The Internet of Things is a promising vision which enables trillions of sensor devices to be connected. A common bottleneck for such devices is the energy supply. Using batteries is expensive, it reduces the device lifetime and has a negative impact on the environment. Energy harvesting provides a more sustainable solution. Unfortunately, environmental energy may not always be available. Due to the random nature of environmental energy it is hard to conduct repeatable tests for such devices. In this thesis we presented a way to test energy harvesting devices in a realistic and repeatable fashion.