CLIC cost estimate CLIC cost estimate

Hans-H. Braun, CLIC-GDE meeting, February 8, 2008

Cost model goals

Methodology

Cost distribution

Future improvements

Goals of cost studyGoals of cost study

Get reasonably precise cost estimate

Identify cost drivers and assure cost conscious design

Get cost model scalable in RF, G and ECM as input for optimization studies

CLIC cost model technical approach

1. For main tunnel equipment (w/o RF structures) steadily improving estimates, but limited accuracy because design is still evolving.

2. For main tunnel RF structures (accelerator and decelerator) estimate based oncost of machining facilities, manpower and material for given production period.

3. Drive beam generator based on CTF3 costs with appropriate scaling except forKlystron + Modulator which are derived from TESLA cost estimate plus some local expert input.

4. Civil engineering and technical infrastructure costs from CERN-TS 2007 studyfor CE same methods and people as for ILC study

5. Basic assumption for all cost estimates is that CLIC will be constructed at CERN.

6. Key input for injectors, damping rings, BDS, control system from NLC studywith some cross checks on recent European projects and gut feeling fudge factors

7. Electricity costs based on present CERN contract

8. Basic tool EXCEL, systematic documentation in EDMS just started

Main Linacs35.2%

Control System3.1%

Beam Delivery System4.6%

Drive Beam Generation13.3%

Civil Eng. & tech. Infr.33.0%

Main Beam Injectors2.4%

Drive Beam Distribution

3.4%

Damping Rings2.3%

Main Beam Transferline

2.8%

An error in BDS estimate of 50% changes total cost by less than 3%

An error in main linac estimate of 50% changes total cost by almost 20%

Priorities have been to get good cost estimate and reduce costs of

1. Main linac modules2. Civil engineering3. Drive beam generation

Standard Linac module cost distribution

Drive beam instrumentation

4%

Alignment and stabilisation

6%

Assembly in the tunnel4%

Drive beam quadrupole

14%

PETS25%

Accelerating structures

19%

Baseline supporting system

1%

RF components21%

Vacuum 6%

PETS

DB Quad

waveguide

acc. structure

supports

A remark on electricity cost

Present CERN electricity cost is less than one third of ILC assumption

For optimization Alexej uses sum of investment + 10 year electricity costDepending on price/kWh electricity cost has a very different weight

ILC electricity cost assumptions per MWh are roughly factor 4 higher

CLI

C 3

.0 T

eV

© F. Willeke

What’ s the yearly impact of 388 MW × 5000h/y grid power on your electricity bill ?

CLI

C 0

.5 T

eV

CLI

C 1

.0 T

eV

CLI

C 3

.0 T

eV

.

What’ s the yearly impact of 388 MW × 5000h/y grid power on your electricity bill ?

CLI

C 0

.5 T

eV

CLI

C 1

.0 T

eV

Site Installation

Access TU Shaft Shaft

Shaft

Shaft

Shaft

Site Installation

Cavern

CavernCavern

Cavern

Main Beam Turn-around Conv.

Main Beam Turn-around

TBM 1 InstallationTBM 2

InstallationMain Beam TU, TBM 1

Main Beam

Site Installation

UTRA, UTRC & DBD

Site Installation

Cavern

Site Installation

Shaft

Shaft

Main Beam Turn-around, Conv.

Main Beam

Main Beam

TBM 1 Installation

Shaft

Site Installation

UTRA, UTRC & DBD

UTRA, UTRC & DBD

Site Installation

Cavern

Site Installation

Shaft

ShaftAccess TU

Main Beam Turn-

Dewatering

TBM 3 Installation

Main Beam

TBM 2 Installation

Main Beam

TBM 4 Installation

UTRA, UTRC & DBD

Outlet

Inlet

Water Transf. TU, TBM 5

Planning assumed to be constraint by CE

This sets the time scale forother components, in particularfor those needing specific production facilities

90 95 100 105 110 115 1200

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

G (MV/m)

Cos

ts (

a.u.

)

Main LinacsCivil E. and tech. Inf.Drive Beam GenerationMiscalleaneousElectricityTotal/2

Cost distribution as function of main linac accelerating gradient for RF=12 GHz

Based on a set of parameters from Alexej, parameter meeting 14.11.06

10 15 20 25 300

0.2

0.4

0.6

0.8

1

1.2

1.4

Frequency (GHz)

Cos

ts (

a.u.

)Main LinacsCivil E. and tech. Inf.Drive Beam GenerationMiscalleaneousElectricityTotal/2

Cost distribution as function of main linac frequency for G=100 MV/m

Based on a set of parameters from Alexej, parameter meeting 14.11.06

Conclusions and Outlook

• Scaleable cost model has been established extrapolating from cost estimate for 2005 nominal parameters. This model has been extensively used in parameter optimization.

• In 2007 CERN TS department completed cost estimate for civil engineering and technical infrastructures. CE as least as good as ILC estimates but some estimates for technical infrastructure are still crude (in particular electrical distribution).

• Presently a number of CLIC design features is only performance driven and needsrevision for potential savings.

• Estimate of subsystems like injectors, DR’s and BDS is presently very crude (but impact on totals is small)

• For many components we have only sketchy designs, naturally this limits the precision of the overall estimate

• More systematic approach with complete PBS and planning in preparation

• Consistent approach for risk analysis, functionality to extract cost by component type and character of cost (i.e. manpower, raw material, production facility…) for parametric studiesand consistent method for inflation correction wanted.

• For cost of some cost drivers with unprecedented parameters (RF structures, Modulators)studies with industry and/or consultants required and foreseen.