OPAL Reactor Training Simulator

Transcription

1 OPAL Reactor Training Simulator Etchepareborda A. 1, Flury C.A. 1, Lema F. 1, Maciel F. 1, De Lorenzo N. 2, Alegrechi D. 1, Damico M. 1, Ibarra G. 1, Muguiro M. 1, 1 National Atomic Energy Commission, Bariloche, Río Negro, Argentine 2 INVAP S.E., Bariloche, Río Negro, Argentine 1. Introduction The use of training simulators in support of operators training is a common practice in nuclear power plants. One of the reasons is that nuclear safety authorities demand the operators the use of training simulators in order to obtain an operation license renewal. In the case of research reactors, the use of training simulators is not a common practice, since safety authorities do not normally demand the use of simulators for operation license renewal. Its use is associated to important research reactors related with production activities. The motivation for its use is related to diverse factors, such as the certification of operators training processes; the reduction of initial and re-qualification training times; and the potential reduction of costs associated to operation errors. In the case of the OPAL research reactor, ANSTO has decided to use a Reactor Training Simulator (RTS) facility, which can be used in support of initial training, re-qualification training, and examination of operators. The RTS consists of several computers running the same Reactor Control and Monitoring System (RCMS) graphical interface. The screens emulate the control room video display units, and with the addition of virtual instruments permit the operator to perform actions over the simulated plant instead of panel switches and pushbuttons used in the real console. In this way operators can become familiar with the RCMS graphical interface, its operation and functions without the need for a fully replicated reactor control room. The RTS specifications follow the applicable sections of the ANSI-ANS 3.5 norm [1]; which defines the functionality and development procedures for full scope nuclear power plant simulators used for operator training and examination. In this paper the general features of the OPAL RTS, together with some tools and procedures used for its development are presented. 2. General features The RTS scope includes a complete set of plant normal evolutions, and malfunctions obtained from the plant design basis accidents list. For these transients the operator is required to take the same action on the simulator to conduct a plant evolution as on the reactor facility, using the reference plant operating procedures. The main plant systems that are under direct operator supervision are included. These are the reactor core, the primary cooling system, the reactor and service pool, the hot water layer system, the reactor and service pool cooling system, the reflector with its main and intermediate cooling system, the secondary cooling system, the reactor protection system, the two reactor shutdown systems, the reactor containment ventilation system, the containment residual heat removal system, the radiation monitoring system and the electrical system. Within these systems both the variables connected to the RCMS and the local variables associated to field actions are modeled, leading to several thousands input-output variables in the plant mathematical model (PMM). The RCMS software used in the RTS is identical to the plant RCMS; a Foxboro I/A Series system [2]. In the OPAL, the control programs of this system are executed in several control processors with direct connection with the field buses. In the case of the RTS, the same control programs are executed in a

2 Foxboro engineering station (AW70X), which contains a process that emulates the operation of the RCMS. The connections with the field buses within the control programs are replaced with connections to the PMM. The control programs emulator and other applications of the Foxboro system are executed within a virtual machine in order to implement in the RTS functionality that is not available in the RCMS, such as pause, fast-slow time execution, and snapshots. Figure 1 presents the layout and the main modules of the RTS. Two RCMS visual display units with touchscreen (Fox TS 20 ) and an additional terminal for the virtual instruments (hardware emulator, HE ) are available for the trainee. Two servers are available for the instructor. One server executes the virtual machine (VM) that contains the emulation of the RCMS (AW70X) and a visual display unit (Fox 20 ). The other server is dedicated to the execution of the RTS main modules: the Model Manager (MM), the Simulation Manager (SM) and the module where the instructor commands the simulation: the Instructor Console (IC). Instructor Server 1 Fox 20 Server 2 IC 17 Trainee VM AW VMM 70X IC M M S M WS1Fox TS 20 WS2 HE TS 20 WS3Fox TS 20 HE Figure 1 Layout and main modules of the Reactor Training Simulator. The MM encapsulates the PMM, which is built using Matlab -Simulink Workshop (RTW) simulation development environment [3]. (MS) and Real-Time The IC incorporates functionality that provides flexibility for the definition of events during a simulation; ranging from plant design basis accidents to normal plant failures, such as a sensor drift or a contact failure. These events can include actions over all plant variables. The main systems models in the PMM are tested against RELAP [4] outputs. The simulated models represented correctly the steady state, tendency and timing of all variables relevant to the plant normal evolutions and malfunctions included in the DBA list. In [5], some modeling approaches and performance tests for both normal evolutions and malfunctions are presented.

3 3. Model Manager Within the MS development environment, the models are developed on graphical windows interfaces based on component libraries, where they can be easily constructed, debugged, tested and compiled to C source code that is embedded into the MM. Specific libraries of components are designed to facilitate the development of the nuclear, hydraulic, ventilation and electrical plant systems models. Further details on the RTS development facility can be found in [6]. In Figure 2 it is presented the characteristics of the interface with the model of the reactor containment ventilation system of a research reactor. The model browser in the left window indicates that the graphical model shown is a subsystem of the chilled water system. This subsystem contains several library components that were dragged and dropped from the library into the window and then graphically connected. Tanks, pipes, isolation valves, actuators and sensors are shown in the portion of the subsystem model that is shown in Figure 2. The library components are implemented through dynamic libraries, allowing lower code size and faster execution time in the model development phase. The library component s parameters are data structures that are linked to plant design input data. These parameters are defined in specific initialization ASCII files with high level Matlab commands. Plant design data is collected into these files, which initialize the geometrical data structures and perform the steady state calculation. Figure 2 Process model development environment with some library components. The library elements that define the inputs and outputs of the PMM are the sensors and actuators. The actuators define the inputs to the model and their use can be configured through a menu. They are used as signal generators during the development phase of the model. These elements can represent a local command signal, such as the local start-up of a pump or manual opening of a valve, or an automatic

4 action of the plant distributed control system setting the aperture of a control valve, or starting an air handling unit. It can also be directed to an emulated hardware of the virtual instruments terminal. These actuators can also be linked to a power source with a pre-defined fail mode, and receive fail signals from the instructor console. The sensors define the outputs from the model and their use can be configured through a menu. The signal can be output to the instructor console to represent a local sensor such as a level glass or a manometer, the instructor then can inspect this variable in the IC. The signal can also be directed to the RCMS, or to an external device, such as an emulated hardware of the virtual instruments terminal. A sensor power signal input and a sensor external fail signal input can also be added. The sensor type can be defined to be digital or analog. Each system of the PMM is individually tested for steady state, normal evolutions, and malfunctions within the MS simulation environment before its integration. All system models are integrated into the PMM, compiled into C source code with the RTW MS toolbox and embedded into the MM. 4. Instructor Console The use of the simulator is oriented to many different training objectives, such as the trainee familiarization with the OPAL RCMS; the demonstration of operative procedures for normal plant evolutions and for component malfunctions; and the evaluation of trainee skills in the operation of the plant. These training objectives are implemented in the simulator through training sessions; and the instructor console is the interface where the instructor defines training sessions. A training session starts with a login where the instructor fills administrative information for the session report, such as instructor and trainee identification and a description of the session. Then the instructor selects any of the available plant initial conditions. An initial condition is a snapshot of the plant at a particular state. Several initial conditions associated to different plant configurations are available to start a simulation session in real-time, or in slow or fast time. Fast time is intended to be used when a slow plant transient is evolving and the operator is required to wait it finish with no action in the mean time. As an example it can be the filling of a tank or the evolution of a controlled temperature in a slow process. Slow time is intended to be used in order to have time enough to explain the trainee an operative procedure with the simulation running; or also to analyze, navigating the plant RCMS visual display units, a sequence of events after a fast transient, such the loss of power to an electrical distribution bar and the triggered automatic protective actions. Once the simulation is started, all the interaction of the trainee with the plant RCMS visual display units and with the emulation of the hardware console is recorded into video files until the simulation is stopped. This feature can be used to document a session used to evaluate the performance of a trainee. It also permits the instructor to stop the simulation and replay it in the case of a trainee mistake in an operative procedure. Another use of the replay facility is to prepare training material such as video demos of operative procedures for both normal evolutions and malfunctions. The replay videos can be viewed with standard free software and without the simulator. With the simulation running the trainee interacts with the RCMS visual display units to operate the plant. When the operative procedures include field actions, the instructor, in the role of the auxiliary of the operator, implements the field actions in the simulator through an action edition interface. Also component malfunctions can be implemented through this interface. This facility gives the instructor access to all local field actions and malfunctions within the scope of the simulator, such as the inspection of field instruments; the reconfiguration of hydraulic circuits acting on local valves; the reset of a motor thermal relay; the opening of a circuit breaker; the blockage of a heat exchanger; the loss of coolant in an hydraulic circuit; the drift of a sensor. The instructor can browse into the plant subsystems, component types within the subsystem, specific component and actions available for the component type. For

5 example the instructor can select the primary cooling system (plant system), within the system select the pump electrical drivers (component type), then select a specific electrical pump driver, and then select the commutation from remote to local (one of the actions available for this component type). In Figure 3, the instructor is setting a drift fault to the sensor S_4020_01 OI_005 from the Neutron Linear Channel of the N-16 detector component of the Nucleonic Instrumentation system for a period time of 10 seconds. The instructor can also select a variable or a set of variables to be dumped to a file where their evolutions in time are recorded. Here, as shown in Figure 3, by checking the Register to dump checkbox, the evolution of the Neutron Linear Channel measurement value is recorded to a file. The grouping of systems, component types and specific components is based on plant tags and are linked to the sensor and actuator objects embedded into the mathematical model. All the sensors and actuators that were linked to the instructor console are presented in the proper subsystem and component type based on their tag. Figure 3 Tasks loading and definition in the instructor console. These actions are grouped into a tasks list and are assigned by the instructor an activation time that contemplates the typical delay associated to the field action or the desired time of activation of a malfunction. If the tasks list is a common plant operative procedure, it can be saved to be reused in other training sessions. All these tasks are documented in the session report. Operations that involve a significant delay among the individual tasks are grouped into procedures, where individual tasks have an activation time defined relative to the initiation of the procedure. The instructor can load (or save) a procedure during the simulation. At any time the instructor can pause the simulation and take a snapshot. This snapshot is stored as a backtrack point for the session, and can be saved as a plant initial condition. A backtrack point is equivalent to an initial condition, but available only within the simulation session. Within the session, the instructor can pause or stop the simulation and resume it from any of the backtrack points defined. This

6 permits to focus the training on a critical part of an operative procedure without having to repeat completely non critical and time consuming parts of it. A training session can be saved in order to be resumed later. When the session is finished, the simulation is stopped, replay video files are generated and can be viewed, and the session report is generated. This report is based on information presented in the log window. This window shows all the relevant events that have been activated on the session. These events comprise: tasks lists, system messages, commands, comments, field actions, malfunctions, out of scope messages. The instructor can select the information to be presented in the report that may also include the RCMS alarms that were activated during the session. 5. Conclusion The general features of the OPAL RTS, together with some modeling tools and procedures used for its development were presented. The RTS includes the same Reactor Control and Monitoring System graphical interface used in the OPAL reactor. This interface, together with virtual instruments that emulate hardware components of the reactor consoles, permit the operator to perform actions over the simulated plant without the need for a fully replicated reactor control room. The RTS model scope includes the main plant systems that are under direct operator supervision. The operator is required to take the same action on the RTS to conduct a plant normal evolution or malfunction as on the reactor facility. The RTS model is built with Matlab, Simulink and Real-Time Workshop, which eases the model development process through the use of the specialized hydraulic, air, nuclear and electrical components libraries. The libraries components model scope proved to be adequate to represent the normal evolutions and malfunctions specified for the RTS. Specific systems or components that require more complex thermo-hydraulic models can be developed through engineering design codes, such as RELAP, or through proprietary models, and then they can be embedded directly into the RTS plant mathematical model. 6. References [1] American National Standard ANSI/ANS Nuclear Power Plant Simulators for Use in Operator Training and Examination. [2] The Foxboro Company, Foxboro I/A Series system. 33 Commercial Street, Foxboro, Massachusetts USA. [3] MathWorks Inc., MATLAB, Simulink and Real Time Workshop. 3 Apple Hill Drive, Natick, MA , [4] Innovative Systems Software, RELAP/SCDAPSIM/MOD3.2. L.L.C, [5] Etchepareborda et al., ANSTO RRR Simulator for Operating Personnel Training, Proceedings of the IAEA International Conf. on Research Reactor (Utilization, Safety, Decommissioning, Fuel and Waste Management), Santiago, Chile, [6] Etchepareborda et al., Reactor Simulator Development Facility for Operating Personnel Training, Proceedings of the IAEA International Conf. on Research Reactor (Utilization, Safety, Decommissioning, Fuel and Waste Management). Santiago, Chile, 2003.