ML20214F079
| ML20214F079 | |
| Person / Time | |
|---|---|
| Site: | Braidwood  |
| Issue date: | 11/05/1986 |
| From: | Kostal K SARGENT & LUNDY, INC. |
| To: | |
| References | |
| OL-A-179, NUDOCS 8705220397 | |
| Download: ML20214F079 (34) | |
Text
,
.t
.a' p-l79 9,4 10-30-86
~.
l0'.e t IG
'MiVC
~'
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION 87 APR 2'2 P6 :29 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD GFFit..T n.
00C C h..
. / tt.E.
In the Matter of
)
- n A:C-
)
COMMONWEALTH EDISON COMPANY
)
Docket Nos. 50-456
)
50-457 (Braidwood Station Units 1 and 2)
)
x REBUTTAL TESTIMONY OF KENNETH T. KOSTAL (ON ROREM Q.A. SUBCONTENTION 2)
(Harassment and Intimidation)
IIUCLIM ftGULATORY C0HI$$108 IY ficial Id. lle.
Ded.tk N'e m#
4/o.s to
,, m
- i,,,
ist.mo V Appucast L'
f(CilVED latenener futCi[D Cent's Offt
~ I*
Centracter DATE
- w#
Oui.,
mi e Neb nont.,
G' I
8705220397 861105 gDR ADOCK 05000456 PDR August, 1986 n
V.)
~
f
',,.[(1 ) 7[CS!.I,Is-l 4
~
t'.
10-30-86 s'
d.
i REBUTTAL TESTIMONY OF KENNETH KOSTAL ON THE BCAP CSR SAMPLES AND ENGINEERING EVALUATIONS FOR ELECTRICAL WORK Q.1.
Please state your full name for the record.
I A.1.
Kenneth Thomas Kostal.
Q.2.
By whom are you employed and in what capacity?
A.2.
I am employed by Sargent & Lundy (S&L) in the capacity of Project Director for the Braidwoo'd Project.
f Q.3.
What are your responsibilities for the~d,esign of the l
Braidwood Project?
A.3.
As Project Director for the Braidwood Project, I am responsible for implementation and technical integrity s
of S&L's design efforts.
I regularly report to the client regarding S&L's performance on the Braidwood Project and on the status of engineering.
I work with Commonwealth Edison (Ceco) and the Project Team to establish significant design parameters and direct i
appropriate application of S&L engineering.
s
\\
i
t 10-30-86 Q.4.
Please state your educational background and profes-
)
sional experience.
A.4.
I graduated from the University of Illinois in 1965 with a BA in Architectural Engineering and in 1967 with a MS in Architectural Engineering.
I have 20 years of experience in the field of civil engineering which includes civil / structural / architectural engi-neering and design work for fossil and nuclear power plants.
My assignments have included 14 units with a total capacity in excess of 10,000 megawatts.
I have'
?
also been involved in numerous studies.
In addition, as a structural engineer, I have performed numerous calculations and assessments relating to welded steel construction which includes the design of welded con-nections.
Prior to joining Sargent & Lundy in 1967, I was engaged by the University of Illinois as an instructor in structural design and as an engineer responsible for structural design and construction f
drawings for light office buildings.
i I am a registered professional engineer in 31 states 1
and I also have a separate structural engineering i
license in the State of Illinois and am licensed in Alberta, Canada.
Presently I am a member of the
+
following organizations:
j s
y I -
l 1
l
t
=
10-30-86 American Concrete Institute American Institute of Steel Construction
)
\\
American Nuclear Society American Society of Civil Engineers Structural Engineers Association of Illinois Western Society of Engineers Q.5.
What is the purpose of your testimony?
A.S.
The purpose of my testimony is to discuss the engi-neering evaluations performed by Sargent & Lundy with,
respect to the various weld discrepancies identified in the electrical construction categories during the Construction Sample Reinspection (CSR). - I also am prepared to respond to questions with respect to the methods of structural analysis used to evaluate the objective attribute discrepancies discussed in Mr. Thorsell's testimony.
In this testimony I also review the basis on which the more highly stressed portions of the CSR samples were selected in the electrical construction categories.
In addition, based on the results of the reinspection and engineering evaluations, I, offer an opinion
,((j regarding the quality of the reinspected L. K.
~]
i 10-30-86 Comstock work.
I discuss the conservatisms in the e
design of Braidwood which are relevant in asserting the adequacy of L. K. Comstock's work and the perfor-mance of the Comstock Q.C. inspectors.
Q.6.
What is the basis for your knowledge concerning the CSR sample selection and engineering evaluations referred to in your previous answer?
A.6.
Since the initiation of the CECO BCAP Program in 1984 I have been intimately involved in the program.
This involvement began with the forming of Sargent &
Lundy's project team dedicated to the BCAP Program.
The project team also included a liaiso,n group who provided a direct communication link between CECO BCAP personnel and S&L project team members.
As Project Director I was assigned to head up this entire effort on S&L's part.
Of particular importance was the effort and involvement on S&L's part associated with the Construction Sample Reinspection portion of the l'
i BCAP Program.
I initially provided input into the CSR l
Program.
This included the sampling selection and the inspection attributes associated with each category.
I was involved in daily, communication with the project team during the CSR reinspections and our engineering i
evaluations.
These evaluations were conducted in 1
.g 4
-e,-
.,.m_.-v,_m-,-e-.-y,-..--,
---y
..w--
.-sr,---,
y
~
10-30-86 accordance with S&L standards which were reviewed and approved by me.
In addition, I conducted weekly meetings at the Braidwood Station with the S&L project team to review the progress of our efforts in support of the BCAP Program.
Also reviewed at these meetings were the evaluations performed by S&L in support of the CSR effort.
I reviewed engineering evaluations performed by S&L for electrical construction categories throughout the duration of the CSR and more recently in preparing this testimony.
In addition to the daily involvement in the program I have made numerous status presentations to the NRC on a monthly basis, reviewing the various engineering evaluations performed by S&L.
Are you famil'r with Dr. Kaushal's testimony, and in Q.7.
a particular the table showing CSR Reinspection Results attached to Dr. Kaushal's testimony, Attachment 2C (Kaushal-3)?
3 A.7.
Yes.
\\.>
D G 1 i
is 0
hji,
t, 10-30-86 y,
W:' \\
je*
s-
- f t *,
Q.8.
Does Dr. Kaushal's testimony and in particular
. C (Kaushal-3) accurately reflect the results of the engineering evaluations performed by S&L for purposes of the CSR in the electrical f
construction categories?
A.8.
Yes.
f Q.9.
Dr. Kaushal states in his testimony that the BCAP CSR element included six construction categories within the scope of L. K. Comstock's work.
Do you agree with' this categorization?
A.9.
Yes.
The six construction categories w't,hin the scope i
of L. K. Comstock's work include:
conduit, conduit hangers, cables, cable pans, cable pan hangers and electrical equipment installation.
Each category is a group of hardware items constructed using similar processes or containing similar types of components.
Thete six construction categories encompass all safety-related work performed by L. K. Comstock.
Q.lo.
Is stress a significant design factor with respect to any electrical construc. tion category?
se,
G _
4
[
[,
10-30-86 t
A.10.
- Yes, for one electrical construction c.ategory, cable f
pan hangers.
All cable pan hangers are uniquely I
designed such that the stress in each component is maintained within code allowables.
Cable pan hangers consist of frame supports, composed of individual structural members, as compared to conduit hangers which are generally simple cantilever members.
Due to the configuration of the cable pan hanger supports and the various loads applied to these supports, cable pan hanger designs are performed on an individual hanger basis.
These hangers are designed to accommodate cable pans assumed to be fully loaded.
Member stresses are evaluated for individual hangers during the design to assure they are within code allowables.
i Q.11.
Why isn't stress a significant design factor for the remaining electrical construction categories?
A.11.
Stress is a consideration in establishing maximum conduit spans; however, design stress levels in con-duit is generally low throughout the entire conduit population due to the typical spans found in the i
plant.
Ths design standards are established based on i
the heaviest cables which could fit into each conduit and the physical properties of the smallest diameter conduit.
These standards establish maximum allowable
-U 1
s
[
7-l l
10-30-06 conduit spans which meet code allowable stresses.
In reality, spans are less than allowed due' to plant conditions and arrangements.
Also in most instances maximum size cables ate not installed in each conduit.
These factors result in conduits stressed to much lower levels than allowed.
For conduit hangers, stress is a consideration but it is not a significant design factor.
Individual conduit hangers have been designed by selection from standard conduit hanger load tables.
The conduit hanger load tables have been established generically i
based on stress calculations assuming bounding condi-tions.
For practical economical reasons the design of conduit supports is a simplified conservative method.
For cable construction category, stress is not a design factor.
Cables are supported throughout their length by cable tray and conduit systems.
Therefore, cable stress is not a criteria for design.
l Stress is not a significant design factor in the cable pan construction category, however, it is considered.
I Allowable cable pan spans have been determined by j
stress calculations based on maximum pan loading and seismic accelerations.
Cable pans are supported by
e ~
10-30-86 hangers at intervals less than the maximum allowable standard spans due to plant conditions and arrange-ments.
Since the actual cable pan spans are less than the established allowable spans, unique cable pan calculations are not performed.
The design margins, which are inherent in this standard design approach, result in actual cable pans installed which are not highly stressed.
L For electrical equipment installation, stress is a i
consideration but it is not a significant design factor.
Electrical equipment mounting details have been designed conservatively based on loads derived from peak seismic accelerations.
Many of'the equip-J ment mounting details are standardized, as was the l
\\
case with conduit supports.
Individual equipment mounting details, designed by selection from standard tables, do not have actual stress calculations com-pleted which represent its actual design loaded condition.
l l
Q.12.
Explain how more highly stressed items were added to the CSR sample for cable pan hangers.
For construction categories for,which stress is a A.12.
significant design factor, efforts were made to ensure
},
t 9-
.=.=
l..
o*
10-30-86 that the total sample contained a population of more t
highly stressed items.
A list of highly stressed hangers was prepared by S&L and issued to BCAP for their use in randomly selecting highly stressed-items.
Additional items were randomly selected from this list by the BCAP Task Force for reinspection.
For cable pan hangers, the BCAP Task Force selected ten additional items in order to assure a minimum of 60 samples were inspected.
This resulted in a total of 68 cable pan hangers which were mors highly stressed.
lt Q.13.
For each electrical construction category, how many reinspections of subjective attributes were performed as part of the CSR?
A.13.
A summary of reinspections and evaluations for subjective attributes is included in Attachment 2C (Kostal-1).
No welding was performed in the conduit and cable construction categories; as such, no subjec-I tive attributes were reinspected.
There were a total of more than 10,000 wel'ds which were inspected to 17 different subjective attributes.
Q.14.
For each construction category,,ow many subjective h
discrepancies were submitted to Sargent & Lundy for J
~
engineering evaluation?
f 10-30-86 O
2 A.14.
Tne total number of sub]ective oiscrepancies submitted to Sargent & Lundy for evaluation are snown below, expressed in terms of inspection points found to be discrepant
(" discrepancy points").
Tne figures in parentnesis snow tne percentage of oiscrepancy points wnen compared to tne entire population of inspection points for eacn construction category.
Conduit Hangers 1737 -441H (0.f t: ([0.bl[.)
D -7 9 ( 0. S 2 0 ) ( 0. S l h Cable Pans Cable Pan Hangers 1640-MM (1.22t)()/g1M Electrical Equipment Installation 6 65.4 u :1.071) 6,4 e y,Y Q.15.
How were the discrepancies associated witn tne subjec-tive attributes evaluated?
A.15.
All discrepancies found during reinspection were sub-jectea to an engineering evaluation by Sargent & Lundy to cetermine tne extent to wnigo the discrepancy affected an item's ability to perform its safety-related design function.
Evaluations were performed by comparing the oiscrepancies with design parameters, tolerances, and design margins using eitner documented engineering judgment or calculations.
Tne majority of i
the discrepancies required simple comparative evalua-tions wnien establisned tnat eacn discrepancy nad a 11 -
Kostal revised 1
l 1
~
- s
?
10-30-86 negligible effect on the items capacity to carry. load I
and no effect on its safety-related design function.
other discrepancies required unique calculations to evaluate capacity reductions in order to determine whether the item would be able to perform its safety-related design function.
Discrepancies were evaluated by first determining whether they had an effect on the capacity of the weld.
Next, if there was an effect, its magnitude was established by subtracting the defective portion when determining the overall remaining capacity of the weld.
When a weld had multiple discrepancies, a' single eval-untion was performed to determine their effect.
In the case of weld defects, for example, there were cases where more than one defect was present on a weld or in a welded connection, but only one evaluation was performed which assessed the combined effect of the individual defects.
Where necessary, adjacent welds were evaluated to determine if the design loads could be accommodated in a welded assembly.
Where this engineering evaluation was done, the adjacent welds were' reinspected to
'.7.',
verify weld quality.
.J
~... - -.
10-30-86 1
Weld discrepancies were evaluated by considering those portions of weld with defects to be totally ineffec-tive. and then calculating the weld capacity of the remaining weld.
In the case of weld size or length discrepancies, the actual weld size or length was used to determine weld capacity.
Disre'garding defective portions of welds is a conservati've approach because even the discrepant portion of the weld exhibits significant load-carrying capacity.
Q.16.
How did S&L derive the inspection point and discre-pancy point numbers which appear in your testimony, Mr. Thorsell's testimony, and Dr. Kaushal's testimony?
i j
A.16.
The process of deriving the discrepancy point and inspection point data is not merely arthimetical; to
(
some extent it is judgmental.
In 1985 Sargent & Lundy engineers and designers reviewed each CSR checklist to l
determine the number of inspection points associated I
with each attribute.
Once the number of inspection -
points for each attribute was established, each completed CSR inspection package was reviewed by Sargent & Lundy personnel familiar with the hardware design and CSR to determine the total number of inspection points for each hardware item.
The number of discrepancy points was then determined by reviewing C
n
-,,---n..v--.
-- - --- -- ~ -
.e.,
. p 10-30-86 9
each CSR observation report.
The results appear in my testimony and Mr. Thorsell's testimony and in
. C (Kaushal-3).
In 1986, S&L was again asked to derive inspection point and discrepancy point data, in two new formats
--on a per item basis and on a weld-by-weld basis.
This required S&L to review and summarize data previously created and selectively repeat the process described in the. preceding paragraph.
The results of I
this 1986 effort a e reflected in Attachment 2C (Kaushal-4).
There is a net difference of three discrepancy points between the total number of discrepancy points derived in 1985 and in 1986.
i How do these methods of engineering evaluation compare Q.17.
l with those used in the Byron QCIRP?
The methods used for engineering evaluation of discre-A.17.
i pancies for CSR are essentially the same as those l
employed in the Byron QCIRP.
The only differences to S&L my knowledge are the following two distinctions.
performed engineering evaluations only on a sample of weld discrepancies found in the electrical work for the Byron QCIRP where as for Braidwood CSR weld discrepancies, S&L evaluated each weld discrepancy.
h b.
w-
,w-.
.-,-,n-,-e-,,,,-.---.-,,,,,e,------
.,--_,,.,----,,,-.,,,,,_,.,-,,m-,
-~e..
10-30-86
+
l i
In addition, for CSR weld discrepancies involving a
)
)
crack, S&L evaluated whether the crack could propagate and, if not, credit was taken for the uncracked length of the weld segment.
This weld propagation assessment was not conducted for the Byron QCIRP weld discrepan-x, cies, where the entire weld segment was to be non-existent, whenever a crack was assumed found.
Q.18.
What,was the purpose of categorizing the reinspection discrepancies as insignificant, notable, and design significant?
A.18.
Discrepancies are installed conditions which deviate I
from the design shown on design drawings.# The sever-ity of a discrepancy can range from an insignificant
~
cosmetic flaw such as a weld splatter, which has no impact on the capacity of an item to perform its function, to a more significant flaw which could result in a component whose stress level exceeds code allowables.
Therefore, the number of discrepancies alone does not convey the actual quality of the work.
To more accurately convey the quality of the work, l
discrepancies have been categorized by level of severity.
"o v
. 1
10-30-86 j
Q.19.
What did Sargent & Lundy's engineering evaluations of the discrepancies for subjective attributes in each of c
the six construction categories show?
A.19.
Sargent & Lundy engineering evaluations of subjective attributes showed that there were no design signifi-cant subjective a'ttribute discrepancies in any elec-trical construction category.
In fact, over 984
%.Y, 00.";) of the total number of CSR subjective attribute inspection points showed no discrepancies at all.
Approximately two-thirds of tne discrepancias
' l.
evaluated were determined to be insignificant.
The remaining one-tnird resulted in notable discrepancies.
Although weld joint capacities were reduced in these cases, the discrepancies did not impair the capability of the items, wnien remained witnin code allowables.
Q.20.
What does the term " design significance" mean?
A.20.
" Design significance" relates to the ability of the structural components to perform their intended func-tion, which is to carry all design loads within code established allowable stress.
Code established allow-able stress is incorporated into the design criteria for all equipment supplied to Braidwood and in parti-cular structural steel beams.
These code established
. Kostal revised
~
3 10-30-86 allowable stresses have been dev' eloped to provide required margins of safety against failure.
Anything which affects the ability of a structural component to perform a function within the code allowable stresses has design significance.
S&L engineering evaluations, performed on L. K. Comstock identified. discrepancies, established that the stress in all components did not exceed the code allowable stress and consequently are not design significant.
Q.21.
What does the term " design margin" mean?
I A.21.
The concept of margin is one that is inherent in the engineering discipline.
Engineers designFa structure such that it is sufficiently stro.sg to withstand the expected forces and stresses with spare or extra strength to account for uncertainties and contingen-I cies.
This extra strength is called cargin.
" Design margin" is the difference b.scween code allow-able stress and actual stress.
Engineers maintain the presence of design margins by ensuring that actual stress is less than code-allowable stress.
For example, connections are designed in groups rather than individually.
l l
O
- a-10-30-86 The most highly stressed connection is designed to be within code-allowable stresses; therefore, all other connections within the group, which are not as highly stressed, have even greater design margins.
Thus, the actual stresses for most connections in the example will be less than those allowed by the applicable code.
There is a second margin in the structural design of connections.
This is the margin that code writers put into the design process in the form of the difference between code-allowable stresses and the failure of a 5
component.
Code writers typically attempt to obtain a margin of approximately two when they write a code.
This means that a structure designed to a' code could carry approximately twice the design load and not i>
l' fail.
These code required margins were not reduced by any discrepancy identified in the CSR.
Q.22.
Please discuss the " notable" discrepancies' identified in each construction category.
A.22.
Notable subjective discrepancies were identified in the construction categories of cable pans, cable pan hangers, conduit hangers.and electrical equipment installation.
There were no notable subjective
<?.
discrepancies associated with cable or conduit
%/
_ i
=
_. =..:..a 10-30-86
~
construction categories.
Tne to11owing is a review of tne types of notaole oiscrepancies founo in eacn of these categories:
Conduit Hangers
(
For the conduit nanger population 28,300 reinspection
'78 points were reviewed; out of tnis population 4e nota-ble discrepancies were identified.
Tnis represents j
less tnan 0.3% of tne entire population.
These nota '
ble discrepancies were related to weld size and weld lengtn.
In all cases the actual stresses were within code allowables.
The minimum design margin remaining
/*
I after tnis evaluation was more tnan 30 percent above l
code allowables.
Tne average design margin remaining.
e y
for all welds witn ciscrepancies was 900 percent above code allowables.
Caole Pans For tne cable pan population 9400 reinspection points 11 were reviewed; out of tnis population HMk notable dis-crepancies were identified.
Tnis represents approxi-
,e mately 0.2% of the entire population.
Tne notable discrepancies were related to cable pan hola down welds.
Six out of 605 cable pan nold down welds i Kostal revised
- y- - -
3 --...
w.-._m
_y-., - - - _ _,,, -, -,,.,,
.e,-._.__w_,._,-.--____--w-,--,,,,,,,,,._-,.m.--._-r-v.,
=:...
.._:==-
10-30-86 inspected by BCAP were missing.
In all cases where a 3
i-missing weld was observed an adjacent weld was present to carry the design load.
Remainder of the notable weld discrepancies were found to be underlength welds.
There were no decign significant discrepancies.
The minimum design margin remaining after this evaluation s
was more than 30 percent above code allowables.
The 4,
average design margin remaining for all welds with discrepancies was 300 percent above code allowables.
4.
Cable Pan Hangers For the cable pan hanger population, 129,000 reinspec-52/1 tion points were reviewed; out of this population-464 notable discrepancies were identified.
This repre-sents less than C 4'/ eof the entire population.
These notable discrepancies were mainly related to weld size, missing portions of welds and weld length.
In all cases the actual stresses were within code' allow-ables.
The minimum design margin remaining after V
evaluation was more than 5 percent above code allow-ables.
The average design margin remaining for all
/
welds with discrepancies was 800 percent above code allowables.
l l
l s
. )
~
=
10-3,0-86 1
Electrical Equiptent Installation For the electrical equipment installation population, 37,700 reinspection points were identified; out of 111 -
tnis population est notable discrepancies were icenti-fled.
This represents approximately of tne I
entire population. 'Tnese notable discrepancies were related to weld size, location, concavity, and lengtn.
In all cases the actual stresses were within code allowables. The minimum design margin remaining after evaluation was more than 10 percent above code allowables.
Tne average design margin remaining for s
s all welos witn discrepancies was 500 percent above code allowables.
i Q.23.
Wnat conclusion about the quality of the reinspected i
l L. K. Comstoon worn can,you draw from tne evaluation of subjective attribute discrepancies identified in the CSR?
A.23.
It is my professional judgment that the quality of the L. K. Comstocx worn on the Braidwood Station is ade-quate.
My judgment is based on two significant elements.
First none of the discrepancies identified witn respect to L. K. ComstoCK work had design signi-ficance.
Secono, tne existence.ot conservative load-ings and assumptions used in tne design of Braidwood Kostal revised !
10-30-86
-4 Station and tne margins innerent in tnat design provioes tne capacity to accommodate tne type of discrepancies identitieo in tne CSR Program.
Tnese types of discrepancies are those that I would have l
expected to fino.
Q.24.
In assessing toe quality of L. K. Comstocn's work, I
wnat significance, if any, do you attaen to tne rate t
of subjective and objective discrepancies identified in tne CSR Program?
A.24.
In my judgement tne results of tne electrical portions of the CSR Program indicated that the rate of discrepancies is extremely low.
Over 98% (9 8.7,4 ) of all inspection points were found to be correct.
Less tnan 2% were found discrepant.
Wnen evaluating tnese bf */o discrepancies tne vast majority o.f tnem +Mt) were determined to be insignificant; meaning, they had no effect or minimal effect on the capacity of tne component to perform its function.
If I were to
[
t examine tne data on a welo or individual item basis, i
the conclusions would still be the same, that is the vast majority or discrepancies associatea witn l
I individual items or welds are insignificant.
Ninety-nine point six percent (99.6%) of all electrical inspection points were found to be eitner correct or l
l Kostal revised '
_._. ~ _
10-30-86 naving' insignificant discrepancies wnien nave little
~
or no effect on tne capacity of tne component to perform its function.
Tne remaining notable discrepancies were evaluated and none were found to be design significant.
Since an extremely low percentage of inspection points were 0.4 7e' found to be notably discrepant (approximately,et) the 4
results suggest to me that the inspectors were ff.nding tne important discrepancies wnicn coula effect tne capacity of the_ component and tnat tnese discrepancies were corrected during the construction process.
In determining tne quality of tne electrical worn tnese percentages were less important :nsn other considerations.
Tney may be useful, however, in I
evaluating tne quality of tne individual inspector's performance.
Tne important consideration, from my perspective, in determining the adequacy of construc-tion, relates to tne impact of discrepancies on t
l, i
reducing the capacity of tne component.
As previously stated the types of discrepancies found wnien could effect tne capacity of tne component were small.
Tne design margin remaining after accounting for notable oiscrepancies were always sufficient to assure electrical components would perform tneir intended
. Kostal revised
~.
lf '..
10-30-86 function and remain within code allowables.
These results indicate to me that the work performed in the electrical area is adequate.
Q.25.
Please describe in more detail the conservatisms in the design of Braidwood which are relevant in assessing the adequacy of L. K. Comstock's work.
l-A.25.
Various conservatisms exist in the design process which provide considerable design margin.
For purpose of discussion, these conservatisms may be summarized into three broad categories:
conserva'tisms attributed l
to code requirements, seismic design conservatisms and 4 :
conservatisms attributed to practical desi.gn/construc-tion implementation.
Conservatism is applied by the design margin that code writers put into the design process in the form of i
allowable stresses.
The code writers typically i
attempt to obtain a margin of approximately two when they write the code.
This means that a structure des l%aed to code could carry approximately twice the design load and not fail.
It should be noted that in our design and evaluation of discrepancies we do not I
encroach on code margins.
24 -
-r.
,----e y
em y---
-y p-,-,,w,,
, - - - -,,+
---r-a,----~---,,m-wen,-----
=.
--~
9 10-30-86
~
The welding code also provides margin indirectly by requiring minimum size and lengths for welds.
An example of this would be the case of welding a 1/2 inch thick gusset plate connection.
Based on design calculations, a 1/8 inch fillet weld may be all that is required.
However, the code requires that welds be a minimum size based on the thickness of the base metal being welded.
j Therefore, even though only a 1/8 inch weld is l
required by design, a 3/16 inch weld has been' specified as required by the code for *1/2 inch thick l
material.
In the second category, seismic design conservatisms have been established in' order to account for seismic forces in a practical manner.
Cable pan hangers are conservatively designed by performing seismic analysis j
for each individual hange.
This individual analysis e
isolates the hanger from the adjacent supports.
The actual cable pan support system consists of cable trays supported by a series of hangers.
This permits load redistribution to adjacent hangers.
A refined computer analysis can be performed for the cable pan support system consisting of the trays and the series The analysis results'in significantly of hangers.
N 9. _..,
10-30-86 lower stresses and provides for additional design margin.
As part of the evaluations performed for the Byron licensing hearings, three cable pan support assemblies were an'alyzed using these more refined techniques.
The re'ault of that analysis showed that even with the
~
inclusion of the most discrepant weld and the corres-ponding impact on the support assemblies, none of the connections or members exceeded the code allowables, even when loaded to twice the design load.
The time and costs involved in the preparation'and application of a more refined computer analysis make this approach less practical.
In addition to the conservatism of enveloped seismic analysis as compared to systems dynamic analysis, the 1
conduit, conduit hangers, cable pan and electrical I -
equipment mountings have all been designed based on peak seismic accelerations.
A more refined analysis in which the seismic acceleration is selected based on the actual component frequency results in a lower seismic design force.
Implementation of any of these refined techniques of seismic analysis in the design phase are less practi
,,e
-3 1
10-30-86 a
e cal base'd on cost economics.
Since a simplified
. conservative method was used in the original design, I
all subsequent design changes can be evaluated with minimal consideration.
If refined analysis had been used at the outset of design, subsequent design changes would have required comprehensive evaluation.
?
Many conservatisms can be attributed to the third 4
category, practical design / construction implementa-tion.
The design margins which exist in the practical design / construction process generally relate to the standardization of components and their connections.
+
This standardization results in more efficient design
., ~
and more economical construction.
For egample, in the area of conduit hanger supports, a series of standards have been developed which are utilized by designers to These standards pick the various component hardware.
y are based on typical member sizes, in which a member size can carry a range of design load.
The purpose of I
standardizing the design is to limit the number of l'
components thereby allowing for ease in construction of the conduit hanger assembly.
In the process of i
standardizing the designs, the maximum allowable load When a is determined for each standard component.
i designer chooses a component he compares the actual design load with the maximum standard load, and in all si
, i
.= -
10-30-86 s.
cases the, standard hanger will have margins remaining.
This process provides,for a more a
economical design as well as for a more economical construction of the components.
It also provides for additional design margins to accommodate construc-tion discrepancies which may occur.
Cable trays are routed and si~ zed prior to detailed cable routing.
The cable trays and their supports are then designed to accommodate the weight of the maximum namber of cables that can fit into the cable trays.
Most of the cable trays and cable tray supports at Braidwood are not loaded as heavily as has been assumed in design and thus most have additional margin.
i l
Similarly, most of the conduits and conduit supports have been designed to accommodate the maximum number of cables that will fit into a conduit.
Since most of
.the conduits are not filled with the maximum number of cables, most of the conduits and conduit supports are not loaded as heavily as has been assumed in design and thus most have additional margin.
?
l
+
+
-+..<
^
L.
.L'.^.
^
10-30-86
'/
The materials used for cable trays, cable tray supports and conduit supports have been specified by d
the engineer to meet minimum strength requirements.
In order to consistently meet these minimum require-ments, manufacturers typically set for themselves higher strength requirements to avoid having to scrap substandard material.
The actual test results indicate that Braidwood electrical hanger materials l
l have an average strength equal to 20% more than the i
minimum strength assumed in the design.
All of the conservatisms discussed in my testimony are applicable to the component design and therefore relate to all the subjective evaluations and in a similar fashion to the objective evaluations addressed i
in Mr.
Thorsell's testimony, i
{
The conservatisms as related in the preceding testi-These mony overlap and are often t'ines cumulative.
conservatisms provide a basis for my judgment l
regarding the adequacy of the work performed by L. K.
l Comstock.
In discussing the results of the Byron QCIRP in its Q.26.
October 16, 1984 supplemental initial decision, the I
Byron Licensing Board observed:
s Ns/
i l
_=
,,1 10-30-86
".The, absence of any design-significant discrep-ancies leads to the conclusion that despite the existence of discrepancies, the original inspectors had sufficient competence not to overlook design. significant construction defects.
.i This conclusion, however, has limited value as a demonstration of inspector competence.
The greater the sum of the design margins, the less it challenges the inspector's competence to discover design-significant defects."
LBP-84-41, 20 NRC 1203, 1212-13.
Would you agree o'r disagree with this statement if applied to the results
- l of the BCAP CSR for electrical construction categories?
A.26.
I agree with the Byron Licensing Board's statement.
a broader statement can be made regarding the In fact, elements required to assure that a component item can 4
i carry its intended design loads.
These elements i
4 include the plant design, the quality of construction and the quality control inspection.
The engineering process includes various elements of r
conservatisms in the development of an individual j
l component design.
These conservatisrs first include the code which has been established in most cases to maintain a design margin of two.
This means that a i
component can accommodate twice its design load prior to failure.
In addition, the engineer standardizes 3
i w
l I
' l l
l i
I
~
10-30-86
-e various. component designs for maximum loading condi-tions and maximum installation tolerances which could The implementation of this standardization occur.
results in very conservatively designed individual components which assures additional margins of safety.
The second element is the quality of construction.
The crafts, through their experience and training, are familiar with the construction of these various components.
This generally results in components being installed to an adequate stanoard to assure the components' ability to perform their. intended func-tion.
s The third element is the quality control inspection performed in the field.- This quality control inspec-tion provides an additional level of review to assure All that the components are constructed adequately.
three of these elements act together to assure that a component item can carry its intended design loads in
~'
The larger the design margins a safe manner.
the better a initially provided by the engineer, component can accommodate deficiencies arising out of In a like the construction installation process.
y manner, the better the construction quality and the more conservative the design, the less challenge there
', c,
- (,r
A a
(
s 10-30-86
~
In the electrical is for quality control inspection.
construction categories at Braidwood, these three elements have acted t6gether to assure that no design significant discrepancies have been found.
Have the conservatisms in design of the electrical Q.27.
construction categories changed since idue CSR cut-off t
I date of June 30, 19847 No, prior to June 1984 and subsequent to that time the A.27.
engineering design methods have remained constant; the conservatisms discussed in this testi-therefore, mony are applicable to LKC prior to and subsequent to June 30, 1984.
Does this complete your testimony?
Q.28.
A.28.
Yes.
f.
\\J)
\\
ATTACHMENT 2C (KOSTAL-1) Revised
SUMMARY
OF SUBJECTIVE DISCREPANCIES ELECTRICAL CONSTRUCTION CATEGORIES PERCENT r
PERCENT PERCENT DESIGN i
CONSTRUCTION REINSPECTION PERCENT INSIGNIFICANT NOTABLE SIGNIFICANT ELECTRICAL CATEGORY POINTS DISCREPANCIES DISCREPANCIES DISCREPANCIES DISCREPANCIE Conduits 0
0 0
0 O
Conduit Hangers 28,300 0 769 c.6: )
OvMr o.%
er2+ 0. '2 %
0 Cables 0
0 0
0 0
4res 0.Sl 0.59 erE4 0.1.*2.
O Cable Pans 9,400 Cable Pan W ).4'l
-eve 6 1.01 N o.41 0
j Hanger 129,000 Electrical Equipment 4ve7 1.60 em O.67 W O.S$
0 Instal,lation 37,700 W l.8o 9;I21 0.34 W 0.4G 0,
s TOTAL 204,400 R
1
.