JP5574815B2 - Reactor power monitoring apparatus and method - Google Patents

Description

  The present invention relates to a power monitoring technique for an operating nuclear reactor.

The boiling water reactor (BWR) can control the output by changing the steam flow rate (void ratio) in the boiling core by changing the core flow rate.
However, depending on the core flow rate and other operating conditions, the neutron flux distribution and flow state in the core may become unstable due to void transport delay in the core and the negative feedback effect due to the negative void reactivity coefficient. It has been known.
As a result of this nuclear thermal hydraulic instability phenomenon, the output and flow rate are greatly oscillated, the heat removal characteristics at the fuel rod surface temperature deteriorate, and the soundness of the fuel rod cladding tube is impaired. Is done. For this reason, when designing the fuel and core of a boiling water reactor, nuclear thermal hydraulic stability analysis is performed to ensure that such vibration phenomenon does not occur in all expected operating areas. Designed to have
And in the area | region where such nuclear thermal hydraulic stability deteriorates, it sets beforehand so that a driving | running | working may be restrict | limited. Also, depending on the type of nuclear reactor, even if it enters into the operation restricted area, there is a type that has the function of reducing the output by inserting a control rod or the like and leaving the operation restricted area.

On the other hand, from the viewpoint of detection and suppression (Detect and Suppress), the output vibration phenomenon is allowed and the output vibration phenomenon caused by nuclear thermal hydraulic instability is accurately detected, and the vibration before the soundness of the fuel is impaired. There are also many nuclear power plants that suppress this.
For this reason, an output vibration detection algorithm called an OPRM (Oscillation Power Range Monitor) using a dedicated detection signal for detecting an output vibration phenomenon has been proposed (for example, Patent Document 1).

Also, there is a known method that performs principal component analysis of vibrations from multiple nuclear instrumentation signals, extracts independent components with different vibration modes, and evaluates the stability of different vibration modes: core integral stability and region stability. (For example, patent document 2).
And, there is a known method for evaluating nuclear thermal hydraulic stability in consideration of the core average neutron flux measurement value (APRM), the delay corresponding to the heat transfer time constant in the fuel rod, the main steam flow rate measurement value, etc. (For example, Patent Document 3).

Furthermore, a device is known that predicts and analyzes the stability of the reduction ratio and issues an alarm when the sequentially detected core stability exceeds the predicted value (for example, Patent Document 4).
In addition, if the indicators that deteriorate the nuclear thermal hydraulic stability of the boiling water reactor are sequentially calculated from plant information such as core power distribution, core flow rate, core pressure, and feed water temperature, and those indicators exceed the set value A technique for issuing an alarm is known (for example, Patent Document 5).
A method of monitoring the stability based on the neutron flux spatial mode distribution obtained by calculating the core eigenvalue from a number of LPRM signals is known (for example, Patent Document 6).

US Pat. No. 5,555,279 JP 2002-221590 A Japanese Patent Laid-Open No. 2002-181984 JP 2000-314793 A JP 2000-121778 A Japanese Patent Laid-Open No. 11-231089

However, in Patent Document 1, the boiling water reactor generally has a decrease in nuclear thermal hydraulic stability as the size, power density, and burnup increase. Cannot be used for type reactors.
In Patent Document 2 to Patent Document 6, nuclear thermohydraulic stability monitoring accuracy has been improved in the past, as a decrease in the margin of nuclear thermohydraulic stability is inevitable as core power and power density improve. It cannot be pursued more than that.

  An object of the present invention is to solve such a problem, and an object of the present invention is to provide a reactor power monitoring technique that improves the monitoring accuracy and reliability of nuclear thermal hydraulic stability.

A reactor power monitoring apparatus according to the present invention calculates a first stability index from time-series data indicating output oscillations of nuclear instrumentation signals output from a plurality of nuclear instrumentation detectors that detect neutrons in a core. A first determination unit that determines whether the nuclear thermal hydraulic stability of the core is stable or deteriorated by comparing the calculation unit, the first stability index and the first reference value, and the first determination A second calculation unit that calculates a second stability index of the core based on the time-series data when it is determined that the deterioration is in the unit, and the output by comparing the second stability index and a second reference value A second determination unit that determines whether or not to perform a vibration suppression operation, and in response to the determination in the first determination unit and the second determination unit, Either vibration information or automatic start signal of vibration suppression device Process the time-series signal that has passed through the low-pass filter even if it is a condition that does not reach the criteria for generating any of the alarm, vibration information, and automatic activation signal of the vibration suppression device. in case of reaching the other criteria set stricter than the reference, the alarm, characterized by Rukoto stepwise generate one of automatic start signal of the vibration information and the vibration suppression apparatus.
Further, in the reactor power monitoring apparatus, a first calculation unit that calculates a first stability index from time-series data indicating output oscillations of nuclear instrumentation signals output from a plurality of nuclear instrumentation detectors that detect neutrons in the core. And a first determination unit that compares the first stability index with a first reference value to determine whether the nuclear thermal hydraulic stability of the core is stable or deteriorated, and the first determination unit A second calculating unit that calculates a second stability index of the core based on the time-series data when it is determined that the deterioration has occurred; and comparing the second stability index and a second reference value, A second determination unit that determines whether or not to perform a suppression operation, and even if it is determined by the second stability index that the suppression operation of the output vibration is unnecessary, it is set more strictly than the first reference value When the first stability index satisfies the third reference value If there was boss, characterized in that the suppression operation of the output vibration is executed.

  According to the present invention, there is provided a reactor power monitoring technique that improves the monitoring accuracy and reliability of nuclear thermal hydraulic stability.

1 is a longitudinal sectional view showing an embodiment of a nuclear power plant to which an output monitoring apparatus according to the present invention is applied. 1 is a block diagram showing an embodiment of a reactor power monitoring apparatus according to the present invention. (A) (B) Explanatory drawing of a region vibration and spatial higher-order mode distribution predicted by the state prediction unit. Waveform graph of vibration impulse response when disturbance is applied to the system. Waveform graph of the nuclear instrumentation signal output from the detector. Explanatory drawing of the statistical processing part, the 1st calculation part, and dispersion | distribution analysis part which are applied to this embodiment. (A) (B) Overlaid graph of vibration period of a plurality of nuclear instrumentation signals. The graph which shows the standard deviation of the vibration period in all the nuclear instrumentation signals, and the graph which shows the standard deviation of the vibration period except an outlier. Frequency distribution graph of vibration period. Frequency distribution graph of reduction ratio. The standard deviation graph of a vibration period which shows the case where nuclear thermal hydraulic stability changes from a deterioration state to a stable state. The frequency distribution graph of the reduction ratio in the time zone of 250 seconds and 300 seconds of FIG. (A) (B) (C) (D) The figure which shows the pattern of the vibration centerline of a region vibration. The figure which shows grouping for area vibration monitoring. The figure which shows grouping for local vibration monitoring. (A) (B) Explanatory drawing of the position of the detector from which a reduction ratio shows the maximum value changing at the time of a region vibration. Standard deviation graph of group vibration period for area vibration monitoring. Comparison graph of spline fitting of time series data and its derivative. The graph which shows the locus | trajectory of the upper side peak value and lower side peak value detected by the peak detection method. Explanatory drawing of the operation | movement flow of a peak detection method. Graph of amplitude detected by peak detection method. Graph of peak values detected by the peak detection method. A graph of the reduction ratio detected by the peak detection method. Operation | movement explanatory drawing of a 3rd determination part. (A) (B) Explanatory drawing of the phase difference detection between each group. The flowchart of the reactor power monitoring method.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
A nuclear power generation system shown in FIG. 1 includes a nuclear reactor 10 that generates steam by heating reactor water by heat generated by nuclear fuel fission, a main pipe 21 that guides the generated steam to a turbine 22, and rotationally driven by the steam. A generator 23 that is coaxially connected to the turbine 22 that converts rotational kinetic energy into electrical energy, a condenser 24 that cools, condenses, and condenses the steam expanded by the work of the turbine 22, A water supply pipe 26 that feeds the liquid by the pump 25 and guides it to the nuclear reactor 10 is formed.

  The feed water returned to the nuclear reactor 10 is heated again as reactor water, and the above-described process is repeated to continuously generate power. And the output monitoring apparatus 30 which monitors the nuclear thermal hydraulic stability in the nuclear reactor 10 is provided so that this electric power generation may be maintained stably.

  The nuclear reactor 10 includes a pressure vessel 11 in which a shroud 15 is fixed in an interior filled with reactor water, a core support plate 17 fixed to the shroud 15, and a core support plate 17 supported by the core support plate 17. Are surrounded by a reactor core 16 and a steam / water separator 13 for separating the reactor water that has passed through the reactor core 16 into a gas-liquid two-phase flow.

Then, as described above, one steam separated from the steam by the steam separator 13 is guided to the main pipe 21 to contribute to power generation, and the other separated water joins with the feed water returned from the feed water pipe 26. The combined reactor water is lowered in a region (downcomer D) sandwiched between the shroud 15 and the pressure vessel 11 by a plurality of recirculation pumps 18 provided in the circumferential direction (only one is omitted in the drawing). To the lower plenum L.
Reactor water guided to the lower plenum L passes through the core 16 again and is heated to become a gas-liquid two-phase flow and reaches the upper plenum U. The reached gas-liquid two-phase flow is again guided to the steam separator 13 and the above-described process is repeated.

  As shown in the horizontal sectional view of FIG. 2, the core 16 absorbs and outputs a rectangular tube-shaped fuel assembly 33 in which a large number of fuel rods (not shown) are stored, and neutrons accompanying the fission reaction. 1 and the nuclear instrumentation detectors 31a, 31b, 31c, 31d for detecting the neutrons as shown in FIG. 1, and upper and lower ends are fixed to the upper lattice plate 14 and the core support plate 17, respectively. A large number of instrumentation tubes 34 are arranged.

  The instrumentation pipes 34 are installed at a ratio of about 1 to 16 fuel assemblies 33. For example, in an improved boiling water reactor having 872 fuel assemblies, 52 instrumentation pipes 34 are provided. ing. The nuclear instrumentation detectors 31a, 31b, 31c, and 31d provided at four locations in the vertical direction of the instrumentation pipes 34 respectively have an A level, a B level, a C level, and a D level according to the height position from below. being called. Then, the reactor water circulating inside the reactor core 16 flows in from the A level, is heated by the fuel, starts boiling, and sequentially reaches the B level, the C level, and the D level while changing the water / steam two-phase state. .

Nuclear thermal hydraulic stability is greatly affected by pressure propagation in this two-phase state of water and steam.
That is, as shown in FIG. 1, the two-phase state (water / steam ratio) fluctuates due to the pressure propagation delay of the reactor water from the bottom to the top, and the nuclear instrumentation detectors 31a, 31b, 31c, 31d , And a phase difference is generated in each of the nuclear instrumentation signals S (Sa, Sb, Sc, Sd) detected from the A level, B level, C level, and D level.

Since the phase difference of the output oscillation in the reactor water flow direction has an action of canceling the response of the nuclear instrumentation signals S to each other, it is possible to evaluate a plurality of nuclear instrumentation signals S at the same height level. It is preferable from the viewpoint of accuracy and reliability of monitoring of hydrodynamic stability.
Further, it is not necessary to monitor the stability from the A level to the D level, and the nuclear thermal hydraulic stability is evaluated at the B level, which is generally said to have the highest sensitivity of the stability monitoring.

  As shown in FIG. 2, the output monitoring device 30 includes a grouping unit 41, a first calculation unit 42, a reference value storage unit 43, a first determination unit 44, a peak detection unit 45, and a second calculation. The unit 46, the second determination unit 47, the third determination unit 48, the statistical processing unit 50, and the variation analysis unit 70 are configured.

The grouping unit 41 divides the plurality of nuclear instrumentation detectors 31 into groups based on information transmitted from the state prediction unit 60.
A nuclear instrumentation signal S processed by a first determination unit 44 and a second determination unit 47, which will be described later, is an individual signal of a nuclear instrumentation detector 31 selected from a group, or a plurality of nuclear meters in group units. It may be an average signal of the device detector 31.

The state prediction unit 60 includes a process calculator 61, a database 62, and a nuclear thermal hydraulic simulator 63.
The state predicting unit 60 configured as described above is a grouping unit for the core power distribution estimated based on the physical model or database, the spatial higher-order mode distribution of the neutron flux, the arrangement information of the nuclear instrumentation detector 31, and the like. 41.

  The statistical processing unit 50 performs an autoregressive analysis method, an autocorrelation function method, or a spectrum analysis on the time series data indicating the output oscillation of the nuclear instrumentation signal S output from the plurality of nuclear instrumentation detectors 31 that detect neutrons in the core. Statistical methods such as law are applied.

  The first calculation unit 42 calculates a first stability index from the time series data of the nuclear instrumentation signal S. Then, in the variation analysis unit 70, a variance or standard deviation indicating variations in a plurality of time-series data is obtained. Note that when calculating the standard deviation, time series data in which the vibration period varies beyond a certain range is excluded as an outlier.

  The reference value storage unit 43 stores a first reference value, a second reference value, and a third reference value that are used as threshold values in each of the first determination unit 44, the second determination unit 47, and the third determination unit 48. It is.

The first determination unit 44 compares the first stability index with the first reference value to determine whether the nuclear thermal hydraulic stability of the core is stable or deteriorated. If it is determined that the state is stable, the variation monitoring mode is continued, and if it is determined that the state is getting worse, the mode is shifted to the peak detection mode. Note that, even after shifting to the peak detection mode, the variation monitoring mode continues in parallel.
The first determination unit 44 determines that the nuclear thermal hydraulic stability is deteriorated when the standard deviation of the vibration period continuously exceeds the first reference value for a predetermined time.

The peak detection unit 45 performs peak detection of time-series data when the first determination unit 44 determines that the deterioration has occurred.
The peak detector 45 fits the set time-series data with a polynomial, and searches for a point where the differential value of the polynomial is zero as a peak. A spline function is applied as a polynomial.
When fitting time series data to a spline function, it is desirable to perform spline interpolation at the sampling interval when the nuclear instrumentation signal S is digitized.

The point where the differential value of this polynomial becomes zero is obtained as follows.
First, a search is performed within a period in which a margin is provided for half of the vibration cycle with the switching point from the first determination unit 44 as the origin. Then, among the sections sandwiched between the plurality of data points constituting the time series data, a section in which the product of the differential coefficients of the data points at both ends is a minus sign is obtained, and the section is further subdivided at the division points. The point where the absolute value of the coefficient is minimum is searched for as a differential value of zero.
Next, it is obtained by repeating the above-described search flow within a period in which a margin is provided for half of the vibration period with the searched point as the origin.

The second calculation unit 46 calculates the second stability index of the core based on the peak detection result of the time series data. Here, the amplitude or reduction ratio of the plurality of time series data is used as the second stability index.
In other words, the searched peak position (time axis) is substituted into the spline function of that section to obtain the peak value (vertical axis), and the difference from the adjacent peak value obtained in the same way is set as the amplitude, and the adjacent peak position The difference in time is taken as the vibration period.

The second determination unit 47 determines whether or not to operate the output suppression device 35 by comparing the second stability index with the second reference value. And when the 2nd determination part 47 does not determine to operate the output suppression apparatus 35 within predetermined time, the operation | movement of a said 2nd calculation part is stopped. Note that the time required to stop the operation of the second calculation unit 46 is desirably set longer than the time required for the determination by the first determination unit 44.
Note that the first calculation unit 42 operates in parallel even when the second calculation unit 46 is operating, and the second determination unit determines that the power control of the core is to be executed within a predetermined time. If not, the operation of the second calculation unit 46 is stopped and only the operation of the first calculation unit 42 is continued.

  Even if the second determination unit 47 determines that the output suppression device 35 does not require an operation based on the second stability index, the third determination unit 48 sets the third reference value that is set to be stricter than the first reference value. When the stability index is satisfied, it is determined that the operation of the output suppression device 35 is necessary.

  Incidentally, in response to the determinations in the first determination unit 44, the second determination unit 47, and the third determination unit 48, as an output vibration suppression operation, an alarm device (not shown), a vibration information providing device (not shown), and an output Any one of the automatic activation signals of the suppression device 35 may be generated step by step.

  Power oscillations due to nuclear thermal hydraulic stability are generated macroscopically throughout the core due to instability of the flow conditions in the fuel channel surrounding the fuel assembly 33 due to reactivity feedback with respect to the dynamic response of the neutron flux. It is a phenomenon. This output vibration is thought to be caused by excitation of the spatial mode of the neutron flux by this reactivity feedback.

  When the excited spatial mode is the fundamental mode, this is called core integral vibration, and the phases of the output vibrations of the core cross section at the same height are basically aligned. In this case, the plurality of nuclear instrumentation signals S measured in the same cross section have almost no phase difference and are not canceled even if they are added, and vibration detection using the average signal (APRM) is sufficiently possible.

  On the other hand, when the excited spatial mode is a higher order mode, it is called region vibration, and according to the spatial higher order mode distribution, there is a phase difference in each nuclear instrumentation signal S in the core section of the same height. Arise. The node of this spatial higher-order mode distribution becomes the center line of vibration, and the phase difference becomes 180 degrees across this center line, and the vibration is reversed.

FIG. 3A shows a spatial higher-order mode distribution in the region vibration, and the two regions b across the center line c of the vibration corresponding to the node as seen in a horizontal section in FIG. , C oscillate with opposite phases, that is, with a phase difference of 180 degrees.
In this case, if a plurality of nuclear instrumentation signals S extending over the two regions b and c are averaged, vibration cancellation due to the phase difference occurs, and the amplitude of the averaged signal is smoothed, making it difficult to detect vibrations. That is, the use of an APRM signal that averages and outputs the entire core is inappropriate for detection of such region vibrations.
Although illustration is omitted, it can be said that the use of the APRM signal is inappropriate for detection of local vibrations that output vibrations in a narrow region around a specific fuel assembly 33 (FIG. 2).

Here, referring to FIG. 4, the reduction ratio, the vibration period, and the amplitude are defined using a vibration impulse response when a disturbance is applied to the system. The peak values are set in order as X ₁ , X ₂ , X ₃ , X ₄ ,..., And the times when these peaks appear are T ₁ , T ₂ , T ₃ , T ₄ ,. For example, a reduction ratio, a vibration period, and an amplitude that are generally used as indices representing the stability of nuclear thermal hydraulic stability are defined as follows.
Reduction ratio = (X ₃ −X ₄ ) / (X ₁ −X ₂ )
Vibration period = (T ₃ −T ₁ ) or (T ₄ −T ₂ )
Amplitude = (X ₃ −X ₄ ) or (X ₁ −X ₂ )
The phase difference is defined as an angle in which a time difference between T _n between a plurality of signals is 360 degrees for one cycle.

If this reduction ratio is less than 1, the impulse response is attenuated, so this system is stable. If it exceeds 1, the vibration grows and the system becomes unstable. Further, when the reduction ratio is 1, the vibration continues with a constant amplitude.
Further, the shorter the vibration period, the faster the vibration grows or attenuates. The reciprocal of this vibration period is generally called a resonance frequency or natural frequency, and the unit is expressed as Hz or cps.

Here, although FIG. 4 can be said to be an ideal impulse response, the actual response of the nuclear instrumentation signal S as shown in FIG. 5 is not an ideal impulse response. That is, high-frequency noise and low-frequency trends are superimposed on the actual nuclear instrumentation signal S, or repetitive vibrations as shown in FIG. 5 do not show monotonous attenuation or growth.
For this reason, even if the peak is detected as it is from the response of the actual nuclear instrumentation signal S and the reduction ratio is calculated from the amplitude, it is not the impulse response of the system, so the reduction ratio is accurately estimated. I can't.

Therefore, in order to obtain the impulse response, it is necessary to estimate the transfer function of the system, and for this purpose, it is necessary to apply a statistical method using a certain data length (the statistical processing units in FIGS. 2 and 6). 50 functions).
The reduction ratio derived from the response of the nuclear instrumentation signal S in FIG. 5 using the estimated transfer function is about 0.7 to 0.8, and this response repeats growth and decay. Regardless, it can be said that the system is stable.
In this case, as long as the fuel soundness is not threatened by the output vibration, it is only necessary to monitor the system so that the instability of the system does not progress.

  However, the reduction ratio is appropriate in that it is an index that can directly evaluate the nuclear thermal hydraulic stability, but its accurate estimation requires a certain length of data. Further, when the vibration state depends on the core position, such as the region vibration, the value varies for each nuclear instrumentation signal S of interest. For this reason, it is difficult to determine whether or not to apply an operation for starting the output suppression device 35 (FIG. 2) or the like and improving the core stability.

  In the present embodiment, the statistical processing unit 50 and the first calculation unit 42 are used, and the vibration period of each nuclear instrumentation signal S is applied as a parameter for monitoring deterioration of stability. Then, the variation analysis unit 70 analyzes the plurality of nuclear instrumentation signals S in the range grouped by the grouping unit 41. In addition, although dispersion | distribution or a standard deviation is mentioned as a parameter | index which shows dispersion | variation, a standard deviation is employ | adopted here from the ease of selection of a criterion.

The nuclear instrumentation detector 31 (LPRM: local output region detection system) is grouped as will be described later, the processing of the statistical processing unit 50 is performed for each group, and the oscillation period of the individual signal is obtained. The standard deviation is obtained.
The standard deviation is compared with the first reference value of the reference value storage unit 43 in the first determination unit 44, and if the determination criterion is satisfied, the peak detection unit 45 is activated based on the determination time. Even after the peak detection unit 45 is started, the vibration period variation analysis by the first determination unit 44 is processed in parallel.

A method for obtaining the vibration period by a statistical method will be described with reference to FIG.
First, to generate a nuclear instrumentation signal digital processing to S, and series data X _t when subjected to a treatment to remove the noise and trend component. The method of the time-series data X value of the correlation function calculated directly autocorrelation function from _t is maximized delay time and the vibration period, it obtains the spectral density due FFT (fast Fourier transform) and autoregressive method The method that uses the reciprocal of the frequency (resonance frequency, Hz) that maximizes the vibration period, the transfer function obtained by the autoregression method, the vibration period obtained from the resonance frequency estimated from the transfer function pole, and the impulse response obtained by the autoregressive method The vibration period is obtained by a method for obtaining the vibration period from the relationship shown in FIG.
Regardless of which method is used, a certain amount of data length (including vibrations of several cycles or more) is required for accurate estimation.

The autocorrelation function is the covariance between the time-series data X _t at a certain time t and the past value X _t−1, and the delay time at which the value becomes maximum corresponds to the oscillation period.
Autoregressive methods for time series data X _t, performs a fit linear as shown in FIG. 6 (1), estimates the autoregressive coefficients a _k (FIG. 6 (2)). Here, _et is Gaussian noise that is out of the fit. Several algorithms for efficiently estimating such autoregressive processes have been devised and are widely used.

The time-series temporal response characteristics are reflected in the autoregressive coefficient. By using this coefficient, it is possible to obtain the reduction ratio, vibration period, and phase difference.
There are methods for estimating the reduction ratio and the vibration period from the autoregressive coefficient by estimating from the impulse response, the spectral density, and the transfer function, respectively.
First, the impulse response is obtained by the equation (3) in FIG. 6 using the autoregressive coefficient.

  The impulse response is the response shown in FIG. 4, and the reduction ratio and the vibration period can be obtained based on the equation (4) in FIG. In order to suppress estimation variation, an average value is used as an estimated value for the reduction ratio and the vibration period.

The autoregressive process is expressed as shown in FIG. 6 (5) using the delay operator Z ⁻¹ (X _t−1 = Z ⁻¹ X _t ) in discrete (digital) time. The reciprocal of A (Z ⁻¹ ) shown in equation 7) is a transfer function from Gaussian noise to time series, and stability information is included in this transfer function.
When this transfer function is used, the spectral density function S (f) is given by the equation (6) in FIG. The frequency f _max at which this spectrum is maximized is the resonance frequency, and its reciprocal 1 / f _max is the vibration period.

The zero of A (Z ⁻¹ ) is the pole of the transfer function, and stability is determined by the positional relationship of this pole in the complex plane, so the reduction ratio and the vibration frequency are estimated by estimating the transfer function pole. be able to.
As shown in FIG. 6 (8), if the transfer function pole = (P _R , P _I ), the reduction ratio and the vibration period are as shown in FIG. 6 (10) from the relationship shown in FIG. 6 (9). expressed. Here, Δt is a sampling period of the time series signal.

Here, a group including a total of N nuclear instrumentation signals S is considered. Then, let T _i (t) be the vibration cycle of the i-th signal. Each vibration period T _i (t) is calculated using the statistical method described above.

FIG. 7A is a response example of each vibration period T _i (t) calculated by the first calculation unit 42 from a plurality of nuclear instrumentation signals S indicating a stable state, and FIG. 7B is an unstable state. It is an example of a response of the state which is moving to. In addition, although FIG. 7 (A) and FIG. 7 (B) are divided into upper and lower parts, they are continuous results.
Thus, it can be seen that both the variation with respect to time of individual vibration periods and the variation between signals are remarkable in the stable state, and both are small in the unstable state.

Here, the variation analysis unit 70 (FIGS. 2 and 6) calculates the variation (standard deviation) σ _T (t) of the vibration period within the group. In FIG. 6, a symbol with a superscript bar indicates an average value within a group of vibration periods in a target time interval (represented by t).

FIG. 8 shows the response of the standard deviation of the vibration period of FIG. 7. The value of this standard deviation decreases from after 450 seconds and converges to a constant value from around 750 seconds where the vibration is sufficiently developed. I understand.
This is thought to be because nuclear thermal hydraulic stability depends on the instability of individual fuel channels and the dynamic instability of density wave oscillation.

That is, the passage time of the two-phase flow differs due to the difference in the two-phase flow state in each fuel channel, and the resonance frequency of the density wave oscillation is usually unique and different for each fuel channel. However, in the process of macroscopic instability growing throughout the core via neutron dynamics, non-linear frequency entrainment occurs and the oscillation period of each individual fuel channel is inherent to macro instability. It is thought to be drawn into the cycle.
For this reason, the fluctuation of the vibration period decreases in the process of instability growing macroscopically, and it is recognized that it is effective to use the fluctuation of the vibration period as a monitoring parameter for nuclear thermal hydraulic instability.

  However, the discontinuous values appearing in the vicinity of 100 seconds and 200 seconds in the graph a in FIG. 8 are due to the presence of outliers shown in FIG. When a sufficient number of nuclear instrumentation signals S are included in the group, the standard deviation has robustness against such outliers. However, since the average value and variance (standard deviation) are not robust statistics in the first place, it is necessary to exclude such outliers when calculating the monitoring parameter.

  As a method for removing such a deviation value, for example, there is a method in which, after first calculating a standard deviation, a vibration period that is deviated by a constant multiple of the standard deviation from an average value of vibration periods is excluded. That is, with C as a constant, the standard deviation is calculated anew from only the data of the vibration period within the range of the left expression 70 in FIG.

The graph line b shown in FIG. 8 shows a case where C = 1 and discontinuous responses due to outliers are removed. Note that the value of C is set in consideration of other stability parameters such as a reduction ratio in advance.
Occurrence of outliers may be due to a measurement system failure or insufficient range adjustment. In that case, it is possible to exclude the nuclear instrumentation signal S from the monitoring group by performing a range check in advance, and it will be excluded until the failure or adjustment deficiency is improved.
Apart from such a case, a deviation value may be calculated in the course of the statistical processing calculation. In this case, the relevant nuclear instrumentation signal S is automatically excluded from the monitoring group only when the outlier value is calculated, but if the outlier value is not calculated, it is taken into the monitoring group again.

Here, the sampling period for the time series data X _t and digitizes nuclear instrumentation signal S has a 25 msec. This sampling period is regarded as the first reference value c when the vibration period of the nuclear instrumentation signal S is aligned due to deterioration of stability, and is indicated by a one-dot chain line in FIG.

In FIG. 8, when the response of the graph line b corrected for the deviation value of the standard deviation is confirmed, the graph line b once falls below the first reference value c around 225 seconds, but immediately exceeds 25 milliseconds, and 470 It falls below around seconds, but rises immediately as well.
Next, the graph line b is below the first reference value c around 540 seconds, and after that, it continues to exceed 25 seconds around 625 seconds, but decreases rapidly after 650 seconds and falls below 10 milliseconds. It has converged to a small value.

  By the way, the reduction ratio actually starts to increase suddenly from around 640 seconds, and a remarkable vibration component is observed in the nuclear instrumentation signal S from that vicinity. Therefore, it is considered appropriate that the first determination unit 44 determines that the stability of the nuclear instrumentation signal S is deteriorated after about 600 seconds and switches to the peak detection mode.

However, in order to exclude a jump value that satisfies the first reference value c by chance, it is not determined only once, but the continuation time is monitored based on a time point when the first reference value c falls below, and a predetermined continuation length is determined. Switch to peak detection mode when it exceeds the limit.
This continuation length depends on the vibration period and the data length used for the calculation, but is set to several times or more of those values. That is, the peak detection mode is switched when the graph line b continues below the first reference value c several times or less.
For example, if the duration is set to 15 seconds in FIG. 8 for 5 times, the peak detection mode is switched around 555 seconds.

  On the other hand, as the first stability index, an association between the standard deviation of the vibration period and the reduction ratio can be used. This is effective for detection when the output vibration becomes apparent in a part of the region, not in the entire core, for example, in the region vibration.

In the vicinity of 450 seconds of the response of the graph b in FIG. 8, the first reference value c is not reached, but the variation of the vibration period is clearly reduced as compared with before that.
FIG. 10 shows the distribution of the reduction ratio of each nuclear instrumentation signal S at the 450 second time point. A plurality of nuclear instrumentation signals S with a reduction ratio of 0.8 are spatially adjacent and are considered to be in a previous stage that develops into a region vibration.

  Here, for example, a case where there are a plurality of maximum attenuation ratios of the nuclear instrumentation signal S or a reduction ratio exceeding 0.8 is set as a starting point of instability. Then, if 30 milliseconds, which is 20% of the sampling period, is set as a reference value, the time point of 450 seconds is determined as the start point of deterioration of stability, and switching to the peak monitoring mode is enabled.

By the way, when stability is recovered after shifting to the peak detection mode due to the determination of stability deterioration in the first determination unit 44, it is also necessary to determine to return to the original vibration cycle variation monitoring mode.
FIG. 11 shows a state in which the stability recovers from the state of stability deterioration with little variation and the variation increases in the response of the standard deviation of the vibration period. That is, in this graph line, 230 seconds or later corresponds to the stability recovery state.

Here, the sampling period is 80 milliseconds, and the standard deviation of the vibration period before 230 seconds is an order of magnitude smaller than that, indicating that the output vibration is developed before 230 seconds.
After this output vibration suppression operation is executed, the standard deviation of the vibration period increases rapidly, exceeds the sampling period of 80 milliseconds around 240 seconds, and thereafter changes to a value around 100 to 140 milliseconds. ing.

FIG. 12 shows the distribution of the reduction ratio of the individual nuclear instrumentation signals S around 250 seconds and around 300 seconds. According to this distribution, a plurality of signals exceeding the reduction ratio of 0.8 remain, and the stability is insufficient. In other words, if the stability recovery judgment criterion is set to a sampling period equivalent to the stability deterioration judgment criterion, there is a high possibility that it will be non-conservative in terms of safety.
Therefore, a larger criterion (stricter) is used as the criterion for restoring stability than the criterion for deteriorating stability. For example, as a criterion for stability recovery using standard deviation, if the sampling period is 1.5 times and the duration is 10 times the period, in the example of FIG. do not do.

Next, the grouping unit 41 will be described.
In ABWR, there are 208 nuclear instrumentation signals S (LPRM signals) in total. In order to monitor the fluctuation of the vibration period, it is desirable to group the nuclear instrumentation detectors 31b (FIG. 1) of the B level having the highest signal average value among the four levels. This is because a phase difference appears in vibration between detectors having different axial levels, and output vibration is likely to occur when the lower power distribution is high.

Note that a group including all of these nuclear instrumentation signals S (LPRM signals) may be prepared, and any of APRM signals divided into eight groups can be used.
However, when the APRM signal is used, it is the core integrated vibration that is expected to be detected in the group. In this case, it is considered that there is substantially no difference from normal APRM monitoring. In some cases, it is necessary to select an operation method in which a channel that does not include an outlier or a failure detector is selected or a channel is switched in the middle.

Generation of region vibrations is greatly influenced by lower distortion of the output distribution, so group them according to the A level or B level from the four detectors 31a, 31b, 31c, 31d (FIG. 1). Is effective in determining the vibration mode from the phase difference.
As shown in FIGS. 13 (A), (B), (C), and (D), the core management is performed on a quarter target, and the spatial higher-order mode distribution generally corresponds to one of these four patterns. Is. Therefore, 8 groups divided by these 4 patterns are prepared for monitoring the region vibration.

  Based on the spatial higher-order mode distribution obtained by the state predicting unit 60, the number of nuclear instrumentation signals S necessary for use in the order of higher-order mode distribution is selected within each group. When there is no higher-order mode distribution, signals are selected in descending order of output distribution, or selected from a fixed arrangement determined in advance, for example, two signals from the outermost peripheral portion and three signals from the inside.

FIG. 14 shows a case where 10 signals are assigned to each group. Here, ten detectors (detectors circled in the figure) that are closest to the node having the higher spatial high-order mode value are selected in each group region.
The pattern shown in FIG. 14 corresponds to the pattern shown in FIG. 13D, and two groups in which 10 signals are assigned to each group are also prepared for the pattern shown in FIG.
If the patterns in FIGS. 13C and 13D are predicted in the high-order mode distribution prediction, it is not necessary to prepare the patterns in FIGS. 13A and 13B. In this case, four groups are selected for area vibration monitoring.

Next, a group corresponding to local vibration monitoring is set.
Since local vibrations are likely to occur in the thermally severest fuel channel, the fuel that is obtained by the state prediction unit 60 when the most radial power distribution is severe and the gross including the effect of the axial power distribution. Select the detector closest to the fuel with the severe power distribution.
If the detectors selected from these two power distributions are different, prepare two different groups, or select one of the fuels with severe fuel health or the severer fuel with higher spatial mode distribution. Or select as appropriate.

As shown in FIG. 15, when the position of the detector d is selected as the fuel for monitoring local vibration, the four detectors e closest to the detector d are selected and combined with the detector d. Group five.
In this case, the stability index (vibration period, reduction ratio, amplitude) is calculated independently from the detector d, and the standard deviation of these stability indices based on the five signals is calculated as a local vibration monitoring signal. Further, there is a method in which the detector closest to the peak of the higher-order mode distribution in the region vibration monitoring group in FIG. 14 is individually set to a group for local vibration monitoring.

The above is a method in which a detector adjacent to a fuel assembly to be monitored is fixed in advance based on the predicted output distribution, higher-order mode distribution, etc., and local vibration is monitored using the signal from this detector. It is.
On the other hand, a method of sequentially selecting a monitoring signal from the group used for monitoring the reactor core integrated vibration is also conceivable. This is a method of selecting, as a local vibration monitoring signal, a signal representing the maximum value of the reduction ratio or the amplitude among the responses of the individual signals from the detectors constituting this group.

16A and 16B show the transition of the maximum reduction ratio of the original nuclear instrumentation signal S used in FIG. In about 1000 seconds, the detector having the maximum reduction ratio changes over the four sections a → b → c → a.
Among these, the detector a located at the lower right of the core has the largest reduction ratio in the section a, and it moves to the detector b at the symmetrical position of the core in the section b, and then to the detector c in the section c. Change and return to detector a in section a.

  The signal having the maximum reduction ratio is regarded as the nuclear instrumentation signal S determined to be the most unstable and used as a local vibration monitoring signal. However, since it is inappropriate to change the monitoring signal frequently, the signal should be changed when the reduction ratio of the monitoring signal suddenly decreases or when a larger reduction ratio continues several times the vibration period. carry out. However, it is fixed after the time when the monitoring mode is switched to the peak detection mode.

As shown in FIG. 3 (A), the region vibration is caused by the deterioration of the stability of the spatial higher order mode of the neutron flux. For this reason, the distribution of vibration modes in the core, that is, the distribution of vibration phase differences is similar to the spatial higher-order mode distribution.
Therefore, the power distribution in the operating state where the core is likely to become unstable and the prediction function of the higher-order mode of the neutron flux space are provided together to effectively monitor the region vibration.

  FIG. 2 shows a state prediction unit 60 that predicts the power distribution and the higher-order mode of the neutron flux space. The process calculator 61 estimates plant parameters based on a nuclear thermal hydraulic physical model while capturing plant data. The nuclear thermal hydraulic simulator 63 of the process calculator 61 is used to calculate the output distribution or higher-order mode distribution under operating conditions where the stability deteriorates.

The operating conditions where the stability deteriorates are low flow rate and high output conditions after the pump trip, and such conditions are obtained in advance by analysis for each operating cycle.
When the nuclear thermal hydraulic simulator 63 does not have a function for calculating a higher-order mode distribution, a distribution calculated offline in advance according to the control rod pattern, burnup, etc. is stored in the database 62, and the distribution under the operating conditions is stored. Is calculated by interpolating from the database distribution.
From the output distribution and higher-order mode distribution thus obtained, the grouping unit 41 sets a signal group for monitoring region vibration and local vibration.

Next, determination of region stability by grouping will be described.
FIG. 17 is a diagram in which the top 10 signals of the higher-order mode distribution as shown in FIG. 14 are grouped using the pattern of FIG. 13C and the standard deviation of the vibration period within the group is calculated. is there. On the other hand, FIG. 8 described above uses an average group of the entire core.
In the vicinity of 450 seconds in FIG. 17, the standard deviation of the vibration period is less than 25 milliseconds, which is the sampling period of the time series data, and this is when the stability criterion in FIG. 8 is multiplied by 1.2. Equivalent to.
In FIG. 16, this group includes a signal having the highest reduction ratio in the section c, which corresponds to the determination of the deterioration of stability based on the maximum reduction ratio of 0.8.

Next, specific examples regarding the peak detection mode will be described.
After detecting the deterioration of stability by monitoring the standard deviation of the vibration period, the peak detection mode can quickly detect the growth of the output vibration in order to avoid the deterioration of the fuel and plant health caused by the output vibration. It is necessary to switch to urgently.
In a state where the stability has deteriorated to some extent, it is considered that the vibration component affecting nuclear thermal hydraulic stability is superior to the noise component.

Therefore, the peak can be detected analytically by fitting the outstanding vibration component to a polynomial that can be analyzed. In this case, since the data length to be used is smaller than that of the statistical method, the responsiveness is also increased.
However, if fitting is performed for the entire data length, it cannot be fit well with a relatively low-order function. Also, if the function is too complex, it will depend too much on the data, and robust fitting will not be possible.

Therefore, a piecewise polynomial approximation for dividing and fitting time series data is adopted.
A typical example is a spline function, and a cubic spline interpolation method that approximates a small section with a cubic polynomial that is most commonly used among them is used. That is, the time of the time series data is divided into sections t _n ≦ t ≦ t _{n + 1} and the section is interpolated by the following cubic equation.
S _n (t) = a _n + b _n (t−t _n ) + c _n (t−t _n ) ² + d _n (t−t _n ) ³

By applying the time series data to the polynomial in this way, the differential coefficient can be easily obtained analytically. If the time when this differential coefficient becomes zero is obtained, it corresponds to the time when the peak (peak or valley) of the vibration appears.
dS _n (t) / dt = b _n + 2c _n (t−t _n ) + 3d _n (t−t _n ) ²

FIG. 18 compares the spline function a with the derivative b obtained by differentiating the spline function a. Here, the sampling interval of the time series data is used in the data section of the spline interpolation. That is, the sampling interval is interpolated by t _{n + 1} = t _n + Δt with Δt as the sampling period.

As a search method for the time when the derivative b becomes zero, since the derivative of the cubic spline function is a quadratic function, it is often difficult to stably obtain the root of the quadratic equation. For this, the following procedure is used.
First, an interval in which a value obtained by multiplying adjacent differential coefficients is negative is searched. The differential coefficients at the end points [t _n , t _{n + 1} ] correspond to the coefficients of the first-order term of the spline function a, and thus are [b _n , b _{n + 1} ], respectively.

  Here, it is necessary to apply a low-pass filter to the time-series data in order to remove fine fluctuations in the estimation error of the differential coefficient. When a section where the differential coefficient is zero is estimated, the section is further subdivided to obtain a differential coefficient, and a point where the absolute value becomes the smallest is taken as a point where the differential coefficient becomes zero. The peak value can be obtained by inputting the time corresponding to this point into the spline function.

FIG. 19 is a graph connecting peaks extracted using this method.
The above procedure is shown by a formula. A section [t _n , t _{n + 1} ] satisfying b _n · b _{n + 1} ≦ 0 is obtained from the linear coefficients b _n and b _{n + 1} of the spline functions of adjacent sections.
That is, a point where the derivative is zero, that is, a vibration peak exists in this section. Therefore, this section is further subdivided. That is, the section [t _n , t _{n + 1} ] is divided into N at equal intervals.

Then, if the points in between are [t _n , t _n + Δt / N, t _n + 2Δt / N ... t _{n + 1} ] = [t ⁰ _n , t ¹ _n , t ^m _n , ... t ^N _n ], By substituting this subdivided time, a differential coefficient is obtained, and a subdivided point whose value is closest to 0, that is, the smallest absolute value is obtained.
That is, assuming that m is changed from 0 to N, m = m _min is determined so that | dS _n (t ^m _n ) / dt | Then, since t _n (m _min ) = t _n + m _min Δt / N is a peak position, S _n (t _n (m _min )) into which this time is substituted becomes a peak value.

The peak is either the maximum value (mountain) or the minimum value (valley). For example, if this is a peak, searching for the next peak with the same method will turn it into a valley, and then the peak. Therefore, if the time at which this peak appears is t _k (m _min ), the time difference t _k (m _min ) −t _n (m _min ) becomes the oscillation period. Further, since the difference between the values of the peak and the valley becomes the amplitude, the ratio of this amplitude and the amplitude of the previous cycle becomes the reduction ratio.

  Thus, the peak detection is repeated until a peak is detected, and when it is detected, the next peak is detected with the point as the origin. This procedure is repeated below. The first origin is the time when the variation monitoring mode is switched to the peak detection mode.

The operation flow of the peak detection method will be described based on FIG.
When the peak detection method is activated by the mode switching (S51), the mode switching time point is set as the origin (S52), and the vibration period T derived by the variation monitoring method before the switching (S53) is T /, which is a half thereof. There is a high possibility that the next peak exists within 2 (S54). The time-series data is taken in one by one after passing through the low-pass filter (S55, S56), and fine fluctuation components having a frequency larger than the target nuclear thermal hydraulic stability frequency (reciprocal of the vibration period) are removed. .

Here, during the half period (with some margin e added) from the peak search origin, cubic spline interpolation is performed with the sampling period as the section width (S57). Then, the spline coefficient b _n of the second-order term is obtained, and a section where b _n · b _{n + 1} is equal to or less than zero, the product of b _n and the adjacent coefficient b _{n + 1} is searched (S58).

If the section is found, then the section is further subdivided by N, that is, the time further subdivided by Δt / N with respect to the sampling period Δt is obtained (S59). Then, a differential coefficient at this subdivided time is obtained, and a subdivided point whose absolute value is minimized is searched (S60).
The point t _n ^min becomes the peak position, and the peak value S _n (t _n ^min ) derived by substituting the point into the original spline function is stored (S61, S62).
If a peak point is found in this way, the next peak is similarly searched using the end point t _{n + 1} of the section including the peak point as a new origin (S61 → S52).

  Here, the amplitude, period, and reduction ratio related to the second stability index are estimated from the stored peak positions and peak values adjacent to each other (S63). The peak is searched alternately for peaks and valleys, so the amplitude is the peak value minus the valley value, but the period is the time difference between peaks between the peak and the next peak or between the peak and the next valley. Become. The reduction ratio is a value obtained by dividing the adjacent amplitude by the previous amplitude.

  Parameters monitored by the peak detection method are the amplitude or peak peak value, and the reduction ratio. That is, at this stage, the stability has already deteriorated to some extent, and the vibration component other than the noise component is dominant in the output response, and the peak is likely to be a response based on the output vibration. It is unlikely that this is a sudden peak.

The peak monitored here includes a plurality of average peaks in a group monitoring an average response, an average peak in a group monitoring a region vibration based on a higher order mode, and an unaveraged peak monitoring a local vibration. .
Since the behavior of these peaks varies depending on the type of output vibration, the criteria for starting the vibration suppression means are also different.

FIG. 21 shows the amplitude response, and FIG. 22 shows the peak (peak) response. Both are standardized by the signal value at the time when the monitoring mode is switched.
The graph a in FIG. 21 and the graph a in FIG. 22 are local vibration monitoring signals, and the graph b in FIG. 21 and the graph b in FIG. 22 are 10 average signals grouped for local vibration monitoring in the region b in FIG. 21 and FIG. 22 are responses respectively extracted from the average signal of the entire core.

In both FIG. 21 and FIG. 22, the peak response of the local vibration monitoring signal is the fastest, the peak of the region vibration monitoring signal follows with a slight delay, and the peak of the average signal follows with a large delay.
This is a natural result because the target output vibration is a region vibration. The judgment criteria for outputting a trip signal for starting the operation of suppressing the output vibration are different from the difference in the characteristics of the signals, the difference in the characteristics of the output vibration mode, and the difference in the impact on the whole plant.

That is, the reference value for outputting the trip signal is set to be small for an average signal having the least sensitivity and a large impact on the entire plant, and to be set large for a signal for monitoring local vibration.
For example, if the amplitude of the local vibration signal is set to 30%, the amplitude of the region vibration signal is set to 20%, and the amplitude of the average signal is set to 10% as the reference value of the amplitude, the amplitude of the region vibration monitoring signal is 20% in FIG. A trip signal is output in the vicinity of 860 seconds when the value is reached.

  If the reference value of the peak value is set to half of the amplitude, the average signal reaches the reference value earliest and outputs a trip signal in the vicinity of 770 seconds in FIG. In this output vibration example, the average output slowly rises (there is a trend of low-frequency components), so that the trip signal is output earliest with the average value signal when judged by the peak value.

By the way, in order to avoid the output delay of the trip signal based on the spline interpolation process (FIG. 20: S57), the value of the signal (average signal of each group and local vibration monitoring individual signal) that has passed through the low-pass filter has a predetermined value. If it exceeds, the trip signal is output even if the trip detection is not made by the peak detection method.
This predetermined value is, for example, a 50% increase in the trip setting value of the peak value, or 15% of the trip setting value of the region vibration monitoring signal.
When the trip determination is made without going through the peak detection method in this way, the stability of the output has already deteriorated considerably (for example, the reduction ratio is 0.8 or more), which is a sudden peak. It is more likely that the output vibration has grown more rapidly than the possibility, which is effective as a backup function.

FIG. 23 shows a response example of the reduction ratio obtained from the peak amplitude of the region vibration monitoring signal.
In a state where the amplitude is small, an increase and decrease in amplitude occur alternately between adjacent peaks, and the instantaneous variation of the reduction ratio is large, which is not appropriate as a monitoring signal. However, the reduction ratio is effective as an index for determining whether or not the vibration is growing rapidly.
A method using an average value of two reduction ratios is effective for a response in which the amplitude alternately increases and decreases. In order to avoid an abnormal value of the reduction ratio at a small amplitude, it is preferable to enable this reduction ratio monitoring when a value smaller than the trip set value by the amplitude, for example, 1/4 of the trip set value is exceeded.

In this case, since it becomes effective when the amplitude exceeds 5% in FIG. 21, the trip by the reduction ratio becomes effective from around 780 seconds, and the trip becomes effective by the abnormally large reduction ratio in FIG. Is avoided.
Here, if the trip judgment value of the reduction ratio from the single amplitude ratio is set to 1.3 and the trip judgment value of the reduction ratio of the average value of the two times is set to 1.2, the trip will occur in the vicinity of 800 seconds. validate. In addition, it is necessary to set a value having a sufficient margin as a criterion for determining the amplitude for enabling the trip based on the reduction ratio to be several percent that is a steady noise level.

Even if none of the trip criteria is reached in the output vibration monitoring by the above peak detection, when this variation converges to a small value in the vibration cycle variation monitoring parallel to the peak detection method It can be determined that nuclear thermal hydraulic vibration is generated with high probability.
In this case, although the peak detection is prioritized, a trip signal is generated based on the vibration period variation monitoring. Thereby, it is possible to suppress output vibration with higher reliability without impairing the fuel soundness or the soundness of the plant.

The standard deviation indicating the variation of the vibration period also depends on the number of signals used for calculation. In general, as the number of signals increases, the value of the standard deviation decreases, so it is necessary to set the set points individually according to the number of signals in the group.
For example, in the case of FIG. 8, since it is an average value signal group, the number of signals is large, and after about 700 seconds, the standard deviation is greatly reduced and converges to a substantially stable value. On the other hand, in the region vibration monitoring group, the number of target nuclear instrumentation signals S is one fifth of the average value group, and the value is proportional to the square root of the reciprocal of the number of signals according to the definition of standard deviation. Therefore, the size is about √5 to 2.24 times.

  FIG. 24 is an explanatory diagram of a method for determining whether or not to output a trip signal based on the standard deviation of the vibration period. Here, the line q that defines the reset condition is set to ½ of the determination criterion for mode switching, and the line p that defines the trip condition is set to ¼ of the determination criterion for mode switching. Thus, as long as the reset condition is set to be looser than the trip condition, the values of the lines that define the reset condition and the trip condition are not particularly limited.

  Time measurement is started from the time when the signal falls below the line p, and a trip signal is output when a predetermined duration K has passed. Even if the signal falls below the line p, the time measurement is reset if the line q is exceeded before the duration time u (for example, 10 times the vibration period) elapses. This is because, in general, the standard deviation increases as the value decreases.

  In FIG. 24, the line p, which is a ¼ criterion, is first reached at around 746 seconds, but it is immediately exceeded. However, the time measurement of the duration time u is maintained, and the trip signal is finally output because the standard of 30 seconds, which is 10 times the vibration period, is continuously satisfied.

  In the core shown in FIG. 25A, an averaged signal in the region a, an averaged signal in the upper region b of the diagonal d, and an averaged signal in the lower region c of the diagonal d are observed. FIG. 25B shows a phase difference graph of three sets of combinations in these three regions.

Between the average signals grouped for monitoring by the diagonal d, the phase difference of the vibration is calculated using a statistical method. The statistical method is to calculate the phase difference from the phase of the spectrum in the latter case from the delay time that maximizes the function value if the former is obtained if the cross-correlation function or its cross spectrum is expressed. it can.
This makes it possible to accurately grasp the output vibration mode, perform optimum monitoring for the vibration mode, and improve the accuracy of the determination described above.

Here, the graph bc in FIG. 25B shows the phase difference between the two region vibration monitoring signals in the region b and the region c, and the phase difference has reached 180 degrees after 600 seconds. It can be seen that this combination is a region vibration that vibrates in opposite phases.
The graph b-a shows the phase difference between the two region vibration monitoring signals in the region b and the region a, and the graph ca shows the phase difference between the two region vibration monitoring signals in the region c and the region a.
The graph ba and the graph ca are not clearly opposite in phase, but since the phase difference in the graph ba is larger, the center line of the vibration that becomes the node of the region vibration is It is assumed that the center line d of the group division is shifted. In the case of core-integrated vibration, almost all of these phase differences do not appear.

As shown in FIG. 25B, when the region vibration is already observed in the vicinity of 600 seconds and then the mode shifts to the peak detection mode, the regions b and c in FIG. 25A are representative. It is possible to preferentially monitor the average signal.
For example, as a trip condition for the peak detection method, if it is not desirable to start the vibration suppression operation when only one of the group average signals is tripped, two or more trip signals are generated. If you want to trigger the operation.

That is, when the average signal of both the region b and the region c trips, it is only necessary to start the output vibration suppression operation assuming that the region vibration has developed. Alternatively, by relaxing the trip conditions of the region b and the region c as compared with the trip conditions of other signals, it is possible to surely trip these two signals together to suppress the region vibration.
If the vibration mode information is obtained by a statistical method before the peak detection method is activated in this way, a reliable trip condition that minimizes malfunction and operation delay based on that information. Can be optimized.

Here, an example of application to the optimization of the output vibration suppression operation control after the stability deterioration determination is shown, but this can be applied to the stability deterioration determination in the same manner.
That is, when monitoring the convergence state of the vibration cycle variation, if information on the vibration mode is obtained, the average signal for monitoring the region vibration is preferentially monitored based on the information. And if the standard deviation of the vibration period in the group of the area | region b and the area | region c falls similarly, it will be judged that area | region stability deteriorated. In this case, the determination criteria for deterioration in stability are relaxed from normal, and it is determined that the stability is deteriorated when both satisfy the relaxed determination criteria.
And the peak detection mode about the group signal of the area | region is started. Thereby, at this stage, the trip condition in the peak detection method of the region b and the region c is slightly relaxed as compared with other signals.

FIG. 26 shows an operation flow of the output monitoring apparatus.
A predicted state of stability deterioration is predicted from the plant information (S71), and the nuclear instrumentation signals are grouped from the output distribution and higher-order mode distribution derived from the result (S73).
Also, digitized nuclear instrumentation signal S is sequentially input as plant information, and appropriate filtering processing is performed on the signal (S72), so that fluctuation components in a frequency band different from nuclear thermal hydraulic output vibration are removed. .

  Grouping is executed based on the signals based on the plant information (S73), statistical processing is performed on the individual signals and the averaged signals (S74), and the standard deviation of the vibration period within each group is output (S75). . Further, the phase difference between the group signals is output (S81), and the vibration mode is determined based on the phase difference (S82).

This vibration mode is first referred to when determining instability from the standard deviation of the vibration period (S76). That is, the optimum group signal for determining instability is selected based on the characteristics of the vibration mode (core integration, region or locality, if the region is the center line of vibration, and if it is local, the position is local).
If instability has already been determined and the peak detection method has been activated (S78), the stability is also determined (S76). Here, if it determines with it being stable, the process (S75) of a standard deviation calculation of a vibration period from a statistical process (S74) will be repeated.

  If instability of the standard deviation of the vibration cycle is determined, an alarm is first sounded (S77), and the peak detection method is activated (S78). However, even if the peak detection method is activated, the calculation process of the standard deviation of the vibration period is continued in parallel. Conversely, if stability is determined in a state where the peak detection method is activated (S76), the peak detection method temporarily stops the processing.

After switching to the peak detection method, the value of the peak (mountain), the amplitude, and the reduction ratio are output from the ratio of the amplitude as monitoring parameters (S79). Based on these monitoring parameters, a trip signal is determined (S80A, S80B, S80C). Further, the filtered signal data that does not pass the peak detection method is also included in the monitoring parameter, and the trip signal is determined (S80D).
These monitoring parameters refer to group characteristics and vibration modes (S82) to adjust the trip determination criteria and perform trip determination (S80). Note that the reduction ratio is added to the monitoring parameter only when the amplitude value exceeds the reference.

When a trip determination (S80A, S80B, S80C) is made according to the trip determination criterion for each parameter based on the peak detection method, a trip signal is input to the logic gate (S83).
This logic gate is composed of an OR gate if the vibration suppression operation is activated by any one of the trip determinations (S80A, S80B, S80C), or vibration suppression if a plurality of trip determinations are not made. If the operation is not activated, an AND gate is used.

  A trip signal for trip determination (S80D) that is directly performed from the standard deviation of the vibration period is input to the OR gate (S84). In addition, the result of trip determination (S80A, S80B, S80C) based on the peak detection method is input to this OR gate. In this OR logic, if any one trip signal is generated, the vibration suppression operation is automatically started (S85). Here, the vibration suppression operation refers to an output reduction operation by inserting a control rod.

  As described above, according to the present invention, the output vibration caused by nuclear thermal hydraulic instability in the nuclear reactor can be detected with high reliability, and the output vibration can be detected without significantly affecting the soundness of the fuel or the plant. This will contribute to safe and efficient operation of the reactor.

DESCRIPTION OF SYMBOLS 10 ... Reactor, 11 ... Pressure vessel, 13 ... Gas-water separator, 14 ... Upper lattice plate, 15 ... Shroud, 16 ... Core, 17 ... Core support plate, 18 ... Recirculation pump, 21 ... Main piping, 22 ... Turbine, 23 ... Generator, 24 ... Condenser, 25 ... Pump, 26 ... Feed water piping, 30 ... Output monitoring device, 31 (31a, 31b, 31c, 31d) ... Nuclear instrumentation detector, 32 ... Control rod, 33 DESCRIPTION OF SYMBOLS ... Fuel assembly, 34 ... Instrumentation pipe, 35 ... Output suppression device, 41 ... Grouping part, 42 ... First calculation part, 43 ... Reference value storage part, 44 ... First determination part, 45 ... Peak detection part, 46 ... 2nd calculation part, 47 ... 2nd determination part, 48 ... 3rd determination part, 50 ... Statistical processing part, 60 ... State prediction part, 61 ... Process calculator, 62 ... Database, 63 ... Nuclear thermal hydraulic simulator , 70 ... variation analysis unit, S ... nuclear instrumentation signals, X _t ... time series Data.

Claims (20)

A first calculation unit that calculates a first stability index from time-series data indicating output oscillations of nuclear instrumentation signals output from a plurality of nuclear instrumentation detectors that detect neutrons in the core;
A first determination unit that compares the first stability index with a first reference value to determine whether the nuclear thermal hydraulic stability of the core is stable or deteriorated;
A second calculation unit that calculates a second stability index of the core based on the time-series data when the first determination unit determines the deterioration;
A second determination unit that determines whether to perform the operation of suppressing the output vibration by comparing the second stability index and a second reference value ;
Corresponding to the determination in the first determination unit and the second determination unit, as the suppression operation of the output vibration, any one of alarm, vibration information and an automatic activation signal of the vibration suppression device is generated in stages.
Even in the condition that does not reach the criteria for generating any of the alarm, vibration information and automatic activation signal of the vibration suppression device,
If another standard set more strictly than the standard by processing the time-series signal that has passed through the low-pass filter,
It said alarm, the output monitoring device of the reactor, characterized in Rukoto stepwise generate one of automatic start signal of the vibration information and the vibration suppression apparatus.   A first calculation unit that calculates a first stability index from time-series data indicating output oscillations of nuclear instrumentation signals output from a plurality of nuclear instrumentation detectors that detect neutrons in the core;
  A first determination unit that compares the first stability index with a first reference value to determine whether the nuclear thermal hydraulic stability of the core is stable or deteriorated;
  A second calculation unit that calculates a second stability index of the core based on the time-series data when the first determination unit determines the deterioration;
  A second determination unit that determines whether to perform the operation of suppressing the output vibration by comparing the second stability index and a second reference value;
  Even if it is determined by the second stability index that the output vibration suppression operation is unnecessary, when it is determined that the first stability index satisfies the third reference value that is set more strictly than the first reference value. The reactor power monitoring apparatus, wherein the operation for suppressing the output vibration is executed. In the reactor power monitoring apparatus according to claim 1 or 2 ,
A grouping unit for dividing the plurality of nuclear instrumentation detectors into groups;
The grouping unit focuses on the estimated power distribution of the core, the spatial higher-order mode distribution of neutron flux, or the region centered on the nuclear instrumentation detector close to the identified fuel assembly. A reactor power monitoring device characterized by In the reactor power monitoring apparatus according to claim 3 ,
The nuclear instrumentation signal processed by the first determination unit and the second determination unit is:
A reactor power monitoring apparatus, wherein the reactor power monitoring apparatus is an individual signal of the nuclear instrumentation detector selected from the group or an average signal of a plurality of nuclear instrumentation detectors in the group unit. In the reactor power monitoring device according to claim 3 or claim 4 ,
The estimation of the power distribution or the spatial higher-order mode distribution of the neutron flux is executed based on a physical model or a database. In the reactor power monitoring apparatus according to any one of claims 1 to 5 ,
The reactor power monitoring apparatus according to claim 1, wherein the first stability index is an index indicating a variation in vibration period of the plurality of time series data. In the reactor power monitoring apparatus according to any one of claims 1 to 6 ,
The reactor power monitoring apparatus characterized in that the second stability index uses an amplitude or a reduction ratio of a plurality of the time-series data. In the reactor power monitoring apparatus according to claim 6 or 7 ,
The reactor power monitoring apparatus characterized in that the vibration period is obtained by applying a statistical method such as an autoregressive analysis method, an autocorrelation function method, or a spectrum analysis method to the time series data. In the reactor power monitoring apparatus according to claim 7 or 8 ,
The amplitude or reduction ratio is obtained by a peak detection method that fits the time-series data set in a section with a polynomial and searches for a point where the differential value of the polynomial is zero,
The point at which the differential value becomes zero is first searched within a period with a margin of half of the vibration period with the switching point from the first determination unit as the origin, and then the vibration is performed with the searched point as the origin. A reactor power monitoring apparatus, characterized in that the search process is repeated within a period in which a half of the cycle has a margin. In the reactor power monitoring device according to any one of claims 6 to 9 ,
The index indicating the variation of the vibration period is calculated by excluding time series data in which the vibration period varies beyond a certain range among all the time series data included in the group. Furnace output monitoring device. In the reactor power monitoring device according to any one of claims 6 to 10 ,
As the first reference value, the sampling period obtained by multiplying the digitized sampling period of the time series data or a coefficient considering the reduction ratio is used,
The first determination unit performs the determination based on a duration indicating deterioration of the nuclear thermal hydraulic stability in the comparison between the first stability index and the first reference value. Output monitoring device. In the reactor power monitoring apparatus according to any one of claims 9 to 11 ,
Using a spline function as the polynomial,
The point where the differential value becomes zero is obtained as a section where the product of the differential coefficients of the data points at both ends thereof is a minus sign among the sections sandwiched between a plurality of data points constituting the time series data. A reactor power monitoring apparatus characterized in that a point at which an absolute value of a differential coefficient at a further divided division point is minimum is searched. In the reactor power monitoring device according to any one of claims 1 to 12 ,
When the second calculation unit is operating, the first calculation unit is operated in parallel,
When the second determination unit does not determine to execute the power control of the core within a predetermined time, the operation of the second calculation unit is stopped and the operation of the first calculation unit is continued. Reactor power monitoring device characterized by In the reactor power monitoring apparatus according to any one of claims 1 to 13 ,
The spatial distribution of the phase difference of the output vibration is obtained based on the phase difference of the time series data from the nuclear instrumentation detector located in the local unstable central region and the peripheral region thereof, and the local vibration, the region vibration, or the core integration is obtained. A reactor power monitoring device characterized by estimating an output vibration mode such as vibration. The reactor power monitoring apparatus according to any one of claims 3 to 14 ,
An output vibration mode is estimated by calculating a phase difference of an average signal in units of the group, and the first reference value and the second reference value are set to be optimum values according to the estimated output vibration mode. Reactor output monitoring device characterized by changing In the reactor power monitoring device according to claim 1 ,
Reactor characterized in that, as a reference for stepwise generation of any one of the alarm, vibration information, and vibration suppression device automatic start signal, the amplitude exceeds a steady noise level. Output monitoring device. In the reactor power monitoring apparatus according to claim 2 ,
The reactor power monitoring apparatus according to claim 1, wherein a duration time determined as the deterioration is measured by the first determination unit, and the operation for suppressing the output vibration is executed when a predetermined criterion is satisfied. In the reactor power monitoring device according to claim 2 or 17 ,
When the determination criterion is satisfied in the peak detection method, a signal that does not pass the peak detection method satisfies the determination criterion, or a condition that satisfies any of the cases where the variation of the vibration period satisfies the determination criterion,
Or, when the peak value, amplitude, and reduction ratio each meet the criteria, either peak value and amplitude, peak value and reduction ratio, or a combination of amplitude and reduction ratio is achieved An output monitoring apparatus for a nuclear reactor, wherein the operation for suppressing the output vibration is executed when a condition for satisfying is satisfied. Outputting nuclear instrumentation signals from a plurality of nuclear instrumentation detectors that detect neutrons in the core;
Calculating a first stability index from time-series data indicating the output vibration of the nuclear instrumentation signal;
Comparing the first stability index with a first reference value to determine whether the nuclear thermal hydraulic stability of the core is stable or deteriorated;
Calculating the second stability index of the core based on the time-series data when it is determined that the nuclear thermal hydraulic stability is deteriorated;
Determining whether to perform the operation of suppressing the output vibration by comparing the second stability index with a second reference value;
Corresponding to the determination that the nuclear thermal hydraulic stability is deteriorated and the determination that the output vibration suppression operation is performed, as the output vibration suppression operation, alarm, vibration information, and automatic vibration suppression device Generating one of the activation signals in stages;
Even under conditions where none of the alarm, vibration information, and automatic start signal of the vibration suppression device has reached the standard, the time series signal that has passed through the low-pass filter is processed so that it is set more strictly than the standard. A step of generating any one of the alarm, the vibration information and the automatic activation signal of the vibration suppression device in a stepwise manner when the other standard is reached. .   Outputting nuclear instrumentation signals from a plurality of nuclear instrumentation detectors that detect neutrons in the core;
  Calculating a first stability index from time-series data indicating the output vibration of the nuclear instrumentation signal;
  Comparing the first stability index with a first reference value to determine whether the nuclear thermal hydraulic stability of the core is stable or deteriorated;
  Calculating the second stability index of the core based on the time-series data when it is determined that the nuclear thermal hydraulic stability is deteriorated;
  Determining whether to perform the operation of suppressing the output vibration by comparing the second stability index with a second reference value;
  Even if it is determined by the second stability index that the output vibration suppression operation is unnecessary, when it is determined that the first stability index satisfies the third reference value that is set more strictly than the first reference value. And a step of executing the operation for suppressing the output vibration.

Images