|
NEURAL NETWORKS WITH PSEUDO-RANDOM DISTRIBUTION OF RELATIONSHIPS USING THE EXAMPLE OF MERCURY ELECTROLYZER OPERATION MODE MODELING |
|---|---|
| รหัสดีโอไอ | |
| Creator | Olga Anatolievna Medvedeva, Aleksandr Nikolaevich Ivanov, Nikolay Danilovich Morozkin, Svetlana Anatol'evna Mustafina |
| Title | NEURAL NETWORKS WITH PSEUDO-RANDOM DISTRIBUTION OF RELATIONSHIPS USING THE EXAMPLE OF MERCURY ELECTROLYZER OPERATION MODE MODELING |
| Contributor | - |
| Publisher | TuEngr Group |
| Publication Year | 2562 |
| Journal Title | International Transaction Journal of Engineering, Management, & Applied Sciences & Technologies |
| Journal Vol. | 10 |
| Journal No. | 16 |
| Page no. | 10A16A: 1-12 |
| Keyword | Neural networks Modeling, Machine learning, Hyperparameters, Electrolysis, Pseudorandom distribution of connections. |
| URL Website | http://tuengr.com/Vol10_16.html |
| Website title | ITJEMAST V10(16) 2019 @ TuEngr.com |
| ISSN | 2228-9860 |
| Abstract | This article analyzed the applicability of artificial neural networks to solve the problems of physicochemical process modeling using the example of the mercury electrolyzer operation mode used in caustic soda production. This paper also described the basic qualities of the existing neural networks and the ways of their training. The authors propose the solution to the problem of modeling based on the networks with pseudo-random distribution of connections. This paper described the architecture of these networks, three learning algorithms are proposed. The implementation of neural networks with pseudo-random distribution of connections was performed by Python programming language. The article presents the comparative learning results of different networks with different sets of hyperparameters. Also, the determination of the optimal settings of neural networks allows achieving high learning efficiency. The resulting neural network model described the electrolysis process adequately in accordance with the available source data. |
