Last edited on 2011-06-01 03:44:30 by stolfi

Fukushima Daiichi
Units #1, #2, #3
Suspected errors in the data

Jorge Stolfi
2011-03-22 (updated 2011-05-29)

Several outlier points in the plots are suspected data readout or transcription errors. However the suspicion was not enough to allow or justify a correction.

Reactor #1

• 2011-03-17 12:50, the CAMS reading in the drywell is given as "4.10×10^-3 Sv/h" in NISA's Release 28. May be "4.10×10⁰ Sv/h" or "4.10×10¹ Sv/h".
• 2011-03-18 03:00, the CAMS reading in the drywell is given as "3.80×10^-3 Sv/h" in NISA's Release 29. May be "3.80×10⁰ Sv/h" or "3.80×10¹ Sv/h".
• 2011-03-18 07:55, the CAMS reading in the drywell is given as "3.75×10^-3 Sv/h" in NISA's Release 30. May be "3.75×10¹ Sv/h" or "3.75×10¹ Sv/h".
• 2011-03-20 15:00, the CAMS reading (A) in the drywell is given as "1.20×10¹ Sv/h" in NISA's Release 36. May be "1.20×10² Sv/h" or "4.20×10¹ Sv/h".
• 2011-03-23 21:00, the SLP1 source gives the torus CAMS "A" as "3.12E+1" Sv/h, while the adjacent values suggest some value closer to 4.50E+1 Sv/h.
• 2011-03-24 11:00, the given temperature "175C" is too low. Perhaps "225C"?
• starting 2011-04-17 12:00 (NISA's Release 98), the drywell CAMS reading (A) dropped by almost exactly a factor of 10 compared to previous values: it was given as "1.07×10⁰ Sv/h" but previous values suggest that the correct reading should be "1.07×10¹ Sv/h". This may have been incorrect carry-over of exponent from previous reading, "9.92×10⁰ Sv/h". The (B) reading is around 2.00 Sv/h, and readings have changed a lot recently, so it is not clear which exponent is correct. The drop in the factor of 10 persisted in several subsequent NISA releases, so perhaps the exponent error was in all previous readings instead.
• 2011-03-23 09:00, NISA release N045 gives 2.0 m^3/h = 33 l/min) as the water input through the water feed line. Preceding value (N044) is 300 l/min, the next value (N046) is 188 l/min. Also 2.0 m^3/h was the flow into the fire extinguisher line until 2011-03-22 15:30, the last reading before the switch to the water feed line.

Reactor #2

• 2011-03-13 9:55 and 10:35, the core pressures given as "1.283 MPa" and "1.263 MPa" are much lower than the previous value (6.08 MPa at 9:25) and the next one (5.85 MPa at 16:00). Interpolation of the smooth trend would give around 5.9--6.0 MPa. On the other hand, there seem to be sudden transient excursions in the drywell and torus pressures at that same time; so the data may be correct after all.
• 2011-03-15 11:42, NISA release N024, the water level "A" is given as "400 mm", may be the "B" reading or should be -1400.
• 2011-03-14 22:40, the SLP2 source gives the water level "A" as "-700mm", the correct value may be -1700 mm.
• 2011-03-16 14:00, the drywell pressure is given as "400 kPA" in the NIRS table, dropping to 75 kPA at midnight. But NISA's News Release N026 claims that the pressure had already dropped to "40 kPA" at 12:25. perhaps the time of the NIRS table is incorrect: instead of 14:00 it should be 12:00?
• 2011-03-17 11:25, the water level is given as "-1800 mm" in NISA's Release N028. That may be a typo; the correct value seems to be -1400 mm. Note that -1800 mm is the correct water level for reactor #1.

Reactor #3

• 2011-03-20 11:00, the water level in the core is given as "-2350 mm" in NISA's Release N036. That is a dip of 350 mm from the adjacent values. However the torus pressure at that time went off scale (greater than 400 kPa).
• 2011-03-20 15:20, the presure in the suppression torus is given as "800 kPa, ~ down scale" in NISA's Release N037. However the previous report gave "400 kPa ~ over scale", so the two values may be guesses for some unknown value above 400 kPa. Indeed most readings between 2011-03-14 and 2011-03-24 are given as "down scale", but seem to be "over scale" instead.
• 2011-04-16 12:00, the drywell CAMS (A) reading was given in NISA's Release N097 as "2.00×10¹ Sv/h", but adjacent values suggest that the correct reading should be around 1.6×10¹ Sv/h.
• 2011-03-17 11:00, the SLP3 source lists the torus pressure as "downscale", but 4 minutes before was 325 kPa. True that it was oscillating wildly, but...
• 2011-05-29 11:00, NISA release N153 gives 9.20 Sv/h for the CAMS drywell "A". Based on adjacent values it is probably 7.20 Sv/h.

Disclaimer

IMPORTANT: These plots, scripts, and data files are provided "as is", for the reader's convenience, with no guarantee whatsoever. There may be errors in the original document and/or in my processing of it.

This page is NOT an official document of the State University of Campinas (UNICAMP). Neither its author (me) nor UNICAMP should be held responsible for any damages that may result from the use of this information. Corrections and suggestions are welcome.