Skip to main content

Über dieses Buch

Welcome to Bavaria - Germany and to the INTERNATIONAL NUCLEAR SIMULATION SYMPOSIUM AND MATHEMATICAL MODELLING WORKSHOP. A triennial international conference jointly promoted by Control.Data, GRS and SCS, which takes place at Schliersee, a small town near the Alps. The aim of the Symposium is to cover most of the aspects of nuclear modelling and simulation in theory and practice, to promote the exchange of knowledge and experience between different international research groups in this field, and to strengthen the international contact between developers and users of modelling and simulation techniques. On the occasion of the Symposium people of scientific and engineering disciplines will meet to discuss the state-of-the-art and future activities and developments. A large number of contributed papers has been strictly examined and selected by the papers committee to guarantee a high international standard. The book contains the accepted papers which will be presented at the Symposium. The papers have been classified according to the following topics: 1. HARDWARE TOOLS 2. SIMULATION-SOFTWARE-TOOLS 3. PLANT ANALYSER 4. REACTOR CORE 5. NUCLEAR WASTE Authors from 9 countries will meet at the Symposium. They work for Industrial Companies, Universities and the Research and Development Institutes so that a broad spectrum of simulation activities is covered: Theory and application, hardware and software, research and operations. The editor is greatful to the authors for making possible the publication of this book, and especially to WOLFGANG F. WERNEB, for the selection of the papers and the contribution to the success of the Symposium.



Introduction Simulation “In the Core”

Before the widespread use of the digital machines, relationships and activities in the scientific-engineering approach were largely determined by the human mind, its formalization and analytic powers. The scientist or engineer who approaches a real world process tries to gain insight or an understanding of the phenomena on the process under study. Today after Chernobyl, which exploded in the world’s worst nuclear accident on 26 April 1986. The desire to prove that the errors which led to it will never be repeated is understandable, but the main impression left by the shuttered villages and abandoned fields in which bushes are begining to sprout is the length of the time necessary to eradicate the consequences of the original accident.
M. R. Heller

The Role of Symbolic Processing in Supercomputing

Most of you are dealing with what I would call a “megaproblem” — a problem that requires the highest performance computer available. Such problems occur mainly in applications where physical phenomena are simulated, such as weather, earthquakes, and ocean currents. Such problems also arise in the design of systems that depend on physical processes but where experimental study is uneconomical, difficult or hazardous, such as the behavior of airfoils in a wind stream, the vibration of buildings and other structures, and the detonation of nuclear weapons. In addition to simulations of physical phenomena, the simulation of economic activities, such as with linear programming, also can use high performance computers. Cryptography is another megaproblem area. Control Data has traditionally provided the tools to solve such problems—both hardware and software.
Robert M. White

The ETA10 Supercomputer System

The ETA Systems, Incorporated ETA10® is a next–generation supercomputer featuring multiprocessing, a large hierarchical memory system, high performance input/output, and network support for both batch and interactive processing. Advanced technology used in the ETA10 includes liquid nitrogen cooled CMOS logic with 20,000 gates per chip, a single printed circuit board for each CPU, and high density static and dynamic MOS memory chips. Software for the ETA10 includes an underlying kernel that supports multiple user environments, a new ETA FORTRAN compiler with an advanced automatic vectorizer, a multitasking library and debugging tools. Possible developments for future supercomputers from ETA Systems are discussed.
Charles D. Swanson

Mathematical Problems in the Simulation of Reactor Plants

The paper gives an overview of the mathematical problems encountered in the simulation of reactor plants, particularly with light water reactors. Major problem areas are:
Numerical methods for the solution of partial differential equations, describing the process models.
Several topics are addressed:
  • ways of spatial discretisation which lead to efficient solution algorithms
  • efficient time integration
  • coupling of different types of model equations
  • algorithms suitable for use on multi-parallel processors.
Principals and methods for the quantification of uncertainties which influence the results of reactor simulation calculations. Several of the commonly used methods are summarized. Criterions for their use are discussed.
W. F. Werner

Nuclear Power Plant Transient Analysis

Plant Data Compared to Simulation Results
For the 1308 MWe NPP Müilheim-Kärlich, with a two loop nuclear steam supply system and once-through steam generators a computer code for the analysis of operational transients and control system optimization has been developed. After completion of startup testing a comparison of code results against field data was made. The comparison shows good agreement and demonstrates that the scope of the simulation and the mathematical models meet the requirements of a best estimate analysis tool for operational transients.
G. Breiling

Development of an Advanced Thermal Hydraulics Model for Nuclear Power Plant Simulation

This paper summarizes the development of an advanced digital computer thermal hydraulics model for nuclear power plant simulation which has been carried out at CAE Electronics Ltd. A review of thermal hydraulics code design options is presented together with a review of existing engineering models. CAE has developed an unequal temperatures-unequal velocities five equation model based on the drift flux formalism.
CAE has selected the model on the basis that phase separation and thermal non-equilibrium are required to simulate complex and important phenomena occurring in systems such as reactor cooling systems (RCS) and steam generators (SG) of nuclear power plants. The drift flux approach to phase separation and countercurrent flow was selected because extensive testing and validation data supports full-range drift flux parameters correlations [1]. The five equation model was also chosen because it conserves important quantities, i.e. mass and energy of each phase, and because of numerical advantages provided by the ease of coupling phasic mass conservation equations with phasic energy conservation equations.
The basis of CAE’s five equation thermal hydraulics model as well as supporting models for convection and conduction heat transfer, break flow, interphase mass and heat transfer are described. Comparison of code calculations with experimental measurements taken during a small break LOCA test with the OTIS facility are presented.
The use of such advanced thermal hydraulics model as plant analyzer considerably improves simulation capabilities of severe transient as well as of normal operation of two phase systems in nuclear power plants.
Richard Moffett

Finite Element Analysis in Computational Fluid Mechanics

This paper highlights and illustrates the fundamental aspects that the finite element discrete approximation method has in development of key requirements for CFD algorithms for fluid mechanics conservation law systems The fundamental decision of trial and test space leads in a natural way to asymptotic error estimates that determine accuracy improvement rate with the choice. The issue of stability, hence dissipation and dispersion error mechanisms, is developed using a Taylor series extension. The choice of time-integration procedure is fundamental and analyzed. Comparisons to alternative theories are made, and representative numerical solutions are cited to add substance to the theoretical developments.
A. J. Baker

Prediction of Fluid Behaviour during Reactor Transient Analysis using Coupled 1D and 3D Models

The detailed understanding of the behaviour of a pressurised water reactor during transient operation is crucial to the design of the reactor.
Hitherto mathematical simulation effort has largely concentrated on one-dimensional loop analysis of the reactor system: this gives a reasonable understanding of the consequences of a transient, such as the blowdown. but does not yield adequate information on features such as pressure effects in the reactor vessel. For greater understanding of such effects, a three-dimensional reactor model is required.
Two- and three-dimensional transient simulations of reactor vessels and of individual components of the circuit have, in the past, been performed in order to provide pressure and temperature fields for input to structural analysis programs. Such decoupled calculations have however tended to underestimate the real situation, since boundary conditions have been specified from stand-alone loop calculations; furthermore, there has been no simple way of incorporating fluid-structure interaction effects.
This paper describes the application of the PHOENICS program to predict the fluid behaviour in a PWR during a hypothetical blowdown. The analysis features two technical novelties. The first is the use of a three-dimensional model for the reactor vessel, directly linked with one-dimensional loop models for the primary water circuits; the second is the dynamic coupling of a fluids code (PHOENICS) with a stress code (ABAQUS) to provide a first step towards the modelling of fluid-structure interaction effects.
D. Kirkcaldy, P. J. Phelps, N. Rhodes

ENEA Nuclear Power Plant Engineering Simulator: Mathematical Models, Perfomances, Opportunities of Usage

Enea, the Italian Commission for Nuclear and Alternative Energy Sources, has acquired an Engineering Tool. For Design and Analysis of Pwr Nuclear Power Plants: an Engineering Simulator, which reproduces in real time the behaviour of the whole plant, presenting all informations available to the operator in Control Room in a compact way.
G. M. Mancini, A. Mattucci

The State of Mathematical Modeling for Power Plant Training Simulators

This paper provides a survey of modeling technology used in power plant training simulators, identifying the capability of existing models, and commenting on the areas where modeling advances are expected to occur.
L. R. Foulke

Concepts and Realization of the KWU Nuclear Plant Analyzer

The Nuclear Plant Analyzer (NPA) is a real time simulator developed from KWU computer programs for transient and safety analysis (“engineering simulator”). The NPA has no control room, the hardware consists only of commercially available data processing devices.
The KWU NPA makes available all simulator operating features such as initial conditions, free operator action and multiple malfunctions as well as freeze, snapshot, backtrack and playback, which have evolved useful training support in training simulators of all technical disciplines.
The simulation program itself is running on a large mainframe computer Control Data CYBER 176 or CYBER 990 in the KWU computing center under the interactive component INTERCOM of the operating system NOS/BE. It transmits the time dependent engineering data roughly once a second to a process computer SIEMENS 300-R30E using telecommunication by telephone. The computers are coupled by an emulation of the communication protocol Mode 4A, running on the R30 computer. To this emulation a program-to-program interface via a circular buffer on the R30 was added. In the process computer data are processed and displayed graphically on 4 colour screens (560 x 512 pixels, 8 colours) by means of the process monitoring system DISIT. All activities at the simulator, including operator actions, are performed locally by the operator at the screens by means of function keys or dialog.
H. Moritz, R. Hummel

The Design, Development and Operation of a Compact Nuclear Power Plant Simulator

This paper discusses the philosophy and technological considerations necessary for constructing and utilizing a plant specific compact nuclear power plant simulator, how it compares to full scope replica simulators, engineering simulators, part task simulators and basic principles training simulators. Included in this discussion are the design process, scope of simulation, the manufacturing process, test programs and experiences with operator training.
Items addressed include the applicability and use of a compact simulator, how well it reproduces the actual reference plant, how well the transferal of knowledge is accomplished and what financial considerations need to be evaluated.
This paper will try to provide the details on just how this type of machine was designed and developed by Westinghouse for the Swiss Utility, Nordostschweizerische Kraftwerke (NOK) AG.
Michael F. Lynch, E. Grimm

Design and Analysis of Nuclear Processes with the Apros

APROS is the product being developed in the Process Simulators project of Imatran Voima Co. and Technical Research Centre of Finland. The aim is to design and construct an efficient and easy to use computer simulation system for process and automation system design, evaluation, analysis, testing and training purposes. At halfway of this project a working system exists with a large number of proven routines and models. However, a lot of development is still foreseen before the project will be finished. This article gives an over view of the APROS in general and of the nuclear features in particular. The calculational capabilities of the system are presented with the help of one example.
M. Hänninen, E. K. Puska, P. Nyström

The Role of Simulation in Control System Design/Modification

Due to the discrepancy between design and actual plant data, controller tuning is required during power test of new plant. Furthermore, after a period of operation time, the aging effect of the sensors and components will cause the system performance to change. And with the improvement of control system hardware, better control algorithm can be implemented to assure the safety of the system operation. Control system tuning/modification is necessary to keep the system at its best performance. So, it is not once in a system life time job.
Simulation play an important role in control system design and modification. Because the reactor itself has a very restricted operation regulation, we can’t afford to modify/tune the control system by applying the “trial-and-error”method directly on actual power plant. Instead, physical plant is modeled in terms of set of mathematical equations in advance. The control system design and analysis are performed by computer simulation.
Besides the computer simulation, we can also verify and try the newly developed algorithm on experimental reactor system. The Taiwan Research Reactor power regulating system modification work is taken as an example to demonstrate the role of simulation. In this work, the model of Taiwan Research Reactor and its power regulating system were setup and analyzed by simulation. Parameter optimization and tuning map technique is applied to optimize the control system performance. Analysis results will be verified on Taiwan Research Reactor. These procedures can also be applied to commercial nuclear power plants.
Sen-I Chang, Shih-Jen Wang, Min-Song Lin

USNRC’s Nuclear Plant Analyzer: Engineering Simulation Capabilities in the 1990’s

The Nuclear Plant Analyzer (NPA) is the U.S. Nuclear Regulatory Commission’s (NRC’s) state-of-the-art nuclear reactor simulation capability. This computer software package integrates high fidelity nuclear reactor simulation codes such as the TRAC and RELAP5 series of codes with color graphics display techniques and advanced workstation hardware. The NPA first became operational at the Idaho National Engineering Laboratory (INEL) in 1983. Since then, the NPA system has been used for a number of key reactor safety-related tasks ranging from plant operator guidelines evaluation to emergency preparedness training.
The NPA system is seen by the NRC as their vehicle to maintain modern, state-of-the-art simulation capabilities for use into the 1990s. System advancements are envisioned in two areas: first, software improvements to existing and evolving plant simulation codes utilized by the NPA through the use of such techniques as parallel and vector processing and artificial intelligence expert systems, and second, advanced hardware implementations using combinations of super-, minisuper-, supermini-, and supermicrocomputer systems and satellite data communications networks for high flexibility and greatly increased NPA system performance.
E. T. Laats

CASMO-3/SIMULATE-3 Core Follow Calculations on Oskarshamn 3

The new Studsvikcore analysis package, comprising CASMO-3 and SIMULATE-3 as the main components, has been used to performcre follow calculations on the Oskarshamn 3 ASEA-ATOM BWR-3000 reactor.
The initial and the second cycle have been followed. The average k-effective value of cycle 1 is 1.0005 with a standard deviation of 0.0005 and of cycle 2 1.0018 with a standard deviation of 0.0009.
At some of the reference points the calculated distribution have been compared to TIP traces. The calculated axial traces agree very well with the measured traces. Both neutron TIP and gamma TIP traces have been analysed. The radial power is predicted within 3–4 % (rms)
K. Ekberg, S. Lundberg

CASMO-3/MBS Benchmark Calculations on RINGHALS PWR

The Studsvik code package CASMO-3/MBS for PWR analysis has been benchmarked against several cycles in the three Ringhals PWR. The results from these calculations show quite good agreement with measured boron letdown curves and power distributions. Cores with different degrees of leakage loading pattern show the same good agreement. The over all RMS deviation in assembly power over 4 cycles is 1.05%, with single cycles going down to 0.71 % RMS.
E. B. Jonsson, M. Eriksson, C. Holmlund, G. Norström

3-D Full Core Calculations for the Long-Term Behaviour of PWR’s

Presently, the most realistic simulation of a pressurized water reactor (PWR) core is by means of three-dimensional (3-D) full core calculations. Only by such 3-D representations can the large scope of axial effects be treated in an accurate and direct way, without the need to perform various auxiliary calculations. Although the computationally efficient burnup-corrected nodal expansion method (NEM-BC) is used, the computing effort for 3-D reactor calculations becomes rather high, e.g. a storage of about 320000 words is required to describe a 1300 MWe PWR. NEM-BC was introduced (1979) into KWU’s package of PWR design codes because of its high accuracy and the great reduction of computing time and storage requirements in comparison to other methods.
The application of NEM-BC to 3-dimensional PWR design is strongly correlated with the progress achieved in the solution of the homogenization and dehomogenization problem. By means of suitable methods (equivalence theory) the transport-theoretical information of the pinwise power and burnup distribution for the heterogeneous fuel assemblies is transferred in a consistent manner to the full core reactor solution. The new methods and the corresponding code system are explained in some detail.
H.-J. Winter, K. Koebke, M. R. Wagner

Numerical Methods for Advanced LWR Core Simulators

The subject of this paper is the solution of partial differential equations for the simulation of the reactor core on high-performance computers. The multi-level methods used for the calculation of core power distributions are well-suited for the considered multiprocessor systems. The implementation of a multigrid nodal diffusion method on a special array configuration of the memory-coupled multiprocessor DIRMU is outlined. The problem of time integration is discussed from the viewpoint of advanced multi-level techniques and higher-order integration methods. For the reconstruction of local power distributions by means of coarse-mesh solutions the variable-order ECD-method is introduced which yields accurate flux expansion coefficients in the center of a calculational node.
R. Böer, H. Finnemann, E. Michel

Evaluation of Systematically Derived Neutron Kinetics Models

A variational principle made stationary by the solution of the two-group, three-dimensional, time-dependent matrix equations and adjoint matrix equations embodied in the nodal code QUANDRY is used to derive a number of more approximate neutron kinetics models varying from point kinetics to a group-dependent, one-dimensional scheme. Numerical tests of the accuracy of these models substantiate the value of using adjoint weighted kinetics parameters. They also suggest that using unity weighted parameters (while not acceptable for all cases) is usually preferable to Galerkin weighting, and that certain one-dimensional models are actually inferior to the adjoint-weighted, point kinetics scheme.
Antonio F. V. Dias, Allan F. Henry

Computer Simulation of the Long-Term Stability of a NuclearWaste Repository in a Salt Dome

Numerical simulation of repository-induced perturbations with in the host rock is of particular significance because licensing procedure for a final repository requires a prior reliable and convincing demonstration of safety. Long-term assessment of the salt barrier integrity cannot be evaluated from experiments alone but only by computations. A proper geomechanical modeling is necessary to evaluate barrier efficiency. The necessity of valid at ion of the geomechanical model is explained. Preliminary design calculations oriented towards problem identification and trend indication are presented.
M. Wallner

Response of Underground Openings to Dynamic Loadings

A Finite-Element program for seismic analysis of underground structures is presented. The numerical formulation solves nonlinear dynamic problems including an infinite medium of rock or soil. As a general solution strategy a hybride numerical method is proposed. Its efficiency is demonstrated by examples. The combination of in-situ dynamic loading experiments and dynamic Finite-Element calculations gains the knowledge of dynamic response of underground structures and provides a basis for predictive numerical calculations.
H.-J. Alheid, K.-G. Hinzen, A. Honecker, W. Sarfeld


Weitere Informationen