Skip to content
DSML.KZ
Publications

Talk · IEEE International Conference on Smart Information Systems and Technologies · May 13, 2026

Predicting Volatile Fatty Acid Concentration in Anaerobic Digesters Using Routine Measurements and AI-Based Machine Learning

Abdulaziz Abdukarimov, Miras Mukazhan, Aruzhan Baisiikhova, Adilet Kakharov, Waldemar Wójcik, Баглан Иманбек

Adilet Kakharov
Open via DOI

Abstract

Accurate prediction of volatile fatty acid concentration in anaerobic digesters is important for monitoring and early detection of reactor instability. This study presents a machine learning approach that predicts total volatile fatty acids using three routine measurements: pH, carbon dioxide partial pressure, and total ammonia nitrogen. A dataset of 533 observations from laboratory-scale experiments was divided into training and test subsets, and missing values in the input data were imputed using an iterative method with a random forest estimator. The final model, trained with an automated ensemble learning framework, achieved strong performance on the independent test set, with a coefficient of determination of 0.8266 and a low mean absolute error. Analysis of feature importance showed that pH contributed the most to prediction accuracy, followed by carbon dioxide partial pressure and total ammonia nitrogen. The results indicate that combining routine online measurements with data-driven modeling provides an effective soft-sensing tool for estimating volatile fatty acid concentration without relying on expensive analytical equipment.**This research was funded by the Science Committee of the Ministry of Science and Higher Education of the Republic of Kazakhstan (Grant No. AP26103739).

Anaerobic Digestion and Biogas ProductionWater Quality Monitoring and AnalysisMicrobial Fuel Cells and Bioremediationvolatile fatty acidsanaerobic digestionanaerobic digesters

More by Adilet Kakharov