3 years ago we presented a talk called "The Future of BIM is NOT BIM, and It's Coming Faster Than You Think. A lot has changed in 3 years, so we felt it appropriate to have a sequel to that talk which aired at Autodesk University in November to highlight all the amazing technology that has come to pass. Please enjoy!
“The world is wide, and I will not waste my life in friction when it could be turned into momentum.” -France E. Willard
Many people including myself have had a hard time introducing new technology into their firm. When I was a BIM Manager, I recall one such case when I tried implementing Dynamo into the architecture firm I was working at. I had just got back from Autodesk University, and had attended Marcello’s class “Practical Uses for Dynamo Within Revit”. I was so inspired during the class, that I remember opening my laptop before Marcello had finished. I had run into that same limitation that so many of us had in Revit of tagging a calculated value. I decided to write a simple get parameter/set parameter script. I was so in the zone, that I remember everyone had left the class, and I was all by myself in the lecture room still coding this script. I was so proud of myself that I had overcome this limitation in Revit. I eagerly came back to show my manager the value of Dynamo and evangelized to him that all our staff needed to learn it.
I remember him telling me “We are architects, not programmers. This is NOT something we should be teaching our staff”. Feeling highly discouraged I went back to my desk and mulled over our conversation. I kept writing more and more dynamo scripts presenting them to him, telling him we could package them with Dynamo player, and the staff would simply need to hit “run”, and not necessarily have to deal with the spaghetti mess under the hood. I had mitigated the friction, and he was eventually able to see the value.
The secret sauce or parallel to getting your idea into the market or to get a new technology implemented at your firm is to Reduce Friction and disrupting the value chain that currently exist. If you can disrupt any current habit or pattern and make it have less friction, consumers will use it. Let’s take for example the progressional process of how we have listened to and consumed music. Pre-2005, you used to go to a brick and mortar music store, purchase your music, drive it home, put it in the cd play, and push play. With the advent of the internet and online shopping, “CD Baby” was born, and you could now order music from the comfort of your home. Fast Forward to around 2015 “Voice” is born, and the idea of a speaker connected to the internet could play any song you desire.
I think most of us, myself included sometimes thing that the AEC industry is somewhat insulated from these technologies, and that they have no real application in our day to day world. I would like to challenge us in this and reference my friend Ryan Cameron’s talk at Data Day when he used an Alexa to filter through design options in project fractal (now project refinery).
Ryan Cameron using an Alexa to sort through design options in Autodesk Project Fractal
“The world’s most valuable resource is no longer oil, but data” – The Economist
Data has just passed oil in the context of value and worth, but what are we doing with the data?
I want you to imagine for a moment that you are walking down a trail in the woods. It’s a beautiful sunny day, and about 72 degrees. Your dog goes off the beaten trail a bit. You call him, but he doesn’t come. You go over to get him, and you notice a few feet from him a shimmering reflection on the ground. It’s gold! Now you could either start harvesting the gold, bringing it to Jarod’s and cash in big, or you could go back to the path, and leave the gold there. This is the equivalent of what each of us do every day on our projects. There is a pluthera of data in our projects, sitting there, waiting to be used, harvested, and turned into something more valuable than oil (or gold)!
How do you currently consume data? The problem with most firms is that they are sitting on gold mine of data but are doing nothing with it. Most firms are consuming the data on their projects in rows and columns, either in a door, receptacle, or finish schedule. Looking at data in this context can be overwhelming and hard to digest.
What if instead you could consume the data in a more consumable way? This is exactly what we’ve done with our new EvolveLAB dashboard.
EvolveLAB Data Visualization Dashboard
Utilizing D3 dynamics gives you an interactive way to leverage this valuable information
This gives you direct access to the information in your project(s) including Model Warnings, In-Place Families, and Department Adjacencies.
The problem with Passive Design
We have been designing passively for years. It’s great for getting our ideas from our head to our hand. However, what if we could leverage the computer the same way we leverage calculator for problem solving? What if we could feed the computer the minimums, maximums, and what we want to solve for?
The Messy Middle
Now, when I say the future of BIIM is NOT BIM, I’m referring to BIM as we know it. We are in this transition phase of moving from Building Information Modeling to Building Information Optimization (Data Driven Design) and Building Information Automation.
Since most firms have adopted BIM (Building Information Modeling), we are now able to manipulate building parametrically utilizing input/output parameters. In this example, we are designing a canopy structure for a Harley Davidson Dealership. The inputs are the spacing, patterning, etc. The outputs are the quantities, tube steel length, and costs. The process is mostly manual and brute force parametric modeling using visual scripting.
Another example is the Hobbs trail project we did for Hufft Design in Kansas City. We went through and set up the scrips to optimize standard tube lengths, bracket connections, and intersecting terminations. The whole process was executed and optimized using Dynamo.
Evolutionary Problem Solving
We were recently brought on to help with a master planning optimization project. The fixed parameters were the site perimeter, and the existing buildings. We had a series of podiums that the proposed buildings would sit on, and a fixed space program including office, retail, and residential units. We fed the computer the maximum building heights, and let the computer solve the master plan’s optimal building proportions.
Evolutionary problem solving using Galapagos
Another approach to designing beyond evolutionary problem solving is optioneering. This takes a more co-design process where humans and computers work together to optimize a building based on a set of parameters.
One analogy I like to use with this approach is when my family and I go to a pizza restaurant in Colorado called Mod Pizza.
You essentially can “optioneer” your pizza. You can pick 8”, 10”, or 12” pizza. Gluten Free Crust, Regular Crust. Choose which toppings you want, etc.
We can use tools like project refinery (formerly Project Fractal)fed by Dynamo to optioneer buildings.
In this airport project example we were hired to help design, we were able to choose what our curtain grid spacing was, average panel height, but more importantly the number of custom panels. This gives us the ability to help the project budget reducing the number of custom panels.
Optioneering using Project Fractal now Project Refinery
Another example of optioneering was one we did for Mortenson Construction and their solar design group. We built a custom parallel coordinate graph chart for them specific to solar design. This gave their engineers the ability to optimize their site based on the power requirements and density of panels.
Solar optimization parallel graph chart built by EvolveLAB for Mortenson Construction
Optioneering can sometimes break down though. Imagine what happens when you have opposing goals. To borrow from Zach Kron’s analogy of optimizing a donut.
Let’s say you want to maximize a circular donuts surface area with sugar, but you also want to maximize the void penetrating the donut. These are opposing goals.
The same can be said of a building. Perhaps you are trying to maximize daylighting while at the same time trying to minimize heating and cooling loads.
Ultimately you end up having to assign “weighted” data.
This is the idea of assigning a higher level of importance to data so that way certain designs have a higher importance.
Computation for MEP
Architects aren’t the only ones benefitting from computational design though. One area that is just starting to bud in the computational design space is MEP. Check out the automation Sean Fruin has done leveraging dynamo player. He is able to Data mine the HVAC information, run air balance calculations, place diffusers, equipment, and calculate pressure drops.
Further Brett Young never fails to impress with his new company M2.XAI. The below software algorithm highlighted from shows Reinforced Deep Learning applied to a multi-trade MEP corridor. This training sequence resulted in an optimized AI solver after 12 hours of computing and 1.2 million design / coordination iterations.
M2.XAI Computational Design for MEP
Now, three years ago I predicted many of these dynamo/grasshopper scripts would start to be replaced by “niche” programs that had a nice UI. Take for example this Revit test fitting tool we built for Architects to be able to generatively design multi-family and hotel projects.
Revit test fitting tool built by EvolveLAB
TestFit.io takes it a step further. They launched in October of 2017. Testfit.io provides a radically new approach to solving site feasibility studies. Harnessing the concept of co-creation, TestFit leverages the knowledge of the user with the efficiency of algorithms. Using a parallel coordinate graph you can optimize your Floor Area Ratio, Net Rentable Square Feet, and Unit Counts.
Building Information Interoperability
Another trend in our industry is interoperability between programs and software agnosticism. Now Autodesk has definitely made a try at this with the interoperability with FormIt and Revit. The challenge is the objects are converted from a Mass in FormIt to a Mass in Revit.
Another popular tool for design in SketchUp. We introduced our own interoperability tool “Helix” which converts a SketchUp model into real Revit walls, doors, windows, roofs, and curtainwall.
SketchUp to Revit Interoperability using EvolveLAB’s Helix tool
However, there are many other opportunities for interoperability with other programs as well. Take for example this confidential project we worked on. We worked to streamline all the data touchpoints.
We needed to track the entire process from concept to installation.
We used Rhino.Inside a WIP program that leverages Rhino/Grasshopper inside of other programs like Revit. We were able to create an interoperability routine that would translate the file from Rhino surfaces to REAL native revit families.
We then developed a dynamo script that would generate a series of QR Codes that would be assigned to the elements of the project for fabrication and installation.
These QR Codes would have unique information applied to the panels and their respective parts that would help track every unique panel and its location on the site.
At the end of the day though, we have to find a way to build it. That’s the whole reason we do what we do. So that way we can make our designs a reality.
Applying the manufacturing process to buildings…
For a very long time, I’ve been asking the question of why we can’t apply mass production processes (similar to the automotive industry) to the building industry. The feedback I always receive is “Well Bill, unlike a Ford Taurus that is exactly the same, each building is unique, so it’s hard to apply scalable mass production techniques to a unique structure”.
Well, what if we could componentize and modularize buildings, and apply those mass-production techniques to the componentized parts? That is exactly what Katerra is doing. The find repeatable projects, and portions of projects, and execute the project as a modular kit of parts.
Katerra modular construction systems
The other problem with traditional construction processes is that we have to contend with weather which can contribute to delays due to rain, snow, etc.
Blue Print Robotics in Atlanta, Georgia is overcoming this challenge by using a similar approach as Katerra by helping teams build projects offsite in their controlled environments helping construction progress all year round.
After the assembles have been assembled, they are put on a flatbed for shipment and installation.
What happens when you start using these construction methodologies with computational design? That is exactly what we did for a client that had a modular structural framing system. We were able to componentize the structure and apply a parametric relationship of those parts to the building.
Computation for Manufactures
Architects, Developers, and Contractors are not the only ones benefiting from computational design. Take for example the assignment of panels to any building geometry.
Everyone is complaining about skilled labor shortages. This is a serious problem plaguing our industry.
The U.S. Bureau of Labor Statistics - "there were about 300,000 vacancies in the construction industry as of June. It's expected to need 747,000 more employees by 2026."
There’s no growth, and there’s a huge demand for potential workers. It’s the perfect storm for a robotics disruption.
3 years ago I highlighted two robots that were hitting the construction site. One of them being SAM, the (Semi-Automatic Mason) Sam cost half a million dollars, and can lay between 800 to 1,200 bricks in a single day – a skilled human mason tends to top out at 500. That’s almost double the number of bricks.
Another bricklayer I highlighted was from Hadrian robots who claim that their robot can lay 1000 bricks per hour!
Now, not only is SAM, and Hadrian robots more efficient, they are never going to get tired, never going to get injured on the site, and are never going to file for unemployment.
Since then, there have been some new robots hitting the construction site from Built Robotics. This includes an Autonomous Dozer, Compact Track Load, and Excavator
All of these machines are connected and ran from an ipad. They come equipped with sensors to help with clash avoidance with humans, obstacles, and each other.
Will robots take over our jobs though?
“In the future, I do think we’ll see a lot of autonomous equipment out on job sites,” said Noah Ready-Campbell, founder and CEO of Built Robotics, “but I also would emphasize that we’ll need plenty of workers, too. People are just much better at improvising and making judgment calls, and that’s critical, especially in the early stages of a product.”
How do you best prepare yourself and your firm for these changing processes?
Alright, so we’ve talked about Computational Design, Evolutionary Problem Solving, Optioneering, Interoperability, and Robotics in Construction. What does all this mean to you? That’s a question you have to answer for yourself. That said, if you are like me and you see all these trends coming, and want to best prepare yourself, how would you do so?
If you are an employer
Invest in people. Find the best talent in the industry, and hire them. Hire people smarter than you.
Invest in Research and Development.
If you are an employee
Invest in yourself. There are plenty of resources on website blogs, LinkedIn Learning, Lynda.com, etc.
Learn to Code, even if it’s only visual programming.
Digital Fabrication and Robotics
Become the expert and bring the information back to your firm
Don’t wait for someone to ask you
I’d like to end us with a reminder by way of a quote from Shajay Bhooshan
Architecture (Bill-I would add Engineering, Construction, Manufacturing, etc.) does not exist in a bubble. Robotics, 3D printing, AI, big data, and so on will have an impact on the industry. Design cannot be simply a matter of intuition when you’re solving complex problems.” - Shajay Bhooshan’s“ Zaha Hadid Architects