Knowledge-based prediction of shear strength of concrete beams without shear reinforcement

Structural engineers heavily rely on computer software to perform structural analysis, and they increasingly computerize design procedures to avoid manual repetitions. To benefit fully from the computerization, it is necessary to utilize the domain knowledge contained in a database such as a concrete shear database used in this paper. A knowledge-based system uses a database of knowledge in combination with its retrieval mechanism such as artificial neural networks (ANN) to imitate problem-solving strategy of human. This paper presents an application of the knowledge-based approach, utilizing the shear database and retrieval of information using ANN. The database can be used more extensively than regression of shear strength that had been reported in other literature. As a demonstration, two models are developed and compared with design equations. The first model estimates shear strength, and the second model systematically provides conservative estimation. Although both models already outperform all existing design equations, they can be easily revised for further improvement whenever additional experimental data sets become available.