In der Reaktorsicherheitsforschung sind auftriebsgetriebene Strömungen von Relevanz für Störfall-szenarien mit Verdünnung der Borkonzentration und für thermische Schockbelastungen des Reak-tordruckbehälters. In der numerischen Simulation der Strömungen werden neben der Berücksichtigung der Auftriebskräfte Quell- und Korrekturterme in die Bilanzgleichungen für die turbulente Energie und die turbulente Dissipation eingeführt. Es wurden erweiterte Modelle entwickelt, in die zusätzliche Gleichungen für die Turbulenzgrößen turbulenter Massenstrom und Dichtevarianz eingehen. Die Modelle wurden in den CFD-Code ANSYS-CFX implementiert. Die Validierung der Modelle erfolgte an einem speziellen Versuchsaufbau (VeMix-Versuchsanlage), mit Einspeisung von Fluid höherer Dichte in eine Vorlage. Als Kriterien für die Validierung wurde der Umschlag zwischen impulsdominiertem Strömungsregime mit vertikalem Jet oder ein vertikales Absinken bei Dominanz von Dichteeffekten herangezogen sowie lokale Konzentrationsmessungen mit Hilfe eines speziell entwickelten Leitfähigkeits-Gittersensors. Eine Verbesserung der Simulation dichtedominierter Vermischungsprozesse mit den erweiterten Turbulenzmodellen konnte allerdings nicht nachgewiesen werden, da die Unterschiede zwischen den Rechnungen mit verschiedenen Turbulenzmodellen zu gering sind. Andererseits konnte jedoch die Simulation der Stratifikation von Fluiden unterschiedlicher Dichte im kalten Strang einer Reaktoranlage deutlich verbessert werden. Anhand der Nachrechnung von Ver-suchen am geometrisch ähnlichen Reaktor-Strömungsmodell ROCOM wurde gezeigt, dass diese Stratifikation von bedeutendem Einfluss auf die Vermischung und somit letztendlich auch auf die Temperatur- bzw. Borkonzentrationsverteilung innerhalb des Reaktordruckbehälters ist. Sie lässt sich nur korrekt simulieren, wenn ausreichend große Abschnitte des kalten Stranges mit modelliert werden. Somit konnte doch eine bessere Vorhersagegenauigkeit der Simulation der Vermischung erreicht werden. In reactor safety research, buoyancy driven flows are of relevance for boron dilution accidents or pressurised thermal shock scenarios. Concerning the numerical simulation of these flows, besides of the consideration of buoyancy forces, source and correction terms are introduced into the balance equations for the turbulent energy and its dissipation rate. Within the project, extended turbulence models have been developed by introducing additional balance equations for the turbulent quantities turbulent mass flow and density variance. The models have been implemented into the computati-onal fluid dynamics code ANSYS-CFX. The validation of the models was performed against tests at a special experimental set-up, the VeMix facility, were fluid of higher density was injected into a vertical test section filled with lighter fluid. As validation criteria the switching-over between a momentum controlled mixing pattern with a horizontal jet and buoyancy driven mixing with vertical sinking down of the heavier fluid was used. Additionally, measurement data gained from an especially developed conductivity wire mesh sensor were used. However, an improvement of the modelling of buoyancy driven mixing by use of the extended models could not be shown, because the differences between calculations with the different models were not relevant. On the other hand, the modelling of the stratification of fluids with different density in the cold leg of a reactor primary circuit could be significantly improved. It has been shown on calculations of experi-ments at the ROCOM mixing test facility, a scaled model of a real reactor plant, that this stratification is relevant as a boundary condition for the mixing process inside the reactor pressure vessel. It can be correctly simulated only if sufficient large parts of the cold legs are included in the modelling. On this way, an improvement of the accuracy of the prediction of mixing processes was achieved.

In der Reaktorsicherheitsforschung sind auftriebsgetriebene Strömungen von Relevanz für Störfall-szenarien mit Verdünnung der Borkonzentration und für thermische Schockbelastungen des Reak-tordruckbehälters. In der numerischen Simulation der Strömungen werden neben der Berücksichtigung der Auftriebskräfte Quell- und Korrekturterme in die Bilanzgleichungen für die turbulente Energie und die turbulente Dissipation eingeführt. Es wurden erweiterte Modelle entwickelt, in die zusätzliche Gleichungen für die Turbulenzgrößen turbulenter Massenstrom und Dichtevarianz eingehen. Die Modelle wurden in den CFD-Code ANSYS-CFX implementiert. Die Validierung der Modelle erfolgte an einem speziellen Versuchsaufbau (VeMix-Versuchsanlage), mit Einspeisung von Fluid höherer Dichte in eine Vorlage. Als Kriterien für die Validierung wurde der Umschlag zwischen impulsdominiertem Strömungsregime mit vertikalem Jet oder ein vertikales Absinken bei Dominanz von Dichteeffekten herangezogen sowie lokale Konzentrationsmessungen mit Hilfe eines speziell entwickelten Leitfähigkeits-Gittersensors. Eine Verbesserung der Simulation dichtedominierter Vermischungsprozesse mit den erweiterten Turbulenzmodellen konnte allerdings nicht nachgewiesen werden, da die Unterschiede zwischen den Rechnungen mit verschiedenen Turbulenzmodellen zu gering sind. Andererseits konnte jedoch die Simulation der Stratifikation von Fluiden unterschiedlicher Dichte im kalten Strang einer Reaktoranlage deutlich verbessert werden. Anhand der Nachrechnung von Ver-suchen am geometrisch ähnlichen Reaktor-Strömungsmodell ROCOM wurde gezeigt, dass diese Stratifikation von bedeutendem Einfluss auf die Vermischung und somit letztendlich auch auf die Temperatur- bzw. Borkonzentrationsverteilung innerhalb des Reaktordruckbehälters ist. Sie lässt sich nur korrekt simulieren, wenn ausreichend große Abschnitte des kalten Stranges mit modelliert werden. Somit konnte doch eine bessere Vorhersagegenauigkeit der Simulation der Vermischung erreicht werden. In reactor safety research, buoyancy driven flows are of relevance for boron dilution accidents or pressurised thermal shock scenarios. Concerning the numerical simulation of these flows, besides of the consideration of buoyancy forces, source and correction terms are introduced into the balance equations for the turbulent energy and its dissipation rate. Within the project, extended turbulence models have been developed by introducing additional balance equations for the turbulent quantities turbulent mass flow and density variance. The models have been implemented into the computati-onal fluid dynamics code ANSYS-CFX. The validation of the models was performed against tests at a special experimental set-up, the VeMix facility, were fluid of higher density was injected into a vertical test section filled with lighter fluid. As validation criteria the switching-over between a momentum controlled mixing pattern with a horizontal jet and buoyancy driven mixing with vertical sinking down of the heavier fluid was used. Additionally, measurement data gained from an especially developed conductivity wire mesh sensor were used. However, an improvement of the modelling of buoyancy driven mixing by use of the extended models could not be shown, because the differences between calculations with the different models were not relevant. On the other hand, the modelling of the stratification of fluids with different density in the cold leg of a reactor primary circuit could be significantly improved. It has been shown on calculations of experi-ments at the ROCOM mixing test facility, a scaled model of a real reactor plant, that this stratification is relevant as a boundary condition for the mixing process inside the reactor pressure vessel. It can be correctly simulated only if sufficient large parts of the cold legs are included in the modelling. On this way, an improvement of the accuracy of the prediction of mixing processes was achieved.

The goal of the work described in this report was the experimental investigation of the mixing of coolant with different quality (temperature, boron concentration) in nuclear reactors on the way from the cold leg through the downcomer and lower plenum to the core inlet in a systematic way. The obtained data were used for the clarification of the mixing mechanisms and form a data basis for the validation of computational fluid dynamics (CFD) codes. For these purposes, experiments on slug mixing have been performed at two test facilities, modelling different reactor types in scale 1:5, the Rossendorf and Vattenfall test facilities. The corresponding accident scenario is the start-up of first main coolant pump (MCP) after formation of a slug of lower borated water during the reflux-condenser mode phase of a small break loss of coolant accident (LOCA). The matrices for the experiments were elaborated on the basis of the key phenomena, being responsible for the coolant mixing during pump start-up. Slug mixing tests have also been performed at the VVER-1000 facility of EDO Gidropress to meet the specifics of this reactor type. The mixing of slugs of water of different quality is also very important for pre-stressed thermal shock (PTS) situations. In emergency core cooling (ECC) situations after a LOCA, cold ECC water is injected into the hot water in the cold leg and downcomer. Due to the large temperature differences, thermal shocks are induced at the reactor pressure vessel (RPV) wall. Temperature distributions near the wall and temperature gradients in time are important to be known for the assessment of thermal stresses. One of the important phenomena in connection with PTS is thermal stratification, a flow condition with a vertical temperature profile in a horizontal pipe. Due to the fluctuating character of the flow, this may cause thermal fatigue in the pipe. Besides of thermal fatigue, a single thermal shock can also be relevant for structural integrity, if it is large enough, especially in the case, that the brittle fracture temperature of the RPV material is reduced due to radiation embrittlement. Therefore, additional to the investigations of slug mixing during re-start of coolant circulation, the mixing of slugs or streams of water with higher density with the ambient fluid in the RPV was investigated. The aim of these investigations was to study the process of turbulent mixing under the influence of buoyancy forces caused by the temperature differences. Heat transfer to the wall and thermal conductivity in the wall material have not been considered. Experiments on density driven mixing were carried out at the Rossendorf and the Fortum PTS facilities.

The goal of the work described in this report was the experimental investigation of the mixing of coolant with different quality (temperature, boron concentration) in nuclear reactors on the way from the cold leg through the downcomer and lower plenum to the core inlet in a systematic way. The obtained data were used for the clarification of the mixing mechanisms and form a data basis for the validation of computational fluid dynamics (CFD) codes. For these purposes, experiments on slug mixing have been performed at two test facilities, modelling different reactor types in scale 1:5, the Rossendorf and Vattenfall test facilities. The corresponding accident scenario is the start-up of first main coolant pump (MCP) after formation of a slug of lower borated water during the reflux-condenser mode phase of a small break loss of coolant accident (LOCA). The matrices for the experiments were elaborated on the basis of the key phenomena, being responsible for the coolant mixing during pump start-up. Slug mixing tests have also been performed at the VVER-1000 facility of EDO Gidropress to meet the specifics of this reactor type. The mixing of slugs of water of different quality is also very important for pre-stressed thermal shock (PTS) situations. In emergency core cooling (ECC) situations after a LOCA, cold ECC water is injected into the hot water in the cold leg and downcomer. Due to the large temperature differences, thermal shocks are induced at the reactor pressure vessel (RPV) wall. Temperature distributions near the wall and temperature gradients in time are important to be known for the assessment of thermal stresses. One of the important phenomena in connection with PTS is thermal stratification, a flow condition with a vertical temperature profile in a horizontal pipe. Due to the fluctuating character of the flow, this may cause thermal fatigue in the pipe. Besides of thermal fatigue, a single thermal shock can also be relevant for structural integrity, if it is large enough, especially in the case, that the brittle fracture temperature of the RPV material is reduced due to radiation embrittlement. Therefore, additional to the investigations of slug mixing during re-start of coolant circulation, the mixing of slugs or streams of water with higher density with the ambient fluid in the RPV was investigated. The aim of these investigations was to study the process of turbulent mixing under the influence of buoyancy forces caused by the temperature differences. Heat transfer to the wall and thermal conductivity in the wall material have not been considered. Experiments on density driven mixing were carried out at the Rossendorf and the Fortum PTS facilities.