Breaking load and bending strength prediction in manufacture of fibre cement composites using artificial neural networks and a flocculation sensor |
| |
Authors: | C. Negro A. Alonso A. Blanco J. Tijero |
| |
Affiliation: | Department of Chemical Engineering, Complutense University of Madrid, Avda. Complutense s/n, 28040 Madrid, Spain |
| |
Abstract: | The optimisation of the flocculation process during fibre cement production is a new key issue for the fibre cement industry. Many companies face difficulties in optimising the flocculant dosage in real time, which leads to product strength losses. This paper shows the feasibility of using artificial neural networks (ANNs) to establish correlations between flocculation data, in-line measured in a Hatschek machine by a focused beam reflectance measurement (FBRM) sensor, and mechanical properties of final composites. The results show a clear relationship between the mechanical properties of fibre cement composites and the flocculation process and that these are determined in real time. Three ANNs have been created to predict breaking load for 48 h and 7 days and bending strength for 7 days, to obtain good correlations between the predicted and the real values. |
| |
Keywords: | A. Fibres B. Mechanical properties C. Process monitoring E. Forming |
本文献已被 ScienceDirect 等数据库收录! |