Coal

Improving the Flexibility of Coal-Fired Power Plants

Issue 9 and Volume 118.

Increasing the flexibility of coal-fired power plants
Increasing the flexibility of coal-fired power plants is an important priority for most power plant managers and utility executives. Researchers have discovered several ways to increase the flexibility of coal.

By Nikhil Kumar, Intertek

Improving operational flexibility of conventional fossil generation should be a top priority for asset owners and operators. Historically, operators could ignore the impacts of cycling on equipment, primarily the risks involved – safety, availability and costs because the need for operational flexibility was minimal and therefore the risk was small. Wholesale market deregulation, increased renewable generation and environmental pressures have impacted how plants are operated; but plant cycling is not a new phenomenon. Plants have always cycled; however, it is the intensity of cycling which has been impacted.

Larger fossil steam units primarily ran baseload, with minimal change in their output. Moreover, the majority of power plants were designed for this minimal flexibility and ran largely baseloaded. However, operating baseloaded does not mean that the unit did not perform any cycling. Spinning reserves or frequency control has often been managed by some of these large steam units. They may not operate in an on/off mode, but still perform cyclic or flexible operation.

A first step to understanding operational flexibility is characterizing cycling operation. Almost all power plants have cycled historically, whether in an on/off mode or load following. A load follow itself could be a small MW change causing relatively small changes in steam temperatures, or a significant load follow with a unit operating at minimum loads. Additionally, with the use of duct firing or supplemental firing output can be increased beyond the maximum rating of the plant for short periods. Each of these flexible operating modes has varying degree of impact on equipment reliability and damage. Figure 1 on pg. 34 provides a visual representation of these operating modes. Therefore, while units have always been cycled, the intensity of cycling has varied from one unit to another depending on economics. Occasionally, existing coal units, which are usually not subject to extensive flexible generation, have been heavily cycled as highlighted in the report from the U.S. National Renewable Energy Laboratory (NREL) and Intertek AIM for 21st Century Power Partnership.

fig 1

Flexible Generation Trends

Before we delve into determining the risks of increased cycling of fossil steam units, it is important to determine the historical trend of flexible operation using the cycling characteristics highlighted above. Some of the questions we seek to answer are – what is the typical cycling profile of power plants in the U.S., do some units cycle more than others, if yes, then why and how; are units performing increased amounts of significant load following, what is the minimum load of operation, etc.

Intertek AIM contributed to a study performed by Electric Power Research Institute (EPRI) reviewing the cycling characteristics of almost 1,000 supercritical and subcritical (coal and gas) steam units in the U.S. (Product ID:3002001180:”Effect of Flexible Operation on Boiler Components: Theory and Practice, Volume 1: Fundamentals”). Figure 2 on pg. 36 shows there has been almost negligible change in the median number of starts on the large subcritical coal units and in fact a downward trend in terms of starts on the smaller subcritical units. This result while not surprising was not expected.

fig 2

With about 30 times more wind generation in December 2011 compared to January 2000, the general expectation was to observe increased cycling on coal power plants, especially since natural gas prices have lead to the coal generation to be on the margin. Interestingly though, there are some units on the grid that do perform increased cycling. The bars shown on the charts (secondary y-axis) show the most cycled unit in the sample for both unit types, and the average number of starts on the most cycled large subcritical coal unit is 46 starts per year and small coal unit is 175 starts per year, both statistics well over the median or even the 75th percentiles.

Clearly, while the vast majority of power plants on the grid have not operated with increased cycling, there is a reasonable sample of power plants (95th percentile) that have almost 2-3 times more on/off cycling. What enables these units to cycle significantly more than the sample? More importantly, if the units aren’t performing start/stop cycling then, which units are providing the flexibility to the grid? The speculation thus far has been that natural gas fired combined cycle plants provide this flexibility, but as suggested by the EIA, the average utilization (capacity factor) of the combined cycle fleet has increased since 2005, suggesting a more baseload operating profile than intermediate or peaking.

If there is a trend of increasing baseload operation, then which units are providing the flexibility? There is no doubt that the demand for flexible generation has increased, but what kind of flexibility? To further examine the operating trends of the units, we evaluate the load following operation of the larger coal units on the grid. Figures 4 and 5 show the number of hours large subcritical and supercritical coal units operates at low loads. Interestingly, the increased low load operation has almost a similar trend as observed in Figure 3 on pg. 38 which shows an increased net generation of wind in the U.S. Even more interesting, is the sharp increase in significant low load operation at both unit types while a decrease in smaller load changes. Moreover, the supercritical units have seen a significantly sharper increase in the low load operation than the subcritical coal units.

fig 3
fig 4
fig 5

In Intertek AIM’s experience load following power plants is far better than starting/stopping power plants. If operators can keep plant equipment hot, without subjecting the equipment to huge thermal shocks, then the long term damage on the equipment can be minimized. However, operating units at low loads is not risk free. The equipment at power plants still undergoes thermal stresses and operators who are tuned to running plants at full load for extended periods of time are not trained for this operating mode. Finally, most fossil steam power plants in the U.S. are much older with aging equipment and archaic control systems providing little to no feedback to operators performing increased low load operation. With financial and environmental constraints on existing fossil fleets, it is difficult to justify large capital retrofits, so what are some of the strategies plants can adopt?

  • Benchmarking – Determine how are we doing compared to peers?
  • Understand the risk – What are the impacts of cycling operation? How have our peers performed?
  • Manage the risk.

The first section of this article has provided data to allow a rudimentary benchmark analysis. Benchmarking your operation to the peer group of units can be a competitive advantage. It is important for asset owners to determine which quartile do their assets lie in? If you determine you are cycling less than your peers, but anticipate more flexible operation in the future, then learn, adapt and be proactive about maintenance.

Understand the risk

Typically, the utilities make their decisions on which units to cycle based on factors such as unit size, age, equipment type, fuel costs, system requirements, production costs, etc. The utility plant operator needs to know the effect of the changes on his plant equipment (such as reliability and life- limiting effects) in order to decide how to allocate and recover costs, as well as how best to spend today’s dollars to avoid future costs. Power plant cycling costs have a large variation and depend on several factors such as, design, vintage, age, and operating and maintenance history. When the system requirements cause power plants to cycle, a key decisions faced by utility power plant operators is to determine how to mitigate the effects of cycling. Some effects of cycling on the performance of a power plant include changes in:

  • Number and intensity (range as well ramp rate) of cycles
    – Startup and shutdown cycles
    – Load follow cycles – mild as well significant
  • Equipment failure rates
  • Maintenance requirements
  • Heat rates and startup fuel usage
  • Temperature and pressure (stress) transients
  • Chemistry requirements

Utilities can and have often either implemented new procedures or installed additional equipment to reduce the effects of cyclic operations on their assets, which may include:

  • Increased monitoring and inspections
  • Improved control systems
  • Systems to bypass steam to the condenser
  • Turbine valve modifications, water induction prevention, and drain automation
  • Increased water treatment, instrumentation, and analysis
  • Boiler and turbine stress analyzers
  • Low Load Operation is essentially a flame stability issue. Some of these improvements will benefit start cycling, but also allow better turndown capability
    – Boiler and burner modifications to lower minimum load capability
    – Improved burner flame stability, flame scanning, and monitoring. Upgrade ignitors.
    – Sliding pressure operation
    – Variable speed drives on pumps and fans
    – Lower primary air flow
    – Fuel supply control logic upgrade
    – Procedures to reduce auxiliary load requirements at low loads

As discussed earlier, the NREL and Intertek AIM report highlighted operation of a power plant with substantial increase in cycling operation – on/off as well as turndown, with fairly limited capital investment to retrofit. However, as seen from the earlier sections of this article, this is not the norm. Moreover, the plant in question had certain design features that allowed for elevated levels of cycling, such as horizontal superheater and automated drains. Equally important was the fact that the plant had performed in a cycling mode for decades allowing procedures for operation and maintenance to evolve and operators to be trained. However, the fundamental difference in the operation of that particular unit was the forced outage rates that were accepted by the operators. The plant averaged almost 20 percent equivalent forced outage rate (EFOR) for a 10 year period, with half of those years with EFOR over 20 percent. The average EFOR for similarly sized coal power plants in the U.S. is no more than 6-7 percent for the same period. Most coal plants play a relatively different role in the U.S. compared to the plant in question, most providing a bulk of system generation.

Figure 6 on pg. 42 shows a risk chart from a small sample of units that implicitly show a relation between cycling and forced outage hours. To help reduce the clutter in the chart, some key units have been highlighted to illustrate the impact of cycling on forced outage rates for different design units. Note that there are several outliers in the figure that represent High Impact Low Probability (HILP) events which are high consequence events resulting in large forced outages. Often these events are directly related to cyclic operation while other times they may be design flaws in the equipment. The figure shows a stark difference in the forced outage rates of a unit designed for cycling, which operates with significantly more cycles per year than a set of baseload units. Most units fall somewhere in between the two trend lines shown in the chart, however, in all cases there is an upward trend of increasing forced outage rates with increased cycling. There is a time lag between failures and cycles, but eventually this trend is observed.

fig 6

Manage the risk

Most operators are well aware of the risks of cycling. Almost all well maintained plants have databases of cycling related failures – thermal fatigue (cracking of thick walled components, header ligament cracking, etc), thermal expansion (tube to header cracking), corrosion (underdeposit corrosion, oxygen pitting), short term overheating, distortion and cracking, etc. On the other hand, it is not uncommon to notice cycling related erosion type damage for example on the pulverizers with mills coming on and off or back end corrosion issues with preheater baskets or bag replacement in the baghouse are not documented as cyclic costs or risks.

Many of the equipment damage effects of cycling discussed above crop up over time. The time delay relationship between an added unit cycle and accelerated component failure might range from several weeks to several years. This raises some very important questions – cycle power plants in the short term without preparing while incurring large future costs, or proactively plan and mitigate the risks. See table for some risk mitigation strategies.

tab 1

Plants designed for cycling have already mitigated some of the long term risks. For example, a 500 MW combined cycle plant in the Intertek AIM database which is a designed for cycling, low pressure unit, with stack damper and exit stack isolation dampers, automated drains, and a bypass system with a condenser steam dump provides daily or twice daily start capability to the utility. In our experience, this unit has relatively lower cyclic damage compared to similarly sized units and also lower forced outage rates related to cyclic operation. Another unit in our database has design features for better startups and for operation at lower loads down to 15 percent of maximum rating from about 480 MW. The unit is a natural circulation subcritical coal unit, with a turbine bypass system, large drain lines to control boiler pressure and a blowdown line. This unit is in the top 5th percentile of the large subcritical coal units.

To prepare and mitigate the risks, plant operators can either implement operational improvements or retrofit their plants. Implementing operational changes can provide significant improvements in cyclic operation of both fossil steam and combined cycle power plants. This is a low hanging fruit, compared to large or even modest capital improvement projects. However, as seen from the two examples, units with cycling design can indeed operate with significantly more flexibility. Retrofitting existing plants can provide long term cost reduction to “some” owners. All units should indeed take countermeasures to reduce cycling damage, but obviously, it doesn’t make sense to retrofit every unit on the grid. Careful selection of individual units in your fleet for better startup, ramping and turndown capability should be made. Also, while fast start capability on units is desirable, lowering turndowns and improving ramping capability is likely to provide the most significant benefit. Load following instead of cycling off is always better and this type of flexibility is also going to pay huge dividends as more renewable generation is introduced on the grid. Not only will lower turndowns allow units to provide flexibility with higher penetration of variable generation, but also provide very important system backup by providing spinning resources on the grid.

Electric power generation like other highly capital intensive industries have unique constraints that don’t always allow them to respond to change quickly. However, in a competitive environment, asset owners have to be proactive and prepare for changes in their operating profile. While the majority of power plants were not designed to operate in a flexible mode, most of them have cycled historically.

Operating power plants in flexible modes is inevitable and necessary. Asset owners should determine long term penalties associated with cycling, implement and integrate new procedures and equipment, prepare and train plant operators as well as maintenance staff and discuss cost recovery mechanisms with market operators. It is important to determine the operating constraints of your unit, without impacting safety, availability and long term cost of operation.

Whether units should be cycled or not is no longer an important question, instead determining the new improved dispatch stack wherein units are operated economically with reduced long term risks from cycling, and new units are designed and procured based on lessons learnt today, will allow us to meet our future energy requirements safely and reliably.

Author

Nikhil Kumar is director of Utilities Risk and Economics at Intertek AIM.

More Power Engineering Issue Articles
Power Engineerng Issue Archives
View Power Generation Articles on PennEnergy.com