Multidirectional Hybrid Simulation Tests at Polytechnique Montreal: System, Challenges and Applications

Transcription

1 Multidirectional Hybrid Simulation Tests at Polytechnique Montreal: System, Challenges and Applications Martin Leclerc, Robert Tremblay, Najib Bouaanani, Youness Mechmachi, Ali Imanpour, Yasaman Balazadeh-Minouei Polytechnique Montreal, Canada Hybrid Simulation Technologies & Methods for Civil Engineering MTS / UC Berkeley March 20-21, 2018

2 Plan 1. MDHTS System 2. Challenges: 1. System stiffness 2. Friction in actuator s swivels 3. Measurements precision 3. Applications 4. Conclusions 2

3 Laboratory at Polytechnique Montreal 3

4 MDHTS Upper Platen kn Actuators Z Rz X Rx Ry Y kn Actuators 4

5 MDHTS Manufacturer: MTS Corporation Full 6 DOF control (X, Y, Z, q X, q Y, q Z ) Mixed mode of displacement/force control Multi-axis testing of substructures quasi-static (cyclic) Pseudo-dynamic testing Hybrid Testing 2 Two displacement feedback signals: Absolute referencing (from actuator displacement) Relative referencing (from transducer s readings) 2 Hybrid configurations: ScramNet (Matlab, Simulink & xpc Target) CSI (MTS) 5

6 MDHTS Total System Capacity Vertical Force, Fz Horizontal Force, Fx & Fy Vertical Displacement, Z Horizontal Displacement, X & Y Rotation in swivels (all) Maximum moment X-X Maximum Moment Y-Y Maximum Moment Z-Z 4 x ±1800 kn (±7200 kn) 2 x ±1000 kn (±2000 kn) ±300mm ±375mm ±7 deg (±0.122 rad) From vertical actuators +/- 2 x 1800 kn x 3 m = ±10800 kn-m From Horizontal actuators +/- 2 x 1000 kn x 4.73 m = ±9460 (Y-actuators at 4 m elevation) +/- 2 x 1000 kn x 9.23 m = ±18460 kn-m (Y-actuators at 8 m elevation) Combined ±20260 to ±29260 kn-m From vertical actuators +/- 2 x 1800 kn x 2 m = ±7200 kn-m From Horizontal actuators +/- 2 x 1000 kn x 4.73 m = ±9460 (X-actuators at 4 m elevation) +/- 2 x 1000 kn x 9.23 m = ±18460 kn-m (X-actuators at 8 m elevation) Combined ±16660 to ±25660 kn-m +/- (1000 kn x 2 m kn * 3 m) = ±5000 kn-m 6

7 Plan 1. MDHTS System 2. Challenges: 1. System stiffness 2. Friction in actuator s swivels 3. Measurements precision 3. Applications 4. Conclusions 7

8 Challenges: System stiffness 8

9 Challenges: System stiffness Adapted From Imanpour (2015) 9

10 Challenges: System stiffness Control on specimen s displacements only (relative displacement) 4 Vertical Transducers + 4 Diagonal Transducers Control Point Fz Mx F y At Control Point: Forces: From Actuators Load Cells Displacements: From Actuators LVDTs & 8 Local Transducers 10

11 Challenges: System stiffness Metrology-grade scanners used to locate encoders in 3D space 11

12 Plan 1. MDHTS System 2. Challenges: 1. System stiffness 2. Friction in actuator s swivels 3. Measurements precision 3. Applications 4. Conclusions 12

13 Challenges: Friction in actuator s swivels Preliminary Hybrid Simulation: Flexible specimen 3 DOF (X, Y & Rx) Unstable control (large oscillations) 13

14 X-dir Force Feedback, kn Challenges: Friction in actuator s swivels Measured Response Displacement Command Specimen Response Friction Response Expected Specimen Response Time, sec 14

15 Challenges: Friction in actuator s swivels Friction characterization: Vertical Actuator PLATEN X Z RY Ctrl pt. 2m Vertical Actuator Horizontal Actuator 4m Direction X Control mode Relative = cyclic Y Relative = 0 Z stepwise incremented Load Rx Relative = 0 Ry Relative = 0 Rz Relative = 0 15

16 X-Force, kn Challenges: Friction in actuator s swivels X Force, Combined Specimen Response (linear) and Friction K F = 140 kn/mm K F = 20 kn/mm 0-10 K s = 2.1 kn/mm K F = 10 kn/mm X-Displacement (Abs), mm 16

17 Single Swivel Friction Torque (kn*m) Challenges: Friction in actuator s swivels Friction in One Vertical Actuator Swivel vs Actuator Load 35 Contact Surface Actuator in Compression Ry-dir - Tension X-dir - Tension Average - Tension Ry-dir - Compression X-dir - Compression Average Compression extrapolated values 10 Actuator in Tension 5 y = x y = x Axial Load in One Vertical Actuator (kn) * When specimen is in tension, actuator is in compression and vice-versa 17

18 Challenges: Friction in actuator s swivels Mitigation Techniques: Friction Compensator Negative friction model (add negative stiffness!!!) OpenSEES model of MDHTS (Ali Imanpour) 18

19 Challenges: Friction in actuator s swivels Mitigation Techniques: Feefback Instabilities Polynomial smooting of feedback signals Quadratic function in MTS controller 19

20 Displacement,mm Force, kn Challenges: Friction in actuator s swivels Bridge Pier, H=2.75m Youness Mechmachi (2016) Hybrid Response Numerical simulation Time (s) Hybrid Response Numerical simulation Time (s) 20

21 Challenges: Friction in actuator s swivels Mitigation Techniques: Load Cells One approach was to add 64 load cells to measure friction at each end of actuator swivel and compensate in real time. 21

22 Challenges: Friction in actuator s swivels Mitigation Techniques: Real-time friction computation Displacement cmd 6 DOF Real-time PC Force fbk Friction Physical substructure Numerical substructure Friction model (Simulink) 22

23 Plan 1. MDHTS System 2. Challenges: 1. System stiffness 2. Friction in actuator s swivels 3. Measurements precision 3. Applications 4. Conclusions 23

24 Challenges: Measurements precision Stiff specimen vs. encoder s resolution Original resolution : mm Unstable control Upgraded resolution: mm Stable control 24

25 Plan 1. MDHTS System 2. Challenges: 1. System stiffness 2. Friction in actuator s swivels 3. Measurements precision 3. Applications 4. Conclusions 25

26 Applications: Hybrid Simulation Testing: Computational driver Finite element model (OpenSees) Physical testing system MTS FlexTest 200 controller Experimental equipment: MDHTS Middleware OpenFresco Interfaces between OpenFresco & Flextest controller MTS Computer Simulation Interface (CSI) Mathworks Simulink Platform (xpc Target) 26

27 Application: Concrete C-shape wall OpenSees/OpenFresco Experimental element 27

28 Application: Concrete C-shape wall Sound specimen: (Youness Mechmachi) Carbon-fiber repaired specimen: (Hamid Arabzadeh) (November 2017: 89hrs test) (March 2018: Tested yesterday!) 28

29 Displ (mm) Application: Concrete C-shape wall 15 Displ. in X direction (mm) at the top of the specimen - Nahanni Earthquake Numerical Prediction 10 Hybrid Testing Simulation Time (s) 29

30 Application: Multi-Tiered Braced Frames Column Biaxial Buckling a) b) Base Tier 1 Level Base Base 30

31 tension Axial Load (kn) compression Application: Multi-Tiered Braced Frames Pure Numerical Simulation Numerical Hybrid Simulation Hybrid Simulation Axial Displacement (mm) 31

32 Conclusions Testing facility fully operational for multi-directional hybrid simulation testing (6 DOF) of large-scale structural components Test programs successfully completed on steel columns and concrete structural members Local displacement control must be used to obtain realistic deformations Friction removal AND smoothing of force s feedbacks must to be combined 6 DOF Hybrid simulation testing on stiff specimens is highly sensitive to system accuracy (noise level) and precision (resolution) to achieve convergence 32

33 Acknowledgements UC Berkeley PEER for OpenSees and OpenFresco Industrial partners: MTS Systems Inc. (Special thanks to Dr. Shawn You) LCL Bridge Products Ltd Research Engineer Romain Siguier and Technical staff of the laboratory at Polytechnique Montreal Researchers from Polytechnique Montreal: Youness Mechmachi, Yasaman Balazadeh-Minouei, Prof. Robert Tremblay and Prof. Najib Bouaanani Researchers from McGill University : Ahmed Elkady and Prof. Dimitrios Lignos Researchers from Concordia University : Hamid Arabzadeh and Prof. Khaled Ghalal Funding from Natural Sciences and Engineering Research Council of Canada (NSERC), Quebec Funds for Nature and Technology (FRQNT), Canadian Foundation for Innovation (CFI) & Government of Quebec 33

34 Thank you 34