Appendix G (LEED Energy) Recipe and Ironbug Integration

Hello All,

We’re happy to announce a couple of new features across the Pollination ecosystem, which are complementary to one another.

Appendix G Performance Recipe

After many of you have asked us for it, we have finally published a recipe to Pollination that can estimate the number of LEED “Optimize Energy Performance” credits that a particular model can achieve! The recipe is called Appendix G Performance and you can find it here on the Pollination platform or by searching for available recipes with the PO Setup Runs component. Just as the name of the recipe implies, it generates a baseline Model (as a HBJSON) from the input Model, which is consistent with ASHRAE 90.1 Appendix G 2016 (and later). This includes adjusting the geometry, constructions, lighting, HVAC, SHW, and removing any clearly-defined energy conservation measures like daylight controls. The recipe even performs an initial EnergyPlus sizing run on the baseline model to ensure that all HVAC equipment efficiencies align with the 90.1 standard.

After the creation of the baseline model, the recipe executes an annual EnergyPlus simulation, performing 4 separate simulations of the baseline model in parallel for each of the 4 cardinal directions per the Appendix G specification. Alongside these baseline simulations, the input Model is also simulated to get the energy performance of the proposed building. At the end, all energy use results are post-processed along with the energy costs inputs to estimate the Appendix G Performance Cost Index (PCI) for the latest 3 versions of ASHRAE 90.1 (2016, 2019, and 2022). An additional step is run to estimate the number of LEED “Optimize Energy Performance” points for LEED v4 (and 4.1), using inputs of electricity carbon intensity alongside the energy cost inputs.

The ultimate output of the recipe is a couple of JSON summaries, which note the Appendix G performance and the LEED Energy performance. The first file is called appendix_g_summary.json and a sample of this is visible below (note the PCI improvement in percent for the latest versions of ASHRAE 90.1):

    "proposed_eui": 112.866,
    "proposed_energy": 3517144.444,
    "proposed_cost": 703428.89,
    "baseline_eui": 235.3,
    "baseline_energy": 7332474.306,
    "baseline_cost": 1214797.19,
    "pci_t_2016": 0.666,
    "pci_t_2019": 0.591,
    "pci_t_2022": 0.574,
    "pci": 0.579,
    "pci_improvement_2016": 13.055,
    "pci_improvement_2019": 2.0219,
    "pci_improvement_2022": -0.880

You will see that it is possible to get “negative improvement”, especially in the later versions of ASHRAE 90.1, where the performance targets are strict. This isn’t always an indication of a poor design since the Appendix G PCI metric is highly susceptible to relative cost differences between energy sources like electricity vs. natural gas. So we recommend that users take care with the energy cost inputs for the recipe and set them to the real values of their project, even though the recipe has reasonable defaults that align with average US prices.

The recipe also outputs a file called leed_summary.json, which contains the
ASHRAE 90.1-2016 PCI for both cost and carbon (GHG) emissions in the format below.

  "proposed_eui": 112.866,
  "proposed_cost": 703428.89,
  "proposed_carbon": 464263.067,
  "baseline_eui": 235.3,
  "baseline_cost": 1214797.19,
  "baseline_carbon": 1577657.766,
  "pci": 0.579,
  "pci_target": 0.666,
  "pci_improvement": 13.055,
  "carbon": 0.294,
  "carbon_target": 0.633,
  "carbon_improvement": 53.511,
  "leed_points": 9

Note the total number of LEED points at the very end is out of the total possible (16, 18, 20) that can be achieved under the “Optimize Energy Performance” category. Altogether, the execution of the recipe looks like this from Grasshopper:

We have provided a Grasshopper sample of a full-building Large Office model (built with Dragonfly) for those who wish to get started testing the recipe. NOTE THAT YOU WILL NEED TO HAVE POLLINATION RHINO 1.25.7 OR ABOVE INSTALLED TO CORRECTLY USE THE RECIPE. You can get this either by downloading and running the latest Pollination Rhino installer or by running the LB Versioner component from Grasshopper. Hopefully, it also goes without saying that the recipe is intended for full-building models that have all of the critical energy simulation properties assigned to them (ConstructionSets, ProgramTypes and detailed HVAC Systems). The recipe won’t fail for cases like a model with ideal air load systems but the results aren’t likely to be meaningful given that an ideal air system is not a real piece of HVAC equipment and it reports all energy usage as district heating/cooling.

All of the methods that the recipe uses are open source and those of you familiar with Appendix G modeling probably already know that we are relying heavily on the OpenStudio Standards gem in order to build the baseline HVAC systems and assign efficiencies to all of the HVAC equipment. So, as is typical with many of our releases, a great deal of thanks is due to the National Renewable Energy Lab and the OpenStudio team. The Ladybug Tools methods that edit the building geometry and assign baseline constructions are also open source and many of them can be found within this new honeybee_energy.baseline subpackage. Lastly, the logic of the recipe itself is open source and is available here under the Pollination Github.

We would like to emphasize again that Appendix G in ASHRAE 90.1 2016 (and later) sets a relatively high standard of energy efficiency and is close to aligning with the AIA’s goal of carbon neutral practices by 2030. When using only the detailed HVAC template systems that are shipped with Ladybug Tools and Pollination Rhino, it can be very difficult to get all available LEED points. For the Large Office sample model provided above, we were not able to get more than 9 LEED points after applying the best energy conservation strategies that we know. It’s likely that, to get all available LEED points, some amount of on-site renewable energy may be necessary for certain buildings. We are working towards a specification of photovoltaic panels in Honeybee that will enable this but, in the meantime, there are likely still a few more LEED points that can be achieved by fine-tuning the efficiency of HVAC equipment and implementing strategies beyond what the template HVAC systems allow. For this reason, we are particularly excited to announce this Appendix G capability alongside a new-and-improved Ironbug integration!

Ironbug Integration with Honeybee Models

Many of you are already familiar with the Ironbug plugin that @mingbopeng has developed and maintains. The plugin provides over 100 components for customizing HVAC systems to a high level of detail, effectively providing the full set of capabilities of the OpenStudio SDK for HVAC. In the past, it was possible to use Ironbug with certain Pollination recipes that could accept the Ironbug-edited OSM file as the input model but other recipes like the comfort mapping and the Appendix G recipe above require a model in Honeybee (HBJSON) format because they pre-process the model with the Ladybug Tools core Python libraries before they simulate them in engines like EnergyPlus/OpenStudio.

For these reasons, it’s with great pleasure that we announce Ironbug has been fully integrated into the Honeybee Model schema (aka. HBJSON). This means that you can assign your Ironbug HVAC to Honeybee Rooms using a new LBT-Grasshopper component called “HB Detailed HVAC’’ like so:


The resulting Ironbug HVAC will be automatically included in any OpenStudio/EnergyPlus simulation of the Honeybee Model and a full specification of the Ironbug HVAC is included in the Honeybee Model JSON (HBJSON) that gets uploaded to the Pollination platform or any online simulations. This enables you to set efficiencies of HVAC equipment down to a very fine level of detail and model many more energy conservation strategies than the template HVAC systems allow. By modeling things like high-performance chillers/ heat pumps, adding heat recovery strategies like run-around loops, and using thermal resources like ground heat exchangers, many of the LEED credits for the recipe above can be achieved.

This is our first step to provide better solutions for modeling detailed HVAC systems as part of our development roadmap for 2023. In the next few weeks, we will be working towards an even stronger integration between Ironbug and the Pollination Rhino plugin, effectively making Ironbug and Grasshopper the default interface for building and editing detailed HVAC systems for your Pollination Rhino models. The video below provides a sneak preview of what to expect with new HVAC integration and how we’ll be making use of Grasshopper as a dedicated HVAC editor.

So stay tuned and, as always, please feel free to open a new discussion topic on this forum for any questions or feedback on the new capabilities!


@chriswmackey Thank you Chris for this amazing update!! this is literally making history, I am really excited to test it. On a side note, is there any plans to include 90.1- 2010?

Thanks, @moelsayed ,

No, not for Appendix G modeling. You can probably still create a pretty close approximation of a 90.1-2010 baseline building using the appropriate ConstrucitonSet/ProgramType/HVAC System along with the new efficiency standard input for the “HB Sizing Parameter” component. But I don’t know of any active green building standard that’s still using the Appendix G from 2010 so this wasn’t really on our agenda. LEED v4 is still on 2016 and most states that regularly maintain their building energy codes are at least at 2013.

What is the reason why you need it for 2010? Are you operating in a state with an old building code and there’s some local provision for an Appendix G analysis?

Hi @chriswmackey,
LEED v4 uses 2010. You can use 2016 under the v4.1 substitute but there are some projects that would result in less LEED points using 2016 versus 2010. Most LEED v4 Baselines are using ASHRAE 90.1-2010 Appendix G.

1 Like

Thanks, @justinshultz .

This is something I didn’t realize since “Optimize Energy Performance” in LEED v4.1 is explicit that 90.1-2016, Appendix G must be used but there isn’t a specific version of ASHRAE 90.1 cited under Option 1 of the same credit in LEED v4. I guess they mention 90.1-2010 somewhere else in the v4 documentation.

It probably goes without saying that 2010 was significantly less strict than 2016 and so the building’s pretty much always going to look better when compared against a poorer-performing standard. But I also see that there’s no evaluation of carbon emissions in the older credit, meaning that electrified buildings will usually look much better in v4.1 given that electricity tends to be more expensive but less carbon intensive than on-site natural gas (at least in the US).

Given that the older credit is kinda incentivizing on-site consumption of fossil fuels, I have to imagine that the USGBC is planning to put a sunset on the ability to submit with the older version at some point.

We can certainly make a version of the recipe that uses 2010 but it’s a lot of work because the definition of a baseline building was not as standardized back then. So I wouldn’t really want to sink so much time into this if the option might not be around that much longer. An alternative would be to put together a sample file that shows how to to use the measure cited above to make a baseline model. It won’t have all of the fanciness of the recipe, with it’s 4 simulations from the different cardinal directions but you could do this yourself with a little extra time and post-processing.

Would you have time for a call about this at some point? It’s just hard to get a sense of how much a 2010 version would get used without a conversation. I sent you a PM about it in the event you’re available.

1 Like

Hi @chriswmackey ,

In v4 EAc1 references EAp2 with the exact 90.1 version.

It says “Follow the criteria in EA Prerequisite Minimum Energy Performance to demonstrate a percentage improvement in the proposed building performance rating compared with the baseline. Points are awarded according to Table 1.”

EAc1 (Optimize Energy Performance) and EAp2 (Minimum Energy Performance) are closely linked, they also share the same worksheet that you need to submit. In the Minimum Energy Performance description you can find:

“Calculate the baseline building performance according to ANSI/ASHRAE/IESNA Standard 90.1–2010, Appendix G.”

I cited these from the LEED v4 New Construction guide:

1 Like

Thanks for the clarification, @furtonb . It’s good to know where it says it.

Would you say that you also find yourself in a position of using LEED v4 instead of v4.1 because you find the older version usually gives you more credits?

1 Like

Yes, same here. Point-hunting FTW! :face_with_peeking_eye:
I don’t have any new LEED projects at the moment, so I can’t say if it would be v4 or later for newer projects.

I would also add in familiarity, consultant companies might have workflows established for v4 and want to stay in comfortable waters. I can give a call to the guys that did the certification procedure, if that would help. The company, that I was working at, was involved as energy modeller with a limited scope, the LEED version decision was already made when we got appointed for the last LEED project.

I don’t know how long v4 and 90.1-2010 will stay relevant, though. This year a new version of 90.1 should be released, right? It would be quite ridiculous to keep an almost 15 year old standard alive for much longer, but someone more familiar with USGBC strategy should chime in, as I’m located in Europe, so my use of 90.1 is solely for certifications at the moment.

For another reference, the baseline building implementation instruction in DGNB’s ENV 1.1 criterion (Life cycle assessment) references 90.1-2013 (or latest):

“Baseline building design characteristics in accordance with the Appendix 5 or characteristics specified in Section G3 of the ASHRAE 90.1 – 2013 (or latest) standard.”
Page 102 in the following PDF:

I suppose there are hints in BREEAM and maybe WELL too, but I’m not sure about the latter, as I’m not really familiar with those frameworks.

As far as I’m understanding, 90.1-2016 and 90.1-2019 are not so different, so for DGNB projects the implemented recipe would be fine (I don’t have a comparison table).


I just wanted to say thanks for the explanation, @furtonb, and I agree that it’s odd to be using such an old standard. It’s not even so much the absolute age of the standard but the fact that there have been 4 revisions since then, which have some pretty important improvements.

Largely because of one of those improvements, the current recipe calculates the performance in all recent ASHRAE 90.1 versions - 2016, 2019, and 2022 (and you are right that 2022 has been released already). We should also easily be able to update it for future versions of 90.1 given that the definition of the baseline model has been standardized since 2016.

But I can also recognize that the transition with LEED is likely to be very gradual and we’ll need something that addresses the LEED v4 / ASHRAE 90.1-2010 issue. At the very least, I will make sure that people can run the “Create Baseline Building” measure that NREL maintains and I’m continuing the conversation about this measure on the original thread that asked about it here:

If the USGBC continues to let people use v4 for submissions, I think we may eventually need to release an Appendix G 2010 recipe since I recognize the measure is pretty painful to use. For those people desiring this type of recipe, I would ask that they give us feedback on the current Appendix G recipe since the experience for a 2010 recipe is only going to be as good as the current one.

I think I may draw a line in the sand with 90.1-2013, though, since creating the baseline building for this version is very complicated. I’ll help people run NREL’s measure for the 2013 case if they needed but I don’t see us sinking the large amount of time it would take to make a 2013 recipe. At least not any time soon.


4 posts were split to a new topic: 1. Handler-python: cannot import emissions from pollination_handlers.inputs

hi, @chriswmackey is it possible to assign an osm file to this recipe? Is there a way I can convert it to hbjason first?

Hey @eanvari ,

The answer is unfortunately, no, the Appendix G recipe does not allow you to connect an OSM as input since we can’t reliably make a baseline model from any OSM. It can do it from any HBJSON, though, so you can try importing the OSM to HBJSON through the Rhino plugin or with the HB Load gbXML OSM IDF component.

Just be aware that the process of importing an OSM to HBJSON is lossy and you will loose things like the HVAC systems from the OSM. This matrix shows everything that can be imported. But you can always try to re-assign the HVAC after importing it. This is why we announced the Appendix G recipe alongside IronBug integration since this now allows you to specify any details of the HVAC system equipment in HBJSON format.

1 Like

Thank you @chriswmackey. I get the point, the matrix is so helpful for my work.

Excellent thread and very lively discussion, thanks for the hard work on this Chris!

Adding to the conversation, a couple of useful resources:

  1. Alternative Performance Rating Method
  1. LEED Interpretation 10481 - v4 BDC IDC equivalency for ASHRAE 90.1-2013

Wondering what are some of the testimonies on the use of this recipe?

Will this recipe work with Pollination grasshopper or do you need to purchase Pollination Rhino?

Hi @pushkarnachan! Welcome to the forum. All the recipes are accessible from the Grasshopper plugin. :slight_smile: