New steam generators for Krsko28 February 2001
A consortium of Framatome and Siemens won the contract to replace the steam generators at Slovenia’s Krsko plant in February 1998. Two years of preparation meant the replacement, in April-June 2000, went as planned.
The single Westinghouse-designed unit at Krsko shut down on 15 April, 2000 for an outage to replace the two steam generators. The plant came back on line as scheduled 62 days later, on 15 June. After the steam generator replacement and other modernisations, the plant’s gross electrical output is now 711MWe – an increase of over 12%.
Detailed work on the outage began a year before it was due to start, when a common scheduling team from plant operator NEK and the Siemens/Framatome consortium was established to define the schedule and check for interface. It was decided to have one common outage schedule including all activities, discussed in monthly sessions. Siemens /Framatome schedulers were permanently on-site from November 1999 to mid-February 2000 to finalise and optimise the schedule, and it was finally ‘frozen’ in mid February. At that time, the schedule consisted of over 7800 activities (6000 for NEK and 1800 for Siemens/Framatome). The Siemens/Fram-atome activities were translated into 124 work orders whose structure had been previously defined and mutually agreed.
The common team updated the outage schedule twice a day, and the updated version was immediately published on the Intranet, so people could see the new situation immediately and react. In case of major deviations, relevant people with no access to the Intranet were informed immediately. For Siemens /Framatome, a so-called “four-day schedule” was issued once a day, showing the present day and the next three days. This was the basis for the daily meetings during the replacement.
RIGGING AND LIFTING
Each old steam generator weighed 320t, and each new one weighed 343t. To lift and lower the existing steam generators from the reactor operating deck (el 115m) to ground level (el 100m), and lift the replacements into place, an outside lifting tower had to be installed to hoist the loads vertically and shift them horizontally. In fact, the lifting tower comprised two towers, two heavy load girders (design load 420t), each 34m long, two cross beams, two 200t wire jacks for horizontal movement and four 200t wire jacks for vertical movements. An auxiliary 10t trolley was attached to the heavy load girders so materials, tools and containers could be lifted into and out of the reactor building.
Engineering studies found that the safety-related items close to the lifting tower – including the refuelling water storage tank and the reactor make-up water tank – might be affected if the lifting system was installed before the reactor reached cold shutdown. As a result, the tower was installed to half of its full height while the reactor was in operation and completed after the reactor reached cold shutdown and the equipment hatch was closed for core off-loading. No transports could be made into the reactor building while the tower was being installed. The lifting system was passed by the regulator, after a functional test, on 27 April.
The steam generators were moved horizontally out of and into the reactor building on a sliding system. The border between outer and inner part of this system was inside the equipment hatch. The outer part was installed as soon as the equipment hatch was closed and all transports into the reactor building had been performed. The inner part was erected as soon as the equipment hatch was re-opened after the reactor had been defuelled and the missile shield was placed back on its operating position above the pressure vessel. The shield had to be replaced because its intermediate storage area interfered with the sliding beams. The heavy runway had two supports on top: a sliding saddle attached to the steam dome, and a tilting device mounted on the channel head so the steam generators could be raised from horizontal to an upright position.
A light transport system with winch-driven trolley was installed before placing the heavy runway, so scaffolding, shielding and insulation could be moved in and out of the reactor building. A smaller system was installed through the emergency hatch (at el 100m) for tools and lead blankets.
When the missile shield was placed back in its operating position above the pressure vessel, a cover was installed above the reactor cavity to exclude foreign items and pollutants during the replacement. The cover was removed after all the work above the reactor operating deck had been completed. The load capacity of the cover was 1t/m, and was also used to store materials and equipment.
The existing polar crane had neither the lifting height nor trolley capacity for the steam generators, and an auxiliary lifting gantry was installed on top of the polar crane girders. The auxiliary gantry consisted of an auxiliary trolley with an axial bearing – to allow rotation – and a 400t wire jack for lifting. The jack was equipped with a spreader beam, which was attached to the steam generators. A complicated load test was not necessary, as the equipment and the crane were shown to be reliable from calaculations in accordance with ANSI 30b. The regulatory authority released the gantry and the polar crane on 4 May. The gantry weighed around 30t.
Steam generator lifting operations began on 5 May and finished on 14 May, when the last of the new generators was aligned.
To ensure there was enough lifting equipment inside the reactor building, and to be independent from the existing polar crane, three auxiliary cranes were installed, one on the top of each cubicle, each with a lifting capacity of one tonne. A fourth crane on the vessel head intermediate storage area had a 1.5t capacity. This allowed materials and tools to be lowered from the equipment hatch through the material opening to the lower elevations with the steam generators in place.
PRIMARY COOLANT PIPES
Two cuts were made in the reactor coolant system, one on each primary nozzle. The replacement steam generators were to be welded using a narrow gap welding process, so the fit of the pipe ends to the nozzles had to be very precise. Ten optical measurement steps were performed:
•The existing system was measured in a previous outage.
•The new steam generators were measured in ENSA’s (the fabricator’s) workshop before shipping, and the results compared with the first step to confirm the two-cut method.
•The steam generators were measured on-site to determine bevel location on nozzles.
•The steam generators were measured after beveling, comparing the results of steps 1 and 4 to define cut lines on the existing system.
•The cut lines on the existing system were confirmed.
•The pipe ends were measured after cutting and after steam generator removal.
•Further measurements were made during pipe displacement.
•The pipe ends were measured in the fit-up position prior to machining, comparing the data to define the new bevel location.
•The setting of reference ring on new bevel location underwent optical measurement.
•As-built measurements were made after welding and NDE.
The measurement system consisted of electronically combined theodolites and a PC for data processing. Between one and four theodolites were used in each measurement and verification calculations were performed simultaneously to detect questionable results. The data was evaluated outside the reactor building.
Before the coolant pipes were cut, each hot leg and crossover leg was locked in its original position by a clamping device equipped with hydraulic jacks. Each hydraulic unit was equipped with fine scale manometers to monitor displacement. As well as blocking the system, the hydraulic jacks were also used to manage pipe displacement fit each pipe-end to the relevant nozzle.
Installing the clamping device was a time- and dose-intensive job. Because NEK was performing other maintenance work on the lines, the pipes could not be filled to reduce the dose level. In addition, shielding had to be removed at the points where the device was installed on the piping.
To maintain the dose to be ALARA, mechanical cutting machines driven by hydraulic power units were used to cut and machine the pipes.
The two cuts were performed simultaneously and in two steps. In the first, a blade cutter completed 95% of the wall cut. In the second step, the blade cutters were replaced by cutting wheels. These do not produce debris, which could enter the primary system, so the primary manways of the old steam generators do not have to be opened to install up a plug to exclude foreign material. Cutting began on 4 May, which was also the start of Siemens/Framatome’s contractual replacement window of 28 days.
Both future joints on each loop were machined in parallel, after they had been decontaminated and optical measurements had been taken.
The bevelling machines were centred in the pipes and adjusted with a reference ring, aligned by optical measurement. The machines were equipped with ball socket pivots and mechanical jacks to allow fine adjustment. Weld edge geometry was given by the GTAW narrow gap welding process.
The pipe ends were decontaminated after the old steam generators were removed to lower the dose level for the following operations. The contact dose rate level inside the open pipe-ends was 60-130mSv.
Decontamination was carried out by sandblasting, which produces less waste than a chemical process. Before the process began, the pipes were sealed with a leak-tight disc so the blasting media could not enter the reactor cooling system.
In the first step, the pipe-end was blasted with electro-corundum to remove the highly-active oxide layer to a length of around 500mm. In the second, it was blasted with glass beads to improve the superficial stress conditions and to smooth the surface.
A closed-circuit system with negative pressure differential was used to avoid release of aerosols. After decontamination, a shielding plug was inserted in each pipe-end to separate the decontaminated area from the non-decontaminated one. A dose reduction factor of 40-60 was achieved.
The pipe-ends were welded to the new steam generator nozzles by a remote-control mechanised GTAW process with a special narrow gap weld geometry. A layer by layer circumferential technique was used, with constant welding parameters around the circumference. The welding stations were placed in a low-dose area at el 100m.
The hot leg and crossover leg were welded simultaneously, so with the last steam generator in position, four welding machines were operating in parallel. Welding was performed in two 10-hour shifts, with four hours to cool down the pipes.
After around 25 layers an intermediate X-ray was performed with the root already ground. After weld completion and build-up, the joints were mechanically ground to a special geometry for ultrasonic testing. No indication requiring repair has been detected on any of the four joints. Once the piping works had been performed, the shielding plugs were removed through the manways via the channel head. Foreign object search and retrieval on the primary side ensured no foreign material was left in the primary circuit. Once cleanliness was assured, the primary manways were released for closing.
The primary manway covers were moved with a swivel arm device that had been specially designed and installed. A multistud tensioner was used for the cover studs.
The primary side was closed and ready for filling on 31 May – 27 days after the first cut had been made.
As well as the main steam and feedwater lines, the replacement affected the auxiliary feedwater line, the blowdown and condensate systems, the primary drain line and all the steam generator instrumentation.
Main steam line
A section of the main steam line was removed to gain clearance for the rigging. The main steam line restraint was also removed, so that auxiliary cranes could be installed. As soon as the new steam generator was in place and the relevant auxiliary crane was dismantled, the restraint was reinstalled and the existing pipe spool was machined and re-welded followed by NDE.
Main feedwater line
The old steam generators had a preheater and the feedwater nozzle was located above the tube plate. The new versions do not have preheaters and the feedwater nozzle is in the conical section – 10m higher than before.
The feedwater line was rerouted to connect it to the new nozzle location and new supports and restraints were installed. The resulting change in the seismic response spectra required new and modified supports in the reactor and intermediate buildings.
The old preheater bypass and warmup line was removed and internals of the feedwater control valves changed to allow higher flow rates. NEK changed the feedwater pump impellers to allow three-pump operation.
Auxiliary feedwater line
A portion of the auxiliary feedwater line was cut out to allow for the steam generator rigging. Afterwards the cut-out portion was replaced by new pipe material with prefabricated pipe spools.
Blowdown system and sampling lines
The blowdown system was completely dismantled to a point near the reactor penetration. It was replaced using new pipe material connected to two new points on the steam generators that would allow a flow rate up to 5% of the nominal feedwater flow.
New seismic response spectra required supports and restraints to be modified or added in reactor and intermediate buildings.
New vent lines were installed at high points and blowdown and condensate systems modified to avoid water hammer effects on restarting the system. The blowdown system is near the channel head, and work on the system had to be closely coordinated to avoid conflict with other work and ensure no ferritic pollution entered the primary system. Sampling lines were replaced by new material and connected to the blowdown line.
Nozzle locations on the new steam generators were different, so instrumentation lines were removed to the transmitters and rerouted with new tubing. Instrumentation lines were socket-welded, so X-rays were not required.
The system is still operated using the three existing condensate pumps,so condensate suction pipe size was increased from 24” to 30” to reduce pressure drop at uprate flows.
Primary drain lines
The new steam generators have drain nozzles at each primary manway and at each water chamber so they can be drained completely. New drain lines with two root valves were installed; a quick coupling allows the connection of a metallic hose to drain the residual water into the floor drain system.
SUPPORTS AND INSULATION
To allow for the new pipe route and rigging clearance, shims on the reactor coolant system restraints, pump tie rods and steam generator supports were removed. For rigging, the upper lateral support ring was dismantled in two pieces and parked inside the cubicle. Gaps were measured in ‘hot’ and ‘cold’ conditions so the shims could be restored.
Once the new steam generators were in place, the upper lateral support rings and shims were reinstalled. During heat-up the gaps were checked, with final measurements at 292°C after a 24-hour stabilisation period. Insulation on the old steam generator was removed and replaced by metal reflective panels. Heat loss measurements on the replaced insulation showed good results, which means the outside temperature of the insulation was significantly decreased. For pipework the insulation was generally reused except where the pipe was rerouted.
Before the replacement began a multipurpose building (MPB) was erected for:
•Interim storage of new steam generators.
•Long term storage of old steam generators.
•Storage of low level radwaste.
•Providing training area with full scale mock up of the lower part of the steam generator.
Jobs like cutting, decontamination, welding, and closing of manways, were all trained on the mockup to familiarise personnel with the plant conditions. After the replacement performance, all kind of equipment and tools used inside the controlled area were transported to the decontamination shop to be checked and decontaminated.
A temporary building was erected to give access to controlled areas and was connected to the fuel handling building, as existing access could not handle the number of personnel. The building was equipped with lockers, showers, toilets and radiation monitoring system to check for contamination. The health physics control point was enlarged and remained in the existing access building.
Monitors in the access building were connected to cameras to check RCS activities. Two more cameras were located above 115m to check activities on the reactor operating deck. This avoided ‘nuclear tourism’, saving dose to staff who may have entered to see what was going on. There were over 38000 entrances and exits through the access building.
Before work began, several dose estimates were made based on radiation levels measured in previous outages, and on the likely manpower and time required for the work.
These calculations and the ALARA principle allowed optimisation of the work processes and the final amount of shielding determined. Dose reduction measures were:
•Training of staff on the mock-up.
•Flushing the primary circuit with H2O2 to decrease the activity in the circuit.
•Performing as much work as possible with steam generators filled and maintaining secondary side water level as long as possible.
•Refilling of crossover leg immediately after the decontamination to the shielding plug.
•Filling the RCS immediately after closing the last manway.
•Only carrying out vital work in the loop rooms prior to refilling the RCS.
•Before dose-intensive work, groups were briefed on radiation conditions.
•Dose-following during replacement to exercise counter-measures.
The final accumulated dose was about 1.4manSv. This was maintained, although more works were performed in the radiation areas than scheduled, and some dose levels were higher than previously measured.