The prediction of physical phenomena that occur on Earth has been a topic of vital interest since the beginning of human societies. Predicting for example lunar or solar movements has allowed the development of navigation and agriculture. With increasing modernization and urbanization around the world, new needs appear for human communities, which makes necessary the construction of systems. Examples are the electricity that keeps our cities powered and the water service that reaches our homes. Lately one of the topics that has received a lot of collective attention is undoubtedly the use of renewable energies such as that from wind, thus beginning an era in which we hope to leave the use of fossil fuels behind.
The efficiency of wind energy d from large turbines depends entirely on wind movement patterns, which have turned hardly predictable due to climate change, caused by storms, hurricanes and rainfall, for example. How can we predict these changes to make better use of the available energy? Recently, a study developed by university researchers in different places has been able to develop a wind pattern prediction system through computational algorithms. The most interesting thing about this is that these instructions have been based on biomimicry, imitating the way in which living entities develop their processes. The document proposes a “recurrent neural network (RNN)” which is capable of accurately anticipating the speed and direction of the wind, which in turn allows determining when the wind turbines should be in operation or not.
This optimization inspired by biomimicry handles different problems through the imitation of nature. Elements taken from it are evolution, swarm intelligence and natural selection. For example, genetic algorithms (GA) are a popular type of optimizers. GA iteratively evolves toward solutions close to an optimal state by following the steps of selection, crossover, and mutattions of a population of candidate solutions. On the other hand, particle swarm optimization (PSO) is another tool of this type. In this case the inspiration comes from the migratory movement of birds and the schooling of fishes. In the case of the wind pattern detection system, the researchers have combined elements of PSO and neural networks to generate a hybrid method. The results obtained were impressive, since they managed to have significantly higher efficiency than other conventionally used algorithms. This allows us to have much more efficient turbines for energy generation. These systems implicitly contain the following life principles:
- Integrate the unexpected
- Incorporate resilience through variation.
- Leverage cyclic processes.
- Use feedback loops.