The future is recorded in the past3 April 2002
Managing a plant effectively requires a good understanding of the history of the plant. This requires an effective data archival system. By John Kalanik and Anthony Maurer
The pressures upon nuclear operators to run their plants in the most efficient manner possible are considerable. The general target that industry seems to be adopting is to get construction costs down to $1000/kWe and operating costs down to 1 cent/kWh. This is a tough target to achieve, but one that is necessary for the survival of the industry. If the nuclear industry is unable to compete on economic terms, then it has no future.
One element of achieving economic competitiveness is to maximise the effectiveness of operating the plant. That requires the plant management to have an accurate picture of how the plant and its equipment operate.
As information technology improves, we are better able to capture, archive, and retrieve more and more data from plant equipment. Perhaps the most valuable element of this is the ability to rapidly retrieve and interpret the data.
One company that is addressing this issue is InStep Software, which has developed eDNA (Distributed Network Architecture), a data historian that captures, time-stamps and archives large volumes of historical and real-time data from equipment and systems operations sources.
From the point of view of a software company that is spread over several different industries, the nuclear industry is slow moving, very demanding, and does not engage in cutting corners. Thus a system designed for the nuclear industry has to be much more thorough and much more reliable than for other industries.
The eDNA system can gather information and store it. With eDNA, time sensitive data is accessible via web or client server. This can be integrated with other systems for preventative maintenance, forecasting and other reporting and analytical functions. The client server is intended primarily for plant engineers and operators, providing sophisticated calculations for plant performance and aids in preventative/predictive maintenance. The web server serves as an executive dashboard, providing web access to data for forecasting and other reporting and analytical functions.
Key to the effectiveness of eDNA is the ability to carry out high-speed compression of a large database, allowing it to process and deal accurately with rapidly changing data. The amount of information passing through the platform is very large. By providing staff outside the control room to access the real-time data, it is possible to do analysis in real-time, and take action in good time.
The driving principle of the system is for it to act as a process historian giving real-time analysis, and for it to act as a platform for providing real-time information. Real-time information is the key to operating a plant optimally, as this allows much greater control and plant optimisation. In order to maximise the speed of calculating and disseminating data quickly and accurately, a compression algorithm is used. The compression system used gives the benefits of supporting a large number of users, making use of inexpensive and readily available hardware. Messaging allows a user to go in, get what information they require, and then leave the network. This ensures that the network is not swamped by excessive numbers of users at any one time.
In addition, data compression means that data can be transmitted at a much greater speed with minimal network impact. During periods of high analysis, compressed data is transferred across the network and uncompressed locally at the client. This ensures that a negative is not imposed on the network during heavy data analysis.
Gathering information at the unit level is a powerful tool for the unit in question. It ensures that the specific features of the unit are taken into account. In addition, it helps the owner optimise plant operation while taking into account the market situation for electricity. Information gathered from all the units owned by the operator enables best practice and technology to be identified quickly and effectively.
Another additional advantage of such a system is that, whereas before there was a requirement for several people trained with specific skills, the system allows them to be replaced by just one person to operate the system.
Real-time and historical data can be converted into information for:
• Analysing plant operations moment-by-moment for quality control.
• Forecasting and predictive maintenance functions that increase plant productivity and efficiency and prevent system and equipment shutdowns.
• Improving the accuracy of power bids on the trading floor.
• Reducing plant operating costs.
• Increasing revenues.
The Eberline Radiation Monitoring System (RMS) was designed in the mid 1980s. This system, while state of the art at the time of development, was noted as having had communication problems with outlying components and remote computer reliability. As a result, eDNA and RadServ were developed by Industrial Peer-to-Peer (Ip2), one of InStep Software's business units. RadServ is a system which overlays radiation monitoring computer equipment with the eDNA architecture. RadServ replaced an outdated, maintenance intensive critical plant system component that had become very unreliable, leading to less maintenance, increased reliability and fewer potential plant shutdowns.
The next stage
According to InStep, the next stage of development of the system is to use the calculation server in a time deterministic mode. In addition, the system is shortly to be used to measure the reactor flux at a BWR, and record large quantities of high-speed, time-stamped data.
Enterprise applications are also being added. It is planned to gather data from many different types of plant, such as nuclear, coal, hydro, and enable the utility to compare the performances of these different plants.
Examples of use
A number of utilities are using the system. These include:
• Southern California Edison
• Southern Energy
• Arizona Public Service
• Dominion Electric
• Ontario Power Generation
• Southern Nuclear
• American Electric Power
• Nuclear Management Corporation
One plant that has adopted the eDNA system is the AmerGen Clinton plant, operated by Exelon. The plant had identified a need to replace non-compliant equipment for radiation monitoring as a result of obsolescence. The eDNA system was installed as a RMS control system (RadServ). There was approximately six months of design, software development and pre-installation testing (which included field unit emulators and spare radiation monitors). Physical installation of the system, which impacted system availability, took about 48 hours. Due to the complexity of this project, system functions were tested/monitored for 10 days prior to declaring the system operational. RadServ is expected to be a solution for RMS to resolve the same significant issue at other nuclear plants.
With the introduction of RadServ, Clinton personnel quickly discovered the accuracy and speed of the eDNA historian. At the time, Clinton lacked a true historian for the plant process computer so the eDNA historian appeared to be a viable solution. Using the eDNA historian and software, Clinton installed a data archival retrieval system (DARS) for its plant process computer and RMS. The DARS system allowed plant-wide evaluation of system/plant real-time parameters at each desktop on site.
For example, on one occasion, the reactor underwent a scram. It was thought that this might be a result of water hammer. However, when the real-time data was analysed, it was found that there was not a water hammer problem. The operator has discovered that the system enables identification of problems rather than simply noting that there was an anomaly that needed further investigation.
The DARS system was in place for approximately six months before senior management made a commitment to fully invest in the system data. The DARS system was removed from service and data fully qualified from source to client tools. Qualified DARS data is now used in the performance of technical specification surveillances by operations, and preventive/predictive maintenance by engineering.