Yield strength prediction for thermoplastic composites based on a Sparse Fuzzy Model

Nowadays thermoplastic composites are commonly used owing to their good mechanical properties, which can be ensured only by the proper mixing of different types of materials. In this paper, we present the results of our studies regarding the fuzzy modeling of the relation between the yield strength and the amount of the used components (ABS, polycarbonate, multiwall carbon nanotube). The initial rule base was created using FCM clustering and the parameters were tuned by RBE-SI that applies a hill-climbing approach and enriches the rule base with new rules if it is necessary. Owing to the possible sparse character of the rule base the fuzzy rule interpolation based FRIPOC method was used as inference technique. The model was validated by applying it to an independent set of test data.