Approaching the design of ILC conventional magnets for the RDR

1
Approaching the design of ILC conventional
magnets for the RDR

• Cherrill Spencer, ILC
• 25th April 2006

2
Current General Layout of a 500Gev ILC
Schematic borrowed from Fred Asiri of
Conventional Facilities new labels added.
31 km overall
superconducting linac
Part of e source, e booster auxiliary source
Beam Deliveries 2 Interaction Regions
e- Damping Ring
TWO e Damping Rings
Cherrills rough count totals 12,600 magnets
150 styles! About 600 quads in the main linacs
75 others will be superconducting. So approx
11,925 conventional magnets.
3
Recall our philosophies for
designing and costing NLC magnets, power supplies
power cables

• NLC magnet designing and estimating of costs
  took place from 1999 to 2003
• Recall our approaches for NLC and consider which
  might be appropriate for the ILC magnet
  designing and costing exercise towards writing
  the RDR
• First cost estimate in 1999 had to be done in a
  few months like we have to now, then we revised
  our counts and costs several times over next 4
  years.
• Look at some old NLC presentation slides from
  July 2000 next 12 slides

4
Roles Responsibilities of the Magnet
Cost-estimating Group 1/3

• ROLES
• Define technical scope of all magnets, power
  supplies, cables and cable trays
• Develop NLC-wide design, manufacturing and QC
  philosophies
• Develop NLC-wide permanent electromagnet
  costing guidelines
• incorporating Management Grp guidelines
• Develop cost estimating processes to fit in with
  current WBS structure -includes validation
• Apply cost estimating processes
• Check our costs been assigned correctly in the
  appropriate WBS elements

5
Roles and Responsibilities page 2/3

• RESPONSIBILITIES
• Following magnet team engimators are responsible
  for estimating the MS and labor costs for all
  activity phases up to, but not including
  installation of listed items
• Ponce Rodriguez- all areas DC cables, magnet
  related IC cables, all cable trays
• Wes Asher- Power Supplies auxiliary equip for
  damping rings beam delivery
• Steve Lowe- all PCD coordination drafting
• Val Nesterov- Power Supplies auxiliary
  equipment for injector main linac
• Bobby McKee- Magnets for damping rings
• Carl Rago- Magnets for injector main linac
• ( No-one) - Magnets for beam delivery
• Cherrill Spencer- responsible for developing or
  interpreting costing guidelines, negotiating
  treaty points, acquiring validating data,
  scrutinizing team member estimates, checking how
  our estimates are assigned in the WBS

6
Roles and Responsibilitiespage 3/3

• plus cost estimates for the controlling
  electronics of the permanent magnet field tuners
  come from Mike Browne, CD.
• The magnet cost estimating group are NOT
  responsible for costing
• Subsystem mechanical assembly drawings
• e.g. magnet plus support drawings
• Any system drawings containing magnets
• installation costs exact treaty point being
  clarified with the Installation TSET
• Movers and supports
• BPMs
• Vacuum chambers
• System engineering beyond typical magnet, PS or
  cables subsystem engineering design tasks that
  make sure our components fit/work with other
  beamline components

7
Define Magnet Technical System

• 1. Understand how magnets and power supplies and
  cables fit into the WBS.
• 2. Respond to functional requirements. Create
  descriptions and technical specs.
• 3. Define boundaries of a generic magnet and a
  generic PS. Make and document treaty points
  with other TSET teams.
• 4. Count how many magnets in each beamline, count
  styles, count PS, cables.
• 5. Develop detailed lists of parts that must be
  included in a magnet or PS cost estimate.

8
NLC-wide Magnet Design Philosophy

• NLC magnets are approx 50 water or air cooled
  conventional electromagnets can be powered by
  off-the-shelf PS, and approx 50permanent magnets
• All engineering, drafting, magnetic
  measurements done by NLC employees
• Our major technical challenges
• Produce 5600 magnets over 7 years
• Make them extremely reliable
• Minimize cost while maintaining performance
• To meet these challenges we will
• Identify failure modes using FMEA,design in
  reliability
• Have uniform standards for common materials such
  as ferrite, steel, conductor, cooling hoses
• Have standard designs for common parts such as
  terminal blocks, coil retainers, manifolds
• Have a restricted list of approved off-the-shelf
  parts water fittings, insulation, epoxies

9
NLC-wide Magnet Manufacturing Philosophy

• Most electromagnets will be fabricated in other
  countries, some by commercial companies or SLAC
  or other HEP labs.
• Most permanent magnets will be assembled at FNAL
  by non-shop technicians
• Need to make an early start on specifying steel
  and identifying steel vendors, ditto permanent
  magnetic materials
• Choose materials to mitigate E.SH concerns
• Will have tightly coordinated controlled
• material procurement tracking
• fabrication processes
• drawing release revision
• Will have on-site high capacity comprehensive
  incoming magnet inspection and measurement
  facility. ALL magnets will be QAd and
  magnetically measured.

10
NLC-wide DC PS Manufacturing Philosophy

• Philosophy driven by large number of systems and
  stringent reliability requirements.
• Performance goals can be met with SWITCH MODE
  power supplies.
• Most PS to be built by commercial companies, are
  within the 1kW-15kW range now available
• Units will be modified production versions
• connectorized to allow replacing failed units
  quickly in system (quick-swap)
• PS controller provides mechanism to adapt
  commercial supplies to NLC applications.
• PS controller will be quite different from
  existing SLAC (PEPII/SLC/FFTB) designs.

11
Develop cost estimating processes- in progress

• Have developed various cost estimating processes
  for magnets, PS and cables.
• Characteristics of these processes, they
• pay attention to our design and manufacturing
  philosophies- permanent methods in progress
• are applied consistently across all 3 beam areas.
• use the tremendous hands-on knowledge of our team
  members to produce realistic estimates.
• currently provide estimates aiming to be 50
  confident, learning how to assign in the WBS
  structure easily and without error.
• are flexible and repeatable so as to quickly
  accomodate changes as we refine costs between now
  and CDR.
• use NLC labor types and hourly rates.

12
Some general cost-estimating assumptions

• We will have designed, built, equipped and
  staffed these on-site facilities for testing,
  final assembly and measurement of all magnets and
  power supplies
• an incoming magnet testing and magnetic
  measurement facility
• a magnet fiducialization lab with a CMM
• a PS pre-assembly and test lab
• an electronics rack assembly facility
• Repetitive tasks are identified, number of labor
  hours per task agreed upon and used by all
• Commonly used parts/materials are identified,
  cost of such items agreed upon and used by all
• EDI estimates include analysis, design
  engineering, some subsystem engineering,
  manufacturing engineering, as well as drafting.

13
CD0.4 Magnet specific costing guidelines

• Electromagnet Specific Assumptions
• Solid steel cores, low carbon steel prices
• Hollow core copper conductor Potted coils
• SLAC style insulation and epoxy
• Agreed upon list of components
• Fabricated offshore use offshore labor codes
• PS Specific Assumptions
• Every group of same style PS has one quick
  spare PS added to cost no 2-for-1 redundancy
• Use 1999 catalog price of existing standard
  commercial PS reduced by 10
• Same controller cost for every PS system
• DC Power Cable Specific Assumptions
• Cable lengths assume cut cover tunnels and
  TEEs shorter cables than in Lehman.
• Cable size are gt NEC reduce heat put into
  beamline housings.

14
Typical Electromagnet Costing Process

• Is a bottom-up process, assume small quantity
• Choose a specd magnet with a layout drawing
• Make a parts list
• Make a drawing list how many, what size.
• Make a BH task list
• Determine materials costs using common prices
• Estimate hours for fabing each part
• For standard parts use NLC list of costs
• use outside vendor estimate or ask SLAC shops
  remember they give high estimates
• ask Magnet Engineer
• use appropriate labor type, use offshore labor
  rates
• Calculate fabrication costs, apply learning curve
  for higher quantity if not inherently in the
  hours estimate
• chose learning curve percentage appropriate to
  the mix of hand assembly and machining
• Estimate drafting hours using NLC list of hours
  per drawing by complexity and size.
• Compare result with historical data on similar
  magnets, both from SLAC or other labs.

15
Cable and Tray Costing Process

• For Lehman did detailed, bottom-up beamline by
  beamline cost estimates
• Material cost based on RS Means Electrical Cost
  Data or vendor quotes
• Installation labor hours based on above RS Means
  - hours/feet length of size X cable Ysize tray.
• For CD0.4 has removed many cables completely ,
  estimated rest in detail
• Engineering estimated at 20 of MS, then divided
  between phases
• Cable sizes-larger than National Electrical Code
• Injector NEC 2
• DR NEC 1
• Main Linac NEC 5
• Beam Delivery NEC 3

16
MOST CHALLENGING ASPECT OF ILC CONVENTIONAL
MAGNETS MAKING THEM RELIABLE ENOUGH

• Consider the availability requirements of the ILC
  as set out in the BCD
• A good idea for each engineer to read Chapter 10
  of the BCD
• Got to this URL and download the Operations and
  Reliability chapter
• http//www.linearcollider.org/wiki/doku.php?idbcd
  bcd_home
• Describes a simulation of the whole ILC that has
  been developed and the models output tells you
  how long the ILC will be down it its components
  have certain mean time between failures (MTBF)
  and certain times to repair (Mean Time to Repair
  MTTR).
• OVERALL ILC UPTIME GOAL IS 85 during the
  official runs of 9 months per year

17
Availability DEFINITIONS
Availability Average ratio of the time that the
system or component is usable to the total amount
of time that is needed.
MTBF (Mean Time Between Failure) MTBF is a basic
measure of reliability for repairable items. It
can be described as the number of hours that pass
before a component, assembly, or system fails.
Failure rate MTBF-1 l
MTTR (Mean Time To Repair) MTTR is the average
time required to perform corrective repair on the
removable items in a product or system.
Availability of N magnets (Availability of one
magnet) N
Expected Downtime in hours (1-Availability) x
Operation hour/year
18
How total allowed downtime of 17
is distributed among tech systems (assuming 17
15)
Magnets allowed 5 of 17gt0.8
In certain scenario each magnets MTBF has to be
20 million hours
All 12600 magnets allowed to be down 0.8 of ILC
running time. Same as need to be up 99.2 of
time!
19
Measuring Electromagnet Availability at Stanford
Linear Accelerator Center

• Obtain magnet failure history (CATER system) for
  5 year period (1997-2001)
• Categorize data into solid wire and water-cooled
  electromagnet types.
• Calculate average beam downtime for different
  types of magnet from failure data.
• Obtain SLAC beamlines runtime schedule for this 5
  year period.
• Count number of magnets in each SLAC beamline
  during specific runtime periods.
• Identify magnet failures that shut down the beam
  from CATER system report for each runtime period.
  CATER is an accelerator failure tracking
  database
• Calculate magnet operating hours by multiplying
  number of magnets by run hours for each period.
• Calculate MTBF, MTTR and availability of one
  magnet for each period.
• Calculate average availability for one magnet
  using all or some subset of the SLAC beamlines
  data.
• This process repeated for switching power supply
  failures over same period.

20
Summary of SLAC magnet PS failure data
solid magnets with solid wire coils water
magnets with water cooled coils small PS
lt12A, lt50V large PS gt12A,gt50V. Time to
repair is the total hours the beam was down for
the stated failures, so MTTR Time to repair/No.
of failures.
21
How the MTBF value can vary depending
on time period studied, choice of magnets

• From SLAC 1997-2001 data water cooled
  electromagnets average MTBF was 1,150,000 hours,
  i.e. about 1/20th of what the ILC needs!
• But look at a different period and eliminate the
  worst offenders
• This 2002 dataset does not include any magnets
  in the 2 SLC damping rings, which have
  notoriously failure-prone magnetsby removing
  them from the dataset one can make the average
  MTBF vastly longer 12,000,000 hours,. We
  understand why the DR magnets fail more
  frequently and would avoid making the same design
  mistakes in ILC magnets.
• IN ANY EVENT, FOR THE ILC, WE CANNOT DESIGN AND
  FABRICATE MAGNETS LIKE WE HAVE BEEN DOING FOR THE
  PAST 40 YEARS AT SLAC.
• We have to carry out a detailed Failure Modes and
  Effects Analysis (FMEA) to learn how to revise
  our magnet designs and fabrication techniques to
  make more reliable magnets

22
When we really get to design the ILC
magnets we will have to do FMEAs on basic magnet
styles

• Failure Mode and Effects Analysis (FMEA) process
  considers each mode of failure of every component
  of the system, identifies their causes and
  ascertains the effects of each failure mode on
  system operation (ALL ILC components should have
  FMEA done on them).
• As we cost estimate the ILC magnets we will have
  to account for the cost of doing FMEAs and paying
  for some higher quality materials and more
  expensive processes.
• The causes of the most severe and likely to occur
  failures of a standard SLAC water- cooled
  electromagnet were identified as (a) water leaks
  and corrosion (b) various assembly errors.
• Design changes were made in the conductor,
  terminals, core numbers of items.

23
FABRICATION FEATURES of E-M QUAD DRIVEN BY
RELIABILITY OR COST NEEDS

• Solid C1006 steel core, 4 quadrants ground on the
  outside 4 pieces bolted together and the 4
  poletips coil pockets EDMd in the same
  operation 0.005mm reproducibility on the
  poletips better coil pocket stability and
  tolerances. Machining cost less.
• Hollow seamless ROUND copper tubing (per ASTM
  B75) used for conductor, several advantages
  compared to square conductor easier to wind not
  prone to twisting does not keystone much
  smoother internal surface with many less crevices
  and defects where corrosion can start allows
  direct attachment to compression fittings. Coil
  winding cost less.
• New style power terminals commercial motor
  disconnects, modified to be brazed onto the
  conductor coil leads cheaper and more reliable
  than custom made multifunction terminals
• Potted coils instead of wet lay-up better
  dimensional stability and water resistance. The
  reduction in number of coils shorting out from
  nearby water leaks worth the higher cost.
• Prototype quad was made and has been run for
  many hours over 4 year period in a measurement
  set-up without any failures