Ethiopia’s Biggest Dam Oversized, Experts Say

Thursday, September 5, 2013

In May, Ethiopia diverted the Blue Nile to begin building its largest dam project to date, the 6,000 MW Grand Ethiopia Renaissance Dam (GERD) – a move that angered Egypt, which fears its water supply will shrink over the many years it will take to fill the huge reservoir. Besides the tensions this huge project is causing politically, there is growing concern that the dam will not produce nearly as much power as it has been designed to. A number of engineers have questioned the dam’s design. Asfaw Beyene, a Professor of Mechanical Engineering and Director of the Center for Renewable Energy and Energy Efficiency at San Diego State University, has written a technical article about the dam being over-sized; here he answers questions about this issue. 

WRR:What would be the consequences if the dam is “over-sized”?

AB: It means that more than half of the turbines will be rarely used. It is like buying a 10-story building for a personal residence. You may fill it a few times a year when you have enough guests, but the rooms will be unused most of the time. The dam height and the natural flow rate of the water are the factors that fix the potential power output. The GERD’s available power output, based on the mean flow rate (the average of river flow throughout the year) and the dam height (145 meters), is about 2,000 MW. There is little doubt that the system has been designed for near-peak flow rate, but that high flow only happens during the 2-3 months of the rainy season. The planned 17 turbines are in excess of what can be produced given the dam height and the river flow rate. Targeting near peak or peak flow rate makes no economic sense. 

Engineers use a calculation called “plant load factor” to describe the ratio of a power plant’s actual output over a period of time, to its potential output if it were possible for it to operate at full capacity indefinitely. In the case of GERD, the load factor for the dam designed to produce 6,000 MW would be about 30%. If it were “right-sized” to 2,000 MW, its load factor would be about 90%.

WRR: What questionable assumptions do you think the GRD engineers made?

AB: I think the dam is sized for the peak flow rate of the river, which lasts just a few months. The peak flow rate of Blue Nile is under 6,000 cubic meters per second (mcs), even exceeding 6,500 mcs once in a while. With 145 meters of dam height, this peak flow can produce about 7,000 MW. The average flow rate of Blue Nile is reported to be much lower. So, given the height of the dam and the flow rate, there is no way the dam can produce 6,000 MW for more than 3 months of the year even if the dam stored the difference between peak-flow and design-flow rates. The only scenario under which the power output will be annually consistent is if the hydroelectric dam is designed for a mean flow, which is about 1,456 mcs. This will provide just less than 2,100 MW.

WRR: What are the economic implications for the dam producing so much less power than it is supposed to, if your predictions are correct about the dam being designed for an overly optimistic river flow?

AB: It is simple: the extra 10 or so turbines will be parked for about 9 months of the year. The size calls for about 7 turbines with 350 MW each. Even if we add one extra turbine for maintenance downtime, the appropriate design target should not exceed 2,800 MW. This assures year-round supply of electricity at almost constant level, also requiring a shorter period for initial reservoir filling. The total price at $800/kW rate (I get this by dividing $4.7 billion by 6000 MW, which is a common approach in the power industry) will be about $2.3 billion dollars – much less than the $4.7 billion for the 6,000 MW. 

What does this mean in human terms? According to the World Bank, Ethiopians use on average of about 200 kWh of electricity per capita per year. A per capita comparison is however less than useful because it shifts with population growth. A better comparison is kilowatt-hours used per household per year, which is about 500 kWh for Sub-Saharan Africa. (For comparison’s sake, the global baseline is around 13,000 kWh/year, and the average US household uses 18,000 kWh per year, including natural gas and electric.) If we assume 500 kWh/year per household, the 4,000 MW of “missing power” could have covered more than 70 million households (not including the cost of transmission lines). If we take a South African household average of 5,000 kWh/year, it could affect cover 7 million households. 

WRR: Do you have a recommendation for Ethiopia in this case?

AB: It is clear that the issue is highly politicized, and the politics seems to suppress legitimate engineering inputs and environmental discussions. My suggestion to the concerned authorities is to make the matter transparent, rethink the number of turbines that are to be installed, and resize the hydroelectric power output by reducing the number of turbines. A few weeks ago I visited a hydropower plant near La Serena, Chile that was turned off because of the drop in water level. The engineers there regretted that they didn’t size the turbines for a much smaller head. The GERD faces the same fate unless the dam’s sizing is corrected. 


Where: Blue Nile, near Sudan border

Dam size: 145m, 1,708m long 

Reservoir size: 1,680 sq km, will hold about 70 bn cubic meters of water (larger than Ethiopia’s largest natural lake)

Resettlement: 20,000 people

Dam Cost: US$4.8bn (equal to about 15% of Ethiopia’s GDP in 2012, and about 60% of the annual budget)