Machine Design and Automation

ugliestmysticAI and Robotics

Nov 14, 2013 (3 years and 7 months ago)

53 views


PMA

Production Engineering,

Machine Design and Automation

K.U.Leuven

Department of Mechanical Engineering

Celestijnenlaan 300 B,

B
-
3001 Leuven, Belgium

Tel: +32 16 32 24 80 Fax: +32 16 32 29 87

www.mech.kuleuven.ac.be/pma

Production Engineering, Machine Design and Automation

2

Overview of presentation


General information


Facts and figures


Research areas


Spin offs and Technology Centre


Scores related research


Mechatronics


Noise and Vibration


Robotics and Intelligent Machines


Research Infrastructure

Production Engineering, Machine Design and Automation

3

General information : Facts and Figures


Staff : 130


Research projects funded by


university: 30%


governments (European + federal + regional): 36%


industry: 30 %


foreign fellowships: 4%


Students :


Engineering degree : 80 per year


Ph.D. :


8 per year

Production Engineering, Machine Design and Automation

4

General information : Facts and Figures


Academic staff (professors)


F. Al
-
Bender


H. Bruyninkcx


J. De Schutter


W. Desmet


J. Duflou


W. Heylen


J.P. Kruth


B. Lauwers


D. Reynaerts


P. Sas


J. Swevers


H. Van Brussel


D. Vandepitte

Production Engineering, Machine Design and Automation

5

Spin offs



Name




Establ.


Area

LMS International



1979




Dynamic analysis

Data Analysis



1988




Monitoring for

Products






maintenance

Krypton Electronic



1989




Measuring & quality

Engineering






control systems

Materialise




1990




Rapid prototyping








products and software

Metris





1995




Reverse engineering








& quality control

Optidrive




1997




Optimization of drive








systems

Production Engineering, Machine Design and Automation

6



FMTC: Flanders’ Mechatronics
Technology Centre


New research centre, operating since October 2003


Initiative of Agoria, the Belgian multi
-
sector federation for the
technology industry, and 14 leading mechatronic companies in
Flanders


Supported by the Flemisch government via the IWT


Mission: to establish the bridge between the academic and
industrial know
-
how in mechatronics


The centre executes industry driven mid
-
term and long
-
term
research projects in:


Machine diagnosis


Modular Machines


Super Performance Machines


Close co
-
operation with PMA in project research and PhD
research


Contact: Prof. Hendrik Van Brussel

Production Engineering, Machine Design and Automation

7

General Information: Research Areas

production

processes

production

machines

machines &
products

intelligent
manufacturing
systems

mechatronics
intelligent
machines &
robotics

P

M

A

noise & vibration
engineering

Production Engineering, Machine Design and Automation

8

General Information: Research Areas


Production processes


Computer integrated manufacturing


Dimensional metrology and quality control


Design of light weight structures


Micro
-

and precision engineering


Automotive engineering


Space technology



Design of mechatronic systems


Noise and vibration engineering


Robots and intelligent machines


Production Engineering, Machine Design and Automation

9


Types of research:

Balance between:


Long
-
term fundamental research


Short
-
term or applied research



Linked to industrial needs:


Fundamental research in co
-
operation with industry


Solving specific problems


Experimental validation of developed
techniques



General Information: Research Areas

Production Engineering, Machine Design and Automation

10

Scores (most) related research

Topics


Design of mechatronic systems


Noise and vibration engineering


Robots and intelligent machines





No strict separation of
research topics


Co
-
operation between
different research groups


{

{

Production Engineering, Machine Design and Automation

11

Scores (most) related research
(cont’d)


Robots and intelligent machines: task planning,
active sensing, sensor based environment modeling
and task execution


Dynamic balancing of high speed machines


Modeling, identification and analysis of (non
-
linear)
dynamic systems


Vibro
-
acoustic modeling and prediction


Active control



Production Engineering, Machine Design and Automation

12




Non
-
Minimal State Kalman Filter (NMSKF) for nonlinear systems


exact Bayesian estimation: pdf state x, given measurements Z
k








and are updated with a KF algorithm


static systems with additive Gaussian measurement uncertainty


limited group of dynamic systems





















'
ˆ
)
(
'
ˆ
'
ˆ
)
(
2
1
exp
~
|
k
k
T
k
k
x
x
g
P
x
x
g
Z
x
p
Robots and intelligent machines

'
ˆ
k
x
'
ˆ
k
P
)
(
'
x
g
x

Production Engineering, Machine Design and Automation

13

Robots and intelligent machines


Application of the NSMKF for sensor based geometric
model building


example: force controlled manipulation


measurements: contact force (6D) and motion (6D)


task: build geometric model of the environment using
measurements


starting from scratch: large initial uncertainties on estimated
states


use primitive contacts (e.g. vertex/face contact)


nonlinear
equations


reduce number of parameters using statistical hypothesis testing

e.g.: two vertex/face contacts reduce to one edge/face contact


example: putting a cube in a corner











Production Engineering, Machine Design and Automation

14

1 vertex/face

2 vertex/face

Robots and intelligent machines

Production Engineering, Machine Design and Automation

15

Robots and intelligent machines


Task specification as a constrained
optimization problem


multiple motion tasks x(t) are defined; task
jacobian J




over
-

or underconstrained task specification


task weighting with e.g. system inertia


implementation using torque control or joint
velocity control


example: automatic generation of a ‘natural’
underconstrained human motion

q
J
x



Production Engineering, Machine Design and Automation

16

Robots and intelligent machines


Task planning for ‘active
sensing’:

a constrained optimization
problem



generate robot trajectory such
that sensor information
collected between start and
goal position yields the most
accurate estimates of the
parameters of the world model



in force controlled manipulation:
‘hybrid’ optimization


generate sequence of discrete
‘contact formations’ (e.g.
edge/vertex)


generate continuous motion
within a contact formation

Production Engineering, Machine Design and Automation

17

Dynamic balancing of reciprocating machinery


Input torque balancing to reduce drive
speed variations:


novel mechanism: cam based ‘centrifugal
pendulum (CBCP)


optimized, designed, and implemented


functions correctly, yields significant
enhancement of dynamic behavior


Counterweight balancing of linkages:


reduce shaking forces and moments
exerted on the supporting frame


reformulated as a convex optimisation
problem


computationally efficient, global optimum

Production Engineering, Machine Design and Automation

18

Dynamic balancing of reciprocating machinery

Production Engineering, Machine Design and Automation

19


Fundamental


Development of on
-

and off
-
line parameter estimation
techniques


Modeling and analysis of local non
-
linear system dynamics,
e.g. friction


Applied


Analysis of road noise transmission in vehicles


Non
-
destructive material identification using mixed
numerical
-
experimental identification techniques, e.g. for
laminates


Model based friction compensation


Experimental robot identification


Modeling, identification and analysis of dynamic
systems

Production Engineering, Machine Design and Automation

20


Physics based friction model


Stochastic model based on asperity interaction
scenario combined with phenomenological
mechanisms: creep deformation, adhesion


Able to describe all observed types of friction
behavior




Development of dynamic friction models and
friction compensation techniques

Production Engineering, Machine Design and Automation

21


Generalized Maxwell
-
slip model for friction
compensation


Development of dynamic friction models and
friction compensation techniques

Production Engineering, Machine Design and Automation

22


Detailed experimental analysis
of friction: tribometer


Friction compensation:
combination of


Model based feedforward
compensation


Disturbance observers



Development of dynamic friction models and
friction compensation techniques

Production Engineering, Machine Design and Automation

23


Fundamental: Mid
-
frequency modeling:


Wave based models


Fuzzy finite element models


Hybrid tools for aero
-
acoustic modeling



Applied


Validation of vibro
-
acoustic modeling
techniques


Models for sound propagation in subsonic
confined flows, e.g. automotive mufflers


Modeling of the drive train dynamics of a wind
turbine


Validation of the FE & BE methods in vibro
acoustic modeling and analysis


Vibro
-
acoustic modeling and prediction

Production Engineering, Machine Design and Automation

24


Fundamental


Development of robust controllers for non
-
linear systems
based on approximate model structures


Optimal decoupling for improved MIMO control design


Modeling and control of Linear Parameter Varying systems


Applied


Active control of exhaust noise of combustion engine


Design of lightweight inertial actuators with integrated
velocity sensor for active vibration control of a thin panel


Active noise and vibration control for machining systems


Active and semi
-
active suspension systems for passenger
cars


Anti
-
sway control for the load of a tower crane


Model
-
based and robust servo
-
control for high
-
performance
drive systems

Active control

Production Engineering, Machine Design and Automation

25


MIMO identification and control design: cumbersome


Decentralised control neglecting coupling: limited
performance


Combine decentralised control with I/O decoupling:


optimised static decoupling


dynamic transformation filter (inverse based control)


Validation: Time Waveform Replication (TWR):


Optimal decoupling for improved MIMO control
design

Production Engineering, Machine Design and Automation

26

Optimal decoupling for improved MIMO control
design

GOAL :

Replicate loads acting

on vehicle (component)

Application:

Time Waveform
Replication

Production Engineering, Machine Design and Automation

27

Optimal decoupling for improved MIMO control
design

Production Engineering, Machine Design and Automation

28

Optimal decoupling for improved MIMO control
design

Production Engineering, Machine Design and Automation

29

Optimal decoupling for improved MIMO control
design

Production Engineering, Machine Design and Automation

30


IWT
-
project with Tenneco/Monroe


Development of robust controller for
active suspension of quarter car


linear model


uncertainty modeling,


H
-
infinity design



Control of active and semi
-
active suspension
systems for passenger cars



Production Engineering, Machine Design and Automation

31

Control of active and semi
-
active suspension
systems for passenger cars



Production Engineering, Machine Design and Automation

32

Control of active and semi
-
active suspension
systems for passenger cars



Production Engineering, Machine Design and Automation

33

Control of active and semi
-
active suspension
systems for passenger cars



Some measured FRF’s and fitted models

Production Engineering, Machine Design and Automation

34

Control of active and semi
-
active suspension
systems for passenger cars



Estimated uncertainty

Production Engineering, Machine Design and Automation

35

Control of active and semi
-
active suspension
systems for passenger cars



Results: PDF of body acceleration and tire force, for: passive
suspension (blue), constant settings of active suspension (red), and
robustly controlled active suspension (green)

Production Engineering, Machine Design and Automation

36

Control of active and semi
-
active suspension
systems for passenger cars



Robust Control



No Control

Production Engineering, Machine Design and Automation

37


IWT project with ARCOMET NV


Anti
-
sway control of the load of a tower crane



Production Engineering, Machine Design and Automation

38


Results


Anti
-
sway control of the load of a tower crane



Production Engineering, Machine Design and Automation

39

Without control




With Control


Anti
-
sway control of the load of a tower crane



Production Engineering, Machine Design and Automation

40


Team Cube : high
-
performance 6
-
DOF shaker table


Multi
-
channel Measurement, data
-
acquisition and analysis
equipment


Multi
-
channel DSP
-
based control systems


Krypton K600 : 6
-
DOF position/orientation measurement
system


Robotics laboratory equipped with 5 industrial robots, 2
mobile learning robots and semi
-
autonomous wheel chairs


Several high
-
performance drive
-
systems based on linear
motor technology


Semi
-
anechoic measurement room


Other experimental test setups, e.g. tribometer, weaving
machine a blank, etc ...


Research Infrastructure

Production Engineering, Machine Design and Automation

41

Research Infrastructure: Team Cube

Production Engineering, Machine Design and Automation

42


Research Infrastructure: Team Cube

DOF
0-250 Hz
Stroke (mm)
Velocity
(m/s)
Acceleration
(g)
Longitudinal
50,8
0,96
6,8
Transversal
50,8
0,96
4,4
Vertical
101,6
0,96
5,3
Pitch

n/a
n/a
Roll
4,5°
n/a
n/a
Yaw

n/a
n/a


3 actuator pairs



82 kN/pair



(1+2)/2 = Z / (1
-
2)/2 = Pitch



(3+4)/2 = X / (3
-
4)/2 = Yaw



(5+6)/2 = Y / (5
-
6)/2 = Roll

Production Engineering, Machine Design and Automation

43

Research Infrastructure: Team Cube

Show

film

Production Engineering, Machine Design and Automation

44

Research Infrastructure: Robot lab

Production Engineering, Machine Design and Automation

45

Research Infrastructure: linear motor
based machines

Production Engineering, Machine Design and Automation

46

Conclusion


Research at PMA most related to Systems,
Control and Optimization:


Robots and intelligent machines


Noise and vibration engineering


Design of mechatronic systems


Characteristics:


Many applications of existing methodologies


Development of new approach in view of
applications


Validation by experimental work


Valorisation in industrial projects