ML19241C065
| ML19241C065 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/05/1979 |
| From: | Wilson T BABCOCK & WILCOX CO. |
| To: | Oelkers E BABCOCK & WILCOX CO. |
| References | |
| NUDOCS 7907260535 | |
| Download: ML19241C065 (2) | |
Text
.
r8 V}l
/
THE BASCCCX & WILCOX COMPANY cc:
J. a. cicude sas
'0WER GENERATION GROU?
G ^
M*7*
Io l
P. E. Ma ola J. A. eize Cc.I.v-..w, 5.fS..v. a_rv.v%r.,. 5 C e.s.. ~c. s,,....
H. D.
..a, ea
. a 2-Frcm fk.m.] bj%
l.*/ f j'[!)
C. T. Rembcugh R. V. Dey.a:s T.
.f..
"I.'.a C"*,
a~"u<..'d.
..G.
.. s C ".. *. ~.s' C LS "t.". "..
c ios 66:.:
Cust.
File No.
TMI-2 or Ref.
Suej.
7,g g._, A,,.D
~L. C'.... A.".'.. O ~ 6..r =.."." C ^ l.'. '...e, r%' '
"a '..". G 0a:e
.s -
r T
^
o
^
L.,,.a,S 0e : :.:. a a.:.2 Ra.,....,..,
AP,...u,, 3 o.,/ 9 g
... a. :...
gn.......,.................,..........,...,.
Ackneviedee mnt:
I have gathered the folleving infor_atica :h ugh c=nve:sations wi:h several individuals having experience vi:h ther:ccouples : hat have been subjected to extremely high te=peratures.
I an indeb:ed to P. E. Ma:cla, R. 3. Stoudt, and Tc Kollie of CFE for their help and cocpera:1ca.
Descrietien of the ? chle::
The incere ther:ccouple design a: the "C-2 core is a grounded junc:ica Chrecel-Alu=el detector.
~he wire dia:eters are appr:xima:ely 10 tils, :he sheath is 61 mila OD Inc:nel cubing, cad the insulation ca:erial is A1.0..
I have gathered infor:ation en reliabili:7 and :acertainty of this :/p$ ef ther:ccouple af ter being subjected to extremely hi;h :e:peratures (> 2000*7).
The problem of gress failure (cpen circuit, shes:h failure, new function or other failure) would be indica:ed by no reading or extremely low, erratic reading. No detec: ors are giving indica ica of g ccs failure.
Given the survival at presen: conditions, the prospec: f: centinued Operation is excellent.
The primary questicas are :he foll: wing:
1.
Are the readings accurate?
2.
Are the errors in :he readings cens12:en: vi:h the hypothesi:ed scencrio of the transient?
Decalibration Phenc':enen:
MNN A pheno =enca cbserved by Or. Kollia in thermoccupies having e'eperienced extre:21f high te=peratures is a dete:1 ora:ed s:ste in which the :her:ccouple gives a s:able but inaccura:e :eading.
~he effect is called decalibra:ica.
- he decalibratica error is randes bu: does follev sc=a : rends.
~'he prica:y dependencies are the following:
1.
Shcath dia:eter The large: the detector, the less susceptible ::
decalib:a:icn.
2.
Sheath =aterial Inconel is be::er than s:ainicss.
~he decalibration 79072 6o3ss o
/
Z. CELKIRS 2
APRIL 5, 1979 error for Incenel is usually nega:ive.
Measured temperature is lever :han actual.
3.
Te:perature Decalibration error increases with terperature.
(10*C roughly dcubles the reac:1ca race of :hu
=cchanis=.)
4.
Ti=e of cxposure Decalibratica error is rcughly linear with e posure ti=c.
5.
Te=perature profile Decalibratica eccurs along the length of the ther:o-couple.
Hence, the error depenca ca the profile cf the elevated te=perature and on the profile of a subsequen: cessurement wi:h the :her occuple.
Other dependencies which are not kncun specifically but are expec:ed :o be s=all are insulation :sterial, length of de:ector, and wire dia eter.
Experience with s= aller diameter detectors (20 tils CD) but si=ilar in other charac: eristics (grounded junction Chrenel/Alueel :hertocouple vi:h Incenel sheath) shev essentially :ero probabili:y of survival a: 1100*C for 10 :s 100 hrs. withou: a =easurable decalibratien. The decalibra:ica errc: : ands :o be n astive (r:ading lever than actual) for this tes t; hovcver, cae sas;le gave a posi:ive error.
The magnitude for vers: case error (:erceberad' was -50*C.
Larger decalibratica errors would be expec:cd for hi;her te pe:atures.
The primary :echanism for decalibratica is ri;ratica of cens:ituents pri=aril chrc iu: in the de:cc:cr thrcugh '.apor phase ::ansfer. The chrczal lead leses chrc iu: and the alueel lead gains chrc 1u.
Selfab tes:s indicating lit:le decalibratien for 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> peried a: 7000*? are net necessarily cenclusive since the te=perature gradiant of the high temperature state does not simulate the transient envir:n=ent nor did the ces: cenditica simulate the prasent enviren-cent.
Cenclusions:
If it can be deter ined that the ther:ccouple is reading accurately ncv (fer exa:ple, if 1: agrees vich outle: RID's), : bis implies that i:a readings are believabic throughcut the ::ansient.
If thre is reason :o believe a ther:occuple is decalibrated,
- here is a hi;h probability that actual t2:pera:ure is greater than 1:s indication.
- Hence, the cluster of high readings shculd not be disregarded.
Since a high percentage of detec: ors survived :he :::nsient, the naximu: tempera-ture did not attain a level for videspread failure.
Recc= endations:
1.
Perform testing simula:ing trancient ccadi:i:ns :o deter =ine the te=pera:ure thresheld for gross f ailure in TMI-2 design de:cc: ors.
2.
Perfor: testing to quan:ify =agni:ude and direc:1cn of decalibra:ica errors for a range of te=peratures si=ula:ing reac:c: environment.
M9190 TL'4:se
_ _,_g