a b s t r a c tIndustrial needs are evolving fast towards more flexible manufacture schemes. As a consequence, it is often required to adapt the plant production to the demand, which can be volatile depending on the application. This is why it is important to develop tools that can monitor the condition of the process working under varying operational conditions. Canonical Variate Analysis (CVA) is a multivariate data driven methodology which has been demonstrated to be superior to other methods, particularly under dynamically changing operational conditions. These comparative studies normally use computer simulated data in benchmark case studies such as the Tennessee Eastman Process Plant (Ricker, N.L. Tennessee Eastman Challenge Archive, Available at 〈http://depts.washington.edu/control/LARRY/TE/down load.html〉 Accessed 21.03.2014).The aim of this work is to provide a benchmark case to demonstrate the ability of different monitoring techniques to detect and diagnose artificially seeded faults in an industrial scale multiphase flow experimental rig. The changing operational conditions, the size and complexity of the test rig make this case study an ideal candidate for a benchmark case that provides a test bed for the evaluation of novel multivariate process monitoring techniques performance using real experimental data. In this paper, the capabilities of CVA to detect and diagnose faults in a real system working under changing operating conditions are assessed and compared with other methodologies. The results obtained demonstrate that CVA can be effectively applied for the detection and diagnosis of faults in real complex systems, and reinforce the idea that the performance of CVA is superior to other algorithms.

Accurate prediction of interfacial friction between the gas and liquid in annular two-phase flow in pipes is essential for the proper modelling of pressure drop and heat transfer coefficient in pipeline systems. Many empirical relationships have been obtained over the last half century. However, they are restricted to limited superficial liquid and gas velocity ranges, essentially apply to atmospheric pressures, and the relationships are only relevant for pipes with inner diameters between 10 and 50 mm. In this study, we carried out experiments in a large diameter flow loop of 101.6 mm internal diameter with the superficial gas and liquid ranges of 11-29 m/s and 0.1-1.0 m/s respectively. An examination of published interfacial friction factor correlations was carried out using a diverse database which was collected from the open literature for vertical annular flow. The database includes measurements in pipes of 16-127 mm inner diameter for the liquid film thickness, interfacial shear stress, and pressure gradient for air-water, air-water/glycerol, and argon-water flows. Eleven studies are represented with experimental pressures of up to 6 bar. Significant discrepancies were found between many of the published correlations and the large pipe data, primarily in the thick film region at low interfacial shear stress. A correlation for the interfacial friction factor was hence derived using the extensive database. The correlation includes dimensionless numbers for the effect of the diameter across pipe scales to be better represented and better fit the wide range of experimental conditions, fluid properties, and operating pressures.

Kernel methods are a class of learning machines for the fast recognition of nonlinear patterns in any data set. In this paper, the applications of kernel methods for feature extraction in industrial process monitoring are systematically reviewed. First, we describe the reasons for using kernel methods and contextualize them among other machine learning tools. Second, by reviewing a total of 230 papers, this work has identified 12 major issues surrounding the use of kernel methods for nonlinear feature extraction. Each issue was discussed as to why they are important and how they were addressed through the years by many researchers. We also present a breakdown of the commonly used kernel functions, parameter selection routes, and case studies. Lastly, this review provides an outlook into the future of kernel-based process monitoring, which can hopefully instigate more advanced yet practical solutions in the process industries.

No abstract

The version in the Kent Academic Repository may differ from the final published version. Users are advised to check http://kar.kent.ac.uk for the status of the paper. Users should always cite the published version of record.

No abstract

Online monitoring of Water-Liquid Ratio (WLR) in multiphase flow is key in petroleum production, processing and transportation. The usual practice in the field is to manually collect offline samples for laboratory analysis, which delays data availability and prevents real time intervention and optimization. A highly accurate and robust sensing method is needed for online measurements in the lower end of WLR range (0%-5%), especially for fiscal metering and custody transfer of crude oil, as well as to ensure adequate flow assurance prevention and remedial solutions. This requires a highly sensitive sensing principle along with a highly precise measurement instrument, packaged together in a sufficiently robust manner for use in the field. In this paper, a new sensing principle is proposed, based on the open-ended microwave cavity resonator and near wall surface perturbation, for non-intrusive measurement of WLR. In the proposed concept, the electromagnetic fringe field of a cylindrical cavity resonator is used to probe the liquid near the pipe wall. Two of the cylindrical cavity resonance modes, TM010 and TM011 are energized for measurements and the shift in the resonance frequency is used to estimate liquid permittivity and the WLR. Electromagnetic simulations in the microwave frequency range of 4 GHz to 7 GHz are used for proof-of-concept and sensitivity studies. A sensor prototype is fabricated and its functionality demonstrated with flowing oil-water mixtures in the WLR range of 0-5%. The frequency range of the proposed sensors is 4.4-4.6 GHz and 6.1-6.6 GHz for modes TM010 and TM011, respectively. The TM011 mode shows much higher sensitivity (41.6 MHz/WLR) than the TM010 mode (3.8 MHz/WLR). The proposed sensor consists of a 20mm high cylinder, with a diameter of 30mm and Poly-Ether-Ether-Ketone (PEEK) filler. The non-intrusiveness of the sensor, along with the high sensitivity in the resonance shift, makes it attractive for practical applications.

This paper has investigated the water holdup and the pressure gradient of water-lubricated transport of high-viscosity oil flow in horizontal pipes. Experimental results on the water holdup and the pressure gradient of water-lubricated high-viscosity oil two-phase flow in a horizontal 1 inch pipe were discussed. Models for the prediction of the water holdup and/or the pressure gradient of core flow or water-lubricated flow were reviewed and evaluated. It was found that the water holdup of the water-lubricated flow is not only closely related to the input water volume fraction but also the degree of the oil phase eccentricity which is attributed to the oil phase Froude number. This can explain the inconsistency of the experimental results with regard to the relationship between the water holdup and the input water volume fraction in the literature. The applicability of the existing empirical or mechanistic models of water-lubricated highviscosity oil flow were discussed and demonstrated. A modified correlation to the water holdup correlation of Arney et al. (1993) which was shown to be exclusively applicable for concentric core flow was introduced for stable water-lubricated flow, including both concentric and eccentric core flows. This correlation was evaluated and a fair applicability was shown. The accuracy of different models for the prediction of the pressure gradient of water-lubricated transport of high-viscosity oil was demonstrated to be not high in general. This is closely associated with the difficulty in accurately accounting for the influence of oil fouling on the pressure gradient.