بهینه سازی انرژی و مدلسازی پیش بینی صنایع پتروشیمی
Abstract
1- Introduction
2- Related work
3- Convolutional neural network integrating cross-feature
4- Benchmark verification of UCI data sets
5- Case study: energy optimization and prediction modeling of petrochemical industries
6- Conclusion
Acknowledgments
References
Abstract
he petrochemical industry is the top priority of the national economy and sustainable development. For the purpose of improving the energy efficiency in the petrochemical industry, an energy optimization and prediction model based on the improved convolutional neural network (CNN) integrating the cross-feature (CF) (CF-CNN) is proposed. The CF can combine the correlation between features to obtain the input of the CNN, which can avoid over-fitting problems caused by fewer features. Then the CNN is designed as a threelayer structure and the Rectified Linear Unit (ReLU) is introduced to achieve better generalization capability and stability with boiler fluctuations in the petrochemical industry. The developed method has better performances of modeling accuracy and applicability than that of the back-propagation (BP) neural network and the extreme learning machine (ELM) on University of California Irvine (UCI) benchmark datasets. Furthermore, the developed method is applied to establish an energy optimization and prediction model of ethylene production systems in the petrochemical industry. The experimental results testify the capability of the proposed method. Meanwhile, the average relative generalization error is 2.86%, and the energy utilization efficiency increases by 6.38%, which leads to reduction of the carbon emissions by 5.29%.
Introduction
The petrochemical industry is the top priority of the national economy and sustainable development. And the ethylene industry is a key part of complex petrochemical production industry. Currently, the ethylene is one of the demanding basic chemical materials [1] and the ethylene production consumes more than 15% of energy (including fuels and materials) in thousands of chemical products [2]. However, when the ethylene is produced by cracking, the total energy loss exceeds 45% [3]. Along with China’s rapid development, the total ethylene production capacity in 2017 increased to 23.21 million tons and the ethylene equivalent consumption grew up to 10% [4].
However, the fast growth of the ethylene production leads to the increase of the energy consumption and carbon emissions, and reduction of the ethylene production efficiency. Therefore, how to increase the ethylene production efficiency and reduce the carbon emissions has become a problem in the world. Nowadays, with the high-speed development of artificial intelligence, more and more energy optimization and analysis models have been used to raise productivity and energy efficiency in complex petrochemical processes. Geng et al. developed an artificial neural network (ANN) based on self-organizing cosine similarity, which overcame shortcomings of the single-hidden layer network in building the ethylene production prediction model [5]. An index decomposition analysis (IDA) method, combined with an ANN and a data envelopment analysis (DEA) was proposed by Olanrewaju et al. to assess energy consumption [6] and the energy capacity of industry in South African [7]. However, the process of extracting features in experiments is very complicated and the above models are not stable for fluctuations in complex industrial production.