M-sequenze based ultra-wideband radar and its application to crack detection in salt mines [Elektronische Ressource] / Ralf Herrmann. Gutachter: Motoyuki Sato ; Reinhard Knöchel. Betreuer: Reiner Thomä
241 pages
English
Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

M-sequenze based ultra-wideband radar and its application to crack detection in salt mines [Elektronische Ressource] / Ralf Herrmann. Gutachter: Motoyuki Sato ; Reinhard Knöchel. Betreuer: Reiner Thomä

-

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres
241 pages
English

Description

IlmenauUniversityofTechnologyDepartment of Electrical Engineering and Information TechnologyM-sequencebasedultra-widebandradaranditsapplicationtocrackdetectioninsaltminesDipl.-Ing. Ralf HerrmannDissertation zur Erlangung des akademischen GradesDoktor-Ingenieur (Dr.-Ing.)vorgelegt der Fakultät für Eletrotechnik und Informationstechnik derTechnischen Universität Ilmenau von Dipl.-Ing. Ralf Herrmann, geb.am 19.09.1978 in SchmalkaldenGutachter:1. Prof. Dr.-Ing. habil. Reiner S. Thomä2. Prof. Dr.-Ing. Motoyuki Sato3. Prof. Dr.-Ing. Reinhard KnöchelVorgelegtam: 07.07.2011Verteidigtam: 21.10.2011URN: urn:nbn:de:gbv:ilm1-2011000344IIPage intentionally left blank.IIIAcknowledgement“Zwei Dinge sollten Kinder von ihren Eltern bekommen:1Wurzeln und Flügel . . . ” (J. W. v. Goethe, 1749 - 1832)thEven though times are very different now than they used to be in the 18 century, somewisdom is even more true today. Without contribution of so many people it is unthinkableto follow an academic path and finish a dissertation.I want to express my deep gratitude to my parents, grandparents, brother, and friends foreverlasting support and encouragement on my way.Besides personal support, good education is invaluable to unfold one’s potential.

Sujets

Informations

Publié par
Publié le 01 janvier 2011
Nombre de lectures 13
Langue English
Poids de l'ouvrage 34 Mo

Exrait

IlmenauUniversityofTechnology
Department of Electrical Engineering and Information Technology
M-sequencebasedultra-widebandradarand
itsapplicationtocrackdetection
insaltmines
Dipl.-Ing. Ralf Herrmann
Dissertation zur Erlangung des akademischen Grades
Doktor-Ingenieur (Dr.-Ing.)
vorgelegt der Fakultät für Eletrotechnik und Informationstechnik der
Technischen Universität Ilmenau von Dipl.-Ing. Ralf Herrmann, geb.
am 19.09.1978 in Schmalkalden
Gutachter:
1. Prof. Dr.-Ing. habil. Reiner S. Thomä
2. Prof. Dr.-Ing. Motoyuki Sato
3. Prof. Dr.-Ing. Reinhard Knöchel
Vorgelegtam: 07.07.2011
Verteidigtam: 21.10.2011
URN: urn:nbn:de:gbv:ilm1-2011000344II
Page intentionally left blank.III
Acknowledgement
“Zwei Dinge sollten Kinder von ihren Eltern bekommen:
1Wurzeln und Flügel . . . ” (J. W. v. Goethe, 1749 - 1832)
thEven though times are very different now than they used to be in the 18 century, some
wisdom is even more true today. Without contribution of so many people it is unthinkable
to follow an academic path and finish a dissertation.
I want to express my deep gratitude to my parents, grandparents, brother, and friends for
everlasting support and encouragement on my way.
Besides personal support, good education is invaluable to unfold one’s potential. Starting
from my teachers at school, the lecturers at the Ilmenau University of Technology during
my years of study to my colleagues at the Electronic Measurement Research Lab I like to
thank everybody for providing knowledge, scientific stimulation, and motivation.
I want to especially thank my supervisors Prof. Dr.-Ing. Habil. Reiner Thomä and Dr.-
Ing. Jürgen Sachs for giving me the chance to work on an interesting research project,
continues scientific guidance, and numerous technical discussions and support. Fur-
thermore, I want to thank Prof. Dr. Motoyuki Sato and his colleagues at the Centre for
Northeast Asian Studies from the Tohoku University in Sendai (Japan) for accepting me
as a guest student and introducing many aspects of radar data processing.
Finally, the work presented in this dissertation was supported by the German Federal
Ministry of Education and Research (BMBF) under grant number 02C1194 from the years
2004 to 2008. I also acknowledge the help of our project partners at BoRaTec GmbH
Weimar and the colleagues from Fraunhofer IZFP Dresden in geological matters and
during measurement campaigns.
Dipl.-Ing. R.Herrmann Ilmenau, July 6, 2011
1Translation — Two things children should receive from their parents: roots and wings ...
R. HerrmannIV
Page intentionally left blank.V
Contents
Acknowledgement III
List of Figures VIII
List of Tables XI
List of Abbreviations and Symbols XII
Abstract XVII
Abstrakt XIX
1 Introduction to non-destructive testing 1
1.1 Non-destructive testing — an emerging research field with many applications . . 2
1.2 A challenging task — inspection of the disaggregation zone in salt mines . . . . 4
1.3 Properties of the disaggregation zone in salt rock and NDT sensors . . . . . . . 6
1.4 Current radar sensors in salt mines and GPR . . . . . . . . . . . . . . . . . . . 8
2 Theoretical analysis and design goals for a new sensor system 11
2.1 Propagation conditions for electromagnetic waves within salt rock . . . . . . . . 11
2.2 Bandwidth and frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.1 Scattering of thin gaps or layers . . . . . . . . . . . . . . . . . . . . . . 15
2.2.2 of small volume defects . . . . . . . . . . . . . . . . . . . . 19
2.3 Mode of operation — electromagnetic wave interaction with salt rock . . . . . . 21
2.3.1 Electro-magnetic surface waves . . . . . . . . . . . . . . . . . . . . . . 22
2.3.2 Radar sensor using antennas . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3.3 Antenna polarisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.4 Dynamic range requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4.1 Dynamic range for UWB time domain signals . . . . . . . . . . . . . . . 27
2.4.2 Setup for a typical measurement situation . . . . . . . . . . . . . . . . . 29
2.4.3 Simplified model for dynamic range simulation . . . . . . . . . . . . . . 30
2.4.4 Results of dynamic range simulation for thin gaps . . . . . . . . . . . . 32
2.5 Summary of design goals for the new UWB sensor . . . . . . . . . . . . . . . . 37
3 UWB system design for detection of sub-mm disaggregation in salt 39
3.1 Overview of available UWB electromagnetic sensor technologies . . . . . . . . . 39
3.1.1 CW/FMCW principle . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.1.2 Vector network analysers . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.1.3 Impulse radar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
R. HerrmannVI Contents
3.1.4 Noise and pseudo-noise principles . . . . . . . . . . . . . . . . . . . . . 46
3.2 The starting point — review of the M-Sequence UWB sensor principle . . . . . 49
3.2.1 RF chipset of basic M-Sequence devices . . . . . . . . . . . . . . . . . . 50
3.2.2 Digital backend of basic devices . . . . . . . . . . . . . . . 54
3.3 New 12 GHz bandwidth M-Sequence UWB sensor . . . . . . . . . . . . . . . . 55
3.3.1 Extended M-Sequence stimulus . . . . . . . . . . . . . . . . . . . . . . 56
3.3.2 Software-defined dense equivalent time sampling with four receivers . . . 60
3.3.3 Frontend for measuring full two-port S-parameters and calibration . . . . 63
3.4 Summary of system design for a 12 GHz bandwidth M-Sequence sensor . . . . . 65
4 Uniform dense equivalent time sampling 67
4.1 General aspects of sampling for UWB sensors . . . . . . . . . . . . . . . . . . . 69
4.2 Analysis of accuracy requirements for uniform dense sampling . . . . . . . . . . 71
4.2.1 Periodic non-uniform sampling with sine signals . . . . . . . . . . . . . . 72
4.2.2 Periodicrm sa with an UWB stimulus . . . . . . . . . . 76
4.3 Calibration of phase states for uniform dense sampling . . . . . . . . . . . . . . 81
4.3.1 Phase shifter control hardware and timing. . . . . . . . . . . . . . . . . 81
4.3.2 Phase calibration . . . . . . . . . . . . . . . . . . . . . . 84
4.4 Summary of phase shifter control calibration for uniform dense equivalent time
sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
5 Device clutter removal by full 2-port calibration 95
5.1 Recapitulation of basic coaxial calibration methods . . . . . . . . . . . . . . . . 96
5.1.1 Example: 3-term calibration . . . . . . . . . . . . . . . . . . . . . . . . 98
5.2 Full two-port calibration of M-Sequence UWB-sensors with 8-term method . . . 101
5.2.1 Choice of calibration method . . . . . . . . . . . . . . . . . . . . . . . 101
5.2.2 Formulation and implementation of 8-term correction . . . . . . . . . . 105
5.2.3 Performance figures of 8-term correction . . . . . . . . . . . . . . . . . 110
5.3 Accuracy analysis of calibrated MLBS system . . . . . . . . . . . . . . . . . . . 116
5.3.1 Calibration repeatability and long-term effects . . . . . . . . . . . . . . 116
5.3.2 Comparison of calibration performance . . . . . . . . . . . . . . . . . . 123
5.4 Summary of coaxial calibration for the new M-Sequence system . . . . . . . . . 129
6 Selected salt mine measurement results 133
6.1 Measurement setup in salt mines . . . . . . . . . . . . . . . . . . . . . . . . . 133
6.1.1 Antenna type and arrangement . . . . . . . . . . . . . . . . . . . . . . 135
6.2 Data processing for detection of disaggregation zone . . . . . . . . . . . . . . . 139
6.2.1 Removal of antenna crosstalk . . . . . . . . . . . . . . . . . . . . . . . 140
6.2.2 Suppression of surface reflection . . . . . . . . . . . . . . . . . . . . . . 142
6.2.3 Visualisation of disaggregation zone . . . . . . . . . . . . . . . . . . . . 146
6.2.4 Typical data processing flow for disaggregation in salt rock . . . . . . . . 149
6.3 Measurement results for an old cylindrical tunnel . . . . . . . . . . . . . . . . . 150
6.3.1 Measurement setup for Bernburg . . . . . . . . . . . . . . . . . . . . . 151
6.3.2 Selected scans in 2D visualisation . . . . . . . . . . . . . . . . . . . . . 153
6.3.3 3D visualisation of a tunnel section . . . . . . . . . . . . . . . . . . . . 154
6.4 Measurement results for a newly cut tunnel in Borth . . . . . . . . . . . . . . . 157
R. HerrmannContents VII
6.4.1 Subsidence analysis of measurement results . . . . . . . . . . . . . . . . 158
6.4.2 Development of the disaggregation zone in a new tunnel stub . . . . . . 160
6.5 Comparison of UWB and ultrasonic sensor results . . . . . . . . . . . . . . . . 164
6.5.1 Properties of ultrasonic prototype sensor for inspection of the disaggrega-
tion zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
6.5.2 Ultrasonic and radar results for a reference profile in Bernburg . . . . . . 166
6.6 Summary of real-world measurements in salt mines . . . . . . . . . . . . . . . . 167
7 Summary and outlook 171
7.1 Summary of UWB sensor development for salt rock inspection . . . . . . . . . . 171
7.2ry of salt rock disaggregation measurement results . . . . . . . . . . . . 175
7.3 Contributions to the state of the art . . . . . . . . . . . . . . . . . . . . . . . . 177
7.4 Aspects of further technical development . . . . . . . . . . . . . . . . . . . . . 179
7.5 Outlook for UWB sensor application in salt mines . . . . . . . . . . . . . . . . 180
7.6 A near-future vision: a new family of UWB sensors for your favourite application 182
A Algorithms for calibration of equivalent time uniform sampling 185
A.1 Definition of cost function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
A.2 Nelder-Mead based optimum search . . . . . . . . . . . . . . . . . . . . . . . . 189
A.3 Successive approximation search . . . . . . . . . . . . . . . . . . . . . . . . . . 190
A.3.1 Implementation and parameters of approximation search . . . . . . . . . 191
A.3.2 Reduction of number of test signal measurements . . . . . . . . . . . . 191
B Summary of full two-port coaxial calibration for S-parameters 193
B.1 Calculation of measured S-parameters . . . . . . . . . . . . . . . . . . . . . . . 195
B.2 Two-port calibration using all 16 error terms . . . . . . . . . . . . . . . . . . . 197
B.3 Reduction of the 16-term error model to 8 terms . . . . . . . . . . . . . . . . . 200
Bibliography 203
Publications with own contributions 214
Theses 217
Thesen 219
Erklärung / Statement of Origin 221
R. HerrmannVIII
List of Figures
2.1 Salt block and measurement setup for investigation of propagation conditions up
to 40 GHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Measurement results for transmission through salt rock with UWB stimuli from
500 MHz to 40 GHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Scattering model for thin gaps in salt rock . . . . . . . . . . . . . . . . . . . . 16
2.4 Radar cross section of spherical scatterers in the far field . . . . . . . . . . . . . 20
2.5 Illustration of SFDR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28td
2.6 Simplified typical measurement case for thin gaps in the disaggregation zone . . 30
2.7 Stimuli spectra X(f) and time domain IRFs x(t) for simulation . . . . . . . . . 33
2.8 Simulated receive signals y(t), stimulus bandwidth 1 GHz . . . . . . . . . . . . 35
2.9ted 20lg(|y(t)|), stimulus bandwidth 5 GHz . . . . . . . . 35
2.10 Simulated receive signals 20lg(|y(t)|),us 10 GHz . . . . . . . . 36
3.1 Principle for characterisation of an LTI system by UWB sensors . . . . . . . . . 40
3.2 Block schematic of a baseband M-Sequence sensor . . . . . . . . . . . . . . . . 49
3.3 Ideal M-Sequence of order 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
th3.4 Section of a 9 order M-Sequence recorded by an oscilloscope . . . . . . . . . . 52
3.5 Illustration of sub-sampling timing in M-Sequence devices . . . . . . . . . . . . 53
3.6 System outline of 12 GHz bandwidth sensor . . . . . . . . . . . . . 56
3.7 Principle and power spectra of extending an M-Sequence stimulus . . . . . . . . 57
3.8 Testing setup for the extended M-Sequence stimulus . . . . . . . . . . . . . . . 58
3.9 Spectrum magnitude of the realised extended M-Sequence stimulus . . . . . . . 59
3.10 Modified receiver structure of the new M-Sequence UWB sensor . . . . . . . . . 61
3.11 Illustration of new dense sub-sampling scheme . . . . . . . . . . . . . . . . . . 62
3.12 Frontend for S-parameter measurements and automatic calibration unit . . . . . 64
3.13 Detailed block diagram of 12 GHz bandwidth M-Sequence sensor . . . . . . . . 66
4.1 Sub-sampling scheme of new M-Sequence sensor . . . . . . . . . . . . . . . . . 68
4.2 Dense sampling of ideal sine signal for the uniform and periodic non-uniform case 72
4.3 Time domain comparison of a uniform and periodic non-uniformly sampled ideal
9 GHz sine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.4 Comparison of power spectra from a uniformly and a periodic non-uniformly sam-
pled ideal 9 GHz sine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.5 Explanation of interpreting a periodic non-uniformly sampled ideal 9 GHz sine . . 75
◦4.6 Section of idealised extended M-Sequence stimulus with f = 9GHz andφ = 45 76c 0
4.7 Illustration of analysed interval for non-uniform sampling of the UWB stimulus . 79
4.8 SCR for different alignment phases φ with random sampling time errors Δt 80samp s i
4.9 Phase shifter control hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
R. HerrmannList of Figures IX
4.10 Comparisonofacorrectlyandperiodicnon-uniformlysampledUWBpowerspectrum 86
4.11 Cost function valuesC for variation of phase shifter control values r and r . 882D 3 4
4.12 Flow chart of phase shifter control optimisation in the new sensor . . . . . . . . 91
4.13 Comparison of initial and optimised UWB power spectrum of the new sensor . . 91
5.1 Decomposition of a measurement device for calibration . . . . . . . . . . . . . . 97
5.2 Error model for 3-term correction . . . . . . . . . . . . . . . . . . . . . . . . . 99
5.3 Main paths of port crosstalk in the new M-Sequence sensor . . . . . . . . . . . 103
5.4 Receiver assignment of the new M-Sequence sensor. . . . . . . . . . . . . . . . 104
5.5 Error model for 8-term correction . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.6 Power spectrum of a calibrated open port . . . . . . . . . . . . . . . . . . . . . 109
5.7 Frequency domain window function used to suppress ill-conditioned frequencies . 111
5.8 Calibration matrix condition graphs for the two device ports . . . . . . . . . . . 114
5.9 Calibration with a defective RF switch . . . . . . . . . . 114
5.10 Phase deviation graphs for the line standard after calibration . . . . . . . . . . . 115
5.11 Illustration of different drift effects on impulse responses . . . . . . . . . . . . . 117
5.12 Acquired data and processing for long-term drift of an open standard in time domain119
5.13 Gain and delay variation of main peak from open standard . . . . . . . . . . . . 121
5.14 Residual amplitudes from switch repeatability measurement of open standard . . 122
5.15 Raw and calibrated reflection measurement of an open by the M-Sequence device 126
5.16 Comparison of calibrated reflections of an open port between M-Sequence and
ZVK sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.17 Comparison of calibrated transmission between M-Sequence and ZVK sensor . . 127
6.1 Scanner concept for salt mine measurements . . . . . . . . . . . . . . . . . . . 134
6.2 Illustration of interaction volume and antenna distance . . . . . . . . . . . . . . 137
6.3 Influence of antenna inclination on surface reflection . . . . . . . . . . . . . . . 138
6.4 Implemented array and resulting radiation pattern . . . . . . . . . . . . 138
6.5 Comparison of crosstalk reference for HH polarisation and crosstalk estimate from
mine data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.6 Example: HH polarised data before and after antenna crosstalk removal . . . . . 142
6.7 Data from a scan in HH polarisation after surface reflection removal . . . . . . . 145
6.8 Time dependent gain function used for visualisation . . . . . . . . . . . . . . . 148
6.9 Typical data processing flow for salt mine measurements of tunnel sections . . . 149
6.10 Test site in the mine of Bernburg . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.11 Measurement setup in . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.12 Example scan tunnel Bernburg, BL2, polarisation HH . . . . . . . . . . . . . . . 152
6.13 scan BL2, pola VV . . . . . . . . . . . . . . . 152
6.14 Example scan tunnel Bernburg, BL2, polarisation HV . . . . . . . . . . . . . . . 153
6.15 Disaggregation zone of a tunnel section around BL2, pol. HH, isothreshold 50% 155
6.16 zone of a around BL2, polarisation VV, isothreshold
50% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.17 Disaggregation zone of a tunnel section around BL2, pol. HH, isothreshold 37.5% 157
6.18 New tunnel stub and measurement setup in Borth . . . . . . . . . . . . . . . . 159
6.19 Amount of subsidence for new tunnel stub within 2 days after creation . . . . . 161
6.20 Interferometric comparison of datasets measured 48 h apart, pol. VV . . . . . . 163
R. HerrmannX List of Figures
6.21 Fully processed dataset measured 54 h after tunnel creation, pol. HV . . . . . . 163
6.22 Comparison of fully processed radar and sonar data from the reference profile at
BL2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
A.1 Power spectrum of the test signal (extended M-Sequence stimulus) . . . . . . . 186
A.2 Time domain shape of unwanted spectral image c(t) used for cost analysis . . . 188
A.3 Cost function value excerpt for variation of r and r . . . . . . . . . . . . . . . 1903 4
A.4 Flow chart for optimisation of phase shifter control by successive approximation . 192
B.1 Abstract system model for two-port calibration . . . . . . . . . . . . . . . . . . 194
B.2 Illustration of T-parameter error terms for 16-term and 8-term methods . . . . . 201
R. Herrmann