JPS5831606B2 - Nuclear reactor abnormality detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is a method of detecting a fuel assembly from an alternating current component (hereinafter referred to as a temperature fluctuation signal) of an output signal of a temperature detector (for example, a thermocouple) installed at the fuel assembly coolant outlet of a nuclear reactor. θ is also related to a nuclear reactor abnormality detection device for detecting abnormalities occurring in the body.

In liquid metal cooled fast breeder reactors, if a fuel assembly blockage accident or a large-scale release accident of gaseous fission products from the fuel rods occurs, the cooling capacity decreases and the coolant boils →
There is a risk that an abnormality will escalate into an accident, such as a fuel melting accident and propagation of the accident to adjacent fuel assemblies.

When these abnormalities occur, it is extremely important for the safety of fast reactors to detect the abnormality at an early stage and take prompt action.

Conventionally devised reactor abnormality detection methods using temperature fluctuation signals include methods that compare the RMS value of the temperature fluctuation signal with an alarm limit value and issue an alarm when the alarm limit value is exceeded, and methods that detect the temperature at the entrance and exit of fuel assemblies. There was a method of determining a cross-correlation function of fluctuation signals and detecting an accident within a fuel assembly from changes in the shape of the cross-correlation function.

However, with the method that focuses on the RMS value of the temperature fluctuation signal, since the change in the RMS value is small, it is difficult to detect local blockage accidents within the fuel assembly. This method has drawbacks, such as the difficulty of installing fuel assembly inlet instrumentation and the ability to detect only relatively large phenomena of flow reduction.

The present invention identifies the waveform of a temperature fluctuation signal using an autoregressive model, calculates the autocorrelation function of the error signal between the temperature fluctuation signal to be monitored, and the temperature fluctuation signal predicted from the autoregressive model, and calculates the magnitude of this autocorrelation function. By determining whether or not the value is within a predetermined value, the technical drawbacks of the above-mentioned conventional devices are eliminated, and a more practical nuclear reactor abnormality detection device is provided.

An embodiment of the present invention will be described in detail below.

FIG. 1 is a block diagram showing the configuration of a nuclear reactor abnormality detection device according to one embodiment of the present invention, in which 1 is a fuel assembly, 2 is a
is a fuel pin, 3 is a coolant flow, 4 is a temperature detector, 5 is an AC amplifier, 6 is a low-pass filter, 7 is an autoregressive model of the temperature fluctuation signal, and using the identified autoregressive model calculate the predicted value of the temperature fluctuation signal, calculate the error signal between the temperature fluctuation signal to be monitored and the predicted value, find the autocorrelation function of the error signal, and compare the value of the autocorrelation function with the alarm limit value. This is an abnormality detection arithmetic processing device for detecting an abnormality occurring within a fuel assembly.

FIG. 2 is an explanatory diagram showing the autocorrelation function of error signals during normal and abnormal times.

The feature of the present invention lies in the configuration of the abnormality detection arithmetic processing device 7, and the principle of the system will be described below.

N time-series signals y1 (i=1.2...-...,N) obtained by sampling the temperature fluctuation signal y when the fuel assembly is in a normal state at equal intervals along the time axis. From, (a)
Create an autoregressive prediction model for the temperature fluctuation signal expressed by the formula.

8 8 In the formula (a), 0Ja1 s a2+..., aM are autoregressive model coefficients determined so that the formula 09, including the order M, is minimized.

Here, εl is the time series signal yk of the temperature fluctuation signal, and (
b) Error signal (
(hereinafter referred to as error signal), that is, the order M of equation (a)
is determined by the extreme value search method, but the calculation can be carried out efficiently if the actual data is analyzed in advance and a guideline for the optimal order is established.

Ham When 8M is given, in formula (A) 0) al, a2,...
..., aM is determined by the least squares method, and the calculation method is shown in formulas (U) to (G).

In the above, 0J in step 3 and 0J in step 5
is an adjustment parameter of this device, which is determined in advance using actual data.

The suffix J indicates data to be collected or processed during reactor abnormality monitoring, and the suffix m in formula (a) is the coefficient a1 r a21..., aM in formula (a) AA A This indicates that the data is used to determine the

Figure 2 is an example showing the autocorrelation function of the error signal when the fuel assembly is normal and when the fuel assembly is abnormal.Curve a is the autocorrelation function of the error signal when the fuel assembly is normal; The error signal Φ autocorrelation function when an abnormality occurs in the fuel assembly, C is the alarm level of the RMS value of the error signal, and d is the alarm level of the error signal 0 autocorrelation function.

When the fuel assembly is normal, ψεε(j) has k of 1 to 2.
Above that, it rapidly attenuates and takes a low value.

This property is also guaranteed theoretically by determining the autoregressive model of the temperature fluctuation signal so that the equation (/→ is minimized).

However, when an abnormality occurs in the fuel assembly and the dynamic characteristics of the temperature fluctuation signal change, ψεε(1
) vibrates greatly and is not damped even if k is made large.

The present invention focuses on this property, and ψ and ε(j)
(3<k(J)0) By monitoring the amplitude, an attempt is made to detect a fuel assembly internal accident.

In addition, the magnitude of ψεε('') contains information about accidents that are difficult to show up in differences in the dynamic characteristics of the temperature fluctuation signal, so by monitoring the value of ψ5ε(J) at the same time, the abnormality detection function can be improved. You can expect fulfillment.

FIG. 3 is a block diagram showing one embodiment of the abnormality detection processing device 7(1) which is the main part of the nuclear reactor abnormality detection device according to the present invention.

In FIG. 3, 8 is an analog-to-digital converter for quantizing the temperature signal, 9 is an arithmetic unit A (for example, a microprocessor) for calculating the equation (A), and 10 is (for calculating the Ha equation). 0 arithmetic unit B (for example, a microprocessor); 11 a timer for generating a periodic pulse signal to cause arithmetic unit B to execute at a fixed period; and 12)
13 is a correlator for calculating the formula (f), 14 is a 0 setter A115 for setting the alarm limit Ld2, and a comparator A116 for comparing Ld2 and ψε~). is the alarm limit Ld2 and ψεε(1)
Since the comparison result ψSε(1) issues an alarm when it is larger than Ld2, the U-fu alarm A117 has an alarm limit Ld1.
The setting device B118 for setting ψ2εl) is -Ld
Comparator B for comparing between 1 and Ld, 0)
119 is an alarm B for issuing an alarm when ψεε('') exceeds the range between -Ld1 and Ldl.

Next, the operation will be explained with reference to FIG.

A temperature fluctuation signal y(1) is extracted by passing a temperature signal detected by a temperature detector 4 installed at the outlet of the fuel assembly through an AC amplifier 5 and a low-pass filter 6.

The analog-to-digital converter 8 is used to quantize y(t), and the arithmetic unit A9 executes the calculation of equation (a).

(b) Parameters in the formula al l a2 j...
, aM are calculated using the calculation unit BIO and updated at regular intervals.

The execution cycle is set by the timer 11. Subtractor 12
are used to calculate the actually observed temperature fluctuation signal y, the predicted value y of the temperature fluctuation signal calculated by the calculator A9, and the stop error signal ε.

Using the correlator 13, the error signal ε, 0) self-interval function p:
Calculate E(1).

Using the comparator A15, ψεε(j) is compared with the alarm level Ld set by the setting device A14, and ψ:E(j)>L
At the time of d2, the alarm device A16 generates an alarm.

Using the comparator B18, check the condition shown by the following formula (IJ), and if the condition (S)2〇 is not satisfied,
The alarm device 819 generates an alarm.

As described in detail above, the present invention calculates the autocorrelation function of the error between the estimated value of the autoregressive model and the actual measured value, and determines whether the magnitude of this autocorrelation function is within a predetermined value. By using the reactor abnormality detection device according to the present invention at Q set to
Since it is possible to simultaneously detect changes in the behavior of the temperature fluctuation signal that do not appear in changes in the S value, it is possible to detect abnormalities that occur within the fuel assembly and to detect phenomena in the early stages of an accident, thereby improving reactor Φ safety. can be improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a nuclear reactor abnormality detection device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram showing the autocorrelation function of error signals when the fuel assembly is normal and when the fuel assembly is abnormal, and FIG. FIG. 7 is a block diagram showing an embodiment of an abnormality detection arithmetic processing device (70). In the figure, 4 is a temperature detector, 5 is an AC amplifier, 6 is a low-pass filter, 7 is an abnormality detection processing unit, and 8 is an analog
Digital converter, 9 is arithmetic unit A110 is arithmetic unit B11
1 is a timer, 12 is a subtracter, 13 is a correlator, 14 is a setter, 15 is a comparator A116, an alarm A117, a setter B118, a comparator B119 is an alarm B.

Claims (1)

[Scope of Claims] 1. A temperature detector installed at the outlet of a coolant that cools a fuel assembly, means for extracting an alternating current component from a temperature signal obtained from the temperature detector, and a temperature signal extracted by the means. Identify an autoregressive model for the waveform of the AC component of A nuclear reactor abnormality detection device comprising an abnormality detection calculation processing device that detects an abnormality occurring within a fuel assembly by determining whether the size is within a predetermined value. 2. Nuclear reactor abnormality detection according to claim 1, characterized in that the abnormality detection arithmetic processing device is configured to issue an alarm when there is a relationship between the arithmetic unit A that calculates Device.