How vulnerable are nuclear reactors to hacking? Third grade review: Atomic decay generates heat! More than one exact way is possible - Natural uranium - Enriched uranium (More Uranium-235, less Uranium-238 which is kind of stable) - U-235 decays to Thorium 231, and so on, ends up as lead eventually - U-236 usually decays immediately, but sometimes doesn't - Plutonium-239 - Other stuff that decays (uncommon) This can make electricity - Radioisotope generators Example: The Demon Core 14 lb, like a bowling ball Made from Hanford-refined plutonium Nickel coated to prevent corrosion Not just a solid sphere! Designed to implode - Just compress to activate! - Squashed inside U-238 sphere in a fission bomb 5% below critical mass Harry Daghlian Neutron reflection expriments - Neutrons and fission How close can we get to critical? Dropped brick Louis Slotin and 7 others Curved neutron reflectors Again, how close can we get? Technique demonstration, like CS lecture Thumb hole at the top? Measure activity when reflector is close "Tickling the Dragon's Tail" Fermi: Continue and die within a year Screwdriver slipped Most people in the room didn't die from the accident Cherenkov light observed: - https://en.wikipedia.org/wiki/Cherenkov_radiation Fate of the Demon Core Too radioactive for Operation Crossroads tests, except maybe last Last test cancelled Melted down and recycled Controlling the Reaction The neutron concentration must be controlled Therefore, we have control rods! Neutron Poison: Xenon-135 and Xenon-precluded startup - Xenon-135 comes from Iodine-135, which causes buildup of Xenon-135 after reactor shutdown - "Burns" off, problem when reducing power Samarium-149 - Builds to equalibrium in about three weeks Gadolinium-157 and more - Can dump neutron poison in during SCRAM - Tend to build up over time Burnable neutron poison included in naval reactors - May be in fuel rods to minimize difference from new rods - Or just use control rods to handle it Example: RBMK layout Graphite tips on control rods - Can maximize effect, because water would absorb neutrons Fuel Management: Generally in rods of some sort RBMK fuel assembly diagram - Can be replaced while reactor is running, remotely Pebble bed, etc What is SCRAM? "Safety Control Rod Axe Man" "Safety Control Rod Activation Mechanism" Maybe both just backronyms Anyway, shut down the reactor as quickly as possible It would also shut down if you spread the fuel out on the floor in a thin layer. Or distributed it over a large area! Getting power from this: Steam turbines, reciprocating steam engines, generators and marine power RBMK diagram (gigawatt-class reactor) Boiler Safety - https://en.wikipedia.org/wiki/Grover_Shoe_Factory_disaster - If lots of time: https://www.youtube.com/watch?v=dgrhbdeXW_E - Google "Crush, Texas" sometime Computer management: The computer can manage control rod insertion to maintain a constant power - Despite Xenon poisoning, etc. Downfall of a Reactor Example: Chernobyl Chernobyl plant: 4 RBMK-1000 reactors, 2 more under construction Steam turbine experiment and maintanance shutdown of reactor 4 Decay heat after shutdown and need for continued power Coolant flow of 7,400 gallons per hour on each of 1,600 channels! - About 3,200 gallons per second total 5,500,000 watt pump handles this - 5.5 MW, but remember, the reactor makes 1,000 MW - Typical house service is maybe 25 kw Backup generator power-on time - Took about a minute at Chernobyl - This was part of the problem at Fukushima - But the Fukushima reactor battle was won! Solution: Steam turbine spin-down energy - Debugging: The generator doesn't work at low speed - This would be the fourth try The plan: At least 700 MW output with turbine at full speed, then perform automatic shutdown - No problems were anticipated - Chernobyl management wanted the new system debugged and installed - Power was available for cooling, so the new system didn't need to work - Day shift workers were instructed in advance and ready Scheduling problem: A different station went offline - Chernobyl was requested to maintain power, and agreed - Preparations for the test were made anyway, including disabling ECCS - 23:04: Clear to shut down - Change of plans: Night shift will perform the experiment Timeline of Experiment: - Power decreased to 700 MW by 00:05 - This causes Xenon-135 poisoning from decaying Iodine-135 - Power kept dropping with no intervention, to 500 MW - Control rods were inserted too far on accident + Unknown cause, Toptunov and Akimov both died - Control rod insertion brought power down to 30 MW, too low for test - Control room disabled automatic control rod control, pulled control rods to upper limits - Increased to 200 MW, extreme Xenon poisoning - Unstable operation, triggered many alarms, preparations continued - Backup pumps were activated, increasing water flow and inlet coolant temperature causing steam bubbles - Very unstable configuration at this point (01:19) - Nearly all control rods were then removed manually, including fail-safe rods which are never removed - Automated system that could SCRAM was disabled by manual removal - 01:23: Time for the experiment! - A minute later: Manual SCRAM, maybe - Control rods take 18-20 seconds to insert, tip does NOT absorb neutrons, insertion causes brief spike - Power spike fractured fuel rods, and blocked some control rods 1/3 inserted - Last reading on control panel was 33,000 MW - steam explosion threw 2000-ton upper plate through building roof - Various reactor guts were thrown out the top of the building - Visible blue light from building - Second explosion: Nobody really knows what caused it, about 10 tons of TNT equivalent - Reactor 3 building on fire at this point, reactor 3 shut down at 05:00 (successfully) Cleanup, liquidators, Pripyat, etc 11 RBMK reactors still operating today, but Lithuania shut down Ignalina due to EU Causes, in summary: Disabling of automatic systems Lack of understanding of Xenon poisoning Inexperienced staff performing experiments (Made it worse) Lack of adaquate containment structure, because this would have doubled the cost Answering the question: How do we blow up a reactor by computer? Start by changing the work schedule to remove anyone competent Deliver fradulant message explaining weird experiment that must be performed Directly causing it with the computer: There are systems to prevent this, and they probably work! Direct sabotage: Control rod insertion mechanism - Keep in mind it probably fails with rods fully inserted - Like semi-trailer brakes, not like car brakes