Search

This work focuses on the implementation of high performance systems to the wastewater treatment of sugar factories. For this purpose, systems with immobilised bacteria were studied. For the hydrolysis of organic matter and denitrification, fluidized bed reactors were used. The nitrification was studied with an airlift reactor system. Both hydrolysis and nitrogen elimination were investigated on laboratory and pilot scales in sugar factories. Although with porous materials higher biomass concentrations are attainable for the hydrolysis (up to 55 kg/m3), for economical reasons sand was used (22.5 kg/m3) for the pilot scale-study. With a pilot-scale reactor (volume 1 m3) the maximum sucrose conversion rate achieved with sand in the first campaign was 52 kg/(m3 d). For the nitrogen elimination on the pilot scale, a system with denitrification and nitrification was combined. The highest performance for the nitrification (reactor volume: 0.68 m3) with pumice as support material was 1.2 kg NH4-N/(m3 d), limiting the whole system. The denitrification rate (reactor volume: 0.12 m3) was four times higher (3.5–5 kg NO3-N/(m3 d). Rules of the modelling of the system are discussed.

SUMMARY The purpose of this paper is to provide a quick overview of neural networks and to explain how they can be used in control systems. We introduce the multilayer perceptron neural network and describe how it can be used for function approximation. The backpropagation algorithm (including its variations) is the principal procedure for training multilayer perceptrons; it is briefly described here. Care must be taken, when training perceptron networks, to ensure that they do not overfit the training data and then fail to generalize well in new situations. Several techniques for improving generalization are discussed. The paper also presents three control architectures: model reference adaptive control, model predictive control, and feedback linearization control. These controllers demonstrate the variety of ways in which multilayer perceptron neural networks can be used as basic building blocks. We demonstrate the practical implementation of these controllers on three applications: a continuous stirred tank reactor, a robot arm, and a magnetic levitation system. Copyright # 2002 John Wiley & Sons, Ltd.

The acceptor reaction of the dextransucrase from L. mesenteroides NRRL B-512 F is of technical interest for oligosaccharide synthesis. In this work the acceptor reaction of maltose, the strongest acceptor known so far, has been investigated at different experimental conditions. The data obtained were used for modeling of dextransucrase catalysis with a kinetic model that had been developed earlier.

The kinetic behavior of soluble and insoluble forms of dextransucrase from Leuconostoc mesenteroides NRRL B-1299 was investigated with sucrose as substrate and maltose as acceptor. To study the parameters involved, a kinetic model was applied that was previously developed for L. mesenteroides NRRL B-512F dextransucrase. There are significant correlations between the parameters of the soluble form of B-1299 dextransucrase and those calculated for the B-512F enzyme; that is, their properties are comparable and differ from those of the insoluble form of B-1299 dextransucrase. Whereas the calculated parameters for high maltose concentrations describe the kinetic behavior very well, the time curves for low maltose concentrations were not described correctly. Therefore, the parameters were calculated separately for the two ranges. Copyright 1999 John Wiley & Sons, Inc.

Dextransucrase catalyses the formation of dextran, but also of numerous oligosaccharides from sucrose and different acceptors, if appropriate conditions are chosen. Much experimental work has been carried out and a scheme of reactions and a mathematical model have been developed to describe the complex kinetic behaviour of the enzyme. A computer program was used to calculate the parameters of the model from a broad range of experimental data, investigating a large number of kinetic tests with the acceptors maltose and fructose. The results lead to design considerations for a continuous reactor system with immobilized dextransucrase to produce leucrose, a disaccharide of industrial interest.

Detection of Mycoplasma bovis by traditional culture methods is rather time-consuming and is often hampered by bacterial contamination. The development of a rapid and specific alternative is, therefore, an important prerequisite in improving the diagnosis of bovine diseases caused by this agent. The authors have successfully used nucleic acid probes containing genomic restriction fragments of M. bovis cloned into the plasmid vector pUC19 for species-specific detection by dot blot hybridisation of small quantities of M. bovis deoxyribonucleic acid (DNA). The problem of direct M. bovis detection from contaminated milk could not be solved using this procedure. Therefore, further research was conducted using in vitro DNA amplification by polymerase chain reaction (PCR). Species-specific nucleic acid probes were sequenced and suitable PCR primers selected. Using the PCR procedure, ten colony-forming units (CFU) were detected from broth cultures and, after DNA isolation, the equivalent of 1 CFU was detected. Direct detection of M. bovis from biological samples proved extremely difficult due to protein interference. It was shown, however, that direct PCR detection from milk is possible after effective protein removal by combined extraction and protease digestion.

Publisher Summary This chapter discusses the kinetics and reaction engineering of oligosaccharide synthesis with dextransucrase. Experiments are performed with free and immobilized enzyme with no significant difference in the reaction kinetics; high efficiencies of immobilized dextransucrase were obtained. The kinetic investigations allow giving a quantitative estimate for the acceptor strength based on the kinetic parameters of the acceptor reaction. In the study, it is found that acceptors will compete with the substrate for the acceptor-binding site. Strong acceptors such as maltose accelerate the overall reaction rate, up to a factor of over two, because of a faster acceptor reaction than dextran formation rate. Further, in the presence of the weak fructose acceptor, the slow acceptor reaction becomes the dominating step at high fructose concentrations resulting in a decrease of the overall reaction rate; maximal rates of leucrose formation were 20% of the reaction rate without acceptor.

Detection of Mycoplasma bovis by traditional culture methods is rather time-consuming and is often hampered by bacterial contamination. The development of a rapid and specific alternative is, therefore, an important prerequisite in improving the diagnosis of bovine diseases caused by this agent. The authors have successfully used nucleic acid probes containing genomic restriction fragments of M. bovis cloned into the plasmid vector pUC19 for species-specific detection by dot blot hybridisation of small quantities of M. bovis deoxyribonucleic acid (DNA). The problem of direct M. bovis detection from contaminated milk could not be solved using this procedure. Therefore, further research was conducted using in vitro DNA amplification by polymerase chain reaction (PCR). Species-specific nucleic acid probes were sequenced and suitable PCR primers selected. Using the PCR procedure, ten colony-forming units (CFU) were detected from broth cultures and, after DNA isolation, the equivalent of 1 CFU was detected. Direct detection of M. bovis from biological samples proved extremely difficult due to protein interference. It was shown, however, that direct PCR detection from milk is possible after effective protein removal by combined extraction and protease digestion.

Mycoplasma bovis, the main causative agent of mycoplasmal mastitis, arthritis and pneumonia in cattle, causes considerable economic losses. Veterinary hygiene measures would be most effective if introduced at an early stage, especially the culling of cows shedding the pathogen for the control of mastitis. It is therefore crucial to ensure that diagnostic methods are available which can perform rapid and specific detection of the agent at acceptable costs. Six different detection methods have been compared and evaluated in terms of performance parameters and suitability for routine diagnosis. Conventional M. bovis isolation and identification from culture is the only technique used for routine diagnosis at present. However, this process is rather laborious and time-consuming, and final results are available only after several days. Enzyme-linked immunosorbent assay (ELISA) techniques can be used to screen for M. bovis antibodies or antigens in clinically-diseased animals. Detection of the agent in subclinical cases was accomplished in pre-incubated samples by an antigen capture ELISA involving a monoclonal antibody. Whole-cell protein patterns generated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis were used to identify and classify field isolates. Nucleic acid hybridizations using probes of defined specificity were conducted both as filter dot blot assay and in solution with ribosomal ribonucleic acid as the target. The latter was found to be potentially suitable for the screening of biological samples, although problems due to high background and reduced specificity remained. Finally, the presence of M. bovis cells in culture supernatant and in milk samples was demonstrated using the polymerase chain reaction. This procedure is potentially superior to all others currently available, due to its high sensitivity, specificity and speed. However, a number of practical problems must be solved prior to full-scale introduction of this technique for routine diagnosis.

Numerical image processing techniques are applied to the restoration of x‐ray pinhole photographs of solid targets irradiated by intense relativistic electron beams. The numerical and analytic bases for image restoration are briefly described. An experimentally determined point‐spread function for the pinhole camera is utilized to eliminate the effects of high‐energy x‐ray degradation of pinhole resolution. It is shown that penetration by hard x rays from filtered spectra through the pinhole edges can lead to apparent broadening of the electron‐beam pinch on high‐Z targets. Elimination of x‐ray shine‐through also permits determination of photon intensity from high‐Z target ablation products, and improves spatial resolution of the x‐ray source.

The Particle-in-Cell method for the numerical solution of problems in fluid dynamics has been extended to the study of shock and rarefaction flows in a multiphase mixture. To test the numerical procedure, we have examined the propagation of sound signals through a mixture in which the theoretical speed is appreciably less than that of either material separately, with results that validate the calculational technique.

This paper presents an overall picture of the Kiwi-A reactor control systems. The techniques of analysis and design of the control systems, the reactor power and propellant flow rate control hardware, the control system performance during tests, and special measurements made on the control systems to check analytical predictions are discussed. Problems presented by future reactors and some new control systems for these reactors are described.

A crosscorrelation method for the study of the small signal response of nuclear reactor systems is presented. The technique and equipment necessary for the implementation of the method are discusse...

This short note describes two courses that dealt with microprocessors in an electrical engineering curriculum. The first course, taught at the junior level at Notre Dame, was in logic design. Training students in error-free microprocessor realization of logic circuits was a major course objective. The course was unique in that the first half was that of a "standard" logic curriculum, while the second half was devoted to microprocessor realization of combinational and sequential logic circuits. This second half included material on the Intel 8080 microprocessor, structured programming, a high level programming language (PL/M) and the application of skills in these areas to create microprocessor logic circuits. The course was self-paced. It achieved its primary objective and received an enthusiastic response from the students. The second course, taught at a senior/graduate level at the University of Tulsa, was titled "Advanced Microprocessor Systems Design." Its objectives were to introduce students to the 16 bit Intel 8086 microprocessor and associated integrated circuits, to develop student skills in both an assembly language (ASM-86) and a high level language (PL/M-86), and to teach the students the concepts of top-down structured design. Each student designed a small microcomputer based on the 8086 and its associated integrated circuits. In both courses, students learned new microprocessor and programming skills quickly and efficiently. As hoped, use of disciplined techniques such as careful problem statement and structured programming led students in both courses to error-free microprocessor logic circuits.

The Particle-in-Cell method for the numerical solution of problems in fluid dynamics was extended to the study of shock and rarefaction flows in a multiphase mixture. To test the numerical procedure, the propagation of sound signals through a mixture in which the theoretical speed is appreciably less than that of either material separately was examined, with results that validate the calculational technique. The stability of an interface between a gas and a fragmented metal impacted by shock or rarefaction was also studied. The resulting calculations give an accurate differentiation between the stable case, for which the interface should stay sharp, and the unstable case, for which there should be interpenetration. Information on both the details of interpenetration and its effects on the overall dynamics is presented. 17 figures, 12 tables.

Abstract The application of parallel computers to reservoir simulation is studied with the aid of two recently introduced Multiple Instruction, Multiple Data (MIMD) computers. The problems posed by reservoir simulators are shown to be particularly well suited for such parallel computers. Fundamental parallel programming techniques using Fortran are presented for the Hetrogeneous Element Processor (HEP) produced by Denelcor and the iPSC Hypercube produced by Intel. These techniques are then applied to two of the most time consuming tasks in reservoir simulation: forming the matrix coefficients and producing a sparse matrix solution. The problem of parallel formation of sparse matrix coefficients is addressed for the single and multiphase cases using both black oil and compositional fluid models. Parallel sparse matrix solution is addressed by considering D4 ordered Gaussian elimination, and the multigrid, conjugate gradient, and Successive Over-Relaxation (SOR) methods. Three parallel algorithms based on SOR are developed to illustrate the possible diversity of parallel algorithms. The red-black and multicolored SOR methods which are often used on vector machines are shown to be easily adaptable to parallel MIMD machine. Matrix partitioning and a new method of iteration pipelining are also presented as parallel SOR methods. The various SOR algorithms, in parallel form, are mapped onto the HEP to illustrate the speedups currently possible on an MIMD machine. Projections are made on the future importance of these types of machines to reservoir simulation. Introduction Over the past decade, significant advances have been made in the computing power available for reservoir simulation. Computer speeds have roughly doubled every two years for the past forty years. This increased computing power has made possible the practical implementation of more advanced reservoir simulator models. For example, use of the time consuming fully implicit method is now a practical alternative to the Implicit Pressure, Explicit Saturation (IMPES) method. Clearly, the advances in computing power are directly related to both the quality of the results obtained and the number of different scenarios that can be considered. In view of these changes, understanding and applying the new generation of supercomputers is of great importance to reservoir simulation. Recent advances in computing power have to a large degree been brought about by "supercomputer" vector machines such as the CRAY and the Control Data Corporation CYBER. There is increasing interest in producing even faster computers which will be highly parallel in nature. Even CRAY Inc. is following this trend by the introduction of it's XM/P computers which offer parallel central processing units. Highly parallel MIMD computers have been studied in research labs over the last dozen years and a few have been made available commercially. These parallel machines, while already in the supercomputer class, can be considered precursors to the next generation of supercomputers. An understanding of the basic capabilities and properties of these parallel machines is crucial to their appropriate application to reservoir simulation. The intent of this paper is to show basic methods of applying parallel MIMD computers to reservoir simulator problems. Also, it is hoped that the paper will demonstrate the great potential these machines hold for reservoir simulation. First the basics of parallel computers and parallel programming will be introduced. Parallel programming concepts are then applied to two of the most time consuming tasks in reservoir simulation: forming matrix coefficients and sparse matrix solution. The forming of matrix coefficients is considered for the single and multiphase cases using black-oil and compositional fluid models. Parallel matrix solution is addressed using D4 ordered Gaussian elimination, multigrid, conjugate gradient, and SOR methods. P. 281^