Structural health monitoring systems consist of a method to measure the structure's performance at a given point in time and a means to interpret the measured raw data in terms of presence and location of damage. This work focuses on the interpretation of the measured raw data. The concept of structural health monitoring that this work supports is a multi-stage damage identification system where the first, or screening, stage is used to indicate when further inspection is necessary, prior to the degradation of structural parameters. Finding low levels of damage is important for this screening stage. This paper describes the application of a previously presented wavelet based algorithm to detect relatively low levels of structural damage. Typically strain sensor data is used in the algorithm. However, the algorithm does not depend on a specific type of data (strain, stress, displacement, etc.), the conditions (load, boundary, etc.) under which the data was acquired, or the geometry of the structure. The structures examined in this paper are steel plates. Sensor data is either obtained from experimentation or finite element models. In the case of finite element analyses, sensor data in the form of non-linear time histories is extracted, thus producing the equivalent of raw sensor data. The wavelet based algorithm makes use of the continuous wavelet transformation and examines how this feature changes as damage accumulates.