MODELLING OF DEEP LEARNING EMPOWERED PREDICTIVE MODEL FOR ETHYLENE PRODUCTION IN HYDROPOWER INDUSTRIES

摘要

In recent times, hydropower becomes low-carbon renewable energy with advanced technology and flexible operation. Hydropower, as a power generation form, has fewer data samples than thermal power generation, which results in the tedious issue of designing a precise production and energy prediction model. So, an efficient production capacity predictive approach by the use of the Deep learning (DL) model is needed for energy optimisation and saving. In this aspect, this paper presents a new DL-enabled predictive model for ethylene production in hydropower industries. The proposed model aims to design a sunflower optimisation algorithm with Long short-term memory (LSTM), called the SFO-LSTM model to predict the productivity of ethylene in hydropower industry. The proposed SFO-LSTM model involves the design of LSTM-based predictive process to estimate the value, which is higher closer to the actual production value. In addition, the hyperparameter optimisation of the LSTM model can be optimally modified by the SFO algorithm and thereby boost the classifier results. A comprehensive simulation analysis is carried out and the results are investigated under varying numbers of hidden layers. The experimental results stated that the SFO-LSTM model has accomplished maximum prediction outcomes under different aspects.