Neural network analysis of the productivity of biogas plants for small agricultural enterprises

Abstract

The article is devoted to the problem of assessing the productivity of biogas plants. The aim of the work is to build intelligent tools for evaluating the performance of biogas plants by determining the output of biogas depending on the properties of raw materials based on the fuzzy inference method according to the Sugeno algorithm. First of all, the output of biogas is influenced by the chemical composition of the raw materials used. The chemical composition indicators were obtained by the authors in the framework of experimental studies. To carry out the analysis, a knowledge base was built on the following parameters: humidity, crude ash content, crude fat content, crude protein content, crude fiber content, nitrogen-free extractive substances content. The fuzzification of its vertices in the section of 2- and 3-term sets has been carried out. Membership functions of fuzzy sets for each parameter are constructed. The fuzzification of the root is defined in 5 categories. A system of rules was compiled based on experimental data, and the biogas yield was calculated depending on the initial parameters. The results obtained can be used in the organization of biogas plants.