2016_1 Maximising profits through efficient pulse load operation master

Maximising profits through efficient pulse load operation

Over the past few years, the share of wind and solar power in many grids has increased rapidly all over the globe. Nowadays, a sizeable proportion of all installed capacity is made up of intermittent renewable energy sources, a fact that introduces a previously unheard of degree of uncertainty into the power systems.

Text: CHRISTIAN HULTHOLM & JAIME LÓPEZ Photo:

Over the past few years, the share of wind and solar power in many grids has increased rapidly all over the globe. Nowadays, a sizeable proportion of all installed capacity is made up of intermittent renewable energy sources, a fact that introduces a previously unheard of degree of uncertainty into the power systems. This makes it increasingly difficult to ensure the essential balance between supply and demand in the power systems of the 21st century. It is clear that, in order to secure system stability, flexibility will be needed more than ever. Since intermittent renewable power sources are incapable of providing said flexibility in the form of on-demand power, future thermal generation will need to bear the burden and be as flexible as possible. One key characteristic of a flexible plant is the ability to ramp up and down quickly, fitting a demanded production time bracket that we refer to as a pulse.

This article is based on the award-winning paper (Highly Commended 2015 in the track “Power Plant Technologies”): “Maximising profits through efficient pulse load operation” presented at the Power-Gen Asia 2015 conference.

Dealing with pulses 

These pulses require the power generation fleet to adapt to them, by raising output for a set period of time and returning to the previous level after the pulse is over. The backbone of most modern power systems is made of so-called baseload technologies: those that cannot easily adapt their output in a short time span or that suffer high efficiency losses if they do so. For example, nuclear power plants and coal power plants, as part of the baseload fleet, are ill-suited to provide on-demand flexible generation so they are not considered in this study.

Gas-fired power, due to its superior ramping capabilities compared to baseload technology, is generally used to supply the needed power for these pulses. However, combined cycle gas turbines (CCGTs) are the most widespread solution of that kind. Although able to ramp up and down in relatively short periods of time, they are still better suited for dealing with stable loads due to their heavy derating under variable load conditions. Since CCGTs are currently the state-of-the-art technology used to deal with load pulses, they constitute the benchmark for the solution we propose in this paper.

Next, a number of different cases of pulse load operation will be reviewed. The considered time frames for the pulses are four hours, eight hours and 14 hours. The plant sizes taken into consideration are 100 MW and 400 MW, representing two of the most usual classes in current power systems.

2016_1 Maximising profits through efficient pulse load operation 2
Fig. 1 - Examples of different pulse loads in a typical daily demand curve
2016_1 Maximising profits through efficient pulse load operation 3
Fig. 2 - 4-hour pulse, ramp-up and -down required by gas-fired technologies under analysis.

Case study: 100 MW power plants

Introduction

The reasoning behind the choice of these concrete examples is their real-life relevance. As it can be seen in Figure 1, the 4-, 8- and 14-hour pulses can be found as recurring patterns in daily demand curves of highly industrialised power systems. Although this concrete figure corresponds to the Japanese daily demand, similar patterns are commonplace in most US power systems.

Assumptions

The operational costs considered include the total fuel costs, as well as the total variable operations and maintenance costs (VOM). Both are calculated for the duration of the startup process, the actual, required period of generation, as well as the shutdown time. Additionally, the cost impact on maintenance from the start-up, i.e. the equivalent operating hours (EOH), is also taken into consideration.

The considered technologies are all gas-fired combined cycle power plants, equipped with air-cooled condensers. The comparison is made between state-of-the-art technology based on internal combustion engines (CC ICE) and gas turbines (CCGT). The former technology is represented by Wärtsilä’s Flexicycle solution, whereas the latter is based on a Frame 6 solution (100 MW case) and on a Frame 7 solution (400 MW case). The performance data has been obtained from the latest available versions of engineering calculation software, namely GT Pro for the combined cycle gas turbines and PerfPro for the internal combustion engines.

The fuel price taken into consideration is the harshest in the region, signifying the most favourable condition for CCGTs, due to their higher nominal baseload efficiency. For that matter, we have chosen the Japan LNG import spot price as of May 2015, averaging at approximately 12.13 USD/MMBTU (41.39 USD/MWh).

2016_1 Maximising profits through efficient pulse load operation 4
Formula
2016_1 Maximising profits through efficient pulse load operation 5
Table 1 - Performance comparison of 100 MW power plants.

4-hour pulse

A typical CCGT power plant has a start-up time of approximately 60 minutes. Hence, the CCGT plant needs to receive the start command one hour before the actual pulse takes place, a situation that in many cases may prove difficult. Moreover, the energy produced during this start-up hour is typically not reimbursed. The same applies for the shut-down time, which is typically 30 minutes. 

On the other hand, a typical ICE in single cycle (SC) mode can be started in only 5 minutes. As soon as the combined cycle equipment is ready, typically in approximately 50 min, the loop can be closed, all while the plant is operating continuously. When it comes to shut-down, the combined cycle ICE plant is able to unload in 20 minutes (1 min when in SC mode).

Hence, the amount of time the ICE plant is run outside the required settlement periods, as defined by the electricity markets, can be minimised efficiently. In other words, the amount of consumed fuel and accumulated running hours can be kept to a minimum.

 Next, the total operating costs of 100 MW power plants during a four-hour pulse will be analysed. As seen in Figure 3, at first, the impact of the start-up costs of the CCGT is considerable. Second, when adding the fuel and VOM costs for the one-hour start-up period, the accumulated costs already exceed USD 13,000 at the time when the settlement period begins. Over the four-hour pulse, the higher nominal efficiency of the CCGT only marginally compensates for these initial costs, compared to the CC ICE. When also considering the costs associated with the shut-down time, the total difference over the full pulse is over USD 10,200, in favour of the CC ICE.

When looking at the four-hour pulse in terms of efficiency, the energy produced during the start-up and shut-down periods must not be taken into account since it is not reimbursable. This energy production is only a by-product of the pulse operation and is not of commercial use. Hence, we will define the overall pulse efficiency as in formula (see separately).

Where we have defined tp = (4, 8, 14) h, depending on the length of the pulse under analysis in each case. Each of the integrals account for the energy produced in one of three periods of time: before the actual delivery of the energy pulse (t<0), during the pulse (0<t<tp) and after the pulse is over (t>tp). Since the early start-up and the delayed shut-down are needed, yet do not provide any monetary value being outside of the contracted load pulse, we will discount the energy produced during said periods by subtracting the integrals that quantify out-of-pulse energy delivery. Predictably, the closer the generation technology is able to match the shape of the demand pulse, the less the energy waste. This equates to an interesting balance between baseload efficiency and operational flexibility, which our overall pulse efficiency (η pT) quantifies. 

Based on this calculation and the previous assumptions, we find that the overall pulse efficiency of the CCGT is 41.0% vs. 46.0% for the CC ICE. We can conclude that, in the case of a 4-hour pulse, the far superior operational flexibility of the internal combustion engines outweighs the higher baseload efficiency of the CCGT.

The second part of our analysis focuses on the economic side of the issue. For energy production in current-day competitive markets to be profitable for asset owners, ensuring that the most efficient solution ‘on paper’ also makes economic sense, is a must.

As we can see in Table 1, the CCGT solution incurs hefty start-up costs, due to the use of start-up fuel and the greatly increased wear and tear of the machinery during the start-up phase, accounted for in the so-called equivalent operating hours (EOH) calculation. This start-up cost puts the CCGT at a disadvantage compared to the combustion engines, which do not suffer any increased wear and tear nor require extra fuel for starting up. Also, being internal combustion engines capable of starting just five minutes before the beginning of the pulse, the amount of fuel spent in generating non-productive (i.e. out-of-pulse) energy is negligible compared to that of the CCGT solution.

The results of the economic analysis are summed up in Figure 3. We can see that for the case of a 4-hour pulse, the steep start-up cost of the CCGT and its need for one hour of out-of-pulse production cannot be compensated by means of its superior baseload efficiency. Hence, for the typical pulse length of four hours, which covers most morning and evening peaks in developed countries, we conclude that the CC ICE solution is not only technically superior (i.e. it yields a higher overall pulse efficiency), but it is also economically superior by means of a reduced operation cost compared to that of a CCGT.

2016_1 Maximising profits through efficient pulse load operation 6
Fig. 3 - 4-hour pulse production – fuel, O&M and start-up costs (100 MW power plants).
2016_1 Maximising profits through efficient pulse load operation 7
Fig. 4 - Pulse efficiencies (including fuel, O&M and start-up costs) for 100 MW power plants.

8-hour and 14-hour pulse

Similarly, eight- and 14-hour pulses have been analysed for the 100 MW power plants. A summary of the overall pulse efficiencies for these periods is presented in Figure 4. Hence, not even during a 14-hour pulse is the higher nominal efficiency of the CCGT sufficient to compensate for the energy waste resulting from a slower start-up and shut-down; the CC ICE solution is still a more efficient option.

Case study: 400 MW power plants

In a similar fashion to that used for analysing 100 MW power plant solutions in the previous chapter, we now present a comparison of the pulse load efficiencies of a Frame 7-based CCGT solution and a scaled-up version of the CC ICE solution based on Wärtsilä’s Flexicycle technology. Table 2 provides a comparison of the technical performance of these solutions.

4-hour, 8-hour and 14-hour pulses

An overview of the calculated pulse efficiencies for the 400 MW CCGT and CC ICE solutions is presented in Figure 5.

It is in this case, where larger power plants are evaluated, that the results turn out most interestingly. During the 4-hour pulse, the overall efficiency is almost three percentage points higher for the CC ICE solution. However, we can observe a near break-even of the efficiencies occurring briefly before 8 hours.

For lengthier pulses, the CCGT becomes the more efficient option, surpassing the CC ICE by a slim margin. However, it is important to notice that even though the pulse efficiency of the CCGT is slightly higher than that of the CC ICE, the total operational costs are still higher. 

In the 8-hour case, although the overall pulse efficiency of the 400 MW CCGT is 0.3 percentage points higher than that of the CC ICE, the overall cost generated by the CCGT is still 8.1% higher, which is a very remarkable difference. Even in the 14-hour case, where the CCGT, best suited for baseload operation, clearly wins in terms of overall pulse efficiency, the CC ICE retains a 1% cost advantage.

2016_1 Maximising profits through efficient pulse load operation 8
Table 2 - Performance comparison of 400 MW power plants.
2016_1 Maximising profits through efficient pulse load operation 9
Fig. 5 - Pulse efficiencies (including fuel, O&M and start-up costs) for 400 MW power plants.
2016_1 unlocking optimum CHP performance 4
Table 3 - Summary of the completed modelling.

Conclusions

As a result of our technical and economic analysis, we can conclude that the shorter a single pulse of power generation is, the greater the importance of reaching full load quickly. This emphasis on flexibility easily outweighs a higher baseload efficiency, and, therefore, the evaluation criteria in power plant projects must be rethought. In summary, we conclude the following: 

  • For short and medium pulses, 4 and 8 hours, an internal combustion engine-fired combined cycle (such as Wärtsilä´s Flexicycle solution) is a more competitive option than both small-scale (100 MW) and large-scale (400 MW) CCGTs.
  • Even though the 400 MW CCGT has a higher average efficiency than Flexicycle during an 8-hour pulse, its total operational costs are higher (due to the start-up costs and increased maintenance needs).
  • For long pulses, 14 hours, CC ICE is a more competitive option than small-scale (100 MW) CCGTs.
  • Even though the 400 MW CCGT has considerably higher average efficiency than CC ICE during a 14-hour pulse, its total operational costs are still higher than that of the internal combustion engine-based solution.

Table 3 serves as a summary of the completed modelling and presents the best solution in both technical (overall pulse efficiency) and financial (overall cost) terms, for each of the scenarios under analysis.

 

White paper

Flexicycle power plants

Flexicycle power plants 2015 PDF

Leave a comment

Load more comments

Related articles