Prediction of material properties of propellant waste modified bricks through microstructures by Topographic independent component analysis coupled 3D Convolution neural networks. Issue 19 (1st October 2022)

• Record Type: Journal Article
• Authors: Mehta, P.K.; Kumaraswamy, A.; Saraswat, V.K.; Chinnadurai, Vijayakumar; kumar, B.Praveen
• Is Part Of: Ceramics international. Volume 48:Issue 19(2022)Part B
• Page Start: 28918
• Page End: 28926
• Publication Date: 2022-10-01
• Subjects: 3D CNN -- Bricks -- Porosity -- SEM images; Ceramics -- Periodicals; Céramique industrielle -- Périodiques; Ceramics; Periodicals; Electronic journals; 666
• DOI: 10.1016/j.ceramint.2022.04.064
• Languages: English
• ISSNs: 0272-8842

Abstract

Abstract: Purpose: Deciphering materials properties from the textural information of microstructural images is still a challenge. Objective: Proposes a Topographic Independent Component Analysis coupled with 3-D Convolutional Neural Network (TICA- 3DCNN) architecture that processes the features from scanning electron microscope (SEM) images to predict the material properties of high energy propellant (HEP) modified bricks. Method: First, high-energy modified bricks with ten different HEP additives fractions is prepared, and their microstructural information is acquired with SEM. Then, SEM images of each group are pre-processed and subjected to the TICA analysis to decipher the distinct higher-order microstructural dependencies, such as co-activation of components. These microstructural co-activation components are subsequently ordered and passed into 3D CNN architecture to map them with the material properties such as water absorption, compressive strength, density, and porosity. An extreme learning machine is employed as a fully connected classification layer in the 3D-CNN architecture. Finally, the performance of the proposed TICA - 3DCNN approach is compared with the traditional 2D CNN frameworks such as Inception-v4 and Faster RCNN. Results: Proposed TICA-3DCNN architecture has performed with higher sensitivity and specificity than Inception-v4 and Faster RCNN architecture in deciphering microstructural textural properties and mapping them with materials properties. Conclusion: The Proposed model can be used in the construction and building materials to predict material properties through microstructural features. Highlights: Predicting material properties of propellant waste modified bricks from the textural information of microstructural images. Current work presents topographic independent component analysis coupled with 3-D convolution neural network architecture and trained on textures to map the material properties. TICA based 3DCNN approach is more effective than traditional 2D CNN frameworks such as Inception-v4 and Faster RCNN. Better sensitivity and specificity values observed for TICA - 3DCNN approach in deciphering microstructural textural properties and predicting materials properties.