United Technologies Research Center (UTRC) delivers some of the world’s most advanced technologies, innovative thinking, and disciplined research to its business units at United Technologies Corporation, as well as external customers around the world. UTRC’s expertise ranges from aerospace, building infrastructure and services, heating and air conditioning, fire and security systems, and power generation. UTRC has been focusing on advancing and maturing technical opportunities to create the next generation of high-performance buildings. In focusing on buildings, consumers of 40 percent of U.S. energy every year, UTRC aims to provide innovative solutions to reduce buildings’ impact and steer its customers toward a more sustainable future.
Improving Building System Models to Enable Deep Energy Retrofits
It has been shown that during normal operations, buildings use up to 30 percent more energy than designed, in part due to uncertainties in the design and simulation process. This unanticipated added energy use results in significant cost increases to owners and tenants. There is no set of tools to accurately quantify and manage the uncertainty in building energy use. In turn, this creates a significant barrier to adopting energy-efficient technologies for buildings- modifications that can drive down energy use with short-term return on investment.
Accurate building energy model parameter values are not always available, such as in the case of equipment efficiencies. Useful patterns of building operation and occupant behavior cannot be precisely determined during the design phase but have a significant impact on energy performance during operation. To accurately predict building energy performance, it is essential to estimate the effects of uncertainty and identify key parameters that impact building performance. That information can then be used to calibrate building models and improve their predictive capability. The information can also be used to inform builders, equipment manufacturers, and building operators, as to needs of tenants and building portfolio managers. An example of parameters considered in a building energy analysis is shown below.
In collaboration with Lawrence Livermore National Laboratory (LLNL) through its HPC Innovation Center, the capabilities of UTRC’s building energy analysis will be expanded to deliver an integrated suite of analysis tools meeting all requirements for rapid building retrofit design. The energy use of Building 101, located in the Philadelphia, PA, Navy Yard, will be examined using Energy+ and TRNSYS, varying more than 900 parameters over different operating values. These parameters account for energy used in heating and lighting, and are highly dependent on the occupant’s use of the building, the weather, equipment properties, and window and wall insulation. These tools will help ensure that design objectives in minimizing energy use without sacrificing comfort are met and maintained during building operations.
In order to best understand the impacts of a large number of building parameters, UTRC aims to run thousands of simulations, exploring the building model parameter space. By parallelizing these simulations, the time required to complete the thousands of simulations will be reduced from weeks to a few hours. The information provided by these simulations will allow building operators to understand which parameters play the most significant roles in building energy usage and how one might vary these parameters within acceptable comfort ranges to affect this usage. This information also will help building operators eager to retrofit their building to gain energy efficiency and reduce operation costs.
UTRC experts in collaboration with Lawrence Livermore computer scientists ran 10,000 whole building simulations for Philadelphia Navy Yard Building 101 to generate data for global sensitivity analysis encompassing 917 building parameters. The analysis identified key parameters for each energy output for the building. An example of analysis results is shown to the right, where parameters affecting the electricity consumption of air handling unit fans are identified. In this example, 91% of the overall sensitivity is due to only three uncertain parameters. Results obtained in Building 101 analysis suggest that for a typical building there are only a few parameters per energy output that impact energy performance significantly. This kind of analysis provides invaluable information to building retrofit designers as it reveals which retrofit solutions will have the biggest impact on building energy performance.
Prior to the hpc4energy incubator
This type of analysis takes 1-2 days when using the current UTRC Linux cluster. With the resources made available at LLNL, the analysis turnaround time was reduced to only a few hours, thus making it more suitable for use by building stakeholders. Consequently, the ability to perform rapid sensitivity analysis for the whole building with a large number of uncertain parameters would greatly benefit the building sector, especially in areas of energy consumption and cost savings.
The hpc4energy incubator provides access to the capabilities of Lawrence Livermore National Laboratory for selected companies in the energy sector to demonstrate the benefits of incorporating high performance computing (HPC) into technology development. Part of Livermore’s broader industrial outreach and economic development initiatives, hpc4energy is supported by the HPC Innovation Center on the Livermore Valley Open Campus.