A novel seasonal grey model for forecasting the quarterly natural gas production in China

Abstract

As a green and clean source of energy, natural gas plays an indispensable role in changing the energy structure and reducing greenhouse gas emissions. In this context, forecasting natural gas production becomes particularly important. In China, natural gas production has obvious seasonal and cyclical characteristics. To this end, a grey prediction model combined with particle swarm optimization (PSO) is proposed: the PSO-based data grouping grey model with a fractional order accumulation (PSO-FDGGM (1,1)). The traditional grey model has shortcomings such as low prediction accuracy and poor adaptability, and its application is limited. Compared with the traditional grey models, this model can be used to predict time series with seasonal characteristics, and the prediction accuracy is higher. Based on the data grouping-based grey modelling method (DGGM (1,1)), this model introduces the modeling method of the grey system model with the fractional order accumulation (FGM (1,1)), that is, combining data grouping with fractional accumulation, and using PSO to optimize fractional-order parameters, which improves the performance and adaptability of the model. In order to verify the effectiveness of the model, the model is compared with the traditional grey model (GM (1,1)), FGM (1,1) model, DGGM (1,1) model, autoregressive comprehensive moving average (ARIMA) model and the PSO-based grey forecasting model (PSOGM (1,1)). The mean absolute percent errors (MAPEs) of the comparison model in the test set are 7.46%, 9.28%, 4.38%, 8.60% and 5.97% respectively, while that of PSO-FDGGM (1,1) model is only 3.19%. The results show that PSO-FDGGM (1,1) has good adaptability and prediction accuracy for seasonally fluctuating sequences. In order to further verify the effectiveness and applicability of PSO-FDGGM (1,1), this paper adopts two ways: adjusting the parameters of PSO algorithm in the new model and applying the new model to the prediction of solar power generation, and the empirical results show that the new model is effective in forecasting seasonal data. Finally, the new model is used to predict the quarterly production of natural gas in China from the third quarter of 2021 to the second quarter of 2024. The forecast results can provide a basis for formulating natural gas production plans and environmental policies.