IR 05000289/1989009

Insp Rept 50-289/89-09 on 890417-21.No Violations Noted. Major Areas Inspected:Measurement Control,Qc,Training & Analytical Procedure Evaluations

ML20247N716
Person / Time
Site:	Three Mile Island
Issue date:	05/25/1989
From:	Jang J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
To:
Shared Package
ML20247N713	List: IR 05000289/1989009 ML20247N711
References
50-289-89-09, 50-289-89-9, NUDOCS 8906060009
Download: ML20247N716 (8)

Text

_ _ _ _ . - - - - _ _ _ _ _ _ _ _

-

.,

-

.

.

.

t U.S. NUCLEAR REGULATORY COMMISSION

, REGION I Report N /89-09 Docket N License N DPR-50 Priority - Category 1 Licensee: GPU Nuclear Corporation P.O. Box 480 Middletown, PA 17057 Facility Name: Three Mile Island Nuclear Station, Unit 1 Inspection At: Middletown, Pennsylvania Ins:tection Conducted: April 17-21, 1989 ds Inspectors: -

[ 6 f-24~-89 M C. Jang, Sdn%r Radiation Specialist date Approved by: .

W R. J. Bore W ChieT, Effluents

@2987 date Radiation Protection Section Inspection Summary: Inspection on April 17-21, 1989 (Inspection Report No. 50-289/89-09)

Areas Inspected: Reutine, announced inspection of the non-radiological chemistry progra Areas reviewed included measurement control, quality control, training and analytical procedure evaluation Results: Within the areas inspected, no violations were identifie gg6060oop890525 Q ADOCK 05000289 PDC

_ _ _ _ _ - _ _ _ _ _ _

_ _ _ - - _ - ___

( .e

-

.

.

..

. -

. i

!

!

l DETAILS j 1.0 Individuals Contact _ed e 1.1 Licensee Personnel l

• T. G. Broughton, Operations and Maintenance Director, TMI-1  !

R. Bybel, Group Chemistry Supervisor, TMI-1  !

• E. Fuhrer, Chemistry Manager, TMI-1 l

• H. D. Hukill, Director, TMI-1  !

L. Lucas, Group Chemistry Supervisor, TMI-1 '

• S. M. Otto, Licensing Engineer, TMI-1

• J. G. Reed, Senior Chemist II, Plant Enginee !'

E. Gliot, Chemistry Instructor, Training Department 1.2 NRC Personnel l

• T. A. Moslak, Resident Inspector j i

• denotes those present at exit meeting on April 21, 1989. j i

2.0 Purpose l l

The purpose of this inspection was to review the licensee's chemistry l program in the following area l l

The licensee's ability to measure non-radiological chemistry l

~

parameters during normal operation The licensee's ability to demonstrate the acceptability of the analytical results through implementation of quality control ,

activities'in the chemistry laboratory

{

3.0 Previously Identified Items j (Closed) Inspection Follow-up Item (50-289/88-30-02): Demonstration of j analytical capability at minimum detectable concentrations. The  !

licensee's analytical capability at these concentrations was demonstrated i by analysis of NRC standards during this inspection. (See SerHon 5.)

4.0 Review of Analytical Procedures I l

The inspector reviewed the following analytical procedures related to the l non radiological chemistry progra )

N1915 Rev. 10 " Determination of Ammonia" N1916.3 Rev. O " Ammonium and Morpholine by I.C. (4000i)" ,

!

,

!

_ ___ _____ __ _________-__ _ _ _ _ _ _ __

- - _ _ _ _ _ _ - _ -

. y

-

.

.

'

(l

N1912.3 Rev. I " Air-Acetylene Flame on the Perkin-Elmer 2380 i Atomic Absorption Spectrophotometer" N1914 Rev. 1 " Determination of Hydrazine in Water (Low Concentration)"

N1913 Rev. 7 " Determination of Molgdate Reactive , Silica" N190 Rev. 8 " Determination of Free Fluoride Ion" N1912.2 Rev. 2 " Graphite Furnace Analysis using the Perkin-Elmer 3030 Atomic Absorption Spectrophotometer"

. Based on the review of the above procedures, the inspector determined that the licensee's procedures were technically adequate. The procedures are required by Regulatory Guide 1.33, which is referenced in Section 6.8.8 of the licensee's Technical Specifications. No violations were identifie .0 Evaluation of the Analytical Capability During the inspection, standard chemical solutions were given to the licensee for analysis. The standard solutions were prepared by Brookhaven National Laboratory (BNL) for the NRC Region 1[ and were analyzed by the licensee using normal methods and equipment. The analysis of standards is used to verify the licensee's capability to monitor chemical parameters in various plant systems with respect to Technical Specifica-tion and other regulatory requirement In addition, the analysis of standards is used to evaluate the licensee's analytical procedures with respect to accuracy and precisio ' After the. licensee provided their lower limits of quantification (LLQs)

for the non radiological chemical parameters to the inspector, the appropriate NRC standards were provided to the licensee with instructions for dilution to achieve solutions at concentrations near the licensee's LLQ The licensee's LLQs were:

Ammonia (Ion Specific Electrode and Ion Chromatography): 500 ppb Fluoride (Ion Specific Electrode): 40 ppb Hydrazine (Spectrophotometry): 10 ppb Silica (Spectrophotometry): 10 ppb Lithium ( AA-Flame): 200 ppb Sodium (AA-Flame): 30 ppb Iron (AA-Graphite): 30 ppb The results of the comparisons are listed in Table The analytical results of the standard measurements comparison indicated that one (1)

out of twenty-four (24) comparisons was in statistical disagreement under the criteria used for comparing results (see Attachment I). This statistical disagreement is insignificant because of the high precision ,

associated with the result., 1 i

_ - - _ _ _ _ _ _ _ - _ - _ _ _ _ - _ - _ _ - - _ - _ _ _ _ _ - - - _ _ _ - _ _ _ _ - _ _ - _ - - _ _ _ _ - - _ _ _ _ _ - _ _ _ .-_-_--_-______--_--____n

, - _ __ _ . _ _ _

r .

.

-

.

.

The results of two chemical parameters (silica and hydrazine analyzed using the spectrophotometry method) did not meet the agreement criteria until they were re-analyzed a third tim The first and second analytical results of silica and hydrazine were biased from the known values but showed high degrees of the precision. These results indicated a potential problem with the calibration standard solution The inspector had discussions with the licensee regarding the preparation of their calibration standards. The licensee stated that their calibration standards may have degenerated. In order to determine 11 calibration standards were the problem, silica and hydrazine standards were prepared and certified by GPUN Laboratory Services during this inspection. The licensee used these certified standard solutions for the instrument calibration. The results of silica and hydrazine measurements were then in agreement. In general, the laboratory environmental conditions (temperature and-humidity) were acceptable. The inspector also noted that the licensee maintained the spectrophotometer regularl .0 Chemistry Laboratory Quality Control Program The inspector reviewed Section 3.1 (Preparation and Analysis of Spiked Samples) of Procedure N1826, " Control of Analytical Performance". The inspector also reviewed licensee's " percent recovery of spike samples" data and associated control charts for anions, cations, boron, hydrazine, and silica. These control charts consist of nean values and acceptance levels. The inspector noted that the licensee has performed these items as required by Section 3.1 of Procedure N182 Reviewing the above control charts, the inspector noted that the licensee should use those control charts for trending analyses of each chemical parameter to better control accuracy of the analytical results-. For example, the average spike sample recovery of hydrazine was about 106%

(which was within the licensee's hydrazine acceptance criteria) in April, 1989. The first analytical results of the NRC hydrazine standard solutions were about 7% higher than the BNL known value. This would have been compensated, had the hydrazine recovery data been utilized to correct for system calibration shifting. This was identified by the inspector through use of the results of the licensee's spike sample progra Similar results were shown in silica analyses.

l The inspector calculated means and standard errors for analytes using the licensee's " percent recovery of spike samples" data and the results are i

,

listed in the following table.

l

_ __ _ _ _ -

_- __ _ _ - .

.

. .

.f Percent Recovery of Spike Samples TMI-1 Chemistry Laboratory April 19, 1989 Number of Percent Analyte Mean Determinations Recovery Range Hydrazine 106 3 6 100.3 - 10 Silica 97 3 10 89.0 - 10 Ammonia 97 6 .10 89.9 - 10 Boron 9 .6 (Manual) 4 98.6 - 9 .8 0.6 (Auto 15 98.9 - 10 Titration)

Lithium 97 4 14 95.0 - 10 Copper 108 10 4 99.0 - 12 Iron 102 9 2 95.0 - 10 Chloride 103 7 18 89.2 - 11 Fluoride 94 3 15 90.0 - 10 Sulfate 100 6 18 85.7 - 10 Sodium 97 5 4 90.0 - 10 (boron 5 90.0 - 10 Matrix)

The highest percent recoveries of silica, fluoride, and sodium were about 100%. These suggested that analytica l results of silica, fluoride, and sodium might be underestimated during April, 1989. The mean and standard error of copper was 108 10% in Aprilji989. This suggested that analytical results of copper might be overestimated about 8% during April, 1989. Although all analytical results of spike samples were within the licensee's acceptance criteria, the bias (using mean values) and the precision (using standard errors) of each analyte measurement could be estimate Based on the above review, the inspector determined that the licensee had a good quclity control program but its results were not being fully utilize The inspector recommended that the licensee use the

_ - _ _ _ _ _ _ - _ _ - _ - _ _ -

. _ - _ _ _ _ , _ . . _ _ _

.

', ',

j 6  ;

!

i spike sample recovery program to evaluate the accuracy of his 1 measurement J 7.0 Training and Analytical Performance

'

The inspector observed on the job training (0JT) in the Unit-1 Chemistry Laboratory. This OJT was conducted by an instructor from the Training Department. The instructcr observed a technician's analytical /calibra-tion performance and then discussed any improvements in techniques and theory with the technician. The inspector discussed with the instructor the purposes and objectives of the OJT. In addition, the analytical performance of technicians was evaluated by the supervisor through the

" percent recovery of spike samples" program (Procedure N1826, " Control of .

Analytical Performance").

Based on the above observation, discussion, and review of the analytical performance data, no violations were identified in this are .0 Exit Interview The inspector met the licensee representatives (denoted in Section 1) at the conclusion of the inspection on April 21, 1989. The inspector summarized the purpose, scope, and findings of the inspectio _ - _ _ _ _ _ _ _ - _ .- _ _ . _ - - .

le

. -

...

.

..

i TABLE 1

.

TMI Unit 1 Capability Test Results Unit = ppb NRC Licensee Ratio Analyte Value Value (Lic./NRC) Comparison

'Ammenia 620 30 581 10 0.93 0.05 Agreement (Ion Chromato- 60216 583 10 0.97 0.02 Agreement graphy) 980 50 1040 30 1.06 0.06 Agreement Ammonia 620 30 627 12 1.01 0.05 Agreement (Specific Ion 60216 590 10 0.98 0.02 Agreement Electrode) 980 50 890 40 0.91 0.06 Agreement Fluoride 45 4 44.3 .98 0.09 Agreement (Specific' Ion 84.6 .71 .99 0.02 Agreement Electrode) 166 3 163.2 .98 0.02 Agreement Lithium 197 4 204 5 1.04 0.03 Agreement ( AA-Flame) 300 7 308 5 1.03 0.03 Agreement 413 10 423 3 1.02 0.03 Agreement Sodium 61 7 63.7 .04 0.12 Agreement ( AA-Flame) 106 6 119.71 .13 0.07 Agreement 158 9 174 2 1.10 0.06 Agreement Iron 37.2 .3 .98 0.03 Agreement (AA-Graphite) 79.2 .3f .98 0.02 Agreement 117 3 120 3 1.03 0.04 Agreement Silica 10.6 .1 0.1(2) 0.95 0.06 Agreement (Spectrophoto- 20.8 .8 0(2) 0.95 0.04 Agreement metry) 31.4 .2 0.1(2) 0.962 0.013 Statistical Disagreement (2)'

Hydrazine 19.9 .9 0.1(2) 1.000 0.015 Agreement (Spectrophoto- 49.9 .5 0.2(2) 0.992 0.011 Agreement metry) 100.0 .0 0.3(2) 0.990 0.010 Agreement (2) Third analytical results (2) Insignificant

,

- - - - - - _ _ _ _ - - - - - _ - . _ - _ .

__ - - - _ - -

_ - -._ ._- _. - . _ _ - _ _ -

l .

.

. ..

,

ATTACHMENT I CRITERIA FOR COMPARING ANALYTICAL MEASUREMENTS This attachment provides criteria for comparing results of. capability tests.

l In these criteria the judgement limits are based on the uncertainty of the ratio of the licensee's value to.the NRC value. The following steps are performed:

(1) the ratio of the licensee's value to the NRC value is computed ratio =. Licensee Value NRC Value (2) the uncertainty of the ratio is propagate If the absolute value of one minus the ratio is less than or equal to twice the ratio uncertainty, the results are in agreemen (ll-ratiol s 2 uncertainty)

2Z= 5' then 3* S*+bY y Z2 x2 y2 2(From: Bevington, P. R., Data Reduction and Error Analysis for the Physical Sciences, McGraw-Hill, New York, 1969)

_