Intelligent ensemble of voting based solid fuel classification model for energy harvesting from agricultural residues

In recent times, utilization of renewable energy resources for transportation and electric power generation can be the sustainable way that reduces the risk of environmental, climatic, economic, political, and security concerns related to fossil fuel combustion. Among the several renewable energy sources, biomass is commonly used due to the fulfilment of ecological compatibility due to the fact that it is attained from plant and animal waste. At the same time, the classification of fuel materials becomes difficult when the material is previously processed or gathered from an environment that makes it challenging to discern. Therefore, with the constant development and diversification of energy harvesting from agricultural residues, there is a requirement to design a model for the classification of solid fuels. The recent developments of machine learning (ML) and deep learning (DL) techniques can be used for the solid fuel classification. The ML and DL models comprise many interdisciplinary areas, such as statistics, mathematics, artificial neural networks, data mining, optimization, and artificial intelligence. With this motivation, this paper designs a new intelligent ensemble of voting based solid fuel classification (IEVB-SFC) model for energy harvesting from agricultural residue. The proposed IEVB-SFC technique involves different stages of operations such as data acquisition, data preprocessing, classification, and ensemble process. At the primary stage, the data preprocessing is carried out in three different ways such as data transformation, class labeling, and data normalization. Besides, the IEVB-SFC technique comprises three different DL models as long short term memory (LSTM), gated recurrent unit (GRU), and convolutional neural network based LSTM (CNN-LSTM). Finally, an ensemble of three DL models takes place by the use of voting technique and thereby determines the appropriate solid fuel class labels, show the novelty of the work. The experimental results showcased the betterment of the IEVB-SFC technique over the recent state of art techniques with the maximum accuracy of 0.97.