Author Archives: Max

Mr.Comfy in IBPSA News 24 4

It’s a great (late to be posted here!) pleasure to announce that IBPSA once again published an article written by me about the latest Mr.Comfy installment released this summer. Read the full article here: Special thanks again to Christina Hopfe, who has graciously supported this project for quite some time now – it’s a great pleasure for me to make a contribution.

In other news, Mr.Comfy development is set to resume sometime late this year / early next year, when new features will be added pending a couple new papers to be written. Development has been on the backburner for a while now while I tried to secure a new professional position, which is still an ongoing quest- need someone? Let me know.

I’ll publish the new version and upcoming feature set description once it has been more formalized on this blog.



Mr.Comfy Call for Example Work

A shot into the dark, this, but since Mr.Comfy has been out there for about a year by now, I’d be very curious what other professionals have created in terms of analysis and/or visualization with the software. If you would like to share results to be published on this site and shared with a wider audience, please contact me and I will gladly put it out there.

I, too, have created some analysis narrative sets with Mr.Comfy  in my recent consulting role and will share that once it’s “official” and the rights have been cleared, which is a difficulty I can hence appreciate. Note that I have no problem with giving your firm coupled exposure if you have something to share, so please don’t hesitate to ask.

IBPSA NYC video posted | Update Schedule

The video grab of my talk “Spatial Analysis of Thermal and Daylight Simulation Data with Mr.Comfy” has finally been made available on the IBPSA NYC chapter’s Vimeo channel. You can download the full slide set here.

I would once again like to take this opportunity to thank Pallavi Mantha of the IBPSA NYC chapter for the kind invitation; it has been a lot of fun to present (even if it was just remotely) to such an interested audience.

On a related note, Mr.Comfy development is on a temporary hiatus while I juggle several jobs at once, so as it stands now no further updates are to be expected at least until early next year. This is unfortunate, but the to-do list and feature requests keep on growing so that once development continues, expect some good things to come out :) An additional writeup of the current feature set is expected to be published in the next iteration of IBPSA News, which I will post here once available.

Mr.Comfy New Features Presentation posted

I’m excited to have been given the opportunity by Baumann Consulting USA’s Denver, CO office to put together a little overview & new features slide set presented to their US energy modeling team. You can download the whole presentation here (PDF, 20mb), titled “Not (only) Visualization: Model-embedded Simulation Data Analysis + Metrics Prototyping with Mr.Comfy”.

As the title suggests, the presentation touches on the new variable prototyping feature by showing a few experimental new metrics such as “Discomfort Solar Gains” and “Multivariable Heat Discomfort”, which were generated in Mr.Comfy from existing variable data through custom expressions and are especially applicable for shading design and unmet load hour explorations. A lot of the custom metrics discovery work is ongoing, so no claim to super metrics discovery fame here, but I still consider them interesting enough to merit sharing them with a wider audience.

Multivariable Heat Discomfort Spatial Mapping of EnergyPlus custom metrics variableSpatial Visualization of Daylight Intensity (log) Distribution and Custom Variable “Multivariable Heat Discomfort”. Monthly breakdowns, occupied hours only.

I will make a more detailed blog post about the metrics in the time to come, however from the few additional tests I’ve made so far it is obvious (at least to me) that stringing together multiple operators to e.g. have the analysis default to the most severe comfort indicator in a group of several makes sense, as they all react differently to individual influences such as activity, radiant temperatures etc.

As another result from the talk, improvements in workflow and imports has definitely made it more strongly onto my agenda; there were plans to add direct *.idf and *.eso parsing, anyways, but good to have this requested from a professional consulting audience, too. Thanks once again to Amir Bazkiaei of Baumann US for the invitation, I’ve had a lot of fun and hope to see some results generated by the team with Mr.C in the future.

Daylight Models in the Daylighting Handbook I

I’m completely stoked that the esteemed Prof. Christoph Reinhart of MIT’s Sustainable Design Lab has included rapid-prototyped physical daylight metrics models created in some of my past classes at the TU Berlin into his recently released book, “Daylighting Handbook I, Fundamentals: Designing with the Sun“.

Thank you, Christoph, it’s a big honor! If you, dear reader, are into daylight simulation, I can’t recommend Christoph’s work highly enough- as the father of Daysim, everyone in the field is probably using his software in some shape or form. Please support him by possibly adding that book to your library, and also check out MIT’s Daylighting OpenCourseWare class, which makes a huge contribution to making information about daylighting in buildings available to a wide audience.

Daylighting Handbook  I RP model shot

Rapid-prototyped Physical Daylight Model with Embedded DAv 300 Metric; design & modeling: Piotr Jardzioch, Jakub Sobiczewski; Prototyping: M C Doelling. Book snapshot courtesy of A Jakubiec.

Mr.Comfy 0.21 released | Custom Spatial Thermal Metrics & Independent Daylight Visualization

Mr.Comfy 0.21 brings two new main features: custom thermal variable expressions and improvements in how the daylight mapping functionality can now be used more independently of the thermal mapping inputs, requiring only a set of Daysim *.pts and *.ill files to be provided. See the full list of changes here. The new metrics functionality is probably more exciting, so let’s see what that is all about:

One of the original development drivers behind Mr.Comfy has always been the insight that specific combinations of design intent, building typology and climate require custom analysis scenarios to achieve highest projected early design-stage building performance. Environmental design procedures should be uniquely tuned to marry process with data interpretation and generation scenarios, and link performance display to geometry. That’s the interstice in which the tool operates; the newest version goes a step further by allowing custom report variable generation from existing E+ report outputs.

CoolingSeasonExcessSolarGainsMonthly performance map of “Zone Discomfort Excess Solar Gains” while space cooling active, custom thermal variable

Take the above example of the “Cupertino” sample building simulated in South Florida’s subtropical climate. A common design question in such a scenario is whether the shading strategy in combination with active cooling always delivers occupant comfort; the custom metric “Zone Discomfort Excess Solar Gains”, displayed above for each zone and month of the year, reveals that due to lower winter sun angles and still quite high outside air temperatures, several building spaces exhibit discomfort. A custom variable expression that only writes solar gains to a new data field if either the zone air temperature is above 26°C or Pierce TSENS thermal comfort index is higher than or equal to 1.8, which indicates heat discomfort, makes it possible to identify the exact location and point in time when this behaviour occurs- which can then e.g. be fixed through different shading strategies. Custom variables thus make it possible to design spatially visualized thermal metrics that are specifically tuned to the climate and building typology at hand. How they work in practice is documented here.

I hope that users will come up with their own metrics that you’re then willing to share here; I’d love to see some applied examples that are not as, well, constructed as the one above :)

The remaining updates for the 0.21 release concern improved thermal display processing speed, fixing the DIVA VIPER thermal parser to work with daylight co-display, and various little addons and improvements. As quite a few things changed under the hood, if you find any bugs, please just let me know.


BSO’14 Paper & Optimization Case Study

Monthly Zone Air Temperature Visualization (from EnergyPlus Data)

Top floor monthly average zone air temperatures (daytime schedule), original studio design state by Christopher Sitzler & Laura de Pedro.

It has been a long time in the making, but I’ve finally presented my paper “Space-Based Thermal Metrics Mapping for Conceptual Low-Energy Architectural Design” (yes, a mouthfull of a title, that) at Building Simulation & Optimization 2014, UCL, London. Download the thing here or check out the scribd preview. It’s somewhat of a milestone for me, as no prior refereed publication about Mr.Comfy existed- and that’s important to achieve.

Apart from a walkthrough of Mr.Comfy features related to architectural performance design cognition (you know, how the colors make sense and why), the paper contains yet another student design optimization case study that I’ve withheld while waiting for the paper to make it out there. Christopher Sitzler and Laura de Pedro did a stellar job not only in the original “Robust” integrated studio, where the design originated, but Christopher also volunteered to further optimize thermal design performance in the followup Mr.Comfy prototyping class, for which he has my deep gratitude.

Top Floor Performance AxonometricTop floor multi-variable performance mapping, base vs. adapted design state Mr.Comfy & DIVA4Rhino visualization. Design: Christopher Sitzler & Laura de Pedro.

So that I don’t retell the paper’s content here, suffice to say that through spatial thermal mapping and analysis, Christopher managed to squeeze performance improvements out of a design that was originally created in an already simulation-embedded design studio and overall did not perform badly at all. Through geometric building fabric adaptations and select building assembly material improvements, especially the glaringly uncomfortable top building floor of a 50-zone design was much improved; shown above are both base and further optimized design states mapped during summer, when daylight utilization, thermal comfort and cooling load needed a lot of work. Ironically, the somewhat dramatic performance problems were originally masked when only looking at whole-building averages, as the top floor is only a relatively small part of the total building area.

All-floor total heating energy use

All-floor spatial visualization of annual total heating energy use, base vs. adapted design state, Mr.Comfy display of E+ data. Design: Christopher Sitzler & Laura de Pedro.

The remaining floors were primarily re-optimized to lower heating energy consumption, as they did not suffer from the dramatic summer overheating apparent in the top floor. The visualization is quite interesting insofar as it not only shows what was changed, but how zones are mutually influential; compound changes are usually hard to spot in their “thermo-spatial” effects and commonly occur in rapid design iterations. Charted in a traditional fashion, total changes in energy use are summarized below.

Base vs. adapted design total energy use

Stacked “traditional” bar chart of total annual heating, cooling, lighting energy use, base vs. adapted design state

The paper finally goes into some detail regarding how students use the tool, and how they thought that spatial mapping improves design cognition (summary: yes, they think it does). Many of the features added to Mr.Comfy since are not included in the publication (unfortunately..) but will be further discussed in a journal paper, once I get ’round to it and figure out what’s next with me.

Thanks for reading~ enjoy.

Ongoing site / doc. updates for 0.21 release

[Edit: 0.21 is out. Blog post/announce etc. later]

As 0.21 is almost done, adds a few new features and will be released soon, you will see the documentation, input glossary etc. undergo some gradual updates in the coming time- hence not all the stuff described in the docs will refer to the still current 0.20 version- sorry for any confusion this may cause. I’ll push the update out as soon as I can, hopefully before BSO14.

What will be in 0.21?

- Custom thermal metrics prototyping through user expressions
- Independent daylight mode (no thermal data / workarounds needed anymore)
- Updated VIPER thermal parser to work with daylight tools
- log daylight display
- Single frequency mode for thermal & daylight components
- Thermal display performance increases (up to ~3x and beyond)

Environmental Performance Intent & Simulation Data Visualization (Case Study)

One of the primary drivers behind the development of Mr.Comfy has always been the idea that by making the invisible visible in a spatial format, design performance thinking and its underlying mental models can be enriched to gain greater intuition (or “tacit” knowledge) – which then aids the construction of explicit knowledge, as e.g. argued by Friedman (2003, see discussion in this paper).

Performance representations, especially when distributed over the incomplete scope of what individual calculation engines can resolve, never fully encapsulate that which a building is intended to do; indeed, one can argue that the same is true for a building in other domains, e.g. the way different artefacts represent it never does justice to the totality of experiences created through a structure’s existence. Add aspects of being in-process, prior to completion, and the waters are even murkier- do tools and representations used to shape something in the making influence the possibility of its creation, of the way it is being “thought”?


South Florida Wildlife Center, 2009 – 2010, 2014 (Perspective Rendering, Author)

Positions on this in design research vary, however as far as the interaction of performance simulation, its data expression and influences on geometric optimization mediated through operative design thinking is concerned, it has always been my experience (and conviction, I admit) that yes, there is no escaping the tangible impact of different modes of model manipulation on design cognition.

Coincidentally, that is one of the reasons Mr.C’s inputs are designed to specifically answer question complexes such as “What is happening in a design, when do the behaviours occur, where do they occur, and how do they compare to simultaneous states in other parts of the intended building?” To answer these enables designers to find out why patterns exist, and through contextual cognition to influence them- quoted directly from my upcoming BSO14 paper on Mr.C, which I’ll share here once it’s published in the proceedings.

Performance representations that are intimately linked to the building geometry they investigate (or make possible) are not quite new- think, for a much overused example, of Gaudí’s catenary models (, or even graphic statics in general, which firmly occupy an almost uncanny position between drawing the intended and deriving its shape from rules – or physical laws – while you’re at it. Now do not get me wrong, I in no way intend to compare my small contribution to building analysis to Gaudí’s (and the Catalan engineers he worked with) genius innovations, but merely intend to through this point open up wider the discussion about the interface of design intent, the way it is encoded in models, including their subsequent representations, and “actual” performance as then derived from spatio-mathematical operations.

Indeed the way performance design intent is often encoded in early environmental design, which is a different beast from structural engineering due to its extreme temporal volatility, is through concept drawings that e.g. contain “magic arrows” of ventilation or shading design- such as the one below, taken from one of my own projects for a wildlife sanctuary in South Florida. Coincidentally, the common appearance of such hybrid drawings in the past TU Design/Sim classes had initially motivated to think up first Mr.C concepts (even though there’s plenty more examples of such representations in the literature, but hey, personal experiences always have a greater impact, don’t they?).

small_wccFormPerformance-2South Florida Wildlife Center “Magic” Ventilation & Shading Section (Author)

Thoughts on the representation/multi-domain performance complex first appeared in a completely forgotten paper for eCAADe 2012; in a sense, the Mr.C representations developed since then attempt to link an understanding of geometry with the explicit – that is, the calculated – knowledge of what the actual performance, mediated through geometric factors, really looks like. The resultant representations can be understood as “magic drawings on steroids”, especially if you consider that typically, in early design simulation models, boundary conditions and input data scope may be extremely limited.

Given later design and simulation stages, that limitation might obviously no longer apply, but I nonetheless suspect that the close relationship of “traditional” environmental design intent sketching and spatial sketch simulation data mapping breeds a greater cognitive proximity of early concept and validation levels, without the disconnect of switching representational domains (e.g. from drawing to chart, from space to abstract projection).


South Florida Wildlife Center Conceptual Energy, Ventilation & Daylight Data Visualization (Author)

As the reader might have picked up on, the previous sketch contained multiple performance domains (daylight/shading considerations and natural ventilation). The common occurrence of multi-domain intent sketching (not just my own, but also that of countless others) reveals the tendency of designers to mentally see buildings as totalities of form and sensual impacts, which the ambient environment co-determines. This observation is what in part and to this day drives the development of Mr.C’s features that allow for the simultaneous display of several datasets at once, including daylight, to construct a spatially displayed totality that might come close to what a designer might envision.

Through this and the facilitation of asking spatio-temporal questions through the tool, I hope I’ve made the point that to me, a visualization is not an illustration, but is “the use of interactive visual representations of abstract data to amplify cognition” (Ware, 2004 p. xvii) – and from understanding one state a designer becomes able to envision others. In one of the next posts, I’ll try to write something about how this type of “cognitive” approach might differ from “genetic” or “algorithmic” optimization – but that is for another time.

Ware, C. Information Visualization, Perception for
Design. San Francisco: Morgan Kaufmann,

Spatial EnergyPlus visualization optimization class result 02

It’s been a while since I updated the blog with any results of the past mapping/optimization class- mainly because I’ve put a lot of them into the EPFL/NYC IBPSA chapter presentation and didn’t yet get around to separate them out. The mapping/optimization case study I’m posting here now wasn’t included in said presentation- that’s why it’s next in line.


Ground floor base vs. adapted design state Mr.Comfy mapping, ‘ROBUST’ studio design, summer 2013. Design: Alan Patrick, Ismael Cárdenas. Simulations, Mapping: Alan Patrick, winter 2013/’14.

Download the whole case study here (PDF, 2.2 mb).

Alan Patrick’s & Ismael Cárdenas’ original design was created in a fully simulation-integrated design studio held in summer 2013 at the TU Berlin. The task was to design a mixed-use office & exhibition building in downtown Berlin, Germany, that sits on a heavily overshadowed, south-facing urban site. Heating energy use reduction is the primary driver in Berlin’s climate (and this particular typology), along with maximization of daylight utilization.

Mr.Comfy mapping of the resultant structure was performed in a followup class, to spatially identify further optimization potentials and visually check for errors in the large whole-building EnergyPlus model (in this case generated with DesignBuilder). Alan especially focused on changing aspects of the ground floor geometry (adding an unconditioned buffer space, increasing glazing U-values etc.) and testing two different variants of skylight geometry on the top floor.

Akin to the previous case study, a match was found between the designer’s assumed mental picture of building behaviour and performance states discovered through mapping; some surprises, however, remained, e.g. the impact of skylight geometry on residual cooling loads. Overall, dynamic visualization appeared to have a positive impact on performance cognition and optimization, squeezing a reduction of projected energy use out of a design that had previously been created with the help of design-centric optimization measures.

Download the whole case study here (PDF, 2.2 mb).