Yongbyon ELWR nuclear reactor - Two renders for 38 North

[duanekemp]

A non-proliferation reconstruction project of the Yongbyon Nuclear Scientific Research Center: Experimental Light Water Reactor. Two exclusive renders for 38 North's website presentation: https://www.38north.org/resources/2021/04/photo-galleries/yongbyon-experimental-light-water-reactor-render/

The only photos available of this structure are taken from 50k to 70k feet elevation. Though the Airbus drones are capable of the 70k ft imagery, I suspect that the 50-60k elevations are more likely. Since the location is well within the North Korean territory, all images are taken from angles that permit the photos but, from outside of the DPRK border.

As well, it appears as though the imagery is treated so that the assets align with actual ground map geographical/topographical points and land marks, such as roads and rivers. This feature, though super for point to point distance calculations and precision, is useful, but vertical heights are skewed and buildings are stretched. Their footprint on the map is good... but, features are distorted and makes vertical estimations challenging. This was a difficulty as the building, after a certain certitude of precision in reference to the photos, needed to be rescaled in the vertical. For this, another known structure (with older horizontal view available) needed to be reconstructed, then compared to the areal imagery, scaled upward to fit the airbus shots for one "confirmable" reference for downscaling the reactor. The second references to confirm height were the shadow lengths from the timestamp and dates of the images. Between several confirmations between these two methods, a reliable vertical height was established.

Keeping this in mind, the project is reconstructed from these blurry, Airbus satellite imagery resources which took myriads of perspectives, sun conditions and seasons to determine actual shapes of smaller assets and to establish actual height from shadows on the ground. Fun work with considerable input and resources from consultants Olli Heinonen and Frank Pabian who assisted in this reconstruction being as detailed as it turned out.

These non-proliferation experts/specialists became enthusiastic during the progressive build and dug deeper for more visual resources to help make it as accurately detailed as possible (with what is known or available). I estimated originally a 1 to 1.5 months for the reconstruction. However, with the collaboration of expert consultants, it took four months until nothing was left to alter, edit or add to without further (much more detailed and closer) reference images. We knew we nailed it when both nuclear expert specialists were excited by the final version.

At this point, without the DPRK releasing photos from the ground or much closer, the current build is the most accurate reconstruction/presentation of this nuclear facility.

From this model, two renders were made exclusive to the think tank "38 North", a project under the East Asia Program at the Henry L. Stimson Center.

The project is huge and at some date, I should be able to share other views and perspectives of the larger model. But, as these have been published, some may find the project interesting at this stage.

Also included are four Airbus images (excerpts from larger areal photos) as examples of reference photos available, from which the model was reconstructed.

[L i am]

Interesting, it seems you are conveying macro information down to someting as detailed as a wiring loom. Must be keeping very busy Nice work and truly well done

[duanekemp]

Thank you, Liam.

The hardest to discern, for detail, was the Transmission yard on the east side and the ambiguously hidden Equipment hatch. Very few images show that north side face with enough detail since most flights are from South Korea or from the east & west. So, that North face gets ignored due to lack of flights of photos from within the DPRK. Was REALLY happy to get those two parts done!

[duanekemp]

To illustrate the vertical height issues:

Model built accurately to the reference photos:

Model after scaling with shadow and secondary building reconstruction as new "independent" reference:

5 MWE reactor building (same research park) reconstruction used for scale comparison:
Photos exist that show direct horizontal profile so was the perfect reference building to use.

By building this other, older reactor that is nearby, there was a physical comparison between known elevations and percentage of stretch needed to match the distortion of the areal photos.

The "skew" is quite significant but, was resolved in this way.

[L i am]

Is the project close to you? I find when I do larger projects I think I have got everything.......but I need to visit the site so many times What app did you use for the arials? If you used Nearmaps it would be a highly defined, up to date asset that has hight information all in one go

[duanekemp]

Hi Liam,

The ELWR reactor is in a part of the most photographed, top secret protected area of the world, yet no source we know of has been able to pinpoint height data from the buildings, though their calculated footprints can be reproduced.

I live in Switzerland but, I don't know anybody who has visited North Korea. Not only that, but as this restricted, sensitive and military protected region is geared towards North Korea's nuclear weapons program, only sanctioned, authorised diplomatic nuclear inspectors have been able to visit (one of the consultants I collaborated with, for example) and photography is absolutely forbidden.

It's like knowing where Groom Lake is but, going on the property of a secret military base/site is illegal and out of the question. This is also why any visit is impossible and as why there are no available on site photos published.

So, here's the coordinates for precision: https://www.google.com/maps/place/Yongbyon+Nuclear+Science+and+Weapons+Research+Center/@39.7959459,125.7543004,448m/data=!3m2!1e3!4b1

[duanekemp]

Thank you Bryan.

[Bryan K]

Very cool and interesting!