Publications Repository - Helmholtz-Zentrum Dresden-Rossendorf

In the framework of the external validation of the thermohydraulic code ATHLET Mod 1.1 Cycle D, which has been developed by the GRS, post test analyses of two experiments were done, which were performed at the french integral test facility BETHSY.
The BETHSY experiment 5.2c investigates the accident procedures in case of a total loss of feedwater at the steam generator secondary side, [3]. In such an accident the emergency cooling of the reactor core with primary bleed and feed, the behaviour of the steam generators in case of dry out and the long time behaviour of the test facility are special subjects of interest. During the experiment the high pressure injection system, the hydroaccumulators and the low pressure injection system were available.
The evaluation of the calculated results shows, that all main phenomena can be calculated in a good quality compared with the experiment. Resulting from various calculations it should be noticed that the quality of the results strongly depends on the modelling of the heat losses of the facility, which were partly compensated by the trace heating. This trace heating was changed several times in the experiment to compensate the changing heat losses. The exact modelling of the resulting heat losses has a strong influence on the course of the whole transient. In this test the unsufficient modelling of the resulting heat losses may be the reason for deviations of the calculated transient from the observed transient.
The results show, that the safety relevant statement of the experiment could be reproduced by the code ATHLET.

In the framework of the external validation of the thermalhydraulic code ATHLET MOD 1.1 CYCLE D, which is being developed by the GRS, post test analyses of two experiments were done, which were performed at the french integral test facility BETHSY. During test 9.3 the consequences of a steam generator U-tube rupture with failure of the high pressure injection and of the auxiliary feedwater supply were investigated. As accident management measures, the depressurization of the secondary sides, first of the two intact steam generators, then of the damaged steam generator and finally the primary depressurization by opening of the pressurizer valve were performed.
The results show, that the code ATHLET is able to describe the complex scenario in good accordance with the experiment. The safety relevant statement could be reproduced. Deviations, which did`nt impose the general results, occured concerning the break mass flow during the depressurization of the damaged steam generator and the description of the failure of the heat transfer to the damaged steam generator. Reasons are hardly to find, because these processes are highly complex.

Texture Analysis of EXAFS-Samples using the Rietveld Method

C. Hennig1, G. Nolze2
1Forschungszentrum Rossendorf e.V., ESRF- ROBL/CRG, B.P. 220, F-38043 Grenoble
2Bundesanstalt für Materialforschung und -prüfung, Unter den Eichen 87, D-12205 Berlin

Until now a lot of EXAFS measurements have been carried out on polycrystalline samples. It is known that powdered samples are characterized by a more or less strong texture, caused by an axial pressure during sample preparation. Owing to the polarized synchrotron radiation, this leads to an incorrect determination of the coordination number Nj, in particular for compounds with an anisotropic coordination center. Exemplarily, a strong polarization dependence can be detected in EXAFS spectra for oriented single crystals containing a uranyl coordination center [1]. The influence of preferred orientation on EXAFS measurements will be shown here for polycrystalline uranyl phosphate hydrate samples. The amplitude function of the EXAFS formula describes this polarization dependency with the term
Njeff = 1/2 Nj (1 + 3cos2*j), (1)
where *j is the angle between the polarization vector of the synchrotron radiation and the interatomic vector between the absorber and backscatterer. The exact calculation of the coordination number Nj is especially difficult because of its strong correlation to the Debye-Waller factor *j. A way to determine the preferred orientation is the use of X-ray diffraction measurements. The powder diffraction technique allows to determine the effective multiplicity of each Bragg reflection. One of the commonly used descriptions of a simple preferred orientation is this given by March and Dollase [2]. There, the preferred orientation is characterized by a single vector and the degree of preferred orientation:
Icorr = Istr (G2cos2 *k + G-1 sin2 *k )-2/3 (2)
For a given reflection hkl, this formula describes the relation between the corrected intensity Icorr and the integral intensity Istr resulting from well-known crystal structure data. Istr will be corrected by the preferred orientation in dependence of the orientation parameter G and the angles *k between the scattering vectors of all symmetry-equivalent lattice planes and the preferred orientation vector , assumed as lattice vector. In contrast, the orientation parameter G is valid for all reflection and must be fitted in a special refinement procedure. Most of Rietveld programs allows the use of the March-Dollase function. Both, the preferred orientation vector and the orientation parameter G should be introduced as additional amplitude correction terms for the calculation of polarization dependent EXAFS measurements on powder samples.

[1] C. Hennig, et al., Z. Krist. Suppl. 15, (1998) 156
[2] Dollase W.A., J. Appl. Cryst. 19 (1986) 267-272

Local Structure Analysis of Uranyl Phosphates and Arsenates using EXAFS Spectroscopy

C. Hennig, T. Reich, A. Roßberg, H. Funke, M. Rutsch,
G. Geipel, H. Nitsche, G. Bernhard

Forschungszentrum Rossendorf e.V., Institut für Radiochemie, P.F. 510119, D-01314 Dresden

First EXAFS measurements were taken on the new ROssendorf Beam Line (ROBL) [1]. This beamline is built by the Forschungszentrum Rossendorf as collaborating research group (CRG) at the ESRF in Grenoble. ROBL comprises a radiochemistry laboratory of type B and a materials research station. The radiochemistry station is designed for X-ray absorption spectroscopy whereas the materials research station uses X-ray diffraction techniques. The monochromator is equipped with a Si(111) and Si(311) water cooled double-crystal system of fixed-exit type which covers an energy range of 5keV-35keV. Higher harmonics are rejected by two Si/Pt coated mirrors.
Some saxonian granites contain uranium as hydrothermal crystallized uranyl phosphates and arsenates. Uranium and arsenic appear also in seepage waters of mine tailing piles from uranium mines and give stable complexes. Local structure analysis of these compounds is nessesary to understand the environmental weathering and mineralization processes.
The investigated samples are natural Cu[UO2PO4]2*nH2O, Ba[UO2PO4]2*nH2O, Ca[UO2PO4]2*nH2O, Cu[UO2AsO4]2*nH2O and synthetic H[UO2AsO4]*nH2O. The crystal structures are generally built up by stable layers of [UO2]2+ and [PO4]3-or [AsO4]3- units and the charge neutrality is achived by different interlayer cations.
U LIII-edge and As K-edge EXAFS spectra were collected in transmission, Cu K-edge EXAFS spectra were measured with a multichannel Ge fluorescence detector [2]. The measurements were carried out with a sample orientation of 0° and 45° to the beam direction in order to investigate the influence of polarization effects.
We have measured small but significant bond length differences within this structural isotypic group. Deviations in the coordination numbers are obtained as a result of polarization effects and moreover different Debye-Waller factors in connection with structural defects. In conclusion, the uranyl arsenate layer structure is nearly independent from the interlayer cation arrangement. Our investigation demonstrates the possibility for compensation the lost angle information in EXAFS spectroscopy by using the radial distribution functions at various absorption edges.

[1] ESRF Beamline Handbook 1997 pp.145
[2] Bucher, J.J., et al., Rev. Sci. Instrum. 67 (1996) 1

There is evidence from recent studies that the brain endothelium (of capillaries and/or larger vessels) may serve as a specific target for serotonin (5-HT). This neurotransmitter is expected to be involved in the regulation of the BBB permeability and/or of the cerebral blood flow via receptor-mediated mechanisms. Effective control of these processes depends on a speedy uptake and metabolism of released 5-HT molecules. To realise this, a similar mechanism of 5-HT uptake as in brain may exist at the BBB. In this study, we have demonstrated using reverse transcriptase-polymerase chain reaction (RT-PCR) that 5-HT transporter mRNA is present in the brain endothelium and that a saturable transport system for 5-HT is functionally expressed at immortalized rat brain endothelial cells (RBE4 cells). These cells take up [³H]5-HT by an active saturable process with a K _m-value of 397± 64 nmol/L and a transport capacity of 51.7 ± 3.5 pmol•g^-1•min^-1. The 5-HT uptake depends on Na⁺, as indicated by the replacement of NaCl by LiCl. The 5-HT uptake was sensitive to specific 5-HT transport inhibitors such as paroxetine, clomipramine, fluoxetine and citalopram, but not to inhibitors of the vesicular amine transporter such as reserpine or tetrabenazine. Our results demonstrate that cerebral endothelial cells are able to participate actively in the removal and metabolism of the released 5-HT, which supports the concept of direct serotonergic regulation of the BBB function.

Up-to-date methods of numerical modelling of random fields were applied to investigate some features of solute transport in saturated porous media with stochastic hydraulic conductivity. The paper describes numerical experiments which were performed and presents the first results.

New reactor designs comprise passive elements for decay heat removal. Computational-Fluid-Dynamics codes are an appropriate tool for the assessment of the efficiency of those components. Most of these codes are based either on the Finite-Volume or the Finite-Element method. Because of the importance for reactor safety these numerical tools have to be thoroughly validated using results from experimental setups.
The governing mechanism in passive components for decay heat removal is natural convection and heat transfer with internal heating. To assess the capability describing mixed convection flow, post test calculations of an IAHR benchmark exercise were performed (Kamide et. al, 1991). The commercial codes CFX-4® and ANSYS/FLOTRAN® were used, representing the Finite-Volume Method and the Finite-Element Method respectively.
This paper presents a discussion of the problems and capabilities of each code to calculate complex flow regimes and temperature fields.

A dedicated PET-system has been integrated into the heavy ion tumour facility at the Gesellschaft für Schwerionenforschung (GSI) in Darmstadt in order to monitor and control the applied dose distribution during the irradiation. The positron camera has strongly spatially-varying system behaviour due to its limited angle design and the presence of detector gaps. Furthermore, the expected counting statistics are limited by the doses applied per therapy fraction.
A three-dimensional Maximum-Likelihood Estimator algorithm has been adapted to this imaging situation. Corrections for activity outside of the field of view, parallax errors, randoms as well as normalization factors have been implemented. The algorithm has been validated using ²²Na reference sources. Resolution, edge detectability, geometrical fidelity of size and position have been chosen as figures of merit. Results of the performance studies as well as an example of the algorithm's application to patient data are presented.

Initiatoren des zweiten Workshops zum Thema „Meßtechnik für stationäre und transiente Mehrphasenströmungen" im Forschungszentrum Rossendorf (FZR) e.V. waren das Institut für Sicherheitsforschung des FZR und das Institut für Prozeßtechnik, Prozeßautomatisierung und Meßtechnik (IPM) an der Hochschule für Technik, Wirtschaft und Sozialwesen (HTWS) Zittau/Görlitz. Der Workshop wurde mit Unterstützung der Deutschen Gesellschaft für Chemisches Apparatewesen, Chemische Technik und Biotechnologie (Dechema) e.V. sowie der Kerntechnischen Gesellschaft (Fachgruppe Thermo- und Fluiddynamik) veranstaltet. In drei Haupt- sowie 13 Fachvorträgen wurden optische Meßverfahren, Impedanzverfahren und Sondergebiete zur Messung wichtiger Größen in einer Zwei- oder Mehrphasenströmung vorgestellt.

For the analysis of boron dilution transients and main steam line break scenarios the modelling of the coolant mixing inside the reactor vessel is important. The reactivity insertion due to overcooling or deboration depends strongly on the coolant temperature and boron concentration. The three-dimensional flow distribution in the downcomer and the lower plenum of PWR's was calculated with a computational fluid dynamics (CFD) code (CFX-4). Calculations were performed for the PWR's of SIEMENS KWU, Westinghouse and VVER-440 / V-230 type. The following important factors were identified: exact representation of the cold leg inlet region (bend radii etc.), extension of the downcomer below the inlet region at the PWR Konvoi, obstruction of the flow by the outlet nozzles penetrating the downcomer, etc. The k- turbulence model was used. Construction elements like perforated plates in the lower plenum have large influence on the velocity field. It is impossible to model all the orifi!
ces in the perforated plates. A porous region model was used to simulate perforated plates and the core. The porous medium is added with additional body forces to simulate the pressure drop through perforated plates in the VVER-440. For the PWR Konvoi the whole core was modelled with porous media parameters. The velocity fields of the PWR Konvoi calculated for the case of operation of all four main circulation pumps show a good agreement with experimental results. The CFD-calculation especially confirms the back flow areas below the inlet nozzles. The downcomer flow of the Russian VVER-440 has no recirculation areas under normal operation conditions. By CFD calculations for the downcomer and the lower plenum an analytical mixing model used in the reactor dynamic code DYN3D was verified. The measurements, the analytical model and the CFD-calculations provided very well agreeing results particularly for the inlet region.
The difficulties of analytical solutions and the uncertainties of turbulence models for the numerical solutions require additional experiments. Therefore a 1:5 scaled plexiglas model is under construction at RC Rossendorf. The model can be used variably for PWR's of KONVOI- and EPR- types including the primary loops. The measurements of the mixing effects will be performed with modern wire mesh sensors based on conductivity differences and LDA is used to measure the flow conditions.

The second workshop on “Measuring Techniques for Steady State and Transient Multiphase Flows” took place at the Forschungszentrum Rossendorf (FZR) from 24 - 25 September 1998. This series of meetings was initiated by the Institute for Safety Research of the FZR and by the Institute of Process Technology, Process Automation and Measuring Techniques of the Hochschule für Technik, Wirtschaft und Sozialwesen (HTWS) Zittau/Görlitz. The workshop was supported by the Deutsche Gesellschaft für Chemisches Apparatewesen, Chemische Technik und Biotechnologie (DECHEMA) and the Section “Thermo and Fluiddynamics” of the Kerntechnische Gesellschaft. Three main lectures and 13 technical papers dealt with measuring techniques such as optical and radiometric methods, impedance measurements, hot film probes and model-based methods for the measurement of important quantities in two or multiphase flows, such as phase fractions, particle size, velocities, material exchange and concentration of components.

To assist in the development of technetium-based radiopharmaceuticals that are useful for the diagnostic imaging of steroid receptor-positive breast tumors, we have synthesized a series of small-sized metal chelates according to 'n+1' mixed ligand, thioether-carbonyl and organometallic designs. In these preliminary investigations, rhenium was used as a model for the radioactive technetium. The metal chelates contain the rhenium metal in several oxidation states, being +5, +3, and +1, and they were attached to 21-substituted progesterone derivatives. A competitive receptor-binding assay (rat uterine cytosol, 0°C) was used to determine the binding affinity of these conjugates for the progesterone receptor. The highest affinity of 9% (RU5020=100%) was obtained with a '3+1' mixed-ligand complex, containing a NMe group as the central donor atom in the tridentate ligand part. This value reflects a relative binding affinity of 75% compared with the parent steroid progesterone.

Analyzing the safety margins in transients and accidents of nuclear reactors 3-dimensional models of the core were used to avoid conservative assumptions needed for point kinetics or 1-dimensional models. Therefore the 3D code DYN3D has been developed for the analysis of reactivity initiated accidents (RIA) in thermal nuclear reactors. The power distributions are calculated with the help of nodal expansion methods (NEM) for hexagonal and Cartesian geometry. The fuel rod model and the thermohydraulic part provide fuel temperatures, coolant temperatures and densities as well as boron concentrations for the calculation of feedback effects on the basis of cross section libraries generated by cell codes.

DYN3D can analyze RIA initiated by moved control rods and/or perturbations of the coolant flow. Stationary and transient boundary conditions for the coolant flow, the core inlet temperatures and boron concentrations at the core inlet have to be given. For analyzing more complex transients the code DYN3D is coupled with the plant model ATHLET of the GRS. Exercises are presented of the extensive validation work for DYN3D. Some examples are shown of application of the code.

Ausgehend von Messungen der Redoxpotentiale in Abhängigkeit von der Wassertiefe im gefluteten Teil des Schlema/Alberodaer Uranbergwerks werden die Konsequenzen bezüglich des Valenzzustandes des Eisens und Urans betrachtet. Um zu zeigen, daß nicht nur die Pyritoxidation als sauerstoffzehrender Prozeß im Grubenwasser in Frage kommt, sondern auch organische Substanzen, wie das Grubenholz einen Beitrag zur Bildung reduktiver Bedingungen leisten, wurde die Redoxsituation im Hochmoor "Kranichsee" am Erzgebirgskamm untersucht.

Im Laborexperiment wurde das Redoxpotential von Fichtenholzlignin in Abhängigkeit vom pH-Wert mittels der direkten Potentialmessung und der Titration mit Hexacyanoferrat(III) bestimmt. Dadurch werden Voraussagen bezüglich der durch Lignin verursachten Metallreduktionen möglich. Es wurde die Eisen(III)-Reduktion hinsichtlich ihres Umsatzes und ihrer Kinetik sowohl spektrophotometrisch mit o-Phenanthrolin als auch elektroanalytisch durch Square-wave-Voltammetrie untersucht.

The influence of mine-wood leachates on the adsorption of U(VI), Th(IV), Fe(III) and Pb(II) (1) on typical rocks and minerals of the mining sites in the German Erzgebirge and (2) on sediments of the Königstein mine (Saxony) and its near field was studied under aerobic condition.

The wood was decomposed by reflux boiling of wood shavings with water. The leachates were analysed for DOC, phenolic and saccharic compounds, molecular weight distribution and content of carboxylic and phenolic hydroxyl groups. The adsorption experiments were carried out equilibrating the geomaterial with the aqueous phase in the presence of U(VI) and DOC for 4-6 weeks. The uranium concentration ranged from 3.2 10-6 to 3.2 10-5 mol/L (0.76-7.62 mg/L) and the DOC content was 8.3-166 mg/L. The distribution ratio (Rs [mL/g]) was determined by liquid scintillation counting of an added 234U tracer (10-40 Bq/sample) after distribution and a careful phase separation.

Uranium adsorption on rocks and minerals from the Erzgebirge: The adsorption from a DOC-free synthetic mine water takes mainly place on diabase and on calcite; about 80% of the uranium is bound on these minerals. The influence of the wood degradation products and pine wood lignin increases the adsorption to 90-95 %. Granite and basalt adsorb from DOC-free mine water 50-60% of the uranium. In the presence of the wood leachate and lignin, the adsorption increases to about 80%. The effect of the wood leachate and lignin is insignificant for gneiss. Very little uranium adsorbs on phyllite. From all solutions practically no uranium is bound ( 2%).

Uranium adsorption on Königstein sediments: It is evident that uranium is preferentially bound to sandstone (91-97%); especially with a high distribution ratio to the limonithe-rich turonian sandstone (142 mL/g). Even under the influence of the organic compounds the sorption ranges from 68 to 93%. Wood leachate and lignin (2-4%) decrease the sorption. The adsorption o ...

The interaction of radium with characteristic sediments of the upper Saxon Elbe river valley was investigated with the aim to provide key data for transport modelling within the frame of mining site restoration. Adsorption measurements, therefore, were carried out on site-specific sediments (different kinds of sandstone, claystone and lime marl) using (1) acidic mine water coming from sulfuric acid residues of the former in-situ leaching process, (2) groundwater that is inflowing into the mine and (3) mixtures of both which simulates certain stages of the mine flooding process. The distribution ratios (Rd) were determined by batch experiments using an 8 weeks' equilibration at aquifer temperature (14 °C). Radium adsorption from acidic mine water onto different kinds of sandstone is characterized by high distribution ratios (3 400 mL/g). When contacted with groundwater, the sorption is decreased more than 20-fold (60-180 mL/g). For claystone and lime marl the distribution ratios amount to 130 and 480 mL/g, respectively, and they are nearly constant in the mine flooding process. The main adsorption mechanisms on the leached sandstone and the clayey sediments were identified as coprecipitation and ion exchange, respectively.

The long-term migration behaviour in the aquifers of the south-east Saxon Elbe basin was calculated on the basis of Hadermanns' equation, considering retardation, dispersion and radioactive decay.

With the background of uranium mine restoration the adsorption of radium on different kinds of sandstone, claystone and lime marl was studied in dependence on such parameters as water composition, acidity, phase contact time and the concentration of radium, barium and sulfate by static batch experiments at the mine temperature of 14°C.

Mit Hilfe neuartiger Gittersensoren können Gasgehaltsverteilungen in einer Zweiphasenströmung im Strömungsquerschnitt mit einer Zeitauflösung von über 1000 Hz gemessen werden. Die Sensoren beruhen auf einer elektrischen Leitfähigkeitsmessung. Zwei hintereinander angeordnete Sensoren erlauben die Ermittlung von Geschwindigkeitsverteilungen der Gasphase. Hierzu werden die Zeitverläufe des Gasgehalts an den einzelnen Meßpunkten der beiden Sensoren individuell durch Berechnung von Kreuzkorrelationsfunktionen ausgewertet. Die erhaltenen Geschwindigkeitsverteilungen können mit den ebenfalls gemessenen Gasgehalten multipliziert und über dem Querschnitt integriert werden. Dadurch erhält man den Gesamtvolumenstrom der Gasphase. Tests an einer mit Wasser-Luft-Gemisch betriebenen Versuchsschleife zeigen die Funktionsfähigkeit dieses Verfahrens.

An overview of the two years of operation of the Rossendorf CYCLONE 18/9 facility is given. Datas of the radionuclide production are presented and improvements at the cyclotron and our experience in operation and maintenance of the CYCLONE 18/9 are described.

The NOKO-Test facility is a rig for the test of passive components of nuclear power plants and is operated at the research centre J'lich. The facility is used for the experimental investigation of the emergency condenser of the SWR 600/1000, which is a new boiling water reactor design, developed by Siemens. The paper describes the numerical investigation of the natural circulation processes at the secondary side of the NOKO facility. Results of CFX-4 calculations are presented. They show, that during the operation of the emergency condenser a stable temperature stratification is evolving. At the lower part of the secondary side of the NOKO test facility even a dead volume is established. The fluid of this volume does not take part in the fluid circulation. The phenomena observed in the calculations were also found in the experimental results. Different changes of the tank configuration were investigated to reduce the effect of the temperature stratification.

In December 1997 the treatment of patients suffering from head and neck tumours with beams of high-energy ¹²C ions started at the Gesellschaft für Schwerionenforschung (GSI) in Darmstadt. By the simultaneous application of positron emission tomography (PET) the treatments are verified in-situ during the irradiation. The PET-based quality assurance procedure is a chain of consecutive steps. The starting point is the three-dimensional treatment planning on the basis of X-ray computed tomograms (CT). During the heavy ion treatment the produced positron-emitter distribution is measured simultaneously. After that the expected temporal and spatial activity distribution is computed. Both images are reconstructed iteratively and filtered. For the analysis the PET and CT images are superimposed.

The fifth dynamic benchmark is the first benchmark for coupled thermohydraulic system/three-dimensional hexagonal neutron kinetic core models. In this benchmark the interaction between the components of a VVER-440 NPP with the reactor core has been investigated.

The initiating event is a symmetrical break of the main steam header at the end of the first fuel cycle and hot shutdown conditions with one control rod group stucking. This break causes an overcooling of the primary circuit. During this overcooling the scram reactivity is compensated and the scrammed reactor becomes recritical. The calculation was continued until the highly-borated water from the high pressure injection system terminated the power excursion.

Several aspects of this very complex and complicated benchmark problem are analyzed in detail. Sensitivity studies with different hydraulic parameters are made. The influence on the course of the transient and on the solution is discussed.

Die durchgeführten Untersuchungen an 50 netzgekoppelten PV-Anlagen (Leistungsbereich < 5 kWp) ermöglichten eine vergleichende Analyse der eingesetzten Hauptkomponenten sowie eine Beurteilung der gewählten Anlagenkonzepte. Danach steht bei den Modulen -neben dem Übergang zu Großmodulen- die Erhöhung der Qualität (STC-Leistung !) eindeutig im Vordergrund. Als brauchbarstes Wandlungskonzept in den Wechselrichtern hat sich die Pulsweiten-Modulation erwiesen. Neben hohen Jahresnutzungsgraden (94 %) sollten künftige Modelle durch einen hohen Eingangsspannungsbereich (bis 600 V) die Installation von Einstrang-Anlagen ermöglichen. Durch die Integration von Komponenten zur Anlagendiagnose (etwa PR-Überwachung) und Erhöhung der Zuverlässigkeit können weitere Ertragspotentiale erschlossen werden. Jahreserträge von deutlich über 800 kWh/kWp sind damit bei gut ausgerichteten Anlagen auch im hier untersuchten Leistungsbereich erreichbar.
Durch Auswertung von Einstrahlungsdaten konnten zudem einige energiewirtschaftlich wichtige Parameter der photovoltaischen Energieerzeugung ermittelt werden.

Seit 1993 wird im Institut für Sicherheitsforschung des FZR ein Photovoltaik-Experimentierfeld betrieben. Neben einer kleinen netzgekoppelten PV-Anlage besteht die Gesamtanlage aus einem Strahlungsmeßfeld und einem Modulmeßfeld. Durch kontinuierliche mehrjährige Kennlinienvermessung wurden die ertragsrelevanten Parameter von Modulen ermittelt. Die jährlichen Energieerträge unterschiedlich orientierter Module entsprechen den Berechnungen der jeweiligen Einstrahlungssummen nach dem PEREZ-Modell.
Die STC-Leistung verschiedener Module wurde nach einem Outdoor-Verfahren bestimmt. Dabei wurden Minderleistungen zwischen 15 und 25 % gegenüber Datenblattangaben gefunden. Das auf den DC-Ertrag bezogene Performance Ratio der einzelnen Module lag zwischen 89 und 94 %. Ein mittleres jährliches Leistungsspektrum -gültig für große Teile Deutschlands und kristalline Module- wurde ermittelt.

Es wird eine Kerbschlagbiegeanordnung vorgestellt, mit der simultan Schlagkraft, Rißöffnung, Schallemission und magnetische Emission gemessen werden können. Grundbaustein der Anordnung ist ein invertiertes Pendelschlagwerk. Mit dieser Anordnung wurden seitgekerbte, vorermüdete ISO-V-Proben aus Stählen mit unterschiedlichen Festigkeits-Zähigkeitseigenschaften geprüft.
Die Untersuchung soll aufzeigen, ob Schallemission zur Detektion der stabilen Rißinitiierung an hochzähen Stählen geeignet ist und folglich auf diesem Wege die dynamische Rißinitiierungszähigkeit zuverlässig bestimmt werden kann. Dazu wurden andere Bestimmungsverfahren herangezogen. Als zuverlässigste Verifizierungsbasis wird die physikalische Rißinitiierungszähigkeit betrachtet, die auf der Grundlage der Messung der Stretchzone und der Rißwiderstandskurve in einem Mehrprobenverfahren ermittelt wird. Außer im unteren Übergangsbereich, in dem nur sehr geringer duktiler Rißfortschritt auftritt, konnte mit Schallemission in allen Fällen stabile Rißinitiierung nachgewiesen werden. Die damit bestimmte Rißinitiierungszähigkeit liefert über das gesamte Spektrum der untersuchten Stähle eine nahezu 1:1-Korrelation zur physikalischen Rißinitiierungszähigkeit. Im Einzelfall treten aber auch beträchtliche Streuungen auf und Rangfolgen werden innerhalb einer Versuchsserie nicht rangordnungsgerecht wiedergegeben. Alle anderen untersuchten Einprobenverfahren schneiden in diesem Vergleich deutlich schlechter ab und sind somit als Verifizierungsbasis nicht geeignet.
Die Untersuchung läßt nicht zwingend erkennen, daß Schallemission tatsächlich direkt mit dem Übergang der maximalen Rißabstumpfung zum stabilen Rißwachstum in Verbindung steht. Offensichtlich werden aber Prozesse detektiert, die im zeitlichen und örtlichen Umfeld der Rißinitiierungsprozesses stattfinden. Somit kann im instrumentierten Kerbschlagbiegeversuch mit Hilfe einer zusätzlich in die Hammerfinne integrierten Schallemissionssonde eine brauchbare Schätzung des bruchmechanischen Parameters dynamische Rißinitiierungszähigkeit effektiv und ohne großen Aufwand erhalten werden.

Für die Validierung von Computercodes zur dynamischen Simulation von Druckentlastungsvorgängen werden experimentelle Daten benötigt. Besonders geeignet sind dabei Versuchserien, bei denen einzelne Parameter variiert werden. Mit Hilfe von Parameterstudien lassen sich Rückschlüsse sowohl auf die Güte der Einzelmodelle als auch des Gesamtmodells ziehen. Solche Daten liegen jedoch meist nur für Druckentlastungen verdampfender oder gasentlösender Stoffsysteme vor. In der Praxis ist aber i.a. die Druckentlastung im Fall durchgehender exothermer Reaktionen interessant. Daher wurde eine geeignete Modellreaktion gesucht. Die thermische Zersetzung von Wasserstoffperoxid zeichnet sich dadurch aus, daß sie einfach durchführbar ist, daß mit Wasser und Sauerstoff unproblematische Reaktionsprodukte entstehen und daß die Reaktion allein durch Wärmezufuhr ausgelöst werden kann.

Im Rahmen der externen Validierung des von der Gesellschaft für Anlagen- und Reaktorsicherheit entwickelten Störfallcodes ATHLET, der in der Version Mod 1.1 Cycle D vorlag, wurden zwei Experimente nachgerechnet und analysiert, die an der französischen Versuchsanlage BETHSY durchgeführt wurden. Im Test 9.3 werden die Konsequenzen untersucht, wenn bei einem Heizrohrbruch die Hochdruckeinspeisung sowie die Not-Speisewasserversorgung der Dampferzeuger versagen und nur die Druckspeicher sowie die Niederdruckeinspeisung zur Verfügung stehen. Als Accident Management Maßnahmen wurde die sekundäre Druckentlastung und schließlich die primäre Entlastung über den Druckhalter untersucht.
Die Analyse kommt zu dem Ergebnis, daß der Code ATHLET in der Lage ist, dieses komplexe Szenario recht gut zu beschreiben. Die sicherheitsrelevante Aussage konnte reproduziert werden. Abweichungen, die jedoch nicht die prinzipielle Aussage in Frage stellten, ergaben sich bei der Berechnung des Leckmassenstromes in der Phase der Druckentlastung des havarierten Dampferzeugers sowie bei der Modellierung des Versagens der Wärmeübertragung dieses Dampferzeugers. Die Ursache dieser Abweichungen ist sehr schwer aufzuklären, da diese Prozesse sehr komplex miteinander verflochten sind.

Im Rahmen der externen Validierung des von der Gesellschaft für Anlagen- und Reaktorsicherheit entwickelten Störfallcodes ATHLET, der in der Version Mod 1.1 Cycle D vorlag, wurden zwei Experimente nachgerechnet und anlaysiert, die an der französischen Versuchsanlage BETHSY durchgeführt wurden.
Das Experiment 5.2c dient der Untersuchung der Notfallprozeduren beim Totalausfall der Speisewasserversorgung auf der Dampferzeugersekundärseite. Spezielles Interesse gilt hierbei den Möglichkeiten der Notkühlung durch primäres Bleed and Feed, dem Verhalten der Dampferzeuger bei sekundärseitiger Austrocknung und dem Langzeitverhalten der Anlage bis zum Einspeisen der einzelnen Notkühlsysteme. Im Verlauf der Transiente erfolgt die Einspeisung sowohl aus dem Hochdrucknotkühlsystem als auch aus den Druckspeichern und dem Niederdrucknotkühlsystem.
Die Auswertung der Rechnungen zeigt, daß praktisch alle wesentlichen Phänome im Verlauf der Transiente korrekt wiedergegeben werden. Als ein wesentlicher Einflußfaktor auf die Qualität der Rechnungen hat sich die Modellierung der Wärmeverluste, welche an der Versuchsanlage teilweise durch eine Zusatzheizung (Trace Heating) kompensiert werden, herausgestellt. Werden die Wärmeverluste im Primärkreis falsch berechnet, so ergeben sich signifikante Abweichungen im Verlauf des Primärdrucks. Da alle Prozesse im weiteren druckgesteuert erfolgen, wirken sich Abweichungen im Primärdruck besonders stark auf den Verlauf der Transiente aus.
Die Ergebnisse der Rechnungen zeigen darüberhinaus, daß die sicherheitsrelevante Aussage des Experiments durch den Code ATHLET reproduziert wird.

The formation of helium induced cavities in silicon during short-time annealing is analyzed by spectroscopic ellipsometry. Specimens implanted with 40 keV He+ ions to a dose of 5×1016 cm–2 are heat treated at 800 °C for times of 1–1200 s by rapid thermal annealing. Spectroscopic ellipsometry is employed to obtain quantitative information on the cavity volume depth profiles. A newly developed formula is used to model the optical multilayer depth profiles. The cavity volume is found to increase during annealing for about 300 s and to decrease for longer annealing times. Over this characteristic time a marked change in the He loss occurs, which has been reported only recently. Swelling of the helium implanted and annealed silicon is analyzed using an atomic force microscope. Step heights are consistent with the cavity volume per unit area obtained from spectroscopic ellipsometry data analysis. The number density of cavities after annealing for 600 s is calculated to be 1.16±0.27×1017 cm–3 and is found to be largely independent of depth in the central part of the cavity layer.

In 1995 at the integral test facility ISB-VVER in Elektrogorsk near Moscow, natural circulation experiments were performed, that were scientifically supported by the Forschungszentrum Rossendorf. These experiments were the first of this kind at a test facility, that models VVER-1000 thermal-hydraulics. Using the code ATHLET, which is being developed by Gesellschaft für Anlagen- und Reaktorsicherheit, pre- and posttest calculations were done to determine the thermal-hydraulic events to be expected and to define and tune the boundary conditions of the test. The conditions found for natural circulation instabilities and cold-leg loop-seal clearing could be confirmed by the tests. In the paper, the experimental results are presented, analysed and compared with ATHLET calculations. The main thermal-hydraulic phenomena were identified and compared with the results gained during similar experiments on VVER-440 test facilities. Besides the thermal-hydraulic standard measurin!
g system, the facility was equipped with needle-shaped conductivity probes for measuring the local void fractions.

A description of the fuel rod behaviour and heat transfer model used in the code DYN3D for nuclear reactor core dynamic simulations is given. Besides the solution of heat conduction equations in fuel and cladding, the model comprises detailed description of heat transfer in the gas gap by conduction, radiation and fuel-cladding contact. The gas gap behaviour is modeled in a mechanistic way taking into account transient changes of the gas gap parameters based on given conditions for the initial state. Thermal, elastic and plastic deformations of fuel and cladding are taken into account within 1D approximation. A creeping law for time-dependent estimation of plastic deformations is implemented. Metal-water reaction of the cladding material in the high temperature region is considered. The cladding - coolant heat transfer regime map covers the region from one-phase liquid convection to dispersed flow with superheated steam. Special emphasis is put on taking into account the impact of thermodynamic non-equlibrium conditions on heat transfer.
For the validation of the model, experiments on fuel rod behaviour during RIAs carried out in Russian and Japanese pulsed research reactors with shortened probes of fresh fuel rods are calculated. Comparisons between calculated and measured results are shown and discussed. It is shown, that the fuel rod behaviour is significantly influenced by plastic deformation of the cladding, post crisis heat transfer with subcooled liquid conditions and heat release from metal water reaction.
Numerical studies concerning the fuel rod behaviour under RIA conditions in power reactors are reported about. It is demonstrated, that fuel rod behaviour at high pressures and flow rates in power reactors is different from the behaviour under atmospheric pressure and stagnant flow conditions in the experiments. The mechanisms of fuel rod failure for fresh and burned fuel reported from the literature can be qualitatively reproduced by the DYN3D model. However, the model must be extended and improved for proper description of burned fuel behaviour.

The fifth dynamic benchmark was defined at the 7th AER-Symposium, held in Hörnitz, Germany in 1997. It is the first benchmark for coupled thermohydraulic system/three-dimensional hexagonal neutron kinetic core models. In this benchmark the interaction between the components of a VVER-440 NPP with the reactor core has been investigated.

The initiating event is a symmetrical break of the main steam header at the end of the first fuel cycle and hot shutdown conditions with one control rod group stucking. This break causes an overcooling of the primary circuit. During this overcooling the scram reactivity is compensated and the scrammed reactor becomes recritical. The calculation was continued until the highly-borated water from the high pressure injection system terminated the power excursion. Each participant used own best-estimate nuclear cross section data. Only the initial subcriticality at the beginning of the transient was given.

Solutions were received from Kurchatov Institute Russia with the code BIPR8/ATHLET, VTT Energy Finland with HEXTRAN/SMABRE, NRI Rez Czech Republic with DYN3D/ATHLET, KFKI Budapest Hungary with KIKO3D/ATHLET and from FZR Germany with the code DYN3D/ATHLET.

In this paper the results are compared. Beside the comparison of global results, the behaviour of several thermohydraulic and neutron kinetic parameters is presented to discuss the revealed differences between the solutions.

An improved synthesis of the enantiomerically pure thioester precursors of ¹¹C](+)-McN-5652 ([¹¹C](+)-1), and [¹¹C](-)-McN5652 ([¹¹C](-)-1) starting from racemic McN-5652 ((±)-1) is described. The synthetic method includes the resolution of (±)-1 by repeated crystallization of the (+)- and (-)-di-p-toluoyltartrates yielding (+)-McN-5652 ((+)-1) and (-)-McN-5652 ((-)-1), each with > 98 % enantiomeric purity. S-Demethylation of (±)-1, (+)-1 and (-)-1, respectively was achieved by treatment with sodium amide at low temperatures (-78 °C) followed by conversion of the intermediate thiols 2 with acetyl chloride to give the corresponding thioester precursors (±)-3, (+)-3 or (-)-3. This demethylation method almost completely suppressed the isomerization of the pharmacologically active trans diastereomer into the inactive cis form. Chiral HPLC analyses confirmed that the S-demethylation proceeded without any racemization. ¹¹C-Labelling of (+)-3 or (-)-3 yields enantiomerically pure [¹¹C](+)-McN-5652 or [¹¹C.

The coupled thermohydraulic system/3D-neutron kinetic code complex DYN3D/ATHLET was applied to an asymmetric main steam line break (MSLB) analysis for the Russian VVER-440 type reactor. Such type of MSLB accidents cause an asymmetrical overcooling of the reactor. In this case the coolant mixing inside the reactor pressure vessel (RPV) has an important influence on the behaviour of the reactor. The code DYN3D includes a special model for the mixing of coolant from different primary loops in the lower plenum of VVER-440 type reactors which can be used in the coupled code, too. This model is based on the analytical solution of the Navier-Stokes equations in the potential flow
approximation in 2-dimensional cylindrical geometry and the diffusion equation for heat transport or soluble poison. The model is validated against experimental results from a 1:5 VVER-440 flow model with air and experimental data from VVERs with all main coolant pumps (MCP) working. Using this model for the coolant mixing in the MSLB analysis, recriticality of the scramed reactor was predicted. If homogeneous coolant mixing is assumed, no recriticality will be obtained.
The stationary three-dimensional
flow distribution in the downcomer and the lower plenum of a VVER-440/V-230 reactor was calculated with a CFD code (CFX-4). For this calculation the RPV from the cold legs inlets, the downcomer, the lower plenum and the lower core support plate was nodalized in detail. The comparison with experimental data and the above mentioned analytical mixing model showed a good agreement for near-nominal conditions (all MCPs are running). However, the comparison between the CFD-results and the analytical model revealed differences for MSLB conditions. After shutdown of the MCPs, natural convection established in the primary circuit. The mass flow rate of the affected by the MSLB loop is approximately twice the value of one of the other loops, a redistribution of the flow below the inlet nozzle of the affected loop is observed. For this reason the temperature field at the core entry cross section has two equal minima next to the position of the concerned inlet nozzle. The temperature distribution obtained by the analytical model has one minimum, just near to the position of this inlet nozzle. The shape of the temperature distribution for MSLB conditions is practically the same like in nominal conditions. The extension of this sector due to the increased mass flow is properly considered by the model.
The core inlet temperature distribution obtained by means of CFX-4 was used to estimate the reactivity effect in the MSLB analysis.

The change of the spreading velocity of ultrasonic surface waves by thin film deposition is descibed by an approximation of the nonlinear equations of motion. The effect of residual stress in deposited films on the spreading velocity is considered. As an example for practical applications, the velocity change of surface waves after TiN-film desposition on steel bars in investigated.

Abstract
In ultrarelativistic heavy-ion collisions at √s > 20 (120) A GeV a copious production of charm (bottom) production sets in which, via correlated semileptonic DD (BB) decays, gives rise to a dilepton yield at invariant mass M ≈ 2 - 3 GeV in excess of the Drell-Yan yield and the thermal dilepton signal from deconfined matter as well. We show that appropriate single-electron transverse momentum cuts (suitable for ALICE at LHC) cause a threshold like behavior of the dilepton spectra from heavy-quark meson decays and the Drell-Yan process and can allow to observe a thermal dilepton signal from hot deconfined matter.

Das schwerste Element, dessen chemische Eigenschaften bisher experimentell untersucht wurden, ist das Element 106, Seaborgium. Kürzlich wurde am UNILAC der GSI in einem vom Institut für Radiochemie des FZR am U-120 vorbereiteten gaschemischen Experiment die Bildung und Hochtemperaturflüchtigkeit von Seaborgiumoxidhydroxid nachgewiesen.
Seaborgium ist wie die anderen Transactiniden nur in Kernreaktionen mit schweren Ionen zugänglich, wobei die Produktionsraten der kurzlebigen Isotope 263Sg (t1/2 = 0,9 s), 265Sg (t1/2 = 7,4 s) und 266Sg (t1/2 = 21 s) bei wenigen Atomen pro Tag liegen.
Daraus ergibt sich die Aufgabe, die chemischen Eigenschaften des Seaborgiums an einzelnen Atomen zu untersuchen. Zur Untersuchung einzelner Atome sind gaschromatographische Methoden besonders prädestiniert. Als chemisch ähnliche Vergleichselemente für die Charakterisierung des Seaborgiums kommen vor allem Molybdän und Wolfram in Betracht.
Die experimentelle Vorbereitung am U-120 verfolgte zunächst die gaschromatographische Untersuchung von Molybdän und Wolfram in Spurenmengen. Dabei wurde nachgewiesen, daß die Gaschromatographie von Oxidhydroxiden des Molybdäns und Wolframs in Quarzglassäulen bei Verwendung von feuchtem Sauerstoff als Reaktivgaskomponente des Trägergases durch dissoziative Adsorptions- und assoziative Desorptionsreaktionen determiniert wird.

Zur methodischen Vorbereitung der aufwendigen Transactinidenexperimente wurde das Seaborgium durch Molybdän und Wolfram modelliert. Wir entwickelten und testeten einen Mehrzonenchromatographieofen für Reaktions- und Chromatographietemperaturen bis 1400 K, der direkt an die Detektionseinheit ROMA (rotating wheel multi detector analyzer) angekoppelt wurde. Zur Direktankopplung der Detektion an die Chromatographie sind die mit über 900 K aus der Chromatographiesäule austretenden Spezies auf Metallfolien zu kondensieren. Nach einer Ausbeuteoptimierung unter Aufrechterhaltung einer mindestens 99 %igen Aerosolrückhaltung wurden für 105Mo (t1/2 =35,6 s) bei einer Chromatographiesäulentemperatur von 1300 K bis zu 70 % der in die Chromatographieapparatur eingespeisten Aktivität auf Aluminiumfolien abgeschieden. Die Verweilzeit in der Chromatographieapparatur betrug ca. 10 s.
Die Auswertung des Seaborgiumexperiments am UNILAC der GSI hatte zum Zeitpunkt der Niederschrift dieses Berichts gerade begonnen, es konnte aber bereits resümiert werden, daß es mit Hilfe des U-120 gelungen war, die Lösung einer äußerst anspruchsvollen Aufgabenstellung der modernen Kern- und Radiochemie gründlich vorzubereiten.

The paper describes the formation of Ge nanocrystals in thin thermally grown SiO₂ layers (d_ox=<100 nm) using implantation of 10¹⁵ - 2x10¹⁶ Ge⁺/cm² and subsequent annealing. Although the implanted Ge depth profile is distributed over almost the whole SiO₂ layer, a very narrow band (typical width 5 nm) of Ge nanoclusters very close but well-separated to the Si/SiO₂-interface is formed by self-organization under specified annealing conditions. A possible mechanism for this self-organization process is discussed including nucleation phenomena, Ostwald ripening and defect-stimulated interface processes. Simple MOS-structures were prepared and the effect of charge storage inside the clusters has been derived from C-V characteristics.

Dünne Filme aus kubischem Bornitrid (c-BN) sind aufgrund ihrer herausragenden thermomechanischen, chemischen und elektronischen Eigenschaften von besonderem Interesse für die Materialforschung. Bornitrid in der kubischen Phase ist das nach Diamant härteste bekannte Material. Anders als Kohlenstoff verhält es sich gegenüber ferrithaltigen Metallen chemisch inert und zeigt auch bei hohen Temperaturen eine vergleichsweise geringere Oxidationsneigung. Wegen seiner hohen thermischen Stabilität und der Möglichkeit, dünne Schichten bei niedrigen Temperaturen zu synthetisieren, ist Bornitrid  als  Hard-Coating-Material für Werkzeuge besonders geeignet. Mögliche Anwendungen für mikroelelektronische Hochleistungsbauelemente ergeben sich aus der hohen Wärmeleitfähigkeit und der weiten Bandlücke (E_g ~ 6 eV). Bornitrid kann mit Beryllium und Silizium n- bzw. p-Typ dotiert werden und läßt sich mit einer Oxidschicht passivieren. Die optische Transparenz im sichtbaren und Infrarotbereich macht das Material zudem für die Oberflächenvergütung optischer Bauelemente geeignet. ...

Der Einfluß plastischer Verformung nach einsinniger Zug- und Biegebelastung auf die Kleinwinkelstreuintensität wurde an technischen Al-Legierungen untersucht.
Bei geeigneter Versuchstechnik lassen sich sowohl mit Röntgenstrahlung als auch mit Neutronen verformungsbedingte Effekte nachweisen. Sie sind im Einschnürbereich einer Zugprobe als Porenstreuung und im Bereich vor der Rißspitze von biegeverformtem, hochduktilem Aluminium als Streuung an Versetzungsstrukturen zu interpretieren. Durch punktförmiges Abtasten mit einem stark fokussierten Synchrotron-Röntgenstrahl können inhomogene Schädigungsstrukturen mit hoher Ortsauflösung nachgewiesen werden. Diese mit SM-SAXS (Scanning Microbeam - Small Angle X-ray Scattering) bezeichnete Technik ist ein Schritt zur Entwicklung eines
Kleinwinkelstreu-Mikroskops.

Small Angle Scattering Characterization of deformation-damaged microstructure
The microstructural changes preceding the ductile fracture were investigated by small angle scattering methods at commercial Al alloys. Within the necking region of tensile deformed specimens additional small angle scattering effects can be measured. They are in agreement with the assumption of pore scattering. The inhomogeneous deformation field ahead the crack tip of a bended and precracked specimen can be monitored by scanning with a highly focussed X-ray beam of a Synchrotron source. In this case the scattering intensity decreases with increasing distance from the crack tip. The effect can be interpreted as scattering at the dislocation arrangements. The used scanning microbeam small angle X-ray scattering technique is a first step to the development of a "small angle scattering microscope".

The results of a project sponsored by the German Bundesministerium fuer Bildung, Wissenschaft, Forschung und Technologie are presented. The Project aimed to improve the safety against embrittlement of VVER-1000 type reactors by a more reliable and accurate determination of the neutron load of reactor pressure vessels. Therefore, six scientist from three Russian research institutions were sponsored to support with their work another BMBF project of the FZR aimed at the same goal. By providing reliable data for the evaluation of ex-vessel neutron activation experiments at two VVER-1000 and formulating the corresponding reactor models a basis has been established for further investigations as well in the FZR as well as in several Russian and Western research institutions. The leading Russian nuclear data library ABBN/MULTIC has been improved and tested. The uncertainties affecting the calculations of the fluence spectrum at the outer boundary of the pressure vessel have been analysed and a spectrum covariance
matrix has been derived. The methodologies for the experimental determination of activation rates and for calculations of fluence spectra and activation rates have been further developed and tested by interlaboratoy comparisons. Measurements of different laboratories were compared with each other, as well as the corresponding calculations. Moreover,
measurements and calculations were compared against each other, partly with participation of further Russian, Czech and Western institutes. The results of the Intercomparisons have been evaluated by the "International Workshop on the Balakovo-3 Interlaboratory Dosimetry Experiment" in September 1997 in Rossendorf. As a result of these works a better evaluation of the reached accuracies was possible and proposals for an improvement of the used methods could be derived.

Die Klassifikation eines chemisch-verfahrenstechnischen Prozesses nach Gefahrenklassen wird unter besonderer Berücksichtigung der Prozeßsicherheit behandelt. Mangels verläßlicher A-priori-Information über den Prozeß (z. B. eines dynamischen Prozeßmodells) werden klassifizierungsrelevante Prozeßeigenschaften (Merkmale) allein aus Prozeßsignalen gewonnen. Fundamentales Hilfsmittel ist dabei die statistische Informationstheorie. Außer durch die Anwendung eines speziellen und - weil zur Kategorie der Assoziativspeicher gehörenden - besonders einfachen Klassifikators wird insbesondere durch die Nutzung der statistischen Informationstheorie erreicht, daß die Klassifizierungsergebnisse im Interesse der Prozeßsicherheit transparent und nachvollziehbar sind; ungeeignete Merkmale können dadurch leicht lokalisiert werden. Wie exemplarisch gezeigt wird, sind Zusammenhänge, die informationstheoretisch zwischen bestimmten Prozeßsignalen nachgewiesen werden, für die Strukturierung eines dynamischen Prozeßmodells nutzbar. Ein solches Modell kann zur Effektivierung der Prozeßmerkmale und folglich zur genaueren Klassifikation genutzt werden.

This paper describes the application of three-layer perceptron networks to the identification and diagnosis of dangerous states in strongly exothermic semibatch reactions. To assess the potential danger of different faults, separate perceptron networks were used for danger assessment and for fault isolation. Results are presented to illustrate the performance of the neural-network approach on real process data obtained from typical faults of a catalytic esterification process which were simulated in a laboratory reactor.