Harnessing the potential of wind energy is a challenging problem, from capture through conversion into electrical power. A key facet of wind farm design is how the turbines are placed. A “typical” wind farm project may encompass 200×300 grid points on the wind map to cover 20 turbines; if you consider the number of ways to position these turbines, the placement problem becomes incredibly cumbersome.
But it’s not just maximizing energy yield that vexes the human brain. For years, the industry studied the problem of how to factor load constraints into the power optimization mix. Turbine loads (described through IEC standards or load models) have been considered too difficult to include in the optimization problem. Instead, consultants and manufacturers manually review and adjust layouts, which not only takes time – it also generates less than ideal results.
Markedslabben AS, a Norwegian wind farm design tool and optimization start-up, cracked the code on the problem. Using FICO® Xpress Optimization Suite, they incorporated load constraints in the layout optimization problem using a MIP formulation with iterative relaxation. The resulting algorithm allows them to solve the problem in just minutes, for a typical size of 20 turbines. The resulting cloud-based solution, WindFarm Designs, helps clients optimize wind power projects to be profitable, compliant and, ultimately, able to help reduce CO2 emissions.
“While industry software previously has been capable of maximizing energy yield by heuristics or genetic algorithms, no one has been able to incorporate load constraints into the optimization,” said Klaus Vogstad, owner of Markedslabben AS. “This means consultants and manufacturers spend a lot of time checking layout solutions for load compliance, and adjust these layouts manually. This process is time-consuming, and also achieves suboptimal results.”
As a result of this game-changing use of optimization, Markedslabben AS was awarded a FICO Decision Management Award for Analytic Excellence by our panel of independent judges. One of those - Dan Ariely, author of Predictably Irrational, and Duke University Professor of Behavioral Economics - said, “Markedslabben demonstrated the power of advanced analytics to solve some of the world’s toughest problems. For their results and their solution to a problem that has frustrated firms before, the judges overwhelmingly selected them as this year’s Analytic Excellence winner.”
