Flexible backup holds the key to Brazil’s energy quandary

The best way for the world to achieve carbon neutrality, or net zero, by 2050 is through the integration of balancing plants – such as flexible engine-based thermal plants – alongside renewable energy and energy storage.

That is the conclusion of a study released on Tuesday by Wärtsilä, which warns of the risks that extreme weather conditions pose to Brazil's electricity system, such as the historic drought that affected the country this year. 

With the growth of its electricity system primarily driven by wind and solar sources, and no prospects for building new hydroelectric reservoirs, Brazil has increasingly relied on thermoelectric plants as a backup.

In 2021 and 2022, the country held two backup capacity auctions aimed at contracting flexible plants – facilities that can be quickly turned on and off to meet system demands.

A new auction, delayed several times, is now scheduled for 2025. However, its assumptions should be revised, argues Wärtsilä's project sales director, Gabriel Cavados, in this interview with BNamericas.

According to Cavados, the government is overly optimistic about how often the contracted thermal plants will be used, leading it to favor technologies that, while cheaper to procure, are more expensive to operate.

BNamericas: How do you see the situation in the Brazilian electricity sector, given the growing participation of renewable sources? 

Cavados: Our expansion is mainly due to wind and solar generation, and we’re not expanding the hydroelectric reservoirs. We therefore need to contract an adequate amount of flexible generation. 

We don't need thermal MWh, except in periods of water scarcity. In the cost of thermal generation, 85% corresponds to generation and O&M [operation and maintenance], the rest is capacity. 

So this 'deadweight' capacity doesn't affect the consumer's pocket. You pay the insurance premium and hope you don't use it. We need to contract capacity that will adequately meet the system's needs in the future. There's no point in hiring coal-fired thermal power for capacity, because it's a slow technology, designed to be used at the base of the system. 

If we have a drought in Brazil every 10 years, the thermal capacity that is contracted will operate for three or four months in a row. So it's not just to integrate the intermittency of wind and solar. It will provide supply when there’s no water in the reservoirs. 

BNamericas: The government has postponed the backup capacity auction until 2025. Are changes to the rules necessary?

Cavados: The variable portion of the plant's cost [i.e. generation cost] is measured at 120 hours per year, giving the impression that the plants will be used very little. We think this is a bit of a mistake. When you don't need to use the plant, it will be really cheap. But if you do need to use it, the bill will be much more expensive. It would be more prudent to consider a slightly higher utilization factor. But if the wrong technology is hired, when the system comes under stress, the cost will be three times higher than it could have been. 

BNamericas: The auction notice hasn't yet been issued. What was the basis of your analysis?

Cavados: I based my analysis on the formula used in the 2021 auction, which assumes 120 hours – about 1.5% of the total hours in a year – as the capacity utilization factor.

There’s a technical note from [federal energy research company] EPE explaining this formula. It highlights that the 120 hours are only an estimate and clarifies that thermal plants must be available at all times for hours as required by the ONS [national grid operator]. Our concern is that when thermal plants are dispatched, the cost of energy will be higher than it could have been.

Among the technologies capable of providing flexible capacity, such as open-cycle gas-fired power plants, we have high-revolution engines, industrial turbines, aeroderivative turbines and internal combustion engines, which is our focus. Each has its pros and cons. Some technologies are cheaper to procure but have very high variable costs when in operation, while others are more expensive upfront but cheaper to run.

I consider high-speed and industrial turbines to be “low-quality” technologies because they offer lower capacity costs but lead to higher costs when the plant is used. Aeroderivative turbines and medium-speed engines are slightly more expensive but much more efficient, with lower CVUs [unit variable costs].

The risk is that if we assume a very low number of operating hours, we’ll end up contracting the cheaper, less efficient technologies.

Our challenge is to build a robust system that can integrate large amounts of renewable energy while reducing costs. The solution is to invest in flexible thermal capacity using natural gas, which has lower carbon emissions and can eventually transition to neutral fuels. Our role is to chart the optimal path for integrating increasingly intermittent energy sources.

I’m not criticizing Brazil’s electricity system. On the contrary, it’s a global benchmark. But we need to constantly revisit the old rules because we’re in the midst of a transition. If we don’t, we risk penalizing the system in the future.

Editor's note: Last week, the ONS received offers from seven thermoelectric plants with flexible conditions to meet peak load demand. The Nova Venécia 2, Maranhão IV, Maranhão V, Parnaíba IV, Ibirité, Três Lagoas and Porto Pecém I plants will now operate more dynamically, adjusting their production according to the variation in demand throughout the day. 

The move, established by an ordinance issued by the mines and energy ministry, contributes, according to the government, to reducing generation costs and optimizing the use of water resources for hydroelectric plants.