1 / 32

Performance of the Control System

Performance of the Control System. C.H. Wang Control Group 2007.5.10. Outline. System Overview System Performance Computer&Network Oracle Database PS Control System Ramping Problems Improvement Summary. VME IOC. CABLE. Control System Overview. Public Network. Firwall.

lenka
Download Presentation

Performance of the Control System

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Performance of the Control System C.H. Wang Control Group 2007.5.10 IMAC 2007.5.10

  2. Outline • System Overview • System Performance • Computer&Network • Oracle Database • PS Control System • Ramping • Problems • Improvement • Summary IMAC 2007.5.10

  3. VME IOC CABLE Control System Overview Public Network Firwall File Servers Console Console Console Console Console Console Ethernet VME IOC VME IOC VME IOC PC PC CABLE CABLE CABLE CABLE CABLE Vacuum AB PLCs RS232 PS Controllers… Cryogenic AB PLCs RF Controllers PPS PLCs Interlock PLCs IMAC 2007.5.10

  4. Control System Overview • Subsystems were put into operation on time before the beam commissioning. • PS control system • Vacuum/Temp. Monitor • Cryogenic control • Timing system • Safety&Interlock systems • Computer system • Network • Oracle database • operation&running are going well. IMAC 2007.5.10

  5. Computer Performance • Sun File Server : 2-node fail over • CPU：8*1.2GHz UltraSPARC® III • Memory：32GB SDRAM Disk：6*73GB • CPU Load：10％～13％ • Memory Load：19%~20% • Sun Workstation/Solaris running SAD • CPU：2*1.2GHz UltraSPARC® III • Memory：4GB SDRAM Disk：2*73GB • CPU Load：15％～18％Memory Load：24%~30% • HP PC/Linux running EDM • CPU：Pentium IV-3.2GHz • Memory：1GB DDR Disk：160GB SATA • CPU Load：40％～49％ • Memory Load；34%~52% IMAC 2007.5.10

  6. Computer Performance • HP PC/Linux: Archiving ~6,000channel • CPU：Pentium IV-3.6GHz • Memory：2GB DDR Disk: 250GB、500GB SATA • CPU Load：40％～49％Memory Load：34%~52% • Oracle Server: ~4000 channel • 2 x Intel Xoen 3.0GHz • Memory： 2GB Disk: 2 x 146GB • CPU Load：40％～49％Memory Load：34%~52% • VME IOC(~30): MVME5100/VxWorks5.4.2 • CPU: 450MHz Memory: 512M RAM • Max. CPU load: ~50% IMAC 2007.5.10

  7. Control Network • Core Switch:Catalyst 4506 (redundant) • Edge Switch: Catalyst 3550/2950 • Two separate VLAN • Network running ok(most case) • Network problem: slow (sometimes) • Occasionally, VME IOCs crash by network trouble • Loosing CA link between OPI and IOCs is a candidate of the cause IMAC 2007.5.10

  8. Control Network IMAC 2007.5.10

  9. Typical Network Traffic IMAC 2007.5.10

  10. Oracle Server • Stores magnet measurement parameters • Stores real-time signal historical data over long period • PS • Vacuum • Temperature Measurement • Beam Instrumentations • Timing • Physics • Signals：~4000 Sample rate:1min. /time IMAC 2007.5.10

  11. Web User Interface IMAC 2007.5.10

  12. Query Results 2006/12/21 15:36:45 to 2006/1228 15:36:45 Beam Current within a week，max.201.2(mA） IMAC 2007.5.10

  13. Query Results Vertical direction display E+ beam current lifetime within a day IMAC 2007.5.10

  14. Oracle DB Problems • Too much data saved • Disk(120GB)/RAM data storage space small • Disk usage • 600MB-700MB/24hours • 10GB/16x24hours • 300GB/year • Archived data not imported into Oracle • Upgrade Oracle Computer Configuration in future IMAC 2007.5.10

  15. PS Control System • Signal： • ~10000个，11VME IOC • Basic Function： • ON/OFF/Status monitor • ON/OFF Command interlock with the setpoint • Alarm interlock with the setpoint IMAC 2007.5.10

  16. PS Control Calibration • Before commissioning, Q、S PS calibration • Calibration: y=kx+b • K:slope; b: Intercept • Y: DCCT; x:PSC code(Hex.) • 10 points： • Linear fit data • Precision: 100ppm IMAC 2007.5.10

  17. PS OPI • BSR、BER、BPR 3rings • 4 panels per ring (BSR/BER/BPR)： • B、Q、S Monitor • B、Q、S ON/OFF • Corrector Monitor • Corrector ON/OFF IMAC 2007.5.10

  18. BPR PS Monitor • Left: region 2、3，Right: region 1、4 • 8row info，Magnet/Knob/Status/Arlam/DESIMON/Setpoint/Current • Ramping/STDZ/Execution Status Monitor IMAC 2007.5.10

  19. PS Control Commissioning • Device commissioning started region by region • On one hand magnets were installed, on the other hand magnet PS remote controls were tested • Short time for All PS control online testing • PS ramping region by region put into operation at the beginning of the beam commissioning • After the commissioning, start to develop the ramping program of the entire ring BSR • Short testing time for ramping program IMAC 2007.5.10

  20. PS Ramping • SR Running: 2006.12.25 –2007.2.2 • Important Issue: Energy Ramping in BSR • Requirements: • All PS in BSR are ramped to the desired value synchronously (from 1.89GeV to 2.5GeV) • Implementation: • PS Control distributed(11 VME IOC) • Ramping program in every IOC • EPICS event to trigger every ramping program • Real-time DB processing slow(60Hz clock) , different CPU load, synchronization problem • Ramping program in a independent VME IOC • Change EPICS clock from 60Hz to 600Hz • Real-time database processing fast • write setpoint to setpoint DB of the all IOCs step by step • Ramping procedure is getting fast IMAC 2007.5.10

  21. B、Q、S PS Control Distribution • PS distribution is different, B PS independent, Q/S PS+ DC PS • 1 DC power to 10 choppers，for example :5 OQ and 5 IQ • Different PSC number in each VME crate: • PVs Num. Different ,Different CPU load • EPICS default clock : 60Hz PVs processing slow • Ramping Program in each IOC: not synchronously IMAC 2007.5.10

  22. Three Stage Testing • Stage 1(first half of Dec. 2006) • A ramping program in every IOC • EPICS IOC clock 60Hz • Stage 2(second half of Dec. 2006) • A ramping program in a independent IOC • EPICS IOC clock 60Hz • Stage3(Jan. 2007) • A ramping program in a independent IOC • Change EPICS IOC clock from 60Hz to 600Hz IMAC 2007.5.10

  23. Stage 1(2006.11-2006.12) • a ramping program in every IOC • Dipole IOC propagate event to all IOCs. • Every Ramping program write setpoint to RTDB step by step • initial and final current first given • The number of steps determinedby • (Desimon-setpoint)/Delta • Delta determined by the option in the OPI • Setpoint=Setpoint+Delta • resulting in ramping relevant hardware. • Problems: • Slow: (1.89GeV to 2.5GeV) • Sometimes, not synchronized • Continuing the ramping, a rapid beam loss was encountered when the energy reached VME crates in the network IMAC 2007.5.10

  24. Stage 2 (06.12) • Re-arrange VME IOC distribution in the network • All IOCs are connected to a way to minimize the unrelated network traffic. • EPICS clock 60Hz • A ramping program in a independent IOC, no synchronization problem • Ramping Slow: 20 min. (1.89GeV to 2.5GeV) • Reason: • EPICS clock 60Hz resulting PVs processing slow • Time interval between two steps must be long enough IMAC 2007.5.10

  25. 1.89Gev 2.5Gev ~20min. Testing Results IMAC 2007.5.10

  26. Stage 3(Jan. 07) • Change EPICS clock from 60Hz to 600Hz • Testing with beam( 2 times)(On Jan. 30) • Current=46mA, slow speed,delta=0.5A, • Ramping from 1.89GeV to 2.5GeV:5 min. no beam loss • Current=47.89GeV, fast speed, delta=1.0A • Ramping from 1.89GeV to 2.5GeV:1.5 min. a little bit beam loss • Testing with beam(on Feb. 4) • Current=250mA,fast speed, delta=1.0A • Ramping from 1.89GeV to 2.5GeV(with 3 stops): ~1.5min. No beam loss • Continuing the ramping, a beam loss was encountered when the energy reached IMAC 2007.5.10

  27. Max. Current 1.89Gev 2.5Gev Current ~6min. ~1.5min. Testing results IMAC 2007.5.10

  28. Max. Current 2.5Gev Current 1.89Gev ~5 min. ~1.5min. Testing results IMAC 2007.5.10

  29. Energy(B current)/Lifetime/CurrentRamping no beam loss IMAC 2007.5.10

  30. Summary • Ramping program has a enough fast speed now • Improvement on Ramping Quality with beam • Different magnets, different PS, different magnetic field response • Improvement next step: • Linear ramping with smaller delta value (for eg. 10mA) • Table ramping • Current table related to the energy will be given by the AP group IMAC 2007.5.10

  31. System Performance IMAC 2007.5.10

  32. Thanks！ IMAC 2007.5.10

More Related