ClimaCell Announces Autodesk Partnership & Short-Term Weather Prediction Model for the Construction Industry

In an interview with Rei Goffer, ClimaCell, CSO we discussed the company’s partnership with Autodesk BIM 360 announced at Autodesk University this week.

ClimaCell is targeting the construction industry with their short-term weather prediction model. Why? “One-half the lifetime of a construction site is very weather sensitive, from the time they dig a hole for concrete, it’s a big factor on daily basis,” said Goffer. “We need to be able to plan accordingly and for each weather phenomenon. That’s why we’re addressing the construction industry.”

ClimaCell also addresses other industries. Goffer is one of three co-founders, with a background in aviation. All three co-founders come from military service where they were very affected by weather. “We knew there was a big gap in the provision of Weather data,” said Goffer.

“We are developing the world’s most accurate granular short-term prediction model,” Goffer said.  “So that basically construction site managers can’t get this from any other weather app or specialized server.”

What differentiates ClimaCell from existing weather data providers?

Goffers reports:

1. Weather data – we have many more weather observations, at much higher quality than any other weather company.
2. Better models – we developed proprietary models for short-term weather prediction, that are faster and more accurate than anything previously available.
3. Computing power – we run our models on GPUs (Graphical Processing Units) rather than on CPU (standard processors), which gives us an edge in terms of competing time and resolution.”

“In almost all cases the issue is with the data, our understanding of what’s going on right now at a granular level with the weather,” said Goffer. “If you ask how can you generate better weather information? Can you use existing technologies, weather stations, satellites, radar, can you build and deploy more of these?”

The answer is yes, but they are not scalable and are expensive. Their sensors are limited. “In the U.S., we have a lot of these networks and the sensors but can be much better.

“We invented Passive Sensing so that we can make weather observations from many different sources, but they are not necessarily dedicated weather sensors on the radar,” said Goffer. “We have many different technologies to get weather data from existing infrastructure devices, such as wireless networks.  We’re analyzing the way signals in wireless networks behave and their behavior is actually affected by the weather. The propagation is affected by atmospheric conditions. If we can see that we can generate with a partial tool and we can get a fine understanding of what’s going on in the atmosphere on the ground level where people live and work on a minute by minute bases down to the street level.”

Goffer said they have many technologies that they do not disclose to get proprietary datasets with which they feed their forecasting model and expand capabilities. The second part is the model so once you have better data you can improve the modeling. ClimaCell creates a proprietary short-term prediction model for which the professional term is “nowcasting” which is 0-6 hours of forecasting. That model is the first model to take advantage of the very granular data set that provides almost minute by minute forecasts instead of hourly forecasts or six hours.

With computing power, there is a huge amount of data and a new model. Traditional software architecture does not support very very granular weather model, so ClimaCell transitions their entire architecture to GPUs.

On a construction site, in 6 hours you get 1 hour intervals of weather news: chance of rain, in vague terms. With a prediction of 70% chance of rain in downtown Boston next to South Station, construction owners want to know if it’s going to rain on their site or not? “Traditional forecasting engines can’t provide that information because they don’t have fine data,” said Goffer.

“Our software can be used on the construction site, and you can see a map of the rain predicted by our engine for the next six hours down to the street level, with five minutes intervals instead of 1 hour intervals,” said Goffer. “It’s Very valuable in terms of real time operations that the construction manager needs.”

Integrated into the Autodesk solution is the ability to automatically have your construction site on ClimaCell software and automatically generate alerts that you define. “So you can say you care about a specific intensity of precipitation and you want to know if it will cross a threshold. Do tell me to send a text message or email to my team. You can take it to the next level and send instructions to people in different functions.”

This way, you can also preset alerts and make decisions that save a lot of time.

ClimaCell does work with other BIM modeling programs other than Autodesk.

How frequently can you set up alerts? “The alert can be set for whenever. Is the rain intensity crossing east?” said Goffer. “You can get that three hours in advance and the granularity. You can’t predict the weather that far ahead like three days ahead. We know enough about the weather because we have a very fine picture of a given time. We have a high certainty 2-3 hours down to the street level.”

“Our software integrates all available weather data sources, coming from government and from other sources. The entire weather market in the U.S. is from mostly government radar and weather stations. We’re the only company market globally that has significant amount of proprietary weather observation. In most places, like dense areas, we’re much better than anyone else, because we have government stuff integrated into our software and our proprietary data is twice as accurate as what you get from radar and ten times the resolution.”

Goffer will not say how they acquire their proprietary weather data, it is the company’s “secret sauce.” Now ClimaCell has weather data for the entire U.S. and by next year they plan to have it for most of the major countries in the world.

Tags: 3D, AEC, architecture, Autodesk, Autodesk Revit, Autodesk University, BIM, building, building information modeling, CAD, climate change, construction, design, engineering, engineers, GIS, infrastructure, laser scanning, mobile, point clouds, reality capture, visualization, weather prediction model

Categories: 2D, 3D, 3D PDF, AEC, apps, architecture, AutoCAD, Autodesk, BIM, building information modeling, civil information modeling, Cloud, collaboration, construction, engineering, field, geospatial, GIS, IES, infrastructure, lidar, mobile, project management, rendering, simulation, site planning, sustainable design, terrain, visualization

This entry was posted on Thursday, November 16th, 2017 at 8:50 am. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.