DEVELOPMENT OF Ci CONTENT DETERMINATION TECHNOLOGY IN VARIOUS KINDS OF DRUMMED RADIOACTIVE WASTES USING A NON-DESTRUCTIVE METHOD APPLICABLE FOR NUCLEAR POWER PLANTS IN KOREA

DuckWon, Kang and Kyoung Rock, Park
Korea Electric Power Research Institute
Radiation Safety Group

ABSTRACT

Various kinds of radioactive waste streams, such as ion exchange resins, water, compacted paper trash, rolls or stacks of air handling filters and solidified waste are produced at nuclear power plants in Korea. These wastes are packed into the drum at the radwaste building and temporarily transferred to a temporary storage building. Several kinds of drums are used for radioactive waste packing at nuclear power plants in Korea. However, estimation of nuclides and their activities in the drummed radioactive waste is still difficult and unreliable. Nuclear act of Korea requires low and intermediate level radioactive waste generated at nuclear power plants prior to transportation to permanent disposal sites to be known in detail. Individual historical records of the radioactive waste should contain information about the activities of nuclides, total activity and the type of waste in the drum. Consequently, with the incorporation of gamma spectroscopy, a fully automated nuclide analysis system which can accurately evaluate the nuclide and activity in the drum was developed. The activities of the representative -emitters(Cs-137, Co-60) which are mixed with several materials in the drum were measured by this system. Then appropriate scaling factors were used to assay the activities of nuclides which could not be directly analyzed by this system. The scaling factors were determined by a computer program developed in Korea. The validity of the scaling factor was checked through comparison with several experimental results of local samples taken from real waste streams. Demonstration drums with similar geometries to real waste drums were used for proving the reliability of Ci content determination in this system. It was performed by putting the standard sources(Co-60, Cs-137) into demonstration drums. In the results, the measurement errors were less than 30% in the various demonstration drums. Therefore, measurements of real radioactive waste drums gave us good results for the homogeneous and non- homogeneous wastes generated at Kori(the name of a place in Korea) nuclear power plant in Korea.

INTRODUCTION

Since Kori unit 1 started commercial operation in 1978, 10 PWRs and 1 PHWR have been in operation in Korea and 4 PWRs, 3 PHWRs have been under construction by KEPCO (Korea Electric Power Corporation) in Korea. Many activities at nuclear power plants generate radwastes of various type, such as ion exchange resins, water, compacted paper trash, rolls or stacks of air handling filters, and solidified waste. All solid radwaste drums are temporarily stored in on-site storage buildings at each site until final disposal facilities are available. The temporary storage capacity is not sufficient to accommodate all radwaste packages. Therefore, the radwaste drums stored in the temporary building have to be transferred to the final disposal sites to be buried permanently in future. Nuclear Act of Korea requires the manifest of radwaste drums generated at nuclear power plants prior to final disposal. It should contain information such as the activities of nuclides, total activity, weight and the type of waste in the drum.

The method to establish an assay technique for determining the kinds of radionuclides and their activities in homogeneous and non-homogeneous radioactive materials non-destructively has been continuously developed since 1992. It consists of a fully automated nuclide analysis assay system, non-destructive analysis and evaluation system of the radioactive wastes in the radwaste drums. The activities of the representative -emitters(Cs-137, Co-60) which are mixed with several materials in the drum were measured by this system. Then appropriate scaling factors were used to assay the activities of nuclides which could not be directly analyzed by this system, such as alpha, beta emitter and low energy gamma nuclide. The scaling factors were determined by a computer program developed in Korea. Furthermore, this assay system can automatically mark the analysis results onto the drum. Also the drum handling machine which was developed could reduce the radiation exposure to workers. It is designed to move automatically the radwaste drums to the measurement position, then analyze the activities of the nuclides in the radwaste drums.

This paper discusses the constitution of the radwaste drum assay system, the evaluation method and results.

MATERIAL AND METHOD

Radiation Measurement

Generally, it is assumed that the radwastes in the drum are homogeneous and self-absorption in the matrix is negligible when measuring the activities of the waste drum. However, above assumptions may give rise to error when the radwastes are evaluated by the assay system. Therefore, to measure accurate Ci content for the waste drum, the main concern is how to reduce the error. The segmented gamma scanning method has been widely used to effectively measure the activity of nuclide in the radwaste drum. In particular, it is suitable for analysis of the low density waste drums. The SGS(segmented gamma scanning)method utilizes the following techniques to reduce the error from non-homogeniety self-absorption effects of the matrix.

• Measurement of the segments which are divided into some parts at the vertical plane to compensate a non-uniformity for the vertical direction
• Rotation of the waste drum to compensate a non-uniformity for a radius of the drum
• Measurement of the transmission in order to correct the attenuation due to a matrix of the sample

The following is the basic principle of segmented gamma scanning method. HPGe detector is shielded to decrease the background radiation and the shielding material is available as the collimator, as shown in Fig. 1. Therefore, radiation from the waste drum enters into the HPGe detector as cone type. We can easily calculate count rate for radiation entering the HPGe detector from radioactive material through the following equation.

where, K₁ = 1-exp(-µh), Zo is net count rate of detector for standard gamma source, It is equal to Zo = k A_o/4S_o2, Ao is activity of standard gamma source, k = calibration coefficient, So = distance between detector and waste drum, We assume that Ca, activity concentration is distributed uniformly over the volume in the waste drum. Fo is the area of sample in the waste drum entering into the detector. Therefore, In Eq. (1), we can see the correlation between detector count rate and activity concentration of radioactive materials. That is, if we know the linear attenuation coefficient for radioactive materials, µ, we can get the concentration for radioactive material from the detector count rate. The equation is the following

In Eq. (2), the values of Ao, Zo, Fo can be obtained through the assay system calibration with standard gamma source. The linear attenuation coefficient, µ is given in the average density attenuation correction, transmission source correction, differential peak attenuation correction.

Attenuation Calculation

The matrix correction is to correct the measured peak efficiencies for sample self- absorption. These peak efficiencies may then be used to provide an improved calculation of sample activities. The three methods of correction used in this paper are Average Density Correction, Transmission Source Correction, Differential Peak Correction.

In each of these methods an attenuation correction factor is calculated for each peak found in the sample, then the individual peak efficiencies are scaled by this amount. These methods is to calculate an attenuation correction factor CF_i , as shown in the following equation.

where, CF_i is the correction factor at energy, E_i

i is the uncorrected efficiency at energy E_i

_c is the corrected efficiency at energy E_i

The CF_i has the following form for each peak

where, µ_i / is the mass attenuation coefficient [cm²/gm] at the energy Ei and (t) is the calculated sample mass per unit area [gm/cm²]. We can use the following three methods to take a value of (t)

Average Density Attenuation Correction

The sample average density is calculated from the sample net weight and the container volume. The density of the sample is then multiplied by the sample average thickness to find the value of (t). This is mainly used in case of the uniform matrix.

where, ad is subscription of average density, M_net is the sample mass, V is the sample volume, t is the average thickness of the sample.2.2 Transmission Source Correction

This technique uses the energy from the transmission source to estimate the density. The transmission for a gamma ray of energy E_i is defined as the attenuation factor.

where x is the mass per unit area for the attanuating material. If we use a transmission source with a gamma ray at energy ET, then the transmission at energy E_i is given by:

Alpha is the ratio of the gamma absorption coefficients at two energies. The transmission for the transmission source gamma peak is found by measuring the initial peak area due to the transmission source without any attenuating material and then measuring the same gamma peak area during the assay with attenuating material present. The transmission measured is then:

where I_T(0) is the transmission peak area without attenuating material, and I_T(x) is the peak area with attenuation. Once the transmission has been calculated then the attenuation correction factor can be found using the approximation

where T_i is the transmission at E_i. Kappa() is a correction factor which will depend on the particular geometry under consideration. For a cylinder, it is equal to 0.823 in the our SGS system. This attenuation factor can then be applied directly to correct the peak efficiency at each energy Ei. The measured transmission may also be used to calculate a (t) value. By using the original definition of the transmission Eq. (4 ) we may define a (t) such that:

then the (.t) for Ei will be

Differential Peak Attenuation Correction

Nuclide that exist in a given sample volume in some case have two peaks that can be identified from the associated spectrum. In those situations, we can use the relative attenuation of these two peaks to obtain an estimate of the sample density. The calculations require that we know the unattenuated intensities of the two lines. These may be obtained from an initial efficiency(response) calibration. Then we can calculate the value of (.t) as the following.

where,

I₁(0) = the unattenuated count rate of the low energy peak
I₂(0) = the unattenuated count rate of the high energy peak
I₁(t) = the attenuated intensity of the low energy peak
I₂(t) = the attenuated intensity of the high energy peak D
µ = µ₁ - µ₂

Attenuation Correction of Non-uniformity Matrix

This techniques is to reduce the error of the measurement due to the non-uniformity of the matrix in the container and the effect of self-absorption in the matrix. This divides horizontal area into some sectors (we selected 8 sector) to correct the non-uniformity of the radioactive materials in the container.

This techniques's algorithm is similar to the geometry mean. If a non-uniformity exists, then a geometric averaging method will be used to reduce effectively an in homogeneity of horizontal area .

where A_n is the collected peak data at a sector of the segmented horizontal area during the rotations of the container.

Constitution of the Radwaste Assay System

The radwaste assay system is composed of a gamma spectroscopy system, an automatic drum transfer system(stacker crane), a ink-jet labeling system and a drum rotating device. These systems are connected to a main computer for the remote control. The simplified processing is shown in the following Fig. 2

Fig. 2. The simplified processing of the system.

The gamma spectroscopy system is to evaluate the activities of the nuclides from radioactive materials. The high purity germanium detector which is manufactured by Canberra Inc has 50% efficiency. The automatic drum transfer system which is developed by our engineer is to automatically transfer the waste drum onto the turn table. The turn table which consists of a part of the waste assay system is rotated to reduce the error of non-uniformity for a horizontal area.

Scaling Factor

Since December of 1984 (one year from publication of 10CFR61), utilities have been required to list the quantities of several very difficult to measure radionuclides on all radioactive waste shipments to the low level waste burial sites. some of these nuclides had never before been measured in low level waste. Furthermore, these difficult nuclides pose analytical problems beyond the capabilities of power plant laboratories. Therefore, as a indirect method, we have used the scaling factors which infer the concentration of one radionuclide from another that is actually measured. That is, the scaling factors are to relate the difficult-to-measure nuclide in the radwaste packages to those that are easily measured. It is also able to utility for the non-gamma emitting nuclides based on the ratio gamma and non-gamma radionuclide measured in the radwaste packages. All of these nuclides are to classify the wastes according to requirements of 10CFR61. The scaling factor determining computer code which is developed in Korea based on the EPRI TR-101960 (RAD SOURCE Volume 1. Part 1:A) (1) and EPRI NP 2412-19 (2) is to used the specific nuclear power plant in korea, at KORI site. This computer code requires an initial setup in which the following plant information is input and stored in the program files:

• Number of Reactor Unit
• Reactor of Coolant Mass
• Purification System Flow Rate
• Reactor Coolant concentration for: I-131, I-132, I-133, I-134, I-135, Cs-134, Cs-137,Co-60

The nuclides related to the scaling factor is fission products, collision products, tramp uranium.

Validation of Scaling Factor

This computer code is a scaling factor prediction code developed for use at nuclear power plant in korea. An important element of the adaptation and use of this code in nuclear power plant is the validation of the code results. The approach taken to validate the code results was to compare them against measurements made in operating plant. But The only scaling factors for the collision products were validated those comparing with the data which is made by EPRI NP-4037 and EPRI NP-5077. We selected a Co-60 nuclide for the representative nuclide of the collision products because the corrosion products have a similar characteristic for a process of the formation and the transportation in the reactor coolant. The nuclides which are to predict a scaling factor among the corrosion products, that is, H-3, C-14, Ni-63, Fe-55, Nb-94, is selected for one. The Fig. 3 is a compar ision of scaling factor for the corrosion products of the data analyzed with our nuclear power plant's samples and the EPRI NP-4037, EPRI NP-5077 data.

Fig. 3. Ni-63 scaling factor comparison.

The validation of the scaling factors for the fission products, Tramp uranium nuclides, was to compare them against realistic sample measurements made in operating plant. The scaling factor generated by the computer code using reactor coolant isotopic concentration data obtained during the test period were plotted against the scaling factors derived from the korea atomic energy research institute measurement. The Fig. 4 shows the comparisons for the Sr-90, Tc-99 and TRU. The plots also shows the plus and minus factor 10 boundary lines established for acceptance within NRC guidance (3) and a relatively random dispersion of the computer code results around the line that indicates a perfect fit with measured values. The comparison between the computer code results and the korea atomic energy research institute measurement were quite good in the plus and minus factor 10 boundary lines.

Fig. 4. The validation of Sr-90, Tc-99. TRU scaling factor.

Measurements of the Radwaste Drums

For the waste assay system which was developed by KEPRI, we accomplished experiments to optimum its operating condition. The experiment methods were to determine a rotating speed (rpm) of the waste drum in order to compensate a non-uniformity and to compare an activity measured from standard gamma sources against the original activity of standard sources. Demonstration drums with similar geometries to real waste drums were used for determining a rotating speed of the waste drum and proving the reliability of Ci content determination in this system. It was performed by putting the standard sources (Cs-137, Co-60) into demonstration drums. As shown in Table I, with the results measured from demonstration drum turning into the speed, we decided that the rotating speed was a 10 rpm as optimization speed for the waste drum. For a reliability of the measurement for the waste drum, three type's drums which have the different density was used. In the results, It was a little different in error of the measurements depending on the source position and its arrangement but the measurement errors were less than 30% in the various demonstration drums. It gave us good results for the homogeneous and non-homogeneous wastes. The Table II is the measurement results of standard sources in the waste drums which have different densities.

TABLE I The Measurement Results Depending on Varying the Rotating Speed of the Drum

TABLE II The Measurement Results Depending on the Different Densities of the Drum

RESULTS AND DISCUSSION

The waste assay system was developed to accurately evaluate the nuclides and their activities in the drum with the incorporation of gamma spectroscopy. The activities of the representative -emitters (Cs-137, Co-60) which is mixed with several materials in the drum were measured by this system. Then appropriate scaling factors were used to assay the activities of nuclides which could not be directly analyzed by this system. We have decided that the rotating speed was 10 rpm to optimize the operating condition of this system. Demonstration drums with geometries similar to real waste drums was made and the activity measurements was performed by putting the standard source into the demonstration drum. In the results, the error was less than 30%. It was quite good results . But because the radioactive material in the radwaste drum is more in homogenious than originally assumed, we will try continuously to make correction for the non-uniformity in order to enhance a reliability of the measurement through various experiments.

REFERENCES

1. KNOLL, GLENN F, "Radiation Detection and Measurement," 2nd ed., New York(1989).
2. PARKER, J.L.,"The Use of Calibration Standards and the Correction for Sample Self-attenuation in Gamma-ray Nondestructive Assay", LA-10045(1984).
3. HUSE, S.T, T.E.SAMSON, "Nondestructive Assay Techniques and Associated Measurement Uncertainties," JNMM. 17(1992)
4. SPRINKLE, JR., J.K., and S.T.HSUE, "Recent Advances in Segmented Gamma Scanner Analysis," LA-UR -87-3954(1987).
5. CLINE, J. E., K. L. WRIGHT, J. W. HOLLCROFT, "Activity Levels of Transuranic Nuclides in Low-level Solid Waste from U. S. Power Reactor," EPRI NP-1494, Science Applications, Inc., Rockville, (1980)
6. BEST, W. T., Updated Scaling Factors in Low-level Radwaste, EPRI NP-5077, Advanced Process Tech. California,(1987).
7. VANCE. J. N., et al., "10 CFR Scaling Factor Analysis Program Users Manual, EPRI NP 2412-19, Palo Alto, California,(1992)