Dead zone analysis of ECAL barrel modules under static and dynamic loads for ILD Thomas PIERRE-EMILE, Marc ANDUZE– LLR

Purpose Analyze the impact of static and dynamic loads on ECAL barrel modules assembly clearances. Installing the ILC and the associated detector(s) in Japan (Kitakami)  Being careful with both static and dynamic load especially earthquakes Systems must be designed to resist vertical loads (gravity) but also horizontal ground motion due to seismic loads (dynamic). To optimize hermiticity (i.e. reduce dead zone), we need to get the best estimate of ECAL modules relative motion to define clearances between ECAL modules (Barrel & Endcaps*) ECAL barrel ECAL Endcap2 ECAL Endcap1   Static and dynamic simulations of ILD are needed to: Define gaps  and  Avoid mechanical contacts over the ECAL barrel lifetime * ECAL endcap modules are not studied here 5th of October 2017 Thomas PIERRE-EMILE

  Analysis Procedure  Z Gap  analysis in STATIC CASE One static load is expected : Gravity Expected to mainly affect ECAL module clearance in phi direction due to vertical loads. Gap  analysis in DYNAMIC CASE One dynamic load is expected : Acceleration spectrum from earthquake Expected to mainly affect ECAL modules clearance in Z direction due to combination of horizontal ground motion  Initial gap between 2 ECAL modules in phi : 2,5 mm   Initial gap between 2 ECAL rings along Z : 1 mm Z 5th of October 2017 Thomas PIERRE-EMILE

Static case model (Gap )

Static case model (Gap ) ECAL module definition and simplifications : ECAL model used is an equivalent 3D solid model (simplified model to reduce calculation time) Simplifications are done according to 12o’clock ECAL barrel module under vertical gravity g Simplifications The corner region where contact between modules is more likely to occur Behavior along z axis is a bit different. Comparable results with about 70% of calculation time réduction When having almost the same deformation at representative points, we can have a significant benefit in term of time consumption (~70%) No analysis can be done regarding stresses Behavior along the z axis not precisely reproduced 5th of October 2017 Thomas PIERRE-EMILE

Static case model (Gap ) Model definition & simplification: Limited to HCAL+ECAL barrel (no impact of other elements) Both SDHCAL and AHCAL are used (Two designs : VIDEAU and TESLA) to compare the impact of the geometry Shell elements are essentially used Every sub elements are supposed to be perfectly fixed together (no relative motion allowed between different parts) Electronic layers are taken into account as point mass SDHCAL model : VIDEAU geometry Total barrel Mass = 750 t* AHCAL model : TESLA geometry Total barrel Mass = 705 t* * Total mass provided here include detection layers 5th of October 2017 Thomas PIERRE-EMILE

Static case results (Gap ) Preliminary Analysis Results: SDHCAL model : Total displacement 0,9 mm Smallest gap between ECAL modules in phi : 2,31mm* AHCAL model : Total displacement 6 mm Smallest gap between ECAL modules in phi : 0,95 mm* SDHCAL design seems to be stiffer and reduces ECAL modules relative motion : Best case to optimize the gap  * Initial clearance value is 2.5mm 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap )

Dynamic case model (Gap ) Model definition : Model based on ILD baseline geometry ; Limited to AHCAL barrel design (less rigid model) and central part of the yoke ; Shell elements are essentially used ; Every sub elements are supposed to be perfectly fixed together (no relative motion allowed between different parts) ; No electronic layer are taken into account yet ; Global damping ratio of 2% (iron based structure) ; Recombination with static loads is not yet applied (only dynamic effect is studied); Acceleration Spectrum : calculated from « Standard Reference Earthquake Parameters » Toshiaki TAUCHI, April 2017 @ Lyon (1) (2) (3) ILD model : Mass = 3100 t* 664043 nodes Earthquake peak : 2-6 Hz (maximum stresses) * Total mass provided here does not include detection layers 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Mesh check under static load: Mesh quality has been evaluated for each sub element to minimize calculation elapsed time (~2h for modal and ~10h for dynamic) 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis results: Eigenvalues* and mass participation factor Mode 1 @ 2,3Hz * Eigenvalues = states of excitation/vibration according to specific fixed frequencies also called modes in that presentation First mode at 2.3Hz involves 68% of the total mass in a back and forth motion. A global motion can be observed 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis results: Eigenvalues and mass participation factor Mode 2 @ 3,05Hz Second mode at 3.05Hz involves also 68% of the total mass this time in a lateral motion. A global motion can be observed 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Mode 7 @ 7.4Hz Preliminary Analysis results: Eigenvalues and mass participation factor Mode 18 @ 17.5Hz Mode 8 @ 8.4Hz A more complex mass repartition is foreseen for the vertical vibration. This seems to involve individual elements rather than a global motion. Modes 7, 8 and 18 account for 71% of the total mass involved 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Mode 3 @ 3.8Hz Preliminary Analysis results: Eigenvalues and mass participation factor Mode 13 @ 13.1Hz Mode 11 @ 10.8Hz Despite its complexity, back and forth rotational motion of the detector seems to be essentially driven by 4 modes. They account for 95 % of the total mass 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis results: Eigenvalues and mass participation factor Mode 9 @ 8.6Hz Rotational motion around beam axis can be divided into several small mass participating modes among which one can find mode 9 (35.4% of total mass). As for the vertical motion this tends to indicate the vertical structure stiffness 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis results: Eigenvalues and mass participation factor Mode 6 @ 7.0Hz Global detector twist is essentially driven by mode 6 which accounts for more than 88% of total mass 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis Results: Eigenvalues summary Mode 1 @ 2,3Hz Mode 2 @ 3,05Hz Mode 3 @ 3,8Hz Mode 6 @ 7Hz Due to the very heavy structure: A large amount of modes at low frequency 6 global modes are included into the range of earthquake peak 2-6 Hz 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis Results: Response spectrum contribution – beam axis (z) With the acceleration response spectrum applied along beam axis, the first mode of the structure dominates: back and forth motion of the full structure Maximum displacement*: 24,9 mm Smallest gap between ECAL rings along z: 0,98 mm Smallest gap between ECAL module along phi: 2,29mm Global detector motion is foreseen. No relative motion along beam axis between ECAL modules. * Displacement values provided only accounts for RS contribution. No recombination with static loads were done 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis Results: Response spectrum contribution - Lateral With the acceleration response spectrum applied laterally, the second mode of the structure dominates: lateral motion of the full structure Maximum displacement*: 17,3 mm Smallest gap between ECAL rings along z: 0,98 mm Smallest gap between ECAL module along phi: 1,89mm * Displacement values provided only accounts for RS contribution. No recombination with static loads were done - No significant relative motion in traverse direction between ECAL modules. 5th of October 2017 Thomas PIERRE-EMILE

Dynamic case model (Gap ) Preliminary Analysis Results: Response spectrum contribution - Up and down With the acceleration response spectrum applied along vertical axis, a more complex mode composition drives the behavior Maximum displacement*: 2,9 mm Smallest gap between ECAL rings along z: 0,98 mm Smallest gap between ECAL module along phi: 2,05 mm The displacement are significantly lower (less than 3 mm). A rather good stiffness is foreseen with regard to vertical loading No Z relative motion between ECAL modules * Displacement values provided only accounts for RS contribution. No recombination with static loads were done 5th of October 2017 Thomas PIERRE-EMILE

Conclusion

Conclusion Gravity and Earthquake are the major load cases that must be taken into account when designing instruments to be installed in Japan. This helps in evaluating stresses impact and dead zone aspects. Static and dynamic analysis are used separately to study the relative motion between ECAL barrel modules (clearances): For gap  : The behavior is dominated by static loads. Both geometries of HCAL should be correct, if the initial gap is 2,5 mm. For gap  : The variation of gap is dominated by seismic loads. No significant relative motion has been detected in response spectrum along the 3 main axis To get a better understanding of the general behavior, further analysis should be done: Combing static and dynamic analysis in order to have the behavior as close as possible to the reality. Introducing electronics layer and coil masses in the full model Adding SDHCAL dynamic analysis to complete comparison Evaluating ground motion repartition to use a more realistic ground motion 5th of October 2017 Thomas PIERRE-EMILE

Thank you for your attention Questions?