Plummer, A. (2015) Design and control of dynamic testing
systems: overview. In: Design and control of dynamic testing
systems, 2015-10-21.
Link to official URL (if available):
Opus: University of Bath Online Publication Store
http://opus.bath.ac.uk/
This version is made available in accordance with publisher policies.
Please cite only the published version using the reference above.
See http://opus.bath.ac.uk/ for usage policies.
Please scroll down to view the document.
UK Automatic Control Council
Overview
Andrew Plummer
(UKACC) Prof University
of Bath
The dynamic testing of structures, components and
materials in the laboratory to determine their mechanical
properties is an essential part of engineering R&D.
Overview
• History
• Examples – automotive, seismic …..
• Model-based design and control
• Control
• Technology
116 ton hydraulic testing machine, built 1866
(David Kirkaldy, Southwark, London)
Victorian engineers: open-loop static testing
Dynamic bump-test,
1911
Rolls-Royce, Derby
Ten years in one day
“Royce took two great
drums armed with cams,
mounted them on an axle,
with the top of the drums
level with the floor,
designed a motor to rotate
the drums, and all facilities
for executing the
destructive tests and
recording them.”
F-16 Durability Testing: 25,000 Hours, 2015
(Lockheed Martin, Fort Worth)
Instron 5000lbf closed loop testing
machine – 1946
• Other developments
– Servovalve developed by MIT/Moog (USA) 1950
– Dowty, were first in Europe (60s) with a servohydraulic
materials testing machine
– Fully digital controllers from 1980’s
CHASSIS DURABILITY AND
DYNAMICS TESTING
The 1960’s
www.doddsandassociates.co.uk
Wide range of vehicles…
Iterative Control – from the 1970’s
Target
signal
vector
wt
Initial drive
signal
vector, rt,1
Inverse
model
Gain
(<1)
Previous
command signal
vector, rt,i-1
+
Command
signal vector
rt,i
Closed-loop
plant
+
Gain (1)
Inverse
model
Error signal vector from previous iteration, et,i-1
Error signal vector for current iteration, et,i
+
-
Response
signal
vector
yt,i
F1 chassis dynamics testing
4 and 7/8 post rigs
Aerodynamic Model-in-the-Loop
Rear ride
height
Front ride
height
Car forward velocity
Aerodynamic
down force
model
Rear down force
Front down force
MiL / Hybrid Testing / Substructuring
Force controlled
actuators example
Disturbances
Interface forces
Actuator
dynamics
External forces
(virtual)
Physical
system
Numerical
model
Sensor
dynamics
Measurement
noise
Interface
displacements
External forces
(real)
SEISMIC TESTING
Shaking table control
E-defense, Japan
Modelling
Modal multi-axis
inverse model-based
closed-loop control
Command
Acceleration
Washout
filter
Command
Position
Low pass
filter
+
Forward
decouple
-
Mode
Inverse
Inverse
decouple
P 1
s
Inverse
kinematic
transform
Valve
matching
Valve and
actuator
P
Decoupled axes

1
s2


 s 2  2  s   2  s 2
w nw
w 

1
I
x 6 axes (X, Y, Z, RX, RY, RZ)
Correction for
valve pressure
drop variation
Motion
estimator
Forward
kinematic
transform
Position
Acceleration
The Role of Modelling
and Simulation
Modelling for closed loop control
Required
velocity Velocity
trajectory
generator
Command
velocity
+
Partial
inverse
model
Closed-loop
compensator
Actual
velocity
Actuator
-
+
Residual
dynamics
model
Predicted
velocity
Pedestrian impact testing
Inverse modelling for iterative control
Hydraulic catapult
for occupant
restraint testing
Sledge
3-Stage Servovalve
4th Valve Stage
Accumulator
Rail
Hydraulic Oil
Actuator
Hydrostatic Bearing
Longitudinal Accelerometer
Test rig control implementation
Test rig
Control System
Valve
drives
Standard digital controller
Specialised control functions
via Simulink Real-Time Code
Virtual systems,
e.g. aero model
Sensors
Controller (and rig) development
Off-line and real-time simulation
Test rig
Control System
Standard digital controller
SIMULINK MODEL
Specialised control functions
SIMULINK MODEL
Virtual systems,
e.g. aero model
Valve
drives
Hydraulic system
SIMULINK MODEL
Rig mechanics
SimMECHANICS MODEL
Sensors
Car mechanics
SimMECHANICS MODEL
Control
algorithms
Closed
loop
compensation
Command
shaping –
non-real time
PID
+First order lag
+Notch filter
+Pressure f/b
+Acceleration f/b
Iterative control
Amplitude and
phase control
Command
shaping
– real time
Command velocity
feedforward
Adaptive
control
Self-tuning and
adaptive PID
Multi-axis
closed-loop
control
(decoupling)
Valve crosscompensation
Three
variable
control
Repetitive
(profile)
control
Adaptive inverse
control
Co-ordinate
transformation
H control
Non-linear
model-based
control
‘Delay
compensation’
Minimal control synthesis
Specimen
motion
feedforward
True modal
control
Proportional + Integral (PI) controller
r
+
-
 Gi 
G p 1  
s 

u
Plant
(valve + cylinder/specimen)
y
Hydraulic
resonance
Position (mm)
2
1
0
-1
-2
0
0.2
0.4
0.6
0.8
1
0.6
0.8
1
0.6
0.8
1
Time (s)
Position (mm)
Valve bandwidths:
50Hz
100Hz
200Hz
2
1
0
-1
-2
0
0.2
0.4
Time (s)
2
Position (mm)
50Hz hydraulic
resonance
1
0
-1
-2
0
0.2
0.4
Time (s)
Resonance compensation
1.5
Position (mm)
1
0.5
0
-0.5
-1
-1.5
0
0.1
0.2
0.3
0.4
Time (s)
0.5
0.6
0.7
0.8
1. acceleration feedback
2. differential pressure or
load feedback
3. a first order lag
4. a notch filter
5. A cross-port bleed
Repetitive control
Target
Wi
1
Achieved
Yi
0.5
0
-0.5
-1
0
1
2
3
4
5
Cycle, i
6
7
8
9
e.g. Amplitude control
Command adjustment: Ri  Ri 1  (Wi  Yi )
10
Adaptive & self tuning techniques
Identification
excitation
rt
Closed-loop
plant
G( z 1 )
z n
+
LMS
estimator
_
G( z 1 )
et
Adaptive inverse control
yt
Motion compensated load control
a
b
y1
y2
+
Valve crosscompensation
_
a
b
PI
+
posn feedback
Position control
+
_
+
PI
load feedback
Load control
Technology
• Actuation:
– servohydraulic good for high force, and
bandwidth
– direct drive (linear) electric motors
– low friction
• Sensing:
– Analogue (LVDTs)
– Digital (ultrasonic, absolute encoders)
– Load cells – inertial compensation
• Mechanical hardware
– Frame/joint stiffness
– Joint friction
– Structural resonance
Summary
• Creating high quality rigs for accurate dynamic
testing in the laboratory, replicating real-world loads
and motions, requires:
–
–
–
–
Specialist design knowledge
High quality components
System understanding
Subtle control algorithms and careful control
implementation (e.g. signal conditioning)
• Improvements aided by model-based design and
control.
• Hybrid testing is a future direction