Conservation of heritage artifacts is a very sensitive task as conservators usually have very little information about the artifacts. Moreover, due to the uniqueness and the historic value of these artifacts, invasive analysis are not always possible. Therefore, without sampling options, conservators are required to use non-invasive diagnostic methods in order to identify the metal characteristics of the artifact. When confronted with an unknown artifact, conservators generate conceptual models of the corrosion forms. These models are based on formal representations of corrosion forms, but are not directly exploitable for drawing hypotheses regarding the underlying metal. This paper presents the design of a data structure generated from the conceptual models which supports the comparison and retrieval of corresponding artifacts. Integrated with a database of heritage artifacts, this data structure offers advanced decision support to conservators confronted with unknown artifacts.