Conceptual structural design today relies heavily on the intuition and experience of the structural engineer, often includes an investigation of similar reference projects, and is mostly a time-consuming and demanding task that is characterized by many iteration steps. This research hypothesizes that conceptual structural design as we know it today can be greatly assisted and improved by using modern machine learning (ML) methods.
Typically bridges are designed, verified and constructed from scratch, depending on the prior knowledge of the involved team of engineers. Hence for most of the existing bridge buildings performance data are not available in digital and / or computational form for further science studies. Thus individual researchers and engineers in practice creating their own methods, finding a case study data set and determining efficacy on their own. Not surprisingly, most of those researcher find positive, yet questionably meaningful results.
As data are the back bone for any data-driven method, using a large, consistent benchmark data set from hundreds (or thousands) of bridges allows researchers to develop and assess novel methods across a heterogeneous data set. If multiple researcher use the same data set, then there can be meaningful comparisons of accuracy, speed and ease-of-use.