where K is shape resistance coefficient of bridge pier, or 0.75 for circular piers. A is the emerged area of the bridge pier, ρ is the water density, and v is the flow velocity. Schulze et al. suggests that normal river flow velocity is around 0.7 m/s with a standard deviation of 0.24 m/s. As a result, the water flow force in this case is around 100-300 N. We took an average of 200 N. In the simulation, water force was simplified as a sine wave varying from 100-300 N with a certain frequency. For a given water flow force of a certain frequency, the dynamic acceleration responses of the pier were calculated using Newmark-β method, a well-known time-integration method. Then the acceleration responses were taken as input to the energy harvesting device and the average harvested power (referred to as average power hereinafter) and accumulated charge per hour (referred to as accumulated charge hereinafter) could be numerically calculated. By varying the water force frequency, one could get average power and accumulated charge with respect to each frequency. Their frequency distribution could be plotted like spectra, the former of which is referred to as average power pseudo-spectrum and the latter as accumulated charge pseudo-spectrum. The term “pseudo” was used because the spectra-like plot defined above were not conventional spectra yielded by a mathematical transform decomposing time-domain responses into frequency components. One observation point at the pier cap was chosen to calculate the average power and accumulated charge per hour harvested by the MEH. Only a single observation point was good enough because the fundamental mode easily excited by the in-plane water force would be in-plane mode, whose pier cap would vibration with the same amplitude. This modal motion was verified in a preliminary dynamic analysis.