Using Bayesian Neural Network for Predicting Fatigue Strength Based on Composition and Process Parameters

2021 ◽  
Author(s):  
R M Chandima Ratnayake ◽  
Arvind Keprate
Keyword(s):  
Neural Network ◽  
Fatigue Strength ◽  
Process Parameters ◽  
Bayesian Neural Network ◽  
Strength Based

Artificial Intelligence Based Approach for Predicting Fatigue Strength Using Composition and Process Parameters

2020 ◽  
Author(s):  
Arvind Keprate ◽  
R. M. Chandima Ratnayake
Keyword(s):  
Artificial Intelligence ◽  
Fatigue Strength ◽  
Process Parameters ◽  
Fatigue Testing ◽  
Simulated Data ◽  
Machine Learning Algorithms ◽  
Coefficient Of Determination ◽  
Gradient Boosting ◽  
Data Set ◽  
Strength Based

Abstract Accurate prediction of the fatigue strength of steels is vital, due to the extremely high cost (and time) of fatigue testing and the often fatal consequences of fatigue failures. The work presented in this paper is an extension of the previous paper submitted to OMAE 2019. The main objective of this manuscript is to utilize Artificial Intelligence (AI) to predict fatigue strength, based on composition and process parameters, using the fatigue dataset for carbon and low alloy steel available from the National Institute of Material Science (NIMS) database, MatNavi. A deep learning framework Keras is used to build a Neural Network (NN), which is trained and tested on the data set obtained from MatNavi. The fatigue strength values estimated using NN are compared to the values predicted by the gradient boosting algorithm, which was the most accurate model in the OMAE 2019 paper. The comparison is done using metrics such as root mean square error (RMSE), Mean Absolute Error (MAE), Coefficient of Determination (R2) and Explained Variance Score (EVS). Thereafter, the trained NN model is used to make predictions of fatigue strength for the simulated data (1 million samples) of input parameters, which is then used to generate conditional probability tables for the Bayesian Network (BN). The main advantage of using BN over previously used machine learning algorithms is that BN can be used to make both forward and backward propagation during the Bayesian inference. A case study illustrating the applicability of the proposed approach is also presented. Furthermore, a dashboard is developed using PowerBI, which can be used by practicing engineers to estimate fatigue strength based on composition and process parameters.

Data Mining for Estimating Fatigue Strength Based on Composition and Process Parameters

2019 ◽  
Author(s):  
Arvind Keprate ◽  
R. M. Chandima Ratnayake
Keyword(s):  
Data Mining ◽  
Fatigue Strength ◽  
Process Parameters ◽  
Fatigue Testing ◽  
Absolute Error ◽  
Processing Parameters ◽  
Coefficient Of Determination ◽  
Low Alloy Steels ◽  
Strength Based ◽  
Alloy Steels

Abstract Accurately estimating the fatigue strength of steels is vital, due to the extremely high cost (and time) of fatigue testing and often fatal consequences of fatigue failures. The main objective of this manuscript is to perform data mining on the fatigue dataset for steel available from the National Institute of Material Science (NIMS) MatNavi. The cross-industry process for data mining (CRISP-DM) approach was followed in the paper, in order to gain meaningful insights from the dataset and to estimate the fatigue strength of carbon and low alloy steels, using composition and processing parameters. Of the six steps of the CRISP-DM approach, special emphasis has been placed on steps 2 to 5 (i.e. data understanding, data preparation, modeling and evaluation). In step 4 (i.e. modeling), a range of machine learning (parametric and non-parametric) is explored to predict the fatigue strength, based on the composition and process parameters. Various algorithms were trained and tested on the dataset and finally evaluated, using metrics such as root mean square error (RMSE), Mean Absolute Error (MAE), Coefficient of Determination (R2) and Explained Variance Score (EVS).

scholarly journals BNNpriors: A library for Bayesian neural network inference with different prior distributions

2021 ◽  
pp. 100079
Author(s):  
Vincent Fortuin ◽  
Adrià Garriga-Alonso ◽  
Mark van der Wilk ◽  
Laurence Aitchison
Keyword(s):  
Neural Network ◽  
Network Inference ◽  
Bayesian Neural Network ◽  
Prior Distributions

scholarly journals Artificial neural network for modeling and investigating the effects of forming tool characteristics on the accuracy and formability of thin aluminum alloy blanks when using SPIF

2021 ◽  
Author(s):  
Sherwan Mohammed Najm ◽  
Imre Paniti
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Process Parameters ◽  
Single Point ◽  
Tool Material ◽  
Deformation Mode ◽  
Geometric Accuracy ◽  
Forming Process ◽  
Tool Shape ◽  
Artificial Neural

AbstractIncremental Sheet Forming (ISF) has attracted attention due to its flexibility as far as its forming process and complexity in the deformation mode are concerned. Single Point Incremental Forming (SPIF) is one of the major types of ISF, which also constitutes the simplest type of ISF. If sufficient quality and accuracy without defects are desired, for the production of an ISF component, optimal parameters of the ISF process should be selected. In order to do that, an initial prediction of formability and geometric accuracy helps researchers select proper parameters when forming components using SPIF. In this process, selected parameters are tool materials and shapes. As evidenced by earlier studies, multiple forming tests with different process parameters have been conducted to experimentally explore such parameters when using SPIF. With regard to the range of these parameters, in the scope of this study, the influence of tool material, tool shape, tool-end corner radius, and tool surface roughness (Ra/Rz) were investigated experimentally on SPIF components: the studied factors include the formability and geometric accuracy of formed parts. In order to produce a well-established study, an appropriate modeling tool was needed. To this end, with the help of adopting the data collected from 108 components formed with the help of SPIF, Artificial Neural Network (ANN) was used to explore and determine proper materials and the geometry of forming tools: thus, ANN was applied to predict the formability and geometric accuracy as output. Process parameters were used as input data for the created ANN relying on actual values obtained from experimental components. In addition, an analytical equation was generated for each output based on the extracted weight and bias of the best network prediction. Compared to the experimental approach, analytical equations enable the researcher to estimate parameter values within a relatively short time and in a practicable way. Also, an estimate of Relative Importance (RI) of SPIF parameters (generated with the help of the partitioning weight method) concerning the expected output is also presented in the study. One of the key findings is that tool characteristics play an essential role in all predictions and fundamentally impact the final products.

Sign in / Sign up

Export Citation Format

Share Document