Safety and Nuclear Energy

The Two Famous Incidents and now Fukushima

Three Mile Island. 900 MWe Pressurized Water Reactor. High-pressure & high-temperature water isolated in closed reactor loop produce, via heat exchanger, steam for turbine.

In emergency shutdowns (SCRAM), the reactors built in US have an emergency core cooling system (ECCS) that pump large amounts of water into a reactor core heating too fast.

Designs of that era had (1) lots of pumps, valves, and water pipes in nuclear reactor and (2) vast arrays of indicators and knobs in control room.

Three Mile Island sequence (03/28/79)

Two weeks before, during routine maintenance 2 valves in auxiliary feedwater pumps (on steam side of heat exchanger) were manually closed and then inadvertently left closed.

Incident began with failure of primary feedwater pump to the heat exchanger. Loss of feedwater decreased heat transfer from primary system with its subsequent overheating and pressure increase.

By safety design, backup systems didn't begin immediately, and a SCRAM occurred ⇒ pressure relief valve opened to release excess pressure.

The three backup feedwater pumps, disabled by the closed valves, didn't start after built-in 15-sec delay.

The pressure relief valve stayed open long enough to lower pressure; then relief valve should have closed. The valve stuck open; not noticed for 2 1/2 hours.

By safety design, emergency core cooling system turned on as pressure dropped, but operators thought ECCS turned on by mistake & manually turned it off. The now uncovered core boiled water and steam escaped through the stuck relief valve.

Operators, seeing core cooling (due to escaping steam) mistakenly turned off water pumps to core, hoping to raise temperature. Big mistake: The core was uncovered for 13.5 hours.

Liquid released by stuck valve ruptured a holding tank seal, ultimately spilling 400,000 gallons of radioactive water onto containment vessel floor!

The final toll:
15-30% of the core was uncovered
45% of the core melted
70% of the core was damaged
20 tons of debris fell to reactor bottom
Considerable radioactivity released [1012 Bq]

Summary: human error was the main cause; design-induced mechanical failure also important.

Some important lessons about the design of control rooms were learned by power plant managers.

Chernobyl



Chernobyl in 1997.

The Chernobyl disaster in April 1986 was much greater than Three Mile Island. Again human error the cause.

Design errors important: no containment vessel and use of graphite moderator.

What started as a reasonable test of the use of turbine generators as source of emergency power for computers essential to control turned into a nightmare as repeated operator mistakes amplified the design flaws. No one will build this design again. Rehearsing the scenario is unfruitful.

The final toll is instructive:
31 people died in the accident.
Total destruction of the reactor.
Radioactivity released over northern hemisphere
      [2 x 1017 Bq].

Where are we now

From April 1986 to March 11, 2011, nuclear energy kept on going. While none were built in US, there were elsewhere -- so much so that the US's nuclear dependence for eletricity was barely noticeable compare to 18 nations before us with France getting 80%.

In the US we kept extending licenses. Without building more plants, the nuclear industry continued to supply about same percent because we got better at running the same reactors. There were small problems -- in Ohio a reactor vessel was nearly eaten through but was caught before any leakage occurred. But the NRC didn't radically revise inspection procedures.

In some countries, improved, safer designs were brought on-line. Their success encouraged USA NRC to authorize a new design. In early 2000s utility companies started filing documents to build new reactors, often on sites with reactors. Economic downturn put process on hold.

Fukushima showed how complacent we had become:


Safety and Nuclear Energy
<http://www.physics.ohio-state.edu/~wilkins.5/energy/Resources/Lectures/nucwaste.html>
[Saturday, 16-Dec-2017 23:20:36 EST]
Edited by: wilkins@mps.ohio-state.edu on Wednesday, 26-Oct-2011 09:12:41 EDT