premature dumps in 2011

Premature Dumpsin 2011

Acknowledgements: A.Macpherson, G.Papotti, M.Zerlauth

M.Albert LHC Beam Operation Workshop December 2011

2

Outline

• Introduction• Data extraction• Dump statistics by mode• Dump causes• Dumps in STABLE BEAMS• Summary

3

Introduction• Every dump generating PM data is

assessed by shift crew using PM Online tool.

• All beam dump characteristics are collected and stored in the PM Database.

• Dumps at 3.5 TeV are systematically reanalyzed by MPS experts.(M. Zerlauth, R. Schmidt, J. Wenninger)

• Analysis of PM data for p+ run: 19/02 (first beam) – 30/10/2011 (last dump) IB1 > 1e12 and IB2 > 1e12 (exclude probes,

MPS tests)

4

PM Online Tool

Classification & Assessment by OP

5

LHC Post Mortem Database

URL: lhc-postmortem.cern.ch

access to DB

DB Query Query builder,Reports &Data download

6

All beam dumps by mode

36.5%

0.8%1.0%

5.6%1.5%5.0%5.8%

35.9%7.9%

INJECTION PHYSICS BEAM

INJECTION SETUP BEAM

PREPARE RAMP

RAMP

FLAT TOP

SQUEEZE

ADJUST

STABLE BEAMS

BEAM DUMP

Total: 482

7

Non-programmed dumps by mode

33.1%

1.1%

1.3%6.9%

1.1%6.1%5.3%

44.8% 0.3%

INJECTION PHYSICS BEAM

INJECTION SETUP BEAM

PREPARE RAMP

RAMP

FLAT TOP

SQUEEZE

ADJUST

STABLE BEAMS

BEAM DUMP

Total: 375 (78%)

Inj. problems: ~10 %

8

Beam dump causesE >= 3.49 TeVIB1 > 1e12IB2 > 1e12

169 263 272

LBDS, PIC , BLM , BIC , SIS

Orbit, BPM, Beam Loss, Collimator adjustments, BCM, Feedbacks

42%

71%

10%

79%

9

Beam dump causes by equipment

E >= 3.49 TeVIB1 > 1e12IB2 > 1e12

“TOP 5”: 77% of dumps in SB

10

Beam dump causes by beam monitoring

E >= 3.49 TeVIB1 > 1e12IB2 > 1e12

11

Systems which saved us…

E >= 3.49 TeVIB1 > 1e12IB2 > 1e12

12

STABLE BEAMS – often short !

13

Dumps in STABLE BEAMS

SEUs by systemCryogenics: 16QPS: 15Coll System: 4PC: 2PIC: 1

14

Fill 1958 (S34) Fill 1996 (P8)

Average: 4.6h

Recovery after dump in SB

15

Proposals for improvements• PM Online tool should offer re-editing

(correction of wrongly assessed dumps)• Include injection scheme in PM-

Database• Do we need to add more categories for

dump classification?“Other” used for 11% of non-programmed dumps

• Should we also systematically analyse dumps at injection energy ?

16

Summary• Beam dump causes due to equipment

failures clearly dominate.• 50% of all fills & “1380b – fills” last <

3h• Average fill duration: ~5h• 22% of STABLE BEAMS - fills were

dumped by SEU effects, mainly on cryogenics and QPS equipment

• The average time from a beam dump in STABLE BEAMS to the next injection is ~4.5h.