GB2129188A - A device and process to differentiate gross exposure of nuclear fuel from leaking fuel rods - Google Patents

Abstract

A nuclear instrumentation system for the detection of cladding failure in nuclear fuel rods which can differentiate gross fuel exposure from merely gas leaking fuel rods by monitoring different delayed neutron precursor groups as distinguished by different neutron energy levels.

Description

SPECIFICATION A device and process to differentiate gross exposure of nuclear fuel from leaking fuel rods This invention reiates generally to nuclear instrumentation and more specifically to instrumentation for monitoring the integrity of fuel rods of nuclear reactor fuel assemblies.

Nuclear reactors contain a fuel core which is a grouping of fuel assemblies each of which has a plurality of fuel pins. A fuel rod is a cylindrical tube which contains nuclear fuel pellets. The metal side of the tube separates the fuel pellets from reactor coolant which flows over the surface of the tube or cladding.

A penetration of the cladding termed a fuel failure, may allow fission fragments, particularly gases, to escape from the fuel pin, and may allow coolant/fuel contact.

Nuclear reactors are expected to experience fuel failure in spite of rigorous quality control and conservative operating procedures. Most of the failures result from pin-hole cracks in the cladding and/or end plug welds. Such failures are detected by analysis of fission-gas outside the core (e.g., in the reactor cover gas) and by observation of delayed neutron precursors in the reactor coolant.

The significance of fuel failures arises due to three factors as follows: (A) Safety: Although failed fuel which only leaks gas is of negligible safety significance, there is concern that the breach in the fuel could become large enough to allow ingress of coolant into the pin, or the escape of some fuel particles. In Liquid Metal Fast Breeder Reactors, sodium reaction with fuel material generates a product which is less dense than fuel, causing an increase in volume and sometimes an increase in the breach size. The safety concern from fuel swelling or fuel release occurs because of possible reduced heat transfer from the pin, or the remote possibility that the particles could accumulate to form a heat-generating flow blockage within the core.

(B) Plant Maintenance: Some of the fuel material entering the coolant could be transported away from the core and deposited in any part of the piping, heat exchangers or the pumps. Such deposits could complicate maintenance operations in and around the components.

(C) Economics: Due to the inevitability of fuel failures, the availability of the reactor is significantly improved by operating with a limited number of failed elements. Thus, if the safety and maintenance concerns can be resolved, there is considerable incentive to operate with failed fuel.

One of the problems faced by reactor instrumentation is to detect and monitor failed fuel in such a manner that safe operation of the reactor is not impaired. This problem can be solved by having the ability to observe changes in failed fuel so that significant increases in exposure of fuel to coolant are clearly and reliably annunciated. Relatively isolated and random cladding failure in nuclear fuel is innocuous. If fuel exposure and fuel release occurs, a possible mode of propagation of the failure may be postulated, although current experience indicates that cladding failures remain localized events.

Analyses indicate that small heat generating blockages can be tolerated without jeopardizing safety. If such a blockage can be reliably detected, reactor safety can be further assured by precluding growth of a blockage beyond tolerable limits.

Consequently, it is desired to provide a method for the detection of fuel failure, such method being capable of differentiating between gross failures and small leaks which result in only fission gas leakage.

A process is described which uses delayed-neutron energy spectrometry to observe changes in failed fuel. Development of criteria to preciude fuel failure propagation via formation of heat generating blockages would be a step closer if the methodology described can be applied in practice. Such a practical application must be demonstrated in an operating reactor and such a demonstration must necessarily precede development of quantitative criteria. The analysis described in this report confirms the theoretical feasibility of the approach. If the practical application measures up to expectation, information obtained from this approach would provide a significantly more direct indication of fuel exposure, and changes thereof, than conventional delayed neutron detector (DND) systems.

The basic principle of this technique is based on the fact that the energy spectra of delayedneutrons emitted from different precursor groups vary. The energy information is lost in conventional detectors because the neutrons are thermalized before detection. Numerous precursor isotopes contribute to each of the six delayed-neutron groups normally used in reactor calculations.

Measurements have shown that the energy spectrum of each group is significantly different from the others. Due to the transit time delay between the points of origin of the precursors and the point of detection, not all the delayed-neutron groups can be expected to be of use. In most reactor systems the useful groups are likely to be the ones with mean half-lives of about 6 seconds, 22 seconds and 55 seconds.

In applying the process described here, three energy intervals were chosen, each optimizing the contribution from one of the groups. Each energy interval would necessarily have contributions from the other groups. It was not possible to define the magnitude and location of the energy intervals from the available information. It is expected that such definition will be part of a routine development phase of the process. The count rates in the three intervals can be related one to another in three different combinations for the purpose of redundancy of data if necessary. The number of contributing groups and the number of intervals chosen do not have to be related to one another. The choice of three intervals is relatively arbitrary, but it provides a means to demonstrate achievement of higher levels of reliability than with the minimum of two intervals.

Relating the count rates in different energy intervals by ratioing eliminates various parameters which would otherwise increase the uncertainty of the observations. The final expressions were obtained in terms only of the parameters which directly affect the desired observations. These parameters are the enhancement factors, hold-up times. and the temperature and power dependencies of each. The actual magnitude of each of these parameters is not important to the feasibility of the process.

The application of the process requires continuous measurements of the count rates from start-up onward. The data at any given time is compared with the same set at a previous time. This type of function is best performed through a computer. The practical application of the process is expected to require an on-line computer.

In practical terms, the measurements to be taken are the count rates in each of the energy intervals and the time rate of change of each count rate. The ratios of these measurements are made as required by the appropriate equations. From the expressions developed in the Detailed Description, a set of three ratios would be obtained at any time during steady-state operation with or without failed fuel in the core, the actual values being representative of the existing condition. If changes occur in any of the measurements which may cause suspicion of a heat generating blockage, an immediate down transient would be initiated and the measurements associated with the transient described in the Detailed Description would be made. After the reactor stabilizes at some predetermined lower power level, the rest of the measurements described would be made.For the conditions described, this procedure gives a matrix of nine values from which a decision would be made regarding the possibility of failure propagation. The criteria for such a decision would have to await the experimental data as described earlier.

It is a principal object of the invention to provide extended operation of the reactor with more failed elements and subsequent longer generation. The ability to operate longer between shut-downs required for extraction of failed elements would contribute to the economic operation of reactors.

The invention resides broadly in a method of detection of failed fuel rods for nuclear reactors comprising sensing neutron radiation in a nuclear reactor characterized in that said nuclear radiation is sensed in a given number of neutron energy levels and differentiating the energies sensed to distinguish between failed fuel rods and leaking fuel rods by detecting neutrons at given energy levels.

Figure 1 (A through F) are graphical representations of the yield of delayed neutron precursor groups (1 through 6) versus neutron energy; Fig. 2 is a composite graphical representation of the total delayed neutron yield from fast-fission of Pu239 versus neutron energy; and Fig. 3 is a schematic of the invented system.

A new process is devised to provide information to differentiate between the signal from heat generating blockages, or other large exposure of fuel, from gas leaking fuel pins. A schematic diagram of the flow system representing a reactor coolant channei, reactor core flow traverse to a detection system and return to the core inlet is shown in Fig. 3. This diagram represents a generalized flow-system to include all types of reactors. There could be several such flow loops in each reactor. The principles described here are separately applicable to each and every one of the loops. The process takes advantage of the following three factors: (1) Delayed-neutron Precursor Half-Lives: Conventionally, the neutrons from fission are divided into six groups called precursors with approximate half-lives as shown in Table 1.From the perspective of reactor instrumentation, the useful precursors are those in Groups 1 , 2 and 3. It is seen that the difference in the rates of decay of the members of these groups is about a factor of ten.

(2) Delayed-neutron Energy Spectrum: The delayed-neutron energy spectrum associated with each of the groups is shown in Fig. 1. The data shown is for 239Pu. A composite of the groups is shown in Fig. 2. The available data indicates that a consensus has not yet been achieved on the detailed appearance of the delayed-neutron spectrum. However, the existence of peaks and valleys of the order of 50 KeV in width appears to be strongly indicated. The existence of such structure in the energy spectrum aids in, but is not necessary for, the implementation of this process.

(3) Precursor Hold-up in Fission-gas Leakers. When fuel is directly exposed to the coolant, almost all the precursors enter the coolant by direct recoil from the surface layer of the fuel. Hence, the precursors are instantly carried away by the coolant. In the case of fission-gas leaking out of a fuel pin, there are two categories of contributors to the precursors found in the coolant. The first category is the set of precursor atoms which recoil out of the fuel directly into the coolant; this contribution is directly proportional to the area of the breach. The second category is associated with contributions from areas other than the exposed surface. The mechanisms which control this contribution are poorly understood at this time. The magnitude of the contribution is expressed as an Enhancement Factor which is a sensitive function of the power level. Associated with the enhancement factor is a hold-up time, which expresses the time required for precursor atoms to be carried from various internal sections of the breached pin to the location of cladding failure. The existence of a hold-up time and an enhancement factor has been verified by observations.

TABLE 1 Delayed Neutron Groups Half-Life Group 1 55 seconds Group 2 22 seconds Group 3 6 seconds Group 4 2 seconds Group 5 0.6 seconds Group 6 0.2 seconds The use of the delayed-neutron precursor half-lives and the hold-up time has been proposed or implemented in reactor systems previously. The novelty of the process described here lies in taking advantage of the energy spectrum of the delayed-neutrons. This factor provides an additional dimension with which to decipher the characteristics of the breached pins or exposed fuel. The use of neutron spectrometry enables obtaining a direct indication of the precursor groups contributing to the neutrons observed. Changes in the precursor mix giving a particular signal can be indicative of the elapsed time between creation of the precursors and their detection.

The application of neutron spectrometry would be a relatively simple matter if the composite neutron spectrum were characterized by clearly identifiable peaks, with a peak available for each group of delayed-neutron percursors. In such a case, the area under a peak would indicate the contribution of one isotope to the total neutron emission rate. Such a contribution could be calibrated to determine an area of fuel exposure to the coolant if the following are known: (a) the local fission rate, (b) the flow distribution between the coolant flow at the point of generation and the flow stream in the system, (c) the delay time between precursor generation and detection.

In the absence of clearly defined peaks (the peak structure of the composite spectrum is not well defined) the neutron spectrum can be used as a relative measure of the delayed-neutron contribution.

The contribution of each group of precursors is different at various energy levels. The count rate at one energy level relative to another is a measure of the delay time because of the different half-lives of the precursors.

To apply this principle in the process described here, discrete energy intervals are counted as indicated in Fig. 3. The purpose of selecting multiple intervals as shown is to provide a means to eliminate by normalization those parameters which it would be impractical to measure locally over the whole reactor. For example, consider the count rate in one of the energy intervals shown (denoted by D) and assumed to originate from a pure recoil source in the core. Refer to Table 2 for definitions of symbols.

TABLE 2 Definitions = = a dilution factor which relates a precursor concentration at the detector with that at the source, dimensionless.

A = Surface area of fuel exposed to coolant, either due to flow through a bed of particles or due to opening up of a cladding breach, cm2.

R1 = range of delayed-neutron precusor "i" fission fragment, cm.

F = fission rate per unit volume, cm-3 set'.

yj = fission yield of contributors to Dl.

Q5 = volumetric coolant flow rate in the channel with the exposed surface, cm3 set'.

tTR = transit time of coolant from the location of fuel exposure to the detectors, sec.

N5 = Precursor density in the coolant of delayed neutron group "i" in the immediate vicinity of the exposed fuel or the opening of the breached pin, atoms per cm.3.

NjD1 = Precursor density in the coolant of delayed-neutron group "i" at the detector location 1, atoms per cm.3.

D1 = Count rate in Detector 1 for the interval associated with precursor group "i", counts per second.

D2j = Count rate in Detector 2 for the interval associated with precursor group "i" after a delay time td with respect to D1, counts per second.

Ar = Decay constant associated with contributors to D1, sec.-l.

Su; = Detector efficiency factor in counts per second per unit volumetric concentration of precursors in the coolant at the detector location.

Similarly, consider the count rate in a different energy interval, given by Dj:

As is evident from Equation (3), by obtaining a ratio between the count rates at two different energy intervals for recoil, the parameters related to flow rate, area of exposure and fission rate are eliminated. The ratio

would remain constant in time unless there is a change in the characteristics of the source which changes any of the quantities in Equation (3). The ratio in Equation (3) exactly expresses the ratio for the condition where no failed pins exist in the reactor core and all the delayed neutrons are due to tramp fissile content in the coolant and the trace fissile content in the cladding, ducts and other core structures.If three intervals are established as shown in Fig. 3, three ratios

can be determined at any time by the different combinations of i and j. A set of these ratios found in a clean core - i.e., one in which all fuel pins are intact -- serves as a reference for a pure recoil source.

The expression for the count rate at the detectors will be changed if breached pins contribute to the signal. This is because Enhancement Factor (EF1) and Hold-Up time (thru) within the pin influence the parameters as follows:

where it is assumed that the enhancement factor could be different for the various precursor isotopes.

The ratio for breached pins corresponding to Equation (3) is:

When Equation (5) is compared with Equation (3), it is evident that the hold-up time has a magnified effect due to its occurrence in the exponent. Aiso, if Equation (5) represents a time of failed fuel operation and Equation (3) a clean core, the ratio of ratios is given by:

The following observations can be made without Equation (6): (1) The ratio is identically 1.0 if taken at two different times when recoil alone contributes entirely to the signal, irrespective of the integrated delayed-neutron count rates at each of the times. This is because tHU = 0 and Enhancement Factors do not apply for this case.

(2) The ratio is a very sensitive function of the hold-up time tHU and also of the difference in decay contants represented by "i" and "j". The ratio could be greater or less than 1.0 depending on the actual numerical values. A numerical magnitude distant from 1.0 represents a high likelihood of only breached pins contributing to the delayed-neutron signal.

(3) If the magnitude of the ratio approaches 1.0 after being distant from 1.0, the tendency may be interpreted to mean that the extent of fuel exposure is increasing. This could mean an increased likelihood for the formation of a heat generating blockage. The expected trends in such a change, or the threshold values applicable to make operational decisions, have not been determined as part of this report. It is expected that experimental data would be needed, apart from analyses, to provide such information.

In a fresh reactor, the signals and ratios between signal levels would be monitored as part of startup procedures. If the detector location is such that the background contribution from the reactor is negligible, the measured readings are due to trace fissile nuclides in the coolant and in the core structures. It is very likely that at start-up, there will be no breached pins in the core. The contribution from the trace fissile nuclides are equivalent to direct fuel exposure because the enhancement factor is 1.0 and the hold-up time is zero. Thus, in a fresh reactor operating at steady-state, a set of reference values of signal level ratios is obtained for each pair of delayed-neutron precursor groups. These reference values correspond to the ratio in Equation (3).As reactor operation continues with irradiation of the fuel, the signal levels would remain contant until occurrence of the first breached pin. The breached pin will be detected by the failed fuel detection system which includes the cover gas monitoring system and the conventional delayed-neutron detection system. It is extremely unlikely that the first breached pin would be immediately accompanied by fuel exposure such as a heat generating blockage. When the first breached pin occurs, the reference set of signal ratios corresponding to Equation (5) is obtained. Every subsequent breached pin would be expected to behave similarly to the first and the signal ratios would be relatively unchanged. The possible variation in signals with a multiplicity of breached pins spatially distributed through the core is not addressed in this report and must be studied separately.However, in principle, the signal ratios corresponding to the breached pin condition could continue indefinitely with no deterioration of the condition of the breaches. Any change in the direction of the fuel exposure condition would indicate that a heat generating blockage, or fuel release may have occurred.

To corroborate the existence of a condition other than simple fission gas leakers, two other means for evaluating the signals were developed involving changing the reactor power level. In the first, the count rate associated with each neutron energy interval is passed through differentiators, giving the time differential of the count rate. In the expressions for detector signals, the changing power is represented by F(t) and the time dependent enhancement factor (due to the dependence on power level) is represented by EFj(t) to include the hold-up time for the breached pin.

In a manner similar to development of Equations (1) through (6):

where the subscript HGB has been substituted for recoil because the transient would be applied due to a concern regarding heat generating blockages. F'(t) is the rate of change of fission rate with respect to time, and EF'(t) is the rate of change of enhancement factor.

Similar to the first case, the magnitude of the quantity expressed in Equation (9) is measured as part of start-up by imposing an appropriate transient on the reactor. Note that Equation (3) and Equation (9) have the same right hand side; this would need to be confirmed in practice. When the first and subsequent breaches occur, the quantity expressed by Equation (12) is found. As long as this quantity is different from Equation (9) [i.e., the right hand side of Equation (13) is different from 1.0 by some specified margin], a heat generating blockage can be precluded. The value for the right hand side of Equation (13) which would be used as the criterion for fuel exposure depends upon the particular reactor system and needs to be measured experimentally.

As part of the measurements made by changing the reactor power level, the following expressions are derived for signals obtained from going to a steady state power level P2 from the initial level of P,.

where the hold-up time has been denoted as being a function of the power level, which is expected to be. Also, it has been separated from the enhancement factor in a manner similar to Equation (4).

Equation (16) merely says that the ratio of signals between P1 and P2 is the same as the ratio of power levels, which is as expected with fuel exposure. This value is modified as shown in Equation (19) for the breached pin by the variation of the enhancement factors and hold-up times as a function of power level. Equation (20) shows the difference between the signals from a breached pin and fuel exposure in response to a power level change.

Table 3 shows the quantities which have been derived to express the difference between the measured signals associated with breached pins as compared with a heat generating blockage. The comparison, which is shown as a ratio, would yield values different from 1.0 only if there is a finite holdup time and/or if there is a difference in the enhancement factor between each delayed group. For the expected application with three delayed groups, nine values for ratios would be obtained when the series of observations are made involving measurements at a steady-state power level Pt, a transient to a power level P2, and a steady-state power level P2. Each of these nine values would be identically 1.0 if the delayed neutrons are from fuel directly exposed to the coolant.If there are several fuel pins distributed spatially through the core, it is not possible at this time to predict the variation of the ratios.

The variation would be expected to be system dependent and needs to be studied separately. However, it is very likely that the pattern of variation applicable to a heat generating blockage would be quite characteristic in its difference from that for breached pins. Thus, with the help of an on-line computer system that continuously monitors plant parameters, it should be possible to establish the combination of signals which would be indicative of gross fuel exposure, and to institute an appropriate response.

An examination of the factors which influence the measurements shows that the signal contributions in each energy interval "i" or "j" are not required to be entirely from the respective precursor group. As long as there is a difference in contribution between one energy interval and another, the ratios would provide useful information with respect to characterizing the delayed-neutron sources.

TABLE 3 Characteristic Ratios of Parameters Ratio of Parameter for Breached Pin to That for Heat Generating Parameter Blockage Counts in Energy Interval "i" EFj (#i - #i) etHU Counts in Energy Interval '1" EF EFj Rate of Change of Count Rate in Energy interval "i" (Transient) F(t)EFj(t) + F(t)EF(t) Rate of Change of Count Rate in Energy Interval "j" (Transient) F(t)EF(t) + F(t)EFj(t) Counts in Energy Interval "i" at Power P1 (Steady state) EF(P A, P2) (P) e e (tHU tHU Counts in Energy Interval "i" EF,(P2) at Power P2 (Steady state) The possibility of a heat generating blockage or gross fuel exposure could invoke a reactor shutdown. From an economic point of view, the shut-down (or, if necessary, a scram) should occur only if a significant safety or maintenance concern exists. Hence, the reliability of the data which forms the basis for shut-down must be maximized. For this purpose, the following are proposed as adjuncts to the process described earlier.

I. Multiple Detectors: As shown in Figure 3, two detectors with a delay time td could be placed in the flow system. This would double the quantity of data on which the control response is based. Also, the reliability of the process is increased.

II. Calibration with Recoil Source: A recoil source may be fabricated in the form of a rod made from a fissile nuclide alloyed with a structurally supporting metal. Other shapes for recoil sources are possible, such as spheres or pellets encased to be held in the core, but directly exposed to the coolant.

Such a source could be used to calibrate the process described in this invention to provide the benchmark matrix of values characterizing a heat generating blockage.

III. Calibration with Breached Pin: A similar benchmark matrix could be obtained by inserting an artificially defected pin into the core. It would be possible to design a fuel pin which contains a recoil source as well as a breached pin simulation. The variation in data with position of the breached pin could also be studied with this device.

Using these adjuncts, it is likely that any specified reliability goal can be achieved. It is recognized that a research and development program would have to precede implementation of the process.

Claims (2)

1. A method of detection of failed fuel rods for nuclear reactors comprising: sensing neutron radiation in a nuclear reactor characterized in that said nuclear radiation is sensed in a given number of neutron energy levels and differentiating the energies sensed to distinguish between failed fuel rods and leaking fuel rods by detecting neutrons at given energy levels.

2. A method according to Claim 1 including the step of time differentiating the sensed neutron radiation count rate.