Calibrated Simulations & ASHRAE G14

7 posts / 0 new
Last post

Hi,
Does anyone have experience with calibrated simulations? Especially, in accordance with ASHRAE guideline 14, or other standards / guidelines.

I've been working on a calibrating and EnergyPlus model to measured gas HVAC data for a single family house with extensive monitoring and detailed information available and I am finding it very difficult to achieve the <= 30% CV RSME that Guideline 14 suggests.

I'm surprised that it is this challenging because I have far more information available than what would likely be available in typical modelling jobs. Eg. ACH rating, measured ground temperature data, full as-built plans, indoor temperature measurements....

So I'm curious to hear about anyone else's experiences with this. Is this guideline generally achievable for modellers?

Thanks,

Hayes Zirnhelt

Hayes Zirnhelt's picture
Offline
Joined: 2011-10-02
Reputation: 0

as far as i'm aware if an item is not covered in the nrel guidelines
then it defaults back to the 90.1-2004 appendix g requirements.

Patrick J. O&amp;#039;Leary, Jr.'s picture
Joined: 2011-09-30
Reputation: 200

Hello Hayes,

I helped write the calibrated simulation section of Guideline 14, including the new Guideline 14, which is about to go out for public review. I also participated in the NEMVP, IPMVP.

Currently, the Guideline 14 committee is completing the final draft and will meet at the Chicago ASHRAE meeting for final discussion before it goes out (Dennis Landsberg is the committee chair, and I've cc'd others to chime in). So, if you wait for a little bit, you'll be able to get the public review draft of the new Guideline 14, which contains lots of new, good stuff.

You'll also find useful material in the IPMVP (latest version), the ASHRAE Performance Metric Protocols, and the ASHRAE Handbook of Fundamentals, Chapter 19, 2009. Dr. Reddy also now has a great text book on statistical or inverse methods for engineers.

In addition, here are some of papers that our Lab has published that address the issue of calibrated simulation. You'll find most of these on-line at our web site:

Song, S., Haberl, J. 2008. ?A Procedure for the Performance Evaluation of a New Commercial Building: Part I ? Calibrated As-built Simulation?, ASHRAE Transactions-Research, Vol. 114, Pt. 2, pp. 375-388 (June ).

Song, S., Haberl, J. 2008. ?A Procedure for the Performance Evaluation of a New Commercial Building: Part II ? Overall Methodology and Comparison of Results?, ASHRAE Transactions-Research, Vol. 114, Pt. 2, pp. 389 ? 403 (June).

Haberl, J., Culp, C., Co-authors, ?Chapter 27: Monitoring and Verification,? Energy Management Handbook, 5th Edition, Wayne Turner, Editor, Fairmont Press, New York (2005), pp. 715 ? 762.

Im, P., Haberl, J. 2010. ?Analysis of the Energy Savings Potential in K-5 Schools in Hot and Humid Cliamtes: Application of High Performance Measures and Renewable Energy Systems?, Proceedings of the SIMBUILD 2010 Conference, International Building Performance Simulation Association ? USA, New York, N.Y. (August).

Cho, S., Haberl, J. 2008. ?Validation of the eCALC Commercial Code-Compliant Simulation Versus Measured Data from an Office Building in a Hot and Humid Climate?, Proceedings of the 16th Symposium on Improving Building Systems in Hot and Humid Climates, Texas A&M University, Dallas, Texas, published on CD ROM (December).

Im, P., Haberl, J. 2008. ?Analysis of the Energy Savings Potential in K-5 Schools in Hot and Humid Climates?, Proceedings of the 16th Symposium on Improving Building Systems in Hot and Humid Climates, Texas A&M University, Dallas, Texas, published on CD ROM (December).

Haberl, J., Davies, H., Owens, B., Hunn. B. 2008. ?ASHRAE?s New Performance Measurement Protocols for Commercial Buildings?, Proceedings of the 8th International Conference for Enhanced Building Operation?, Berlin, Germany, published on CD ROM (October).

Kissock, K., Haberl, J., Claridge, D. 2003. ?Inverse Model Toolkit (1050RP): Numerical Algorithms for Best-Fit Variable-Base Degree-Day and Change-Point Models,? ASHRAE Transactions-Research, Vol. 109, Pt. 2, pp. 425-434.

Haberl, J., Claridge, D., Kissock, K. 2003. ?Inverse Model Toolkit (1050RP): Application and Testing, ASHRAE Transactions-Research, Vol. 109, Pt. 2, pp. 435-448, 2003.

Sreshthaputra, A., Haberl, J., Andrews, M. 2004. ?Improving Building Design and Operation of a Thai Buddhist Temple,? Energy and Buildings, Vol. 36, pp. 481-494.

Abushakra, B., Haberl, J., Claridge, D. 2004. ?Overview of Literature on Diversity Factors and Schedules for Energy and Cooling Load Calculations (1093-RP),? ASHRAE Transactions-Research, Vol. 110, Pt. 1 (February), pp. 164-176.

Claridge, D., Abushakra, B., Haberl, J. 2003. ?Electricity Diversity Profiles for Energy Simulation of Office Buildings (1093-RP),? ASHRAE Transactions-Research, Vol. 110, Pt. 1 (February), pp. 365-377.

Reddy, T., Haberl, J., Elleson, J. 1999. ?Engineering Uncertainty Analysis in the Evaluation of Energy and cost Savings of Cooling System Alternatives Based Upon Field Monitored Data,? ASHRAE Transactions-Research, Vol. 105, Pt.. 2, pp. 1047 - 1057 (June).

Haberl, J., Bou-Saada, T. 1998. ?Procedures for Calibrating Hourly Simulation Models to Measured Building Energy and Environmental Data,? ASME Journal of Solar Energy Engineering, Vol. 120, pp. 193 - 204 (August).

Haberl, J., Abbas, M. 1998. ?Development of Graphical Indices for Viewing Building Energy Data: Part 2,? ASME Journal of Solar Energy Engineering, Vol. 120, pp. 162 - 167 (August).

Haberl, J., Abbas, M. 1998. ?Development of Graphical Indices for Viewing Building Energy Data: Part 1,? ASME Journal of Solar Energy Engineering, Vol. 120, pp. 156 - 161 (August).

Haberl, J., Thamilseran, S., Reddy, T., Claridge, D., O?Neal, D., Turner, D. 1998. ?Baseline Calculations for Measuring and Verification of Energy and Demand Savings in a Revolving Loan Program in Texas,? ASHRAE Transactions-Research, Vol. 104, Pt. 2, pp. 841 - 858 (June).

Haberl, J., Thamilseran, S. 1998. ?Predicting Hourly Building Energy Use: The Great Energy Predictor Shootout II: Measuring Retrofit Savings,? ASHRAE Journal, Vol. 40, No. 1, pp. 49 - 56 (January).

Reddy, T., Haberl, J., Saman, N., Turner, W., Claridge, D., Chalifoux, A. 1997. ?Baselining Methodology for Facility-Level Monthly Energy Use - Part 2: Application to Eight Army Installations,? ASHRAE Transactions-Research, Vol. 103, Pt. 2, pp. 348 - 359 (June).

Reddy, T., Haberl, J., Saman, N., Turner, W., Claridge, D., Chalifoux, A. 1997. ?Baselining Methodology for Facility-Level Monthly Energy Use - Part 1: Theoretical Aspects,? ASHRAE Transactions-Research, Vol. 103, Pt. 2, pp. 336 - 347 (June).

Ruch, D., Chen, L., Haberl, J., Claridge, D. 1993. ?A Change-Point Principal Component Analysis (CP/PCA) Method for Predicting Energy Usage in Commercial Buildings: The PCA Model,? ASME Journal of Solar Energy Engineering, ASME, New York, N.Y., Vol. 115, pp. 77 - 84 (May).

Haberl, J., Bronson, D., Hinchey, S., O'Neal, D. 1993. ?Graphical Tools to help Calibrate the DOE-2 Simulation Program to Non-weather Dependent Measured Loads,? ASHRAE Journal, Vol. 35, No. 1, pp. 27 - 32 (January).

I also teach calibrated simulation to the students that take my graduate simulation class. In a nutshell, you can do so by using the following simple rules-of-thumb (these are based on Norford/Hseih's work in N.J.):

1. Force the electrical loads to equal whatever data you have, i.e., hourly is best, or use daily or monthly (i.e., see the new ASHRAE Research Project by Abshakra and Reddy for more on this). Force the occupancy and internal loads as well. Shut off infiltration if there is an AHU.

2. Use real weather data (i.e., roll you own weather file), or use the closest TMY or TMY2 or TMY3 files (but do not match the temporal data).

3. Get the average hourly weekday/weekend electrical profiles to match.

4. Use daily heating and cooling from the appropriate hourly report vs real daily data plotted against average daily temperatures. Turn everything on in the simulation (i.e., do not use occupied unoccupied schedules unless you're really sure of this).

5. Look at the results (graphical) and take a good guess at what to tweek, repeat 1 - 4 as needed.

6. Spend a little time as needed to get the envelope described. Use a simple structure since systems and plant will change answers by +-100 - 500%, versus architectural features that change things by only 20 - 50%.

7. Ask for help.

Finally, follow the +- differences in Guideline 14 or the IPMVP and don't forget that your building could be messed up, which means that your simulation will needed to be messed up to match it.

We're currently working on automated calibrated methods for residential, with some thoughts about the same for commercial. Others have similar tools as well (i.e., see the ASHRAE literature for these).

Hope this helps.

Jeff S. Haberl, Ph.D.,P.E., FASHRAE

Jeff Haberl2's picture
Offline
Joined: 2011-10-02
Reputation: 200

Hayes --
When calibrating residential models, I find that tweaking things like shade
schedules, internal mass, and solar mass coupling is quite helpful. You
also need to implement some natural ventilation if that's happening. Also,
you should always distrust your measurements until you have a better feel
for what the error bound is like. How did you measure ACH? With a blower
door? You'll be lucky to have the infiltration component right to within
30%. Do your wall constructions compensate for two dimensional heat
transfer effects? How well do you think you have these quantified? What
about your ducts? And your air conditioning equipment? Don't overconstrain
yourself.

Remember also that you are using a simulation model, not actually solving
the complete set of heat transfer equations. The model is just a model --
it won't calculate the correct total energy consumption unless
you calibrate it to do so. Remember that different simulation software
modeling the same highly simplified building test cases vary by 10-20% or
more (see BESTEST results comparing ESP-R, EnergyPlus and various DOE2
generations). Simulation models work great for showing relative differences
in energy consumption, not so much for absolute energy consumption.

My general approach for residential building calibration looks something
like this:

1. Build out your model using your best available data.
2. Nail the lighting and equipment loads.
3. Nail the hot water loads.
4. Compare your annual and monthly heating and cooling consumption to
measured and look at the direction each need to move, then start making
tweaks to your model. I try to turn all of knobs that have similar impacts,
and keep them all in what looks like a reasonable range.
5. Take a look at the hourly profiles for different seasons and look at
interactions of solar, mass, and natural ventilation. You can tweak these
to shift your shape some.
6. Last, I do what feels like the big cheat, which is to start messing
with the daily and monthly thermostat schedules to get things fine-tuned. I
can generally keep things within a degree or two of measured temps. This
step will almost certainly be necessary to get your results lined up.

In the end, I think it's fairly difficult to get a residential model
calibrated to the hourly CV(RMSE) of 30% for the HVAC end uses, even for an
experienced residential modeler. If your target is based on total
consumption, then you should not have a problem if the non-HVAC loads are
smooth enough. I don't know for sure, but it appears Guideline 14
simulation guidance was not developed with calibrating to a single
residential single family home in mind. It's not something people really
do. We calibrate to models of aggregations of homes when we're evaluating
residential energy efficiency programs for utilities.

I'd be very curious to hear about other calibration methods/standards used
by experts in the field. Prof. Haberl's response was great.

Justin Spencer's picture
Offline
Joined: 2011-10-02
Reputation: 1

Justin,

Interesting comments on calibration, especially for residential. I'm afraid my previous comments were slanted towards commercial/institutional buildings in cooling dominated climates. Hence, I did not mention DHW, nor some of the finer details that you suggested.

We've been developing a method that follows the BESTEST EX protocols, and a statistical database of building characteristics, to construct a first "blind" pass of a home, knowing only the square footage, zip code and year built, we assume all the rest of the characteristics for the house from the database and usually get close to the actual heating/cooling energy use (surprise) of the house. Early work on this can be found in the MS thesis by Piljae Im. Our current work is similar to what Prof. Kissock is doing at U. Dayton for his model-driven, residential audits.

Jeff S. Haberl, Ph.D.,P.E., FASHRAE

Jeff Haberl2's picture
Offline
Joined: 2011-10-02
Reputation: 200

RE: [Bldg-sim] Calibrated Simulations & ASHRAE G14

There are probably occupancy variations that took place in the house that
cause variance in the energy usage for the measured case, and it may be
difficult for you to capture those effects in the model.

Other building occupancies (commercial, institutional, etc.) stick more
closely to their typical usage patterns and the benchmark CV should be
achievable in most cases.

It should be for residential also if you have enough data about the usage
to tailor the operating schedules.

You might want to start out with a gas usage and HDD regression to identify
if there are any months that jump out as atypical from your modeling,
irregular usage periods, etc.

DHW could be significant also, maybe you can get water usage data for the
house to see if there were periods when occpuants were using more
showers/sinks/laundry/etc.

One other idea might be to see how sensitive the CV is to the measured ACH,
which could vary during times of the year, not sure if your measurement
covered multiple data points (summer/winter/occpupied/unoccupied).

Hope this helps!

David S. Eldridge

David S Eldridge's picture
Joined: 2011-09-30
Reputation: 2000

Hi Justin,

Great response - thanks! I think you brought up a few very important points that I may not have given enough consideration to. Such as the infiltration model, I am using a blower door test ACH rating @50Pa, then converting to regular pressure by dividing by 20 (KP model). Then in EnergyPlus I am using the coefficients from DOE-2 (only a windspeed coefficient) - maybe this is unrealistic as this model would have no infiltration when windspeed = 0.

I also assumed ACH of the attic = 10 at 50 .. not sure if this is realistic, and not sure if its significant. And I have no air exchange between the basement zone and the main zone. I did not account for 2d heat transfer effects, but did compute the overall thermal resistances taking all framing members into account (1/Rt = 1/R1 + 1/R2). Do you think that doing a THERM analysis would increase the accuracy significantly?

What do you mean by solar mass coupling?

I am only calibrating the gas consumption, and the home is a research facility with simulated occupancy (appliances and lights that turn on and off according to a computer program). So it is a "best case" modelling scenario, where I know a lot more about the house than what would typically be known or available.

Interesting point abou the BESTEST results. From that I wonder what accuracy I can assume EnergyPlus has?

Thanks for your help,
Hayes

Hayes Zirnhelt's picture
Offline
Joined: 2011-10-02
Reputation: 0