A Fault Detection and Isolation Framework for Repeatable and Comparable Experimentation

There is an extensive literature available about condition monitoring relying on multi-dimensional data-driven system models and mappings, including proposal of new methods and algorithms, comparison of state-of-the-art methods, and stateof- the-art revisions. But, when practitioners start to implement their own software to carry out their research, there is a lack of articles in the literature with detailed documentation about how to design a framework for repeatable and comparable experimentation. We propose a design for repeatable and comparable experimentation on the field of Data-Driven Residual-Based Fault Detection and Isolation. The framework has already been used for several experiments, with successful results, eliciting features such as (i) decreasing of developing times, (ii) facilitating of configuration management, and (iii) facilitating of collection and comparison of results