With the vastness of existing railroad infrastructure, there exist numerous road crossings which are lacking warning light systems and/or crossing gates due to their remoteness from existing electrical infrastructure. Along with lacking warning light systems, these areas also tend to lack distributed sensor networks used for railroad track health monitoring applications. With the power consumption required by these systems being minimal, extending electrical infrastructure into these areas would not be an economical use of resources. This motivated the development of an energy harvesting solution for remote railroad deployment. This paper describes a computer simulation created to validate experimental on-track results for different mechanical prototypes designed for harvesting mechanical power from passing railcar traffic. Using the Winkler model for beam deflection as its basis, the simulation determines the maximum power potential for each type of prototype for various railcar loads and speeds. Along with calculating the maximum power potential of a single device, the simulation also calculates the optimal number and position of the devices needed to power a standard railroad crossing light signal. A control system was also designed to regulate power to a battery, monitor and record power production, and make adjustments to the duty cycle of the crossing lights accordingly. On-track test results are compared and contrasted with results from the simulation, discrepancies between the two are examined and explained, and conclusions are drawn regarding suitability of the device for powering high-efficiency LED lights at railroad crossings and powering track-health sensor networks.© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.