APPLICATION OF ARTIFICIAL NEURAL NETWORK AND SENSITIVITY ANALYSIS TO ENHANCE BIOETHANOL YIELD IN A BUTADIENE PLANT

The paper discusses bioethanol's uses as a consumer product and its potential in butadiene production for vehicle tire manufacturing. The primary aim of this research is to improve bioethanol production by using sensitivity analysis in the extractive distillation process of a butadiene plant. The study specifically focuses on optimizing essential variables in this process, such as the molar reflux ratio of the azeotropic column and the number of stages in both the azeotropic and extractive distillation columns using Aspen Plus V11. Given the complexity of bioethanol production through extractive distillation, precise modelling is crucial for increasing efficiency and yield. To address this challenge, the study introduces Artificial Neural Networks as a modelling approach.

1. APPLICATION OF ARTIFICIAL NEURAL NETWORK AND SENSITIVITY ANALYSIS TO ENHANCE BIOETHANOL YIELD IN A BUTADIENE PLANT Nigerian Society of Chemical Engineers 53 r d NSChE Annual International Conference and AGM EDO/DELTA, Nigeria. 2nd –4th, November, 2022 Daniel Aliyu KEJI , Bomafinitamunopiri Adango LONGJOHN , Mufliah G. OMOFOYEWA, Abdulwahab GIWA

2. PRESENTATION OUTLINE INTRODUCTION AIM LITERATURE REVIEW METHODOLOGY RESULTS CONCLUSION REFERENCES

3. Introduction Bioethanol, derived from organic crops like sugarcane and corn, is a renewable fuel, unlike finite fossil fuels such as coal and oil. It's a cleaner alternative to gasoline, releasing CO 2 when burned, but the emissions are offset by the CO 2 absorbed during plant growth, reducing net greenhouse gas emissions and combating climate change ( Farrell , 2006).

4. Introduction Cont’d. Neural networks, inspired by the human brain, are versatile machine learning algorithms ( Simon , 2009) . They are valuable in chemical engineering for modelling intricate relationships in large datasets. Also, sensitivity analysis, employed in fields like chemical engineering, assesses how changes in input variables affect system or model outputs.

5. aim This study also explores the application of ANN and sensitivity analysis in modelling and optimizing the extractive distillation process for bioethanol production.

6. Literature review AUTHOR TITTLE CONCLUSION Nitsche and Gbadamosi (2017) Extractive and Azeotropic Distillation, Practical Column Design Guide Development of various techniques employed for pure ethanol recovery. Zabed et al. ( 2014) Bioethanol Production from Fermentable Sugar Juice Explained various factor that affect the bioethanol production. Rathi et al . (2023) Exploring Neuromorphic Computing Based on Spiking Neural Networks: Algorithms to Hardware.  Explains the difference between traditional modelling approach and Artificial Neural Network.

7. METHODOLOGY Feedstock   Butadiene (98%) Juice Extraction Process Fermentation Process Extractive Distillation Butadiene Separation Broth Figure 1: Block Diagram of Butadiene Production

8. METHODOLOGY C ont’d . Figure 2: Process Flow Diagram of Bioethanol Process

9. METHODOLOGY Cont’d. The parameters chosen (manipulated variables) include: Temperature of feed Vapour fraction of feed Molar reflux ratio of the azeotropic and extractive distillation column Number of stages of the azeotropic and extractive distillation column The tool used for the artificial neural network modelling of the extractive distillation process for the production of bioethanol in a butadiene plant was MATLAB R2022b, using the nftool . Table 1: Ranges of Input for Manipulated Variables Variable Start point End point Molar Reflux ratio 0.8 6 Number of Stages of Column B23 20 30 Number of Stages of Column B24 14 17

10. METHODOLOGY Figure 3: Neural Network Structure

11. Results and discussion Figure 4: Result of the Optimized Bioethanol Mole Fraction at the Extractive Distillation

12. Result and discussion c ont’d . Figure 5: Plot of Mole Fraction of Ethanol Recovery against Column Specification of Number of Stages and Mole Reflux Ration in Column B23

13. Result and discussion Cont’d. Figure 6: Performance of the Trained Data

14. Result and discussion Cont’d. Figure 7: R Values of each of the Trained Data

15. Conclusion

16. references Farrell , A. E. (2006), “ Ethanol can contribute to energy and environmental goals ”, Science , 311(5760), 506-508. Rathi , N., Chakraborty , I., Kosta , A., Sengupta , A., Ankit , A., Panda , P. And Roy , K. (2023), “Exploring neuromorphic computing based on spiking neural networks: Algorithms to hardware.  ACM Computing Surveys ,  55 (12), pp.1-49. Nitsche , M. and Gbadamosi , R. (2017), “Extractive and azeotropic distillation”,  Practical column design guide , pp.153-164. Zabed , H., Faruq, G., Sahu, J.N., Azirun , M.S., Hashim, R. and Nasrulhaq Boyce, A. (2014), “Bioethanol production from fermentable sugar juice”,  The scientific world journal ,  2014 . Simon, H. (2009), “Neural network and learning machine”, Pearson Education , Inc, New Jersey , 3 rd edn , pp 2.

17. THANK YOU