Elvianto Dwi Hartono, Anang Lastriyanto, Elok Zubaidah, Yusuf Hendrawan
Heat penetration significantly influences the efficiency and sustainability of various thermal systems. The development of a heat penetration prediction model using Artificial Intelligence (AI) aims to understand and optimize heat transfer processes to maintain honey quality during ohmic heating. The utilization of Neural Network (NN) methods, particularly Long Short-Term Memory (LSTM), for time-series data processing involves key processing variables, such as time, frequency, initial temperature, final temperature, current, and heat rate values. By applying multivariate LSTM for prediction, three data models were developed, each consisting of three submodels. Additionally, parameter tuning was performed, including the number of training data points, the number of neurons in the hidden layer, density, the number of epochs, and batch size. Models B1 and B3, with their specific parameters, demonstrated the best performance in predicting heat penetration, characterized by optimal error functions and minimal error propagation. These models successfully identified data patterns for predicting heat penetration, proving their effectiveness. © by the authors.
Department of Agroindustrial Technology, Faculty of Agricultural Technology, Universitas Brawijaya, Malang, Indonesia; Department of Robotics and Artificial Intelligence Engineering, Faculty of Engineering, Universitas 17 Agustus 1945, Surabaya, Indonesia; Department of Biosystem Engineering, Faculty of Agricultural Technology, Universitas Brawijaya, Malang, Indonesia; Department of Food Science and Biotechnology, Faculty of Agricultural Technology, Universitas Brawijaya, Malang, Indonesia
