Building Innovation Hub Logo

Predictive Energy Modeling

Get the Basics on Predictive Building Energy Modeling in DC

Just like financial planning services can help business owners understand where to invest their money to maximize profit, predictive energy modeling can help building owners and managers understand how to maximize operational efficiency. This can help owners and managers lower operating costs, improve energy performance, achieve sustainability goals, and meet local requirements such as those spelled out by the District of Columbia’s Building Energy Performance Standards (BEPS).

Established in Title III of the Clean Energy DC Omnibus Act of 2018, BEPS sets energy performance targets that are unique to different property types. The goal is to drive down energy consumption of DC’s building stock to help meet the energy and climate goals of the Sustainable DC Plan. These goals include reducing city-wide greenhouse gas emissions and energy consumption by 50% by 2032, and achieving net-zero carbon by 2050. For the latest on BEPS, read the Hub’s blog post explaining what’s been released and what’s to come.

This primer provides building owners and building simulation practitioners with a basic understanding of the predictive modeling process within the BEPS context. It includes key considerations, targeted project outcomes, recommended deliverables associated with the process, additional resources, and relevant case studies.

Jump ahead:

Primer for Owners

Primer for Practitioners


Primer for Owners

1. How does BEPS impact me?

Buildings covered by the BEPS must benchmark their buildings by including their properties’ programmatic description and annual utility bills into the EPA’s free online ENERGY STAR Portfolio Manager. This software compares the building’s source energy use against peer buildings nationwide and assigns it a corresponding ENERGY STAR score (if eligible to receive a score). Buildings with an ENERGY STAR score or source energy performance exceeding the BEPS targets meet the Standards, while those lower than the target will be placed in a compliance cycle. Buildings that do not improve their performance enough to meet the Standards by the end of the compliance cycle are subject to penalties of up to $7,500,000.  To avoid these penalties, building owners facing a BEPS compliance cycle must make and implement plans to reduce their building’s energy consumption within the time allotted.

For a detailed summary of BEPS, read the Hub’s blog post.

2. What is predictive building energy modeling? How is it different from compliance modeling?

Predictive modeling is an analytical technique wherein a building performance analyst creates a sufficiently accurate and detailed virtual representation of a building to reliably predict future building behavior, like energy consumption and utility costs. It is distinct from compliance modeling, a related analytical technique for determining a new building’s compliance with energy code or third-party certification (e.g. LEED) requirements. Code compliance models do not reliably predict energy use and utility cost, because they are often based on standardized assumptions for building operating parameters. The compliance modeling process predicts a building’s energy use based on these standardized assumptions and compares it against that of a hypothetical code-minimum “baseline” building design. If the compliance model shows the building design performs better than the “baseline” building, it’s considered code-compliant, even if the real-world building ends up using more energy than predicted once it is built and occupied. This is problematic because not only is the building not performing as anticipated, but it could mean that the building will not meet the BEPS.

Instead of relying on reference values, predictive model simulation inputs must accurately reflect operating aspects, such as occupancy and operating hours or plug-in equipment power and usage profiles. The model can then be used to estimate the energy and utility cost savings impact of deploying building energy conservation measures.

3. How can existing building owners use predictive modeling to identify strategies to help meet the BEPS requirements?

Existing buildings have historic energy usage available in the form of monthly utility bills, previous benchmarking data, and sometimes energy consumption hourly trend data from an on-site building automation system or energy management system. Building performance analysts use this historic energy data to calibrate the predictive model and improve its accuracy in evaluating future energy use under assumed conditions both before and after implementing energy conservation measures (ECMs).

Building owners can contract a building performance consultant to provide predictive building energy model services and obtain ECM implementation cost estimates from contractors and by consulting past project cost data. The consultant will develop a predictive energy model to estimate the utility cost savings from efficiency upgrades, which can vary from building to building, but can potentially include:

  • Operational modifications to temperature setpoints and HVAC control strategies such as occupancy controls
  • Sub-metering
  • Implementing deferred maintenance
  • Window and roof replacements, wall insulation
  • Air sealing
  • LED lighting fixture retrofits
  • Lighting control upgrades and plug load controls
  • HVAC equipment upgrades and high-efficiency replacement
  • High-efficiency water heaters
  • Low-flow sink and shower fixtures

The building performance consultant’s scope should include consideration of ECM first costs and associated annual utility cost savings to chart a feasible, cost-effective roadmap to meeting the standard. Additionally, future market conditions or regulations may factor into capital investment decisions, such as electrification.

4. How can building developers use predictive modeling for their new construction, major renovation, and adaptive reuse projects in the planning and design phases to help meet the BEPS requirements?

Predictive modeling can be useful when developing strategies to meet the BEPS for planned new construction, major renovation, and adaptive reuse projects, but some modifications are required for new buildings, buildings without historic energy data, or for projects lacking clear definitions of planned building operations.

For example, a speculative office building may not be able to confirm the future tenants’ planned operations during the base building design phase. However, developing targeted future tenant scenarios is essential to accurately predict energy use and designing for energy-efficient operations. Thus, discussing the possibilities with the building owner is critical. Modeling various use scenarios will provide a comprehensive understanding of how different tenants could impact the overall building’s performance and the energy conservation measures that would apply. For example, a call-center tenant that operates with the lights on overnight might benefit more from highly efficient lighting than a typical office tenant that operates primarily during weekday daytime hours with the lights off at night. The typical office tenant may benefit more from occupancy sensors that switch off receptacles powering computers, printers, and copiers during unoccupied times. See the Hub’s Green Leasing article for more information on how to reduce uncertainty in tenant operations.

New project development teams can account for uncertainty in building operational parameters and design for energy efficiency by incorporating a sensitivity analysis. This involves re-running the model with different possible values for uncertain inputs to understand the resulting impact on energy performance. Common operational scenarios to assess using sensitivity analysis during the predictive modeling process for new projects include:

  • Different daily operating hours
  • Seasonal operation and occupancy fluctuations
  • Partial occupancy levels
  • Different indoor lighting levels
  • Different indoor thermostat setpoints
  • Hotter- or colder-than-average weather
  • User error (e.g. forgetting to switch off the lights or close the window)

Building design and development teams should consider incorporating the strategies that maximize energy savings under:

  • A wide range of potential future operating scenarios
  • The most likely potential future operating scenarios

Sensitivity analyses quantify expected energy use, but do not eliminate uncertainty. Once building construction is complete and the building is occupied, many unknowns can be solidified, and predictive modeling can be performed again. In this way, predictive modeling is a reusable tool for ensuring building performance will comply with BEPS.

5. What are key considerations for building owners looking for predictive modeling services?

Building owners can maximize the benefit of incorporating predictive modeling into their building management and development processes by considering the following:

  Existing Buildings New Construction & Major Renovations
Identify Project Goals Include your project’s timeline, budget, and the BEPS performance targets and periods impacting the building in your:
Current Facility Requirements (CFR) Document or Systems Manual Owners Project Requirements (OPR) Document
Gather Building Information Compile documentation required for an effective predictive modeling process, including but not limited to:
  • Project schedule
  • CFR document
  • Utility bills
  • As-built drawings & specifications
  • Renovation documentation
  • Equipment submittals
  • Facility condition assessments
  • Energy audit reports
  • Maintenance logs
  • Remote access to building automation system

(include information relating to tenant spaces, as relevant)

  • Project schedule
  • OPR document
  • Basis of Design (BOD) narratives
  • Concept drawings & specifications
  • Project meeting minutes

(include information relating to tenant spaces, as relevant)


  • Begin the predictive modeling process early in the project, when the cost of changing design decisions is low and the ability to shape the project outcome is high, typically during or immediately after the early concept/ predesign phases.
  • Engage present and future building stakeholders to better understand building operation, design, and energy conservation measure costs:
    • Building occupants
    • Building management
    • Facility staff
    • Design team members
    • Construction team members
  • Review existing project documentation, requesting additional information from the project team as necessary
  • If code modeling is also required, identify the critical building inputs that need to be better defined and tested via sensitivity analysis during early project design phases
  • Develop predictive building model to calculate the energy and utility cost impacts
  • Perform site walkthrough
  • Identify potential ECMs based on existing building conditions
  • Calculate energy and utility cost saved for each ECM using the predictive model
  • Identify cost-effective ECMs with acceptable payback periods
  • Identify potential opportunities for electrification
  • Evaluate efficient design strategies in terms of energy and utility cost saved
  • Identify uncertain building parameters and test the predicted energy and utility cost savings’ sensitivity to different operating scenarios
Hire an Experienced Consultant Consultants must:

  • Have a proven track record of delivering predictive models that include validation by monitoring and measurement, not just code models
  • Have achieved relevant certifications such as the ASHRAE Building Energy Modeling Professional (BEMP)
  • Understand how ECM implementation and efficient design strategies affects aspects of building performance beyond energy, including indoor air quality, occupant comfort, and maintenance requirements.
Monitor and Verify
  • During regular operation, make sure that facility staff maintains building operational parameters that match the recommended values identified during the predictive modeling process.
  • Anticipate the need for future validation by incorporating monitoring requirements into your retrofit plan or building design that can verify the performance improvements expected by the predictive modeling process.


Predictive Modeling Case Study: An Existing Office Building Facing a BEPS Compliance Cycle  

In 2020, an 8-story, 130,000 SF  office building located in northwest DC entered their 2019 utility bills and building information into EPA Portfolio Manager as required by the Clean Energy DC Omnibus Act of 2018 and received an ENERGY STAR score of 66. This was below the BEPS cycle 1 office building minimum score of 71. As a result, the building was on track to be placed in a BEPS compliance cycle. If the building did not improve performance enough to meet the standard or compliance requirements by the end of the cycle, it could potentially face fines up to $1,300,000. Concerned about these potential ramifications, the building owner, a property management company, made the decision to proactively develop a plan for the property to increase its energy efficiency, and thus its ENERGY STAR score, in order to meet the BEPS  before the end of the compliance cycle in 2026.

To this end, the building owner included the BEPS as a consideration in their existing capital improvement and budget planning process and hired a building performance analyst to undertake a predictive modeling process and identify a cost-effective path to achieve the site energy reduction target. Based on a site visit, occupant survey, project documentation review, and monthly utility bill analysis, the building performance analyst identified ten (10) potentially relevant energy conservation measures (ECMs) ranging in cost from $600 – $1,100,000. The ECMs were classified as “low cost” if under $20,000, and “capital intensive” otherwise.

The building performance analyst created a simulation energy model using building parameters identified during the information gathering process. They ensured that the model reflected the actual 2019 building energy usage with a monthly variation less than 5%. The consultant then used the model to predict the energy and utility cost savings of implementing each ECM.

Based on the site energy savings and simple payback period of implementing each measure, the building performance analyst and building owner categorized each measure as ‘go’ or ‘no-go’ to meet energy reduction goals in a cost-effective manner. This resulted in a package of five (5) ECMs to meet the project goals:

  • Low-cost measures:
    • Convert existing fluorescent lighting fixtures to LED
    • Optimize existing chiller controls
    • Re-align existing solar thermal pool heater collector tubes
  • Capital-intensive measures:
    • Retro-commission existing air handling units
    • Install roof-mounted solar photovoltaic (PV) array

This ECM package had an estimated $233,00-$285,000 implementation cost and an expected $16,200 annual utility cost reduction. The anticipated Energy Star score was 75, which would allow the project to meet BEPS Cycle 1 requirements and position it to potentially meet future BEPS cycle requirements. These measures would also reduce operations and maintenance overhead from current levels, improve occupant thermal and visual comfort, and provide the owner with solar renewable energy credits, tradable commodities obtained from owning and operating a PV system. During and after implementation in 2021-2022, the building performance analyst will monitor the building utility bills on a quarterly basis to substantiate energy savings achieved by ECM package implementation and make additional recommendations if needed.

Primer for Practitioners

This primer provides building energy simulation practitioners who are familiar with the compliance modeling process with a basic understanding of the predictive modeling process and how it differs. It also includes essential resources published by industry authorities relevant to predictive modeling, descriptions of predictive modeling software tools, and key considerations to remember during the predictive modeling process.

1. How is predictive modeling for BEPS different from the code modeling required by the District when obtaining a building permit?

While building code modeling involves virtually simulating building energy performance, it does not accurately predict real-world energy performance or utility cost, but instead seeks to evaluate a proposed building design’s compliance with energy code requirements. The code modeling process does this by calculating building’s energy performance modeled per energy code rules, against that of an imaginary “baseline building with the same shape, program, and operating parameters as the building in question, but with code-minimum building enclosure and mechanical, engineering and plumbing systems. If the code model-calculated building performance is better than the baseline performance by a specific relative threshold, the building design complies with code, even if the real-world building uses more energy after construction and occupancy.

This potentially discrepancy between anticipated and actual energy use occurs because compliance modeling energy use inputs are hypothetical. Code modeling rules allow the use of industry-standard reference values for important operational parameters like weather data, envelope airtightness, building utilization schedules, and internal loads if they are modeled identically in the proposed design and baseline model. As a result, there is often a difference between code model-calculated building performance and real-world performance. This flaw is particularly risky for building owners and managers facing a BEPS compliance cycle, because building performance compliance is measured in real-world data. For these clients, increased input fidelity, sensitivity analyses, and/or calibration techniques involved in predictive modeling are required to ensure they avoid hefty, unexpected fines.

2. What standards, references, and guidelines are applicable to the predictive modeling process?

  • ASHRAE 209 – describes an iterative method for applying building energy modeling throughout the building design process
  • ASHRAE 189.1 – provides minimum recommendations to achieve high-performance green buildings
  • ASHRAE Handbook of Fundamentals – includes climate data, reference tables, information on building systems, envelopes, and materials, and building load and energy calculation methodologies (primarily used in North America)
  • ISO 52016-1 – includes climate data, reference tables, information on building systems, envelopes, and materials, and building load and energy calculation methodologies (primarily used in Europe)
  • IPMVP – measurement and verification process requirements including projects incorporating predictive modeling

3. What kind of software and building performance modeling tools can facilitate the predictive modeling process?

There are hundreds of software tools available in the market for simulating various aspects of the performance of buildings and building systems. The United States chapter of the International Building Performance Simulation Association (IBPSA-USA) maintains a comprehensive directory of building energy software tools (BEST) used by building simulation professional practitioners and researchers around the world.

Building performance modeling software tools that can facilitate the predictive modeling process fall into two categories:

  • First-principles models
  • Empirical models

First-principles models reflect physical laws such as mass balance, energy balance, and heat transfer relationships. Tools based on first-principles techniques seek to develop a reliable representation of reality in a virtual environment. While highly accurate, first-principles models require a large volume of detailed building inputs, which may be unknown and difficult to source.

Empirical models are developed using real-world measured data collected by sensors and metering systems, and sometimes from the results of detailed first-principles simulations. Examples of empirical models range from simple curve fit and regression models to machine learning models. While faster and easier to deploy than first-principles models, the applicability of empirical models is typically limited within the ranges of data used in their development.

Choosing the appropriate simulation tool with granularity appropriate for the project is essential to improving project outcomes. For example, a first-principles model is likely well-suited for the predictive modeling process for a laboratory project with complex HVAC systems and energy sub-metering included, while an empirical model may be better for an existing office building with simple HVAC systems and monthly utility data.

4. What are key considerations to keep in mind during the predictive modeling process?

  • Responsibilities and required due diligence of predictive modeling practitioners are as follows:
    • Work with the building owner early in the project to clearly establish measurable project goals.
    • For new construction, discuss planned building design parameters and potential operating scenarios with the building owner, architect, engineer, and other design team members.
    • For existing buildings, collect and document operational parameters and trend data.
    • Validate the data based on engineering judgment and measurements where necessary.
    • Develop a building energy model and provide a systematic analysis of the building energy usage.
    • Implement a quality control process to verify model inputs and validate model outputs.
    • Identify potential energy conservation measures (ECMs), evaluate impact using building energy model, and cost estimates for implementation typically provided by other parties including design teams or energy service contractors.
    • Calibrate the energy model to utility bills or hourly measured data for existing buildings.
    • Perform sensitivity analyses on model results for new buildings including new construction, major renovation, and adaptive reuse projects.
    • Document ECM energy and financial performance impact results.
    • Facilitate decision-making during the new building design or retrofit implementation planning process.
    • Assist the owner and measurement and verification service provider with validation of predicted ECM impact post-implementation.
  • Effective strategies for collecting the information required for the predictive modeling process:
    • Participate in regularly scheduled design team meetings for planned new projects, request inclusion on all design document issuances and access to the online project file system, if relevant.
    • Request copies of monthly utility bills for existing projects, for multiple years where available (note that utility rate structures may differ seasonally).
    • Request pertinent building documentation including owners project requirements (OPR)/current facility requirements (CFR), project drawings and specifications, systems manuals, equipment submittals, maintenance logs, facilities condition assessments, and benchmarking reports. In older facilities, this information may only exist in print form.
    • Perform site visits to identify and assess major energy-using building components, equipment counts, building envelope condition, and current building operations.
    • Conduct building stakeholder interviews with building occupants, facility staff, management, ownership, and visitors to better understand current building operations. In new building situations where future occupants and facility staff aren’t available, consider whether the building owner might have access to building occupants or facility staff from similar buildings within their portfolio.
    • Request trend data from the building automation system (BAS), where relevant:
      • Configure BAS workstations to allow remote access by outside users upon request.
      • Set BAS workstations to record trend data by default.
      • Where available, obtain energy meter and submeter trends.
      • Where available, obtain operational trends including motor speeds, setpoints, temperatures, humidity levels, pressure levels, and flow rates.
      • Always validate measured trend data to remove outliers, signal noise, and other undesirable artifacts.
  • Temporal resolutions appropriate for the predictive modeling process

The time step of a predictive model should be as granular as the measured data planned for use in validating model-predicted performance. For example, where monthly utility bills will be used, a monthly energy model is sufficient. Where hourly BAS-measured trend data is available, an hourly energy model is required.

  • The role of simulation calibration in the predictive modeling process

The predictive model calibration process identifies building model input parameters resulting in a good match (for example, as defined by the IPMVP) between the results and available historic utility bills and/or measured trend data.

  • Relevant results and outcomes from the predictive modeling process to include in reporting and presentations?
    • Utility bill analysis
    • Comprehensive list of building parameters used as energy model inputs
    • Annual and monthly predicted energy usage by energy type
    • Annual and monthly predicted utility cost by energy type
    • Energy use breakdowns by type and end-use
    • Energy and utility cost savings impact of each identified ECM

5. What are effective strategies for communicating with building owners during the predictive modeling process?

Building owners will rely on the results and the recommendations derived from the predictive energy modeling process to inform significant financial and business decisions. It is, therefore, crucial that the entire process is clearly structured and transparent. Input parameters and assumptions need to be documented so that it is clear where they are used in the model and what impact they have. This can be done through sensitivity analyses.

It is also crucial that building owner and energy performance consultant have a clear understanding the project goals (accurate prediction of future energy use and meeting the BEPS), how they will be achieved (utilization of predictive energy modeling by an experienced building performance analyst), and the workflow and process to get there (as outlined above in section number 4). There must be multiple control points in the process when the building owner and building performance analyst discuss the status, unclear inputs and assumptions, and any unforeseen circumstances. The building performance analyst must be able to document and communicate the technical information in a clear and concise form so that the building owner can make informed decisions and adjust expectations where necessary.

Generally, energy performance consultants and building performance analysts need to understand that the potential ramifications of failing to meet meeting BEPS requirements are highly critical to developers and building owners. Practitioners serve as advocates for incorporating an effective and reliable predictive energy modeling process when considering the impact of the Building Energy Performance Standards on both new and existing building projects.

linkedin facebook twitter

Questions or Feedback?