Retrofitting existing buildings represents the fastest way to reduce energy consumption for Department of Defense (DoD) installations. However, the current retrofit process is labor intensive, is scoped at the level of a single building or small group of buildings, prioritizes equipment selection based on initial cost, provides solutions that depend on the expert performing the assessment, and results in 10-30% reduction in energy use through conventional energy conservation measures that are component- or sub-system-based, such as lighting system and central cooling/heating plant upgrades.

To address these gaps, United Technologies Research Center (UTRC) developed an energy performance analysis methodology and tool to analyze solutions with the potential to reduce energy use in existing DoD facilities by 50% or more. The Deep Energy Efficiency Retrofit Analysis Tool (DeepRetro Tool) is a user-friendly tool aimed at speeding up the building energy audit process—from data gathering to running supporting tools—by 10 times, and accuracy within 10% of total measured or reported annual energy consumption, and within 15% of measured or reported annual building energy consumption broken down by major energy sources (e.g., gas, electricity, oil, steam) when the building characteristics are well known. The methodology is also aimed at providing a portfolio analysis capability that can analyze thousands of buildings simultaneously.

The DeepRetro Tool was applied to 33 DoD buildings and validated against relevant test cases provided by American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) Standard 140 for the evaluation of building energy analysis computer programs. Inter-model validation was performed, comparing the DeepRetro Tool results with results from eQuest models for eight buildings located at Fort Bragg, North Carolina. Additionally, measured data was used to achieve model validation and verification for another 25 buildings located at Joint Base San Antonio (JBSA), Texas, within Fort Sam Houston, Randolph Air Force Base (AFB), and Lackland AFB.

The DeepRetro Tool and methodology includes physics-based building, heating, ventilation and air conditioning (HVAC) systems and energy-conservation measure (ECM) models to estimate facility energy use, and tools to suggest and rank building-retrofit ECMs based on system technical, environmental, and economic performance. The analysis of retrofit potential is performed by a rapid screening of individual ECMs separately, followed by the analysis of ECM packages, in which ECM inter-dependencies are considered.

In addition to the deep retrofit analysis of individual buildings, the DeepRetro Tool can analyze an entire portfolio consisting of multiple buildings. This capability is made possible by a matching algorithm that selects relevant building parameters from existing databases.

The DeepRetro Tool provides unique technical capability, including uncertainty quantification and sensitivity analysis, which are used both to improve the accuracy of the building and system models by identifying the audit inputs that are most important to capture accurately, and to determine the range of validity of the proposed retrofit solutions. The tool incorporates an automatic calibration capability that tunes unknown or uncertain input parameters, using metered utility data or sub-metered data, to create a more accurate model of energy use.

A graphical user interface (GUI) was developed to support the detailed retrofit analysis of individual buildings or a set of buildings for which characteristics are well known (detailed portfolio analysis).

By using the DeepRetro Tool and methodology, the baseline energy performance of an individual building has been evaluated in less than 1 minute, given an audit input file. A building retrofit analysis has been achieved in about 1 hour of computational time with minimal human intervention. Additionally, the entire DoD building portfolio of approximately 250,000 domestic buildings has been analyzed in less than 10 hours, using a computer cluster. The portfolio analysis detects trends and selects the best retrofit measures involving integration of energy efficient envelope, lighting, HVAC, and control systems for each market sector (e.g., climate zone, building type, size, etc.).

While the DeepRetro Tool only requires 11 basic building characteristics as inputs, accuracy significantly improves when additional building parameters, systems, and schedules are input into the tool. When building characteristics and schedules are well known, an accuracy of 10% has been achieved when comparing tool results against known and reliable annual energy consumption in a building. The analysis is enabled by an easy-to-use, intuitive user interface.

Future implementation issues that may have to be addressed to use or deploy the DeepRetro methodology at DoD installations include verification of the pertinent Certification of Networthiness (CON) permits to enable installation of the DeepRetro software on DoD computers, integration of the DeepRetro Tool with DoD installation-specific databases and metered data that can provide additional accuracy of results, and a broad training and support program for DoD energy managers and policy makers who will use the DeepRetro Tool and methodology.