Operating Flexibility of Power Plants With CCS

Background to the Study

Most assessments undertaken by IEAGHG and others have assumed that power plants with CCS will operate at base load. It is now becoming clear that in many cases CCS plants will need to be able to operate flexibly because of the variability of electricity demand, increased use of variable renewable energy sources such as wind and solar and poor flexibility of some other low-CO2 generation technologies such as nuclear. However, relatively little work has so far been published on this subject.

IEAGHG has commissioned Foster Wheeler Italiana to carry out a study to review the operating flexibility of the current leading power generation technologies with CCS and to assess performance and costs of some techniques for improving flexibility. This overview of the report was written by IEAGHG.

Scope of Work

The study assesses the flexibility, performance and costs of several examples of power plants with CCS but it is recognised that there are many other potential design options with different degrees of flexibility. The study covers the following leading technologies for power generation with CCS:

  • Ultra-supercritical pulverised coal (USC-PC) with post combustion capture using solvent scrubbing
  • Natural gas combined cycle (NGCC) with post combustion capture using solvent scrubbing
  • Integrated coal gasification combined cycle (IGCC) with pre-combustion solvent scrubbing
  • Pulverised coal oxy-combustion

The study makes use of baseline plant performance and cost data from earlier IEAGHG studies, taking into account cost inflation that has occurred since those studies were undertaken.

The following techniques for improving flexibility and increasing peak power output were assessed:

  • Turning off CO2 capture
  • Storage of CO2 capture solvent
  • Storage of liquid oxygen
  • Storage of hydrogen
  • Storage of CO2 or solvent to provide a constant flow of CO2 to transport and storage

The report also includes a brief overview of energy storage techniques for large scale electricity generation.

Results and Discussion

Operating flexibility of power plants without CCS

Typical flexibilities of power plants without CCS are summarised in Table 1. It should be noted that actual flexibilities of power plants depend on the plant design and the preferences of vendors and operators.

Table 1 Typical operating flexibilities of power plants without CCS

NGCC USC-PC IGCC
Minimum load, % 40-50 30 50
Hot start-up time, hours 0.75-1 1.5-2.5 6-8
Cold start-up time, hours 3 6-7 80-100
Ramp rate, % per minute 4-6 (40-85% load)2-3 (85-100% load) 2-3 (30-50% load)4-8 (50-90% load)3-5 (90-100% load) 3-4

The flexibility of NGCC plants has improved in recent years as suppliers continue to respond to customers’ requirements for greater flexibility and modern NGCCs are typically capable of fast start-up, shut–down and load cycling. The minimum operating load is usually determined by the increasing environmental emissions at low loads.

USC-PC plants are also characterised by low minimum operating loads and good cycling capabilities and start-up times. In contrast, IGCC plants have relatively low cycling capabilities, high minimum load and long start-up times although faster start-up may be possible if an auxiliary fuel is used in the gas turbines.

Operating flexibility of power plants with CCS

There is currently relatively little information in the public domain on operating flexibility of CO2 capture processes and more practical research and dynamic modelling is needed. This report provides illustrative information on CCS plant flexibilities but it should be recognised that flexibilities depend to some extent on the needs of the operators and there is a trade-off between flexibility, costs and efficiency, which is explored to some extent in this report. The characteristics of electricity systems in future may be significantly different to those at present, so it is important that there is a dialogue between CCS process developers and electricity system planners, modellers and operators to ensure that CCS processes are designed to have the appropriate degree of flexibility.

One of the general constraints on part load operation of CCS plants would be the CO2 compressors which would typically be limited to around 70% turndown. Higher turndown could be achieved by recycling compressed CO2 but this would impose a significant energy penalty, as the compressor would still be operating at 70% load even when the power plant was turned down further. It would therefore be advantageous to have multiple CO2 compressors, which may be required anyway due to size limitations, particularly in multiple train power plants. This report is based on power plants that include one or two power generation units. Larger plants with multiple units and common air separation and CO2 compression may provide improved part load performance.

NGCC and USC-PC with post combustion capture

The introduction of post combustion CO2 capture may impose additional constraints on the startup and fast load changing of a power plant but techniques are available to overcome these constraints. In an NGCC plant the gas turbine starts up more rapidly than the heat recovery steam generator (HRSG) and the steam turbine. The regenerator in the CO2 capture plant requires steam from the HRSG or steam turbine and the regenerator needs to be heated to its operating temperature. To avoid constraints on start-up time and to avoid CO2 emissions during start up, the CO2 absorber could be operated using lean solvent from a storage tank and the CO2 rich solvent from the absorber would be stored and fed to the regenerator later. This would enable an NGCC or USC-PC plant with CO2 capture to start up and change load as quickly as a plant without capture. This technique is evaluated in the report.

Oxy-combustion

The main constraint on flexibility of a pulverised coal oxy-combustion plant is the air separation unit. The minimum operating load of the cold box is around 50% while the minimum efficient load of the main air compressor is around 70%. At lower loads, part of the compressed air would generally be recycled to the compressor feed, which imposes a substantial efficiency penalty. This could be avoided in a multi-train plant in which one or more of the compressors could be shut down.

The maximum ramp rate of the ASU is typically 3% per minute but the boiler can typically ramp at 4-5%. The difference between the ASU oxygen supply rate and the boiler demand for a 50%-100% ramp is less than 10 tonnes for a 500MWe plant and this can be satisfied by using stored liquid oxygen (LOX). The LOX storage tank can be refilled during times of reduced power plant load. Around 200 tonnes of LOX storage would typically be included in the plant for the safe change-over from oxygen to air firing and in case of a ASU trip, so no additional LOX storage would be needed to satisfy the ramp rate.

IGCC

As mentioned earlier, the flexibility of IGCC plants without capture is relatively poor but the addition of capture is not expected to significantly affect the flexibility because for example the changes to the design of the acid gas removal plant have no impact on the plant flexibility. Plants with capture will however have reduced part load efficiency for example due to the lower efficiency of CO2 compression at part load which is discussed earlier.

Part load efficiencies

The efficiencies of power plants with CO2 capture at part load are shown in Figure 1.

Image

Figure 1 Part load efficiencies of plants with CO2 capture

The efficiency reduction for operation at 50% load is 3.1 percentage points for the PC plant with post combustion capture. This is higher than for a plant without capture, mainly due to the need to maintain the pressure of the steam extracted from the turbine for the CO2 capture plant, the lower efficiency of CO2 compression and miscellaneous changes within the capture unit. The efficiency reduction for PC oxy-combustion is similar at 3.8 percentage points. The main reasons for the higher efficiency reduction in this case are the lower efficiencies of the ASU and CO2 compressors.

The part load efficiency reduction for NGCC and IGCC depends mainly on the performance of the gas turbine and the data in this report are based on a model of gas turbine that has a relatively high part load efficiency loss. In recognition of the increasing importance of plant flexibility some gas turbine vendors are introducing turbines that have improved part load performance, as illustrated in the main report.

The data points in Figure 1 for NGCC at 50% load and IGCC at 56% load are for operation with both of the gas turbines turned down. The data point for IGCC at 48% load is for operation with one of the gas turbines shut down and the other operating at 100% load, which is significantly more efficient. This operating mode could also be used for NGCCs but it was not analysed in this study.

Assessment of techniques for improving flexibility

Turn off or turn down of CO2 capture

The net power output of a plant could be increased by turning down or turning off the CO2 capture and compression units and emitting more CO2 to the atmosphere. The ability of a plant with capture to ramp up power output could in principle be better than that of a plant without capture if the load of the capture unit was reduced at the same time as the load of the power generation unit was increased. This study assessed the option of turning off capture but various intermediate options involving turning off or turning down parts of the capture plant may also be attractive.

Turning down or turning off capture would increase emissions of CO2 to the atmosphere so regulations would have to permit CCS plants to emit more CO2 during times of peak power demand. This would for example require emission performance standards to be assessed over long periods such as a year. To comply with performance regulations it may be necessary to capture a higher percentage of CO2 during normal operations to compensate for the extra emissions when the capture plant is turned off. The feasibility and costs of doing this have not been assessed in this study.

Turning down or turning off post combustion capture would reduce the plant’s internal consumption of electricity and the low pressure steam that would otherwise be consumed by the capture unit could be used to further increase the net power output, provided the plant was built with the necessary extra low pressure turbine capacity.

Turning off capture in IGCC plants is less straight forward than in plants with post combustion capture because the CO2 capture unit is an integral part of the acid gas removal (AGR) unit which also removes sulphur compounds from the fuel gas. However, it is possible to tune to a certain extent the CO2 capture rate by varying the solvent circulation rate flowrate in the AGR unit, in order to absorb sufficient H2S while only absorbing part of the CO2. With this strategy the capture rate range at which it is possible to operate is limited by both the AGR design and the flexibility of the gas turbine to accept a variable fuel composition. In the plants considered in this study the captured CO2 that is available at high pressures from the AGR is fed to the gas turbines. This enables the quantity of nitrogen that has to be compressed for use in the gas turbines to be reduced, which reduces the compressor power consumption and hence increases the net power output of the plant. CO2 that is available from the AGR at low pressure is vented to the atmosphere but changes to the plant need to be made to reduce emissions of trace components in the vent stream, particularly H2S and CO, to environmentally acceptable concentrations. In this study two techniques were assessed:

  1. Modification of the AGR to improve the purity of the CO2 vent stream.
  2. Include a partial oxidation unit and an activated carbon bed to clean-up the CO2 vent stream.

The modified AGR case has the higher peak power output and efficiency during peak load operation and a lower capital cost but it has a lower efficiency during the time when CO2 is captured.

Only qualitative assessment of turning off capture in oxy-combustion plants was considered. The option of continuing to capture CO2 while turning down the ASU and using stored oxygen in the boiler, which is discussed later, was expected to be more attractive than short term switching between oxygen and ‘air-firing’ modes.

The results of the analysis of turning off capture are summarised in Table 2. The specific emissions for peak power generation shown in this table are calculated in the following way:

Image

Where:

Ep is Emissions for peak generation, t/MWh

Er is Emissions from the reference plant operating with capture, t/h

Ev is Emissions from a plant venting CO2-containing gases, t/h

Pr is Net power output of the reference plant with capture, MW

Pv is Net power output when venting CO2-containing gases, MW

Specific costs for peak generation are calculated in a similar way.

Table 2 Turning off CO2 capture

NGCC PC IGCC
Increase in power output with no capture, % 15.9 27.4 6.4
Thermal efficiency, %
Reference plant with capture 50.6 34.8 31.4
Plant with capability to turn off capture 50.2 34.2 31.1
Plant with capture turned off 58.6 44.3 33.5
Capital cost
Change in cost per kW of normal output, % +5.8 +3.9 +0.5
Change in cost per kW of peak output, % -8.7 -18.5 -5.6
Cost of extra peak power capacity, €/kW 354 322 213
CO2 emissions
Tonnes CO2 per MWh of extra peak power 2636 2944 10450

It can be seen that having the capability to turn off capture increases the capital cost of the plant (per kW of normal power output), mainly because of the need for greater steam turbine capacity, but the cost per kW of peak power output is lower. The net capital cost per kW of extra peak power generation capacity is relatively low, probably less than the cost of other types of peak generation capacity such as simple cycle gas turbines but the specific emissions of CO2 per kWh of extra peak power generation are high, particularly for IGCC. Including the ability to turn off post combustion capture reduces the net efficiency of the plant during normal operations because the low pressure steam turbine is oversized to enable it to use the extra low pressure steam that is available when capture is turned off. The turbine therefore operates at non-optimum conditions when the capture plant is operating. To avoid this efficiency reduction a separate steam turbine could be installed to use the low pressure steam that is available when capture is turned off. This approach was adopted in the solvent storage cases described later.

The economic viability of turning off capture would depend on the carbon emissions cost, the number of hours per week that capture is turned off and CO2-rich flue gas is vented and the peak electricity prices during the time when capture is turned off. The relationship between these parameters for a base load PC plant is shown in Figure 2. Peak power costs would be slightly lower for turning off capture in an NGCC than a PC plant.

The peak power price will be determined by the cost of alternative peak load generation techniques, including simple cycle gas turbines and energy storage (pumped hydro, compressed air energy storage, batteries etc). Determining the costs of these techniques was beyond the scope of this study but in Figure 2 of this overview the costs of a simple cycle gas turbine (SCGT) plant are included for comparison with the costs of turning off CO2 capture. The SCGT plant was assumed to have an efficiency of 40% (LHV), a capital cost of €450/kW, and an emission cost of €50/t of CO2. Two SCGT cases are shown, one based on natural gas at €8/GJ and the other based on distillate oil at the current price of €16/GJ.

Image

Figure 2 Economics of turning off CO2 capture (PC plant)

The overall cost of generation increases as the number of hours per week that CO2 capture is turned off is reduced because the fixed costs associated with turning off capture (Capex and O+M) are attributed to a lower number of MWh of peak power. It can be seen that for an emission cost of €50/t of CO2, turning off capture is less economically attractive than an SCGT, although the costs are broadly similar if oil has to be used as the fuel for the SCGT. The economic advantage of the SCGT becomes greater at higher CO2 emission costs, because the specific emissions associated with capture by-pass are higher than for an SCGT.

Solvent storage

Solvent from post combustion capture can be stored during times of peak power demand for regeneration during times of lower power demand. This reduces the requirement for other peak generation capacity. The extra generation during peak times would have low CO2 emissions, unlike the alternatives of by-passing CO2 capture as described earlier, or using peaking plants such as simple cycle gas turbines without CCS. Solvent storage in IGCC was not assessed in this study because the Selexol solvent would have to be stored at high pressure and it was expected that the costs would be high compared to other techniques e.g. liquid oxygen storage.

Foster Wheeler discussed the practicality of CO2 solvent storage with some leading technology suppliers, including MHI, Aker Clean Carbon and Alstom. These companies all confirmed the technical feasibility of storing solvent, provided the temperature of CO2-rich solvent is maintained at or slightly below the absorber bottom outlet temperature to avoid degassing. High rates of degradation are not expected, degradation would be mainly due to the reaction with oxygen, so nitrogen or CO2 blanketing would always be considered. MEA-water solution that would be stored in capture plants is not flammable but solvent is toxic and the stores are potentially large, as discussed later, so it may not be acceptable at all locations.

Regeneration of stored solvent could take place during times of ‘base load’ operation or during times of low power demand when the power plant is operating at part load. The operating mode of the plant would determine the required capacities of the solvent storage tanks and the solvent regeneration and CO2 compression equipment. If the plant is required to operate only at ‘base load’ the solvent regenerator and CO2 compressor would need to be oversized to cope with regeneration of the solvent from ‘peak load’ operating hours. If the plant is expected to operate for some of the time at reduced load, the stored solvent could be regenerated during these times and the regenerator and compressor would not need to be oversized. If a plant is expected to regularly operate at substantially reduced load at night and at weekends, the solvent regenerator and CO2 compressor could be undersized, i.e. they could be made smaller than in a normal base load power plant, thereby reducing capital costs. However, such a plant would not have the ability to operate at base load for long periods of time and this may not be attractive to the plant owner.

Two operating scenarios described below were assessed in this study as an illustration but it is recognised that in reality power plant operations will depend on many external factors which may change during the operating life of a plant. PC plants were assumed to be operated at higher load factors than NGCC plants at night and at the weekend because their lower marginal operating costs would put them higher up the operating ‘merit order’. The ‘weekly’ and ‘daily’ scenarios involve different amounts of solvent storage and peak load operation.

  1. Daily storage scenarios
    1. PC plant: Operation at peak load for two hours during the weekday day-time, normal full load for the remaining 14 hours of the day-time and 50% load for 8 hours of night-time and all weekend. Stored solvent is regenerated during the night-time.
    2. NGCC plant: Operation at peak load for two hours during the day-time, normal full load for the remaining 14 hours of the day-time and shut-down during nighttime and weekend. Stored solvent is regenerated during normal day-time operation.
  2. Weekly storage scenarios
    1. PC plant: Operation at peak load for 16 hours during weekdays and operation at 50% load during 8 hours of night-time and all weekend. Stored solvent is regenerated during the night-times and weekend.
    2. NGCC plant: Operation at peak load for 16 hours during weekdays and shutdown or operation at the minimum load required for solvent regeneration during night-time and weekend.

In the weekly scenarios the ‘peak’ times are almost half of the total hours. For the PC plants, if solvent regeneration was completely switched off during peak times in these scenarios the amount of CO2-laden solvent to be stored would be extremely large. Also the regenerator would have to be substantially larger than in the reference plant and it may be difficult to provide sufficient steam for the regenerators during the off-peak times when the plant is operating at 50% part load. In the weekly scenarios assessed in this study the solvent regeneration was therefore reduced by only 25% at peak times. Two alternatives were assessed:

  1. Reduced regenerator size. The regenerator is about 85% of the size in the reference plant, which enables all of the stored solvent to be regenerated during off-peak times
  2. 100% regenerator size. There is no reduction in the size of the regenerator, which would enable the plant to operate for long periods at 100% load if required. To minimise the capacity of the storage tanks the regenerator is operated at full capacity during the weekday night time, and it is operated at lower throughput during the weekends.

The lower capital cost of storage tanks and stored solvent in alternative 2 is greater than the extra cost of a larger regenerator. This lower capital cost and the greater flexibility to operate at full load means that alternative 2 is preferred, so results for this are presented in this overview.

In the NGCC weekly scenario, if solvent regeneration was completely switched off during peak times the amount of CO2-laden solvent to be stored would be extremely large, although less so than in the PC plants because gas fired power plants have lower specific CO2 production. It is possible to store 50% of the solvent during peak times without having to oversize the regenerator. Solvent is regenerated at off-peak time by operating one of the two gas turbines at minimum environmental load. As with the PC plant, the lowest cost and most flexible option is to have a 100% sized regenerator.

In the daily operating scenario, solvent regeneration is shut down completely during the 2 hours of peak operation and all of the CO2–rich solvent produced during this time is stored. In the PC plants the stored solvent is regenerated during the night time when the plant is operating at 50% load. In the NGCC plants the stored solvent is regenerated during the remaining 14 hours of daytime operation, which requires the regenerator to be over-sized by about 14% compared to a capture plant without solvent storage. The NGCC plants shut down overnight and at weekend.

Solvent storage has very little effect of the thermal efficiency except for the NGCC weekly scenario, in which one of the gas turbines has to operate at minimum environmental load at off-peak times to regenerate solvent. The solvent storage tanks are conventional sized tanks as used at oil refineries but they are nevertheless large, particularly in the weekly scenario. As an example, in the NGCC daily scenario four tanks each of which is 27.4m diameter and 12.8m high are required.

Table 3 Storage of post combustion CO2 capture solvent

Power plant type NGCC PC NGCC PC
Storage scenario Weekly Weekly Daily peak Daily peak
Hours per week of peak output 80 80 10 10
Increase in power output at peak times, % 6.2 4.8 12.1 22.2
Thermal efficiency, %
Reference plant efficiency, 100% load 50.6 34.8 50.6 34.8
Reference plant time weighted average efficiency 50.6 33.6 50.6 33.6
Storage plant time weighted average efficiency 45.3 33.5 50.5 33.6
Capital cost
Change in cost per kW of normal output, % +19.6 +6.1 +9.3 +5.8
Change in cost per kW of peak output, % +12.6 +1.2 -2.6 -13.5
Cost of extra peak generation, €/kW 3116 2891 752 589
Solvent storage
Quantity of solvent storage, 103m3 286 199 30 46

The overall economics of solvent storage are complex because there are substantial changes in the electricity output at various different times. An electricity price profile at different times is needed, which is beyond the scope of this study. However, an initial assessment of the economics can be made by comparing the capital cost of solvent storage and alternative means of generating peak load electricity. In the weekly scenario the capital cost per kW of additional peak generation capacity is greater than the cost of the reference power plant, which indicates that this scenario is unlikely to be attractive. In the daily scenario the capital cost per kW of additional peak generation capacity is less than the cost of the reference plant but it is probably higher than the cost of the leading alternative technology for peak load generation, namely simple cycle gas turbines. Solvent storage may be attractive in this scenario, depending on fuel prices, carbon emission costs and the electricity price profile.

Liquid oxygen and air storage

Storage of liquid oxygen (LOX) in oxy-combustion and IGCC plants can provide a boost to the peak power output by reducing the power consumption for oxygen production. During the times of peak power demand the power plant is operated at full load, the air separation unit (ASU) is operated at minimum load and the rest of the oxygen required by the power plant is taken from a LOX store. In the oxy-combustion plant the LOX is vaporised by condensing liquid air which is then stored and in the IGCC plant the stored LOX is vaporised using LP steam. During off-peak times the power plant is operated at part load but the ASU is operated at a higher load to enable the LOX store to be re-filled. Performance and cost data for PC oxy-combustion and IGCC plants with oxygen storage are shown in table 4.

An alternative that was evaluated in the report but which is not shown in this overview involves having a smaller capacity ASU which is operated at constant load. This option would reduce the capital cost and oxygen storage requirement but it would give a smaller boost to the power output at peak times. The plant would also not have the flexibility to operate at full load for long periods of time, similar to the post combustion cases with a reduced size solvent regenerator mentioned earlier.

The minimum efficient turndown of an ASU air compressor is 70% and the minimum turndown of the cold box is around 50%. In IGCC, turndown of the main ASU air compressor to 70% would give only a marginal increase in net peak power output. The ASUs are therefore configured to have two smaller air compressors, one of which is turned off during the time of peak demand and the other is operated at 70% load. Having multiple compressors increases the capital cost but provides greater opportunity for high peak generation. Half of the compressed air for the ASU in the IGCC plants is provided by extraction from the gas turbine, which earlier studies and practical experience has shown results in relatively high efficiency, good operability and low costs. When the power plant is operating at part load, less air is available to the ASU from the gas turbine compressor. To operate the ASU at full load more air has to be provided by the ASU’s own air compressors, so an additional compressor is provided for each ASU.

In the oxy-combustion case shown in table 4 there are two 50% capacity ASUs, each equipped with two 60% capacity main air compressors. During peak times one of the main air compressors per train is turned off but the ASUs are kept in operation because it is not feasible to shut down the ASU cold box due to its long start-up time. In the oxy-combustion plant only liquid oxygen and liquid air need to be stored but in the IGCC plant liquid nitrogen also has to be stored, as nitrogen is required for the gas turbine. Nitrogen accounts for more than half of the total storage volume.

Table 4 Storage of oxygen

Power plant type PC-oxy IGCC PC-oxy IGCC
Storage scenario Weekly Weekly Daily Daily
Hours per week of peak output 80 80 10 10
Power output
Increase in output at peak times, % 5.3 7.7 5.8 10.5
Thermal efficiency, %
Reference plant efficiency, 100% load 35.5 31.4 35.5 31.4
Reference plant time weighted average efficiency 34.0 29.5 34.0 29.5
Storage plan time weighted average efficiency 34.8 30.0 34.3 28.9
Capital cost, €/kW
Change in cost per kW of normal output, % +2.5 +2.7 +0.9 +1.4
Change in cost per kW of peak output, % -1.5 -4.6 -4.6 -8.2
Cost of extra peak generation, €/kW 1573 928 381 336
Storage of liquid oxygen and nitrogen/air
Quantity stored, 103m3 12.1 24.0 0.8 3.4

The volumes of storage are much smaller than in the solvent storage cases but vessels have to operate at cryogenic temperatures.

The capital costs of peak generation are relatively low because unlike the earlier cases no additional power generation equipment has to be installed, instead the increased peak power is achieved by reducing the plant’s ancillary power consumption. Although the capital costs per kW of normal power output increase, the costs per kW of maximum peak output decrease, particularly for the daily storage scenarios. The capital cost of the extra peak generation capacity in the daily storage scenarios is competitive with simple cycle gas turbines and the storage option has the advantage that extra peak generation has low CO2 emissions. This preliminary analysis indicates that oxygen storage should be an attractive option for providing additional peak generation.

Hydrogen-rich gas storage

The flexibility of IGCC plants could be improved by storing surplus hydrogen-rich fuel gas produced during off-peak times. The stored hydrogen could be used to generate electricity at peak times or it could be supplied to other energy consumers. This would have the practical and economic advantages of enabling the gasification plant to continue to operate at full load at all times. The leading option for hydrogen storage would be underground salt caverns, which are a proven and relatively low cost technique for large scale hydrogen storage. Some liquid nitrogen would also be stored to satisfy the needs of the gas turbine. Performance and cost data are given in Table 5. The increase in peak power output per unit of gas turbine capacity is relatively small (3.3%) but the increase per unit of gasification plant capacity is greater (26.0%). The overall capital cost per kW of peak capacity is 8.5% lower than the reference IGCC plant. The capital cost of the extra peak generation capacity is negative because the capital cost of the plant is lower and the peak output is higher, although it should be noted that the plant would be unable to operate at continuous full load because of the under-sized gasification plant.

Table 5 Storage of hydrogen

Power plant type IGCC
Storage scenario Weekly
Hours per week of peak output 80
Increase in power output at peak times, %
Per unit of gasifier capacity 26.0
Per unit of gas turbine capacity 3.3
Thermal efficiency, %
Reference plant efficiency, 100% load 31.4
Reference plant time weighted average efficiency 29.5
Storage plant time weighted average efficiency 29.7
Capital cost, €/kW
Change in cost per kW of normal output, % -5.5
Change in cost per kW of peak output, % -8.5
Cost of extra peak generation, €/kW negative
Storage of hydrogen and nitrogen
Quantity of hydrogen stored, 103m3 working volume 100
Quantity of liquid nitrogen stored, 103m3 7.2

The hydrogen storage volume is relatively small for a typical modern salt cavern store, for example about 5% of the capacity of a hydrogen storage cavern being built in Texas. This study focussed on coping with sort term (up to a week) variability in electricity demand. The relatively low cost of underground hydrogen storage means that this technique could also be cost effective for smoothing out longer term seasonal variability in electricity demand.

Another case was assessed in which the gasification and CCS is operated at continuous full load, a constant flow of high purity hydrogen for other consumers is maintained at all times and some of the hydrogen rich gas from the CCS plant is stored at off-peak times. Details of this case are provided in the main report.

Constant flow of CO2 to transport and storage

Variation of the throughput of a CO2 capture plant would result in variation of the flowrate of CO2 to the transport pipeline and storage site. Little information is currently available on the ability of dense flow pipelines and storage wells to accept variable and intermittent CO2 flows and the effects may be site specific. Two techniques for providing a constant flow of CO2 were assessed, in case this should turn out to be required:

  1. Buffer storage of compressed CO2
  2. Buffer storage of CO2-rich solvent, combined with a reduced solvent regenerator capacity

In Case 1 it was assumed that CO2 would be stored in cylindrical pressure vessels. If longer term storage was required and suitable geology was available near the power plant site it may be worthwhile considering an underground temporary buffer store.

Providing CO2 buffer storage for the NGCC and PC plants with the ‘weekly’ operating scenario described earlier (in the section on solvent storage) would increase the plant capital cost by €30-40/kW. This cost could in principle be offset by a reduction in the size and cost of the CO2 pipeline (and injection wells), for example in the NGCC case the cost savings for a 100km dedicated CO2 pipeline would more than offset the cost of CO2 storage. However if a small pipeline was built the plant would not be able to operate at continuous full load for long periods of time. The modest extra cost of installing a full capacity pipeline may be considered worthwhile to maintain the option to operate the plant at high load factors if required.

Case 2 (reduced capacity solvent regenerator and buffer storage of CO2 capture solvent) was found to be substantially more expensive than Case 1 (storage of compressed CO2).

Expert Review Comments

Comments on the draft report were received from seven reviewers who have expertise in the power industry, oxygen production, IGCC project development, and research on post combustion capture and CCS plant flexibility. IEAGHG and the contractor reviewed the comments and various detailed changes were made to the report. The contribution of the reviewers is gratefully acknowledged.

In general the reviewers thought the report was of a high standard. Some reviewers emphasised that many operational issues still need to be considered in detail and more dynamic modelling and optimisation of the control of power plants and capture units is needed. This was emphasised more in the report.

Some reviewers expressed concerns that the load profiles originally assumed for the flexibility assessments may not be optimum as they resulted in excessive amounts of solvent storage, which raises economic, safety and regulatory concerns. To address these comments, additional cases involving short term peaking operation and substantially lower quantities of solvent storage were evaluated. More part load operation cases were also assessed and the oxy-combustion case with oxygen storage was modified to also include liquid air storage, to address reviewers’ comments.

Conclusions

  • CCS may impose additional constraints on the flexible operation of power plants but in general there are ways of overcoming these limitations. A plant with CO2 capture may even be able to ramp up its net power output more quickly and produce more peak generation than a plant without capture, using the techniques considered in this study.
  • The efficiency penalties for part load operation are expected to be somewhat greater for plants with CO2 capture than plants without capture, for example around 3 percentage points at 50% load for a pulverised coal plant with post combustion capture compared to around 2 percentage points for a plant without capture.
  • Increasing the power output by turning down or turning off the CO2 capture unit may be an attractive technique for short periods, depending on the peak power price and CO2 emission cost but preliminary analysis indicates that simple cycle gas turbines may be a lower cost option for peak load generation. Regulations would need to allow the resulting increase in CO2 emissions, for example by averaging emission performance standards over a long period. Some additional equipment, particularly steam turbine capacity, would have to be installed to obtain the full benefit from turning down or turning off the capture unit, which would increase the capital cost. Turning off capture could increase the net power output by 27% for a pulverised coal fired plant and 16% for a natural gas combined cycle plant.
  • Storing CO2–rich solvent and regenerating it at a later time may be attractive as a way of increasing power plant ramp rates and for increasing the net power output during short term peaks in power demand. However, the large quantity of solvent that would have to be stored would mean that operating at peak output for longer periods of time would not be attractive. Plants could be built with a wide range of storage volumes, solvent regenerator sizes and peak power generation capacities; selecting the optimum would be a difficult commercial decision. Storing solvent could increase the net power output by 22% for a pulverised coal fired plant and 12% for a natural gas combined cycle plant.
  • Liquid oxygen and air/nitrogen could be stored in oxy-combustion and IGCC plants to improve flexibility and increase net peak generation by 5-10%. From an economic perspective this is expected to be a relatively attractive option for short term peak power generation.
  • Hydrogen produced in IGCC plants with pre-combustion capture could be stored for example in underground salt caverns, which are commercially proven. This would enable the gasification and CCS equipment to operate at continuous full load and only the combined cycle plant would need to operate flexibly to cope with variable power demand. This would be a significant practical and economic advantage for non-base load power generation. Underground hydrogen storage would be suitable for longer-term as well as short term storage, which could be an advantage particularly in electricity systems that include large amounts of variable renewable generation.
  • Compressed CO2 could be stored at capture plants to reduce the variability of flows of CO2 to transport and storage, if this is found to be necessary. Buffer storage of CO2 would enable a smaller capacity CO2 pipeline to be built but this would constrain the ability of the power plant to operate at continuous full load, which may not be commercially attractive.

Recommendations

  • IEAGHG should assess the ability of CO2 transport and storage systems to accept variable and intermittent flows of CO2.
  • IEAGHG should undertake further work to determine the requirements for CCS plant flexibility, including collaboration where appropriate with other organisations that are undertaking modelling of electricity systems that include other low CO2 technologies.
  • IEAGHG should validate the methodology and results of this study when further information becomes available from plant dynamic modelling and pilot and demonstration plant operation.
  • IEAGHG should propose further reviews and studies on CCS flexibility when appropriate.