The videotape presents an overview of the TMI-2 accident and the resulting clean-up project, through 1989. It includes a review of the reactor core structure, operation Quick Look, reactor vessel head lift and plenum removal, the damaged core rubble bed, core boring, lower head inspection, defueling using various tools, shipping damaged core components to Idaho National Engineering Laboratory, and ex-vessel defueling. Produced by GPU Nuclear, 1989. (Copyright 2001 by First Energy Corporation as successor to the copyrights of GPU, Inc.).
(24 minutes in length)
TMI-2 Clean-up Highlights Video Transcript
The 1979 accident at Three Mile Island Unit 2 severely damaged the reactor core. During the accident the pilot upper relief valve or pur located at the top of the pressurizer stuck open. This allowed water in the form of steam to flow from the cooling system to a drain tank and eventually to the reactor building basement floor.
This loss of cooling water uncovered the core and allowed it to overheat. The loss of coolant was stopped and the reactor system brought back under control. But, not before the core sustained significant damage.
Overview of an undamaged reactor
At this point, an overview of an undamaged reactor is in order.
This diagram shows a cut away view of the inside of the reactor. The reactor vessel is a large steel tank. It contains the fuel core and related components. At the top of the reactor there is a 170 ton cover bolted to the vessel.
This reactor vessel head contains the mechanisms that move the control rods in and out of the core to control the fission process. It can be unbolted and removed to open up the reactor. Inside the reactor vessel on top of the core is a structure called the Upper plenum assembly. Now, this serves as a guide for the control rods as well as directing the flow of coolant. This can also be removed to clear the way for defueling. At the center of the reactor is the core region containing the nuclear fuel. There are 177 assemblies of fuel rods arranged in a grid pattern. Over 200 fuel rods make up each assembly. These are held together by a stainless steel end fitting at each end and by spacer grids distributed along the length of the assembly. At the top of the fuel assembly is a spider which holds the control rods as they are move in and out of the core.
The core is supported by a cylindrical core support assembly, also known as a CSA. Below and around the CSA is a space where water enters the reactor and flows upward through the core to cool it.
Post-Accident Conditions
At bottom of CSA is a series of flow distributors and support plates. These plates evenly channel water through the core. Early information on post-accident conditions was obtained by visual inspections performed with miniature underwater television cameras. The first opportunity to view the core region damage was in 1982 when one of these miniature cameras was inserted through a control rod mechanisms sent down through the plenum and into the upper core region. The operation was known as “Quick Look.” The pile of fragments was resting on the remains of the core. The top of the rubble bed was five feet below the top of the original core region.
Much of the debris was shapeless and non-recognizable. Scattered about in the debris were some recognizable components, springs, pieces of control rod assemblies, tops of fuel elements, and rods of various lengths. The void region was five feet deep and extended in many places all the way out to the edge of the core. Before further investigation and defueling could begin, the head and the plenum had to be removed from the reactor vessel. The head was removed in July of 1984 and placed on a shielded storage stand in the reactor building. The plenum was removed from the reactor vessel in May 1985. It was transferred to a storage stand in the flooded deep end of the refueling canal.
The damaged area of the underside of the plenum was inspected during the transfer operation.
This was the first time the whole grid plate and the associated damage was clearly visible. The foamed appearance of the stainless steel is caused by the interaction of the steel with steam at high temperature.
Defueling the Reactor
To defuel the reactor, workers stand on a six inch thick stainless steel, rotatable work platform mounted on top of the reactor vessel. Through a slot in the platform, they use tools with 30 foot long handles to accomplish defueling operations. In October of 1985, workers began to remove the tops of some of the standing fuel bundles and other debris to allow installation and operation of the canister positioning system.
This is a carousel type device suspended under the work platform that holds five defueling canisters.
The top of the collapsed fuel core looked like this. The fuel rods and large pieces were picked up individually and placed in canisters. This was the “pick and place” defueling phase. As defueling progressed, the rubble and loose debris were removed using a tool called the spade bucket. The spade is the straight part of this tool. It dug into the rubble and the bucket was then closed on the rubble to lift it into the canister. A hard layer discovered earlier by probing with a pointed rod was reached. Defuelers found that their hand operated tools could not break through the hard solid mass.
Core Bore
In addition, a wall of solid material was discovered around the periphery of the core.
In July of 1986 after most of the loose debris had been removed, the core’s stratification sampling project, better known as “core bore” was begun.
A specially modified drilling rig provided by the Department of Energy was mounted on top of the reactor vessel and hollow core drill bits were used to bore into the damaged core. From each of ten locations, a drill string containing a long cross sectional sample of the core was extracted and placed into a defueling canister. This was later sent to the Department of Energy facilities in Idaho for further study. Videotape surveys were then performed in each 3 ½ inch hole. The middle of the core region was found to be a solid mass of once molten material that extended as deep as five feet in the center of the core. This mass was found to be quite hard and brittle, apparently ceramic in nature.
Visible in this mass, particularly near the top were metallic streaks and clumps. These were apparently partially melted end fittings and other structural components. No significant voids were seen in the 10 locations examined. At the bottom and sides of this mass agglomerated material consisting of fuel rods and pellets surrounded by once molten material was observed
This is a transition zone between the once molten mass and the intact fuel rods and fuel rod stubs
Under all of this are stubs of fuel rods. They are shiny indicating they were under water throughout the accident and not exposed to the oxidizing steam environment. Videotape inspections of the lower CSA show that in most places there was no damage and little debris between these plates.
East side of the reactor, however, a large amount of once molten materials is observed.
The material appears to have flowed down from the flow holes in these plates. This material was visible from 4 different core bore inspection locations, all on the east side of the reactor.
The actual route or routes were by the core material may have reached the core head was not found at this time. The lower head of the reactor vessel was videotaped via two access paths. In 1985, cameras were lowered through the annular space between the core support assembly and the reactor vessel wall.
On the way down the visible external surfaces of the core support assembly were inspected to see if any structural damage occurred.
Not only was there no damager of any kind visible on the core support assembly, there was little or no debris accumulation in this region.
In the lower head region, however, piles of rubble like debris were found. It was estimated that there are 10 to 20 tons of rubble in the lower head.
Also visible was debris hanging down through some of the flow holes in the flow distributor head. The material appears to been molten and possibly fractured on cooling. On the north side of the lower head, was found a wall of debris standing 15 inches high and over five feet wide. The second inspection path put a camera right in the center of the lower head. During the core bore operation, a camera was lowered all the way through the lower CSA and the flow distributor head into the lower head region.
The debris in the center of the lower head was granular and was piled up almost to the floor distributor head.
A summary of conditions in late 1986 showed the bed of rubble removed, exposing a mixture of once molten resolidified debris in the lower center of the core.
This was five feet thick in the center and one foot thick at the edges. Below and around this mass were the remains of fuel assemblies still in their original forms and locations varying in height from one foot to over five feet.
In addition, a large amount of debris made its way to the lower head region. In October 1986 a large diameter flat faced drill was used in the Department of Energy drilling rig to break up the solid mass in the center of the core. Refuelers resumed loading the resulting debris.
The drilling operation left a number of large number of rocks, 200-900 pounds each, which would not fit into the defueling canisters. There they operated an impact chisel was used to break these rocks into small pieces that would fit into the canisters.
Most of the rubble and rocks created by the core drilling operations were cleared away. On May 5th 1987 a video inspection of the core cavity was conducted. This view is looking down at the top of the debris bed in the core cavity. A few loose fuel rods and gravel-like debris is visible.
But the cavity has been picked clean of large pieces of debris. At the top of the picture is the peripheral solid mass of material known as the donut. Here we see more of the donut. The donut is the ring of material outside the radius that the drill could reach. The whole central region of the core was a solid rocklike formation before the core drilling.
A little higher the donut ends and there are geometric al intact fuel rods. The remains of the horizontal inconels spacer grids are visible across the bundles and some debris is visible between the rods. The wall behind the fuel rods is the core baffle plate which encloses the fuel core cavity.
Removing Stub Assemblies
An air lift tool was used to remove to remove the loose debris that was left on the top of the stub fuel assemblies. With clear access provided, the defuelers began to remove the stub assemblies.
A clamp tool was used to grab the top of the first two assemblies and lift them out.
With a path to the bottom of the assemblies now available a more efficient tool, the fuel assembly puller, was used to lift each assembly from under the bottom end fitting.
By mid-July 42 assemblies had been removed. As procedures and tools were improved, and operator skills increased the pace accelerated. 94 assemblies were transferred to canisters by the end of August 1987. {One of} the tools improved was the fuel assembly repuller, seen here about to engage a fuel assembly. The spike that was added plunges into the top of the assembly preventing it from falling off the puller
The assembly is then lifted from the grid.
We see the use of this improved tool with one of the new tools in this defueling sequence. A stub assembly is held vertically by the improved fuel assembly core.
The second assembly came loose with the first. the only time this happened. The new fuel assembly handling tool grabs the second assembly and breaks it loose. keeping it vertical. This assembly is transported to the carousel where the funnel guides it into the canister. Entering November, of the original 177 assemblies, Only 12 remained in the south east sector of the core. These were the most difficult to remove because of molten material that solidified in this area.
At the end of the year only portion of one assembly remained. No longer recognizable as a fuel rod assembly, but as a 200 lb. mass, it has resisted all efforts at removal with the current defueling tools. When removed, the original core region will have been defueled.
Cleaning up debris
As a result of defueling operations, a large amount of debris relocated to the lower CSA and lower head as illustrated in the model.
Late 1987 video inspections show the hole in the baffle plates that resulted from the flow of molten material. It measures five feet vertically and almost 2.5 feet horizontally.
An estimated 10,000 lbs. of debris is located behind the baffle plates. Inside the reactor, the remaining fuel is generally located in areas that are hard to reach. Much of it is in the bottom head of the reactor vessel. In order to get to this material, it was necessary to cut sizeable holes in the five plates which constitute the lower core support assembly. To do this, two different tools were used. The top plate is the lower grid rib section upon which the fuel elements rest.
The core drilling machine, which had been used last in 1987 to obtain cylindrical core samples of the melted fuel in the reactor, was modified to cut stainless steel. A drill bit which which worked like a hole cutting saw was used to sever the ribs.
The plate was cut into four sections this way. The pieces were inspected for adherent fuel, cleaned and removed from the reactor vessel.
Two pieces were counted with gamma detectors to verify that little or no fuel remained on them. All four pieces were placed into one of the core flood tanks for storage.
Although the core bore machine worked well for the first plate, fuel debris near the stainless steel reduced its efficiency. Consequently, it was decided to cut the remaining plates with an automated plasma arc cutting system. The plasma arc cutter was tested on one of the baffle plates which surround the core region. The test demonstrated the system was fully capable of cutting stainless steel plates under 40 feet of borated water.
The core bore machine was used to drill around the incore guide tubes and support posts to separate them from the next plate using the plasma arc system the lower grid distributor plate was then cut into pieces, cleaned, gamma scanned and stored in the core flood tanks.
The biggest challenge for the plasma arc cutter was the lower grid forging, a 13 inch thick slab of steel with numerous flow holes.
This massive plate was severed into four pieces by cutting through the ligaments between the flow holes.
The pieces were then cleaned, gamma scanned, and placed into the core flood tank with the others.
At the end of 1988 and into early 1989 disassembly of the lower internals continued using the automated plasma arc cutting system. The in-core guide support plate was cut up and removed from the reactor vessel.
And the flow distributor was also cut up. The final task for the plasma arc system was the cutting of the reactor baffle plates. While the lower internals were being disassembled, shipping of the filled canisters to Idaho continued. During 1988 and early 1989 a total of 55,000 pounds were shipped bringing the total to 77 % of the core removed and 71% shipped to Idaho since the beginning of 1988.
Sound of train
Additional defueling and debris cleanup
Another defueling effort continued in parallel this year. Less than one thousand pounds of core debris was transported out of the reactor vessel during the accident into remote locations within the reactor cooling system. Continuing ex-vessel defueling efforts concentrated on three of these areas.
In the pressurizer, a vacuum system was used to remove loose debris which had collected at the bottom. After the loose material had been vacuumed out, there was still some rock-like material remaining. This material was removed one piece at a time by the mini-rover, a miniature submarine.
Mini- rover is equipped with a color television and a small grapple tool
In the J legs, the short pipes which connect the coolant pumps to the steam generators, miniature Geiger Muller probes were deployed.
Measurements were made to determine the quantity of core material in the area which is inaccessible to the defueling tool. In the steam generator, similar measurements were made to determine the small quantity of fuel in the lower heads.
This material is also inaccessible to the defueling tools. Through additional defueling and measurement efforts the remaining ex-vessel debris has been removed or shown to be of little concern.
By removing the five plates in the lower core support assembly access was gained to the15, 000 pound mass of resolidified fuel debris located in the lower head. This mass was broken up using a 40 foot long chisel much like one would break up a large block of concrete. After the break up, the debris was removed using an air lift tool. With this tool a new record was set - 32,000 pounds of fuel debris was removed in a single month. When most of the debris was removed from the lower head, badly damaged and melted in-core nozzles and some cracks in the welded stainless steel liner of the reactor vessel were discovered.
New tools were designed and built at TMI to inspect and measure these cracks and the size of the badly damaged nozzles. A miniature underwater color television camera was used for the first time during this inspection. The height and width of the nozzles was determined.
The crack was probed with a device a little smaller than an 1/8 of an inch in diameter. This probe could be inserted into the cracks at the depth of approximately 1/16 of an inch.
The cracks were gone over with a wire brush. Cleaning them provided proof that the cladding itself was damaged. Eventually, pieces of the vessel, including these cracks, will be cut out and examined at laboratories in the US, Japan, and Europe.
864 bolts were removed from the baffle plates using a hydraulic impact wrench and a specially designed drill. Following bolt removal, the plates were moved to gain access to the 10,000 pounds of fuel debris which had been deposited behind them during the TMI 2 accident.
Some of the debris located behind the baffle plates was nicknamed the meatloaf. The meatloaf and other fuel debris was broken up using a 40 foot long handled tool to strike, break and dig at the debris. This brought the pieces down to size for vacuuming.
With the size of the fuel debris reduced an underwater vacuum system was used to load debris into fuel cans. Now this vacuum is very similar to the underwater vacuums used in swimming pools, only much longer. You see, it was required to lift the debris from under more than 40 feet of water.
With the baffle plated defueling almost complete, only about 10,000 pounds of fuel debris remains to be removed from the reactor vessel. The remaining fuel in the reactor vessel will be removed by the end of 1989.