The accident of March 11th at Kakrapar was as Wellesley said of the battle of Waterloo, “A damn close run thing”. Although you would not guess that from reading the reports that have appeared in newspapers that have talked about “a small leak” in the primary heat transport system.
What is the Primary Heat Transport System
Once the chain reaction starts a lot of heat is produced in the core of the reactor. The primary heat transport system removes this heat from the core and takes it to steam generators for normal power operation. Two circulating pumps on each side of the reactor are connected to one reactor inlet header, from where coolant is directed to coolant channels through 153 inlet feeders. The coolant from the coolant channels flows to a reactor outlet header through 153 outlet feeders. There are no valves in the main circuit of primary coolant system. The temperature of the coolant inside the core varies from 249oC at inlet to 293oC at the outlet. As no boiling is allowed inside the core, pressure is maintained at 87 kg/cm2. From outlet headers coolant is carried to four steam generators.
When the reactor is in a shutdown state and primary circulating pumps are not available it is the elevation difference between the core and steam generators that provides driving head for coolant to flow to the steam generators.
This design has certain safety advantages over the light water cooled and moderated reactors that are in operation at Koodankulam and proposed for construction at Mithi Vhirdi, Kovada and Jaitapur. Although many eminent people (Even Dr kalam amongst them) have waxed eloquent regarding the safety features of Koodankulam reactors, by and large the Indian nuclear community has privately considered PHWR to be one of the safer designs. One of their best points is the fact that fuel can be added or taken out while the reactor is working. On power refuelling helps in maintaining low excess reactivity. There are two independent fast acting shutdown systems of diverse nature. Reactivity and shutdown mechanisms in the moderator are un- affected by the disturbance in the coolant channel. Large and subcooled inventory of moderator can act as an ultimate heat sink. Thus preventing severe core degradation under accidents. Water surrounding the calandria in calandria vault/shield tank system can hold the fuel channels in debris within the calandria under loss of coolant accident (LOCA) with failure of emergency core cooling system (ECCS) and moderator system.
Although it needs to be said that CANDUs have had quite a few failures of the pressure tubes in Canada as well as other countries.
What happened on March 11
The morning shift starts at 7. People come, show their gate pass, change their gear, go to the canteen to have breakfast; generally get ready to start working. In the meantime, two people from chemistry and two from health physics are supposed to collect samples from the operating reactor. Consequently the reactor building was unoccupied. This is what the management claims. I wonder what happened to the four people whose jobs did require them to be present at the time of the accident (9AM) and it is interesting to find that the whole shift takes more than two hours just to get ready for work!
At 9 O’clock pressure in the primary heat transport system (PHTS) suddenly fell. The people in the control room who tried to remotely see what was happening found themselves unable to make anything out since all they saw in the cameras was a blank wall of steam. The Loss of Coolant Accident (LOCA) had begun.
The PHWR design that is the mainstay of the Indian nuclear programme is a derivative of the Canadian CANDU design. It uses natural uranium as fuel and heavy water as both the coolant as well as the moderator.
LOCA is serious in any nuclear reactor, but in a CANDU it can be catastrophic. Unlike a PWR (Pressurized Water Reactor) (Koodankulam) or a BWR (Boiling Water Reactor) (Tarapur), CANDUs have a positive void reactivity coefficient. What this jargon means is that as the temperature rises in the absence of cooling following a LOCA water begins to boil. Voids are created. In a light water reactor like at Tarapur or Koodankulam, this means that not many neutrons get slowed down. In the absence of slow neutrons to cause further fission, the chain reaction automatically stops. This is an example of a passive safety system that does not need any active intervention on part of the operators. However, in a CANDU, the slowing down of neutrons is done by the moderator heavy water which is separate from the coolant heavy water. A loss of coolant accident here, does not affect the slowing down of neutrons. The chain reaction continues unless it is actively interupted. These reactions are taking place in timescales of nano to micro seconds and human reaction times are too slow for effective action. The RBMK type of reactor that exploded at Chernobyl also had a positive void coefficient of reactivity although it used graphite instead of heavy water for moderation.
There are two independent shutdown systems that work automatically on a signal from the reactor. First, there are Cadmium control rods that fall into the reactor and absorb neutrons. Secondly there is an injection of Lithium pentaborate that is put into the moderator. Boron is a very effective absorber of neutrons. In the absence of neutrons the chain reaction comes to a halt. If both these shutdown systems had not automatically worked for any reason, it would have been goodbye to South Gujarat. Thankfully both these systems did work and the reactor was shut down.
However, shutting down of the reactor is not the end of the story. The reactor during its operation produces many fission products. These fission products continue to give off heat. The halting of the reactor just puts a stop to the production of new fission products. It does not in any way affect the heat being given off by the fission products already in existance. The amount of this heat known as the decay heat is about seven percent of the heat that was being produced while the reactor worked. Unless this heat is somehow cooled, there can still be a disaster as happened in Fukushima. Shutdown reactors can and do explode unless cooling continues.
Emergency core cooling system (ECCS) is provided to remove core heat following loss of coolant accident. This system operates in three phases incorporating high pressure heavy water accumulators, intermediate pressure light water accumulators and low pressure – long term recirculation system. For catering to smaller leaks in the primary coolant system, a separate system called small leak handling system is provided.
When the ECCS was checked before the commencement of reactor operation in 1993, it did not meet its design specifications. At the time Dr P K Iyengar was the DAE head and he wanted the reactor started in a hurry (before personal retirement) without retesting to establish ECCS integrity. Read all about it in Anumukti.
The inside of a nuclear reactor is a harsh environment. Think of it as a giant, complicated pressure cooker. But pressure cookers don’t have large number of neutrons moving hither and thither, continuously hitting everything in their path. Years of this neutron bombardment takes its inevitable toll. Materials age; they become brittle like glass and can shatter without a moment’s notice. To prevent such events from happening, frequent monitoring is necessary.
The coolant heavy water is kept under pressure to prevent boiling. This heavy water is circulated through the reactor in 306 pressure tubes. The temperature inside these pressure tubes is around 2500 to 2900 centigrade. These pressure tubes are enclosed in another set of thinner tubes called the calandria tubes which contain the moderator heavy water. However, the calandria tubes are not pressurised and are at a much lower temperature.(~700 C)The two concentric tubes are kept separated from each other by four spacers (garter springs). The pressure tube and the end fitting are connected together by the rolled joint method. If the rolled joint process has not been carefully performed, the pressure tube is likely to be strained and excessive stresses are produced. These residual strains induce formation of metal hydrides and lead eventually to delayed hydride cracking. Till 1983 the pressure tubes that had failed had done so in a stable manner called leak before break. First small cracks developed and these cracks grew through the wall. There were small leakages of heavy water in the annular gas between the pressure tube and the calandria tube. These leakages were detected, the reactor was shut down and the tube was replaced before breakage.
On August 1st 1983 a Zirconium 2 pressure tube at Pickering nuclear plant broke without warning. Analysis of the cause revealed that one of the spacers (garter springs) between the two tubes had been placed wrongly and as a result the hot pressure tube had sagged and touched the cold calandria tube. The local cold spot in the pressure tube had induced a massive diffusion of hydrides to that spot and this is what had lead to sudden cracking. Further studies led to a change in the material of the pressure tubes from Zircalloy-2 to a Zirconium-Niobium alloy that was supposedly better at withstanding such cracking. This coolant channel replacement was done in Canadian CANDUs in the eightees.
In 1994, Pickering A was the site of Canada’s worst accident at a commercial nuclear station. OnDecember 10, 1994, a pipe break at Pickering reactor 2 resulted in a major loss of coolant accident and a spill of 185 tonnes of heavy water. The Emergency Core Cooling System was used to prevent a meltdown. About 200 workers were involved in the cleanup. The reactor was restarted 14 months later.
Replacement of coolant channels in Indian reactors was started at Rawatbhata in RAPS 2 in 1997. Since then, Kalpakkam, Narora, Kakrapar and Kaiga have all had en masse coolant channel replacement. Nuclear Power Corporation claimed that the reactors were now as good as new and would last another 50 years. Thirty years at full power. The coolant channel replacement work at Kakrapar was completed in 2011. Instead of lasting another 50 years as promised, it took only 5 years of operation for this break to occur and dump 70 tonnes of heavy water on the floor of the fuelling machine vault. That is just Rs 300 crores down the drain but truly that is just a “small leak” as memorably characterized by the station director in a country where crony capatalists abscond with thousands of crores outstanding in their names.
The behaviour of the station authorities was fully along expected lines. They of course, did not follow their own emergency preparedness plan manual. They know as well as everyone else that this weighty document is for public display purposes only. Although the event took place at nine in the morning they did not inform local civic authorities till well after three in the afternoon. Kakrapar plant falls on the border of Surat and Tapi districts. The Tapi district collector whose jurisdiction includes the colony of Kakrapar workers was kept totally in the dark days after the accident.
The head of the Department of Atomic Energy Sekhar Basu set a shining example to the rest of the team by saying that “radioactivity was confined within the reactor building itself and no radiation escaped outside the plant”. This despite the fact that venting was done to remove radioactivity from the confines of the containment building. Radioactive particles in India know their place. Outside the plant premises they disappear into a dimension undetectable through radiation monitors.
Even ten days after the incident, the accident is not fully under control. Heavy water and ordinary water is still coming out from some unknown tubes. The reactor is in a cold shutdown state, but the reactor building is still cordoned off. Only workers with “Green” clearance are allowed to go in. The entry of casual workers (more than 80% of the whole workforce) is still not allowed. Although construction work at units 3 and 4 is still continuing. Contractors are being asked to get workers with Green clearance to come and work in Kakrapar.
The future implications of this “small leak” are unfortunately many and generally unpleasant to the cushy world of nucleocrats. First of all they need to unambiguously fix the cause of the break. Why was there no indication of a leak before the break? Is their monitoring at fault? Was the material used in the Zircalloy Niobium tubes up to par? Or does this same problem exist in all the other CANDU type reactors in India? And so on and on…
What one needs at such a time is a tough regulator who can ask the nuclear establishment to halt operations so that answers to such questions can be explored. When the fire in the turbine building at Narora in 1994 had almost brought Chernobyl on to the Ganga ghats luckily we had Dr A Gopalakrishnan to ask tough questions. We need somebody with his gumption now. What we have instead is Mr Shiv Abhilash Bharadwaj who made a surprisingly sensible comment when he said that, “until the exact cause of the leakage is not pinpointed and till AERB is convinced whether the “failure is specific to the Kakrapar atomic reactor or is it a generic problem all caution has to be exercised”. He of course, immediately followed this up saying that “As of now he is not ordering an all-out shutdown of Indian heavy water atomic reactors simply because the backup safety systems worked ‘perfectly’ at Kakrapar.”
The country would be safer giving such people money to exercise their gambling skills at established places like Las Vegas or Macau instead of in the AERB office at Mumbai.