Fuel Element ThermoMechanical Analysis Using the MGTP Code

1
INTERNATIONAL ATOMIC ENERGY AGENCY Technical
Meeting On On-Line Condition Monitoring of
Equipment and Processes in Nuclear Power Plants
Using Advanced Diagnostic Systems Knoxville,
Tennessee, USA June 27 30, 2005 BWR ONLINE
MONITORING SYSTEM BASED ON NOISE ANALYSIS Javier
Ortiz-Villafuerte, Rogelio Castillo-Durán
Gustavo Alonso-Vargas Instituto Nacional de
Investigaciones Nucleares jov_at_nuclear.inin.mx,
rcd_at_nuclear.inin.mx, galonso_at_nuclear.inin.mx Gabr
iel Calleros-Micheland Central Nucleoeléctrica de
Laguna Verde Comisión Federal de
Electricidad gcm9acpp_at_cfe.gob.mx
2
Nuclear Energy in MEXICO

• Two BWR/5 GE, 675 MWe any unit

3
Introduction

• Noise Analysis
• It is the Monitoring and diagnosis of dynamic
  properties of the fluctuations (periodic or
  random) of signals detected by some measurement
  device
• It is an effective tool for signal analysis due
  to its sensitivity
• the system is not disturbed by measuring a
  dynamic parameter.

4
Introduction

• Dominant Sources of Noise in BWRs are
• Mechanical vibrations of control rods, single
  fuel rod or bundles, and structures.
• fluctuations in temperature, pressure, coolant
  flow, and coolant boiling

5
Introduction

• Noise Diagnostic
• Detection of possible anomalies by continuous or
  periodic control of statistical parameters.
• a).- Base Signature
• The signature of the plant changes slowly with
  reactor operation conditions and plant life,
  throughout a fuel cycle
• b).- Statistical Parameters
• Power Spectral Density (PSD) and
• Cross Power Spectral Density (CPSD).

6
Monitoring System

• Data Collection
• The data collection module is the SIIP, Integral
  System of Information Process of the LVNPP.
• Signal Processing
• The signal processing module takes the original
• signal to extract the noise information.
• Evolutionary Matrix

7
Data Collection

• Capacity for the acquisition of around 3600
  signals at different sampling rates
• One to 250 samples per second
• Signals are acquired in digital form and stored
  for their consultation and historical analysis.
• The enormous amount of signals handled by the
  SIIP, causes that the sampling frequencies are
  just about the minimum necessary to analyze an
  event in the frequency domain.

8
Signal Processing

• DC level was extracted.
• Savitzky-Golay method is applied.
• Normalizaded Power Spectral Density is obtained

9
Evolutionary Matrix

• Evolving systems arise as an answer to the
  necessity of developing information systems whose
  mathematical models reflect the real system in
  the closest possible way, and that are able to
  support and absorb in real time the changes that
  happen in reality.

10
Evolutionary Matrix

• The evolutionary matrix consists of a number of
  m-element patterns, each in one row, a
  reinforcing parameter (h), and the meaning of
  each pattern in the matrix. The evolutionary
  matrix thus has a (m2)n structure, where n is
  the maximum number of patterns allowed.

11
Evolutionary Matrix

• n order to compare the new NPSD against each
  pattern already in the evolutionary matrix, the
  following equation is used

where S is the eucledian distance between the
incoming vector and one of reference, and m is
the number of data points of the spectrum in the
frequency domain.
12
Monitoring System

• This methodology, based on noise analysis, can be
  implemented in a monitoring system, which would
  allow, in principle, carrying out a continuous,
  automatic "on line" control of the plant, with
  the purpose of determining possible anomalous
  behaviors, and without interfering with its
  normal operation.

13
Problems at BWRs

• Vibrations
• Blockage at Jet Pump Inlet Nozzle
• Jet Pump Structural Faults
• Fracture in the Elbow of the Inlet Riser
• Fracture of the Support Arm of the Inlet Riser
• Fracture in the Inlet Riser

14
Structural Faults Reports

• SIL No. 220 (GE)
• SIL No. 330 (GE)
• SIL No. 551 (GE)
• SIL No. 605 (GE)
• IE Bulletin 80-07 (NRC)
• IN 97-02 (NRC)

15
BWR 5
16
Jet Pump
17
Structural Failure Evaluation

• Jet pump flaw evaluation procedures have been
  implemented in some BWR plants.
• The purpose of these procedures is to develop
  allowable continuos through-wall flaw sizes for
  all of the jet pump boundary pressure.

18
Vibrations

• Problems of vibrations have occurred with the
  power uprates of the reactors
• Susquehanna 2 was a abnormal vibration in the jet
  pumps due to a water leakage in the mixer and the
  diffuser junction of the jet pump

19
Events At LVNPP

• Partial Blockage at Jet Pump 11
• Partial Blockage at Jet Pump 6
• Malfunction of the Opening/Closing Controller of
  a Recirculation Flow Control Valve
• Detection of a Faulty Data Adquisition Card

20
Jet Pump Analysis

• Problems
• Turbulence
• Instrumentation

21
Partial Blockage at Jet Pump 11

• In 2002, in Unit 2, the jet pump 11 showed a
  pressure drop greater than 10 respect to the
  average of all the ten jet pumps on its same loop
  (loop B).
• During the following fuel reload (2003) the
  problem was eliminated.

22
Jet Pump 11
Typical signals of the pressure drop of jet pump
11 of the LVNPP Unit 2 of four years, and the
base signature determined by the evolutionary
matrix.
23
Jet Pump 11
Pressure drop signals of jet pump 11 of the LVNPP
Unit 2 during normal operation, and the case of
partial blockage of the jet pump inlet nozzle.
24
Partial Blockage at Jet Pump 6

• Later that year (2002), in the same Unit but on
  the other recirculation loop (loop A), the jet
  pump 6 also presented a fall in its pressure
  drop, with respect to the average of the rest of
  jet pumps on its loop.
• During the following fuel reload (2003) the
  problem was eliminated.

25
(No Transcript)
26
Jet Pump 6
Typical signals of the pressure drop of jet pump
6 of the LVNPP Unit 2 of five years, and the
base signature determined by the evolutionary
matrix.
27
Jet Pump 6
Pressure drop signals of jet pump 6 of the LVNPP
Unit 2 during normal operation, and the case of
partial blockage of the jet pump inlet nozzle.
28
Malfunction of the Openning/Closing Controller of
a Recirculation Flow Control Valve

• The third event analyzed here occurred in 2001,
  in Unit 2. An alarm of high scale on The Average
  Power Range Monitors (APRMs) occurred on
  different occasions.
• Seeking up different causes for possible positive
  reactivity insertions, it was concluded that the
  most probable source was the recirculation flow.
  Then noise analysis was applied to many different
  signals from different equipments on the
  recirculation loop.
• A deeper analysis showed a 0.5 Hz peak in the
  signals from the opening/closing positioning of
  the recirculation flow valve.

29
Malfunction of the Openning/Closing Controller of
a Recirculation Flow Control Valve

• Tracking this peak in signals from other
  equipment, it was found that one of the two
  recirculation flow signals presented the same
  peak, while the other did not. These two signals
  come from two sensors located in the
  instrumentation elbow.
• Maintenance to these two sensors showed that one
  of them was not properly calibrated. Since the
  signals of the recirculation flow are involved in
  setting the threshold of the high scale alarms on
  the APRMs, the improper calibration caused the
  alarms to start.
• Once the problem was solved, it was clear that no
  positive reactivity insertions events actually
  happened and no alarms occurred again.

30
Malfunction
Base signature and offnormal behavior of a
recirculation flow signal. The deviation from
the normal pattern was due to a miss-calibration
of a sensor. No actual physical event occurred.
31
Detection of a Faulty Data Adquisition Card

• The fourth event occurred in 2004, in Unit 1.
  During the monthly jet pumps pressure drop noise
  analysis, it was detected in the power spectra
  that the even jet pumps had a different behavior
  that those from the odd jet pumps, even though
  they come in pairs odd-even.
• In order to establish the cause of such
  difference, the reactor engineering reports were
  reviewed to look up for any problems on the
  operation of the recirculation loops or
  maintenance works. However, no indication of the
  possible cause of the difference in the power
  spectra was found out from the reports.

32
Detection of a Faulty Data Adquisition Card

• An important issue was that the source of the
  signals for all even pumps is the same data
  acquisition module, and similarly for the odd jet
  pumps.
• Then, other signals coming from the suspicious
  module were analyzed, showing again spectra
  different from the normal behavior.
• Finally, maintenance to this suspicious module
  solved the problem, and all the spectra of all
  the different signals showed again the historic
  patterns.

33
Faulty Data Adquisition Card
Base signatures of jet pumps 11 and 12 of Unit 1,
and detection of the malfunction of the
acquisition module that collects data of jet pump
12. No actual physical event occurred.
34
Conclusions

• A methodology based on noise analysis techniques
  to monitor abnormal behavior has been implement
• The methodology has been tested for off-line
  analysis
• The offline examples cover a wide variety of
  events that can occur in a nuclear power plant
  the detection of two different events of partial
  blockage at jet pump inlet nozzle malfunction of
  the opening/closing controller of a recirculation
  flow control valve and detection of a faulty
  data acquisition card. The events occurred at the
  two BWR Units of the Laguna Verde Nuclear Power
  Plant.

35
Conclusions

• The monitoring system is based on the analysis of
  the noise or fluctuations of a signal from a
  sensor or measurement device. Firstly, the base
  signature of the equipment or component is
  determined from historic records. Then, a
  real-time comparison of the Normalized Power
  Spectrum Density function of the signal can be
  performed against previously stored reference
  patterns in a continuously evolving matrix.
• Currently the methodology presented here is being
  considered for online implementation. Possible
  improvements or alternative analysis of the
  signals have also been introduced.

36
Acknowledgments

• Comisión Federal de Electricidad provided signal
  data of LVNPP.

37
Instituto Nacional de Investigaciones Nucleares