The LHC Beam Interlock System

1
M. Zerlauth
0v1
2

Outline

• HCC campaign 2009 is progressing very well with
  gtgt number of tests performed / shift
• Many changes and improvements of test plans,
  powering procedures, powering specifities, tools,
  have/had to be done for this years campaign
• LHC powering system commissioning tools by now
  a rather complex construct
• Presentation summarizes and answers/provides
  hints to some FAQ to further minimize number of
  failing tests and (avoidable) test repetition
• Focus on operational procedure
• Not included are implementation-issues ie,
  errors in sequences, test-plan, LSA, etc

Machine Interlock Systems
3

Conditions before starting a test

• Next step not correctly proposed by sequencer
  (especially for PIC1 / PIC2 of mains)
• Login as lhcop
• SOC needs to be reserved for your console
• Circuit must not be locked (ie due to previous
  failure)
• Is previous test completed? Ie the analysis
  signed
• Check correct short status in SOC editor and
  Simulation/normal mode of power converter

Machine Interlock Systems
4

Conditions for starting a test

• Circuit not being ready for testing is most
  frequent reason for a failed test
• Multitude of conditions required, most frequent
  missing (gt90) are
• Circuit still locked / super-locked -gt PIC
  supervision
• QPS controller not yet ready (ie a reset is
  needed) or 13kA switches are still open -gt PIC
  supervision / Circuit synoptic
• CRYO conditions (temporarily) lost -gt PIC
  supervision
• UPS conditions (temporarily) lost -gt PIC
  supervision
• What the sequencer is capable/allowed doing in
  2009
• Closing of 600A switches
• Resetting of faults in PIC/PC

Machine Interlock Systems
5

Conditions for starting a test

• To avoid unnecessary repetition (and follow-up)
  of failed tests, it is strongly recommended to
  verify the correct state of the circuit before
  launching ANY test, e.g. using Circuit Synoptic

Sequence will work
Machine Interlock Systems
6

Conditions for starting a test

• To avoid unnecessary repetition (and follow-up)
  of failed tests, it is strongly recommended to
  verify the correct state of the circuit before
  launching ANY test, e.g. using Circuit Synoptic

Sequence will NOT work
Machine Interlock Systems
7

Conditions for starting a test

• To avoid unnecessary repetition (and follow-up)
  of failed tests, it is strongly recommended to
  verify the correct state of the circuit before
  launching ANY test, e.g. using PIC Supervision

Machine Interlock Systems
8

Conditions for starting a test

• Be careful with Circuits that have already been
  frequently repeated (especially tests where
  heaters are fired, etc..)
• Dangerous tests PIC2 CIRCUIT QUENCH, PLI.f,
  PNO.f
• Check in HCC pages history, now of test
  repetition also in overall view

Machine Interlock Systems
9
Conditions for starting a test

• New version of HCC show now the current
  repetition number for the test currently under
  execution
• By default, green as long as rep lt2, red if
  repeated more often
• New MPP strategy 2 training quenches allowed,
  afterwards open NC and change I_PNO

10

Reasons for failures during sequence

• In addition to quenches, trips, etc the
  following things might make a sequence
  fail/time-out
• Wrong login in sequencer, resulting in RBAC
  exception -gt PIC FGC is now under RBAC, only
  lhcop will work from CCC island
• QPS controller not reset -gt Error message like
  QPS.PVSS ST_LOGGING not OK -gt Reset QPS
  controller before test
• Mismatch of parameters in LSA and power converter
  (out of limits exception) -gt ask EPC expert to
  re-synchronize or CO to correct LSA
• Exceeding I_ERR or I_EARTH limits Considered as
  warning only (sequence will halt but not fail) -gt
  Confirm with EPC expert and in case accepted skip
  step in sequence
• CRYO conditions (only CRYO_MAINTAIN), UPS
  conditions (only HW signal) -gt Can be checked in
  PIC supervision or PIC history buffer

Machine Interlock Systems
11

Reasons for failures during sequence

• Asserting pop-ups too quickly (ie if action not
  yet completed) might result in timeouts at a
  later stage
• Re-trip of agent after automatic reset (e.g. in
  RQ6.R7B1 ) -gt Circuit synoptic
• PM acquisition in agent blocks (RQ4.L8) -gt
  Circuit synoptic

Machine Interlock Systems
12

Signing a test

• Overall outcome of a HCC test is logical AND
  between successfully performed sequence PMA
  comments
• Thus if sequence failed for any reason, even a
  passed PMA will not alter the test outcome
• For special cases (ie long sequences that failed
  with minor errors), outcome should be altered by
  CO expert BEFORE signing the test
• Automated analysis
• Automated path introduces additional 10 minute
  delay to make sure data is present
• Currently only PIC2 for 120A and 600A signed
  automatically
• Test should normally be signed 15 minutes after
  the end-time of test
• If after this time test does not appear in PMEA
  nor in history -gt Call LV-expert
• If test is signed by experts (not visible in PMEA
  anymore) but not updated in HCC pages (ie next
  step proposed) -gt Call LV-expert

Machine Interlock Systems
13

Unlocking of Circuits after Failing test

• Test execution must be sequential, when UNLOCKING
  circuits in SOC editor after failure, make sure
  no PMA is pending
• Test repetition before completion of analysis of
  previous test could be dangerous and breaks MTF
  upload (requires for sequential upload)

Machine Interlock Systems
14

Re-commissioning of circuit

• Pointer defines commissioning status of a
  circuits LAST_PASSED_TEST
• If circuit should be re-commissioned (e.g. after
  HW change)
• Request point owner to move back this pointed in
  the SOC editor to the required step (where to
  start re-commissioning)

Machine Interlock Systems
15

Cleaning-up

• If sequences have been e.g. repeated by mistake
  (and step was already performed correctly)
• Complete the test (if not long test, critical,
  etc..)
• Abort test immediately tell CO expert to delete
  test in LSA before signed and copied to MTF
• Others?
• Id be happy for your feedback of other
  reoccurring issues
• Try to include new things answers in these
  slide publish on HCC pages and others?

Machine Interlock Systems
16

• FIN

Machine Interlock Systems