A stochastic model for the publishing and citation process is discussed in <A HREF="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3043246/">Detecting h-index manipulation through self-citation analysis</A>, from the fraud-detection point of view of the OP. The functional form of the h-index curve following from this model is a sum of incomplete Beta functions, with parameters being the productivity of the researcher and the career length.

One telltale signal is the "humpback", resulting from self-citation of a paper close the the h-index.

<IMG SRC="https://i.sstatic.net/RP3ok.png" WIDTH="400"/>