SMART POWER GENERATION – THE ROAD TO THE FUTURE
Jacob Klimstra

Jacob Klimstra Consultancy, Twijzel, The Netherlands

ABSTRACT
The expected addition of a substantial amount of electricity generation capacity based on wind turbines and solar photovoltaics in Turkey makes the crucial task of balancing electricity production with demand increasingly difficult. Energy storage over longer time spans is prohibitively expensive, so that the generators other than those based on wind and solar have to do the balancing.  Smart power generation with fast response to requested output changes is needed to comply with the new boundary conditions. Only smart power generation can enable the introduction of a large amount of renewable electricity while maintaining high fuel efficiency, low emissions and good reliability.

INTRODUCTION

The power sector faces rapidly changing boundary conditions, in Turkey as well as in the rest of the world. Fuel prices show large fluctuations over time, even within a year, with a steadily increasing tendency. This is because a rapidly expanding global economy eagerly needs more of the energy commodities that are scarce by nature. Political turmoil in fuel exporting countries has immediate effects on fuel prices. In addition, there is the uncertainty about the price of greenhouse gas emissions. Yet, the technical life of a power plant easily exceeds 40 years, and many years of operation are needed to recover the connected capital investments.

Decisions made today can result in building power stations that are not competitive anymore after a short time.
Another factor changing the power sector is the large-scale introduction of renewable electricity generation by windmills and solar cells. The output of these renewable sources largely depends upon the weather and the time of the day, and is therefore not dispatchable. This means that other power plants have to compensate for the variability in output of the renewable electricity sources.

Consequently, the ‘other’ power plants will experience more variability in output, which reduces their capacity factor and increases the need for rapid ramping up and down of their power output. This not only increases the specific capital costs of the ‘other’ generators, but also tends to lower their fuel efficiency.
The future power plant portfolio that provides countries with electricity will therefore be different. Flexibility in output, quick starting and stopping without severe impact on maintenance costs and maintenance time, as well as high fuel efficiency in a wide load range is needed. At the same time, the reliability should be excellent since modern society requires a high availability of electricity. Data interchange and manufacturing processes depend on an uninterrupted supply of power. In addition, quick adaptation of the installed power capacity should be possible in order to respond to economic growth. Smart power generation is the ultimate solution for this.

ELECTRICITY DEMAND PATTERNS

Electricity demand patterns during the day, the week and the seasons depend on the local climate, the wealth level and the nature of the local economy. Wealthy countries with hot summers and poorly insulated buildings need much air-conditioning in summer. Summer demand during the daytime can then exceed winter demand by a factor two. For instance, electricity demand in Abu Dhabi peaks at 5.6 GW in August, while maximum demand in January is only 2.5 GW. Per capita use of electricity of 17,000 kWh per year in Abu Dhabi is about a factor 7.5 higher than that in Turkey. Summer daytime use of electricity in Texas, USA, can be twice as high as during the night, again due to the use of air conditioning. Countries with an economy based on heavy industry have a much higher base load for electricity over the year than countries where trade and services dominate the economy.* This work is sponsored by Wärtsilä Power Plants Turkey has a rapidly developing economy, moving substantially in front of countries such as Thailand and Brazil. Like in those other two countries, Turkey’s electricity demand pattern is still relatively smooth over the seasons, indicating a limited use of air conditioning.
Figure 1 gives the absolute peaks in demand for every month in 2009. The highest peak occurred in summer due to cooling by air conditioning, while a slightly lower peak occurs in winter caused by electrical heating and more demand for electric light. Until a few years ago, the maximum peak always occurred in winter: the steadily growing wealth level in Turkey clearly results in the installation of more air conditioning and other day-time electricity users such as computers. It is expected that this trend will continue during the coming decades, resulting in considerably higher peaks in demand during the summer. The ratio over a year between the absolute maximum in daily electricity use and the minimum is close to two in Turkey.
Figure 1:  Absolute peaks in electricity demand in Turkey during the months of 2009 [1]
Figure 2 shows the power demand pattern on May 11, 2010, a month where electricity use in Turkey is relatively low.

The average load is slightly more than 24 GW. According to figure 2, a continuous base load is present of roughly 20 GW. The intermediate load is about 6 GW, and lasts from 9 o’clock until roughly midnight. The morning ramp up to intermediate load amounts to three GW/h in this example. This means that the output of the intermediate load machines has to increase with some 1% per minute, which is doable if properly scheduled. Turkey has quite some hydropower, which is very suitable for fast ramping up and short-term peaking. Three relatively small peaks in electricity demand are present, varying in size between 1 and 2 GW. The absolute peak in demand of 28 GW is already higher than the maximum of 25 GW experienced in May 2009 (see figure 1). This illustrates the steady growth of electricity use in Turkey. Thus far, the dynamics in power demand in Turkey were not excessive so that the current portfolio of generators could handle the process of balancing electricity production and demand.  
Figure 2: Hourly power demand in Turkey on May 11, 2011 [1]

EFFECT OF WIND POWER ON ELECTRICITY SYSTEM DYNAMICS

The total amount of installed wind-turbine power in Turkey is about 1.5 GW now. That is still very small compared to the total generating capacity of 50 GW (April 2011). However, based on an increasing global competition for fuels, concern about greenhouse emissions and a steady expansion of the Turkish economy, the government decided to accelerate the introduction of renewable energy sources. The intention is that Turkey will have 20 GW of wind power by the year 2023. Extrapolating the growth in electricity use of roughly 10% per year might raise base load to some 40 GW by 2023. That means that during periods of favourable winds, wind power can then cover 50% of base load. Favourable winds generally cover large areas [2], so that a reduction in wind speed will considerably affect the output of the wind turbines.
In order to see what effect wind power can have on the dynamics of a country’s electricity production, we will use Denmark as an example. In Denmark, the installed wind power capacity almost equals base load (see figure 3). Time spans with almost no wind are interspersed with occasional high peaks that can cover base load. The capacity factor of wind power in Denmark was 30% during the 24 days’ time span in figure 3. This is about the same capacity factor as predicted for wind power in Turkey. Denmark can only cope with this high variability in wind –based electricity by exporting the high output peaks to its neighbouring countries. Neighbours Norway and Sweden have much hydropower based on high-altitude reservoirs and they use so-called pumped hydro for storage of any excess electricity. Another example of an area with much wind power is South Australia, where wind power covers about 17 % of electricity use in the state. It is characterised by a large variability during the day, with a counterproductive output pattern compared with demand.
South Australia uses inefficient simple-cycle gas turbines to cope with the large and rapid variations needed to cover variability in electricity demand and to compensate for the swings in wind output. If wind has to cover 20% of annual electricity demand in a country, while the capacity factor is 30%, occasional wind power peaks of more than 60% of average electricity demand will occur. Therefore, almost everywhere in the world, substantial wind power capacity imposes an increasing variability on the output of the ‘other’ generators.

Figure 3: Variable wind power in Denmark compared with electricity demand [2]

Turkey has also much hydropower (≈ 14 GW), which might be suitable to compensate for wind variations. However, a large part is based on run-of-the river systems that have no substantial storage facilities. It is never the intention to switch off river-based hydropower in case of high winds. Hydropower based on high-level basins is not yet equipped with pumping systems for acting as a buffer for wind-output variability. This means that in the future, fuel-based power stations in Turkey will apparently experience much more variability in output than traditionally.

EFFECT OF SOLAR PHOTOVOLTAICS ON ELECTRICITY SYSTEM DYNAMICS

Expectations are that the capital investment for photovoltaic generators (PV) for converting solar radiation into electricity will become that low that the resulting electricity costs will soon reach grid parity. Grid parity means that it will become economic for private consumers and commercials to install PV since its electricity costs might equal or even be lower than that from the distribution grid.
Turkey is a country with a high level of annual solar irradiation. However, like in most countries in the world, the daily output of PV in Turkey is much higher in the summer months than in winter, witness the diagram of figure 4. In addition, PV cells have no output during the night when electricity demand is still substantial. Consequently, the installed capacity of the other generators cannot be lowered in case of much PV capacity. Yet, PV lowers the utilisation factor of the other generators and thus ultimately increases the capital costs per kWh.  

Figure 4: Solar irradiation in Turkey, depending upon the month of the year [3]

Germany, an example of a country with much PV, has already close to 15 GW of PV installations. In case of maximum sunshine, the peak PV output there equals some 10% of electricity demand. Germany is therefore an interesting example of what effect PV electricity can have on electricity system dynamics. Figure 5 shows that on a sunny and warm day in the service area of German Transmission System Operator Amprion, PV power nicely helps to cover peak demand caused by air conditioning. Its effect is currently rather a smoothing of the dynamics of the other generators. PV in Germany has an average capacity factor of only 6%. That means that installing 10 times as much PV capacity, then covering 10% of average annual electricity demand in Germany, occasionally results in output peaks of 160% of average electricity demand. Therefore, as soon as 10 times as much PV power has been installed as currently, which is the intention of the policy makers, huge negative effects occur on system dynamics. Base load for the other generators can disappear around noon, while a very fast ramping up of the other generators is required to meet electricity demand when the sun sets. This requires generating capacity that can respond rapidly to changes in requested output.  
Figure 4: Example of the effect of solar PV power on the output of conventional power generators in Germany
Storage systems based on electrochemical batteries to smooth the output peaks of PV systems on a daily basis add roughly 50% to the price of electricity originating from PV. Storage of high outputs of PV electricity in summer by pumped hydro for use in winter, raises the costs of PV electricity to roughly € 400/kWh, which is excessive. Storage of electric energy over a time span of e.g. 10 days will already increase the price of electricity by some 12 €cts/kWh. These extra costs are high compared to the current production price between 6 and 8 €cts/kWh with conventional power plants. Compressed air, another storage technology promoted by policy makers, can only store 29 MJ per m3 of cavern volume at 70 bar. Methane has an energy density of 2.5 GJ/m3 at 70 bar. It is therefore much better to use natural gas and biogas as back up fuels for the variability of wind-based and PV-based generators. It is an illusion to presume that fuel-based generators can be made obsolete in the coming decades. Nevertheless, fuel-based generators need to be very flexible in the future and their utilisation factor will drastically decrease. Smart power generators are needed to comply with the emerging boundary conditions for fuel-based generators.

SMART POWER GENERATION

Smart power generators have much more flexibility than the current portfolio of generators. They must be very fast in ramping up and down to deal with the drastically increased dynamics in balancing of electricity supply and demand. They must be able to run in a wide load range without suffering from poor fuel efficiency. They must be able to undergo frequent starts and stops without experiencing additional wear and tear. Box 1 summarises the properties of smart power generators.
Box 1: Properties of a smart power plant
Steam-based power plants, or power plants which output is determined by a large fraction of steam-based output such as a gas-turbine combined cycle (GTCC), are inherently slower than a generating set based on gas engines. If properly designed, gas-engine-based generators can deliver full output within 5 to 10 minutes. This is illustrated in figure 5.  
Figure 5: Comparison of power output increase of different generating techniques after a 5-day’s stop.
A power plant based on gas engines in a multiple units in parallel configuration (cascading) has full-load fuel efficiency in an output range between 10% and 100%, while its combined reliability is unsurpassable [4]. Moreover, its specific maintenance costs
(€cts/kWh) stay fairly constant over the full output range, in contrast with power plants based on single units. Next to that, a modular desig consisting of multiple identical units in parallel allows an easy expansion of capacity if economic conditions require more output

IN CONCLUSION

Economic and environmental arguments dictate that much more wind-based and solar-based electricity generation capacity will be installed in Turkey, as is also the case in the rest of the world. The variability in output of such renewable electricity sources as well as their season-dependent output complicate the delicate task of balancing electricity supply and demand. Therefore, the portfolio of generators other than wind and solar needs more flexible generation capacity than what traditional power plants can provide. Smart power generation based on modern gas engines with high fuel flexibility and fast output response offers the required properties. With a sufficient number of smart power plants in a generating portfolio, an undisturbed supply of electricity can be guaranteed, even with much renewable electricity generation.  

REFERENCES

1.    www.eias.gov.tr
2.    Klimstra, Jacob and Markus Hotakainen, ‘Smart Power Generation’, Avain, Helsinki, 2011, ISBN 978 951 692 846 6
3.    www.eie.gov.tr
4.    Klimstra, Jacob, ‘Cascading: a new approach to electricity generation in the distribution system with maximum efficiency and availability as well as minimum costs’, 2006 China International Conference on Electricity Distribution’,  (CICED 2006), (CP527), Beijing, China, 17-20 Sept. 2006, ISBN 0 86341 638 1

SUMMARY

Turkey will rapidly increase its wind-turbine capacity to a scheduled 20 GW by 2023. Most probably, also a considerable amount of solar photovoltaic capacity will have been installed by then. The variable power output of the two renewable generating technologies results in a more complicated task of balancing electricity production with demand. Wind power in Turkey might have an average capacity factor of 30%, meaning that covering 20% of annual electricity demand by wind can result in occasional peaks of 60% of average electricity demand. That can push a considerable amount of the other generators from the grid. Solar PV may even have a capacity factor of less than 15%. That means that if 10% of annual electricity demand has to be covered by PV electricity, again peaks of 60% of average demand can occur. The peaks from wind and PV can also coincide. During the night, solar PV produces nothing.
Therefore, other generators will have to be much more flexible, with higher ramping up and down of their output. Traditional power plants will then suffer from low fuel efficiency and increasing wear and tear. Both effects result in higher fuel costs and higher maintenance costs per kWh produced. Next to that, their utilisation factor decreases resulting in higher capital costs per kWh. Smart power generation can guarantee proper balancing of electricity production and demand at the lowest possible electricity costs. Power plants based on multiple generating units in parallel driven by modern gas engines operating in cascade mode offer full-load fuel efficiency in a power range between 10 and 100% of output. Ramping up and down of output is unsurpassable fast. Frequent starting and stopping hardly affects maintenance requirement and maintenance cost, since the engine parts are not subjected to fatigue, cavitation and creep. Moreover, the combined reliability of such a power plant is very high. In addition, increasing and even decreasing the installed power capacity is simple because of the modular character of the individual generating units and their relatively short building time.