18 May 2011
Grid operators are always busy forecasting and matching the supply from the generators to the demand from consumers. Over a day, the demand profile of a typical city grid goes from a low base load (mostly at night time) to a high peak demand (mostly during office hours). Supply is managed by “dispatching” generating assets i.e., making them run harder during peak load demand.
When asked about solar energy integration into the grid, they shared that their number one concern about intermittent sources such as solar energy is that it’s not dependable during peak load conditions.
The production of electricity from solar sources depends on the amount of light energy in a given location and point of time. Solar output varies throughout the day and through the seasons, affected by factors such as the cloud cover. When a small amount of solar generators are connected to the grid, the grid operator can manage the variations in one of the following traditional ways:
- Automatic generation control (AGC) or frequency regulation to respond to variation on the order of seconds to a few minutes;
- Activation of “spinning reserves” to respond to variation on the order of minutes to an hour;
- Activation of peak-power generation (usually referred to as reserve margin capacity) to respond to hourly variation.
However, grid operators worry when solar energy forms a significant portion of the generation mix (say 10-30%). The concern is that such a scenario is unmanageable due to the fact that spinning reserves and peak power generation capacity may not be sufficiently available. The conclusion by some grid operators is that solar energy cannot really provide major share of electricity generation, which could be bad news for grid-connected solar. This leads to the question of whether there is an upper limit on grid-connected solar systems that is actually much smaller (e.g., <10%) than what many forecast today?
Mark A. Delucchi and Mark Z. Jacobson report that there are several ways to alleviate the above concern of grid operators.
1. System design approach
One of the ways to address this issue would be to interconnect several geographically-dispersed systems, which smoothes out electricity supply by cancelling the effects of intermittency in each. This works over a large geographical area with different climates. The other way is to use complementary and non-variable energy sources (such as hydroelectric power) to fill temporary gaps between demand and solar generation. This, however, requires that such energy sources are available. One could also think of over-sizing the peak generation capacity of solar units to minimise the times when available solar power is less than the demand. The solar unit output could also be pre-programmed and controlled (peak shaving) to reduce the degree of fluctuation to some extent. All of these would mean extra costs for the designed solar energy system.
2. Energy storage
The intermittency can be tackled by the storage of electric power at the site of generation and using the stored power when solar output is reduced. Various storage technologies are now available for energy storage: batteries, fuel cells, hydrogen gas, compressed air, pumped hydroelectric power and flywheels. Some solar thermal systems make use of heat storage using water, hot oil or molten salt to produce power when the sun is not shining. Energy storage technologies are still quite expensive, although their costs are declining steadily. Another option is to store electric power in electric-vehicle batteries and use it for load management, known as "vehicle to grid" or V2G.
3. Smart grid
A less costly option is to use “smart” demand-response management to shift flexible loads to a time when more renewable energy is available. This, however, is not straightforward considering the socio-economic aspects of implementation necessary to make it successful, and consumers may not like their appliances being managed by grid operators. Therefore, it is necessary to find a demand response approach, such as variable motor drives, where a resource with sufficient capacity to be partially ramped up and down will have no significant effect on the consumers.
4. Forecasting and control
Forecasting the weather (sunlight) gives grid operators more time to plan for a backup energy supply when a solar generation might produce less than anticipated. A study by GE Energy in the US found that state-of-the-art wind and solar forecasting can reduce operating costs by $0.01–$0.02/kWh.
A combination of one or more of the above approaches can possibly be used to manage the intermittency of solar energy grid-connected systems. This will surely add to the implementation costs and hence affect the grid-parity goal for solar energy integration. On the brighter side for grid operators, solar plants suffer much less downtime compared to conventional central power generation and when a distributed solar generation unit is down, the impact on the grid is much less compared to when a central generation unit is down (due to equipment outage or supply disruption).
Written by Nilesh Jadhav, Contributing Editor – India.