Table of contents

For the last eleven years, Measurement Science and Technology has awarded a Best Paper prize. The Editorial Board of the journal believes that such a prize is an opportunity to thank authors for submitting their work, and serves as an integral part of the on-going quality review of the journal.

An Editorial Board working party, comprising Patrick Gill (Chairman), Ralph Tatam and David Birch, was convened to determine a single contributed paper describing new and significant work, well aligned with the measurement scope of the journal, and presented in clear and rigorous form. They received a number of recommendations from the Editorial and International Advisory Board Members, and they would like to record their thanks to the Members for these recommendations, as they form an all-important first stage in the assessment process. There were responses from some six Board Members, together with a number of inputs from the science community. In total, there were 15 papers nominated. To aid the process, additional information in the form of the 2000 MST papers top rated by referees, and the top papers ranked by most electronic accesses, was accessed. Reviews, and papers which included a Board Member as an author, were automatically excluded. From the totality of nominations and working party deliberations, there emerged a clear winner.

Thus the paper recommended by the working party for the MST Best Paper Award for 2002 is:

• 'Electric potential probes - new directions in the remote sensing of the human body' by C J Harland, T D Clark and R J Prance, 13 163-169 (2002)

The paper is well structured with a good outline of the background to the work, a clear description of the device, followed by a range of measurement data showing the non-contact signals achievable. The capability of the device is compared with other more conventional techniques. The authors point to the potential for high resolution body mapping techniques with arrays of such probes. The concept is highly innovative, and topical, and could have very significant implications for health monitoring in the future. Its significance has been demonstrated very clearly through the largest number of electronic downloads of the article in the 2002 issues of MST, way in excess of the next most popular paper. The paper is well written in a way to appeal to a large cross-section of the readership and a wider audience. We had no hesitation in recommending this paper for the MST Best Paper Award for 2002.

Advanced sensors and instrumentation systems are becoming increasingly important in the understanding, control and optimization of combustion processes across a wide range of industrial sectors. To bring together industrialists and academics to discuss the latest developments and trends in this particular area, the ISAT (Instrument Science and Technology) Group of the Institute of Physics organized a highly focused one-day technical meeting, which was held at the Rutherford Conference Centre at the Institute of Physics in London on 9 October 2002. The event was co-sponsored by the Combustion Physics Group of the Institute of Physics, the Institute of Energy, the Institute of Measurement and Control, the Institution of Electrical Engineers, the Coal Research Forum and Sira Ltd. The event was attended by 55 participants, 24 of whom were from industry.

The special feature in this issue (on pages 1099-1158) brings together a collection of some of the papers that were presented at the event. Technical topics covered, though wide ranging as reflected in part by the diversity of the papers, demonstrate recent developments in the fields and possible approaches that may offer solutions to a broad range of measurement problems.

Sensors and instrumentation systems are required at every stage of a combustion process from fuel injection through flames to combustion emissions. On-line measurement and quantitative characterization of fuels fed into a combustion process are required in many applications. For instance, on-line continuous sizing of pulverized fuels in coal-fired power stations has been desirable for many years. Carter and Yan describe a digital imaging based system that has been developed for the on-line continuous monitoring of fuel particle size distribution. Digital imaging techniques, as reported by Shao et al, have also been applied in the automotive industry to achieve quantitative characterization of diesel sprays. Advanced flame monitoring using digital imaging techniques is being widely explored. The paper entitled `Combustion monitoring using infrared array-based detectors' by Carter and Cross presents such an approach in which a simple imaging sensor combined with digital signal processing algorithms is used. `On-line combustion monitoring on dry low NO_x industrial gas turbines' by Rea et al describes the application of digital signal processing techniques to on-line continuous monitoring of combustion dynamics of a gas turbine at a leading power generation organization. The paper by Pal et al describes the recent development of fibre Bragg grating sensor-based devices and their potential applications to temperature monitoring of combustion processes. On-line monitoring of near burner slag deposition using a hybrid neural network system is reported by Tan et al. Monitoring and characterization of combustion emissions have been a topical subject in recent years. Johnson et al describe the challenges in the development of techniques to characterize particulate emissions from aircraft. When taking measurements using low-cost scaled down test facilities it can be difficult to maintain similar conditions to those found in the full-size set-up. Jermy et al report the problems encountered in `Operating liquid-fuel air blast injectors in low-pressure test rigs' and discuss strategies for maintaining continuity under scaled down conditions.

I hope you enjoy reading the papers and find them both interesting and informative. As guest editor, I would like to offer my sincere thanks and appreciation to all the authors, referees and publishing staff of Measurement Science and Technology for their contributions and support, which made this special feature possible.

Most experimental flow investigations are carried out under imposed steady-state flow conditions. The major reason for this is that there is a lack of experimental facilities to impose well-controlled time-dependent inlet and outlet conditions on flows. There is apparently no equipment available to supply, in a well-controlled manner, the mass flow rate for time-dependent internal flow investigations. The work described in this paper remedies this situation. It introduces the basic ideas for a mass flow rate control system for time-dependent laminar and turbulent flow investigations. A first unit was built for controlled mass flow rate variations in the range of 0–217.8 g min⁻¹ under atmospheric conditions corresponding to 0–180 l min⁻¹. With this first unit the authors demonstrate that the basic ideas put forward in the paper can be used to build mass flow rate control units for experimental fluid mechanics studies. The flow rate can be simply controlled by a voltage input. The system was designed to work up to frequencies of 125 Hz.

For a quantitative study of acoustic emission signals or for the analysis of transient waveforms the transfer function of the sensors both in Rayleigh and longitudinal wave sound fields has to be known. To this end, the reciprocity calibration is applied to measure the response of acoustic emission sensors to the normal displacement velocity of Rayleigh and longitudinal waves. In this method neither a reference acoustic source nor a calibrated sensor is needed. The sensitivity is obtained through electrical measurements only, with three transducers alternately working as the sound source and receiver. Our contribution to this well-established calibration technique (the reciprocity method) comes from the use of spectral analysis techniques with short pulses, which enables the efficient determination of both the phase and the amplitude of the sensitivity in a broad frequency range. This study focuses on three points. First, we compare this method with laser interferometry calibration. Then we use the reciprocity technique to study the following points: the aperture effect and the fact that for quantitative acoustic emission analyses it is very important to take account of the material under test.

Current trends in miniaturization of microelectromechanical systems (MEMS) require the use of smaller and smaller components. Development of these microcomponents, e.g. microbeams in an accelerometer and membranes in a microphone, requires state-of-the-art test and measurement methodologies to inspect the deformation of the microcomponents for further understanding of their mechanical properties. In this paper we describe a system developed for testing deformation of a microbeam in an accelerometer under point-force load and a membrane in a microphone under applied voltage. The technique is based on optical interferometry. A collimated monochromatic beam is directed into an air wedge consisting of an optical reference plate and the microcomponent under test. The resulting interference fringe patterns from the air wedge are captured by a CCD camera mounted on a long working-distance microscope and subsequently stored in a computer. The fringe patterns that are related to deformations of the test object are analysed by a simple algorithm for recording both integral and fractional fringes. From the fringe pattern, deformation of the microbeam and membrane in the sub-micrometre range are obtained. The proposed method is potentially applicable to the in situ inspection of microcomponents in MEMS.

We describe the process of depositing gold-black on thin, freestanding pyroelectric detector substrates and compare this with previous work documented in the literature. We have evaluated gold-black coatings on thin, freestanding pyroelectric detector substrates by means of scanning electron microscope, Fourier transform infrared spectrophotometer reflectance, and spectral responsivity measurements. Spectrophotometric measurements indicate that reflectance at normal incidence varies by less than 1% at wavelengths shorter than 2.5 µm and by less than 10% at 10 µm. These results are correlated with the spectral responsivity of the detector and demonstrate that radiation not reflected by the gold-black is absorbed by the detector element. We have evaluated gold-black coatings as a function of position at two wavelengths and found variations of less than 1% at 1.25 µm and less than 5% at 10.3 µm, which demonstrates that spatial uniformity can be coating dependent. Gold-black coatings exposed to a 193 nm wavelength excimer laser were evaluated by visual inspection for damage and determined to have a damage threshold of approximately 38 mJ cm⁻².

In this paper we describe the application of an electric potential sensor to the ambulatory monitoring of the human electrocardiogram (ECG). We show that a high resolution ECG can be acquired using two of these sensors mounted wristwatch style, one on each wrist. These sensors, which do not require a real current conducting path in order to operate, are used non-invasively without making electrical contact to the subject. Furthermore, their sensitivity and low noise floor have made it possible to detect a peak which corresponds, in timing, to the His bundle depolarization—a feature not normally seen in conventional surface ECGs. We predict that these new devices will rapidly find application in the areas of clinical medicine and ambulatory monitoring.

This paper describes a new data acquisition system and reports its qualification testing. This system, tested as part of a study of two-phase flow, is able to store the useful part of the recorded signal, in our case the phase indicator function. The signal can then be processed further (in order to perform flow pattern identification and characterization for instance). This system has been tested by comparing its results with those obtained from a previous real-time data acquisition system. Cross-validation between these two systems is checked for the void fraction, the vapour time and the interface velocity. Finally, the system performance is tested by studying the influence of the duration of data acquisition on the measurement accuracy.

A new kind of high resolution, high accuracy comparator for dimensional measurements and its application for the measurement of the thickness of gauge blocks is presented. This comparator consists of two scanning tunnelling microscopes (STMs) whose tips probe the opposite sides of a sample. The STM tips are fixed on two high quality mirrors that are moved by piezoelectric actuators. The positions of the mirrors are measured by a high resolution laser interferometer, thereby coupling the length measurements to the unit of length. The detection of the sample's surface is performed by the STM probes and utilizes the high dependence of the tunnelling current on the tip-to-sample distance, which follows an exponential function and reaches about 1 nA at a tip-to-sample distance of less than 1 nm. This technique allows one to determine the position of the surface with a higher degree of accuracy than is possible using mechanical or optical probes.

The determination of a distance consists of two consecutive displacement measurements by laser interferometer without and with the sample between the two tunnelling tips. The difference of these measurements corresponds to the object's length once environmental conditions such as temperature and refractivity are taken into account.

In this paper we describe the construction and alignment of the instrument, the optimization of various components and experimental investigations of some parts of the measuring process, show first experimental results of the complete comparator obtained on thin gauge blocks and discuss the uncertainty budget.

To improve the fuel consumption and exhaust emission for gasoline engines, the GDI (gasoline direct injection) system was spotlighted to meet these requirements. Many researchers have performed studies to investigate the spray characteristics and the mixture formation of the GDI injector. In this work, we studied the spray characteristics of a gasoline direct injector by using entropy analysis and particle image velocimetry (PIV) methods. The entropy analysis is based on the concept of statistical entropy, and it identifies the degree of homogeneity in the fuel concentration. The PIV algorithm was developed to elucidate the correlation between entropy and vorticity. From the applied results on a direct injection gasoline spray, we could find that the direct diffusion phenomenon was a dominant factor in the formation of a homogeneous mixture downstream of the GDI spray, especially under vaporizing ambient conditions, and the mixing phenomenon was also progressed by momentum exchange with induced air. In addition, these results revealed that entropy analysis and PIV are very effective methods for the analysis of the mixing process, and the entropy values increase with the progress of uniformity in the diffusion process.

A new architecture for large-scale time multiplexing of fibre Bragg grating sensors (BGSs) is presented. Multiplexing of fibre BGSs often requires a trade-off between measurement performance and the number of sensors per fibre. Our technique utilizes optical amplifiers to increase the available system power for making high-sensitivity sensor measurements on a serial BGS array containing a large number of sensors. The optical amplifiers allow the system to be interrogated with a much lower-power source and simple measurement method. The design process for a variety of sensing conditions is illustrated and the performance is experimentally verified with a 100-sensor system.

Two-dimensional visualization of a liquid fuel film on a quartz cylinder liner was performed in a spark ignition engine rig. The visualization method was based on laser-induced fluorescence and total reflection. Using a quartz cylinder liner and a special quartz prism, only the liquid fuel on the cylinder liner was visualized. The calibration method was developed to quantify the fuel film thickness on the quartz cylinder liner. The calibration result showed that the fluorescence intensity increased linearly with the fuel film thickness on the quartz cylinder liner. Using this method, the wall-wetting fuel mass on the cylinder liner was quantitatively measured for different valve lifts and injected fuel masses in the test rig. The maximum fuel film thickness on the quartz cylinder liner was about 45 µm for a valve lift of 8 mm and an injection duration of 10 ms.

A sensor capable of measuring the amount of oxygen (an unwanted component that is only present because of improper filling or seal failure) within an argon-filled insulated glass window has been designed, built and tested successfully. It operates by using the optical absorption of oxygen in the atmospheric A-band centred at 762 nm. Light emitted by an argon-filled surface glow discharge lamp is Zeeman-tuned on and off an oxygen absorption line using an AC-modulated electromagnet. In the presence of oxygen, the change in the measured intensity of the lamp, obtained using standard demodulation techniques, is proportional to the oxygen column density. Measurements using an industry-standard insulated glass window indicate that the sensor can measure the amount of oxygen in a nominally filled argon window (with a window gap of 10 mm) with a precision of ±0.18% oxygen using a 16 s integration time.

This paper demonstrates the use of a simple laser schlieren technique to obtain flow information in a two-dimensional separated compression ramp-induced shock-wave boundary-layer interaction. Tests were made at a freestream Mach number of 9 and freestream Reynolds number per unit length of 2.078 × 10⁵  m⁻¹. The importance of this technique in studying hypersonic flows is unique since the run times of hypersonic wind tunnels are of very short duration. The method is based on the schlieren principle and uses a parallel sheet of a low-power (15 W) diode laser and an array of very fast response (4 ns) photodiodes. Although the arrangement detects an integral part of the signal from the fluctuating density gradients across the span of the flow, it yields significant insights into the details of the flow structure. The details of the flow field are discussed using time-dependent fluctuating density gradient profiles, the related power spectra and autocorrelation and cross-correlation functions in the interaction region.

The operating temperatures of the actual fibre in Tm:silica and Tm/Ho:fluorozirconate (ZBLAN) fibre laser systems pumped by diode lasers have been measured using a thermal imaging technique. A maximum operating temperature of 192 °C was measured in the Tm:silica fibre and 75 °C was determined to be the highest temperature the ZBLAN fibre could withstand before damage became likely. The outer polymer jacket was found to considerably increase the fibre temperature. Periodic oscillations were found in the output of the Tm/Ho:ZBLAN fibre at high output powers with periods between 12 and 27 ms. Both fibres were found to operate simultaneously at several wavelengths a few nanometres apart and the output wavelength was constantly fluctuating over time. The typical output wavelengths of the Tm:silica and Tm/Ho:ZBLAN fibres were 1.96 and 2.07 µm. Thermal effects were found to be minimal in Tm:silica fibres for incident pump powers of up to 36 W, whereas damage was possible in Tm/Ho:ZBLAN fibres at pump powers above 10 W.

This paper describes the concept and development of an innovative combination of conventional fluxgate magnetometer readout electronics with the control loop of a sigma–delta modulator in order to achieve a new magnetometer design that provides direct digital output without the use of a separate analogue-to-digital converter chip. The new concept is especially aimed at the measurement of extraterrestrial magnetic fields within a dynamic range of approximately ±2000 nT aboard scientific space missions which are affected by high radiation doses.

A digital domain model of the new magnetometer electronics is presented, which is essential for a successful hardware implementation, and the test results of a single-axis prototype are discussed.

The test results show that the fluxgate and sigma–delta modulator control loops can be merged for dynamic ranges up to ±2000 nT without significant deterioration of the overall performance of the magnetometer. The remaining quantization noise in the signal bandwidth (10 Hz) was minimized to below the sensor's noise level (7 pT Hz^−1/2 at 1 Hz) due to the noise-shaping effect of the sigma–delta principle. An offset stability of 0.25 nT over four days was achieved and the linearity error is less than ±3.3 × 10⁻⁵ even though the fluxgate sensor is not kept at near-zero field as for traditional fluxgate magnetometers.

A technological model of this new fluxgate magnetometer concept will be built for a test flight aboard the NASA discovery mission DAWN.

The paper describes a method for measuring the geomagnetic field vector using a scalar magnetometer and a three-axis flux-gate magnetometer rotating about two defined axes. The proposed method is an alternative to DI-flux measurement (D: declination; I: inclination), which is based on a theodolite-borne one-component flux-gate magnetometer. The basic idea is that the calibration of the magnetometer and the determination of the field in the direction of the rotation axis can be carried out using the same procedure. Requirements concerning the three-component magnetometer and the mechanics will be discussed. The method presented will be compared with the DI-flux procedure.

We consider in this paper the general problem of ultrasonic characterization of materials by means of analysing the dependence of attenuation on the frequency and depth of the backscattering noise. Some theoretical analysis is included to define the procedure and to gain insights into the suitability of the approach. From the depth- and frequency-dependent attenuation diagrams we may derive material signatures to be used for classification-oriented problems or derive parameters to be correlated with material properties. A particular case is considered: the characterization of cement pastes. For this case we propose the use of attenuation profiles as material signatures, and we show that the area of the profile exhibits good correlation with the porosity measured by destructive methods.

A novel self-mixing interferometer with a double external cavity was proposed to measure absolute distance and microscopic displacement. The technique developed uses an additional reference reflector to form an external dual cavity between the laser diode and the target. A fast Fourier transform phase detection technique was applied to analyse the self-mixing interferometry signal and a reference external cavity was used to compensate the error due to optical frequency fluctuation. Using these techniques, a distance resolution of 1 mm and displacement resolution down to 10 nm can be obtained.

Nonlinearities and hysteretic behaviours of piezoelectric actuators used in scanning probe microscopes (SPM) lead to uncertain position determinations of the moving probe in all directions. Therefore, probes must be controlled with nanometre resolution and the three dimensions must be traceable to the unit of length for metrological investigations. Many positioning systems exist but only a few are implementable into SPM or have unsatisfying dynamic properties as well as Abbe errors.

A hybrid implementation of a capacitive and a novel three-dimensional interferometric position measurement system has been developed to overcome these problems. The theory and technique of this holographic interferometry are given and the characteristics of the components used for implementation in the final set-up are analysed. A realized SPM head including this technique, in addition to a capacitive position system, as well as its properties are discussed.

A novel chemical image sensor developed for liquid component analysis is proposed in this paper; using it, pH values ranging from 1 to 12 and six kinds of metal ion, namely Cu²⁺, Fe²⁺, Fe³⁺, Ca²⁺, Zn²⁺, and Mg²⁺, can be detected qualitatively and quantitatively. The sensor applies the principles of optical chemistry and microfabrication technology to detect the ion concentrations in the solution, and has the advantages of high sensitivity, reduced contamination, a lower sample volume required, and the capability of detecting several indices at one time. Moreover, three multivariate data analysis methods are suggested in the paper for treating the raw data acquired from the microbeads, and predicting the results. The study demonstrates that the principal component analysis is capable of classifying six kinds of cation with success. Both partial least-squares regression (PLS) and artificial neural networks (ANN) can be used to compute the pH values quantitatively; furthermore, the PLS method has the advantage of requiring fewer iteration steps than the ANN approach.

A theory of diffraction-limited spot size is developed for infinity-corrected microscope optics. Previously reported formulae were originally derived using a single-lens system. In addition, the previously reported relationship between f-number and numerical aperture assumed a paraxial approximation and was limited to air-immersion lenses. Here, a new relationship between f-number and numerical aperture is developed, and is valid for all numerical apertures and all immersion media. In addition, a new theory is developed that estimates the effective numerical aperture of an oil-immersion lens when imaging into fluid of a lower refractive index, such as water. The results indicated that when imaging into water, high numerical aperture NA = 1.0 or 1.2 water-immersion lenses provide comparable and sometimes better diffraction-limited resolution than NA = 1.4 oil-immersion lenses. In addition, when imaging into water, water-immersion lenses may provide superior image quality, because they are corrected for aberrations resulting from the water/glass interface.

A method to calibrate hot-wire X-array probes in the low-speed (under 3 m s⁻¹) range, has been proposed and analysed. It is based on a second-order polynomial fit describing the yaw dependence of the probe sensor response. The method has been extended to the case of variable fluid temperature using the same, simple, data reduction algorithm. The temperature can be varied simultaneously with velocity in a calibration procedure. Two small X-array probes, with an additional cold sensor to measure the temperature, were constructed in order to test the proposed method. The testing was performed in constant and variable temperature fluid flow. High accuracy of the proposed method has been proven in the most demanding condition of low-speed flows.

In this paper, we suggest a technique for carrying out a high-resolution time–frequency analysis based on a modified version of the Wigner–Ville distribution. The impulse response backscattered by an aluminium and steel tube immersed in water is analysed using this method. As will be demonstrated, this approach has the desirable attributes of eliminating all interference terms manifested on the time–frequency plane and of allowing one to visualize more clearly the evolution of the reduced frequency ka of the symmetric S0 and antisymmetric A1 circumferential waves over time. This analysis allows us to determine much more accurately the cut-off reduced frequency of the antisymmetric circumferential wave A1.

Ultrasonic resonator cells for liquid attenuation and sound velocity measurements down to 70 kHz are described. The resonators are provided with easy-to-obtain concavely shaped shaving mirrors as acoustic reflectors and separated piezoelectric devices for the coupling of the cell to the electronic set-up. The advantages of this construction are shown. The mode spectrum of the resonators is discussed and measurement and evaluation procedures, also considering higher-order satellite peaks, are presented. Possible sources of experimental errors are specified and examples of measurements are given to illustrate the favourable performance of the cell design.

Two special design features of capacitor banks for driving pulsed magnets are proposed; these are dimensioned by Laplace transforms. The stray capacitance and the inductance of the wiring can result in sharp voltage pulses with amplitude of up to twice the initial charging voltage. A simple RC filter is designed which efficiently suppresses these overvoltages. The second feature is a circuit that modifies the pulse shape in order to facilitate magnetization measurements on superconducting samples where the magnetic response not only depends on the field but also on the rate of the field sweep. This is achieved by adding a suitable combination of inductors and capacitors to the capacitor bank circuit.

The feedback charge capacitance–voltage method (FCM) (Mego 1986 Rev. Sci. Instrum.58 2798) is a pure time-domain technique originally used for measuring the quasi-static capacitance of semiconductor devices. The method is based on processing the output of a charge-to-voltage converter in response to a double-step (pulse) excitation of the device. Using a transient voltage processor comprising three gated integrators connected to a mixing unit, steady-state (leakage) current may be the source of severe experimental error in capacitance. First, there is a parasitic charge during the pulse that causes an error in sampling the baseline while activating the first channel for an aperture of finite duration Δt. Second, any uncompensated leakage present after the pulse leads to a charge increment representing what is called conduction loss in the frequency domain. Measures to be taken towards minimizing both errors are provided, based on a simultaneous action of active leakage current compensation and second-order filtering. A simple hardware solution for optimizing the dynamic range of the FCM under the leakage is provided. A non-instrumental FCM error connected with the constant dielectric loss in diamond thin films, due to the anomalous kinetics j(t) ∝ t⁻¹ of the transient current, is analysed. The latter causes inaccuracy in assessing the instantaneous (geometrical) capacitance of diamond-based Schottky diodes. A proper gating of the charge-to-voltage converter is suggested for removal of the predicted error.

A new fully automated powder diffraction furnace that operates from 20 to 1600°C under controlled atmosphere or vacuum is reported. The diffraction furnace provides a large heated volume and stable sample positioning, thus minimizing systematic errors in the powder diffraction patterns and temperature control. The thermal expansion of alumina over the temperature range of 20–1420°C was used to evaluate the temperature and sample positioning errors. Analysis of 16 datasets collected over the course of 7 months, including some full-pattern Pawley refinements and Rietveld analysis, demonstrates that the thermal expansion of Al₂O₃ can be determined with good accuracy up to at least 1420°C. The specimen displacement was determined as a function of temperature using full-pattern fitting and remains well under 100 µm across the 20–1420°C temperature range. Heating rates of up to 1200°C min⁻¹ can be achieved without overshoot. Cooling rates of 600°C min⁻¹ are possible at temperatures above 1300°C, while cooling rates of approximately 60°C min⁻¹ are possible down to ∼700°C. Atmosphere control is possible, and oxygen partial pressures of less than 10⁻⁵ atm can be obtained routinely under a blanket gas.

On-line particle sizing is desirable in many industries, particularly in the power generation industry. This paper presents the principle and design of a prototype, digital imaging based, instrumentation system that can measure size distribution of particles that are largely opaque. The system, which is intended as an industrial rather than a laboratory solution, operates on-line and is non-intrusive. The image processing sequence is discussed in some detail. Experimental results obtained using two common types of particulate material demonstrate the viability of the system. The use of static images has allowed the instrumentation system to be calibrated. The linearity and repeatability of the system are quantified in addition to direct comparisons between the measurements made and those from proven, off-line, industrial solutions. The effects of the illumination sub-system on the spatial sensitivity of the viewing field are discussed along with the turndown ratio. The discrepancies between reference and static measurements are also identified and discussed. The results indicate that the system has good linearity and excellent repeatability and that practical on-line measurement of particle size distribution can be achieved using this technique. As presented the system can be used with particles in the range 150µm–25 mm.

This paper presents the application of digital imaging and image processing techniques for the quantitative characterization of diesel sprays. An optically accessible, constant volume chamber was configured to allow direct photographic imaging of diesel sprays, which were generated from a six-hole nozzle in a non-evaporating and pressurized environment. A high-resolution CCD camera and a flash light source were used to capture the images of the sprays. Dedicated image processing software has been developed to quantify a set of macroscopic, characteristic parameters of the sprays including tip penetration, near-and far-field angles. The spray parameters produced using this software are compared with those obtained using manual methods. The results obtained under typical spray conditions demonstrate that the software is capable of producing more accurate, consistent and efficient results than the manual methods. An application of the imaging processing software to the characterization of diesel sprays for a valve covered orifice nozzle is also presented and discussed.

A new generation of low-cost infrared array-based imaging systems have been put to trial within a combustion control instrumentation project. Being pyroelectric based, the device can be operated as a turbulence sensor if the scene is viewed directly or as a thermal imager by interposing a mechanical chopper. From both scientific and engineering aspects the strong carbon dioxide emission band at 4.4 µm is an ideal target for the sensor and will always be observable from carbonaceous flames in reasonable contrast against a hot background. Data from such an array are processed using powerful algorithms resident on a local digital signal processor, so promising a very low-cost route to a final instrument. A wide range of algorithm classes has been developed: initially in a 'Windows' environment but later transferred to the processor; these make use of both temporal and spatial features in the image stream. The most promising control algorithms demonstrated to date involve statistical measures of pixel activity.

To reduce the NO_x emissions levels produced by industrial gas turbines most manufacturers have adopted a lean premixed approach to combustion. Such combustion systems are susceptible to combustion-driven oscillations, and much of the installed modern gas turbines continue to suffer from reduced reliability due to instability-related problems.

The market conditions which now exist under the New Electricity Trading Arrangements provide a strong driver for power producers to improve the reliability and availability of their generating units. With respect to low-emission gas turbines, such improvements can best be achieved through a combination of sophisticated monitoring, combustion optimization and, where appropriate, plant modifications to reduce component failure rates. On-line combustion monitoring (OLCM) provides a vital contribution to each of these by providing the operator with increased confidence in the health of the combustion system and also by warning of the onset of combustion component deterioration which could cause significant downstream damage.

The OLCM systems installed on Powergen's combined cycle gas turbine plant utilize high-temperature dynamic pressure transducers mounted close to the combustor to enable measurement of the fluctuating pressures experienced within the combustion system. Following overhaul, a reference data set is determined over a range of operating conditions. Real-time averaged frequency spectra are then compared to the reference data set to enable identification of abnormalities. Variations in the signal may occur due to changes in ambient conditions, fuel composition, operating conditions, and the onset of component damage. The systems on Powergen's plant have been used successfully to detect each of the above, examples of which are presented here.

Fibre Bragg gratings (FBGs) of type I and IIA were fabricated in Ge-doped and B–Ge co-doped fibres using a 248 nm excimer laser and their performance characteristics were tested and compared with those of a chemical composition grating (CCG), written in a fluorine–germanium doped fibre, over a wide range of temperatures. Long-term testing (more than 600 h) involving a series of step-wise incremental temperature changes shows for the first time the potential of FBGs for high temperature measurement applications (up to and beyond 1100 °C), this depending on the type of FBG involved and the material and composition of the substrate fibre (the CCG was observed to be the most durable at very high temperatures). These gratings are likely to be useful for the simultaneous measurement of strain and temperature over these higher temperature ranges.

This paper is concerned with the development of a system to detect and monitor slag growth in the near burner region in a pulverized-fuel (pf) fired combustion rig. These slag deposits are commonly known as 'eyebrows' and can markedly affect the stability of the burner. The study thus involved a series of experiments with two different coals over a range of burner conditions using a 150 kW pf burner fitted with simulated eyebrows. These simulated eyebrows consisted of annular refractory inserts mounted immediately in front of the original burner quarl. Data obtained by monitoring the infra-red radiation and sound emitted by the flame were processed to yield time and frequency-domain features, which were then used to train and test a hybrid neural network. This hybrid 'intelligent' system was based on self organizing map and radial-basis-function neural networks. This system was able to classify different sized eyebrows with a success rate of at least 99.5%. Consequently, it is possible not only to detect the presence of an eyebrow by monitoring the flame, but also the network can provide an estimate of the size of the deposit, over a reasonably large range of conditions.

Particles emitted from aircraft play a role in the formation of contrails and it is essential to characterize them to understand the physical and chemical processes that are happening. Current methods for measuring aircraft particulate emissions study the reflectance of samples collected in filter papers.

A series of experiments to more fully characterize particulates has been performed on a small-scale gas turbine engine. An intrusive sampling system conforming to current ICAO regulations for aircraft emissions was used with a scanning mobility particle sizer (SMPS). Non-intrusive measurements were made using laser induced incandescence (LII) and samples were taken from the exhaust to analyse using a transmission electron microscope. Results obtained from different techniques showed good agreement with each other. As engine power conditions increased, both the SMPS and LII indicated that the mass of soot had decreased. Differences were observed between measurements of diluted and undiluted samples. The mean particle size decreased with dilution but the size distribution became bi-modal.

The study has shown how significant the sampling environment is for measuring particulates and careful techniques need to be used to ensure that accurate, consistent results can be obtained.

For reasons of economy, gas turbine atomizers are often tested at lower pressures and temperatures than exist in the engine. The spray behaviour differs in the rig from that in the engine. For good measurement practice it is necessary to find rig operating conditions at which the aspect of the two phase flow being measured (e.g. patternation, droplet size) is similar in the rig to that in the engine. We describe a number of strategies for finding operating conditions which allow good measurement. These conditions reduce the air and fuel flow rates below engine values but conserve certain flow quantities. We define an error function to quantify how well the rig conditions match key non-dimensional groups.

We consider an example of a typical aero gas turbine airblast atomizer operated in a rig at 323 K. To preserve patternation it is desirable to conserve the Reynolds (Re), Mach (Ma) and Stokes (Stk) numbers. A perfect match is not possible but we find a number of acceptable compromise strategies. The match improves as the rig pressure increases up to ∼600 kPa where Re and Ma are matched and Stk is correct to an order of magnitude. Further pressure increase give little improvement. It may be possible to match Stk by using a different fuel.