XÂY DỰNG MÔ HÌNH MẠNG NƠ-RON NHÂN TẠO ĐỂ XÁC ĐỊNH NHANH NỒNG ĐỘ DUNG DỊCH NaOH DỰA TRÊN PHỔ TRUYỀN QUA CỦA CHÙM TIA GAMMA NĂNG LƯỢNG THẤP 59,54 keV
Nội dung chính của bài viết
Tóm tắt
Nghiên cứu này trình bày một cách tiếp cận mới để xác định nhanh nồng độ của dung dịch NaOH bằng cách đo sự suy giảm của chùm tia gamma năng lượng thấp 59,54 keV phát ra từ nguồn 241Am và sử dụng mạng nơ-ron nhân tạo (ANN) huấn luyện trực tiếp trên các đặc trưng của phổ gốc. Đầu tiên, mô phỏng Monte Carlo với các thông số chính xác của cấu hình đo thực nghiệm, bao gồm đầu dò NaI(Tl) ghi nhận phổ gamma truyền qua dung dịch NaOH với các nồng độ khác nhau, được sử dụng để tạo dữ liệu phổ. Sau đó, dữ liệu phổ này được trích xuất trực tiếp mà không cần các bước xử lý trung gian để huấn luyện và tối ưu mô hình ANN. Hiệu suất dự đoán của mô hình tối ưu được so sánh với phương pháp đường chuẩn (CF) được tạo ra từ mối liên hệ giữa tỉ số lnR giữa diện tích đỉnh truyền qua giữa nước và dung dịch NaOH với nồng độ tương ứng. Kết quả thu được cho thấy mô hình ANN đạt độ lệch trung bình dưới 2,8 % so với các nồng độ tham chiếu trong phạm vi tập thực nghiệm dùng để kiểm tra, trong khi đường chuẩn cho độ lệch trung bình trên 6,0 %. Ở vùng nồng độ thấp (C = 5,0%), mô hình ANN cho độ lệch chỉ 2,5%, tốt hơn nhiều so với độ lệch khoảng 12,6% từ đường chuẩn. Điều này cho thấy việc kết hợp phép đo truyền qua với tia gamma năng lượng thấp và mô hình mạng nơ-ron nhân tạo là cách tiếp cận đầy hứa hẹn để xác định nhanh nồng độ với độ chính xác cao cho các dung dịch kiềm.
Từ khóa
gamma truyền qua, mô phỏng Monte Carlo, MLP-ANN, NaOH, nồng độ
Chi tiết bài viết
Tài liệu tham khảo
Dorn, M., Kareth, S., Weidner, E., & Petermann, M. (2024). Electrical conductivity of lithium, sodium, potassium, and quaternary ammonium salts in water, acetonitrile, methanol, and ethanol over a wide concentration range. Journal of Chemical & Engineering Data, 69(4), 1493-1502. https://doi.org/10.1021/acs.jced.3c00691
Goorley, T., et al. (2012). Initial MCNP6 release overview. Nuclear technology, 180(3), 298-315. https://doi.org/10.13182/NT11-135
Huynh, D. C., Nguyen, Q. H., Nguyen, T. M. L, Vo, H. N., Tran, T. T., 2019. Validation of gamma scanning method for optimizing NaI(Tl) detector model in Monte Carlo simulation. Appl. Radiat. Isot. 149, 1–8. https://doi.org/https://doi.org/10.1016/j.apradiso.2019.04.009
Huynh, C. D., Truong, S. T., Le, T. T. N., Nguyen, L. T., & Hoang, T. D. (2021). The first result in the determination of the percentage concentration of sulfuric acid solution based on the gamma transmission technique with an energy of 662 ke. Science & Technology Development Journal: Natural Sciences, 5(2), 1179-1188. https://doi.org/10.32508/stdjns.v5i2.1010
Nguyen, T. D., Hoang, T. K. T., & Hoang, D. T. (2024). Determining the concentration of base solution based on gamma transition technique combined with Monte Carlo simulation and artificial neutral network: preliminary results. HCMUE Journal of Science, 21(1), 162. https://doi.org/10.54607/hcmue.js.21.1.3925(2024)
Nguyen, T. T. L., Nguyen, H. D. K., Huynh, D. C., Tran, T. T., Huynh, T. P., & Hoang, D. T. (2024). Determining the thickness of a thin aluminum sheet using the transmission measurement of X-rays with varying energies: A comparative analysis between calibration curve fitting and artificial neural network approaches. Nucl. Instrum. Methods Phys. Res. A., 1068, 169740. https://doi.org/10.1016/j.nima.2024.169740
Perry, R.H., Green, D.W., Maloney, J.O., 1997. CHEMICAL ENGINEERS ’ HANDBOOK SEVENTH Late Editor, Society.
Rhoades, J. D. (1993). Electrical conductivity methods for measuring and mapping soil salinity. Advances in agronomy, 49, 201-251. https://doi.org/10.1016/S0065-2113(08)60795-6
Sbroscia, M., Sodo, A., Bruni, F., Corridoni, T., & Ricci, M. A. (2018). OH stretching dynamics in hydroxide aqueous solutions. The Journal of Physical Chemistry B, 122(14), 4077-4082. https://doi.org/10.1021/acs.jpcb.8b01094
Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural networks, 61, 85-117. https://doi.org/10.1016/j.neunet.2014.09.003
Sinfield, J. V., & Monwuba, C. K. (2014). Assessment and correction of turbidity effects on Raman observations of chemicals in aqueous solutions. Applied spectroscopy, 68(12), 1381-1392. https://doi.org/10.1366/13-07292
Stefanski, J., Schmidt, C., & Jahn, S. (2018). Aqueous sodium hydroxide (NaOH) solutions at high pressure and temperature: insights from in situ Raman spectroscopy and ab initio molecular dynamics simulations. Physical Chemistry Chemical Physics, 20(33), 21629-21639. https://doi.org/10.1039/C8CP00376A
Sue, K., & Arai, K. (2004). Specific behavior of acid–base and neutralization reactions in supercritical water. The Journal of supercritical fluids, 28(1), 57-68. https://doi.org/10.1016/S0896-8446(03)00010-X
Tong, A., Tang, X., Liu, H., Gao, H., Kou, X., & Zhang, Q. (2023). Differentiation of NaCl, NaOH, and β-Phenylethylamine Using Ultraviolet Spectroscopy and Improved Adaptive Artificial Bee Colony Combined with BP-ANN Algorithm. ACS omega, 8(13), 12418-12429. https://doi.org/10.1021/acsomega.3c00271
Truong, T. S., Dang, H. A., Huynh, D. C., Nguyen, T. T. H, Lam, D. N., Nguyen, T. K. A., Tran, T. M. D. & Hoang, D. T. (2021). ANN coupled with Monte Carlo simulation for predicting the concentration of acids. Applied Radiation and Isotopes, 169, 109563. https://doi.org/10.1016/j.apradiso.2020.109563