Rizqi Fitri Naryanto, Mera Kartika Delimayanti
Machine learning was utilized in the study to construct prediction models for predicting Net Output Power based on various characteristics such as moisture content and other materials. The temperature of the gasifier and the air-to-fuel ratio are other crucial factors in the prediction model. According to the findings, a Neural Network could accurately forecast the Net Output Power Yield based on biomass properties. In addition, the relative contribution of various percentages of elements on biomass, such as carbon (23.3-55.8%) and volatile materials (47.8-86.3%), was shown in the form of diverse feedstocks. Structural information was more essential than element compositions for biomass characteristics to predict Net output power yield effectively. RapidMiner software was used to perform the machine learning prediction method in this study. For the performance evaluation of the prediction model, the RMSE (Root Mean Square Error) was used. For predicting net output power from different feedstocks of solid biomass fuels, these models are the best available. The current study brought fresh insights into biomass characteristic analysis, which improved output power. This research indicates that machine-learning approaches can be implemented to predict the net outputs power successfully. © 2023 Author(s).
Department of Mechanical Engineering, Universitas Negeri Semarang, Semarang, 50229, Indonesia; Department of Computer and Informatics Engineering, Politeknik Negeri Jakarta, Depok, 16425, Indonesia
