comment
Initial information issued by Kansai Electric and the Japanese nuclear regulator, the Nuclear and Industry Safety Agency, on the circumstances leading to the Mihama 3 tragedy (see this month’s news) tends to raise more questions than answers. It is of course all too easy to be wise after the event, nevertheless a particular concern must be how it was that the piece of carbon steel pipe that ruptured had never been inspected before, particularly in view of this material’s extremely well known and widely documented proneness to premature thinning under PWR condensate conditions. It is of little consolation that at the time of the accident preparations appeared to be in hand for non-destructive testing of this very section of pipework to take place within a matter of days, during the next periodic inspection. This was after over a quarter of a century of plant operation.
The pipe in question, within the secondary cooling circuit and apparently part of a steam supply system serving auxiliaries such as air conditioning and seawater desalination, was well down the pecking order in terms of nuclear safety – the event only rating 0+ on the International Nuclear Event Scale, despite four immediate fatalities.
But in terms of what might be called industrial safety the accident is clearly significant, and not just for nuclear plants but for any facility, including fossil-fired stations, using carbon steel in high energy water/steam pipework. Indeed, in the wake of the accident, NISA has very sensibly instructed licensees operating thermal power plants with steam turbines generating 1MWe or more to provide reassurance on the state of pipework considered prone to thinning due to erosion/corrosion (specifying main steam, reheat, condensate, feedwater, steam extraction and drain system pipework).
From what we know at the time of writing the Mihama 3 event has some worrying similarities to the catastrophic feedwater system pipe failure that occurred at the Surry unit 2 pressurised water reactor in the USA on 6 December 1986, also taking four lives. This event, where the pace of corrosion took the industry by surprise, placed the issue of “erosion/corrosion” or “flow assisted corrosion” in carbon steel pipework near the top of the agenda – at least for a time – and prompted the adoption of extensive wall thinning monitoring programmes.
The 1986 Surry event demonstrated that under a range of conditions of piping configuration, temperature, pressure, pH, and flow, not infrequently found in the secondary systems of pressurised water reactors, the rate of wall thinning of carbon steel pipework was much greater than expected, and that the effect was exacerbated by the kind of local turbulence and high flow velocities potentially created downstream of restricting orifices, flow control valves, flow meters and reducers and in elbows and tees.
In a pivotal Generic Letter on “erosion/corrosion-induced pipe wall thinning”, issued by the US Nuclear Regulatory Commission in May 1989 to all US nuclear plant licensees, the NRC drew attention to a number of other instances of the phenomenon besides the 1986 Surry event and reported that “incidents of pipe wall thinning or rupture because of erosion or erosion/corrosion have been reported at many other nuclear power plants.” In many of these cases, according to the Generic Letter, the licensees had inspected their two-phase lines for some years, but it was not until the Surry incident that they started to examine some single-phase lines, and discovered wall thinning in them. Instances mentioned in the May 1989 Generic Letter, in addition to the 1986 Surry 2 accident, included: a pipe rupture at Haddam Neck in March 1985 downstream of a feedwater heater control valve; discovery of extensive wall thinning in straight sections of main feedwater piping at Trojan in June 1987; and further instances of very rapid wall thinning in the feedwater system of Surry unit 2 discovered during the September 1988 outage.
Unfortunately in 1991 the NRC was prompted to issue a further notice on the subject alerting operators to “continuing erosion/corrosion problems affecting the integrity of high-energy piping systems and apparently inadequate monitoring programs.” Incidents mentioned in this 1991 notice included: rupture of moisture separator drain system pipes at Millstone 3 in December 1990; a steam leak in a feedwater regulating valve bypass line at San Onofre in July 1990; rupture due to thinning of a straight section of piping downstream of a control valve in the LP heater drain (LPHD) system at Surry unit 1 in March 1990; and rupture of a flow measuring orifice flange in the main feedwater system of Loviisa unit 1 in Finland in May 1990. The NRC notes that, “for all of these events, system temperature was in the range of 280 to 445 degrees F, system pressure was 500 to 1080 psi, flow was 9 to 29 feet per second and the piping material was carbon steel. Also in each event, flow turbulence was present.” These conditions envelope those at Mihama 3, while the March 1990 Surry 1 event illustrates the importance of a comprehensive monitoring programme in a way that seems to have particular relevance to the Mihama 3 case: “The licensee’s erosion/corrosion monitoring program included the LPHD system and provided for inspecting the wall thickness of the pipe elbow located immediately downstream of the failed piping. However, the program did not provide for an inspection for the short section of piping between the elbow and the level control valve.”
So the Mihama 3 event suggests that, despite numerous similar instances of wall thinning in nuclear plants over the years, the problems associated with corrosion of carbon steel pipe carrying water and/or steam in high energy systems have not been fully appreciated and tackled. Once again it seems that there has been a failure to fully take on board the lessons of previous incidents at other plants in other countries and to learn from mistakes.
It is essential that power plants in general, and the nuclear industry in particular, revisit the erosion/corrosion issue and provide assurance that monitoring and inspection procedures and programmes are up to the job and that the degradation mechanisms are fully understood – particularly in view of our increasing reliance on ageing pipework.