Basic Principle's Simulator of Embalse Nuclear Power Plant

Si desea esta página en español, pique aquí

Introduction

Since exists fast and cheap computers, with great store capacity, it has been increased the use of simulators in the training of space craft pilots and nuclear power reactor or another complex plant operators, where an operation error can lead to catastrophic consequences.

A basic principle's simulator of a nuclear power plant consists in a set of programs that resolves, numerically and in real time, the equations that governs the dynamic behavior of the simulated plant.

These kind of tools are used, mainly in the first stage of nuclear power plant operation staff training, giving a meaning of the physic phenomena that governs the plant. They also are used as teaching tools for control methodologies. They are used for students and professionals training, and also for experimented operators to enhance their efficient to cope with abnormal or accidental situations that could take place in the plant.

These simulators are designed with single models of the plant, reducing the control room desk to a set of graphic monitor computer stations, through which the user interacts with the simulated plant.

In the frame of United Nation Project for the Development (PNUD) sponsored by International Atomic Energy Agency(AIEA) in joint with the Comisión Nacional de Energía Atómica (CNEA), it began the development in 1985, in Control Process Division of Centro Atómico Bariloche, a simulator of this features.

History

First it was made the calling simulator architecture. This architecture consists in a set of programs, in the simulation environment, that manages the process synchronization and makes the simulation runs in real time.

This architecture was used first to simulate the RA-6 research reactor, situated in Centro Atómico Bariloche, with satisfactory results.

In parallel it was developed graphic systems for the information presentation to show the simulator output.

Then, it was used the architecture to finally introduce an Embalse Nuclear Power Plant (CNE) model, called MANUVR. This code was used in the Electric and Control Systems Department of Gerencia de Ingeniería (CNEA) to analyze changes in plant control systems.

Then it was taken (model and architecture) to run on a UNIX operating system, in a Personal Computer. Before, it was in a Micro VAX II under VMS operating system.

It was made several matching works between the simulator model outputs and the original code outputs.

Finally it was implemented in the same Nuclear Power Plant, in the Capacity Area, at the beginning of 1993.

Design methodology

As design criteria it was took into account the prop constrains of real time systems, that lead with particular features that must be satisfied, as:

• Managing several concurrent and synchronized processes.
• The simulation must run no more than in real time.
• Flexible and adaptable structure to the diverse kind of plants that one want to simulate.

Of several analyzed methods for real time software systems, DARTS method (Design Approach for Real Time Systems) was adopted for the architecture design. This method allows to give a modular software coding in agreement with this kind of systems.

A detailed report of this method has been described in: Gomaa, H.: "A Software Design Method for Real-Time Systems", Communications of the ACM, Vol. 27, Num. 9, September 1984.

Architecture features

The simulator architecture was designed in such a way that it at lest two persons interact with it:

• an instructor, that stands the initial conditions, introduces fails in the system's components, etc, to evaluate the second person response.
• an operator, that makes the control actions over the simulated plant, trying to realize the previous fixed operation objectives.

The simulator presents a menu set in the display to the instructor. These menus enable him to make the following instructions: begin, stop, continue and finish a simulation; configure the global memory; define the initial conditions of a particular simulation, etc. Also these menus, there are two visors that indicates to the instructor the last performed instructions over the simulator and the simulator state (UNCONFIGURATED, CONFIGURED, DEFINED LATENT, RUNNING and STOPPED). In agreement with the simulator state, there are operations that the instructor can do and others that can not. Some of those operations allow the instructor to change the simulator state.

It has be designed a real time data presentation system as the operator user interface. This system uses another computer, and communicates with the simulator computer through an Ethernet line. The simulator sends it the variables to show to the operator, and wait for receive the operator control action over the plant.

This graphic presentation system was increased with other functions that the specific ones for the simulator. For that reason this system became another project. This data acquisition and presentation system has been called DISPLAYER.

DISPLAYER showing secondary circuit variables of simulator.

Model features

The simulator has a thermohydraulic model. The modeled components are:

1. Primary circuit: The primary circuit model consists in only one loop (in the plant are two, each one passing twice into de nucleus) with: a reactor, a pressurizer, a steam generator (the plant has four), and a main pump (the plant has four). Mass and energy balance equations are used to determine the involved variables, without using the momentum balance equations.
2. Nucleus: Only the heat transfer from the fuel element to the coolant is considered in the model. The temperature profile of an average fuel element is determined. The nuclear power, in agreement with plant operation mode (NORMAL or ALTERNATIVE), is directly calculated from the control result (instantaneous and ideal control) or extracted from a temporal table. If a reactor setback, stepback or trip, is produced, the nuclear power value is determined from special tables for each case.
3. Pressurizer: Control, heaters and discharge and relief valves are modeled. The pressure is determined using the ideal gas model ( PV = nRT ) in the steam compression case. In the expansion case it takes a thermodynamic volume that includes steam and water zones that are in equilibrium with steam (saturated water).
4. Auxiliary circuit: Bleed, feed and purification circuits are modeled. Hydraulic resistances, the heavy water storage tank, feed pumps, valves, filters and the degasser-condenser tank are modeled too.
5. Secondary circuit: ASDV, CSDV and relief valves, start-up and main valves of the feedwater are modeled. The steam flow to the turbine (turbine power), in agreement with the plant operation mode (NORMAL or ALTERNATIVE), is extracted from a table or directly determined from the control (instantaneous and ideal control). In case of turbine trip, the turbine power is extracted from a special evolution table. Feedwater pumps are also modeled. The model can simulate one pump shutdown, but not its start-up. The steam generator is modeled with the mass and energy balance equations.
6. Control: The model has: primary circuit and steam generator pressure controls, pressurizer and steam generator level control, primary circuit inventory control, and turbine and reactor power controls (in agreement with the plant operation mode).

DISPLAYER showing the primary circuit mimic of simulator.

Simulated transients

The model can simulate several transients, as:

• Full power operation, in normal or alternative mode. The user can change the power with ramps, in a range among 30% to 100% of full power plant.
• Reactor and turbine trips.
• Class IV lost, that stops: main pump, feed auxiliary circuit pumps and steam generator feedwater pumps.
• The model automatically does reactor stepbacks, setbacks and trips.
• It can be modeled 30 failures: little LOCAs (lost of coolant accident), pump shutdowns, some valve failures, regulation system malfunctions, etc.

Model constrains

1. The valid power range is between 10% and 100% of full power plant.
2. The model is not valid in problems where the primary circuit transient shows a strong dependency with the momentum equations (for example if exists strong compression or decompression liquid zones, or high steam title).
3. The model is not valid in transients where the steam generator level is too small (when it is seen negative re-circulation flows).
4. The model is not valid in transients where pressurizer level is letter than heater's level.
5. Problems with primary circuit asymmetries can not be simulated.
6. Little lost of coolant accident (LOCA) can be simulated (with flow similar to the bleed flow), but bigger ones can not.

Celso Alberto FLURY fluryc@cab.cnea.gov.ar

o

Félix MACIEL PALACIO macielf@cab.cnea.gov.ar