One I've been thinking about: the building design industry's ability to couple mechanical energy systems to passive energy strategies and building geometry/mate

One I've been thinking about: the building design industry's ability to couple mechanical energy systems to passive energy strategies and building geometry/material. Mechanical engineers have figured out all sorts of clever ways to couple different parts of the HVAC system together ensure maximum efficiency. Air is recirculated, if air can't be recirculated, just the sensible heat or moisture is captured and added to supply air. There are increasingly complicated loops of supply and return cooling/heating water to reduce the energy cost of exhausting waste heat or absorb heat from outside conditions. All of which is calculated with building energy models, and can easily be parameterized and optimized for specific load scenarios. And yet, anything that is the architect's domain is never as tightly coupled. So you end up with mech engineers recommending these sort of plug-and-play systems, architects implementing fairly generic "sustainability" strategies, and the efficiency gains from coupling the two are lost. Off the top of my head, here are some reasons for this (beyond just the disciplinary division between mechanical engineers and architects): - lack of sophisticated computational geometry methods that plug into the energy simulators - the huge upfront cost of developing a mechanistic simulation models - the dependency of energy modeling on complex fluid modeling and the prohibitively huge computational cost associated with that - the inability to capture local microclimate conditions that can inform hyper-localized building design - the lack of a database of building energy models with accurate geometry/material options that would allow architects to predict energy impacts of geometry - the general ambiguity of classifying "shapes" and "geometry", it seems impossible to break down geometry into feature vectors for ML or statistical analysis. I don't know if this actually has a huge energy impact, but my instinct is that there's a lot of potential "coupling" that could occur that is left on the table because the building design industry tends to silo the active/renewable energy systems. Granted, there's plenty of activity in the industry around this, but at this stage it's all based on the idea of hooking up existing state-of-the-art simulation modelers to CAD geometry modelers, and then doing fairly simple optimization design solution space searching between the two. As the divide between the two software "silos" (energy simulation/CAD models) starts to break down, and more sophisticated methods to break down the solution space trickles down into the industry from ML/data science, I think this should lead to a scenario where we can implement more tightly coupled building designs.