Production capacity analysis and energy saving of complex chemical processes using LSTM based on attention mechanism

Abstract

The production data of complex chemical processes are multi-dimensional, uncertain and noisy, and it is difficult to directly control raw materials consumption and measure the product quality. Therefore, this paper proposes a production capacity analysis and energy saving model using long short-term memory (LSTM) based on attention mechanism (AM) (AM-LSTM). The weights of the results sequence in the hidden layer, which have great influence on final results in the output layer, are calculated by the AM. Then the production prediction model is built using the LSTM to extract features of the input data and multiple time series results of the hidden layer. Compared with the common LSTM, the multi-layer perceptron (MLP) and the extreme learning machine (ELM), the applicability and the effectiveness of the proposed model is validated based on University of California Irvine repository (UCI) datasets. Finally, the proposed model is applied to analyze the production capacity and the energy saving potential of the purified terephthalic acid (PTA) solvent system and the ethylene production system of the complex chemical process. The experimental results verify the practicability and accuracy of the proposed model. Furthermore, the results offer the operation guidance for production capacity improvement through saving energy and reducing the energy consumption.