Fatigue life estimation of an all aluminium alloy 1055 MCM conductor for different mean stresses using an artificial neural network. (November 2020)

• Record Type: Journal Article
• Authors: Kalombo, R.B.; Pestana, M.S.; Freire Júnior, R.C.S.; Ferreira, J.L.A.; Silva, C.R.M.; Veloso, L.A.C.M.; Câmara, E.C.B.; Araújo, J.A.
• Is Part Of: International journal of fatigue. Volume 140(2020)
• Publication Date: 2020-11
• Subjects: AAAC 1055 MCM conductor -- Fatigue life estimation -- Artificial neural network; Materials -- Fatigue -- Periodicals; Materials -- Fatigue; Periodicals; 620.1122
• DOI: 10.1016/j.ijfatigue.2020.105814
• Languages: English
• ISSNs: 0142-1123

Abstract

Highlights: 27 new fatigue data for an AAAC 1055 MCM conductor by varying mean stress. Use of ANN to estimate the effect of the mean tensile stress on the fatigue resistance. Estimation of constant life diagrams by using ANN. Training of the ANN with three S-N curves of the AAAC 1055 MCM conductor at a given level of mean stress. Reliable results with a lower computational cost using 15 neurons in the hidden layer. Abstract: This study is aimed to employ artificial neural networks (ANNs) to predict the fatigue lives of the All Aluminium Alloy Conductor (AAAC) 1055 MCM overhead conductor, considering different values of stretching load (mean stress). For ANN training, three Wöhler (S-N) curves are generated for a conductor/suspension clamp assembly. Twenty-seven fatigue tests are carried out with stretching loads, related to everyday stress (EDS) of 17%, 20% and 25.6% of the conductor's ultimate tensile strength (UTS), corresponding to mean stresses of 48 MPa, 54 MPa and 73 MPa, respectively. Constant life diagrams (at 10 5, 10 6, 10 7 number of loading cycles) for the AAAC 1055 MCM overhead conductor are built using the ANN. The results confirm that the ANN can accurately predict the fatigue lives of an overhead conductor for various levels of mean stress also when limited experimental data is used for training.