JP3268097B2 - Reactor stability monitoring device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a reactor stability monitoring apparatus used for monitoring core power stability of a boiling water reactor.

[0002]

2. Description of the Related Art A boiling water reactor (BWR) includes a boiling channel composed of a gas-liquid two-phase flow in a core portion thereof, and the boiling channel has an instability mode caused by a density difference, that is, a density difference oscillation. There is an inherent flow oscillation phenomenon called. Although this is called channel instability, in the actual BWR core, neutron dynamics are added to this, and this appears as a phenomenon in which the neutron flux level in the core, that is, the core output oscillates, and this is called core instability. I have.

[0003] Generally, core instability is an unstable event that oscillates integrally in the core according to a neutron flux fundamental mode that originally exists most stably, and is a thermo-hydraulic instability. In the channel instability, a spatially inhomogeneous higher-order mode of neutron flux is excited in the reactor core and a phenomenon of oscillating at different phases in space is also observed. This phenomenon is called domain instability.

[0004] As an index indicating the stability of the core, a reduction ratio which is a ratio of adjacent amplitudes of the response is used. When the width reduction ratio is less than 1, it means that the vibration is attenuated, so that it is stable. Conversely, when the width reduction ratio exceeds 1, it indicates that the vibration grows and is unstable. Reduction ratio is 1
In the case of, the amplitude is constant and the vibration is continued.

However, in practice, due to the non-linearity of the system, even if the reduction ratio exceeds 1, the amplitude growth stops at a certain point, and the system shifts to a limit cycle vibration, which is a vibration having a constant amplitude. Therefore, limit cycle oscillation is a completely different oscillation phenomenon from harmonic oscillation in a linear system.Since the actual system has nonlinearity, the definition of the attenuation ratio is not the same as the attenuation ratio in the sense of linearity. May be different.

[0006]

Under such circumstances, when monitoring the stability of a plant on-site, at present, the reduction ratio is determined by a statistical method assuming linear stationary and normality of the time-series data. And monitor the transition. Therefore, in the event of a transient that deviates from stationarity,
Alternatively, it is difficult to accurately estimate the width reduction ratio in a region near the stability limit where the nonlinearity becomes remarkable.

[0007] When the unsteady state and the nonlinearity become remarkable, it is difficult to accurately estimate the core stability. In particular, it is necessary to monitor the stability margin near the oscillation where the nonlinearity becomes remarkable and the state of development of the vibration. Conventionally, it is difficult.

The present invention has been made in view of the above circumstances, and has as its object to provide a reactor stability monitoring apparatus capable of detecting a core instability event and improving safety and operation rate. Aim.

[0009]

According to a first aspect of the present invention, there is provided a reactor stability monitoring apparatus for detecting a neutron flux in a reactor core. And sample this neutron flux detection system signal,
A correlation dimension and a Lyapunov exponent for calculating a correlation dimension and a Lyapunov exponent from the sampled time-series data, and calculating a rate of change from the values of the correlation dimension and the Lyapunov exponent. A comparison arithmetic unit that compares a database storing correlation dimensions and Lyapunov exponents corresponding to core stability evaluated in advance, and a stabilization measure determination unit that takes necessary stability improvement measures based on the comparison calculation result. It is provided.

According to a second aspect, the correlation dimension / Lyapunov exponent calculator according to the first aspect sequentially calculates the transition of the correlation dimension and monitors a process in which the value gradually approaches 1.0. Features.

[0011]

According to the first aspect of the present invention having the above structure, the correlation dimension and the Lyapunov exponent are calculated from the sampled time-series data, and the rate of change is obtained along with the values.
By comparing them with reference values stored in a database in advance, the stability of the core can be accurately monitored.

According to a second aspect of the present invention, the development of the vibration state to the limit cycle is monitored by sequentially calculating the transition of the correlation dimension and monitoring the process in which the value gradually approaches 1.0. It becomes possible.

[0013]

Embodiments of the present invention will be described below with reference to the drawings.

According to the present invention, in addition to the width reduction ratio previously required,
Correlation dimensions and Lyapunov exponents are used as indices indicating stability. Then, in describing the embodiment of the present invention, a correlation dimension and a Lyapunov exponent will be described.

That is, the correlation dimension is a kind of fractal dimension reflecting the fractal structure of the dynamical system of the system, and is widely used because the calculation is relatively easy. To obtain the correlation dimension, first, the correlation integral shown in the following equation is obtained.

[0016]

(Equation 1) Here, X _i is a time series data vector obtained by an operation of embedding based on the observed time series data. H is a Heavisite function. The correlation integral obtained in this way is given by the norm r between the time series data.

[Mathematical formula-see original document] When it is assumed that the relationship of C (r)-r [ ^mu] (2) is obtained, this index [mu] is called a correlation dimension. This correlation dimension means that when the measure (norm) used for observation is changed, the number of orbits of the dynamic system existing in a certain phase space is similar to the scale of the measure, that is, the spatial fractal This is a dimension that appears when the gender is extracted from the time-series data by the operation of embedding.

The embedding is an operation of rearranging data vectors at regular intervals from the time-series data at arbitrary intervals (referred to as embedding dimensions). The embedding operation is repeated until the correlation dimension converges on the embedding dimension.
This correlation dimension is closely related to the stability of the system,
It depends purely on the dynamical system that describes the system, irrespective of the non-stationarity and nonlinearity of the system.

That is, the relationship between the stability and the correlation dimension can be explained as follows. If the system is sufficiently stable, there is no special dominant degree of freedom in the time-series data, and noise becomes dominant. Since noise has a high degree of freedom, the correlation dimension tends to take a large value when it is stable.

On the other hand, when the system becomes unstable, it is considered that a specific trajectory in the degree of freedom becomes unstable, and the trajectory gradually suppresses noise components and becomes dominant. Therefore, the degree of freedom decreases as the state becomes unstable, and the value of the correlation dimension reflecting the degree of freedom also decreases. From such a relationship, it is considered that there is a close relationship between the stability of the system and the correlation dimension obtained from the time-series data as the output thereof.

The definition of the Lyapunov exponent is as follows. As shown in FIG. 2, one trajectory X (t) is considered, its initial value X (0) and another trajectory slightly separated from it are taken, and the distance between the two points is represented by d ₁ (0). ). Since the distance between two points rapidly increases with the development of time, the distance between the two points becomes d ₂ (0)
And follow the trajectory with that point as the initial value.

By repeating this operation every τ seconds, the magnification between the orbits can be obtained. Then, since the distance between the orbits is successively increased with each magnification, the magnification for nτ seconds is

(Equation 3) It turns out that it becomes. This expansion rate is exponential to time (development time).
Then the time evolution of the distance between two points is approximately

## EQU4 ## d (t) to exp (λt). This increase rate λ is called a Lyapunov exponent. To specifically evaluate the increase rate λ,

(Equation 5) Will be calculated.

Therefore, by monitoring the values of the correlation dimension and the Lyapunov exponent and their temporal transitions, it is possible to accurately estimate the stability of the system and its transitions by monitoring only the reduction ratio. It becomes possible.

FIG. 1 is a block diagram showing one embodiment of a reactor stability monitoring apparatus according to the present invention. As shown in FIG. 1, a large number of neutron detectors 2 are disposed in a core 1.
A local output area monitor (LPRM) signal 3 is extracted from each of these neutron detectors 2 as an analog signal. These LPRM signals 3 are collected by an adder 4 and averaged equally every about 20 signals to obtain an average output area monitor (A
PRM) signal 5. Normally, the width reduction ratio is calculated from the APRM signal 5 by statistical processing.

Each LPRM signal 3 and APRM signal 5
The analog signal is digitized using an A / D converter 6, and the digitized LPRM signal 3 is first converted into an appropriate pre-process by a pre-processor 7, that is, noise using a Hanning window or the like. The components are removed.

Since the noise component acts in the direction of increasing the correlation dimension, the removal of the noise component is important in the evaluation of the correlation dimension, but the relative relationship of the correlation dimension to the stability is as follows.
It does not change for the target stability. For this reason, if the processing method is unified including the database to be compared, there is no problem in the stability evaluation. Therefore, in this embodiment, the preprocessor 7 is a general low-pass filter.
However, when evaluating the Lyapunov exponent, no filter is used.

The correlation dimension / Lyapunov exponent calculator 8 calculates the correlation dimension using the correlation integral of equation (1) and the Lyapunov exponent using equation (5) from the time-series data after the preprocessing. N in equation (1) is the number of time-series data points, and it is necessary to have enough numbers to obtain these values with sufficient accuracy. Therefore, these values are determined in batches at certain time intervals. Although this time interval depends on the sampling interval, it is about several tens of seconds assuming that the same sampling interval as that of the conventional stability detector is used.

Therefore, all the time-series data sampled within this time interval is temporarily stored in the memory 9. As can be seen from the form of the equation (2), the correlation dimension is obtained as a slope of the logarithm of the correlation integral with respect to the logarithm of the norm.
The correlation dimension / Lyapunov exponent obtained here is stored in the memory 9, and the rate of change is calculated from the difference between the stored value and the value one batch before, which is also stored in the memory 9 at the same time. The database 11 stores a correlation dimension and a Lyapunov exponent for the previously evaluated stability in a table format.

Then, these databases, the sequentially calculated correlation dimension, Lyapunov exponent, and the rate of change thereof are compared in the comparator 10. If the stability exceeds the threshold through the decision unit 12, Judgment signal 1
3 is issued, a signal is sent to the stabilization measure determiner 14, and a stabilization measure signal 15 is issued. A control rod (selection control rod) preset as a stabilization measure is inserted by the stabilization measure signal 15, and the output is reduced, so that the stability is reduced to a threshold value or less.

As described above, in the present embodiment, the values of the correlation dimension and the Lyapunov exponent in the typical core state and their fluctuation rates are stored in a database in advance, and a determination criterion is set based on the database. Monitoring becomes possible.

That is, in the present embodiment, the observation signal is sampled, the correlation dimension and the Lyapunov exponent are obtained from the sampled time-series data, the values are stored to obtain the fluctuation rate, and these are compared with the database. The core stability is determined, and necessary stability improvement measures are taken based on the results.

Next, the operation of the present embodiment will be described.

First, a first monitoring example of monitoring the stability from the correlation dimension and the value of the Lyapunov exponent will be described. FIG. 3 shows the relationship between the correlation dimension / Lyapunov exponent serving as a database and the core stability. As the stability deteriorates (the attenuation ratio increases), the correlation dimension decreases, and the oscillation point (the attenuation ratio is 1 or more)
It can be seen that the decrease rate of the correlation dimension increases as approaching. Conversely, as the stability deteriorates, the Lyapunov exponent increases, indicating that the relationship with the stability has the same tendency as the reduction ratio.

However, the relationship between the correlation dimension and the Lyapunov exponent with respect to the reduction ratio is not linear in FIG.
Can be seen from That is, the correlation dimension has high sensitivity near the oscillation point. Also, it can be seen that the value of the Lyapunov exponent increases exponentially with the increase of the reduction ratio.
Such a feature is very effective in detecting stability deterioration with high accuracy.

Accordingly, the relationship between the core stability and the correlation dimension / Lyapunov exponent which is obtained in advance and stored in the database, the correlation dimension / Lyapunov exponent which sequentially reflects the current core state, and the values before them. By comparing the difference with the step value, the current stability,
Further, a change in stability in the future is predicted, and when they exceed a reference value, an alarm (alarm) or necessary stabilizing means is activated.

Next, the second characteristic in the vicinity of the oscillation point at which the characteristic of the index introduced in the present invention becomes prominent in the nonlinearity.
An example of monitoring will be described. The limit of the stability evaluation based on the reduction ratio in such a state will be described.

FIGS. 4A and 4B show two examples of output vibrations, and FIG. 5 shows the results of sequentially evaluating the reduction ratio from the vibration data. In the vibration example of FIG. 4A, the width reduction ratio has almost reached 1.0, indicating that the limit cycle vibration has been developed. On the other hand, in the vibration example of FIG. 4B, although a slight variation appears in the width reduction ratio,
It shows a value close to 1.0 in terms of the width reduction ratio, and it is understood that the state is close to the oscillation point.

That is, it is difficult to determine the difference between the two vibration states only by comparing the width reduction ratios. When this vibration example is evaluated by the correlation dimension, it can be seen that the two vibration states are clearly different as shown in FIG. That is, the vibration example in FIG. 4A is a developed limit cycle vibration, and the amplitude and the period are almost constant, whereas the vibration example in FIG. 4B is extremely unstable and has an oscillation limit. Although it is in a state close to the above, the limit cycle oscillation has not been reached and the amplitude is not constant.

Therefore, by sequentially evaluating the transition of the correlation dimension and monitoring that the value gradually approaches 1.0, it is possible to monitor the development process of the vibration state to the limit cycle vibration. Is possible. This is shown in FIG. These are all the results of obtaining a change in amplitude and a change in a correlation dimension corresponding to the change from a plurality of vibration states having different stable states. Although the amplitudes are different from each other, the correlation dimension is 1.
It gradually approaches a value close to zero.

On the other hand, there is a state where the correlation dimension only stays at around 1.5 at one point, and in this state, the amplitude is not constant, and the oscillation limit just like the vibration example in FIG. , But not yet in the limit cycle. Such a state is a linearly unstable state, but is still a non-linearly stable state, and is a state that cannot be determined by the evaluation using the reduction ratio.

Therefore, by constantly monitoring the difference between the value of the correlation dimension and 1.0 and its change, it is possible to evaluate the development process of the vibration state, and to thereby predict the future growth of the amplitude. Becomes possible. FIG. 8 shows the overall flow described above.

In the above embodiment, the correlation dimension and the Lyapunov exponent are calculated, the fluctuation rate (change rate) is calculated together with the calculated values of the correlation dimension and the Lyapunov exponent, and the value of the correlation dimension and the Lyapunov exponent and the fluctuation rate are calculated. The stability is monitored by comparing with a reference value previously stored in a database. In particular, monitoring the fluctuation rate utilizes the characteristic that the correlation dimension and the Lyapunov exponent have nonlinear sensitivity to the stability as shown in FIG. 3, and the sensitivity increases as the stability deteriorates. ing.

Here, the reference value is set in multiple stages,
When a lower reference value is exceeded, only an alarm is activated, and when a higher reference value is exceeded, the stabilizing means (selection control rod insertion) is automatically activated. The above is the flow based on the first monitoring example. On the other hand, in the second monitoring example, attention is paid to the value of the correlation dimension, and the difference between the value and 1.0 is monitored. If the difference exceeds the reference value, it is determined that a developed limit cycle vibration has occurred, an alarm is activated, and the stabilization measure is activated.

[0043]

As described above, according to the first aspect of the present invention,
According to the reactor stability monitoring device described in the above, the correlation dimension and the Lyapunov exponent are calculated from the sampled time-series data, the values of the correlation dimension and the Lyapunov exponent are also calculated, and the correlation dimension and the Lyapunov exponent are calculated. By comparing the index value and the fluctuation rate with reference values stored in the database in advance, it becomes possible to monitor the stability of the core more accurately and more accurately and more flexibly than in the past, and especially the Lyapunov exponent Since the value increases exponentially with the increase of the width reduction ratio, it is possible to detect stability deterioration with high accuracy and high sensitivity, and to improve safety and increase the operation rate. it can.

According to the second aspect of the present invention, the transition of the correlation dimension is sequentially calculated, and the process of the value gradually approaching 1.0 is monitored, thereby monitoring the process of developing the vibration state to the limit cycle. Thus, it is possible to effectively evaluate a vibration state close to the oscillation limit, which cannot be determined by the evaluation using the attenuation ratio, and it is possible to predict the growth of vibration and amplitude in the future.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing one embodiment of a reactor stability monitoring device according to the present invention.

FIG. 2 is an explanatory diagram showing a method for obtaining a Lyapunov exponent.

FIG. 3 is a diagram showing a relationship between a correlation dimension, a Lyapunov exponent, and stability.

FIGS. 4A and 4B are diagrams each showing an example of output vibration.

FIG. 5 is a diagram showing a comparison of the reduction ratios of the vibration examples of FIGS. 4 (A) and (B).

FIG. 6 is a diagram showing a change in a correlation dimension in the vibration examples of FIGS. 4 (A) and 4 (B).

FIG. 7 is a diagram showing a comparison between a change in amplitude and a change in correlation dimension.

FIG. 8 is a flowchart showing the entire processing flow.

[Explanation of symbols]

 Reference Signs List 1 core 2 neutron detector 3 local output area monitor (LPRM) signal 4 adder 5 average output area monitor (APRM) signal 6 A / D converter 7 preprocessor 8 correlation dimension / Lyapunov exponent calculator 9 memory 10 comparison operation Device 11 database 12 judgment device 13 judgment signal 14 stabilization measure judgment device 15 stabilization measure signal

Claims (2)

(57) [Claims] A neutron detector for detecting a neutron flux in a reactor core; a neutron flux detection system signal sampled; a correlation dimension and a Lyapunov exponent determined from the sampled time-series data; A correlation dimension / Lyapunov exponent calculator for obtaining a fluctuation rate from the value of the correlation dimension and the Lyapunov exponent value, a correlation dimension corresponding to the fluctuation rate and the core stability evaluated in advance.
A reactor stability monitoring device comprising: a comparison operation unit that compares a database storing Lyapunov exponents; and a stabilization measure determination unit that takes necessary stability improvement measures based on the comparison operation result. . 2. The correlation dimension / Lyapunov exponent calculator sequentially calculates a transition of a correlation dimension, and calculates a value of 1.0.
Monitoring the process of gradually approaching to
A reactor stability monitoring device as described in the above.