Generalized model of prediction of natural gas consumption

We herein present the basic characteristics of a model developed at the Distribution Division of ENARGAS, intended to predict the natural gas consumption of the major cities of Argentina. This model is an extension and generalization of the original model we developed to predict the daily gas consumption in the short range (1 to 5 days). In its original form, the model has been used successfully to predict the daily gas consumption of the Greater Buenos Aires (GBA) area and most of the major cities of Argentina. The model is able to predict the consumption 1 to 5 days in advance with 10% of uncertainty. The new version presented here, extends its applicability to the intermediate range (1 to 5 years). It allows us to estimate the annual peak consumption, load factors and the optimal transportation capacity for a given region of interest. We also present a novel procedure to obtain the distribution of daily consumption from the monthly consumption. This procedure can be used to obtain the parameters of the model, the peak consumption and load factor of the different types of gas consumers for a given region, using the information of consumption that can be obtained from the monthly billing.

[1]  C. G. Lambe,et al.  Applied mathematics for engineers and scientists , 1959 .

[2]  Ronald H. Brown,et al.  Development of artificial neural network models to predict daily gas consumption , 1995, Proceedings of IECON '95 - 21st Annual Conference on IEEE Industrial Electronics.

[3]  Modeling the Belgian gas consumption , 1997, 1997 European Control Conference (ECC).

[4]  Alireza Khotanzad,et al.  Combination of artificial neural-network forecasters for prediction of natural gas consumption , 2000, IEEE Trans. Neural Networks Learn. Syst..