The LOBI Integral System Test Facility Experimental Programme

The LOBI project has been carried out in the framework of the European Commission Reactor Safety Research Programme in close collaboration with institutional and/or industrial research organizations of EC member countries. The primary objective of the research programme was the generation of an experimental data base for the assessment of the predictive capabilities of thermalhydraulic system codes used in pressurised water reactor safety analysis. Within this context, experiments have been conducted in the LOBI integral system test facility designed, constructed, and operated (1979–1991) at the Ispra Site of the Joint Research Centre. This paper provides a historical perspective and summarizes major achievements of the research programme which has represented an effective approach to international collaboration in the field of reactor safety research and development. Emphasis is also placed on knowledge management aspects of the acquired experimental data base and on related online open access/retrieval user functionalities.


Background
The LOBI project originated from a reactor safety research and development contract between the European Commission (EC) and the former Bundesminister für Forschung und Technologie (BMFT) of the Federal Republic of Germany.On the basis of contingent and perceived safety requirements, the BMFT decided in 1972 on the need of an integral system test facility for thermal-hydraulic investigations relevant to accident conditions in pressurized water reactors (PWRs) of German design.
A general international consensus of opinions emerged in the early 70s on the need to provide reliable methodologies for a realistic estimate of emergency-core-cooling system (ECCS) performance which was being questioned as large power reactors were being introduced.Due to the limited sophistication of safety codes and to the lack of relevant experimental data for assessing their predictive capabilities, sufficient conservatism was prescribed in the safety evaluations of design basis accidents (DBAs), such as loss-ofcoolant accidents (LOCAs), in order to account for worstcase uncertainties.This entailed stringent licensing requirements and undesirable operational constraints on nuclear power plants.
Within this context, reactor safety research and development programmes were formulated at the international level to improve the understanding and modelling capabilities of the thermal-hydraulic phenomenologies governing the course of a LOCA or of any other anticipated abnormal occurrence in water cooled reactors.Emphasis was placed mainly on deterministic methodologies supported, as appropriate, by probabilistic risk assessment studies with the aim to better understand accident progression and to substantiate the request for the eventual relaxation of over-conservatism in some safety analysis acceptance criteria.
The European Commission has been engaged in nuclear safety research activities since the signing in 1957 of the treaties establishing its precursors; that is, the European Atomic Energy Community and the European Economic Community.In line with its institutional role, the action of the Commission has been mainly devoted to a systematic effort in promoting the harmonization of safety practices and methodologies among the member states.
On the basis of its tender of May 1973, the Joint Research Centre was charged by the German BMFT with a R&D contract which envisaged the design, the construction, and the operation of an integral system test facility for the investigation of thermal-hydraulic phenomenologies pertinent to PWR large break LOCAs.The contractual agreement was signed in December 1973 and was then renegotiated in 1982 extending the original research objectives to the small break LOCA and Special Transients scenarios.
In its final configuration, the LOBI experimental data base consists of 70 tests (Table 2) covering a wide range of accident scenarios relevant to the safety analysis of PWRs.

Research Objectives
The LOBI research programme, as initially conceived, has been mainly oriented towards the generation of an experimental data base relevant to risk-dominant accidents and transients in PWRs.Specific research objectives included: (i) the identification and/or verification of basic phenomenologies governing the thermal-hydraulic response of an integral system test facility for a range of conditions relevant to LOCAs and special transients in PWRs, (ii) the generation of an experimental data base for the independent assessment of the predictive capabilities of large thermal-hydraulic system codes used in water reactor safety analysis.
The experimental programme has been supported by comprehensive code application and assessment activities.ATH-LET (DRUFAN), CATHARE, RELAP4, RELAP5, and TRAC have been largely used either within JRC or by outside organizations for test design and test prediction calculations.Development and application of advanced two-phase flow measurement techniques to support the execution of the experimental programme have constituted an integral part of the overall research strategy.At the time, a considerable effort has also been devoted to the development of an IBM version of the RELAP5 code which, together with various model improvements introduced at the JRC, has been instrumental in enabling the code calculation capabilities of many organisations within and outside the EC.The most stringent concern in the operation of nuclear power plants is to ensure that its inventory of radioactive material remains safely confined at all times, both during normal and off-normal conditions.In order to satisfy this compelling requirement, nuclear power plants have to comply with a set of design and construction standards, operational constraints, and safety regulations.
The many different safety features which are engineered in each nuclear power plant to compensate for any departure from normal operating conditions should an accident occur, must, irrespective of the cause, provide sufficient safety margins to prevent the release of radioactive material in excess of prescribed limits, and ensure the protection of the operators and the general public.
The emergency core cooling system (ECCS) is the most important safety feature of a water-cooled reactor.Its main purpose is to maintain the reactor in a coolable geometry in the event of loss of coolant caused by a break in the main cooling system or by the operation of the safety relief systems in the case of intact circuit faults.To meet this widely accepted requirement, the ECCS should have the capabilities to (i) provide sufficient core cooling by removing both residual stored energy and decay heat at a rate such to limit the maximum fuel cladding temperature to a value less than the prescribed limit thereby preventing the cladding from loosing its structural integrity, (ii) have sufficient capacity, diversity, reliability, and redundancy to provide core cooling under all conceived accident conditions.
The ECCS safety features of a pressurised water reactor consist of a set of independent subsystems which include the high pressure injection system (HPIS), the accumulator injection system (AIS), and the low pressure injection system (LPIS); each subsystem is characterised by redundancy of equipment and flow paths.At present, PWR safety analysis is based on the evaluation of ECCS performance during design basis accidents (DBAs) which include large break LOCAs, small break LOCAs, and special transients resulting from intact circuit faults.

The Experimental Installation
The LOBI test facility was designed to represent a fullpower high-pressure integral system test facility 1 : 700 scale model of a 4-loop, 1300 MWe PWR, Figure 1.It incorporates the essential features of a typical PWR primary and secondary cooling system.The test facility was commissioned in December 1979 and was operated until June 1982 in the MOD1 configuration for the investigation of large break LOCAs; it was then extensively modified into the MOD2 configuration which incorporates design and instrumentation features best suited for the characterization of phenomenologies relevant to small break LOCAs and special transients.
The measurement system comprises a total of about 470 measurement channels.It allows the measurement of all relevant thermohydraulic quantities at the boundaries (inlet and outlet) of each individual loop component and within the reactor pressure vessel model and steam generators.A process control system allows the simulation of time-or pressure-dependent parameters such as main coolant pump hydraulic behaviour, core decay heat release, and safety injection flow rates.A fast running data acquisition system complements the experimental installation.

Mechanical Components.
The test facility comprises two primary loops, the intact and the broken loop which represents, respectively, three loops and one loop of the reference PWR.Each primary loop contains a main coolant circulation pump and a steam generator.The simulated core consists of an electrically heated rod bundle arranged in a square matrix inside the pressure vessel model.The primary cooling system which is shown schematically in Figure 2 operates at normal PWR conditions, approximately, 158 bar and 294-326 • C pressure and temperature, respectively.Heat is removed from the primary loops by the secondary cooling system which contains a condenser and a cooler, the main feedwater pump, and the auxiliary feedwater system.Normal operating conditions of the secondary cooling system are 210 • C feedwater temperature and 64.5 bar pressure.an annular downcomer, and an externally mounted upper head simulator are additional major components of the overall pressure vessel assembly.The reactor core is simulated by an electrically heated rod bundle consisting of 64 rods arranged in an 8 × 8 square matrix inside the flow shroud; heater rod bundle dimensions are reactor typical.The heater rods are directly heated hollow tubes (material 1.4948 DIN 17007), and the rod wall thickness within the heated length is varied in 5 steps to achieve a chopped cosine-shaped axial power distribution.
The upper part of the rod bundle which extends entirely into the upper plenum is formed by hollow nickel tubes connected to the upper power plate.The lower part is formed by nickel rods and flexible nickel braids which extend partially into the lower plenum where they connect to the lower-power connecting ring.The heat dissipated within these "unheated" regions amounts to about 14%.
Nine grid spacers of original design are placed along the heated length; five additional spacers are mounted in the upper unheated part of the rod bundle.Ceramic segments are arranged inside the core barrel tube forming a square flow shroud which extends over the heated length region of the rod bundle.
The upper head is simulated by an external vessel connected to the upper plenum and to the upper downcomer.Volume as well as height and relative elevations of the reference plant upper head are preserved.In the initial MOD1 version of the test facility, the annular downcomer formed by the pressure vessel and the core barrel tube had a gap width of 50 mm which was later decreased to 12 mm to better represent fluid volume distribution.

Steam Generators.
The LOBI test facility contains two shell and inverted U-tube type steam generators having a geometrical configuration similar to that of the reference plant.In the MOD1 configuration, the steam generators were designed to preserve heat transfer capabilities without proper simulation of secondary side fluid distribution.In the MOD2 configuration, the steam generators were designed with the aim to better represent thermal-hydraulic phenomenologies of interest in intact circuit faults.
The overall scaling ratio which required a capacity ratio of 3 : 1 between the intact and the broken loop steam generator led to a heat transfer exchange power of 3.96 MW (24 U-tubes + 1 installed spare) and 1.32 MW (8 U-tubes + 1 installed spare) for the intact and broken loop SG, respectively.
Each steam generator consists of a single cylindrical pressure vessel with an annular downcomer separated from the riser region by a skirt tube.This tube is supported above the tube plate and carries the coarse separator; a fine separator is arranged in the uppermost part of the steam dome.The U-tubes are arranged in a circle within the riser region around an axially mounted filler tube, with the U-bends crossing over one another above it.This design permits cross-flow between cocurrent and counter-current legs of the U-tubes over their entire length and heat and mass transfer between riser and downcomer to account for the recirculation characteristics of the prototypical system.An adjustable throttle device is installed at the lower end of the downcomer to properly set the recirculation ratios.
Feedwater is directed into the downcomer by a "J-nozzle" feed ring sparger and flows downward mixed with the recirculation water returned by the coarse and fine separators.The steam water mixture leaving the bundle region flows upward into the coarse separator where the moisture is partially removed by centrifugal separation and returned to the downcomer.Additional separation is attained in the fine, box-grid type separator from where saturated, practically dry steam flows into the outlet nozzle.On the primary side, fluid enters the U-tubes through an inlet chamber and flows first upward and then downward exchanging heat with the secondary fluid.

Main Coolant Pumps.
The main coolant pumps of both loops are centrifugal type pumps having a specific speed of 29.2 (DIN).The two pumps are equal in size and are, therefore, operated at two different speeds such as to yield the two different steady-state mass flows of 21 Kg/s and 7 Kg/s for the intact and broken loop at the same pressure head.A special control and drive system allows variation of the pump speed in the forward and backward directions over a range of ±8500 rpm.Since the locked rotor resistance of the MCPs is less than the equivalent reactor pump, there are provisions for the insertion of an additional flow resistance at the outlet of each pump.To account for the injection of the main coolant pump seal water, a closed loop seal water compensation system is installed.

Pressurizer.
The pressurizer design is geometrically similar to that of the reference plant; however, it is scaled in volume but not in height.The surgeline rises within the pressurizer and leaves it radially.The pressurizer is provided with normal and additional heaters; the spray system is simulated with cooling coils placed in the steam region.There are provisions for connecting the surgeline to either the intact or broken loop hot legs.The simulation of poweroperated relief valves (PORVs) as well as safety relief valves (SRVs) is provided in the pressurizer relief line.

Primary Loop Pipework.
The main coolant pipes connecting the major primary loop components have inner diameters of 73.7 mm and 46.1 mm for the intact and broken loop, respectively.Measurement inserts are installed at the inlet and outlet of each major component forming integral part of the main pipework.Since the main coolant pumps are equal in size with an inlet-outlet diameter of 65 mm, special cross-sections adapters are installed at the inlet-outlet of each pump to fit the main loop pipes.
3.1.6.Break Assembly.The break assembly consists of a Tshaped insert with the break orifice housed in a recess machined in the insert.The break assembly can be connected to the main coolant pipe at the selected break location; for example, cold leg, hot leg, or pump seal.Pressurizer breaks or inadvertent opening of the valves are simulated by an orifice inserted in the relief line.A proper connection can be established between primary and secondary systems at the tube plate elevation to simulate steam generator tube rupture (SGTR) sequences.

Safety Injection Systems.
The LOBI-MOD2 emergencycore-cooling system (ECCS) comprises the high-pressure injection system (HPIS) and the accumulator injection system (AIS).As required, the low-pressure injection system (LPIS) could also be simulated.Provisions are made for cold leg, hot leg, or combined cold and hot leg ECC injection in both primary loops.In the MOD1 version of the test facility, only the accumulator system was simulated.
The HPIS water is supplied by a positive displacement pump driven by a variable speed motor.The pump is rated for a maximum flow of 0.39 Kg/s at a total head of 200 bars.
A special speed control system provides appropriate flow regulation to match the reactor HPIS-expected performance.
Properly designed and calibrated throttling devices are installed in the main injection line to provide the required partitioning (depending on particular simulation needs) of the injection rate between the broken and the intact loop.
The AIS is composed of two accumulators, one in each loop.The accumulator of the intact loop has three times the volume and water capacity of that of the broken loop.The total volume of each accumulator is scaled to that of the reference plant having one accumulator for each loop; gas space, water volume as well as height and elevations are preserved.Both accumulators are rated for a maximum pressure of 60 bar and a temperature of 50 • C.
Additional safety injection systems consist of the volume control system (VCS) and of the auxiliary feedwater system (AFWS).The VCS consists of a feed pump and a water preheating system which allows the control of the injected water at the prevailing cold leg fluid temperature.Similarly, the AFWS consists of a feed pump an a water preheating system as the injected water is preheated to generally about 130 • C to prevent thermal shocks in the SG feed line and J-nozzle feed ring.

Measurement System.
The LOBI test facility in both the MOD1 and MOD2 configurations has been fitted with a comprehensive measurement system.Relevant thermalhydraulic quantities were measured at the boundaries of each major component and within the reactor pressure vessel model and the steam generators.
Measurement in the primary loop pipework is performed within the measurement inserts at the inlet and outlet of each major component, that is, the reactor pressure vessel model, the steam generators, and the main coolant pumps.The inserts in the horizontal pipework are properly instrumented in the lower and upper part of the flow crosssection to characterize eventual flow and thermal stratification phenomena.Fluid and wall temperatures, absolute and differential pressures, fluid velocities and density as well as flow direction indicators are generally provided at each measurement insert.
Measurement of temperatures, pressures, and differential pressures are extensively made along the downcomer, lower plenum, rode bundle section, and upper plenum flow paths.Fluid velocity and fluid density are measured at the rod bundle inlet box.
Each rod bundle is supplied with three thermocouples brazed into grooves of 0.8 mm depth and 10 mm length machined into the outer surface of the heater rod tubes and then led through the wall to the inside of the tubes; they leave the rods through the open upper end.
The LOBI-MOD2 steam generators are instrumented to provide a maximum of information on both the magnitude and location of the heat transfer process taking place between the primary and the secondary systems.In particular, the instrumentation is concentrated in the region of the lowest U-bend and immediately above the tube plate in order to detect changes in heat transfer regime.
Steam generators measurements include fluid temperatures, U-tube wall temperatures, and differential pressures on both the primary and secondary side.Primary side instrumentation is applied to two representative U-tubes in each steam generator; the highest U-tube is fitted with temperature sensors whereas differential pressures are measured in the shortest U-tube.Special Pitot tubes are installed in the downcomer of the steam generator secondary side to measure fluid velocities at three peripheral positions.
Measurements in the secondary system is concentrated in the feedwater and in the steam lines at the inlet and outlet of each steam generator.Feedwater line measurement includes fluid temperature and velocity whereas in the steam line fluid temperature and volumetric flow is measured.A special spool piece with low-density measurement capability may be mounted at the outlet of the broken loop steam generator for detecting eventual carry-over in those tests involving blowdown of the secondary side.The pressurizer and the surgeline are provided with fluid temperature and differential pressure measurements and a full-flow turbine.
All ECC injection lines are provided with full-flow turbines and fluid temperature measurements.Mass flow measurement is also provided in the main coolant pump seal water injection lines.The break flow measurement system consists of a condenser-catch tank system.Initially, the high energy flow is condensed in the cold liquid pool contained in the catch tank; thereafter, long-term low-flow energy removal is ensured by a small 250 KW condenser.The break flow in its integral form is obtained by the incremental mass of the catch tank.
A special tailored data acquisition system based on state of art information technologies available at that time is used to record all measurement signals; the sampling rate can be varied up to 20000 samples/s as appropriate for the representation of the experimental information especially in fast evolving transients.Selected data are then read back from the analog tape, converted into engineering units and stored on 300 Mbyte disk modules ready for normalization and correction.
The data acquired at very high sampling rate are discretised and reduced to about 1000 data points over the entire experiment time range; each reduced value is simply an average over the time interval between a discrete set of data points.This is required in order to have a manageable set of data for storage and further processing and analysis.The reduced data in its final form are stored on the corresponding normal-time experimental data tape.As appropriate, special selects are made for the analysis of experimental time intervals characterized by fast transients and stored on a related short-time data tape.A real-time loop control and monitoring system provides defined set of time ordered operations to represent components response during the simulated transient.Control of main coolant pump speed, core heating power, high-pressure injection system flow rate and steam generator level is performed on a real-time basis in response to predefined trip signals.

Scaling Criteria
The LOBI test facility was scaled to preserve, insofar as possible or practical, similarity of thermal-hydraulic behaviour As general scaling principle, a power-to-volume scaling criterion was adopted in the design of the facility to ensure the preservation of the specific power input into the primary fluid, Table 1.To meet general scaling requirements, the test facility was designed to preserve, taking into account the selected 1 : 700 power to volume scaling ratio, the following main parameters: (i) core power to system volume ratio, (ii) volumes and relative volumes of individual components, (iii) rupture size to primary system volume ratio, (iv) pressure drop and temperature distribution along main flow paths, (v) height and elevation of major components, and (vi) core and steam generators heat transfer surfaces.
The elevation of the major components were maintained at full height with the exception of the pressurizer which, while preserving the total volume and the steam to liquid volume ratios, was somehow shortened to allow increased radial dimensions to accommodate the internal heaters.The core and steam generators heat transfer and flow areas were matched to the scale factor.Strict adherence to the power-to-volume scale factor would have resulted in unacceptably high wall frictional pressure losses in the primary loop pipework which was appropriately shortened to increase the pipe diameter in order to match the expected pressure drop in the reference plant.
In the MOD2 configuration of the test facility, special emphasis was given to the scaling of the steam generator primary and secondary sides due to their importance on the thermal-hydraulic evolution of small break LOCAs and special transients.In particular, volume ratio, heat transfer surface-to-volume ratio, hydraulic resistances and elevations, and height of the shortest U-tube were preserved.
A major exception to the general scaling concept is the design of the reactor pressure vessel model annular downcomer.The test facility has been configured with a downcomer of two different gap widths.Initially, a downcomer gap of 50 mm was installed in the MOD1 configuration to prevent ECC bypass which is largely influenced by hot wall delay and counter-current flow limitation phenomena; this, however, resulted in a 6.3 times too large a downcomer volume and, as a consequence, in an atypical thermal-hydraulic system response during large break LOCAs.The downcomer gap width was later changed to 12 mm which again, was a technical compromise between a 7 mm volume-scaled and a 25 mm pressure drop-scaled downcomer.
The LOBI test facility, as any other scaled test facility, has inherent distortions with respect to the reference plant which may impair the typicality of some results.The power-tovolume scaling concept results in a design which exhibits a basically one-dimensional thermal-hydraulic response, components high surface area to fluid volume ratio, and large metal mass to fluid volume ratio.The structural stored energy and system heat losses are important contributors to distortions in those components, such as the reactor pressure vessel and steam generator downcomers, where the coupling between wall heat transfer and fluid flow is at time dominant.System heat losses may significantly influence primary as well as secondary side energy removal especially during the long-term phase of a small break LOCA or intact circuit fault simulations.The LOBI test facility exhibits larger heat losses relatively to the reference plant due to design (higher surface area to volume ratio of fluid retaining components) and operation constraints (main coolant pump seal and instrument cooling); typically, heat losses in a full size plant account for about 0.05% of the nominal thermal power whereas in the LOBI test facility it accounts for about 1.5%.
All in all, the experimental results acquired in the LOBI test facility cannot be directly extrapolated to full-size plants; they provide, however, a reference data base for the understanding of governing thermal-hydraulic phenomenologies and for the assessment of the predictive capabilities of system codes used in water reactor safety analysis.

Project Evolution
The LOBI project has evolved over a time period which has seen a very intensive international effort in reactor safety research and development.In the early '70 s, the level of understanding of LOCA phenomenologies and the attendant computational capabilities were rather primitive.A very limited experimental data base from integral system tests conducted in an initial, very crude configuration of the USNRC semiscale test facility and early one-dimensional versions of system codes such as RELAP were available.
Since then, a very large experimental data base has been gathered in several integral system test facilities, and the computational capabilities have matured encompassing a number of best-estimate codes such as RELAP, TRAC, ATHLET, and CATHARE.In the meantime, two major reactor accidents occurred in TMI-2 (March 1979) and Chernobyl (April 1986) power plants which have had a significant impact on the deployment of nuclear reactor technologies.

Execution of the MOD1 Programme.
The LOBI test facility became fully operational in December 1979 with the execution of the first 200% cold leg break LOCA experiment which was used for an international pretest prediction exercise (PREX).Early in the programme and in response to the TMI-2 accident which occurred in March 1979, new research priorities were formulated to emphasize small break LOCA and special transients tests in the follow-up programme.
After the execution of an initial test series and due to experimental evidences on the atypical influence of the large downcomer (and hence large vessel water inventory) on the system thermal-hydraulic response, the original 50 mm gap width downcomer was replaced with a downcomer having a smaller 12 mm gap width (and hence better scaled vessel water inventory), Figures 3 and 4. In the meantime, while the small downcomer was being procured, an interim test programme was carried out to assess test reproducibility, break geometry, and size effects on the course of a large break LOCA.
The experimental programme with the small 12 mm downcomer was initiated in March 1981 with test A1-66, a 200% cold leg break LOCA with cold leg ECC injection.The LOBI-MOD1 experimental programme was then concluded in June 1982.From December 1979 to June 1982, 28 experiments were carried out including 25 large break LOCA and 3 small break LOCA tests.

Execution of the MOD2 Programme.
After extensive modifications to the test facility, the experimental programme was resumed in April 1984 with the facility in the MOD2 configuration.The first small break LOCA test, a 1% cold leg break LOCA which was used for the OECD-CSNI International Standard Problem 18, was performed in  As the LOBI-MOD2 experimental programme was evolving, the catastrophic Chernobyl accident took place in April 1986; this, however, due to the peculiarities of the RBMK reactor type accident had no significant impact on the LOBIestablished research priorities.
The last experiment of the BMFT contractual programme was executed in November 1989 with the termination of the CEC-BMFT contract in December 1989.Thereafter, the experimental programme was exclusively dedicated to the execution of tests from the Community matrix.With the execution in June 1991 of test BL-06, a 1% cold leg break LOCA designed to address the pump on-off issue, the execution of the envisaged LOBI experimental programme was successfully concluded.Significant signatures of small break LOCA tests and special transients are depicted in following Figures 5, 6, 7, and 8.

International Collaboration.
The international context in which the LOBI research programme has been carried out has offered an opportunity for a close collaboration among delegates of national research laboratories.It has also provided an independent forum for the exchange of concerns and expertise among the participants contributing thus to the harmonization of national views on reactor safety-related matters.

Counter Part Test Programme.
Large system codes used in reactor safety analysis are generally benchmarked against experimental data from scaled integral system or separate effect test facilities; comparison of code-predictive capabilities of simulated accidents or transients in full-size plants would be desirable, but this is, clearly, prohibitive for obvious economic and practical considerations.Controversy, thus, arises when the predictive capability of a system code is scaled up from a small size test facility to the full-size real plant.
It is, therefore, desirable although not strictly required to assess code predictive capabilities against a set of data obtained from different scale test facilities under similar initial and boundary conditions.This, to a certain extent, would decouple the assessment process from physical assumptions emphasizing, instead, the relevance of the geometrical scaling parameters especially on the qualitative rather than quantitative evolution of the thermal-hydraulic phenomenologies.
Within this context, a number of tests of the MOD1 and MOD2 experimental programmes were defined and executed as counter part to similar tests performed in other test facilities such as Semiscale, PKL, BETHSY, LSTF, and SPES.The collaboration has also been extended to posttest analysis of the results.

International Standard Problems Prediction Exercises.
The very first test of the LOBI-MOD1 experimental programme, test A1-04, was used for a special type of blind standard problem exercises, the LOBI preprediction exercise (PREX); 16 participants from various EC member states and the USA submitted calculations using a number of system codes.
The first small break LOCA test of the LOBI-MOD2 experimental programme, test A2-81, was designated by the Committee on the Safety of Nuclear Installations (CSNI) of the Organization for Economic Cooperation and Development (OECD) as International Standard problem 18 (ISP-18); 27 participants from European and North American organizations provided prediction calculations with 12 codes or code versions.

Management of the Experimental Data Base
Management of research data is an issue being debated at the national and international levels by industrial, institutional, professional, and academic research organisations; specifically, preservation and archive as well as access and retrieval of research data are seen as underlying principles of good scientific practices especially in the governance of publicly funded research programmes.
The European Science Foundation (ESF) has addressed the issue of data accumulation, handling and storage in its policy briefing on Good Scientific Practices In Research and Scholarship [1] recommending: data are produced at all stages in experimental research and scholarship.Data sets are an important resource, which enable later verification of scientific interpretation and verification.They may also be the starting point for further studies.It is vital, therefore, that all primary and secondary data are stored in a secure and accessible form.
Archival and dissemination of research data has also been addressed inert alias in the Berlin Declaration [2] promoted by the Max-Planck Gesellschaft (MPG) together with representatives from several international research organisations who have recommended: our mission of disseminating knowledge is only half complete if the information is not widely and readily available to society.New possibilities of knowledge dissemination not only through the classical form but also and increasingly through the open access paradigm via the Internet have to be supported.
Since the inception of the LOBI project, it was anticipated that a proper data analysis and documentation management system would be of fundamental importance in order to ensure that over time the investment of public resources would be beneficial for the nuclear community at large.In view of the continuous advancement of computer hardware and software technologies that are making storage/retrieval techniques rapidly obsolete, particular attention has been placed on conversion of old format onto new media and machines enabling potential users to access and retrieve the data.Currently, the LOBI analytical and experimental data base can be accessed via Internet at http://asa2.jrc.it/.Each experimental data set comprises the digital data file, the Quick look Report (QLR), the Experimental Data Report (EDR) and as appropriate the Test Prediction Report (TPR).As appropriate AVI or MPEG files showing temporal evolution of collapsed liquid levels are also available.The overall data base is complemented by Test Facility Description Reports and components as built drawings (Figure 9).

Conclusions
The LOBI Project has represented an important contribution to reactors thermal-hydraulic safety research.A comprehensive data base relevant to the understanding of governing phenomenologies expected in PWR accident conditions and to the assessment of system codes used in water reactors safety analysis has been provided.As structured, the LOBI project has represented an effective approach to international collaboration in the field of reactor safety research and development.In addition, the EC-JRC context, in which the research programme has been carried out, has provided an independent forum for a systematic exchange of technical and scientific information among experts from EC member states and for sharing best practices in the field of watercooled reactor safety analysis.

Figure 1 :Figure 2 :
Figure 1: Photographic view of the LOBI test facility.

Figure 4 :
Figure 4: Primary system pressure and heater rod temperature response to a large 200% cold leg break LOCA (downcomer gap width 12 mm).

Figure 5 :
Figure 5: Heater rod temperature and vessel-collapsed liquid level in a small 10% cold leg break LOCA.

Figure 7 :
Figure 6: Display of collapsed liquid levels, heater rod temperature, and system pressures at 477 s into a 6% SBLOCA.

Figure 8 :
Figure 8: Primary and secondary system pressures in a 0.4% steam generator tube rupture.

Figure 9 :
Figure 9: Typical STRESA data set for each LOBI test.

Table 1 :
LOBI test facility design and operational parameters.