nAvIsworks - putting the AI in my vision of the ultimate upgrade

[eblackburnTDSKF]

I spent 30-plus years creating content on the architectural side of AEC and have used BIM since 1996. I switched to construction and have spent the last year overseeing our VDC department as we perform BIM coordination and clash detection. I had a bit of an epiphany the other day after thinking to myself, "there has to be a better way!" Allow me to explain.

You see, the Revit models that architects and engineers produce for the purpose of obtaining approvals from AHJs and for generating construction documents usually only communicate the design intent. Particularly with HVAC equipment, there is an assumption made as part of the base bid spec to "model" an RTU (or other) from a preferred manufacturer. But as long as the specifications allow for other manufacturers to be used if they meet performance requirements, what often gets installed is entirely different. It's up to the trade subcontractors to model the fabrication drawings and to use that model to be clashed against the other trade disciplines 9with each of them modeling in their own siloes). Only when all of the trades have their individual models completed can the GC check for clashes and coordinate resolution of those. Because it takes significant time and skill to model, the trade subs have an illogical attachment to their first attempt. This has the unintended consequence of increasing resistance to modifications required to achieve coordination when a simple reconfiguration might fix the problem without much fuss or effort. I have witnessed many meetings where the ductwork was originally modeled in a square configuration that created conflicts when a rectangular profile would work much better. Navisworks, from the beginning, has used a "find, fix and repeat" workflow. And bridging the gap between design intent and fabrication often results in sub-optimal routing and wasted resources/extra costs. There are Revit plugins that optimize ductwork routing, or plumbing routing, but these are isolated to just one discipline at a time.

So what is my grand vision? What if Navisworks could add the power of AI and generative design to actually solve and optimize all of the individual models at the same time? It would recognize the available space, prioritize by sequencing need, and adjust all of the individual elements to arrive at the best solution. At one end would be an AHU with parameters for performance per space served, and at the other end would be the terminal locations in the ceilings, walls or floors. In between is the route of the ductwork and all of the obstacles it has to avoid such as structure, rated walls or floors (so it can add fire dampers). It would have the power to explore multiple options to arrive at the best solution using the fewest materials. And what if there were better universal and modular standards for equipment so that what was included in the base model could also be used for fabrication, regardless of manufacturer?

As model coordination tools migrate to the cloud (with fewer abilities) there is still a strong need for powerful clash detection such as Navisworks. But if Autodesk could add this type of AI power to it, it would be a very compelling product and be much more efficient tool.

[eblackburnTDSKF]

I spent 30-plus years creating content on the architectural side of AEC and have used BIM since 1996. I switched to construction and have spent the last year overseeing our VDC department as we perform BIM coordination and clash detection. I had a bit of an epiphany the other day after thinking to myself, "there has to be a better way!" Allow me to explain.

You see, the Revit models that architects and engineers produce for the purpose of obtaining approvals from AHJs and for generating construction documents usually only communicate the design intent. Particularly with HVAC equipment, there is an assumption made as part of the base bid spec to "model" an RTU (or other) from a preferred manufacturer. But as long as the specifications allow for other manufacturers to be used if they meet performance requirements, what often gets installed is entirely different. It's up to the trade subcontractors to model the fabrication drawings and to use that model to be clashed against the other trade disciplines 9with each of them modeling in their own siloes). Only when all of the trades have their individual models completed can the GC check for clashes and coordinate resolution of those. Because it takes significant time and skill to model, the trade subs have an illogical attachment to their first attempt. This has the unintended consequence of increasing resistance to modifications required to achieve coordination when a simple reconfiguration might fix the problem without much fuss or effort. I have witnessed many meetings where the ductwork was originally modeled in a square configuration that created conflicts when a rectangular profile would work much better. Navisworks, from the beginning, has used a "find, fix and repeat" workflow. And bridging the gap between design intent and fabrication often results in sub-optimal routing and wasted resources/extra costs. There are Revit plugins that optimize ductwork routing, or plumbing routing, but these are isolated to just one discipline at a time.

So what is my grand vision? What if Navisworks could add the power of AI and generative design to actually solve and optimize all of the individual models at the same time? It would recognize the available space, prioritize by sequencing need, and adjust all of the individual elements to arrive at the best solution. At one end would be an AHU with parameters for performance per space served, and at the other end would be the terminal locations in the ceilings, walls or floors. In between is the route of the ductwork and all of the obstacles it has to avoid such as structure, rated walls or floors (so it can add fire dampers). It would have the power to explore multiple options to arrive at the best solution using the fewest materials. And what if there were better universal and modular standards for equipment so that what was included in the base model could also be used for fabrication, regardless of manufacturer?

As model coordination tools migrate to the cloud (with fewer abilities) there is still a strong need for powerful clash detection such as Navisworks. But if Autodesk could add this type of AI power to it, it would be a very compelling product and be much more efficient tool.

[Logos_Atum]

Taking a look at old European houses, I´d rather skip the idea of AI and make people more aware of what´s really been built to last and conserve energy. It can be trained to do fractal like, supporting superstructures within once solid, now 3D printed parts for all I care, but anything other than that will (having a look at what became of architecture in many a field) really not help with what´s wrong. The necessary time a house has to last is hardly worth a discussion I think. Remember? Houses built to last a winter with the wood it can store, no mold, hardly flammable, cool in summers... everything that requires a sense for the environment as well, not a block copied a thousand times, non the matter how the climate for example is where you place it?

I bet we can go on for a while without seeing sense in even more artificial intelligence.

We´ve long sice started patching rubbish with active ventilation, cooling, additional heating. That was punishable. Building houses that need you to burn wood. Many houses look like are built to cost you tons of energy until they rot, and that´s probably after 50 years already. But sustainable design sadly is more of a marketing thing rather than the strength to take everything into account you have to take into account considering your senses. It´s more like a great plan (business strategy) to sell maintenance products all along the way for your "bleeding heart". Unable to let go and patch and patch it over and over again.

Ai isn´t a tool yet. And thats all it must become.

[Logos_Atum]

Taking a look at old European houses, I´d rather skip the idea of AI and make people more aware of what´s really been built to last and conserve energy. It can be trained to do fractal like, supporting superstructures within once solid, now 3D printed parts for all I care, but anything other than that will (having a look at what became of architecture in many a field) really not help with what´s wrong. The necessary time a house has to last is hardly worth a discussion I think. Remember? Houses built to last a winter with the wood it can store, no mold, hardly flammable, cool in summers... everything that requires a sense for the environment as well, not a block copied a thousand times, non the matter how the climate for example is where you place it?

I bet we can go on for a while without seeing sense in even more artificial intelligence.

We´ve long sice started patching rubbish with active ventilation, cooling, additional heating. That was punishable. Building houses that need you to burn wood. Many houses look like are built to cost you tons of energy until they rot, and that´s probably after 50 years already. But sustainable design sadly is more of a marketing thing rather than the strength to take everything into account you have to take into account considering your senses. It´s more like a great plan (business strategy) to sell maintenance products all along the way for your "bleeding heart". Unable to let go and patch and patch it over and over again.

Ai isn´t a tool yet. And thats all it must become.