Simulation and Control of Benzoic Acid Recrystallization Experiment with Artificial Neural Network
GAO Guang-Qin1, HUANG Hong-Mei2, XIE Pu-Hui1
1. College of Science, Henan Agricultural University, Zhengzhou 450002, China;
2. College of Chemistry and Materials Science, Sichuan Normal University, Chengdu 610006, China
Abstract This paper described the encryption experiment of activated carbon ratio (A) and solvent volume (C) in optimum experimental condition of orthogonal experiment design and analysis, and built a relationship model between experimental conditions and the yield and quality of benzoic acid with artificial neural network in order to simulate yield and quality of benzoic acid and control experimental conditions. The research results showed that the proposed artificial neural network model (BARE) had high predictive power, and accuracy could reach above 99% while fast flow filter paper was used. Based on this model, the optimum control program compiled by MATLAB could accurately determine the activated carbon ratio and solvent volume were 2.2% and 72 mL respectively, and the corresponding highest yield of benzoic acid was 0.809 0. It could display continuously the change of the benzoic acid yield with the change of activated carbon ratio and solvent volume, and determine the optimum point of experimental conditions on computer. In experiment teaching, guiding and controlling the student's entity experiment with the model could improve the teaching quality.
GAO Guang-Qin, HUANG Hong-Mei, XIE Pu-Hui. Simulation and Control of Benzoic Acid Recrystallization Experiment with Artificial Neural Network[J]. Chinese Journal of Chemical Education, 2016, 37(12): 22-26.
