Presentation is loading. Please wait.

Presentation is loading. Please wait.

Control Status Readiness Report

Similar presentations


Presentation on theme: "Control Status Readiness Report"— Presentation transcript:

1 Control Status Readiness Report
Eugenia Hatziangeli on behalf of the CERN Accelerator and Beams Controls Group Collimation project The choice of the FE electronics was PXI (an extension of the Compact Pci). This is not part of the standard CO FE catalogue, so the responsibility of the FE software fell on the side of the equipment responsible. On top of this a fesa server is written which access the collimators like a black box (from the point of view of Controls). So from the CO point of view there is no difference between collimators or BI (access via CMW and FESA). The "Labview" disease is less of a problem. We have a JAVA based application program and a FESA based front-end infrastructure. There is a possibility to bypass all this and to control the collimatorsfrom a specific labview (test) application. This door will be closed by our security team. 12 June 2008 LHC – MAC

2 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary 12 June 2008 E. Hatziangeli AB/CO

3 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary 12 June 2008 E. Hatziangeli AB/CO

4 LHC controls infrastructure – Overview
The 3-tier architecture Hardware Infrastructure Software layers Resource Tier VME crates, PC GW & PLC dealing with high performance acquisitions and real-time processing Database where all the setting and configuration of all LHC device exist Server Tier Application servers Data Servers File Servers Central Timing Client Tier Interactive Consoles Fixed Displays GUI applications Communication to the equipment goes through Controls MiddleWare CMW Applications Layer Client tier Business Layer Server tier CMW Hardware CTRL DB Resource tier 12 June 2008 E. Hatziangeli AB/CO

5 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary 12 June 2008 E. Hatziangeli AB/CO

6 Front-Ends Hardware and Software
Hardware Installations All Front-end controls equipment in place (> 200 VMEBus systems and > 250 industrial PCs) WorldFIP infrastructure operational for PO, QPS, Cryogenics and BI (400 km network, nodes) General Machine Timing (GMT) network operational, including transmissions for LHC Collimators and for LHC Experiments Front End Software Architecture (FESA) FESA V2.10 Framework operational, including support for machine critical settings, transactional commands and in-depth diagnostics of RT behavior (via Alarms system LASER) Front-End FESA classes developed by AB equipment groups (> 250 classes deployed on > 400 front-ends) Deployment process =>AB-CO supports 3 last FESA releases All industrial PCs running now Linux O/S On-going Actions Two major tendering exercise for the procurement of AB front-end hardware (adjudication during CERN FC in September 2008) LHC Alarm Service 12 June 2008 E. Hatziangeli AB/CO

7 Accelerator Databases Readiness
Off-line Databases Online Databases Layout Racks & electronics incorporated up to a high level of detail Layout data is now used as foundation for the controls system Still to do More data is being captured relating layout and assets information Tools for data maintenance still to be put in place Operational Settings Data model enhanced to cover functional extensions for Role Based Access (RBAC) , XPOC, Sequencer PS Controls Renovation requirements Logging Service New database hosting since Mar. 08 Common logging infrastructure for the complete accelerator chain Sustained increasing logging requirements for HWC& beam data Improved data retrieval tool 12 June 2008 E. Hatziangeli AB/CO

8 Service Availability and Data Security
Additional server for testing: Standby database for LSA Controls Configuration LSA Settings E-Logbook CESAR HWC Measurements Measurements Logging 2 x quad-core 2.8GHz CPU 8GB RAM CTRL CTRL 11.4TB usable Clustered NAS shelf 14x146GB FC disks Clustered NAS shelf 14x300GB SATA disks Service Availability New infrastructure has high-redundancy for high-availability Deploy each service on a dedicated Oracle Real Application Cluster The use of a standby database will be investigated objective of reaching 100% uptime for LSA Secure database account granting specific privileges to dedicated db accounts DIAMON agent on Oracle Application Servers For all CO databases CO puts the UR, and pays for the hardware IT chooses the hardware, hosts, supports and maintains Database: CO requirement IT chooses the material,CO pays, IT hosts and maintains and supports without extra cost SLA of IT to the suppliers, Dell 24x7 support included in the purchase cost, after 3 years we need to decide disks working hours 12 June 2008 E. Hatziangeli AB/CO

9 Industrial Controls Industrial Controls for LHC have reached a high level of maturity All systems, fully deployed for HWC in 2007, are presently in their operational version Machine protection (PIC, WIC, QPS, Circuits) Collimator Environment Monitoring Package (temperature, water cooling) Survey Cryogenic controls Cryogenics Instrumentation Expert Tools (CIET) Most of the SCADA applications have been ported to Linux The front-end FESA software has been ported to Linux Migration to last version of FESA (v2.10) to be done for next shutdown The interface toward logging database has been consolidated DIAMON is used for diagnostics PLC agents are available, tested and ready to be deployed PVSS diagnostics will be soon available Most of the SCADA applications have been ported to Linux (and so usable from all operator consoles) CIET provides part of the data to the main Cryo controls (temperature) SCADA system, on top of FESA, on top of FIP GW (CIET, QPS, Survey) Rest of industrial controls are the classical PLS and the SCADA on top to control the system Maturity : 1 year of operations encountered several problems and solved many bug Linux: one uniform platform for OP 12 June 2008 E. Hatziangeli AB/CO

10 Cryogenics Control System
Local & Central Control Rooms SCADA Data Servers LHCA LHCCA LHCB LHCCB RM PA Profibus DP WorldFIP Return Module S78 & S81 Main Dryer Comp 4.5K Comp 1.8K LN2 Buffer Comp 1.8K Comp 4.5K Main Dryer Surface QSAA QSCA QSCCA QSCCB QSCB QSDN QSAB UCB 4.5K Cold Box 4.5K QSKA QSRB QSRA Shaft QURA Several system distributed around the machine that need to constantly interchange information via Ethernet Same color indicate same component CB 1.8K Cavern CB 1.8K QURCA QURCB QUI Connection Box Alcoves RM78 Sector 78 (3.3 Km) Sector 81 (3.3 Km) RM81 12 June 2008 E. Hatziangeli AB/CO Tunnel

11 Cryogenic Controls Reliability
The operation of cryogenics sectors has revealed a high risk dependency of the cryo control system on the reliability of the Technical Network Steps taken to reduce the dependencies PLC architecture was rationalized  no dependency on Ethernet of the cryogenics control loops for production equipments Architecture of the network components was optimized  minimum dependency on communications equipment (switches) Powering of network component was checked homogenization where possible with the cryo powering Work ongoing Identify the weak network components and improve (fiber–copper) Consolidate the restart of communication after a network failure Ensure interventions on Technical Network (hardware & software) are carefully planned and agreed with Operation All components are put in the same switched when they were in several All powered with proper power and all switched from the same Consolidate the restart of communication after a network failure (connection between GUIs and server on loss of communication is not correct they block Tunnel exception WorldFip based architecture imposed the use of FEs and then a Ethernet connection between those FEs and the CRYO PLCs. This communication is unavoidable with this architecture and conditions strongly the reliability of the controls. The Ethernet dependence is also influencing FEs: boot server CIET, QPS comms: CMW name server 12 June 2008 E. Hatziangeli AB/CO

12 Machine Interlocks Beam Interlock System Powering Interlock System +
for Protecting the Equipments for Beam Operation for Protecting Supra-ConductingMagnets and Normal Conducting Magnets Beam Interlock System (VME based) Powering Interlock System (PLC based) + Safe Machine Parameters system (VME based) Warm Magnet Interlock System (PLC based) Fast Magnet current Change Monitors (FMCM) 12 June 2008 E. Hatziangeli AB/CO

13 Powering & Warm Magnets Interlocks
Powering Interlock Controllers 36 units of PLC based system protecting ~800 LHC electrical circuits monitored via PVSS Supervision Operational and daily used during HWC Warm magnet Interlock Controllers 8 units of PLC based system protecting ~150 LHC normal conducting magnets monitored via PVSS Supervision Operational and daily used during HWC 12 June 2008 E. Hatziangeli AB/CO

14 Monitored by Operational Application
Beam Interlock System Beam Interlock Controllers 19 VME systems and ~200 connections with most of the LHC systems Monitored by Operational Application User Permit is received by all systems which give an input to BIS Everything is installed and tested (System tests) Integration tests in progress. Test are done point by point No RBAC for the moment, BIS is self protected and the safe beam flag will unmask the signals automatically at the hw level if the condition change and we have high intensity or energy. Individual System Tests successfully performed on going BIS Commissioning (involving all User systems) done in // with HWC 3/8 points already performed BIS will be ready for the machine checkout… 12 June 2008 E. Hatziangeli AB/CO

15 (TT40 incident in 2004) 12 June 2008 E. Hatziangeli AB/CO
No marks or damage on magnet flanges Beam Vacuum chamber cut (outside view) ~110 cm (inside view) Ejected material opposite cut (inside view) 12 June 2008 E. Hatziangeli AB/CO

16 Fast Magnet current Change Monitors - FMCM
Successful collaboration with DESY DESYdevelopment + CERN adaptation First units successfully used during the SPS Extraction tests and CNGS runs in 2007 currently being re-commissioned for 2008 runs Installation and commissioning in progress 12 monitors deployed in the LHC (+ 14 in Transfer Lines), including ALL septa families LHC installations to be completed next month (12 devices) 1st version of FESA class and Java supervision available since June 2007 minor consolidation work in progress View of FMCM board I (A) Must talk to Brunofor the GUI shots ********* The FMCM measures difference in current in a very short timescale. Min deviation that FMCM triggers is ~ 2x10-4 FMCM will be one of the systems to send PM data to the PM server in case of an event. FMCM: reaction time and % current change detection level => values were set out of simulation failure of supplies by simulating exponential current decay and what happens to the beam in those cases So different values were for different types of magnets. In general is ~ 3x10-4 of Nominal current (0.1per mille of nominal ) up to 5x10-3 for other types Reaction time starting from 50 microsecond up to 1milli second, but it will trigger as soon as it see the change of current. I (A) 10 ms 500 ms FMCM <103 PC current FMCM trigger  0.1% drop ! 12 June 2008 E. Hatziangeli AB/CO time (ms) time (ms)

17 Timing System major components
The LHC central timing Master, Slave, Gateway using reflective memory, and hot standby switch The LHC Injector chain timing (CBCM) Master, Slave and Gateway using reflective memory, and hot standby switch Timing is distributed over dedicated network to timing receivers CTRx in front ends LHC and SPS safe machine parameter distribution 12 June 2008 E. Hatziangeli AB/CO

18 Safe Machine Parameters
The SPS and LHC safe beam flags and beam energy are distributed on the LHC timing network Work needs to be done for the final system to be ready The CTR timing receiver modules are able to distribute the beam energy and Safe Beam flags without any software to ensures higher reliability All timing receivers are monitored by DIAMON Powerful diagnostics for receivers simulated beam energy is distributed on the LHC timing network 12 June 2008 E. Hatziangeli AB/CO

19 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary 12 June 2008 E. Hatziangeli AB/CO

20 Post Mortem - Towards Beam Commissioning
Global PM Analysis:Global Event sequence, summaries, advised actions, event DB,… Global event sequence Machine Protection OK Advised Actions Validation of machine protection features Pre-analysis of PM buffers into result files Flagging of interesting systems/data reduction Database catalogue Individual System Analysis & Checks IPOC-BIS Event Sequence BLM, BPM > threshold Circuit events I/XPOC Data completeness and consistencycheck at system and global level (minimum data, configurable) 0) Specification are written for data completeness and consistency defining the amount of data in case of a global post mortem event. This means did we get everything ensuring all appropriate data is there in a common place. This is the most important thigh to have place for the beginning with some viewers to view the data. With BI the work is advanced as the use the same template to send data to the PM (50MB data fpr the BLMs) 1) DATA: BLM, BPM tests data, Circuits events, XPOC 2) First analysis mechanism to be in place 3) Global analysis is for later and Advised action (ready to take beam again or not)-> SIS to allow or block injection Concerns PM data could be lost if there is a power cut in the FE but maybe possible to reconstruct of having the other devices of the same type Even BIC is not protected against this Upon beam dump / self triggering, systems start pushing data to PM system, Logging, Alarms, etc… BLM BPM FGC QPS PIC/WIC BIS XPOC FMCM 12 June 2008 E. Hatziangeli AB/CO 20

21 Post Mortem Readiness Many tools are ready for HWC and tested on sectors 5-6 and 7-8 Implemented since December 2007 Automatic Test Analysis on three applications (expert override possible) Calculated result parameters sent to Sequencer for MTF upload. Well defined GUI for each test step. Test results electronically signed by role using RBAC Event recognition with Event Builder Redundant services for PM collection and data Scalability tests with first beam clients started (BLM, BPM) To do for 2008 Parallel sector commissioning still to be tested For the 600A circuits many steps are still to be automated. Further validation tests with beam clients to be done (BLM, BPM, RF, etc…) Extend framework from HWC to beam operation Implement higher level of Automated Test Analysis Data completeness checks & Individual system tests Automatic tests analysis (PNO) Test results electronically signed by role using RBAC (5 different roles defined) Event recognition with Event Builder (Quench, Fast Power Abort, …). Redundant service (both receiving the PM data at the same time from the equipment) Analysis tools get data from master. (2 HW services in Meyrin and CCR) Up to now only 2 fronts in ll was Ok but several sectors in the near future might bring surprises but we do not anticipate any issues from experience. Will there be bottlenecks? Implement 1st and 2nd level of automated analysis (Data completeness checks and Individual system tests for main clients) 12 June 2008 E. Hatziangeli AB/CO 21

22 Readiness of HWC sequencer
First version of the sequencer deployed in early 2007 Many new versions with improvements deployed since HWC Sector 7-8 May-Jul Winter 08, 5-6Spring 08 ~ 35 sequences written and maintained by 3 HWC experts Sector 4-5: over 1700 sequences executed, in 5-6 over 600 Essential tool for HWC Overall it works well and satisfies the requirements Sequencer (the tool) is complete, no important new features needed Sequences (the tests) are maintained by HWC experts Ready for multi-sector / multi-front HWC Used in multi-front operations for over a year Recent experience in multi-sector operations “normal” HWC is done in sector 78 training quenches are done in sector 56 No scalability issues are anticipated Experience in multi-sector is more recent: for the last couple of weeks, “normal” HWC is done in sector 78 while training quenches are done in sector 56 on the Main Bend magnet. However, I don’t forsee any scalability problem for the sequencer at all. HWC Sequencer Overview 12 June 2008 E. Hatziangeli AB/CO

23 Logging Service Readiness
Logging for Operation Data logged from PS Complex, SPS, CNGS, LHC HWC, LHC , any type of equipment Processes run continuously on dedicated machines Monitored through Alarms system LASER & DIAMON & diagnostic application Logging for equipment commissioning Dedicated service, running in the environment of the specialist Aim: validation of the equipment behavior before operational deployment No interference with Operational logging Requirement for a watchdog system (coming weeks) For critical data (INB, CNGS neutrino events, ...) continuous monitoring of data logged in DB, generation of a specific alarm 12 June 2008 E. Hatziangeli AB/CO

24 Software Interlock System - SIS Overview
Very useful system to anticipate failures and gives early alarms Accommodates complex interlock logic Complements BIS (hardware) as protection system Proved to be reliable tool for operations Excellent experience in SPS (900 parameters monitored) 12 June 2008 E. Hatziangeli AB/CO

25 SIS for LHC Gives 2 Permits for Injection BICs(Beam1 & Beam2)
All PCs not HW interlocked (~ 800, orbit correctors, warm magnets) Current of separation dipoles and MCBX orbit correctors Ring & injection screens (only IN when mode inject-dump) Extraction screens Circulating beam intensity limit Gives Permit for the LHC ring (dumps the beam - initially alarms) Integrated field of orbit correctors (beam dump energy tracking) Extraction screens combined with intensity + energy Orbit at TCDQ Future work RBAC integration Critical settings monitoring (MCS) from LSA Refinement of the configuration as we progress with LHC Current of separation dipoles and MCBX orbit correctors (IR1, 2, 5, 8 exp. protection) MCBX – orbit correctors MCS - management of critical setting 3 Permit trees (2 inj and 1 ring) 12 June 2008 E. Hatziangeli AB/CO

26 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary 12 June 2008 E. Hatziangeli AB/CO

27 Role Based Access (RBAC) Overview
Need to prevent Well meaning person from doing the wrong thing at the wrong moment Ignorant person from doing anything at any moment Application RBAC RBAC Token: Application name User name IP address/location Time of authentication Time of expiry Roles[ ] Digital signature (RBA private key) CMW client FESA CMW server Access MAP T Server Configuration DB Authentication: User requests to be authenticated RBAC authenticates user via NICE user name and password RBA returns Tokento Application Authorization: Application sends token to Application Server (3-tier env.) CMW client sends token to CMW server CMW server (on front-end) verifies token CMW server checks AccessMapfor role, location, application, mode 12 June 2008 E. Hatziangeli AB/CO

28 Management of Critical Settings - MCS
Need to ensure Critical parameters, which can compromise the safety of the machine can only be changed by an authorized person and nobody else are what they are supposed to be MCS ensures Critical parameters are only changed by authorized person RBAC for Authentication & Authorization It signs the data with a unique signature to ensure critical parameters have not changed since the authorized person has updated it Public-private key digital signatures 12 June 2008 E. Hatziangeli AB/CO

29 LHC Controls Security Panel - LCSP
The LHC Controls Security Panel is mandated to address all the technical and non-technical issues concerning AB security for Controls Take responsibility for the RBAC data (ROLES and RULES) Ensure all critical parts of the machine are protected Take responsibility for the CNIC actions reduction of Trusted list, change of operational account passwords,.. CNIC: Computing and Network Infrastructure for Controls 12 June 2008 E. Hatziangeli AB/CO

30 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary See next talk 12 June 2008 E. Hatziangeli AB/CO

31 LHC controls infrastructure Scalability
Tests =>Preliminary results, issues & foreseen solutions Systems which scale to LHC full load Software Interlock System SIS Data Concentrators (BLMs, BPMs) Alarm system LASER (new architecture) Controls Middle Ware (CMW) /Java API Parameter Control (JAPC) Diagnostic & Monitoring tool DIAMON Systems potentially critical (tests ongoing- results mid June) Post Mortem Logging service Scalable to LHC load from all clients except LHC BLM Preliminary limit : ~ 5000 parameters/second Bottleneck : SQL calls management by Oracle Server 12 June 2008 E. Hatziangeli AB/CO

32 Failures in Central Timing
Tests have been performed to validate the behaviour of the Controls Infrastructure when the Central Timing crashes These “crash” timing tests are on going Results The behaviour of the control system with no timing is correct The application programs, servers and front ends recovered without manual intervention when timing returns 12 June 2008 E. Hatziangeli AB/CO

33 RBAC Dry Runs The LHC Controls Security Panel (LCSP) is preparing an RBAC dry-run end June/early July The RBAC default behavior is changed to “Access with no RBAC token is refused” Property not protected is not authorized All Equipment servers will be loaded with RBAC access maps Typical applications will be tested LHC 2-tier & 3-tier applications LHC Core controls (LSA) Background servers, concentrators Fixed Displays 12 June 2008 E. Hatziangeli AB/CO

34 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary 12 June 2008 E. Hatziangeli AB/CO

35 Alarms (LASER) for LHC An important increase in expected alarm events
Required availability 365 days/24h Alarm console extended Allows for dynamic grouping of alarms New alarm definition database schema Ensures the data quality by reducing redundancy and protecting against incomplete data Alarm server modified fundamentally to allow Fast response to increase in load Increased resilience to external failures and improved diagnostics tools LASER console LASER CORE LASER source LASER DB 12 June 2008 E. Hatziangeli AB/CO

36 DIAgnostic & MONitoring System - DIAMON
DIAMON provides Software infrastructure for monitoring the AB Controls Infrastructure Navigation Tree Group View Monitoring Tests Details Repair tools Easy to use first line diagnostics and tool to solve problems or help to decide about responsibilities for first line intervention On the left via the navigation tree the OP have access to the various groups of monitored items (grouped by Acce, location, function, …anything allowed) In the group view the summary state of each item in the group is indicated The detailed panel provided information for the selected item – Op can make the 1st line diagnostic through this and proceed to repair 12 June 2008 E. Hatziangeli AB/CO

37 Outline LHC controls infrastructure – overview
Status Report on Core Controls Front ends Hardware and Software Databases Industrial Controls Machine Interlocks The LHC timing system Core Services and Applications Post mortem Sequencer for HWC See next talk for Beam Sequencer Logging Software Interlock System LSA (see next talk) Controls Security Role Based Access Management of Critical Settings LHC Controls Security Panel Controls Infrastructure Tests Deployment on LEIR, SPS, LHC TL Dry runs - Commissioning Scalability Tests LHC Timing Crash Tests RBAC tests Monitor and Diagnostics LASER DIAMON Injector Renovation Summary 12 June 2008 E. Hatziangeli AB/CO

38 Injector Controls Renovation - Status Report
Injector Controls Architecture - InCA Architecture validation with critical Use Cases Check interfacing of the various components LSA core Standard CO components for Acquisition Standard PS and LSA Applications interfaced to the core Check data flow Low-level trim and monitoring values of correctors Orbit correction using LHC Beam steering application High-level trim + drive a front end Results Whole data flow validated Architecture closer to the final one Injector Complex FE Renovation – 2nd half of 2008 A "strategic" plan for the renovation of the FE controls infrastructure is due by mid-2008 Development and validation of new Front-end solutions in view of their first deployment in 2009 don’t check the data validity 12 June 2008 E. Hatziangeli AB/CO

39 Summary The LHC controls infrastructure had been targeted for readiness for an engineering run at 450 GeV in November This goal has been met. The ongoing hardware commissioning and the extensive use of programs and databases (“learn by doing”) have significantly changed the specifications and the resulting follow-up and work has been done. Additional functionality has been prepared in network security (RBAC) - diagnostic tools (DIAMON) 12 June 2008 E. Hatziangeli AB/CO

40 End 12 June 2008 E. Hatziangeli AB/CO

41 Accelerator Databases
Off-line Databases Data not needed directly to interact with the accelerator Database serves as source for on-line databases Synchronization mechanisms propagate the data for consistency Many database services, e.g. Layout database Online Databases Data needed instantaneously to interact with the accelerator Database is between the accelerator equipment and the client (operator, equipment specialist, software developer) Many database services, including APIs, and applications LSA – Accelerator Settings database MDB – Measurement database LDB – Logging database CCDB – Controls Configuration LASER – Alarms database Part of the offline db Assets (MTF) Documentation (EDMS) Drawings (CDD) Part of the online database TIM – Technical Infrastructure Monitoring database E-Logbook – Electronic Logbooks CESAR – SPS-EA Controls 12 June 2008 E. Hatziangeli AB/CO

42 Offline Database Readiness Status
Layout database Scope includes accelerator equipment (ring, injection & dump lines) and related electronics (tunnel and surface buildings) and field bus connections Racks & electronics incorporated up to the level of detail of the equipment group requirements Layout data is now used as foundation for the controls system Still to do More data is being captured relating layout and assets information Tools for data maintenance still to be put in place – data needs to be kept up-to-date by the equipment owners Layout DB: (Cryoinstr, BLM, BPM level of details very high (navigation down to the lowest level of details down to the card installed) More data is being captured relating layout and assets information=> important as machine goes to exploitation 12 June 2008 E. Hatziangeli AB/CO

43 Database Readiness Status
Operational Settings Highest-availability, Fast response Data model enhanced to cover functional extensions for Role Based Access (RBAC) , XPOC, Sequencer PS Controls Renovation requirements (InCA) Logging Service High-availability Currently >109 records/day ~30GB/day ~10TB per year New database hosting in place since March 2008 Common logging infrastructure for the complete accelerator chain Sustained increasing logging requirements for HWC and beam related data Improved data retrieval tool (TIMBER) in line with end-used needs Now we have ~5 TB of logged data HWC will go down and beam operations will increase in data. Operational Settings: last year model stabilized and this year enhanced to cover InCA requirements, + RBAC, XPOC => confidence in the model) Logging Sustained increasing logging requirements for HWC and beam related data – currently +109 records/day ~30GB/day (BLMs tests with beam related data….) today above 1 billion records a day with its timestamp or 30 GB => 10 Tera bytes per year…) The main issue here is to indentify correctly the important information that needs to be in the database, level of confidence increase data should decrease and stabilize I need a picture to relate the part What happened if db is off line and not Operational settings is the most critical because ether LS operators will not be able to controls and sett the devices Statistics (23 Aug 2007) LHC LOG scheduled outages time took and impact on operations – 1 occasion per year from min to some hours there were no settings. If no setting in LHC the OP will dump depending in which part of the cycle exists and HW interlocks will kick in if there is We do not need the DB to stop the beam – promise to look in the stand by databases (< 10 Gbytes data for LSA) double the infrastructure and do not use just a failover in catastrophe case. Everything is now clustered since march , (see the slide in 8th May TC) mother board exchanged, things happened without the end users noticing, disk are hot pluggable, double everything,.. Scheduled interventions (quarterly security patches) by IT are agreed for convenient time for OP. Logging is buffered before they send it to the logging DB to the measurement DB If all DB goes down this indicated a serious 12 June 2008 E. Hatziangeli AB/CO

44 SMP: Safe Machine Parameters
Generation and distribution of critical parameters over the Timing network Interfaces with: ← BETS (Beam Energy Tracker system) ← BCT system BLM system Kicker system Experiments Beam Interlock System Beam Energy, Safe Beam Flags, Stable Beam Flag, etc… BLM receive the energy to adapt the thresholds accordingly Kickers to adapt the strength of the kick to the energy of the beam Experiments to judge the right time to put their movable devices in the beam BIS using the safe beam flag to know if some of the BIS signals can be masked by OP (low energy and low intensity) For SPS generated flag will indicate safe extraction. Timing is sending the safe Beam Flags , 10HZ aand there is a timeout in the BIS system and when it does not receive the flag it makes the flags unsafe. SPS TI8 is ready and installed and tested during the dry run SMP is also critical with high intensity – basic version - for the SPS: First version already in operation for the LHC: First version will be installed in June 2008 Final version for both machines are planned for 2009 Two SMP variants: one VME controller for LHC another one for SPS View of generic SMP board 12 June 2008 E. Hatziangeli AB/CO

45 Post Mortem for HWC MTF LSA
Event Analyser PM events waiting To be signed by experts HWC sequencer Result of test Result Current cycle MTF LSA The sequencer is currently used in 3 different places: HWC, SPS and LHC dry runs. This slide shows its usage for HWC, laters slides will show the SPS/LHC usage. The light blue parts are the same as on the slide before, The brown parts are specific to the HWC usage, and are going to be explained next. Let’s first explain how a typical HWC test works, The purpose of HWC is to check if electrical circuits work well. A circuit is composed of one or more power converters (FGC), the PIC and the QPS. A typical HWC test sequence executes a series of test steps and checks if everything works fine. 1) The sequencer starts the HWC tests sequence by loading information about the tests to be executed and parameters of the electrical circuit from the LSA DB. 2) Then it sends commands to this equipment to drive it through a series of tests steps. It may take some measurements and provokes a PM event. If a PM event is triggered, the equipment sends data to the PM analysis system, where experts check if everything went as expected. When the sequence has finished, the sequencer collects all results of the test (amongst others from the PM analysis). This data is stored into the LSA DB for real-time follow-up and sent to the MTF system for long-term archival. The HWC sequencer is sending the current cycle to be executed by the equipment Current cycle produces a PM event (+ logging data) which are stored in their corresponding DB Sequencer sends the current cycle to event analyser. Sequence tells the EA which tests are run on which circuits. The list of the PM event waiting to be analysed are displayed. The OP will start the test analysis application and pass or fail it (and sign it) The selected test shows two experts having signed the test and one still has to sign. RBAC is used for the signature RBAC protects that the right people sign for the corresponding tests – no cross signatures allowed One of the analysis application is thePiNO (Power To Nominal) for the current to nominal The red curve shows the current ramp to nominal. V_Meas is the voltage measured on the PC (the jump seen at the beginning of the ramp is the inductance seen by the circuit seen at the beginning of the ramp U_LEAD voltage on the leads. Based on these we can measure the resistance of the circuit.The lead voltage sees only the resistive part of the circuit (increase a bit on the plateau due to heat) Notice that testing of the limits are done during the analysis of the data and the green colour indicates specific tests are passed but the OP will define the end result (OP assisted analysis) Expert sign by clicking pass or fail The second screen shot shows the electronic sign window: On any Analysis application, clicking on the Pass or Fail button, this window opens and the expert enters his NICE login name and using RBAC we check if this person has the correct role to sign for the test. This protects against errors and gives traceability. Viewing & Test analysis applications Data collection, pre-processing and storage Equipment Post Mortem Logging DB 12 June 2008 E. Hatziangeli AB/CO

46 The Logging Service LOGGING MEAS DB DB LOGGING DB
Contain relatively slow time series data Store data online for at least 20 years MEASUREMENT DB Raw time series data, up to 2Hz rate Store data for 7 days Send filtered data of interest to Logging DB each 15 minutes LOGGING DB Filter each 15 min MEAS DB LOGGING process LOGGING process LOGGING process 60 106records / day DEVICE CMW METER browsing & data extraction TIMBER browsing & data extraction LOGGING : resp for acquiring data from the devices of the machine buffering and writing them to the DB Eventually perform data processing on data Objective: by extracting data from the Loggting DB, end users are able to do a post mortem analysis of a specific event which occurred at a given time in the machine. For ex. to be able to reconstruct an accident by correlating a beam loss to a glitch of the current in a power converter. Configuration of JAPC MM: 2 ways of configuring (list of devices to be monitored) JapcMMdepdg on the client LHC HWC from a pre-defined list of devices in the LSA DB for LHC HWC (necessary as the device set changes from one test to the next) Xml file (for other clients where the device set is static) Config of logging: from a config file: to map the value received from the japcMM to the corresponding variable in the DB Databases: Very close collaboration with ABCODM specially with Chris. 2 DBs + JAVA API between logging and Measurement DB: under DM responsibility. MEAS DB: designed for short term storage of data and allows high frequency loading. LOGGING DB: data flushed every 15 min from MEAS DB to LOGGING DB.Transfer of long term marked variables to Logging DB for permanent storage Based on JapcMMresp for subscription mechanism JAPC Intensive usage of factories by Jakub use of japc multiple factories by Roman: TGM Intermediate servers SPS sems DIM Connection to Ilya DEVICE DEVICE 12 June 2008 E. Hatziangeli AB/CO

47 LCSP – RBAC actions Simplification of the definition of RBAC Rules
Introduction of device and property grouping Reduce number of rules and enhance maintainability of RBAC data Equipment State definition Equipment will only have 2 RBAC states : OPERATIONAL or NON-OPERATIONAL This state is derived from the LHC Machine Mode distribution Standardize the ROLES definition : OP group -> responsible for OP roles Roles used for get/set/monitor when STATE is OPERATIONAL Equipment groups -> responsible for expert roles Roles used for get/set/monitor when STATE is NON OPERATIONAL A PIQUET role per equipment group, OP managed => all rights all time Roles are empty Filled only by OP when an intervention is required Equipment State definition : first version is foreseen for mid-June 12 June 2008 E. Hatziangeli AB/CO

48 LCSP – CNIC actions The generic accounts (more than 500) are reviewed with the aim of reducing its number and forcing more frequent password changes Enforce the CNIC policies by reducing the number of Trusted hosts Replace their functionalities with either Clusters of Windows Terminal Servers (WTS) Locally managed consoles where AB/CO manages the installation Trusted hosts that have access to the Technical Network from the General Purpose Network Software packages and patches by AB/CO on Locally managed console 12 June 2008 E. Hatziangeli AB/CO


Download ppt "Control Status Readiness Report"

Similar presentations


Ads by Google