Predicting monthly natural gas production in China using a novel grey seasonal model with particle swarm optimization

Abstract

Accurate prediction of short and medium-term monthly natural gas production in a country is the basis for understanding the supply capacity of natural gas in different months, and for the timely adjustment of natural gas production and import strategies. In China the monthly production of natural gas has obvious seasonal and cyclical variations, thus the use of a traditional grey prediction model is not very effective. As a result, a novel grey seasonal model is proposed in this paper. This is the Particle swarm optimized Fractional-order-accumulation non-homogenous discrete grey Seasonal Model (PFSM(1,1) model). This model enhances the adaptability to seasonal fluctuation data in two ways: the seasonal adjustment of the original data, and improvement of model self-adaptability. We use monthly natural gas production data of China for the period 2013-2018 as samples to predict those for the period 2019-2023. To demonstrate the PFSM(1,1) model does indeed exhibit better predictive capability, we also use the Holt–Winters model and a seasonal GM(1,1) model to predict monthly natural gas production, and compare the results with the model proposed here. The prediction results show that monthly natural gas production in China will continue to increase throughout the 2019-2023 period, that the peak-to-valley differences in monthly production values will also increase, and that the seasonal variations in production will become increasingly pronounced. Moreover, although Chinese production of natural gas is increasing, it will still be difficult to meet future demand, and hence the gap between supply and demand will increase year by year. We conclude that China needs to develop a more complete import plan for gas to meet expected natural gas consumption.