ML19322C744
| ML19322C744 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 11/21/1979 |
| From: | Seminara J LOCKHEED MISSILES & SPACE CO., INC. |
| To: | Chipman G NRC - NRC THREE MILE ISLAND TASK FORCE |
| References | |
| TASK-TF, TASK-TMR NUDOCS 8001220706 | |
| Download: ML19322C744 (6) | |
Text
, -.
t e
L) s *
?
Cx0C ? VJ MISSILES
~
8c E5 P /\\ C; E C: C) h4 P A N Y.
l l F4 C:.
I November 21, 1979
Dear Gordon:
Attached is the detailed critique of the ESSEX first draft.
In sunmary:
o Essex did a good job.of evaluating the TMI-2 control roer in terns of causal elements in the nan-nachine aspe::s.
That work is sound and should be expanded with good photodocunentation tha; is already available.
The comparison between TMI-2 and the other two plants is far too shallow o
to draw meaningful conclusions of the type nade.
A more representative selection of comparison plants is needed.
Use the EPRI :!?-30? report and the Zion evaluation to bolster ycur conclusiens.
I think the wrong con-clusions have been drawn.
The report seems to say that..!MI-2 is significantly worse than the other two plants and hence we should haveLexpected an accident. My view is that none of the p:wer plant centrol ro::s (the,, twenty that I have seen) is properly designed in terns of the nan-na: hine interfa:e considering the nature of the potentially disastrous censequences. We should be concluding that a more fermal and systematic approach to design-ing for the operator is needed across NRC/A-E/NSSS/ Utility reaLns of responsibility.
s
- The NRC comes off as an innocent bystander. What has the NRC done to safeguard the public in terns of the hunan factors aspects of the control Are there any human factors people (with bone fide credentials) roen.
on the staff? Has a Branch been created to deal with the development of human factors data, application of EF yardsticks in the review process, raise the awareness level of inspectors for EF problems, etc.? The answer is N0!
When EPRI NP-300 surf aced several years ago, why didn't the NRC take some p sitive steps in the EF area? In ry conta:ts with NR folks there has been no understanding of EF.
Io the people I have met EF is something of a " black art."
o The report leaves ne in a quandry as to what :: do with respect to plants already in existance, those on the drawing boards, and those of the future.
Should everyone go to CC and Duke and try to copy their style of nanagement? Do we shut down all the existing plants until the control roons are redesigned? Can the CRs be enhanced in some significant way?
Do we make some quick fixes in added displays, training, procedures and say the problem has gone away?
I will now review your draft of the report over the holidays.
My best wishes for a most p:.,.
' Turkey Day.
8001220 1
v(___
i A SU B 5IDI A RY oF LoCKHEED AIR CR A FT Co RPo R ATioN 1111 LOCK H EE D W AY e SUNN YV AL E. C ALIFo RNI A e 94088
,/
e
W CO'!SNTS ON FIRST DRAFT SU~EITTED EY ESSET. DATF.D CCTO E R 22, 1979 PACE PARAF:RAPH REMARE 5
1 Hunan f acters is the scie. hee and techno1cgy cf applying behavioral and physiological...
7 2
Sane as above 11 3
This pararraph nay be misleading. What have been identified as applicable human engineering criteria were not generally recccnited as belenring to any systenatic fccus on the man-nachine inter-face.
.H6w many interviews acninistered?
Preakdown of 1?
2 inte rviewees?
1L 3
In discussing " Industry Standards" you might note that no bena fide human facters specialists were invc1ved in fernulating these standards.
2 What w as the rcle cf ne ':RC in anplying the SRP.
To ny knerledge the NRC had no specialists id Eunan factors and ne yardsticks fer assessire hunan engir.eering aspects of control rocn design If the KRC had an ir. pact in revieving man-nachine aspects of control rcen designs, the twenty designers that I interviewed were not aware of it.
~O.
2 I don't believe Calvert Cliffs and Oconee in contrast to TMI-2 constitute a representative sample en which to base cenparisens. Teu should factor in the five plants covered in ro"? PF-309. It w:uldn't hurt to visit several otners as time and noney allcw.
33-62 Connents on these paces were covered during cur revenber 15-16 neetings
-3 7
It nicht also help te :. ave preftssicnal hunan engineering /ineastrial cesign assistance in :ne analyses, design and review precesses.
In nany cases the EP/IP tean assunes the responsibilit::
for the neckup and all interratien en the CR design effort. Perhaps this is a nore critical factor than the distinction between a "systens engineering" vs a " discipline-recuirenents" approach.
Lock what CC did with mirrer imaging, backpanels, annunciator panels, etc. despite the so-called systems engineering approach.
I would put my noney on a " discipline-recuirements" type approach if the Lead Engineer had a stong HF/ID conponent en the prcject desien tean vs. a " systems engine er-ing" approach with no fonnal human f actors or ID.
POE PARA.
PFARK 6h 1
Incorrect assunption that the utility will do a cetter job than A/E.
The worst plant that I havE'seen (not T:2-2) had the createst utilit r participation it. the design process.
The utilities know less acout E? than the MSSS vendors but client preference deninates CR design ir nany cases and destrys good design features proposed oy tne NSSS vendors.
I Please read my desian practices survey. We need te educate the utilities red the NRF, so that good H? design is cemanded fren designers.
6L 2
- 'ho says the A-E has little responsibility for operational conce rns ? Nhy can't an A-E apply the "s.. stens apprcach"?
- his is tctally erronecus.
fh 2
What is wrcng with the CR rcflecting one design philosophy (assuning that it is the correct philoscphy)?
This whole pare reflect: tne nction that the A-Es are the villains in the piece.
Have,"ou looked at CRs designed by BF'i That Ernenburg sinulater design is nething to write hone aoout.
65 (3)
Don't know what you mean by extraordinary documentatien.
fr
?
Don't oelieve that you ever nake the ecse fer one ferm of ncnacerent vs ancther.
7 95
c)
While CC iTc'etter than T:2-2, there are too many glaring-deficiencies te hold it up as a paracen of virtue.
si
- 7 The writer has a tendener to acuate a ncekup with Hunan enginee ring. While HF specialists tend to use mockups for design and evaluation, the mockup.doen't in and of itself nake for rood human f actors engineering. There was no human engineering in any formal sense in any of the three plants ceing discussed.
66 3
Again you are saying that the client influences plant design in a positive way.
This is not necessarily so.
ou can have a stubborn, ignorant client who places v
CRTs on standup consoles because it is the utility's chilesephr the keep the operatcrs on their feet at all L es.
Sene utilities insist on placire contrels at floor level within easr kick range, etc. etc.
if (d)
'Cr single cut TM -? nanarenent fer lack of consideratien to hunan er-ineerin,- f actors.
The p :clen is epecenic.
Don't understand what ou are saying in last sentence.
'O (g)
Even before TMI-2, the standard was a joke in the industry.
There were no H? psople on the connittee that put the standard toFether.
1
' P AFE PARA.
REMARKS 71 2
How did you establish that all three control rooms were designed.for one man operation? While many tech specs call for one nan operation, the utilities generally have two men on hand per shift.
The basis for the one-nan determination is oescure.
It is not cased on work load analyses during different operational modes.
In fact many operations call for the one operator to be in two places at once.
73 2
You haven't nade the case.
CC and Oconeee should have been compared very specifically in terms of the causal elenents uncovered in the TMI-2 accident.
76 (e) lack of rench and visibility were not the problems at TMI-2. That peint is being made? See acove remark.
In my view all three power plants reviewed are not eptimized from a hunan factors standpoint. I can envisien a serious accident happening at CC or Oconee.
The fr.ct that the latter may be slightly better overall than IMI-2 doe:r't reassure me.
The point is that all three plants ellow for human error and not that two are good and the third is peer.
(a)
Please recheck the ved or creen status of annunciator lights _to_ see if the code has any meaning that we might accept as useful. Vest centrol recms previde red or green indicator lights fer valve position. Are these' red /creen notations consistent with the annunciator system coding? I suspect that CC is contradicting your rule #2 below on this page.
77 Fig. 16 I don't believe you found lh neanings for red at TMI-2.
These proccoly can be subgrcuped into two or three meanings.
76 1
Green means no-flow of fluid, electricity, etc. and hence is consistent 78 2
Are vou really happy with the coding of any of the annunciator systems observed?
7f 3
The real point is that deviant conditions are not manifest at a glance as wculd oe the case with the Green Soard concept.
There is no loric built into the disclay systen se that the operator is told that a velve is open but should be c1csed, etc.
~6 (b)
The operator nust memorize "I$ mediate Actions" and should always have available procedures for guidance
'if needed.
79 1.
pidelity ---What the writer thinks is re asonable to assume here differs from my experience.
81 1
State-of-the-art for utilities but not for military 3
settings
PAGE P AP.A.
REMAES 61 2
Are you saving that written proce[u'res are a thing of the past?
$7 (h)
I do:0t agree with tha inplicatiens of the summary, namely, that TYI-2 was so inferimr to CC and Oconee that we snouldn't be surprised that an accident happened at TMI.
The fact is that TMI-2 is quite typical of power plants in general.
We should be highlighting the fact that none of the three plants reviewed measure up to the degree of attention to human factors recuired of nuclear power.
The comparison between the three plants was too superficial to oe mean-ingful.
68 2
Catercrical disrecard is not too strong for me.
- However, I wcuic no: cirect snis language solely at IEI.
90-9h Most of these observations of deficiencies are noted in EPEI EP-309.
You should draw from earlier work to substantiate the case that nuclear power plant control roons in general do not live up to hun.an engineering standards and that this cannot be tolerated in so vital and donzerous a case.
110 2
You are being unfair to Met. Ed.
There is a general ignorance in the industry regarding procedures.
In any case, it doesn't necessarily follow that because operators are not in the loop to update procedures then Met.. Ed. is clearly downgrading operators.
129' 3
Fuzzy statement.
We could rely on training if it were adequate and the control room was optimized.
112 2
Very weak statement on selection.
Is it true that most operators are former Navy? I know that many are cut am not sure that Most are.
T::I is a Navy ship out many
~
plants are not.
There is a strong need for valid selection techniques for operators.
There isn't even a decent 4
job analysis of the operator's role. There is no reason to believe that the NRC licensing process is a valid screening approach. In f act, there are indications to the Contrar?.
12 1 3
Sone words might be said about nerctive transfer of train-ing, the value of standardication in control reca designs, the use cf identical procedures in the simulator and the operational control roon.
126 2
None of the present simulators provide performsnce scores.
l FURI has funded General Physics to come up with an I
objective performance measurement approach using the T7A simulators.
127 2
These generalizations apply to the whole industry and not just TMI
PAGE PARA.
REM E S 131 2
" Thoughtful" should bet systematic huEah factors analysis.
111 h.2 Why should trainers be involved in the writing or the revision of procedures? Have you checked the cualifications of trainers?
I would prefer that operators reviewed or initiated procedures given some guidance in the fornatting of procedures.
132 3-h The approach to training paralleled the industry's approach to information display.
This is not a uniquely TIG problem or deficiency.
133 1
Same as above 13h 2
Is there sone fornula based on evidence for determining what percentage of the training should be sinulation?
13h 5
If your last two sentences are true, then there are many disasters on the horizon. You are indicting the whole industry.
138 3
The report ends on a ncte of total dispair.
Do we shut down all operational centrol rooms because proposed quick fixes fail to address fundamental problems? Is there a way out of this pit short of halting power generation until new control roons are developed? If TMI-E can't ce fixed or patched up, neither can M6st~of today's CRs.
9 e
i