English
Menu

Mihama Unit 3 Pipe Rupture (2004): Thinning Downstream of an Orifice

On August 9, 2004, a secondary-system condensate pipe in the turbine building of Kansai Electric Power’s Mihama Unit 3 ruptured; five people working nearby were killed by the escaping steam and hot water and six were injured. The rupture was downstream of a flow-measuring orifice, where erosion–corrosion had thinned the carbon steel wall from its original 10 mm to 0.4 mm at the thinnest point. The location had been missing from the wall-thickness inspection list from the start.

What happened

Date
August 9, 2004, at about 15:28
Location
Turbine building of Mihama Unit 3, Kansai Electric Power Co., Japan
Pipe
Secondary condensate pipe (loop A, from the No. 4 low-pressure feedwater heater to the deaerator), downstream of a flow-measuring orifice. Carbon steel (SB42), outside diameter 558.8 mm, nominal wall 10 mm
Conditions
About 140 °C, 0.93 MPa, flow 1,700 m³/h (design conditions 195 °C, 1.27 MPa)
Wall thickness
0.4 mm at the thinnest point of the rupture (minimum required by the technical standard: 4.7 mm)
Casualties
5 killed, 6 injured
Investigation
Nuclear and Industrial Safety Agency (NISA), final report on the secondary piping rupture at Mihama Unit 3, March 30, 2005

Sequence of events and the investigation’s findings

The accident occurred during preparations for the unit’s 21st periodic inspection, when a total of 105 people were working in the building. Workers near the ruptured pipe were exposed to the escaping steam and hot water. The reactor shut down automatically in response to the secondary-system leak.

The rupture opening measured about 515 mm in the axial direction and about 930 mm around the circumference. The crack is reported to have grown first in the axial direction and then around the circumference. According to an information notice of the U.S. Nuclear Regulatory Commission (NRC), the same location on the loop B pipe was 1.8 mm thick.

The chain of causes from the materials side

Mihama Unit 3: the chain of causes from the materials sideAn orifice constricts the flowTurbulent flow downstream140 °C water flows over carbon steelThe protective oxide film dissolvesErosion–corrosion10 mm → 0.4 mm at the thinnestMissing from the inspection listThe thinning goes unnoticedFar below the 4.7 mm requiredThe pipe cannot hold the pressureThe pipe ruptures; steam escapes
Mihama Unit 3: the chain of causes from the materials side

The final report concluded that the pipe wall gradually thinned during operation by so-called erosion–corrosion until it lacked strength and ruptured. The orifice constricted the flow, exposing the pipe wall downstream to turbulent flow. The nuclear industry also calls this phenomenon flow-accelerated corrosion (FAC). The NRC noted that the inner surface of the rupture showed a fish-scale-like pattern while the bottom of the pipe kept its nominal thickness, which it described as characteristic of flow-accelerated corrosion.

Flow-accelerated corrosion proceeds as iron oxide on the surface of carbon steel dissolves into the flowing water and is carried away by the turbulent flow. According to a presentation published by the UK Institute of Corrosion, the thinning rate varies with temperature in a bell-shaped curve (peaking at about 140–160 °C in single-phase water flow), is faster at lower pH, and slows as the chromium content of the steel increases.

The rupture location had already been left off the list when Mitsubishi Heavy Industries prepared the inspection list under the PWR management guidelines in 1990, so the thinning beyond the usable limit went unnoticed. Registration omissions numbered 3 at Mihama Unit 3 and 42 across Kansai Electric’s 11 units. Japan’s White Paper on Nuclear Energy identified the root cause as a wall-thinning management error involving three companies (Kansai Electric, Mitsubishi Heavy Industries and Nihon Arm), in which locations needing management were missing from the original list and were not corrected before the accident.

The clues that revealed the cause

Open the guide with this case’s clues

Opening the guide with these clues ranks flow-induced wall thinning at the top.

Related failure modes: Flow-induced wall thinning (erosion–corrosion)

What changed afterwards: design, rules and inspection

Lessons

Check your understanding

When the corrosion current density of iron is uniformly 1.0 µA/cm² over the whole surface, which value is closest to the rate at which the surface recedes? Assume iron dissolves as Fe²⁺, with a molar mass of 55.85 g/mol, a density of 7.87 g/cm³, a Faraday constant of 96485 C/mol, and one year as 3.15×10⁷ s.

  1. AAbout 7.7 µm/year
  2. BAbout 23 µm/year
  3. CAbout 12 µm/year
Show answer and explanation

Answer: C. About 12 µm/year

The mass dissolved per unit area per second is iM/(nF) = 1.0×10⁻⁶ × 55.85/(2 × 96485) ≈ 2.9×10⁻¹⁰ g/(cm²·s); dividing by the density of 7.87 g/cm³ gives a recession rate of about 3.7×10⁻¹¹ cm/s. Multiplying by one year gives 3.7×10⁻¹¹ × 3.15×10⁷ ≈ 1.16×10⁻³ cm/year, about 12 µm/year (11.6 µm/year). About 23 µm/year is the mistake of taking n = 1. About 7.7 µm/year calculates with n = 3, as if iron dissolved as Fe³⁺, which does not match the stated conditions. Because current density and corrosion rate are proportional, “1 µA/cm² is about 12 µm/year” for iron is a handy conversion when estimating service life from a corrosion current.

In a steel pipe, a corrosion current in which iron dissolves as Fe → Fe²⁺ + 2e⁻ flowed at an average of 0.10 A for one year (365 days). Which value is closest to the mass of iron dissolved? Take the molar mass of Fe as 55.85 g/mol and the Faraday constant as 96485 C/mol.

  1. AAbout 0.61 kg
  2. BAbout 0.91 kg
  3. CAbout 1.8 kg
Show answer and explanation

Answer: B. About 0.91 kg

From Faraday's law, the dissolved mass is m = ItM/(nF). One year is 365 × 24 × 3600 = 3.15×10⁷ s, so the charge passed is Q = 0.10 × 3.15×10⁷ = 3.15×10⁶ C. Since iron goes to Fe²⁺, n = 2, and m = 3.15×10⁶ × 55.85/(2 × 96485) ≈ 913 g, about 0.91 kg. About 1.8 kg is the mistake of taking the number of electrons as n = 1. About 0.61 kg calculates with n = 3, as if iron went to Fe³⁺, which does not match the reaction given. Even a corrosion current of only 0.10 A loses nearly 1 kg of iron in a year, enough to perforate a pipe if concentrated locally.

What is the main reason stainless steel resists rusting in air and water?

  1. AAdding Ni makes the standard electrode potential of the alloy as noble as gold or platinum, so it hardly dissolves
  2. BA dense oxide scale several millimeters thick forms on the surface and seals it off from water and oxygen
  3. CIt contains about 11% or more Cr, forming a passive film a few nanometers thick that re-forms by itself
Show answer and explanation

Answer: C. It contains about 11% or more Cr, forming a passive film a few nanometers thick that re-forms by itself

Because stainless steel contains roughly 11% Cr or more (10.5% by the standard definition), a passive film consisting mainly of Cr oxide, only a few nanometers thick, forms on the surface and prevents further dissolution of the metal. What matters is that in environments containing oxygen the film spontaneously re-forms even when scratched. Passivity is therefore readily maintained in oxidizing environments such as nitric acid, but breaks down easily in oxygen-poor environments, chloride-rich environments and reducing acids (such as hydrochloric acid). The potential of the alloy as a whole does not become as noble as gold or platinum; the main role of Ni is to stabilize the austenitic structure. A thick oxide scale several millimeters thick forms by high-temperature oxidation, cracks and spalls easily, and is entirely different from a passive film.

Related terms

Passivity Faraday's law of electrolysis Stainless steel

Main sources

  1. Nuclear and Industrial Safety Agency, final report on the secondary piping rupture at Kansai Electric Power’s Mihama Unit 3, March 30, 2005 (Japanese; Atomic Energy Commission meeting material)
  2. Atomic Energy Commission of Japan, White Paper on Nuclear Energy (section on the Mihama Unit 3 accident; Japanese)
  3. U.S. NRC, Information Notice 2006-08: Secondary Piping Rupture at the Mihama Power Station in Japan
  4. JSME, history of the power generation equipment codes program (Japanese)
  5. Failure Knowledge Database: Steam Eruption from Nuclear Power Plant Cooling System
  6. R. Beheshty, Flow Accelerated Corrosion: Synergistic effects of the influencing parameters (Institute of Corrosion)

Dates, numbers of people, dimensions, temperatures and other values were checked against the official reports above and sources that cite them.

Other case studies

Back to the failure analysis guide