U.S. Nuclear Reeulatorv Commission Operations Center Event... 09/10/2015 Part 21 (PAR) Event#

U.S. Nuclear Reeulatorv Commission Operations Center Event Renort
09/10/2015
Event#
Part 21 (PAR)
Notification Date / Time: 08/11/2015 09:55
Event Date / Time: 07/28/2015
Last Modification: 09/10/2015
Rep Org: ROTORK CONTROLS
Supplier: ROTORK CONTROLS
Region:
City:
County:
State:
Paop I
Docket #:
Agreement State:
License #:
1
ROCHESTER
51309.
(EDT)
(EDT)
Yes
NY
NRC Notified by:
HQ Ops Officer:
Emergency Class:
10 CFR Section:
21.21(a)(2)
PATRICK SHAW
MARK ABRAMOVITZ
NON EMERGENCY
INTERIM EVAL OF DEVIATION
PART-21 NOTIFICATION
-
Notifications:
RAY POWELL
ANTHONY MASTERS
ROBERT ORLIKOWSKI
BOB HAGAR
PART 21/50.55 REACTORS
Ri1DO
R2DO
R30O
R4DO
EMAIL
MICRO SWITCH INTERMITTENT VARIATION IN RESISTANCE
The following report was received via fax:
'On June 4, 2015, Rotork Controls Inc. opened a formal Part 21 [10 CFR 21.21] investigation into a self-identified
anomaly relating to a basic micro switch
-
Pt No N69-921, description 'V12'. The anomaly is intermittent variation
in electrical contact resistance and was first observed during the factory acceptance test of a Rotork safety related
NA Range Electric Actuator; also referred to as an electric Valve Operator.
"Rotork and the switch manufacturer are currently characterizing switch population contact resistance to establish
whether an unsafe condition could exist as defined under 10 CFR 21.
* **
UPDATE AT 0958 EDT ON 9/10/2015 FROM PATRICK SHAW TO MARK ABRAMOVITZ
** *
The following report was received via e-mail:
"Rotork has concluded the investigation and based on test data determine an unsafe condition can exist as
defined under 10OCFR21. Testing has established .that a small percentage of the switches supplied against Rotork
purchase order P0116932 may exhibit an open circuit condition when operated. A slightly larger percentage of the
population may exhibit high contact resistance ranging from 0.5 Ohm to open circuit. The anomaly affects the
normally open and normally closed contacts equally. Contact resistance anomalies are random, not permanent,
and appear to be unaffected by accumulated cycles. Additionally when a switch is oPerated its electrical state
remains constant until the next operation.
"Probabilities have been established by resistance measurements and by dynamic testing using voltages
09/10/2015
U.S. Nuclear Regulatory Commission Operations Center Event Report
Page 2
•representative of customer applications. The dynamic evaluation predicts an open circuit probability of 0.0008%
per operation. To date, no switch failures have been reported from any customer sites. Below are listed all orders
provided to customers containing V12 switches from the defect batch. Report ER791 provides full details of
Rotork's technical investigation."
Notified the R1DO (Gray), R2DO (Shaeffer), R3DO (Pelke), R4DO (Drake), and Part-21 Group (via e-mail).
rotor
w
Cotro~s
Rotork Controls, Inc.
675 Mile Crossing Blvd.
Rochester, New York
14624
U.S. Nuclcar Regulatory Commission,
Washington, DC 20555-0001
tel: +1585 247 2304
lax: +1585 247 2308
www.rotork.comn
if~ookcr
From: Patrick A. Shaw, P.E
To:
NRC Operations Center
Date: September 9(11, 2015
Subject: Rotork Controls Inc. Part 21 Notification, V12 safety related micro switch, Pt No N69-921
Dear Sir/Madame,
On June 4, 2015 Rotork Controls Inc. opened a formal Pt 2l investigation into a self-identified anomaly
relating to a Basic micro switch - Pt No N69-921, description "V12". The anomaly is intermittent
variation in electrical contact resistance and was first observed during a factory acceptance test of a
Rotork safety related NA range electric actuator; also referred to as an electric Valve Operator.
On July 28 2015 Rotork Controls Inc. wrote to the U.S.NRC requesting a 40 day extension to the 60
day investigation period permitted under 10CFR2 1. The extension was granted and the investigation
expiration date was extended to September 12 th 2015:
Rotork have concluded the investigation and based on test data determine an unsafe condition can exist
as defined under I10CFR21. Testing has established that a small percentage of the switches supplied
against Rotork purchase order P0116932 may exhibit an open circuit condition when operated. A
slightly larger percentage of the population may exhibit high contact resistance ranging from 0.5 Ohm
to open circuit. The anomaly affects the normally open and normally closed contacts equally. Contact
resistance anomalies are random, not permanent, and appear to be unaffected by accumulated cycles.
Additionally when a switch is operated its electrical state remains constant until the next operation.
Probabilities have been established by resistance measurements and by dynamic testing using voltages
representative of customer applications. The dynamic evaluation predicts an open circuit probability
of 0.0008% per operation. To date, no switch failures have been reported from any customer sites.
Below are listed all orders provided to customers containing V12 switches from the defect batch.
Report ER791 provides full details of Rotork's technical investigation.
Page 1 of 5
rotorn
Controls
Rotork Controls Inc. have provided ER791I so that U.S. Licensees may evaluate each application wherc
Rotork equipment is affected by this Part 21 notification.
Below this letter is a summary pursuant to 10CFR Part 21 .21(d)(4), a list of all actuators & spare
components containing V12 switches from P0116932, and a copy of ER791.
Sincerely
Patrick A. Shaw, P.Ew
Quality Assurance Manager
Tel (585) 770-1019
Patrick.Shaw @rotork.com
•
President
Page 2 of 5
/
rotorw
Controls
Required informlation as per 10CFR Part 21.21(d)(4) follows:
(i)
Name and address of the individual or individuals informing the Commission.
Patrick A. Shaw, P.E.
Rotork Controls Inc,
675 Mile Crossing Blvd,
Rochester, New York
14624
(ii)
Identificationof the facility, the activity, or the basic component suppliedfor such facility
or such activity within the United States which fails to comply or contains a defect.
Rotork part number N69-921I micro-switch V12/3252 (RSI104) manufactured by Johnson
Electric (formerly Burgess) during 2013.
Rotork Controls Limited P0116932
(iii)
Identificationof the firm constructing the facility or supplying the basic component which
fails to comply or contains a defect
Rotork Controls Ltd
Brassmill Lane
Bath, England
BAI 3JQ
(iv)
Nature of the defect orfailure to comply and the safety hazard which is created or could
be created by such defect orfailure to comply.
In 2013 Johnson Electric changed V12 micro-switches assembly locations. The first batch
(P0116932) from the new location was found to contain small particles of plastic case
material that are loose within the switch housing and may lodge on the contacts. Switches
manufactured prior to 2013 are not affected and due to batch control traceability within
Rotork, switches from the affected batch have been traced.
The manufacturer's acceptance criteria for the switch contact resistance is less than 25
milliohms. Contact resistance has been found to randomly vary between compliant (less
than 25 milliohms) to non-compliant (greater than 25 milliohms to occasionally open
circuit).
From a sample of 1613 switches, the manufacturer established a 0.11% probability that a
switch may exhibit an open circuit condition when operated. Rotork in-house testing within
actuators, using voltages representative of in-service levels, established a 0.0008%
probability that a switch may exhibit an open circuit condition when operated. Loss of
continuity is random, not permanent, and low in probability. To date, no V12 switch inservice faults have been reported by the industry.
Page 3 of 5
rotorw
Controls
(v)
The date on which the information of such defect orfailure to comply was obtained.
7"h September 2015
(vi)
In the case of a basic component which contains a defect or fails to comply, the number
and location of these components in use at, suppliedfor, being suppliedfor, or may be
suppliedfor, manufactured, or being manufacturedfor one or more facilities or activities
subject to the regulationsin this part.
.Row
1
2
Quantity
6
18
Serial Number
851470010101
B53470010301
Description
90NA1 WD 2023VNO
7NA1E/IWN3 WD 2023VNO
Shipped
3/12/15
1/28/14
3
12
B54087010101
30NAX1E WD 2023VNO
3/20/14
4
5
18
12
B54087010201
854177040101
30NAX1E WD 2023VN0
'14NA1E WD 2"021VNO
3/20/14
2/25/14
6
7
12
854330010101
14NA1E WD 2021VNO
NAlE/5E 12SW V12 AOP
4/25/14
5/14/14
120
B544440101
(QTY 10)
8
6
B544440401
NAlE/SE 65W V12 AOP
12/23/13
9
10
12
12
854727010101
854727010201
14NAIE WD 2021VNO
14NA1E WD 2021VN0
4/25/14
4/25/14
11
12
13
12
12
B54727020101
854727020201
8/5/14
8/5/14
3/14/14
14
18
12
B548420101
856398010101
14NAIE WD 2021VNO
14NA1E WD 2021VNO
NA1/5 SWITCH M ECH
(QITY 3)
14NA1E WD 2021VNO
15
16
17
12
18
18
B56398010201
14NA1E WD 2021VNO
12/29/14
12/29/14
856515010101
B57239010101
11NAlE WD 2023VNO
16NA1E WD 2023VNO
12/29/14
3/12/15
18
19
18
18
857239010201
B57239010301
16NA1E WD 2023VNO
16NA1E WD 2023VNO
3/12/15
3/12/15
20
21
18
857239020101
2/9/15
12/3/14
1
12
6
12
12
12
8576870101
013359010101
013385010101
013962010101
013962010201
013962010301
90NA1E WD 2023VNO
MODID 14/16NA5 G/CASE
MAINT KIT POST 78
7NA1E/IWN3 WD 2024VNO
7NA5E WD 2020VNO
7NASE WD 2021VNO
7NA5E WD 2021VNO
7NA5E WD 2021VN0
___
___
22
23
24
25
26
439
Sum
Page 4 of 5
_______
1/17/14
12/11/13
6/4/14
6/4/14
5/4/14
rotorw
Controgs
(vii)
The correctiveaction. which has beeni, is being, or will be taken; the name of the individual
.or organization responsiblefor the action; and the length of lime that has been or will be
taken to complete the action.
By this notification Rotork is notifying all utilities and listing affected customer orders.
The manufacturer Johnson Electric has introduced cleanliness checks at applicable stages
of switch assembly. Rotork has already obtained replacement switches and further batches
arc now being manufactured.
(viii) Any advice related to the defect or failure to comply about the facility, activity, or basic
component that has been, is being, or will be given to purchasers or licensees.
U.S. Licensees with installed or in storage actuators, switch mechanisms or Add-On-Pak
(AOP) spares modules containing VI2 switches identified in section (vi) should evaluate
the impact of a high contact resistance or occasionally open circuit condition on safety
related systems.
(ix)
In th~e case of an early site permit, the entities to whom an early site permit was
transferred
Not Applicable
Page 5 of 5
rotork
Page 1 of 20
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
Date:
ER791 Issue 1
9 th
September 2015
REVISION SHEETv
Privang Jadav
ORIGIN
DATE OF ISSUE
09/09/2015
TITLE
Part 21 Investigation..Intermittent operation of V12 Micro-Switch N69-921
Prepared
Checked
Priyang Jadav
Product Engineer
Nuclear
Approved
Kevin Sweet
Stuart Adam
-
Senior Engineer
Nuclear
-
Engineering Manager
-
Nuclear
Issue No
Date
Revision
By
1
09/09/15
Original
PUJ
rotorw
Approved by Engineering
Rotork Controls Limited
Nuclear Engineering Report No.
ER791 Issue 1
Date:
9 th
September 2015
CONTENTS
1.
Summary ...................................................................................................
3
2.
Deviation Description .....................................................................................
3
3.
Micro-Switch Function ....................................................................................
3
4.
Review of Orders Affected................................................................................ 4
5.
Root Cause Analysis .......................................................................................
6.
Johnson Electric Micro-Switch Testing .................................................................. 6
7.
ERA Micro-Switch Testing ...............................................................................
10
8.
Rotork In-House testing .................................................................................
14
9.
Conclusion ................................................................................................
20
5
rotork
Pg3of2
Approved by Engineering
Rotork Controls Limited
Nuclear Engineering Report No.
1.
ER791 Issue 1
Date:
9 th
September 2015
Summary
This report documents work performed to support a Part 21 investigation opened by Rotork Controls
Inc. ref. tracking number NCR164927. Measurements and tests were performed to establish the
operational reliability of a Rotork safety related micro-switch containing a manufacturing defect. Noncompliant switches were identified from a batch of 4997 switches supplied to Rotork in 2013 against
purchase order P0116932. The investigation has confirmed that the potential defect is only attributed
to the aforementioned 4997 switches and that switches orders prior to this are not affected.
1281 switches were supplied to customers prior to detecting the non-compliance. A percentage were
supplied to North American customers by Rotork Controls Inc. 3716 switches in stock at Rotork Ltd.
were returned to the supplier for analysis. From the 3716 switches, 1613 switches were tested by the
supplier, who established a 0.11% probability that a switch may exhibit loss of continuity when
operated. However, in-house testing in actuators using voltages representative of in-service levels,
establish a 0.0008% probability that a switch may exhibit loss of continuity when operated. Test data
indicate loss of continuity is random, not permanent, and low in probability. To date, no V12 switch
in-service faults have been reported by the industry.
2.
Deviation Description
On June 4, 2015 Rotork Controls Inc. opened a formal Part 21 investigation into a self-identified
anomaly relating to a Basic micro-switch
-
Part No. N69-921, description "V12". The anomaly is
intermittent variation in electrical contact resistance and was first observed during the Factory
Acceptance Test of a Rotork Safety Related NA Range electric Actuator; also referred to as an electric
Valve Operator.
3.
Micro-Switch Function
The micro-switch has four functions:
*
Rotork actuators can be set to operate to a maximum torque level of opening or closing a
valve. When the torque sensing mechanism in the actuator registers this maximum torque,
the V12 micro-switch will be tripped to turn off the motor.
*
Rotork actuators can also be set to operate to a maximum travel limit. The mechanism inside
the actuator will register when the maximum travel in the open or close direction is reached
and the V12 micro-switch will be tripped to turn off the motor.
*In the Add-on-Pak (AOP) the switch is intended for indication purposes but can also be used
for interlocks and permissives to start other equipment such as pumps and valves. AOP
switches can be set to trip at any point during valve travel. The AOP can also be used for torque
switch bypass. If the circuit is "open" then the MOV could stop before achieving "end of
travel".
*
Switches must be "closed" to initiate travel.
rotorweP
eo
Approved by Engineering
Rotork Controls Limited
Nuclear Engineering Report No.
ER791 Issue 1
Date:
9
th
September 2015
4. Review of Orders Affected
The following are details of the switches affected:
*
Order no. P0116932 was supplied to Rotork in 3 batches, Rotork Lot references LC009150,
LC009162 & LC009205.
*
All switches supplied to customers are from LC009150.
*
Suspect switches have date code 3513K to 4013K stamped at location shown below. Code
refers to week (35 to 40) and year (2013).
Figure 1: Location of date code on switch housing
rotore
PageS
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
5.
of 20
ER791 Issue 1
Date:
9 th
September 2015
Root Cause Analysis
Defect: Intermittent and/or high contact resistance of the N69-921 V12 Micro-Switch.
Why
Why
Why
-
micro-switch did not function as intended
micro-switch contacts were open circuit
microscopic plastic particles lodged between the contacts created an air gap
Why-
the action of the switch plunger can potentially disturb plastic particles within the switch
housing and cause them to become lodged between the electrical contacts
Why
inadequate de-flash and cleaning prior to switch assembly
-
Following the moulding operation housings require de-flash and cleaning operations. In 2013, for
P0116932, these operations were transferred from the housing manufacturer to Johnson Electric. An
investigation by Rotork QA and Johnson Electric determined the intermittent switching was attributed
to inadequate de-flashing and cleaning of the plastic housing at Johnson Electric. All switches supplied
against P0116932 are thus suspect. Microscopic plastic particles may randomly lodge between the
contacts and cause the resistance to exceed the Johnson Electric 25mf) acceptable limit, and possibly
loss of continuity.
Figure2: Uncleanedflash inside of ho using
Figure 3: Plastic particles around the contact
rotorwaeo
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
6.
Date:
ER791 Issue 1
9
th
September 2015
Johnson Electric Micro-Switch Testing
1281 switches have been supplied against customer orders worldwide. The remaining stock of 3716
switches from the defect batch was rejected back to the supplier for further analysis. The aims of
analysis were:
*
To investigate whether a method could be established by which non-compliant switches can
be identified and segregated from compliant switches. Readings exceeding the
manufacturer's 25m0 end-of-line acceptance criteria were classed as non-compliant.
*
To determine whether high resistance or open circuit condition occurs on the normally open
(NO) contact, normally closed (NC) contact, or both.
*
To determine whether the number of switch operations has an influence on contact
resistance.
*
To collect data which can be used to calculate probability of switch failure i.e. open circuit
condition.
Test Method
*
Each switch was operated 10 times. At each operation the NC and NO contact resistance was
recorded with a milliohm meter.
(This gives 10 NC and 10 NO resistance readings)
*
Test voltage
Test current
=
=
9VDC
5mA
*The switch was then operated a further 50 times to check if the number of operations
influenced the contact resistance value. The NC and NO contact resistance was recorded at
every tenth operation.
(This gives 5 NC and 5 NO resistance readings)
Total no. of operations per switch = 60
Total no. of resistance readings per switch
=
30
Test Results
Of the 3716 switches, 1520 were tested to above test method.
1520 x 30 readings per switch
= 45600
readings
A further 93 switches were operated only up to the point where high resistance/intermittent
operation was recorded. The total number of readings across the 93 switches were 911.
Total number of switches tested
Total recorded resistance readings
=
1520 +93
= 45600 +911
=
1613
= 46511
rotork
Page 7 of 20
Approved by Engineering
Rotork Controls Limited
Date:
ER791 Issue 1
Nuclear Engineering Report No.
th
9
September 2015
Table 1 characterises the resistance test results from this data sample. The milliohm meter can register
readings up to 2000 maximum. Readings over this limit are classed as open circuit. The column 'No.
of Readings' is presented graphically in Figure 4.
Resistance Range
No. of
Readings
No. of Readings as
%of Total
Cumulative Frequency
above 25mC)
0 < R < 25m0
25 < R< l00mO
100 < R < 200m0
200 < R < 300m0
46148
240
30
27
99.220%
0.5216%
0.065%
0.058%
300 <R•<400m0
400 < R _<500mO
500m0 < R < 10
1 <R•<20
2 <R <50
5 <R _<2000
2000+ (Open Circuit)
4
0.009%
301
5
1
0.011%
0.002%
306
307
manufacturer acceptance
_____________Readings
___________
240
270
297
3
0.006%
1
0.002%
1
0.002%
51
0.110%
Table 1: Johnson Electricresistance test data
310
311
312
363
N69-921 SWITCH RESISTANCE POPULATION
Data based on sample of 1613 switches from batch of 3716
400
46148 switch operations within manufacturing acceptance criteria (green)
350
300
0
250
363 switch operations outside
manufacturing acceptance crteria (orange)
-• 200
,,_150
0
6 1o0
50
0
00
a
2
U,
N
VI
0
V
0V
V
2o,
Resistance Range
Figure 4: Johnson Electric resistance test data in graphicalf•orm
U
rotorw
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
Date:
ER791 Issue 1
9 th
September 2015
The majority of non-compliant readings are < 500mO. In the range 50 < R < 200 there was one reading
at value 21.90. The effect of this in a 48V, l0mA circuit and a 120V, 30mA circuit is calculated below:
V=IxR
V=IxR
48 =0.0l0 xR
R =48 /0.010 =48000
120 = 0.030 x R
R = 120 /0.030 =40000
If circuit resistance increased due to high switch contact resistance of 21.90:
4800 + 21.9 = 4821.90
48 = Lx 4821.90
I = 48 / 4821.90 = 9.95mA
4000 + 21.9 = 4021.90
120 = I x 4021.90
I = 120 / 4021.90 = 29.84mA
The calculations above show that:
* In a 48V, l0mA circuit, a contact resistance of 21.90 would reduce the current to 9.95mA.
* In a 120V, 30mA circuit, a contact resistance of 21.90 would reduce the current to 29.84mA.
The high resistance values captured thus have negligible effect on the circuit current in above
examples. The primary concern is thus the open circuit condition. Table 1 and Figure 4 show 46148
compliant operations and 51 open circuit operations. The probability of a switch displaying an open
circuit condition is thus:
P(A)
=
51 / 46511
=
0.0011 or 0.11% (operation based data)
Although Rotork are treating the whole batch of switches as suspect, Johnson Electric test data shows
165 switches exceeded the 25m0 acceptance criteria. The probability of selecting a non-compliant
switch is thus:
P(B)
=
165 / 1613
=
0.1023, or 10.23% (population based data)
Summary
Resistance testing demonstrates the occurrence of high resistance or open circuit condition is random.
The anomaly occurs on the NO or NC contact, or both. Once the anomaly has occurred, correct switch
operation may revert on the next operation, or it may revert after a number of operations. Data shows
no trend between the number of operations and the number of anomalies. Some switches displayed
an anomaly on the first operation while others displayed an anomaly on the final ( 6 0th) operation.
Most switches passed the resistance test. However, due to the random nature of the anomaly it is
possible that on the
6 1 St
operation an anomaly could occur. The segregation method cannot provide
absolute assurance.
Contact resistances in the order of tens of Ohms are considered to have negligible effect and most
anomalous readings were below this value. However, the utility must evaluate this point. The open
circuit condition is a concern. Test data based on resistance testing indicate a 0.11% probability per
operation that any switch may exhibit an open circuit condition.
rotorw
~o
Rotork Controls Limited
Nuclear Engineering Report No.
Approved by Engineering
ER791 Issue 1
Date:
9 th
September 2015
Resistance testing is performed at 9VDC. Rotork extended the investigation to establish whether
supply voltage level influenced switch operation reliability. ERA, an independent test laboratory,
performed the evaluation. The results are discussed in the next section.
rotorw
Page 10of2
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
Date:
ER791 Issue 1
9 th
September 2015
7. ERA Micro-Switch Testing
5 switches that lost continuity during Johnson Electric test were sent to ERA to establish a voltage level
above which the switches would operate reliably. The conjecture for testing was that switches in open
circuit condition have particles lodged between the contacts which create an air gap that may be
bridged by the formation of an arc if the voltage potential is sufficiently high.
At ERA the open circuit condition could only be replicated on 2 of the 5 switches. On one switch a DC
voltage ramp was applied while in open circuit mode, and on the other an AC voltage ramp.
The other 3 switches, which had shown open circuit mode during Johnson Electric test, showed no
fault at ERA.
Test Method
*
Similar to Johnson Electric test, the switches were operated 25 times. At each operation the
NC and NO contact resistance was recorded with a milliohm meter.
(This gives 25 NC and 25 NO resistance readings)
*
Test voltage
Test current
<
5VDC
< 10mA
*
The switches were then operated a further 50 times. The NC and NO contact resistance was
recorded at every tenth operation.
(This gives 5 NC and 5 NO resistance readings)
*
During operations if loss of continuity was captured, the switch was kept in the fault mode
and a voltage ramp applied.
For DC ramp test:
Voltage representative of in-service levels
Current representative of in-service levels
Required level of resistance during ramp
For AC ramp test:
Voltage representative of in-service levels
Current representative of in-service levels
Required level of resistance during ramp
=
48VDC
= 10mA
=
48 / 0.010
=
48000
=
40000
= 120VAC
= 30mA
- 120 /0.030
rotorl(
Page
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
Date:
ER791 Issue 1
1of 2
September 2015
9th
DC Test Results
A contact resistance reading of 60k0 was captured on the NO contact. A voltage ramp up to 48VDC
was applied to the switch. The switch started to conduct at 15V, stopped conducting at 23V, and
started to conduct again at 32V (Figure 5).
60kC2 Contact Resistance, 48VDC Ramp
14
13
55
12
50
11
10
9
C" 40
8
0u
•,
30
C)
6
4
10
3
15
10
0
2
5
9
7
8
9
10 11 12 13 14 15
Time (set)
16 17 18 19 20
21 22 23 24 25
-- VO
=(re!
Figure 5: Voltage ramp up to 48VDC when 60kD resistance was captured
On the next operation of the same switch, a contact resistance reading above 2M0 (milliohm meter
limit) was captured on the NO contact and a voltage ramp was again applied to the switch. This time
the switch started to conduct at 10V, stopped conducting at 32V, and started to conduct again when
the voltage was further increased to 64VDC (Figure 6).
>2M0 Contact Resistance, 48VDC then to 64VDC Ramp
95
14
3
-
12
11
10
9
8
C
6
U
4
_iii
0 10
I
20
30
40
50
80
20
80
Time (set)
90
100
110
120
130
140
-- VC
--
Cjrren¶
Figure6: Voltage ramp up to 48 VDC then increased to 64VDC, when >2MD resistancewas captured
rotorl(
ae1
Rotork Controls Limited
of 20
Approved by Engineering
Date:
ER791 Issue 1
Nuclear Engineering Report No.
9
th
September 2015
AC Test Results
AC voltage tests were carried out on a different switch. A contact resistance reading above 2M0Q
(milliohm meter limit) was captured on the NO contact. A voltage ramp up to 120VAC was applied to
the switch. The switch started to conduct at 104V (Figure 7). However, with the switch in the same
state, a subsequent ramp was applied (Figure 8) and the switch failed to conduct.
>2MC) Contact Resistance, 120VAC Ramp (1)
34
32
30
28
26
24
120
10
100
90
22
20 •"
>i
70
•
60
E
18
1604
,0
>
40
14
30
12
10
20
10
0
1
2
3
4
5
6
7
8
8
10
Time (see)
11
12
13
14
15
16
17
18
-- VAC -I-l
•ren
Figure 7: First voltage ramp up to 120 VAC when >2Mf) resistance was captured
>2MC2 Contact Resistance, 120VAC Ramp (2)
34
32
130
120
30
110
28
26
1301
90
24
22
20 •"
E
50
12
40
10
8
> 0
>
30
6
4
2
0
20
10
0
0
1
2
3
4t
5
6
7
8
9
10
Time (sec)
11
12
13
14
35
1615 I
--
~
16
--
urn
Figure 8: Second voltage romp up to 120 VAC when >2MC2 resistance was captured
rotori
ae1
Rotork Controls Limited
Nuclear Engineering Report No.
of 2
Approved by Engineering
ER791 Issue 1
Date:
9 th
September 2015
Summary
Tests have revealed that when a voltage ramp is applied to a switch that is already in open circuit
mode, the switch will not always operate reliably. A safe DC or AC operating voltage could thus not be
established.
However, when the voltage ramps were applied, the switch was static and fault condition already
established. It was thought that in dynamic operation, mechanical impact of the contacts and the heat
from the arc may both act to clean the electrode contacts. Based on this the investigation was
extended to study the likelihood of an open circuit condition arising when switches are being used in
an actuator at voltage levels representative of the application. The testing and results are discussed
in the next section.
rotorw
Pae1
Rotork Controls Limited
Nuclear Engineering Report No.
f2
Approved by Engineering
ER791 Issue 1
Date:
9 th
September 2015
8. Rotork In-House testing
Rotork in-house tests were conducted to test suspect switches in the most representative manner of
how they would be utilised in service. The reason for testing in a manner that represents the
application is based on the hypothesis that mechanical impact of the contacts and the heat from the
arc may both act to clean the electrode contacts thus switch reliability in application may be better
than Johnson Electric resistance measurements suggest.
A batch of switches including those that passed and those that were found non-compliant during tests
at Johnson Electric were called back to Rotork for further testing. Switches were assembled in the
Switch Mechanism, Figure 11, and Add-on-Pak (AOP), Figure 12, of 2 nuclear actuators.
The actuators completed a number of life tests to at least 4000 cycles Open-Close-Open. In between
each life test the switches in the Switch Mechanism and AOP were removed and replaced with more
suspect switches.
rotorw
Page 15 of 20
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
ER791 Issue 1
Open auxiliary
switch (OAS2)
Open auxiliary
Date:
9 th
September 2015
~Close
auxiliary
switch (CAS2)
~Close
auxiliary
_____
switch (CASi)
switch (OAS1)
Open torque/limit
•----Close torque/limit
-
switch (CTS)
switch (OTS)
AOP indication switches
AOP indication switches
IAS7
IASi - IAS6
Figure 12: Photo of Add-on-Pack (AOP)
-
IA512
rotorI
Page 1
of 20
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
ER791 Issue 1
Date:
9 th
September 2015
Switches OTS and CTS were operated at 24VDC, 2.6mA average.
Switches OAS1-2 and IASI-IAS6 were operated at 48VDC, 16.9mA average.
Switches CAS1-2 and IAS7-IAS12 were operated at 120VAC, 12.5mA average.
A reduced voltage life test was also completed using 38.8VDC, 13.7mA and 80VAC, 6.6mA.
Voltages above 24VDC were applied through relays as illustrated in Figure 13.
•,POWER SUPPLY
V12 TEST
SWITCH
•.jVMODULE:
I•-
eod
Figure 13: Typical schematic showing V12 switch test
The operation of each switch was monitored during test. The data captured showed which switches
behaved intermittently and when. The testing is summarised in Table 2.
A total of 84 different switches were tested during tests 1 - 9. Of these:
*
*
*
28 were switches that showed no fault during Johnson Electric tests. These also showed no
fault during representative testing in an actuator.
51 were switches that intermittently exceeded 25m0 during Johnson Electric tests, having
readings in the range 26m03 to 2.8903.
5 were switches that showed open circuit condition. This includes the 2 switches that were
used for DC and AC voltage ramp test at ERA. The 2 were used in Test 9 and no faults were
reported.
In Test 1, the AOP was fitted with switches that showed no fault during Johnson Electric tests. The
actuator completed 5537 cycles after which, in Test 2, all switches other than OTS and CTS were
replaced by more switches that showed no fault. In both Test 1 and 2 the switches that passed Johnson
Electric resistance test did not display open circuit behaviour during actuator operating life.
Between Test 3 - 7 the same actuator was used with the same switches. Switches OTS and CTS were
the same as Test 1 and 2. All other switches were those that showed high resistance during Johnson
Electric tests. The total actuator cycles in Tests 3
-
7 were 15686. The purpose of these tests was to
see what effect the voltage levels had on switch failure. In Test 3 there were 3 intermittent operations
recorded and in Test 7 there were 4, both tests conducted at low voltage 38.8VDC and 80VAC (limited
by the minimum voltage capability of the relays). In Test 8 more high resistance switches were tested
and there were 13 fault occurrences at nominal voltage 48VDC and 120VAC. These tests show that
the fault can potentially occur at nominal and reduced voltage.
rotorl(
ae1
Rotork Controls Limited
Approved by Engineering
No f
actuator
cycles
Actuator
Date:
ER791 Issue 1
Nuclear Engineering Report No.
Test
Ref.
of 2
OAS1-2 &
OS&
IAS1-6
CTS testDecitoofes
test
voltage
voltage
CAS1-2
& IAS712 test
voltage
No. of
intermittent
operations
recorded
9 th
September 2015
Decitoofes
IAS1-12 were switches that
showed no fault during
Johnson Electric tests.
a
5537
14NAIE
24VDC
48VDC
120VAC
0
CS2adOSswitches were those that
showed high resistance
during Johnson Electric
___________tests.
2
14NA1E
4000
3 1NA1
400
24VDC
2V0
48VDC
388VD
__________
4
14NA1E
4000
24VDC
12OVAC
0
8OAC
3
(See Note 1)
________tests.
48VDC
120VAC
0
This test is a repeat of Test
3 with same switches but
0
This test is a repeat of Test
4 with same switches
0
This test is a repeat of Test
5 with same switches
4
This test is a repeat of Test
6 with same switches
at nominal voltage.
_________running
5
14NA1E
1913
24VDC
38.8VDC
80VAC
at reduced voltage.
________running
6
14NA1E
4000
24VDC
48VDC
120VAC
at nominal voltage.
________running
7
14NA1E
1773
24VDC
38.8VDC
80VAC
(SeeNot
2)
8
9
7NAIE
14A1E
400
4000
From Test 1, all but
switches OTS and CTS were
replaced with those that
showed no fault during
Johnson Electric test.
From Test 2, all but
switches OTS and CTS were
replaced with those that
showed high resistance
during Johnson Electric
24DC
8VD
12VAC
13
(See Note 3)
24VDC
48VDC
120VAC
0
running at reduced voltage.
From Test 7, all but
switches OTS and CTS were
replaced with those that
showed high resistance
during Johnson Electric
tests.
Switches IAS4-6 and lAS1112 were those that showed
"OFF" during Johnson
Elcrcts.Alohe
switches were those that
showed high resistance.
Table 2: Summary of Rotork in-haouse switch testing with nuclear actuators
rotor
Page 18of2
Rotork Controls Limited
Nuclear Engineering Report No.
Approved by Engineering
ER791 Issue 1
Date:
9
th
September 2015
Note 1:
Switch at OAS2 (at 38.8VDC) failed to close circuit during cycle 865.
Switch at IAS8 (at 80VAC) failed to close circuit during cycle 28 and cycle 954.
Note 2:
Switch at OAS2 (at 38.8VDC) failed to close circuit for 4 consecutive cycles from cycle
42.
Note 3:
Switch at OAS2 (at 48VDC) failed to close circuit during the following cycles; 279, 280,
281, 283.
Switch at IAS2 (at 48VDC) failed to close circuit during the following cycles; 276, 3647,
3651, 3655.
Switch at IAS9 (at 120VAC) failed to close circuit during the following cycles; 6, 43, 45,
47, 158.
All switches that failed (Note 1 - 3) were those that were found to have intermittent high resistance.
There were no failures of switches that had previously exhibited open circuit.
For the calculation of failure probability over actuator life only the first 4000 operations of each switch
are relevant. A switch operates once per actuator cycle. Based on Table 2, the total number of relevant
operations are:
Test 1
Test 3
4000 operations of 18 switches.
Test2
400 peraion of16 sitces.
4000 operations of 16 switches.
Demonstrates failure is random and the 25m£0
tetdecsfu.
Test4
Swich-7oeraionsfro
With Test 1 - 3 demonstrates that failure is
tes 47
are not included in the
probbilty
alcuatinsrandom
Test 8
Test 9
and not affected by accumulated cycles.
4000 operations of 16 switches.
Demonstrates failure is random and the 25mC)
4000 operations of 18 switches.
test detects fault.
Table 3: Relevant operationsfor calculation of switch failure probability
It should be noted that the switches used in representative testing were not randomly selected. The
selected switches were pre-identified as being compliant or non-compliant as per Johnson Electric
test.
rotorl(
ae1 f2
Rotork Controls Limited
Approved by Engineering
Nuclear Engineering Report No.
ER791 Issue 1
Date:
9 th
September 2015
Probability of a switch exhibiting open circuit in application is given by:
P(C)
P(B) x P(D)
=
Where:
P(B)
= Probability of selecting a non-compliant switch i.e. >25m02. Calculated in section 6.
P(D)
= Probability of a non-compliant switch displaying open circuit when installed in an actuator
The probability of a non-compliant switch displaying open circuit, P(D), is determined using data from
Rotork in-actuator representative tests 3, 8, and 9. The total number of switch operations across these
three tests are computed below and displayed in Table 3:
2(4000 x 16) + (4000 x 18)
=
200,000 (total operations of tested switches classified non-compliant)
The total number of observed open circuit operations in these three tests are:
3 + 13
=
16 (see Table 2)
P(D)
=
(16 / 200,000)
P(C)
=
0.1023 x 0.00008
=
0.00008, or 0.008% (operation based data)
=
8 x 10.6, or 0.0008% (per switch operation)
In some applications when a valve is initially stroked two switches are required to function correctly,
the torque/limit switch and the torque bypass switch. The probability that either of these goes open
circuit is:
P(2 Switches)
100 - 0.0016
=
=
2 x P(B) = 0.0016%
99.998% probability of success
The probability that any one of the 18 switches in an actuator fitted with a switch mechanism and 12
switch AOP could go open circuit is:
P(18 Switches) = 18 x P(B) -0.0144%
100
-
0.0144
=
99.986% probability of success
Summary
The probability P(C) of a switch displaying open circuit in application is considerably lower than the
probability P(A) of a switch displaying open circuit during Johnson Electric test. These result suggest
that switch operation is more reliable than resistance evaluation indicates, It also supports why no inservice faults are reported by industry and why Rotork Controls Inc. eventually observed the anomaly
based on 100% resistance inspection of all V12 switches fitted to actuators.
rotorI(Pge2 of
Approved by Engineering
Rotork Controls Limited
Nuclear Engineering Report No.
9.
2
ER791 Issue 1
Date:
9 th
September 2015
Conclusion
Testing indicates that occurrence of high
probability of a switch exhibiting an open
switches, there is 0.11% probability that a
house tests show that in application there
open circuit when operated.
resistance or open circuit condition is random with a low
circuit condition. Based on Johnson Electric testing of 1613
switch may exhibit open circuit when operated. Rotork's inis a 0.0008% probability that a switch operation may exhibit
Thermal & radiation effects could not be evaluated but are considered not pertinent given the
contamination nature. The housing material is qualified for high temperature and radiation levels thus
the material properties of contaminating particles are thought to be unaffected by these factors.
To date there have been no in-service faults reported.
Rotork cannot evaluate application risk. Utilities must evaluate their in service switches and determine
if a high contact resistance or open circuit conditions could potentially affect safety related functions.