The fourth generation

    Power plant operators are increasingly demanding common platforms for power plant and turbine control systems. PEi looks at how OEMs are evolving their products to meet these demands.

    Dr. Rainer M. Speh, Siemens Power Generation, Germany

    If you ask power producers just what their key requirements are for instrumentation and control systems in power plants and for enterprise management, you will nearly always get the same answer. The right information at the right time and the right place is what they expect, and that for minimum investment and maintenance costs.

    To achieve compatibility for the various different functionalities, the existing software systems are provided with an additional, outer layer. In this context we talk of wrapping. To date there has not yet been any new development of an instrumentation and control (I&C) system that reflects the basic architecture of the Internet, the so-called three-tier architecture, right from the design stage. The benefits of architecture of this kind are obvious. On the one hand, this matches the mainstream of the IT world and thereby enables seamless integration of enterprise-wide communication. This is an absolute must for information provision in liberalized energy markets. On the other hand, this also yields additional benefits for system maintenance and management in terms of low total cost of ownership.

    As a solution Siemens Power Generation (PG) has developed the SPPA-T3000, which it believes is the first web-based power plant I&C system in the world that is truly worthy of this name.

    Through the ages

    In the late 1960s, computers made their debut in power plants. The first application was a sequence-of-event recorder based on a mainframe computer in a nuclear power plant.

    The next generation of computing in power plants was introduced at the end of the 1980s, using local networks that enabled client/server architecture. This is still the structure in use in most control systems current today.

    With the introduction of the Internet, not only in daily life but also in the business environment, the third generation of control systems was developed as an extension of the client/server architecture mentioned above. The resulting system structure would come to be known in the industry as ‘web-enabled’. But it is also a fact that this third generation of control systems is still made up of several different subsystems. This will ultimately lead to increased maintenance costs in the long run.


    Figure 1. The development of the different generation control systems over the last 50 years
    Click here to enlarge image

    The latest trend in general automation as well as in power plant controls is a system structure called ‘web-based’, also known as the fourth generation of control systems. The cornerstone here is the basic architecture of the Internet with its three tiers: the presentation, the processing, and the data tier. This configuration is reflected as early as the system design phase and not added by wrapping existing software as is done for web-enabled systems.

    It can essentially be stated that the I&C structures described here for the third and fourth generations support the trend toward increasing centralization. This applies not only for technical facilities, but also for organizational structures. A good fifteen years ago, distributed configurations were all the rage, and now the current trend is more and more back to centralized structures. One of the reasons for this is surely the enhanced information acquisition and provision capabilities that are now available. Modern communication networks such as corporate and public networks provide a real low-cost, reliable platform. Any information, anytime and anywhere at nearly no cost is not longer a vision, it’s reality.

    In liberalized energy markets an additional factor that is becoming increasingly significant in addition to the standard considerations of high reliability and a long lifetime for I&C systems is not merely the input of the maximum amount of data but far rather the input of the right information and thereby the important information into the decision-making process in good time and at the right point. This demands not just knowledge of the technical background, but application-specific know-how and process expertise to separate the important from the unimportant and enable a meaningful customer-specific solution to be offered.

    System structure

    SPPA-T3000, the latest power plant I&C system from Siemens PG, is the first fourth-generation automation system worldwide. Based on experiences gained in over 250 plants, which have been equipped with the web-enabling extension web4txp since 1999, a system has been completely designed according to three-tier architecture.

    The uppermost tier is composed of thin clients. In addition to the recommended and tested devices for the control room, these can be terminal devices of any kind, such as notebooks, desktop computers or personal digital assistants (PDAs). The only precondition, besides the communication capabilities via TCP/IP and an operating system, is a browser that supports Java applets. No further software installations are necessary for use as an operator terminal, engineering workplace or diagnostics client. Everything the user needs on his/her front-end device is uploaded when the central application server is started. This means that no specific maintenance or update work is required for these applications.

    The use of thin clients also makes possible seamless integration in the existing infrastructure of an enterprise and the use of existing systems without any additional costs. This means that no specific hardware or software is required for the human-machine interface outside the control room.

    Two different types of server constitute the middle tier, also known as the processing tier. The application server is a fault-tolerant server. It is based on a two-processor design (SMR) and is fully redundant. Its availability is >99.999 per cent. One application server is required for each plant or machine (unit).

    One or more automation servers are used depending on the size of the power plant and the task to be implemented. These are based on proven hardware from the market leader in automation engineering, i.e. from the Simatic S7 400 series. This ensures system reliability and longevity of investment.


    Figure 2. The SPPA-T3000’s three-tier structure is mapped onto a control system
    Click here to enlarge image

    The fault-tolerant network that has been tested and utilized over the years takes the form of open rings and is based on Industrial Ethernet. It is used as a communication network between thin clients and application servers on the one hand and to automation servers on the other. Transmission speeds up to 1 GB/s can be realized in line with the currently accepted standard. Optical or wire-based media are used depending on the switches selected for implementation.

    All I/O modules are assigned to the data tier. Components from the Simatic range that have been proven over the years are also used here. I/O modules that contain a processing function are also available. These feature an intrinsic capability for operation under emergency conditions. The connection of an auto/manual station is a simple operation if required. Special modules for turbine I&C with processing functions for short cycle times even shorter than 5 ms, e.g. for position controllers, complete the I/O spectrum.

    All I/O modules are connected to the automation servers via a redundantly configured Profibus, which can be operated at up to 12 MB/s. Both copper wire and fiber-optic cables can be used as a transmission medium.

    Software architecture

    As already explained, standard PCs and PDAs can be utilized for the user interface in the presentation tier as well as the recommended thin clients. The use of different operating systems and browsers is also possible in theory. However, due to the high distribution level of Windows, the use of the most recent version of the software and of Internet Explorer as a browser is recommended, as this configuration has been tested intensively and released.

    The real advantage of the thin client lies in its universality, which permits all functions, such as operation, engineering and diagnostics, to be performed on one machine. This standard work environment is referred to as a Workbench and makes an individual work environment possible on each front-end device.

    The majority of the power services are processed on the application server. The mechanism of the fault-tolerant server makes a single-system image available to both the operating system, Windows Server 2003, and to all applications. This means that no further measures for redundant operation need to be implemented in the application software, as the software only ever sees a single hardware unit. For this reason, standard software, such as spreadsheets, can also be made available in redundant configuration, and the SPPA-T3000 system can also run on a single hardware configuration without any modification. This provides an extended homogeneous solution for subsystems that are not presently implemented in redundant configuration.

    Available data

    The project container constitutes the core of the application server. All the configuration data, current parameters and status information are stored centrally in a hierarchical file structure in the form of XML files. Unlike previous systems, the management of redundant data is not necessary. This design also ensures system-intrinsic data consistency. In theory it is also very easy to make these data available to downstream processing programs independently of the system, e.g. in the office environment, using additional middleware.

    OPC interfaces in server and client versions are also available for the exchange of I&C information. Other interfaces for the connection of black box applications, e.g. for ash handling, are provided via additional communication modules and are connected to the network parallel to the automation servers. Like all other system components, these can also be configured redundantly.

    The actual web server, the servlets for the representation of plant displays and the various engineering functions are also components of the application server. These include session and user management, the report generator and alarm management. All the applications of the application server are programmed in Java and are therefore by definition independent of the hardware and the operating system. This ensures investment security for both the user and the supplier in the fast-moving world of information technology.

    The on-board runtime container is a special feature of the application server. This allows applications to be generated and tested during plant operation with the help of an easily created simulation environment without any effect on operation. When testing has been completed successfully, the new functions can then be loaded in the real-time runtime containers of the automation servers and executed.


    Figure 3. SPPA-T3000 faceplates: As well as being able to display several languages in parallel for different users, colours, time and date formats and other display characteristics can also be defined
    Click here to enlarge image

    It must be noted here that the load procedure of the automation servers has been drastically shortened thanks to the modern software architecture. If we take an application comprising 1000 function blocks and 8000 links as an example, it takes a mere 15 seconds for the necessary data to be loaded into a runtime container, where they can be executed immediately. This procedure can take minutes or even hours in conventional systems.

    Overall it can be said that a very robust design has been achieved through the selected software architecture. Both the runtime containers and the function blocks, which are known as automation function instances (AFI) in SPPA-T3000, are software objects that outwardly exhibit the same basic structure. The containers represent the runtime environment in which the AFIs can be executed with little or no interaction. Advantages resulting from this structure include online expansion and modification capabilities. Operating experience with the first plants implemented in this design has been more than impressive and provides convincing arguments for this approach.

    User benefits

    Increased centralization in organizations and in technology has been one of the mega-trends of the past five years. Driven by enterprise-wide internal networks and the Internet as a global communication medium, the right information can be supplied to the decision processes in good time. Technical preconditions include efficient information processing and the thin clients described above as a front end.

    SPPA-T3000 supports this mega-trend in a unique way. Thanks to the selected software architecture and the very powerful user management, different users can be allowed individual views of the information and the possible action radius of each individual user can be accurately defined. In addition to the assignment of roles to individual persons, access rights can also be set up according to process engineering areas. Data inconsistencies become a thing of the past because only one instance of each item of information is stored in the system. The resulting data consistency advantages are also apparent in the reduced engineering workload.


    Figure 4. SPPA-T3000 has additional advantages, including a transparent, lean system structure
    Click here to enlarge image

    The use of open communication standards in the area of wireless communication also opens up further options. Mobile terminal devices make all information available locally, not only for the purposes of commissioning but also during operation and troubleshooting. If a suitable infrastructure is planned, a mobile operator even becomes possible.

    Interface to applications in the power plant itself, such as to a computerized maintenance management system (CMMS), and on the enterprise control level to enterprise resource planning systems (ERP) can be implemented at any time using SPPA-T3000. The system thus constitutes the foundation stone for the success of the company by providing the right information at the right time and in the right place.

    Cost of ownership

    Modern I&C is characterized by the use of accepted industry standards. The advantages offered by this are obvious, e.g. standardized spare parts management and no necessity for special maintenance skills. Continual performance improvements in the IT world are an additional bonus.

    The unique software architecture of SPPA-T3000 offers other benefits. The subsystems of present-day I&C systems are completely replaced by a transparent, lean system structure. The maintenance required for different platforms with different operating systems and different life cycles finally becomes a thing of the past.

    No posts to display