Hazardous conditions associated with performing the Full-Length High- Temperature (FLHT). Severe Fuel Damage Test No. 2 experiment have been analyzed. Major hazards that could cause harm or damage are (1) radioactive fission products, (2) radiation fields, (3) reactivity changes, (4) hydrogen generation, (5) materials at high temperature, (6) steam explosion, and (7) steam pressure pulse. As a result of this analysis, it is concluded that with proper precautions the FLHT- 2 test can be safely conducted.

Battelle, Pacific Northwest Laboratory is conducting severe fuel damage experiments for the U.S. Nuclear Regulatory Commission to provide data that are being used to bench-mark or validate computer codes like SCDAP and MELPROG. These experiments are designed to study the early phase of a light water reactor (LWR) core uncovery accident from the time of initial uncovery (boil-away) until peak temperatures reach [approximately]2500K. A description of four in-reactor, full axial length, high-temperature (FLHT) LWR fuel bundle tests is given including the geometries and test conditions. The test conditions increase in severity from test to test by increasing the peak temperature attained in the fuel bundle and/or by increasing the operating period at the peak temperatures. Thus the degree of damage progresses from slight in the early test to extensive in later tests. Salient features of these tests include the use of twelve 3.7-m-long fresh and preirradiated fuel rods that are held in position by standard 17 [times] 17 (either Inconel or Zircaloy-4) grid spacers. Two of the fuel bundles were operated for brief periods at typical commercial power ratings just prior to the boil-away transient.

This report presents the final safety analysis for the preparation, conduct, and post-test discharge operation for the Full-Length High Temperature Experiment-5 (FLHT-5) to be conducted in the L-24 position of the National Research Universal (NRU) Reactor at Chalk River Nuclear Laboratories (CRNL), Ontario, Canada. The test is sponsored by an international group organized by the US Nuclear Regulatory Commission. The test is designed and conducted by staff from Pacific Northwest Laboratory with CRNL staff support. The test will study the consequences of loss-of-coolant and the progression of severe fuel damage.

This document presents an assessment of the severe accident phenomena observed from four Full-Length High-Temperature (FLHT) tests that were performed by the Pacific Northwest Laboratory (PNL) in the National Research Universal (NRU) reactor at Chalk River, Ontario, Canada. These tests were conducted for the US Nuclear Regulatory Commission (NRC) as part of the Severe Accident Research Program. The objectives of the test were to simulate conditions and provide information on the behavior of full-length fuel rods during hypothetical, small-break, loss-of-coolant severe accidents, in commercial light water reactors.

The COBRA-SFS and HYDRA thermal analysis computer codes were used to perform ''pretest'' or ''prelook'' calculations on the CASTOR-1C, REA-2023, CASTOR-V/21 and TN-24P spent fuel storage casks. HYDRA results were compared to 24 different test runs, while COBRA-SFS results were compared to 25 runs. These comparisons included tests with three different fill media, vacuum (low-pressure nitrogen), helium, and nitrogen, in both horizontal and vertical casks. The mean difference in peak clad temperature for of all comparisons was +10/sup 0/C for HYDRA and +3/sup 0/C for COBRA-SFS. Their respective standard deviations were +-10/sup 0/C and +-11/sup 0/C. These comparisons have demonstrated that even with the limited amount of experience in modeling the thermal characteristics of casks, both codes do an excellent job of predicting peak clad temperatures for a wide variety of cask configurations.

The study reported here is one of several efforts to evaluate and qualify the COBRA-SFS computer code for use in spent fuel storage system thermal analysis. The ability of COBRA-SFS to predict the thermal response of two single assembly spent fuel heat transfer tests was investigated through comparisons of predictions with experimental test data. From these comparisons, conclusions regarding the computational treatment of the physical phenomena occurring within a storage system can be made. This objective was successfully accomplished as reasonable agreement between predictions and data were obtained for the 21 individual test cases of the two experiments.

Large-particle hydrotransport flow phenomena are analyzed. These include particle suspension mechanisms, relative in situ slip between phases, and shear between separated solid/liquid flow regions. These phenomena are related to the mechanistic force balance modeling approach to predicting nominal and minimum horizontal pipeline transport conditions. Results of four series of tests in a 35 m-long, 0.165 m-diameter plexiglass pipeline are presented. These tests were designed to investigate isolated flow phenomena related to lower limits of operation of large-particle transport. They included measurements of the particle size dependence and flow geometry dependence on interfacial friction factors, initiation of plug flow motion, and plugging tendencies of flow through upward-turning elbows. Numerous flow phenomena and parameter dependencies are included in analysis of predictions made using the force balance modeling approach. Comparisons with data are made where possible and a step-by-step solution procedure is presented. Results show agreement with some data sets and disagreement with others. Most available data, however, are neither extensive enough nor accurate enough to form complete conclusions as to the minimum operating conditions predicted with the force balance modeling approach. Additional data from a systematic experimental program in a once-through pipeline system are necessary to further verify the modeling approach. A limited qualitative analysis of methods proposed to unplug large particle pipelines is presented. Methods assessed include pulsed air injection, vibration augmentation, friction reducing surfactant injection, and mechanical or hydraulic boring machinery. Advantages and disadvantages of each method are assessed qualitatively. Laboratory and field testing are required to determine applications.

Recent efforts in coarse-particle hydrotransport modeling include the development of a physically-based model that separates the particles into two components - suspended load and contact load. The analyses of the two components are coupled by interfacial shear resulting from the components' differing bulk velocities. In support of these modeling efforts, the interfacial friction factor for water flowing in a pipe over well-characterized coal beds was evaluated experimentally. Experiments were conducted in a 32.5-m long, 16.4-cm ID Plexiglas pipe. Coal samples of anti d/D = 0.05, 0.10, and 0.20 were placed inside the pipe at three different bed depths. Pressure taps at four axial locations measured the pressure drop as water flowed over the bed. This paper describes the apparatus, experimental and data-reduction procedures, and the results achieved. A correlation of interfacial friction factor in terms of average particle diameter suitable for modeling purposes is presented.

Interactive flow phenomena have been analyzed to support the mechanistic force balance modeling approach to large particle hydrotransport. Particular emphasis is given to analyzing particle saltation and suspension criteria and to physical modeling of interfacial shear and throughbed flow. The force balance modeling approach has previously been concluded to offer the best physical basis for scaleup calculation of large particle hydrotransport pressure drop and lower limits of operation. Improvements are still needed, however, in describing large particle suspension transition for d/D>0.05. Alternate approaches to modeling interfacial shear are discussed and results presented. Insufficient data are available to conclude which approach is most adequate. Throughbed flow was investigated experimentally and concluded to be negligible under certain conditions. Physical parameter dependency of the force balance model approach on sliding friction, loose packed bed fraction, and specific gravity are discussed as well as operating parameters such as delivered concentration, particle size, and pipe size.

This report describes an independent evaluation of the hydrodynamic data obtained with coarse coal in 3- and 6-in. recirculation loops. The measured data are compared with several recently published predictive methods to evaluate data quality and determine data trends. A limited parameterization of these models to investigate their sensitivity is also presented. In recirculating loop tests, velocity ranges of 2 through 15 ft/sec were used to pump solid concentrations of up to 45% by weight. Pressure drop data were obtained for particle-to-pipe diameter ranges to 0.25 under various flow conditions. The predictive methods used to analyze the hydrodynamic data obtained include mechanistic models and empirical correlations. For each method, predicted versus measured headloss gradients were plotted to evaluate the ability of each method to reproduce the parameter trends exhibited by the pressure drop data. Of the five methods selected for analyzing the data, three could be used to predict the head loss gradients with some certainty, including a sliding bed model that provided predicted results equal to a regression analysis of the original data. The original correlation of the data was found to be capable of reproducing 90% of the measured results to within +-30%. It was therefore concluded thatmore » the original correlation and the sliding bed model provided sufficient accuracy to be used as a means for estimating the headloss for coarse particle coal systems in 3- and 6-in. lines. The success of the predictive methods in reproducing the parameter trends indicate that the data obtained were reasonably accurate and reliable. It is is expected that mechanistic models can become a valuable design tool for predicting hydrotransport performance over ranges of parameters.« less