APPLICATIONS & BENEFITS OF DESKTOP SIMULATOR

Although simulators have traditionally been for training operators, they are becoming increasing attractive for design, optimization, process diagnostics and other non-training purposes. Here is a cross-section of applications where simulation was used to achieve specific goals for a wide range of applications, and has derived substantial benefits:

• Desktop Simulator Training as Complement to Full Scope Simulator Training

Equipped with similar training capabilities as a full scope training simulator, the primary training support functions for desktop simulators are to complement the full scope training simulator in the areas of "knowledge acquisition" training: "know why" and "know why not". Complex plant abnormal incidents can be demonstrated to the operators, engineers and safety reviewers very early on in the design and training process. They can familiarize themselves with the plant design, and evaluate possible logistics of mitigating the incidents, using the desktop simulator, along with online information, animation, visualization provided by the appropriate engineering tools.

According to one trainer, the implementation of the desktop simulation training as a complement to a full-scope training simulator program has resulted in the following benefits:

• Classroom knowledge accumulation is shifted to a much deeper knowledge acquisition with early hands-on interaction with the simulator.
• Training program is shortened with the supplement of Desktop Simulator training. More importantly, time away from the plant for training is reduced.
• System by system training with the Desktop Simulator, conducted with a self-paced CBT (Computer Based Training) courseware without a full-time instructor, has provided practical, gradual and measurable goals to achieve accelerated success in the full-scope simulator training conducted in the final phase of program.

• Optimize Plant Operation

A steady-state boiler performance simulation model was developed for a 600 MW fossil-fired power station. Simulation results provided guidelines for optimizing the unit and improving the heat rate.

Unit heat rate is affected by several parameters such as coal quality and fineness, combustion air flow, combustion efficiency as measured by the unburned carbon in the ash, steam pressures and temperatures, performance characteristics of the fans and coal mills, and performance of the air preheater.

According to the plant personnel, as a result of better training using the heat rate simulator:

• A reduction of 25 BTU/KWH was achieved.
• Projected fuel savings with the improved heat rate was in excess of U.S. $ 1 million over a ten year period.

• Validate Control Software

Many utilities are replacing older control systems with state-of-the art computer based systems. Unfortunately, control-system replacement typically increases unit trips and reduces availability during the early years of operation while controls are being debugged and operators are adapting to the radically changed Man-Machine Interface (MMI). A nuclear power plant, however, modernized the control system of the plant without sacrificing initial performance.

To do this, the utility relied on compact Simulator technology to train the plant operators and validate control-system software before the new installation was completed. The high-fidelity process simulation is interfaced to actual control-system operator display consoles, so that operators see the same process graphics and experience the same dynamics as the real plant following the upgrade. It was reported that:

• The commissioning time for the new control system was shortened by at least three weeks due to the familiarity and the confidence of the operators in the new system.
• Annual outages were reduced, yielding two days of additional full-load generation annually.
• The savings with shorter commissioning time and reduced forced outages had generated at least several million dollars U.S. revenue over one year.

• Develop Strategies

Loss of feedwater supply is a serious challenging event in power plant operation. This is particularly so in plants with once-through steam generators which hold small inventory of water. In one nuclear power plant, the loss of one of the two main feedwater pumps at full power caused the reactor coolant system to heat up and tripped the plant within seconds, resulting frequent, unwanted forced outages.

To alleviate the problems caused by the loss of feedwater supply, control strategies were developed using desktop simulation that enabled the simulated plant to survive a main feedwater trip from 100 % full power without operator's intervention. After intensive simulator testing under every possible plant upset scenarios, the control strategies were implemented into the plant control computers and successfully allowed the plant to respond automatically to the loss of main feedwater pump without placing the plant in danger of a reactor trip.

• Human Machine Interface Testing and Validation

Proper implementation of a quality-controlled design process for Human Machine Interfaces (HMI, or MMI) is a critical element in all engineering design disciplines. Particularly in nuclear power plant operations, regulatory standards are established to ensure all operator interfaces, displays, alarms and annunciations follow sound human factor engineering (HFE) principles and should be carried out in all phases of the design cycle.

In one nuclear power plant design activity, desktop simulator was used to support a mock-up control centre used for designing a new generation of control room complete with advanced Plant Display System (PDS), alarm CRTs, distributed control system (DCS) and panel devices.

To satisfy human machine interface testing and validation, the simulator played an important role. For example, in a simulated plant transient, the dynamic simulated plant data were transmitted to the Plant Display System (PDS) via a data highway. High performance graphics, designed according to an established HFE quality plan, were displayed at the PDS, for evaluation with respect to their effectiveness on operator's decision support functions. Factors such as display completeness, correctness, complexity, speed of induced human response, as well as the clarity for displaying plant evolutionary states were evaluated.

• Advancing Operator Decision Support Tool

A nuclear research organization has used an engineering simulator to develop an advanced operator decision support tool for a nuclear power plant using Artificial Intelligence (AI) and Artificial Neural Network (ANN) technologies. The tool is used to help operators identify malfunctions in plant operation. By acquiring data from simulator runs, the tool was able to "learn" how different transients, malfunctions, and equipment failures affect plant process parameters.

Hundreds of test runs were made at various power levels and operational modes and dozens of failures were inserted in the simulator and analyzed by the tool. After this "training", the tool was able to detect and identify off-normal plant conditions almost immediately, helping operators to take pre-emptive actions to mitigate any failures and prevent their misdiagnosis from complicating the recovery.

• Process Diagnostic

Process diagnostic, or broadly identified as process disturbance management, is one of the most important and difficult tasks that the operators should perform in the event of a plant upset caused by malfunction of a control device or an equipment. It is not only because process diagnostic is a rather complex problem, particularly when dealing with thousands of instruments and equipments, but also because the incumbent process disturbance problem, and its evolutions, often far exceed the human cognitive ability, both in terms of the speed and the depth of correlating events into reliable pattern recognition. Failure to cope with process disturbances not only leads to plant outages, but also may have serious impact to public safety.

To circumvent this problem, one utility has developed a real-time operator advisory system, based on utilizing a high fidelity dynamic simulation model for tracking the real process online. Any discrepancy between the monitored process variable and the simulated process variable will trigger a "search" in the "knowledge base", which was constructed for the purpose of fault diagnosis.

Promising success was reported in the implementation of such advisory tool for a feedwater system in a nuclear power plant. The advisory system was able to routinely identify valve "sticking" or "leakage" problems associated with the system, when these problems were not serious enough to be detected by conventional alarm and monitoring system. The early detection of these problems, followed by maintenance, has substantially improved the reliability of the system, and for a large part, contributed to the much improved availability of the plant.