Register to reply

Would a universal open-source blueprint/design posted online?

Share this thread:
ensabah6
#1
Dec6-08, 12:55 AM
P: 716
Would a universal open-source blueprint/design posted online be helpful to engineers?

One reason nuclear power expensive is that there is little duplication in designs, and design costs are expensive.

Would following what Linus did for Linux kernel help?

First, one standardized universal reactor design, that all new reactors being build would have to conform to exact specifications. Parts should be interchangeable, and where possible, standardized with existing electrical generator parts. (much as pc atx boxes all use same type of parts, power supplies, etc)

Second, posting the blueprint online, with specifications and engineering details, for everyone to see an develop, until it gets to the point where engineers around the world agree as being acceptable to go into production. Free computer simulation software should also be open sourced. Linux and other open source projects are like this. anyone, including anti-nuke activists can look at the blueprint online. hopefully nuclear engineers, physicists, chemists and others can critique the model, and foresee possible problems.

I don't deny there would be no easy way to prevent criminals from looking at the blueprint for security breaches.

This "research" program can be partially funded with gov't money, and it would NOT be copyrighted. It would be a set of universal specifications and standards of compliance in reactor design.

Personally I'd like to see some kind of IFR design as open source.

There's Open RISC which is analogous to this in a hardware implementation.

This should help lower barriers of entry in R&D costs. Is there any reason this couldn't work, and nuclear facilities can't be exactly duplicated everywhere? If a company agrees to build a nuke in exact specifications as outlined online, they should be in return put on the fast-track to approval and operation and funding.
Phys.Org News Partner Science news on Phys.org
'Office life' of bacteria may be their weak spot
Lunar explorers will walk at higher speeds than thought
Philips introduces BlueTouch, PulseRelief control for pain relief
Coin
#2
Dec6-08, 02:01 AM
P: 587
Hmmm... well I don't really know anything about nuclear reactors, but I do know about software, so I'll comment from that perspective. Here's the thing.

One reason that "open source" is possible is that there is zero cost to duplicate a piece of software. You can type "this piece of software is free to reproduce" in the README and it is true-- software is information, there is no manufacturing cost associated with distributing it to someone. Open RISC / OpenCores describe hardware, sort of, but they essentially work on the same principle-- they assume you have access to something like an FPGA. I do not know if Open Cores would really be possible in a world without FPGAs and such.

When you leave this world of pure information where software and the OpenCores IP lives and start talking about things that have to actually be manufactured, this isn't really true anymore. It makes less sense to talk about giving someone the IP for free because reifying that IP into a physical object, even a prototype, costs them money. If nobody will be consuming the IP unless they are willing to pay X dollars to manufacture it, then this implies nobody will be consuming the IP unless they are also willing to pay Y dollars for the IP. In this world I think you're less likely to see something like linux / open cores develop than something like a consortium or a patent pool, where there are a group of people working together or sharing resources but they are doing so under the understanding that any IP exchanged will be paid for. (These patent pools often are set up with standard licensing terms, so they have the quasi-open-source property of being "free as in speech" but not "free as in beer", in the sense that you have to pay to use the patent but the owner of the patent can't turn down your license to use the patent or practically engage in discriminatory pricing toward comers they don't like.)

So then you get to the point of something like a nuclear reactor. Not only does it cost money to build a reactor-- you can't build a nuclear reactor, no matter much money you have, without obtaining radioactive material, the sale and movement of which is tightly controlled. The idea of "many eyes making bugs shallow" or whatever makes much less sense when you cannot actually attempt to test the IP at hand unless you are a government or large institution, and anyone who legitimately has the ability to look at the blueprints and understand them is very likely to have a job offer with such a government or large institution already.

Then there's this:

I don't deny there would be no easy way to prevent criminals from looking at the blueprint for security breaches.
You seem in this post to be working under the assumption that copyright is the only thing restricting the flow of reactor blueprints. Aren't most reactor blueprints classified? I mean, is there any nation on earth where it's legal to distribute nuclear reactor blueprints even if you own the copyright? (And to be cynical, could any nation on earth make such distribution legal without the U.S. immediately commencing bombing?)

Basically, you start with what sounds to me like a good idea here:

First, one standardized universal reactor design, that all new reactors being build would have to conform to exact specifications. Parts should be interchangeable, and where possible, standardized with existing electrical generator parts. (much as pc atx boxes all use same type of parts, power supplies, etc)
But then you use the wrong model for how to proceed from that idea, I think. The model of community-driven open source development doesn't seem to practically apply to the situation at hand. But there are models within the consumer-electronics engineering world that do seem to fit possibly well-- models that apply in a situation where players are not likely to participate unless they were specifically invited and are receiving something tangible in exchange for doing so.

I again know little or nothing about nuclear engineering or physics in specific, but the idea of a standardized "base" reactor design sounds like it would address a real problem you describe (little duplication in design, every part has to be custom-made I assume, each reactor a unique snowflake) and make it easier for companies to work together, just as it does with things like atx boxes. A government or intergovernmental program would, as you suggest, I imagine be a logical way for such a standard to come into existence, and I think it would make sense from the government's perspective to make some parts of the standard free or nominal to license (or free on condition of contributing IP to the pool, or something).

One problem I do see is that reactors seem to be built few and far between, or in spurts-- checking the internet it appears nobody has built a nuclear reactor in the United States since 1990? Standards seem to make a lot less sense if they'd possibly be obsolete by the time the second implementation was built. On the other hand, if there were some kind of program where lots of nuclear reactors were being aggressively built over a short time, as I'm sure some people would like to see happen, then this sort of consortium approach might well make sense!

....again, a real nuclear engineer might not have the same perspective I do here.
vanesch
#3
Dec6-08, 02:57 AM
Emeritus
Sci Advisor
PF Gold
P: 6,236
The principles of the core nuclear design for many reactors is publicly known. By that, I mean, the disposition of the fuel, the composition of the fuel, the material. So there is no "classified secrets" in a nuclear power plant. For instance, there is a book on the design of the French 1300 MW series, which you can buy and read: http://www-instn.cea.fr/article.php3?id_article=220
I have the book, it contains a rather in-depth description of the construction of this reactor - without giving away of course the blueprint. But there is more than enough material in there to know all the points of principle. So no "classified stuff".

However, as pointed out, the exact blueprints, with all the engineering details, are like the blueprints of a 747 or something. It is licensed technology. There are commercial stakes.

Concerning software, most nuclear simulation software is free of charge, but somewhat classified. You have to have some form of clearance. Most nuclear engineers have access to them. The reason is that this kind of software can help you design a weapon to some extend (only to some extend).

I wonder what would happen if somebody started an open source project on nuclear simulation.

Coin
#4
Dec6-08, 04:39 AM
P: 587
Would a universal open-source blueprint/design posted online?

Quote Quote by vanesch View Post
The principles of the core nuclear design for many reactors is publicly known. By that, I mean, the disposition of the fuel, the composition of the fuel, the material. So there is no "classified secrets" in a nuclear power plant. For instance, there is a book on the design of the French 1300 MW series, which you can buy and read: http://www-instn.cea.fr/article.php3?id_article=220... the exact blueprints, with all the engineering details, are like the blueprints of a 747 or something. It is licensed technology. There are commercial stakes.
Huh, that is interesting. So how possible would it be for a private individual to read over this book and then derive their own blueprints from it, say with the intent of releasing their independent blueprints into the public domain or other commons in the way ensabah6 suggests? Or is creating such blueprints something that would require simulation software or prototyping ability to really do?
Astronuc
#5
Dec6-08, 07:16 AM
Admin
Astronuc's Avatar
P: 21,915
Quote Quote by ensabah6 View Post
Personally I'd like to see some kind of IFR design as open source.

. . . .

This should help lower barriers of entry in R&D costs. Is there any reason this couldn't work, and nuclear facilities can't be exactly duplicated everywhere? If a company agrees to build a nuke in exact specifications as outlined online, they should be in return put on the fast-track to approval and operation and funding.
There are safety issues associated with nuclear technology, as well as certain applications of software and microchips. There is a lot of upfront R&D (and experimental testing) such that one cannot simply sit down and design a reactor. Each large component is optimized in terms of material and geometry. The materials and components are made to strict standards, and the quality control/assurance is based on the military quality control standards (MIL-STDs)

Challenges to a nuclear designer include the technical requirements, i.e. the fuel must generate heat while maintain its geometry, and it must retain fission products (although that doesn't always happen). Maintaining geometry ensures the fuel can be properly cooled so that it doesn't overheat (which could lead to failure and release of fission products) and that the reactor can be controlled (shutdown when necessary) so as to avoid failure of the fuel.

The core design is optimized as much as possible to minimize power peaking in order to maintain acceptable margins to technical limits. But then to get the most efficient design, one would to push the fuel as close to the technical limits as possible without failing the fuel.

One can take a design and scale it. I posted a link to EBR-II, which gives a nice overview of the fuel, core and plant.
http://www.physicsforums.com/showpos...&postcount=112

But scaling is not trivial. The EBR-II was a small unit (62.5 MWthermal/20 MWe), and not very efficient. A lot of testing was done, but not a huge scale.

FFTF was larger (400 MW), but it unfortunately was not built with an electrical generation system, but instead simply dumped the heat into the atmosphere. It was strictly a test reactor. A lot of great reasearch was done there, but when it was shut down, a lot of experimental work was simply trashed, i.e. literal 'dump in the garbage'.
gmax137
#6
Dec6-08, 10:02 AM
P: 861
"I don't deny there would be no easy way to prevent criminals from looking at the blueprint for security breaches."

Many people still believe that nuclear power is closely tied to nuclear weapons. This notion has been perpetuated by the anti-nuclear crowd, since the 1960s. I have worked in nuclear power for almost 30 years and I can tell you the only information that is "secret" in the sense that the government cares who knows it is information related to the security procedures at the plants (like, how many people are in the guard shacks, or where the power for the gates comes from). The "blueprints" for the power plants are largely available to anyone who wants to ask for them at the NRC Public Document Rooms. What is not available to the public are the "proprietary" documents developed by the various manufacturers. These are the detailed manufacturing drawings, the manufacturing test procedures, and the computer codes they use to analyse the plant behaviour. These proprietary materials are not available for the same reasons you can't just ask Ford for drawings of their pistons, or ask CocaCola for their recipes. It has nothing to do with government secrets or how to build weapons.
vanesch
#7
Dec7-08, 01:35 AM
Emeritus
Sci Advisor
PF Gold
P: 6,236
Quote Quote by Coin View Post
Huh, that is interesting. So how possible would it be for a private individual to read over this book and then derive their own blueprints from it, say with the intent of releasing their independent blueprints into the public domain or other commons in the way ensabah6 suggests? Or is creating such blueprints something that would require simulation software or prototyping ability to really do?
Well, as gmax137 said, it is not because you have a rather detailed description of the principles of all important components in a specific car that you have the genuine blueprints to have the car pieces made in a workshop, or the specific commercial pieces that are used. You could of course put together an engineering team and have them work out all those details, but then that amounts to nothing else but designing a new car. There are no fundamental secrets behind nuclear power plants, but there is, as in any sophisticated engineering, a lot of corporate know-how which is not made public. Compare it to, say, micro electronics: there are no fundamental secrets on how a microprocessor works, and you can find fairly detailed descriptions of the principles behind the architecture of most of them. But that still doesn't give you the lithographic masks to produce them. You even don't have the hardware description code. But no part of it is really "secret" in the sense that you couldn't re-devellop it with enough effort.


Register to reply

Related Discussions
Any open source alternatives for CAD and simulation? General Engineering 2
Open source question Engineering, Comp Sci, & Technology Homework 0
Open Source Math software Computing & Technology 5
Open Source Solaris Computing & Technology 1