Forecasting U.S. shale gas monthly production using a hybrid ARIMA and metabolic nonlinear grey model

doi:10.1016/j.energy.2018.07.047

Energy

Volume 160, 1 October 2018, Pages 378-387

https://doi.org/10.1016/j.energy.2018.07.047 Get rights and content

Highlights

•
Forecasting U.S. shale gas production can help better understand the gas market.
•
Better forecasting U.S. shale gas production requires better forecasting model.
•
A linear forecasting technique is used to correct nonlinear predictions.
•
The proposed technique performs more accurate than other three techniques.
•
The proposed technique can better predict shale gas and other fuels production.

Abstract

Changes in shale gas production directly determine natural gas output in the United States (U.S.), and indirectly impact the global gas market. To better forecast shale gas output, we hybridized a nonlinear model with a linear model to develop a metabolic nonlinear grey model–autoregressive integrated moving average model (or MNGM-ARIMA). The proposed hybrid forecasting technique uses a linear model to correct nonlinear predictions, which effectively integrates the advantages of linear and nonlinear models and mitigates their limitations. Based on existing U.S. monthly shale gas output data, we applied the proposed hybrid technique to forecast U.S. monthly shale gas output. The results show that the proposed MNGM-ARIMA technique can produce a reliable forecasting results, with a mean absolute percent error of 2.396%. Then, using the same set of data, we also ran three other forecasting techniques developed by former researchers: the metabolic grey model (MGM), ARIMA, and non-linear grey model (NGM). The results of the comparison show that the proposed MNGM-ARIMA technique has the smallest mean absolute percent error. This indicates the proposed hybrid technique can produce more accurate forecasting results. We therefore conclude that the proposed MNGM-ARIMA technique can service us better forecasting shale gas output, as well as other fuels output.

Introduction

The U.S. shale gas production in 2016 doubled compared to 2011, reaching 16,582.41 billion cubic meters [1], which has been heavily impacting on energy markets in the U.S and the world [[2], [3], [4]]. The large-scale exploitation of shale gas production has exerted influence on the energy market in U.S. from the following three aspects. First of all, it changed supply dynamics for U.S. natural gas, leading to changes in the structure of supply and demand [4,5]. More in detail, the surge in shale gas production has met natural gas demand and caused the U.S. to overtake Russia as the world's largest gas producer. Secondly, the development of the supply side of natural gas has promoted the independence of natural gas in the United States [6]. On the basis of self-sufficiency of natural gas, in the first half of 2017, the U.S. became a natural gas net exporter. Finally, the increase in shale gas supply caused a drop in U.S. natural gas prices [7,8]. In 2016, natural gas prices reached 1.61 USD/million British thermal units. This price was one-third to one-fifth the Asian market price and hit a new 20-year low [9,10]. In addition to its impact on the United States, the import and export of domestic natural gas also exerted influence on the world energy market. The changes in shale gas production directly determine the net exports of natural gas from the U.S., resulting in an increase in export and a decrease in imports of global natural gas energy [11]. Better predicting future U.S. shale gas production can provide reference information about the natural gas market in the U.S. and the world.

To better forecast U.S. shale gas output, we hybridized a nonlinear model (Metabolic nonlinear grey model, or MNGM (MNGM (1,1, α)) with a linear model (Autoregressive Integrated Moving Average, or ARIMA) to develop the MNGM (1,1, α)-ARIMA forecasting technique. The method adopts the principle of correcting nonlinear prediction values using a linear model and applies two different models to a combined model. This method broadens the scope of application for the prediction model, and provides a reference value to predict data with large jumps. The forecasted result will provide a reasonable and effective reference value for global energy policy makers and natural gas market participants.

The remainder of this paper is organized as follows. Section 2 reviews the existing literature in this field. The related methodologies of the MNGM (1,1, α)-ARIMA are introduced in Section 3. In Section 4, the forecasting process and evaluation of prediction merit are illustrated with the data of the shale gas production of the United States. The conclusion is given in Section 5.

Section snippets

Literature review

Since the predictive model used in this paper is the MNGM (1,1, α)-ARIMA model, this study will comb the existing researches with several sub-models involved in this combined model, which include grey model, the improved grey model and the ARIMA model. In addition, clues to the literature review will be given at the end of this section. Based on this, the idea of constructing this combined model will be clearly presented.

Methodology

The computational principle of the combined MNGM-ARIMA model is closely related to each single model, MNGM and ARIMA. For the calculation formula, the formula of this combination model is made up by two. For the calculation flow, the steps of the combined model are connected by both. In this section, the calculation steps will be followed to demonstrate the model construction process. Before that, the respective calculation steps of the two single models are briefly described here.

For metabolic

Empirical results

In this chapter, the implementation of the MNGM (1,1, α)-ARIMA model for shale gas production forecasting will be demonstrated. The monthly production data for shale gas period 2014–2016 used in this paper were derived from the U.S. Energy Information Administration (EIA) [1]. A 2D area plot of the monthly data for this period shows there is no seasonal trend in shale gas production, further indicating the reasonableness of developing forecasts using monthly data. The chart below shows that

Conclusion

This study developed and applied a metabolic nonlinear grey-autoregressive integrated moving average (MNGM (1,1, α)-ARIMA) model to forecast U.S. shale gas monthly production. This new model combines linearity and non-linearity to overcome the bias of a single model in predicting results. In addition, time series corrections significantly broaden the scope of application for the combined model. This makes it reliable when applying monthly data for forecasting. This provide guiding value for

Acknowledgement

The authors would like to thank the anonymous reviewers for their careful and thorough reading of this manuscript and for the thoughtful comments and constructive suggestions, which help to improve the quality of this manuscript. This work is supported by the Shandong Provincial Natural Science Foundation, China (ZR2018MG016), the Initial Founding of Scientific Research for the Introduction of Talents of China University of Petroleum (East China) (YJ2016002), and the Fundamental Research Funds

References (42)

Q. Wang et al.
Research status of shale gas: a review
Renew Sustain Energy Rev
(2017)
Q. Wang et al.
Natural gas from shale formation – the evolution, evidences and challenges of shale gas revolution in United States
Renew Sustain Energy Rev
(2014)
S.P.A. Brown
Natural gas vs. oil in U.S. transportation: will prices confer an advantage to natural gas?
Energy Pol
(2017)
Q. Wang et al.
Natural gas from shale formation: a research profile
Renew Sustain Energy Rev
(2016)
A. Saussay
Can the US shale revolution be duplicated in continental Europe? An economic analysis of European shale gas resources
Energy Econ
(2018)
S. Karmakar et al.
Grey fuzzy optimization model for water quality management of a river system
Adv Water Resour
(2006)
U. Kumar et al.
Time series models (Grey-Markov, Grey Model with rolling mechanism and singular spectrum analysis) to forecast energy consumption in India
Energy
(2010)
H.T. Pao et al.
Modeling and forecasting the CO emissions, energy consumption, and economic growth in Brazil
Energy
(2011)
C.S. Lin et al.
Grey forecasting model for CO 2 emissions: a Taiwan study
Appl Energy
(2011)
D. Akay et al.
Grey prediction with rolling mechanism for electricity demand forecasting of Turkey
Energy
(2007)

P. Zhou et al.

A trigonometric grey prediction approach to forecasting electricity demand

Energy

(2006)

H. Zhang et al.

Application of grey modeling method to fitting and forecasting wear trend of marine diesel engines

Tribol Int

(2003)

H. Zhao et al.

An optimized grey model for annual power load forecasting

Energy

(2016)

S. Arora et al.

Rule-based autoregressive moving average models for forecasting load on special days: a case study for France

Eur J Oper Res

(2018)

S. Barak et al.

Forecasting energy consumption using ensemble ARIMA–ANFIS hybrid algorithm

Int J Electr Power Energy Syst

(2016)

C.C. Wang

A comparison study between fuzzy time series model and ARIMA model for forecasting Taiwan export

Expert Syst Appl

(2011)

EIA

U.S. Energy Information Administration

(2017)

A. Paylor

The social–economic impact of shale gas extraction: a global perspective

Third World Q

(2017)

O.L. Anderson

Shale gas revolution or evolution: opportunities and challenges for europe

ONE J

(2017)

J. Gao et al.

Design and optimization of shale gas energy systems: overview, research challenges, and future directions

Comput Chem Eng

(2017)

M. Harleman et al.

Natural resource ownership, financial gains, and governance: the case of unconventional gas development in the UK and the us

(2017)

Cited by (109)

Novel method for total organic carbon content prediction based on non-equigap multivariable grey model
2024, Engineering Applications of Artificial Intelligence
Predicting total organic carbon content plays a crucial role in shale gas reservoir evaluation. To address multi-variable, imperfect logging data and non-equigap characteristics, this study developed a multi-variable grey derivative model for predicting the total organic carbon content trend. To begin with, the non-equigap factor is introduced and the parameter identification of the new model is established on the basis of least square method. Then, to reduce fluctuations in the original data, the weighted geometric average weakening operator is introduced. In addition, considering the system stability, the derivation method is used to calculate the time response formula of the model. Furthermore, two validation cases are provided to verify the validity of this novel model. Finally, the model is applied to the Fuling shale gas field in China for practical forecasting. Compared to other models, such as the grey model, statistical model, and artificial intelligence model, the proposed model achieved higher prediction accuracy in the three cases. Specifically, this model obtained 95.62%, 95.88% and 94.55% prediction accuracy for total organic carbon content. Results that the new model effectively addresses the limitations of previous studies that focus on single-parameter or equal spacing issues.
Grey prediction of carbon emission and carbon peak in several developing countries
2024, Engineering Applications of Artificial Intelligence
The carbon emissions in BRICS countries (Brazil, Russia, India, China and South Africa) account for nearly half of global carbon emissions. To explore the future carbon emissions in BRICS countries and the carbon peak in China and India, the grey Gompertz model with new information priority accumulation is proposed. Intelligent algorithms are used in the parameter optimization. The new model not only increases the influence weight of new information on the predicted results, but also changes the predicted trend of traditional grey prediction model. The effectiveness of the new model for the carbon emission forecast is verified through the known data set. According to the predicted results, the carbon emissions in Russia, Brazil, South Africa tend to stabilize over the long term. The carbon emissions in China and India are rising slowly. China will achieve the carbon peak in 2029 with 10.2 billion tons of carbon emissions. India will not achieve the carbon peak by 2030. The current situations of the carbon emissions in BRICS countries are analyzed. It is concluded that the energy structure, industrial development and the urbanization need to be paid attention to. Detailed policies and suggestions are given in the fourth section. The above research results indicate that the new model can predict the limit value of the carbon emission. The predicted results can provide data support for the policy formulation in carbon reduction.
A fusion gas load prediction model with three-way residual error amendment
2024, Energy
Accurately predicting gas load is crucial for optimal planning and scheduling of natural gas production. Existing machine learning or deep learning-based prediction methods primarily aim to improve accuracy. However, these methods face challenges such as overlapped features, significant background noise, and residual errors. In response to these challenges, we proposed a fusion prediction model, known as the Double Temporal Convolutional Network (TCN) and Light Gradient Boosting Machine (LightGBM) with Three-way Residual Error Amendment (DT-LGBM-3WREA). To address the overlapped feature issue in time series data, we utilize the STL module for data decomposition. To handle background noise, the network incorporates a double TCN module specifically designed for trend and seasonal components. Thus, it can better avoid the transmission of noise between modules. To address residual error, we introduce three-way decision methodology and develop the 3WREA module, which aims to avoid local optima and random factors from perturbing the prediction accuracy. This approach mitigates the influence of remainder item on overall prediction performance. Experimental results demonstrate that the DT-LGBM-3WREA model excels in gas load prediction scenarios, surpassing state-of-the-art methods.
COVID-19 pandemic, Russia-Ukraine conflict and shale gas development: Evidence from fractional integration
2024, Heliyon
Although there are papers on the persistence of energy series including the persistence of shale gas, the impact of recent developments such as the Covid-19 pandemic and Russia-Ukraine conflict have been rarely explored in the existing literature. This paper examines the structure of shale gas production in the U.S. by looking at the degree of persistence across different areas, with the aim to determine if shocks in the series are permanent or transitory. Using fractional integration methods (which unlike the conventional methods, allow for the determination of the persistence of energy and non-energy series in a robust manner), and different subsamples that include the Covid-19 pandemic and the Russia-Ukraine war, our results indicate that there is a substantial decrease in the integration order in the total shale gas production in the U.S. as well as in four other plays-Haynesville, Permian, Utica and Eagle Ford. However, no differences are observed with respect to the Russia-Ukraine war. There is another group of four series (Marcellus, Niobrara-Codell, Woodford and Rest of US 'shale’) with a very small reduction in the degree of persistence and another group of three series with almost no reduction at all in the order of integration (Barnett, Mississippian and Fayetteville). Several implications in terms of policy are reported at the end of the manuscript.
A non-linear grey Fourier model based on kernel method for seasonal traffic speed forecasting
2024, Communications in Nonlinear Science and Numerical Simulation
This article introduces a non-linear grey Fourier model utilizing the kernel method to better characterize seasonal traffic speed with fluctuating trends. The linear trend term in the grey Fourier model is replaced with feature mapping, resulting in a non-linear form. Utilizing Lagrange multipliers and a kernel function, the parameter estimation is transformed into a linear equation-solving problem. Order selection is carried out via power spectrum analysis, and hyper-parameters are determined using Bayesian optimization. Two numerical examples from real problems are presented to validate the model’s performance. The results indicate that the developed model outperforms nine models across four categories. Specifically, the model is applied to predict traffic speed in Guangzhou, and comparative results with four neural networks demonstrate its commendable performance.
A multi-period-sequential-index combination method for short-term prediction of small sample data
2024, Reliability Engineering and System Safety
Based on multi-period-sequential-index combination (MPSIC) technology, three forecasting methods (auto-MPSIC, IV- MPSIC, MSEI-MPSIC) were proposed for short-term prediction of small sample data. Natural gas datasets, coal datasets, electricity datasets and atmosphere datasets were separately tested by using MPSIC method, and then impact of weighting factors, forecasting accuracy analysis were carried out for MPSIC method as well as other comparative methods. The results showed that, auto-MPSIC method was partial to use statistical indicators, such as peak-to-peak, average, root mean square, to decrease prediction error, and meanwhile was also inclined to use sequential index at time of t_i-₁ next to t_i to improve prediction accuracy. It was also concluded that: the proposed MPSIC method could achieve higher prediction accuracy compared with other methods; the robustness of auto-MPSIC method was slightly better than that of IV-MPSIC and MSEI-MPSIC under condition of noisy data, which was attributed to an adaptive weight allocation technology considering statistical distribution of forecasting errors.

View all citing articles on Scopus

View full text

Forecasting U.S. shale gas monthly production using a hybrid ARIMA and metabolic nonlinear grey model

Highlights

Abstract

Introduction

Section snippets

Literature review

Methodology

Empirical results

Conclusion

Acknowledgement

Renew Sustain Energy Rev

Renew Sustain Energy Rev

Energy Pol

Renew Sustain Energy Rev

Energy Econ

Adv Water Resour

Energy

Energy

Appl Energy

Energy

Energy

Tribol Int

Energy

Eur J Oper Res

Int J Electr Power Energy Syst

Expert Syst Appl

U.S. Energy Information Administration

The social–economic impact of shale gas extraction: a global perspective

Third World Q

Shale gas revolution or evolution: opportunities and challenges for europe

ONE J

Design and optimization of shale gas energy systems: overview, research challenges, and future directions

Comput Chem Eng

Natural resource ownership, financial gains, and governance: the case of unconventional gas development in the UK and the us