Table of contents

The problem of how to meter oil - water - gas mixtures has been a significant one in the oil industry since the early 1980s. Since then, considerable research has been conducted into the development of a three-phase flowmeter suitable for use in an offshore environment. This work discusses why three-phase flow measurement is important, the principal strategies and technologies which may be used to meter three-phase flows, and reviews the status of some currently available solutions.

In this paper it is shown that it is sufficient to analyse a small region of an exponentially decaying signal in order to characterize it. A technique based on the singular value decomposition (SVD) method is developed to analyse time series consisting of an exponential function and a noisy background. Techniques for analysing time series that consist of more than one decaying exponential function are also discussed.

This paper presents the development of a CAD-based and vision-aided precision measurement system. A new coordinate system alignment technique for coordinate measuring machines (CMMs) is described. This alignment technique involves a machine vision system with CAD-based planning and execution of inspection. The determination method for measuring datums for the coordinate measuring technique, using the AutoCAD development system, is described in more detail. To improve image quality in the machine vision system, a contrast enhancement technique is used on the image background to reduce image noise, and an on-line calibration technique is applied. Some systematic errors may be caused by imperfect geometric features in components during coordinate system alignment. This measurement system approach, with its new measuring coordinate alignment method, can be used for high-precision measurement to overcome such errors.

A new method of gear measurement is introduced in this paper. It is based on the traditional gear single-flank measurement, but it uses a special master gear with three kinds of measuring tooth and a tranmission tooth. The first kind of measurement tooth has a projective involute which is used as a reference. When it is engaged in mesh with the tooth of the workpiece, the deviation of transmission reflects the involute profile error of the workpiece. When a projective helix on the second kind of measurement tooth is meshed with the tooth of the workpiece the lead error can be obtained. Because the third kind of measurement tooth is the same as that used by a traditional single-flank tester, the tangential composite error can be easily obtained. The transmission tooth - whose thickness is reduced - is used to make the contact ratio less than one. When the new master gear is meshed with the workpiece for several turns the pitch error of the latter can be measured. Experimental results using equipment based on the new technique have shown that the method is feasible and efficient.

A new thin-film heat-flux gauge was designed and fabricated on three different substrate materials. Forty pairs of Pt - Pt/10% Rh thermocouple junctions were deposited in a circular pattern on the same plane of the substrate. Over the thermocouples, 5 and 10 m thick thermal resistance layers were deposited to create a temperature gradient across those layers. Calibration and testing of these gauges were carried out in an arc-lamp calibration facility. The heat flux calculated from the gauge output is in good agreement with the value obtained from the pre-calibrated standard sensor. A laser was also used to test the steady-state and dynamic responses of the heat-flux gauge. During the steady-state test, the time constant for the heating period was 30 s. The frequency response of the heat-flux gauge was measured in the frequency domain using a laser and a chopper. The responses from an infrared detector and the heat-flux gauge were measured simultaneously and compared. It was found that the thin-film heat-flux gauge has a dynamic frequency response of 3 kHz.

High-precision critical-angle refractometers, with automatic refractive index computation and readout, are now available from several manufacturers. In such instruments light is incident from within a transparent reference medium, typically an optical prism, onto a sample material in good optical contact with the prism. The critical angle of reflection is measured at the interface between the prism of known refractive index and the sample material of unknown refractive index. Critical-angle refractometers are calibrated for transparent samples but are commonly used for measurements of refractive index in optically absorbing or optically heterogeneous materials. The instrument will then read an apparent refractive index which differs from the true refractive index of the sample. The optics of the critical-angle refractometer are investigated to quantify errors which arise from the neglect of absorption and heterogeneity in the transparent sample interpretation of the critical angle. Refractive index reading errors will be important when they become larger than the refractive index precision of modern instruments, about . In critical-angle refractometry it is shown that this occurs for samples of surprisingly weak absorption. The critical-angle and Brewster-angle methods are compared for optically absorbing and heterogeneous samples. Errors arising from sample absorption in the Brewster-angle method are shown to be much less than in the critical-angle method.

We describe a room-temperature scanning magnetic microscope of novel design which can be used at a spatial resolution of a few microns to image information stored on magnetic inks used in a variety of printing processes. We suggest that this imaging technique could find general application in the inspection of currency notes and credit cards.

A new self-calibrating algorithm is described that succeeds in reconstructing an almost error-free wavefront from only three interferograms. The algorithm is based on the assumption that the optical phase, taken modulo , is quasi-uniformly distributed in the range [0, ) over the field of the interferograms. When the actual reference phases differ from those considered in the phase computation program a non-uniform histogram of the computed phase results. An analysis of this histogram allows a fitting procedure to find the actual phase shifts. Eventually an accurate shape of the wavefront can be calculated or a corrected signal can be sent to the phase-shifting device.

A reconstruction technique to calculate accurate tomograms of a sample, at a voxel scale, from tomographic experiments that involve probe - sample interactions of any degree of complexity is proposed. The properties of the reconstruction technique that accomplish this are outlined. The first guess to the solution is calculated directly from the reconstructed experimental projection data. To improve the accuracy of the approximate solution at every iteration, projection data are calculated by simulating the tomographic experiment, rather than by using a projection matrix. Calculating the ratio of the reconstructed experimental projection data to the reconstructed simulated projection data provides correction factors at every voxel. The approximate solution is multiplied by these correction factors. This correction formulation is identical to that used in the image-space reconstruction algorithm technique. High spatial resolution and accurate solutions are achieved by not implementing any form of smoothing. Instead, a novel technique is used to reduce the noise in the tomograms substantially. We call this reconstruction technique the discretized image-space reconstruction algorithm. This reconstruction technique provides a means to calculate the mass density and elemental composition tomograms of microscopic samples properly, utilizing the wealth of information measured in scanning transmission ion microscopy and particle-induced x-ray emission tomography experiments. To demonstrate the efficacy of this reconstruction technique, examples of scanning transmission ion microscopy tomography experiments are presented.

A technique is presented that allows simple and potentially automated determination of the optical path-length imbalance in ESPI systems that utilize source-wavelength modulation to obtain phase shifting. The method is based on modulating the emission wavelength of the laser synchronously with the framing rate of the CCD/image processing board and recording the average speckle intensity of the image. The injection current is increased until a , or multiple of , change in the phase has been achieved. Repeating this measurement for two different optical path-length imbalances of known difference allows determination of the laser diode coefficient and the optical path-length imbalance without the use of additional spectrometers.

A new phase modulation spatial light modulator (SLM) characterization procedure is presented, based on the analysis of the interference fringes of a Mach - Zender interferometer arrangement, by means of correlation and Fourier transform methods. This, accompanied by an amplitude calibration technique that makes use of the same experimental set-up, gives a general measurement procedure for SLM full complex characterization.

As an experimental application, two different operation curves of an Epson LCTV are determined, in order to use these configurations for an optical pattern recognition procedure: a phase-only high-efficient joint transform correlator.

For determining the efficiency of a vacuum-UV monochromator with a multiplier type PMT-142 in the range 110 - 450 nm two different methods for absolute and six methods for relative calibrations were used. The results of all relative calibrations except for one agree to within an accuracy close to the experimental error in the range of their mutual overlap. The accuracies of the relative efficiency curves were found to be 18, 14, 7, 3, 6 and 10% for the ranges 110 - 120, 120 - 140, 140 - 250, 250 - 350, 350 - 400, and 400 - 450 nm, respectively. Absolute calibrations were performed at 147 and nm with an accuracy of 8% in both cases. The differences in the grating illuminations at the monochromator applications and in the calibration were taken into account. The influences of the accuracies of all operations necessary for the absolute calibration on the overall calibration error were thoroughly analysed. If there are significant differences between the geometries of the system illumination during calibration and during application, they may be the main source of error.

A simple instrument for measuring edge angles is proposed, in which a light sectioning method is introduced. It has the advantage of simplicity, and consequently provides easy operation and an inexpensive instrument. A specimen is tilted in parallel planar laser beams. V-shaped lines are drawn by the laser beams on both sides of the specimen and are observed through a microscope. The tilt angle of the specimen can also be measured by the parallel planar laser beams. The edge angle can be calculated from the observed apex angle of the V-shaped lines and the tilt angle of the specimen. Measurement error factors are analysed and the measurement result is compensated. The measurement error of the instrument is estimated to be about an average and in variance after compensation. Although the instrument is capable of measuring the edge angles of many kinds of object, it is most suitable for measuring the edge angles of knives. The accuracy of this novel instrument may be sufficient for the quality control of knives.

A high-resolution sea-bed profiling system is described. The multiple transducer array (MTA) consists of a 37 element linear array of ultra-sonic (5 MHz) transducers. The first prototype MTA described herein measures two-dimensional bedform profiles over a length of 45 cm with approximately 1 mm vertical resolution and 2 cm horizontal resolution. Complete profiles can be recorded every five seconds. Laboratory and field tests of the MTA show the system's capability to accurately measure bedforms of known geometry and the ability to work under conditions with relatively high suspended sediment concentrations. Previous high-resolution profiling systems have either had moving parts, or have been unable to obtain the resolution of the system described here.

An experiment to extract water from wet air has been realized with a particular dielectric composite material (CM). The experimental results indicate that the typical S-shaped isotherms of CM are analogous to the data obtained by adsorption of water vapour on various mineral adsorbents, showing the fundamental nature of the phenomena involved. We have found that for 80% relative humidity ratio, for example, and for 1 g of CM, the mass of sorbed water at thermodynamic equilibrium is 2 g.

To characterize the CM by an electrical method, the microwave conductivity is measured at 8.62 GHz as a function of the mass of water sorbed from atmospheric humidity, using the cavity perturbation technique applicable to thin films in a cylindrical cavity tuned on the magnetic mode. The experimental results show that microwave conductivity presents a maximum value at the formation of the first layer of monomolecular sorbed water on the surface of CM (BET mass). On the other hand, the microwave reflectivity (MR) is examined against mass of sorbed water. The CM has MR dB (4% of reflection) at the BET mass. The conductive CM has the potential for lightweight shielding in an environment containing high-speed electronics circuits.

The lack of adequate models for feedback-controlled emission current in hot-filament cathode ionization gauges (IGs) makes it difficult to achieve precision control over a wide range of emission currents (0.01 - 10 mA). A nonlinear mathematical model for transforming the filament heating power into emission current is developed. Experimental and theoretical results are shown. An electron emission regulator is described which was designed for use in the calibration of IGs, mass spectrometers and vapour rate measurements with IGs.

This paper describes the design, implementation and testing of a pair of digitally programmable current sources which permit the controlled movement of a silicon monolithic x-ray interferometer over a range of with a resolution of better than 1 part in 500 000. Ideally a single coil-magnet actuator placed in line with the main axis of movement would ensure translation but no rotation. However, a number of factors produce parasitic twist which reduces the contrast of x-ray fringes. Two coil-magnet actuators have been placed symmetrically off the main axis of the monolith and by independent control of the currents it has been possible to compensate for asymmetries and obtain linear translation with negligible rotation. The measurement of x-ray fringes demonstrates a controlled movement of about 0.6 nm over a period of 30 min. Various techniques are considered which provide a suitable source of mA culminating in a design where each source has coarse and fine digital-to-analogue converters (DACs) with local feedback provided by an embedded proprietary 22-bit analogue-to-digital (ADC) module. Each source has a single-chip microcontroller which accepts serial messages from a computer to coordinate activity. Finally, future designs of multiple programmable current sources are considered.

The quality of coating on pharmaceutical tablets was inspected by a computer-generated hologram which was fabricated by an e-beam lithographical technique. An optical sensor for coating inspection of a tablet is proposed.

.

We present a simple method of switching a nitrogen laser with three parallel, self-breakdown spark gaps by incorporating them into a two-stage Blumlein circuit. These spark gaps are preionized by ultraviolet radiation from an auxiliary spark which suppresses their breakdown jitters and improves their temporal sychronization. The breakdown time delay of these parallel spark gaps enables strong ultraviolet preionization of the laser channel. As a result of these improvements, the laser output is doubled and is more reproducible than that obtained using the one-stage Blumlein circuit.

A combined experimental unit for high-temperature viscosity, density and surface-tension measurements in oxide melts (glasses, metallurgical slags and coal ash) is described. The high-temperature viscometer is based on the Brookfield device whereas the density and surface-tension measurements are based on the maximum-bubble-pressure method. A vertical tube furnace is capable of reaching temperatures up to . An inert (argon) atmosphere allows the use of non-standard, molybdenum spindles for viscosity measurements and permits investigation of melts susceptible to oxidation. The ranges for the measured values are 0.1 - 460 Pa s for viscosity, for surface tension, and for density. Since the range of the pressure transducer can be changed for measurements of higher surface tension and density, the experimental unit can also be used for density and surface-tension measurements in liquid metals and mattes. These high-temperature measurements are accurate to within for viscosity, for surface tension and for density.

A signal conditioning circuit based on a relaxation oscillator is presented for use with resistive transducers. Both the frequency and the duty-cycle of the output signal carry independent information coming from a pair of different sensors. Namely, the frequency of the output signal changes linearly with the resistance deviations detected by a Wheatstone bridge, while the duty-cycle is dependent on the resistance of a second sensor.

The main objective of these two volumes is to provide a comprehensive up-to-date review of the subject of surface finish. The surface finish range is limited to 10 nm rms. These surfaces could be classified as optical surfaces. The author, J M Bennett, is a highly respected researcher in this field. This edition brings together typical papers, some new (1991) and some old (1919), related to surface finish. Quite a few of them could be considered as milestone papers. For researchers this edition is a good starting point with essential papers which in turn contain numerous references.

The book starts with a historical background and an overview. This is followed by sections on instrumentation and measurements. A large section is dedicated to the various theories on scattering of electromagnetic waves. The book finishes with chapters on surface statistics, various American standards for optical surfaces and optical polishing. This book is suitable for both theoreticians and applied scientists. Although the volumes were published in 1992, the contents are by no means out of date. The reproduction of all the papers is of a good standard.

Briefly, the contents are as follows. Section II describes the characterization of surface finish. This section contains practical information about achievable surface finish ranges produced by various production processes. It also contains a list of commercially available surface finish measurement machines and their applications. Section III contains publications dedicated to instrumentation and indicates their useful application. Chapter 1 describes conventional optical microscopes (dark-field, bright field, Nomarski, interference, stereoscopic, photon-tunnelling etc.). Also papers on scanning electron microscopes and more recent scanning probe microscopes, that can make a variety of measurements on atomic scale, are included. Chapter 2 contains contributions on optical and mechanical profilers as well as area-measuring profilers. Chapter 3 addresses scatterometers and finally chapter 4 is a collection of papers on intercomparisons between the various techniques. Sections IV and V are dedicated to measurements. Section IV concentrates on measurements made by the instruments operating in the visible light region, as described in section III. Section V contains chapters on measurements using ultraviolet and x-ray scattering. Also other measurement techniques, some for rougher surfaces, are discussed here. Amongst these alternative techniques are speckle and ellipsometry. Section VI is dedicated to the various scattering theories for small and large roughness ranges. It also contains comparisons between and experiments. The last two sections, VII and VIII, are very much of practical use. In section VII, papers discuss the extraction of surface statistics, as would have been obtained from surface profiles, from light scattering. Section VIII contains various American standards for testing of optical surfaces. This section finishes with four contributions on manufacturing optical surfaces by polishing.

Data fusion is a topic which is currently causing much interest. In NDT, the combining of data from a number of independent sources related to the same test component can provide a more complete and accurate assessment of any flaws present. Ultrasonic testing, radiographic and eddy current inspection, as well as other methods, can provide the required diversity of information. However, in order to obtain meaningful results, techniques must be used which go beyond the simple addition or superimposing of the various data sources. Within this context this book by Xavier Gros is very timely, as it provides an opportunity for those interpreting NDT data to obtain an insight into data fusion methods.

Overall it is a very useful and interesting book containing many suggestions for data fusion which can form the basis of further study, and very complete reference sections at the end of each chapter. Whilst this is a major strength, it also a weakness in that individual chapters appear isolated and do not form part of a coherent whole. This is particularly true in the early review chapters where different topics appear as a list of items rather than as a properly argued line of thought. This also results in a number of repetitions in different chapters concerning POD and ROC curves as well as Bayesian inference.

The first two chapters introduce the subject of data fusion and provide a mathematical framework on the subject independently of the NDT context. Based on probability theory, conditional probabilities are used to characterize the various sensors. For example, for eddy current depth measurement, given a flaw of a certain depth, what is the probability of detecting a signal of a particular value? The application of Bayesian inference techniques and Dempster - Schafer evidential reasoning allows the combination of various probabilities associated with different measurements in order to obtain a higher level of confidence in the parameter of interest, in this case depth. The information given in this chapter is comprehensive and well presented. My only criticism is that it tends to presuppose a background knowledge of probability resulting in gaps in the development of the theory which requires further study. It would have been helpful to include more basic introductory material and examples, if only for revision purposes.

The third chapter contains a very basic introduction to a wide range of NDT methods including liquid penetrant testing and acoustic emission. These and other examples are not relevant to the data fusion cases discussed in the book and no real insight is obtained as to how such sources of information can be used in a data fusion scheme. The material is well known to the NDT practitioner and tends to detract from the natural progress of the book. On the other hand, the following chapter on scientific visualization is a very good introduction to the subject with important aspects such as visualization tools and colour maps being highlighted. It includes the important point that different colour scales, thresholds and representations of the same data can produce remarkably different images. The user has to discover how to get the most from the data. Again visualization is considered separately from data fusion and it is only when the case studies are presented do we see its real relevance. In fact it is very important for showing the results of the fusion process especially when fusing data at the pixel level from 2D scanned images.

The case studies given in the last two chapters, are used to illustrate the application of data fusion to NDT data taken from both real and artificial flaws. In the first example the binary decision of a defect being present or not present is demonstrated with impact-damaged composite material used in helicopter blades. Although a number of NDT techniques applicable to composite inspection is described along with the types of damage encountered, the data fusion mainly relates to data taken from eddy current systems. A particularly nice example of eddy current data taken from a scan over a composite plate is given in terms of a visualized 2D image. In this case the raw data map is compared with that of a conditional probability map showing the probability of a defect being present given that a particular level of signal is being observed. The resulting image shows a considerable improvement over crude averaging in removing noise and in distinguishing two clear defective areas. Although the technique depends on knowing the probability of a defect being present and the conditional probability of observing a particular signal when a defect is present, there is insufficient detail given to obtain a full understanding of how these a priori probabilities were derived.

The second example is a comprehensive study of data fusion applied to weld specimens examined using eddy current, ultrasonic and radiographic methods. In this case data fusion is used to estimate both the depth and length of a series of representative flaws including both artificial slots and real flaws such as lack of sidewall fusion, slag line and crack defects.

Much detail is given including the importance of deriving prior probability distributions for the sensors used. For eddy current crack depth measurement a series of calibrations slots is used and the results compared to Gaussian normal distributions. The same approach is taken for characterizing ultrasonic probes of different angles used to size linear flaws for length. I was somewhat disturbed to see the suggestion that for sensor calibration, slots could safely be incorporated in the parent metal of welded components since it was unlikely that cracking would occur in this region due to external loading.

As to the data fusion itself a number of different scenarios are presented including the use of Bayesian theory to improve the results from one eddy current and multiple eddy current sensors. Whilst a very useful and interesting section the description of the different probability terms being used is somewhat difficult to follow. Of particular interest are sections dealing with the fusion of data to determine the accurate length of a flaw. The outputs used included data taken from ultrasonic probes of different angles and also data taken from ultrasonic, eddy current and radiographic systems. However the important point is made that fusing data from different techniques does not necessarily increase the belief in a particular measurement, since limited prior information on a given NDT method will only increase the uncertainty in the results obtained from the overall fusion process.

The chapter concludes with a very interesting discussion on pixel-level fusion of both eddy current images related to surface flaws such as HAZ cracks, and ultrasonic C-Scan images of internal flaws such as slag lines. Gros demonstrates that naive arithmetic operations do not really work on pixel data, but that a priori information to weight images prior to addition, multiplication and subtraction can produce dramatically improved results in the separation of real features from non-relevant ones. In addition, the use of complementary techniques such as the ones above can be used to form composite images revealing both surface and internal features.

I would certainly recommend this book to engineers and scientists interested in applying data fusion methods to NDT, since it contains lots of options for data fusion, which genuinely give small improvements in accuracy or confidence limits. My only reservation is that for some methods such as ultrasonic testing, obtaining sufficient a priori information on representative flaws may in fact be difficult.

In recent years advances in technology, especially in the field of quantum optics, have made possible repetitive quantum measurements on a single quantum object, an issue that remained an academic curiosity until 25 years ago. This book, first published in 1992, provides a tutorial introduction to this field. The book is coauthored by an experimentalist and a theorist: the former - Braginsky - was the first experimentalist after Weber to develop techniques for gravitational-wave detection; the latter - Khalili - contributed with Braginsky to develop the formal apparatus to describe the experimental devices for measurement schemes.

After the first part, intended as a pedagogical appendix of a standard textbook, and devoted to the modern approach to quantum measurement, the second part is a review of some new methods of quantum measurements for macroscopic objects. The central problem is the detection of a small time-dependent classical force that acts on a quantum probe. This is the typical situation that occurs in gravitational-wave detection, where the probe is a macroscopic moving mass weakly coupled to a classical wave.

The problem of gravitational-wave detection and the limitation to accuracy of measurement of position of a moving mass - the so-called standard quantum limit (SQL), which is a consequence of the random back-action of the measuring device on the mass - triggered the development of new quantum nondemolition (QND) measurements, i.e. measurements where the measured quantity is left unperturbed after the measurement. SQL and QND measurements (both Braginsky inventions) permeate the whole book as a transversal chapter. The aim is to give simple criteria and recipes to design a measuring device in order to optimize the accuracy of the measurement, and, at the same time, to minimize the perturbation that the device exerts on the measured object, so that it will not interfere with successive measurements on the same object. The continuous measurements are regarded as the limiting case of a sequence of repeated discrete measurements, when the interval between the measurements is much smaller than the characteristic time-scale on which the measured quantity changes. The analysis of continuous measurements give to the authors the opportunity to illustrate the watchdog effect and the quantum Zeno paradox (real effect), while deriving the master equation for the evolution of the object state. Then, an entire chapter is devoted to the analysis of devices for continuous measurement of small mechanical displacements.

The book is easy to understand, and can serve as a complementary reading for graduate level courses in quantum mechanics. The material of the book can be useful for all scientists and engineers involved in researches related to high-sensitive quantum measurements, quantum information theory and quantum computation. The reader, however, should be aware that this book presents a very particular point of view, with no room left to very relevant concepts in quantum measurement theory, as, for example, squeezing, quantum amplification, joint-measurements of non-commuting observables and phase-measurements. More regrettably, the list of references is unsatisfactorily incomplete, with no mention of alternative developments, for example, regarding the SQL for measuring the position of a free mass. Here, schemes to breach the SQL have been proposed in the literature (Yuen H P 1983 Phys. Rev. Lett. 51 719, Ozawa M 1988 Phys. Rev. Lett. 60 385), whereas from the book one could conclude that such SQL would be unsurpassable. Also, a criticism is in order of the way of deriving some SQL's, based solely on the Heisenberg relation, even in consideration that the aim of such derivations is mostly pedagogical.