Long heavy-haul trains are now a reality, especially for ore transportation. In some railways, compositions of up to 330 wagons are in service, requiring several locomotives. Trains like that travel long distances, sometimes through cities or in uninhabited regions. They are driven by just one driver which must keep the whole train working safely on the track. The wagons don’t have any source of electrical energy to power sensors and to transmit their signals to the locomotive; nor wireless communication. In fact, in some of these railways, there is no internet along with the track out of the cities. One important indicator of the safety of the train is the force between the wagons during the trip, through the shunting. Using strain gauges to measure these forces is a possible solution and ultrasonic stress sensors (UST) is a suitable alternative. UST with L cr waves requires a low amount of energy and can be employed in rusty and dirty places. However, they also need an energy source. Wind and solar solutions are not always adequate because, unfortunately, there are places where these components have economic value and they can be stolen. A possible source of energy to power the USTs could be the Vibration Energy Harvester (VEH). These simple and not expensive systems can be built in small packs, giving the energy to measure the forces and transmit the data to the locomotive or designated sites along the track. This work aims to evaluate the possibility of using VEH to power USTs to measure the forces between the wagons during the journey. Knowing that the oscillation in the shunting has a very low frequency, the work intent to optimize a multi-beam VEH to be able to capture the highest amount of energy possible, in a very small arrangement, using genetic algorithm. The result shows that VEH is an adequate alternative to power autonomous UST sensors.