Design and Optimisation of the ISIS TS1 Upgrade Target Dan Wilcox High Power Targets Group, RAL 6 th High Power Targetry Workshop, 12/04/2016

Acknowledgements Goran Skoro and Stuart Ansell – Neutronics Leslie Jones – Target Engineering Dan Coates – TRAM Engineering David Jenkins – ISIS Target Design Group Leader And many more…

Overview Background Neutronic requirements Design and optimisation Comparison: current target vs upgrade proposal Future work Conclusions

Background – ISIS Facility Target Station 1 Opened in 1984 Receives 4/5 beam pulses (quasi-40Hz) 200uA beam current (160kW) Proven reliability over many years Target Station 2 Opened in 2007 Receives 1/5 beam pulses (10Hz) 40uA beam current (32kW) High neutronic efficiency Synchrotron 800MeV proton energy Up to 240uA beam current (192kW) Pulsed at 50Hz

Background - TS1 Upgrade Project Will develop and upgrade key elements of TS1 including moderators, reflectors and target infrastructure Aims to achieve a factor of 2 average gain in useful neutron flux for a modest cost (≈1 or 2 new instruments) This will be achieved by using modern software and analysis techniques to improve neutronic efficiency without increasing the beam power The upgrade must not compromise the current level of reliability An essential part of this will be the development of a new and more efficient spallation target Approved by ISIS Facility Board in December 2015, implementation currently scheduled for

Target fluid, thermal and structural design Dan Wilcox Water moderator flow simulations Colin Souza Moderator can optimisation Yanling Ma Support beam structural simulation Wolfgang Fischer (FEA Solutions Ltd.) Reflector thermal design Colin Souza FEA for TS1 Upgrade

General Neutronic Requirements Target cross section similar to beam cross section –Material outside the beam will absorb more neutrons than it produces –Minimise volume of non-target material Water anywhere (apart from moderators) is bad for pulse width; particularly important as ISIS is a short pulse facility Total output = (useful neutron production) x (facility uptime)

Neutronics for TS1 Upgrade For 200µA we need a plate target –Thermal stress limits solid rod targets to <60µA –More highly segmented targets e.g. ‘cannelloni’ are not neutronically worthwhile until ≈1MW (too much water) Cannot modify existing instruments or beamline infrastructure –Important to create vertical space for a water pre-moderator –Existing beam transfer line is not suitable for an elliptical beam Use existing materials which we have experience of, unless there is a compelling neutronic reason not to

Baseline Target Proposal Cylindrical plates of Tungsten clad in Tantalum Parallel flow heavy water cooling

Optimisation Strategy Detailed FEA simulation of current TS1 target and TS1 upgrade concept, using the same methods and assumptions –We know the current TS1 target achieves an acceptable lifetime (4-5 years, limited only by thermocouple failure) –Many difficult to model conditions such as pre-stress and radiation damage will be the same for both targets Get as close as possible to the neutronic ideal without exceeding stress levels in current TS1 target

Plate Thickness Optimisation Reducing the number of target plates reduces the volume of water and tantalum, and the cost of manufacturing –Current TS1 has high stress in plates 1 and 2, but not in 3-12 –Use FEA to optimise the stress distribution across all plates Result: 10 plates required at 200µA –Front 4 plates are limited by stress in tantalum (< current TS1 level) –Back 6 plates are limited by surface temperature (< 100°C to protect adjacent components)

Cooling Manifold Design Cooling manifold requirements: –Even flow distribution between plates –Minimise pressure drop –Minimise water volume –Must not overheat if one channel is partially blocked Straight manifold without a divider was found to give the best flow distribution – and acceptable performance with a blocked channel Full TaperStraight With Divider Straight No DividerHalf Taper

Target Pressure Vessel Contain pressurised cooling water –Meet requirements of PD5500 (British pressure vessel design code) –Minimise volume of material used Two materials under consideration: –Tantalum: better for neutronics, can weld directly to plate cladding –Stainless Steel: less expensive, easier to manufacture, better weld reliability

Beam Window Optimisation Minimise stress in target beam entry window –A thick window will have higher thermal stress –A thin window will have higher stress due to pressure ANSYS DesignXplorer used to automatically find the best trade-off –Two inputs: minimum thickness and radius of curvature –One output: minimise Von Mises stress –Steel and tantalum cases considered

Current Target Proposal 10 cylindrical plates of Tungsten clad in Tantalum Parallel flow heavy water cooling Steel and Tantalum options for outer vessel Manufacturing features added by Leslie Jones – see talk on Wednesday

ISIS Target Comparison Image credit: Leslie Jones, ISIS Target Design Group Mass of material (kg)Current TS1TS1 UpgradeReduction Stainless Steel % Tantalum % Tungsten % Total %

ISIS Target Comparison TS1TS2TS1 UpgradeDesign Limit Peak Temperature (deg C) N/A Peak Heat Flux (MW/m^2) Tungsten Stress [Beam Only] (MPa) Tantalum Stress [Beam Only] (MPa) Tungsten Stress [HIP+Beam] (MPa) Tantalum Stress [HIP+Beam] (MPa)200* 75 Notes: Tantalum will yield at 200MPa TS1 and TS1 upgrade were modelled with MCNPX+CFX+ANSYS, 200µA TS2 was modelled with FLUKA+CFX+ANSYS, 40µA HIP pre-stress modelled assuming 500°C ‘lock-in’ temperature

Future Work Make a final decision on Tantalum vs Steel outer vessel Build a physical model of target cooling flow to validate fluid and thermal models Simulate accident cases (off-normal beam, decay heat plus cooling loss, etc.) in more detail Prototyping and manufacturing tests – see talk by Leslie Jones

Conclusions Worked closely with neutronics and manufacturing groups to understand their often conflicting requirements Developed a target concept to provide the best possible trade-off Used FEA optimisation to get as close as possible to the neutronic ideal without compromising reliability –60% less material used overall – reducing manufacturing cost, neutron absorption and volume of radioactive scrap –Stresses lower than current TS1, which is proven to be reliable