The Journal of Grey System ›› 2026, Vol. 38 ›› Issue (5): 29-48.

Previous Articles     Next Articles

Seasonal Grey Bernoulli Model Based on Fourier Series for Energy Production Forecasting

  

  1. 1. School of Mathematics and Statistics, Shanghai Polytechnic University, Shanghai, 201209, P.R.China
    2. College of Computer and Information Engineering, Shanghai Polytechnic University, Shanghai, 201209, P.R.China
  • Online:2026-11-16 Published:2026-11-16

Abstract: Energy production is influenced by multiple factors and exhibits a complex interplay of nonlinearity, seasonality, and randomness, complicating accurate forecasting. To meet this challenge, a novel seasonal Fourier grey Bernoulli model (SFGBM(1,1)) is proposed that integrates a Fourier series into the nonlinear terms of the grey Bernoulli model. The parameter estimation and forecasting formula are derived. By independently optimizing the truncation orders of the cosine and sine harmonics, the Fourier basis reproduces seasonal variations with the minimal number of components, yielding a maximally concise structure. The fractional exponential order of the Bernoulli term simultaneously captures the inherent nonlinearities of the series. These hyperparameters are selected jointly by a hybrid of the flower pollination algorithm (FPA) and grid search. Six synthetic scenarios that span varying degrees of nonlinearity and periodicity, together with two real monthly series (natural-gas production and wind-power generation), demonstrate that the proposed model simultaneously captures seasonality, nonlinear dynamics and high-frequency amplitude changes. It outperforms benchmark grey seasonal models, nonlinear grey models, statistical models and machine-learning methods, yielding the lowest in-sample and out-of-sample errors.  The SFGBM(1,1) model achieves test MAPEs of 5.02% for wind power generation in Xinjiang and 1.69% for Chinese natural gas production, improving upon the best benchmarks by 28.05% and 29.45%, respectively.

Key words: Fourier series, grey Bernoulli model, seasonal fluctuations, energy production forecasting