Simulation of CO2 Solubility in Polystyrene-b-Polybutadieneb-Polystyrene (SEBS) by artificial intelligence network (ANN) method

3rd International Conference on Global Sustainability and Chemical Engineering

工业技术;化学工业

Sharudin, R.W.^1,2 ; Abdulbari Ali, S.^1 ; Zulkarnain, M.^2 ; Shukri, M.A.^1

Faculty of Chemical Engineering, Universiti Teknologi MARA, Shah Alam

40450, Malaysia^1

Institute for Infrastructure Engineering and Sustainable Management (IIESM), Faculty of Civil Engineering, Universiti Teknologi MARA, Shah Alam

40450, Malaysia^2

关键词: CO2 solubility;    Detention time;    Gradient descent;    Levenberg-Marquardt;    Network training;    Regression coefficient;    Training time;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/358/1/012009/pdf
DOI  :  10.1088/1757-899X/358/1/012009

学科分类：工业工程学

来源: IOP

【 摘 要 】

This study reports on the integration of Artificial Neural Network (ANNs) with experimental data in predicting the solubility of carbon dioxide (CO2) blowing agent in SEBS by generating highest possible value for Regression coefficient (R2). Basically, foaming of thermoplastic elastomer with CO2 is highly affected by the CO2 solubility. The ability of ANN in predicting interpolated data of CO2 solubility was investigated by comparing training results via different method of network training. Regards to the final prediction result for CO2 solubility by ANN, the prediction trend (output generate) was corroborated with the experimental results. The obtained result of different method of training showed the trend of output generated by Gradient Descent with Momentum & Adaptive LR (traingdx) required longer training time and required more accurate input to produce better output with final Regression Value of 0.88. However, it goes vice versa with Levenberg-Marquardt (trainlm) technique as it produced better output in quick detention time with final Regression Value of 0.91.