LHC Post Mortem Workshop - I

This talk will cover the following items:
 -  PM data model
 -  Client API
 -  PM server
 -  Data processing and SDDS conversion
 -  Performance and scalability
 -  Current status

The presentation will cover:
- Data arrival and event building
- SDDS format and its implementation for PM
- PMX method for data description and control
- SDDS converter - generic version
- Possible enchancement of the converter
- LabVIEW application/framework for individual data module (PMM)
- PMM data locator
- PMM SDDS ascii/binary loader
- Internal data classes
- Data viewing
- Data analysis
- Automatic analysis
- Diagnostic tools
- Conclusion

LASER will provide alarm event information to the PM 
system in the case of a PM event. A first solution, 
agreed between the LASER and the PM teams at the 
end of 2005, will be described. Since then, the LASER 
system has evolved which opens up other possibilities 
to integration. These solutions will be discussed as well 
as the questions they give rise to.

This presentation will explain briefly the purpose, scope
and architecture of the LHC Logging Service.  More detail
will be given on the interaction with the Post-Mortem system
including naming conventions and enforcement, data lifetime
policy, combining and correlation of slow logging data and
external transient data.  Finally some ideas and
possibilities will be discussed such as the use of the
Measurement Service and storing of PM summary information.

The commissioning of the warm part of the superconducting
circuits of the LHC started in 2005 with the short-circuit
tests of the power converters where the non-superconducting
elements of the circuits are being commissioned together
with their associated general services.
Once the circuits are at their operation temperature
and before powering them, the interlock system will be
validated (PIC tests). 
 The overall commissioning of the superconducting
circuits will start in February 2007 with the first powering
up to nominal current of all the magnets in Sector 7-8. 
 This talk will introduce the sequence of steps and
detailed procedures which lead to the powering of the
different superconducting circuit types, the powering
strategies designed to be ready for 450GeV beam
commissioning on schedule and the needs of the hardware
commissioning team for diagnostics and to ensure the
integrity of the hardware.

Effective commissioning of the LHC hardware demands a
well-designed set of high level software tools, which is
required for the equipment performance analysis and
validation. The challenge includes a large amount of
equipment integrating heterogeneous systems like powering,
energy extraction, distributed magnet protection systems,
cryogenics and vacuum with their distributed instrumentation
as well as the technical services. Various operational
conditions must be dealt with like the superconducting
magnet quench phenomenon and quench effects, including their
constraints on the next powering cycle while respecting the
destructive power stored in the magnet system. The level of
the commissioning of the main ring superconducting magnet
system will depend not only on the time allocated to the
commissioning, but also on the availability of the high
level software analysis tools.
 The required tools for various phases of the LHC
start-up will be elucidated and discussed. The role of newly
created Main Ring Magnet System Performance Panel (MPP), in
view of the definition of the high level software tools for
the equipment commissioning and performance analysis will
also be briefly addressed.

Three components of the Post Mortem Analysis are already
used by the equipment support teams. This talk will present
the status and the modes of operation for each of them. Then
the present architecture will be detailed, followed by the
implementation dedicated to the Hardware Commissioning.

The first Post-Mortem requirements have come from the
needs of the individual systems involved in the first phase
of the hardware commissioning using short circuit tests. The
second phase of powering the circuits, involving systems
such as vacuum, cryogenics and DFB’s will extend the
requirements of analysis to a new scale.
 This talk will show how we plan to include these new
analysis requirements into the present framework, how it
interfaces with the sequencer and how the analysis could
trigger on spontaneous events. Important aspects, such as
modularity, flexibility, sequencing and scalability will be
covered.

For each beam injection into the LHC a well-defined 
series of beam quality checks needs to be made, 
starting in the SPS just before extraction and in the LHC 
immediately after injection. These checks will be 
dependent on the beam type, intensity and position in 
the filling sequence, and will use transient data which 
must be acquired and analysed at the appropriate time 
and within a specified time window. The requirements  
in terms of functionality, response times and scope are 
described, and the equipment subsystems identified. 
Potential issues are discussed.

Each dump action must be followed by an XPOC which 
is launched automatically and is designed to verify that 
the dump was correctly executed. If an anomaly is 
discovered during these tests, the XPOC must withhold 
the User Permit to the BIS (via a software channel). The 
XPOC comprises beam instrumentation and other 
signals which will come from the logging and Post-
Mortem system, or direct from the equipment.
 The XPOC must be triggered by the dump action, must 
retrieve and analyse key data and make a comparison 
of the relevant parameters against specified reference 
values, and then give or withhold the User Permit 
according to the result. The requirements in terms of 
functionality, response times and scope are described, 
and the equipment subsystems identified. Data types, 
reduction, volumes and rates are estimated.

After an emergency dump a general Post-Mortem request will
be issued to acquire transient data from a variety of
systems. The analysis of a Post-Mortem event may take from
minutes to many months, depending of the desired level of
details. Key data must be however presented in a way which
allows for simple and efficient fault-finding. Operation
crews must be presented with clear information to indicate
of operation may continue or if expert interventions are
required after the emergency beam dump. Key equipment and
instrumentation data required to identify the source and
causes of an emergency abort are described. Various
experiments and measurements will also require the
possibility to make ad-hoc acquisition of some transient
beam and possibly equipment data, in order to diagnose and
solve specific problems and to cope with unforeseen
difficulties. An attempt is made to outline the different
transient data required for general operational purposes,
together with the requirements for triggering and
acquisition which are distinct from the general Post- Mortem
data.

In addition to systematic transient data acquisition, 
operation of the LHC will also require the possibility to 
make ad-hoc acquisition of some transient beam and 
possibly equipment data, in order to diagnose and 
solve specific problems and to cope with unforeseen 
difficulties. An attempt is made to outline the different 
transient data required for general operational 
purposes, together with the requirements for triggering 
and acquisition which are distinct from the general Post-
Mortem data.

A post-mortem timing event distributed by the LHC 
machine timing system is used to freeze the PM buffers 
of a large fraction of the LHC equipment. This event 
must be generated automatically whenever the BIS is 
issuing a beam dump request by changing the state of 
the beam permit signal. This presentation outlines the 
present ideas on how to generate the PM timing event. 
The issue of PM event suppression in the case of single 
beam dumps or special operation modes like 'inject and 
dump' will be addressed.

Post Mortem will be the key to mastering the full 
complexity of LHC Operation and the interaction 
between systems. Many systems will be involved in full 
optimisation and understanding of performance. Today 
a few systems are providing data to validate and 
understand hardware commissioning. This must be 
extended giving priority to obtaining essential 
information related to achieving first collisions. 
 This talk will review systems involved, discuss the 
nature of the information to be provided and attempt to 
identify some priorities. The vacuum system will be 
examined to demonstrate how these demands are 
being met.

The RF acceleration (ACS) and transverse damper (ADT) 
systems will supply post-mortem data at various 
acquisition rates. The PLCs controlling the power 
systems acquire at a few Hz, while high-speed 
digitizers and acquisition buffers embedded in the low-
level hardware acquire transient signals at 80 
MSamples/s over time periods ranging from a few 
milliseconds to several hundred milliseconds. The high 
speed acquisitions in particular will result in high 
volumes of data, and some local data analysis and 
reduction may be necessary to alleviate this. An 
overview of the available data signals will be 
presented, along with tentative requirements on data 
analysis, logging and alarms.

Reliable operation of LHC injection, tune/aperture and 
LBDS kicker systems relies on continuous on-line and 
off-line surveillances of their critical operational 
characteristics. Different acquisition techniques like 
trends logging, shot-by-shot logging or fast transient 
recording will be used to acquire and record the diverse 
types of signals existing within kicker systems. 
Correlation between the acquired data will be done 
through a precise time-stamping of the data acquisition 
time coupled with an internal management of the 
possible acquisition trigger sources.
 The structure of the different post-mortem buffers 
will be presented for each kicker system with estimation of 
their volume and a description of the different acquisition
analysis and recording mechanisms. In addition, the
triggering logic will be described and the remaining
open-issues linked mainly to the distribution of post-mortem
event(s) will be highlighted.

The LHC collimation system is responsible for providing 
clean beam conditions and hence to assure the 
protection the equipment in the LHC.  A failure of the
collimation system may trigger a beam dump to avoid magnet
quenches.
 The post mortem data of the collimation system 
supplies the following information
 •	Demanded and actual positions of all 
collimator jaws (millisecond accuracy)
   Note: information on the actual positions is 
provided by resolver, position and gab lvdt's as well as end
switches and anti-collision switches).
 •	Temperatures of the jaws
 •	Jaw vibrations over a period of a few seconds 
before and after the beam dump
 •	BLM transient data during a collimator 
movement.
 •	Command history
 The first analysis of the collimator post mortem 
data must assure that there were no internal failures in 
maintaining the actual collimator positions.
 A second analysis in combination with information 
from beam loss, beam position and beam profile monitors 
should validate that the collimation efficiency was as 
required.