PSI - Issue 56

B. Kalita et al. / Procedia Structural Integrity 56 (2024) 105–110 B. Kalita/ Structural Integrity Procedia 00 (2019) 000–000

106

2

failure prediction methods, such as inconsistency in the prediction and inability to solve complex nonlinear damage mechanics [2-4]. Using a backpropagation neural network (BPNN) and extreme learning machine (ELM), Raja et al. [4] have estimated the FCGR of Al 2014 alloy and concluded that ELM better on the dataset. Nowadays, in the era of Industry 4.0, data can be shared and retrieved from anywhere around the globe using cloud-based data storage methods [5]. The term machine learning (ML) refers to the application of sophisticated algorithms developed using programming languages like Python to teach a computer to perform a given task accurately. Machine learning techniques are way too easy and flexible as compared to designing numerical equations because of their non-linear activation functions [6, 7]. 2. Materials and Methods The 17-4 PH stainless steel is a martensitic alloy that can be precipitation hardened when used in traditional manufacturing processes. Within an L-PBF process, it will exhibit up to 5% lower yield strength and 6% higher elongation [8]. Due to their popularity in various structural applications, high fatigue resistance is particularly desirable for 17-4 PH SS. Therefore, many recent studies focused on the fatigue crack initiation behavior of L-PBF 17-4 PH SS [11]. In this investigation, specimens were made from Stainless steel 17-4 PH powder that had been atomized with argon using the laser powder bed fusion (LPBF) method. In order to build the ML models, relevant experimental Fatigue data of 17-4 PH SS alloy is collected from the literature. Electrical discharge machining (EDM) was used to create compact tension (CT) specimens from fabricated samples in the EOS M290 in an argon-shielded environment [9]. CT samples containing the same measurements were also fabricated from wrought 17-4 PH stainless steel plates as it is pertinent to contrast the mechanical characteristics of L-PBF 17-4 PH SS with its wrought equivalent. Based on the orientation of the notch in regard to the building direction, L-PBF 17-4 PH SS specimens were machined [10-12]. After machining, L-PBF 17-4 PH SS specimens were put through the heat treatment processes [13, 14]. Contrary to samples treated to 1072 MPa yield strength, 1132 MPa ultimate various deformation behaviors, the specimens treated to 1300 MPa yield strength, 1375 MPa ultimate strength, and 0.16 true fracture strain have more strength and ductility [15]. It is essential to determine the constituent phases to analyze the crack growth behavior related to microstructure because of their varied deformation characteristics. For this, the TCFE9 thermodynamic database for several kinds of steels and alloys based on Fe, including stainless steels, was used [16]. 2.1. Fractography and the fatigue crack growth test (FCG) On a testing device for servo-hydraulic systems, FCG experiments were carried out. Low-stress fatigue on CT samples under load control, at room temperature, using a sinusoidal loading waveform with a load ratio R = 0.1 and a frequency f = 10 Hz until failure experiments (also known as high cycle fatigue regime) were carried out [15]. Because pre-cracking was conducted under the same stress as for the fatigue crack growth rate testing, there won't be a crack tip plastic zone size disparity to contend with. The load amplitudes ranged from 4275 N to 1575 N. The recorded CMOD values, the load range, and the observed cycle count were used to determine the FCG rate, da/dN, and stress intensity factor ranges ( Δ K), in accordance with ASTM E647. Four CT samples from each group were subjected to FCG testing to address variability. The following variables are the key factors affecting Fatigue Crack Growth rate: • Average stress impact (mean stress effect) • Environment • Short cracking effect • Underloads and overloads The stress-life method can be used to determine a material's fatigue life, the crack-growth method, the strain-life method, and probabilistic methodologies. It may be based on techniques for crack or life growth.