The use of power grids is increasing around the world, and with it, a few hurdles need to be jumped over. The rapidly developing power grids we use today continuously work on harnessing as much solar, or wind power, among other methods, as possible.
This trend is due to rise each year, meaning that more and more power grids will be needed.
Furthermore, these grids will need to work efficiently, and at their best capacity.
However, renewable energy generation technologies are variable and not always fully dispatchable. A team of international researchers led by the Singapore University of Technology and Design have put their heads together to find a solution.
Renewable energy works from uncontrollable resources that fluctuate. The classical optimal power flow (OFP) options in place sometimes offer inefficient power generation policies, which lead to line overloads and mass power outages.
Not a fun predicament to be caught up in.
The current OFPs are computed based on very simple predictions of expectations and generation levels for a specific time frame. Even though these predictions can be quite accurate for traditional power grids, they're unpredictable when it comes to renewable generators.
Regardless of the large investments in the industry, power outages due to uncertain renewable power generation occur regularly. This goes to show that a new approach is needed, as the current strategy solely based on technological improvements of the transmission lines is no longer sufficient.
What has the team discovered?
This is where an international team of researchers from the Singapore University of Technology and Design (SUTD), the Politecnico Di Torino in Italy, and the Pennsylvania State University in the U.S. comes in.
The researchers have proposed a new probabilistic dispatch strategy for these modern power grids that lowers generation costs, but also ensures that constraints in the power network are secure. This method prevents overload or power outages.
"One of the advantages of the probabilistic approach pursued in this research is to avoid conservatism associated with the existing methods. Instead of requiring that the network constraints are satisfied for all possible values of uncertainty, we pushed the boundaries and allowed for a small well-defined risk of constraint violation to develop this new approach," explained lead author Dr. Mohammadreza Chamanbaz, Senior Research Fellow, SUTD.