A better electricity system for the developing world

Conventional, centralized electricity networks are the norm in the developed world. But, as Christoph Vitzthum argues, this model is now rather outdated for developed countries ” and quite inappropriate for developing countries.

Christoph Vitzthum

The typical challenge in the developing world is lack of everything, including electricity. When striving to feed the increasing power demand to enable social reform and industrial growth, local decision makers face the question of which route to take. Shall we copy the model of the rich countries, or could there be another, maybe better way? The need to reduce CO2 emissions presents a new, additional challenge, difficult even for the richest of nations.


The present energy infrastructure of the developed countries was mainly created during the monopolistic utility era of the past. The utilities had the power to decide what kind of capacity to construct and how to construct the grid. There was practically no competition. The main challenge for many utilities was to get construction permits for building new generation capacity ” the construction itself was practically risk free as they could turn their cost structure into a solid power tariff. Capital was easily available and cost competitiveness and overall system cost optimization were not the main concerns.

Engine room at the watrsila-supplied on-site power plant at the Barrick gold mine in the US
Click here to enlarge image

Most western utilities had very detailed planning for all their actions. They believed strongly ” and often still do ” in the ‘economies of scale’ principle, which basically means the bigger, the better. The bigger the plant, the better the efficiency and the lower the specific cost. The higher the grid voltage, the lower the losses. They constructed large baseload coal, hydro, nuclear and gas combined cycle plants by the shoreline and built a strong high voltage grid all over the country to ‘evacuate’ the power. The typical outcome was a rigid, over dimensioned, albeit reliable, electricity system.

Today the market situation and rules have permanently changed. Progressive modern utilities have left the past behind and are striving towards a modern, competitive energy system. However, one remainder of the past, still to some extent maintaining the economies of scale thinking, is to look at and calculate power plant and grid investments separately, as if they had nothing, or very little, to do with each other. Quite often the ‘evacuation’ of power from a new, remote, large power plant to the consumers requires new grid investments, which could in many cases, at least partly, be avoided by optimizing the whole system and not trying to produce all power, included the short time peaking and power reserves, remotely at baseload power plants.

Installing flexible, high-efficiency peaking and grid stability generation capacity closer to the load pockets of cities and industrial areas would be a natural way to go, but it does not fit the model of old, large-scale utility thinking.


A typical electricity system built by a rich country monopoly utility ‘wasted’ capital in the following ways:

  • Excessive baseload capacity was constructed. This was possible as there was an ensured return on asset investments.
  • The grid has been sized to transmit the full peak power from large remote power plants to the consumption centres in cities.
  • Efficient peaking capacity was not constructed, instead the so-called peaking plants were typically based on largest possible simple cycle industrial gas turbines located in critical points in the grid. This capacity functioned mainly as an emergency reserve at grid nodes and was hardly ever used as it has such poor heat rate and relatively long starting time. (Of course this is partly because flexible power plants were not available on the market in the past.)
  • Large, steam-fired power plants, running on part load, were used for frequency and load control (unless hydro power is available) and fast starting dynamic reserves were not built.

Due to this, the western power systems can be quite vulnerable when excessive amounts of wind power come into the system. It is difficult to regulate quickly with nuclear and coal-fired steam power plants along the constantly changing, non-dispatchable wind conditions, but the problem has been solved with the very strong grid which can transmit the wind power over long distances to remote countries.

When a developing country looks at the model developed by the utilities during the monopoly era, they see the reliability as a positive thing but they can hardly afford the luxury of the ‘wastes’ described above.


Could the developing countries bypass some of the challenges that the western systems are facing in the changed market conditions of today? Indeed they can!

Baseload capacity and the main high-voltage grid

Baseload capacity strives to use cheap and readily accessible fuels, and typically has a relatively high cost per MW of installed capacity. Hydro, nuclear and coal-fired plants are widely used in stronger grids to produce the baseload. This kind of capacity, whatever the plant type and fuel, should optimally be constructed to operate on full load most of the time i.e. the load of the main grid should be higher than the baseload capacity most of the time ” see Figure 1. Having such expensive power plants standing by or operating on part load is a ‘waste’.

Figure 1. Load duration curve showing the base, intermediate and peaking load segments, and the necessary contingency reserve. Note: The capacity between the installed capacity line and the load curve is standing by!
Click here to enlarge image

The daily load variations are often quite large, especially in cities with offices and other similar consumption that sleeps at night ” see Figure 2. The load reaches a minimum at night and a peak some time in the early evening, or maybe there are two peaks, a day peak and an evening peak. If we want to produce all the energy needed during the peaks in the large baseload power plants, the baseload plant capacity has to cover the whole load curve ” and a large portion of it needs to be standby during the night as there is no load for it. Then the main high-voltage grid also needs to be sized for the full grid load, capable of reaching all consumption points with adequate capacity.

Figure 2. Typical (USA) daily 24-hour load curve
Click here to enlarge image

The past ‘truth’ used to be that this is what a grid should look like. But when looking at the whole system from a modern economical point of view, it is not the best option. The baseload capacity should be designed to match the part of the load curve that is ‘ON’ most of the time. In Figure 2 this is about 31,000 MW, with a growth and contingency reserve. Some of the baseload plants may be located inside the load pockets (cities) thereby further reducing the main grid capacity needs.

Click here to enlarge image

The main grid needs to be capable of transmitting the baseload portion of the country ” or control area ” load curve; not more. The main grid becomes clearly less expensive than the traditional ‘full 52,000 MW grid’. It connects the baseload power plants to the main consumption centres. The past truths say that this sacrifices system reliability, but as modern local decentralized power plants offer similar reliability and availability as the large remote ones, this is no longer true.

Intermediate and peak load capacity

The capacity for the load range above 31,000 MW in the graph (up to the maximum load of 52,000 MW plus reserve) needs to start and stop daily, and should be located close to the consumption centres, thereby relieving the main grid from transmitting these shorter-term peaks. Steam power plants are widely used for this today as they exist in large quantities in developed countries, but their several hour starting times and slow ramp-up and down characters are hardly optimal for this load segment.

The fuel and technology of this local generation capacity needs to be selected based on access to fuels, the size of consumption centre load, and the form of the load curve. The capacity needs to be able to regulate the output in a relatively wide range, keeping high part-load generation efficiency. In an optimum case the capacity should consist of several units; the smaller ones may be connected directly to the medium voltage system.

Increasing wind power capacity puts more demands on the load changing capability of the generation capacity. When wind power increases or decreases its load continuously, and sometimes quite rapidly and unpredictably, this capacity needs to be able to follow the changes within minutes.

Grid stability capacity

Every grid needs reserves to keep the system stable under changing load conditions, and in contingency situations when something goes wrong (plant shutdown, grid failure etc.). This capacity needs to be dynamic, capable of starting and stopping very fast, able to vary the load in a wide range continuously, and to operate with good efficiency on any load for extended periods. These plants need to be located in the vicinity of critical consumption centres, and on grid nodes.

The more wind power capacity in a grid, the more this kind of fast dynamic capacity will be needed to keep the system operational, even during the most extreme weather changes.


As the market conditions change ” fuel prices, load growth, emerging CO2 costs, non-dispatchable wind power entering the grid sooner or later ” the future looks a bit more hazy. In such market conditions power generation assets with flexibility of fuels, operation characters and modes would be highly appreciated.

Arkay on-site power plant in India
Click here to enlarge image

Among the rich nations this has already been understood, especially in the US where most new generation projects today are for flexible multi-purpose capacity that can provide baseload, intermediate load, peaking and grid stability capabilities, all from the same plant. Such capacity was not constructed during the rigid old utility era while there is more than enough baseload capacity already.


Utilities in the developing world face a major challenge in developing their electricity systems. Many active sales agents come to them selling equipment for the ‘western’ concept, and help them to get financing for it. Sometimes it may be difficult for the developing countries to see that in fact they have the relative luxury of going directly for the optimum and bypassing some of the flaws of the western model, which was created under very different competitive terms and market conditions from those prevailing today in the world.

The parameters to optimize at the same time are:

  • access to power for the whole nation
  • reliability of supply
  • ensured access to fuels and fuel flexibility
  • economical competitiveness i.e. cost effectiveness, elimination of ‘wastes’
  • system flexibility and preparedness for load growth.

One of the often heard arguments for the centralized utility model has been that such a system is stable and manageable, and that inserting many smaller generation units into the grid would make it instable and repeatedly cause failures. Over the years this ‘truth’ has been invalidated in many countries, such as Denmark, where a large number of small power plants (including gas CHP and wind) have been added to the grid by developing more balanced tariffs that make them competitive. Modern automation offers the technology to manage grids with multiple generation assets, even very small ones. There is really no reason to think that a more distributed power generation model would not work in practice ” it just needs to be well engineered and controlled.

Another development is that today there are multi-purpose power plants in the market, capable of operating continuously on baseload or part load, with high efficiency, but also starting and stopping rapidly, and ramping up and down continuously without any degradation or increase in maintenance costs. These power plants can operate initially on heavy fuel oil (which in many developing countries still is the only practical, accessible and decently priced power generation fuel), but preferably on gas as soon as natural gas infrastructure in installed. They offer a huge opportunity for the developing world as they not only can function as base load plants, but also as the local intermediate, peaking and grid stability plants of the more optimized system of the future.

It is quite obvious that one needs to visualize the desired end state before one can see which steps need to be taken in order to move to the right direction. Developing a software model of the complete electricity system with present and projected power plant capacity and the whole grid is possible today. With such a model one can compare all the parameters listed above, add, remove and modify power plants and grid connections in order to work out the optimum. I would bet that there will be many surprises when developing such a model, and the end result might well be that going for a copy of the western waste model is not the optimum approach.

The developing world has a unique opportunity to look for a new, more modern and thereby competitive solution for its power infrastructure. It needs to have the courage to challenge the past truths and look at the whole system, despite the external pressure its countries may face, trying to push them to repeat what has been done before. Modern western utilities are searching for the same optimum, but from a very different starting point.

Click here to enlarge image

Christoph Vitzthum is Group Vice President, Power Plants with Wartsila, Helsinki, Finland and the Chairman of the World Alliance for Decentralized Energy (WADE).
e-mail: christoph.vitzthum@wartsila.com

No posts to display