Keeping CC plants profitable in the forbidden zone1 June 2008
In the prevailing electricity market conditions many combined-cycle power plants must be operated under conditions for which they were never intended. To help run this kind of plant efficiently a family of process optimisation software solutions has been devised. They require no changes to the plant process engineering, but the potential savings for operators are on a significant scale.
The power plant instrumentation and controls division of Siemens Energy has developed a family of process optimisation solutions, for power plants being operated in a regime for which they were not originally designed, under the name SPPA-P3000 (Siemens Power Plant Automation Process Optimisation 3000). The basis for it is has been derived from an evaluation of the company’s experience as a power plant operator and constructor with over 600 GW of installed power. The solutions are based purely on software and require no changes to the plant process engineering, but the potential savings for operators can be very great. Two of the solutions are described in detail here.
Operating combined-cycle power plants in a mode other than that intended at the time of their construction frequently results in a situation where the installed automation is no longer able to cope with increased operational requirements.
The following are four commonly met examples of changed operating conditions.
• Where a plant originally intended for a low number of startups per year is now sometimes started up and shut down on a daily basis;
• Where combined-cycle power plants originally designed for operation under base-load conditions are subject to frequent load changes during operation as dictated by market conditions, that is, by load despatch;
• If plants which were not designed for the purpose are used for frequency support;
• If the significance of plant efficiency becomes greater owing to a marked increase in gas prices.
In any of these situations the existing automation system may soon come up against its limits.
To meet these various cases, a total of ten different solutions in the Siemens power plant automation (SPPA) portfolio can be used individually or in combination to optimise power plant operation. They are shown schematically in Figure 1.
Use of proven automation concepts
These requirements have been taken into account in combined-cycle power plants built by Siemens in recent years, and as a result the design and automation concepts further developed under the name ‘Advanced FACY’. Many of the automation principles that have been implemented very successfully in steam power plants for 20 years can also be readily employed in combined-cycle power plants. These include model-based, predictive ‘feed forward’ structures, the consistent exploitation of inherently stable processes and the decoupling of highly intermeshed sub-processes. The result is extremely stable operating behaviour which still features a flexible and fast response.
The SPPA-P3000 solutions have several advantages:
• They can be used for Siemens and non-Siemens power plants alike.
• They can be used for plant configurations of any kind, eg single-shaft or multi-shaft configurations, 1x1, 2x1 or 3x1 plants.
• The solutions can be run on a specially developed optimisation computer compatible with any I&C system available on the market. Only minor changes to the existing I&C are required. This computer uses the platform of the new SPPA-T3000 I&C with its ‘Embedded Component Services’ (ECS) and the functions HMI, engineering, I&C diagnostics, archiving, field integration and alarm annunciation.
The two solutions "Fast start" and "Frequency control" are described in detail below as examples, for experience has shown that great enhancement potential is possible during startup. The provision of frequency support has already proved very lucrative in many countries.
The ‘Fast start’ solution
Many combined-cycle power plants were designed as base-load power plants, in line with the market situation. No great importance was attached to plant startup and shutdown for this reason, as they were expected to occur only rarely.
However today's market situation sometimes requires frequent startups and shutdowns. Existing automation systems can soon reach the limit of their capabilities in this situation.
• The permitted thermal stress limits for thick-walled components are not fully exploited and sometimes they are not even measured.
• The lower-level open loop controls are
• Sophisticated closed-loop controls, such as steam temperature control, are not designed for startup operation – that is, they must be operated manually.
In consequence unnecessarily long and non-reproducible startups mean that the plant takes longer than necessary to reach full output, thus reducing the revenue from delivered power.
The ‘Fast start’ solution optimises and co-ordinates all the components essential to startup. It permits repeatable startup of the combined-cycle power plant in the shortest possible time without violating permitted thermal stress limits for thick-walled components, such as drums and headers. The solution comprises three modules:
Module 1 – predictive load margin computer
The most important variables influencing the startup of a combined-cycle power plant are the following:
• The heat influxes into the heat recovery steam generators, which are either determined by the load control of the gas turbines or by the diverter dampers, if implemented, which direct some of the flue gases past the heat recovery steam generators during startup.
• The steam pressure and steam flow in the heat recovery steam generators, which are determined by the bypass stations or startup valves, and
• The steam temperatures which are controlled with the aid of attemperation systems
Setpoints which permit the fastest possible startup of the combined-cycle power plant without violating the material limits must be calculated for these influencing variables.
The rate at which these three influencing variables can be increased to the values required for turbine startup is limited by the thick-walled components of the heat recovery steam generators, such as drums and headers. The thermal inertia of these components represents the main obstacle to exploiting permitted thermal stress limits. For this reason, straight response of the influencing variables to the measured thermal stress limits would yield deviations far in excess of specifications. Up to now the permitted thermal stress limits have not been exploited for this reason and in some cases they have not even been measured.
Improved exploitation of permitted limits requires a predictive calculation based on the current component condition. This cannot be achieved using classical methods, however, due to the complex models for thermal stress limits and the non-linear influence of the heat influx.
The solution is to base plant startup on dynamic online models of thermal stress limits in thick-walled components. Predictions make it possible to define setpoints for heat influxes into heat recovery steam generators, for startup valves or bypass stations and for steam temperature controls such that permitted thermal stress limits are fully exploited but not violated. Figure 2 shows this calculation with reference to steam temperature control. The calculation is exclusively model-based, and for this reason no costs for backfitting differential temperature instrumentation are included. The described method has been implemented successfully for over ten years in power plant applications.
The current thermal state of the turbine determines the requisite steam parameters in the heat recovery steam generator before turbine startup can commence. In other words, a specific strategy is defined for the given condition of the heat recovery steam generator and the turbine in each case instead of merely offering three general strategies for cold, warm and hot starts.
Module 2 – optimisation of lower-level open loop controls
In many existing plants the open loop controls have not been designed for automated startup. The operating personnel are called on to do more work in such plants. This human factor means that startup strategies vary depending on shift manning. Subsystems are often started up sequentially even though a parallel start is also possible. Startup times vary widely.
In an initial step, the operating experience of the customer is analysed at the plant. These results are combined with Siemens’ expertise gained in other units and are incorporated in the automation package to meet operational requirements. This procedure results in a considerable reduction in and major standardisation of startup times.
Module 3 – optimisation of lower-level closed loop controls
An analysis of implemented plants shows that critical control loops such as drum level or steam temperature controls continually result in difficulties leading to delays and sometimes even to situations where these controls have to be operated manually. Such difficulties occur primarily during startup. They sometimes result in considerable delays, which also vary from shift to shift.
Implementation of tried and tested control concepts means that the controls can also be operated automatically during startup. Drum level controls, for example, can also be operated automatically as single-element controls due to the stabilising effect of the feedback for control valve position. Once feedwater and steam flow measurements become available, there is a smooth changeover to three-element operation.
The operation of bypass stations and startup valves is fully automatic from startup through to the time when all steam generated is supplied to the steam turbine and up until the commencement of normal load operation.
An example of benefit analysis for a standard combined-cycle power plant with an output of 400 MW and an efficiency of 50 % demonstrates that higher revenue is possible through market-driven startup. The following factors were assumed: load-controlled startup, ie the diverter damper is used to regulate the flow of gas turbine exhaust to the bypass stack/heat recovery steam generator; 50 starts per year; and an electricity price of 80 r per megawatt hour. Based on these assumptions, the steam turbine startup time can realistically be reduced by 30 minutes. The result is an increase in revenue of 270,000 r per year and as much as 500 000 r in the case of temperature-controlled startup.
The ‘Frequency control’ solution
The primary frequency control and spinning reserve required by the transmission network operator is provided exclusively by the gas turbine in combined-cycle power plants with conventional control systems. This is due to the fact that the steam turbine is operated in natural sliding pressure mode and consequently only reacts to load changes in the gas turbine.
Older combined-cycle power plants especially are either totally incapable of or only minimally capable of providing primary control capacity due to the ramp rate limitations that apply for their gas turbines. A lucrative source of revenue remains untapped for this reason.
A solution is to co-ordinate gas and steam turbine operation predictively based on dynamic models of gas turbine, heat recovery steam generator and steam turbine behaviour so as to enable the steam turbine to perform frequency support. Lower-level controls may need upgrading in some cases. The frequency control solution therefore comprises two modules, as follows.
Module 1 – frequency support from steam turbine
The HRSGs are operated in natural sliding pressure mode in many existing plants. In other words the steam turbine only reacts passively to load changes in the gas turbine. The rapid load changes required for frequency support cannot be achieved for this reason.
The frequency control module was developed with a view to reducing the gas turbine primary control component and the stress imposed on the gas turbine. This option allows targeted utilisation of the steam turbine for the purposes of primary control. The concept is based on throttled operation of the steam turbine and rapid provision of positive primary control capacity by this turbine. Small-magnitude frequency deviations are compensated solely by the steam turbine. The gas turbine on the other hand is only activated gradually in the event of major frequency deviations, which are relatively rare. This results in a significant reduction in the amount of stress imposed on the gas turbine, as shown in Figure 3.
This throttled operation reduces the efficiency of the steam turbine. However, the effect is at least partly compensated by the fact that the gas turbine can be operated with a higher output and consequently with greater efficiency. In some cases the effect is even more than cancelled out.
Module 2 – optimisation of lower-level closed loop controls
Frequency support greatly increases the dynamic requirements placed on the lower-level controls. Some critical control loops, such as drum level or steam temperature control, are no longer able to cope with these increased requirements.
The implementation of well-tried control concepts permits controls to be upgraded for high dynamic requirements. The pressure dependence in drum level controls is taken into consideration, for example. The cascade controls still to be found in many steam temperature controls are replaced by the superior two-circuit principle with an online superheater model and predictive factoring of the influence of gas turbine power and of supplementary firing, where implemented.
This solution can also be expressed in terms of euros and cents if a number of assumptions are made. The key data for this calculation are based on a standard combined-cycle power plant with an output of 400 MW, an efficiency of 50 % and an annual operating time of 5000 hours. Full-load operation accounts for 2000 hours. This full load must be reduced by 5.3 MW in order to provide primary frequency control and spinning reserve. In the case of revenue of 115Euros per kW per year and frequency control in the range +/- 8 MW, this yields an additional revenue of Euros 640 000 per year.
The eight other solutions for power plant applications named in Figure 1 have been tried and tested and boast similar benefit analyses. Benefit can be improved if they are combined, but the solutions that can be implemented in any particular power plant and the benefit that they are capable of yielding can only be accurately determined by a thorough on-site analysis of the plant.