Abstract:

Digitalization is transforming the labor market. Many tasks that until recently were being performed manually are being now automated and executed by smart machines thanks to artificial intelligence techniques. For this reason, companies are increasingly demanding workers with knowledge in data science, machine learning, neural networks and artificial intelligence in general. However, there is little training in this field, being relegated mainly to the last courses in computer engineering and some specialized masters. Therefore, training in these tools should be extended first to the rest of engineering courses and then to the rest of degrees, so that graduates are much better adapted to the needs of the labor market.

In order to maximize attention and educational success, it is important to propose interesting and motivational didactic proposals and challenges. On the other hand, young people and in general all citizens are increasingly aware of the need to reduce energy consumption, reduce pollution, and contribute to the improvement of the environment to reduce the effects of climate change. Therefore, didactic proposals focused on making a cleaner and greener world awaken the interest and increase the motivation of students, making learning more meaningful.

In this work, an emerging renewable energy, wind power, is used as a transversal content on which pivots a didactic proposal where the two objectives mentioned above converge: training in an artificial intelligence technique, neural networks, and contributing to a challenge related to the improvement of the environment. It is worthy to note that neural networks can be used to predict malfunctions of the turbine, forecast the wind to plan the operation of the wind device, estimate the expected output power depending on the wind and check for misalignment, etc. The use of these technologies can contribute to improve the efficiency of wind turbines, reducing dependence on coal and other polluting energy sources, and therefore improve the environment.

The teaching proposal consists of using neural networks to predict the behavior of wind turbines and detect if there are anomalies in the performance. The application is simple but it allows the student to learn about a current and new technology, neural networks, and also to understand how power is generated from wind. To carry out this didactic unit, students must first study the structure and operation of neural networks, learn the differences between regression and classification problems, and use neural networks to forecast the behavior of wind turbines. On the other hand, students should also receive an introduction to the operation of wind turbines and to the main parts of the device: power, torque, rotor speed, wind speed, etc. With this training, students receive a dataset that they can use to visualize the signals, processing them and understand how the wind system performs, and train different types of neural networks to make predictions. Computational software that already have those tools is used.

The proposal is organized based on learning objectives, key materials, skills, time schedule, and an evaluation to optimize its impact.

Keywords:

Learning-by-doing, motivational education, computational tools, neural networks, wind energy.