used as a serious daylighting design tool. The main reason is because the program does not consider how interior features of a building impact daylight penetration. For this reason, it was necessary to commission another specialized study to determine the impact of various daylighting aperture configurations in the main exhibit hall.

As more design professionals embrace the integrated concept, there is increasing demand for analysis software that can inform a wider array of design decisions. Ideally, energy, airflow and lighting analysis would be done within a single simulation environment. While such software is not currently available, there is increasing interest in the field of interoperability between computer programs that enable data from one program to be fed into another program. For example, many architects use solid modeling software to develop visualizations of what different building massing alternatives will look like. The data that is input to this type of program can also be exported to DOE-2.1E if the right software tools are used. VisualDOE, a graphic user interface for DOE-2 that was developed by Charles Eley & Associates, can accept thermal zoning data if it is contained in a valid document exchange format (DXF) file. Usually, some intermediate CAD processing is needed to define the thermal zones, but once this has been done the zones can be imported to VisualDOE. The advantage is that it saves the modeler the time of inputing the same geometric data that the architects input to their solid modeling program. An additional advantage is consistency between the two different pieces of analysis, as well as a reduction in the chance of data entry errors.

Certain lighting and CFD software programs can also import DXF data. This makes it conceivable for energy, airflow and lighting analysis to be performed based upon a single set of geometric assumptions. While much effort must still be expended to develop all of the program inputs for each different simulation program, reducing the effort associated with geometric inputs is a significant advantage for green building programs that employ a variety of green building features.

Implications for the future

In the future, the concept of interoperability is expected to go much further than mere geometric data sharing among disparate programs. Eventually, it is expected that computerized drafting programs will provide heating and cooling load calculations, lighting analysis, energy use estimates and even thermal comfort predictions for the eventual occupants. One question that arises from pondering the use of such a tool is which member of the design team—architect, mechanical engineer, electrical engineer—would take ownership of a model that contains design information that affects so many disciplines? That issue seems to be a minor one that can easily be overcome, however, when one tallies the advantages that a highly interoperable suite of analysis tools would present to a high-performance project such as the David Lawrence Convention Center.

Energy and cost comparison between base and proposed design

Fuel Type Proposed Building Budget Building Proposed/Budget
Energy (10^3 BTU) Cost ($/yr) Energy (10^3 BTU) Cost ($/yr) Energy (%) Cost (%)
Electricity 153,308,496 $1,266,215 197,476,235 $1,623,236 77.6% 78.0%
District chilled water 27,797,328 $699,567 57,209,784 $1,317,640 48.6% 53.1%
District steam 32,540,000 $324,098 23,959,000 $239,143 135.8% 135.5%
Electricity (Regulated Loads) 71,687,058 $592,082 115,854,797 $952,315 61.9% 62.2%
Steam (Regulated Loads) 32,540,000 $324,098 23,959,000 $239,143 135.8% 135.5%
Chilled Water (Regulated Loads) 27,797,328 $699,567 57,209,784 $1,317,640 48.6% 53.1%
Total (Regulated Loads Only) 132,024,386 $1,615,748 197,023,581 $2,509,098 67.0% 64.4%