A novel model for predicting natural gas consumption - S.Gil and J. Deferrari- IGRC Vancouver 2004 1 A novel model for predicting natural gas consumption S.Gil ‡ and J. Deferrari Distribution Division of ENARGAS. ENARGAS – Natural Gas Regulatory Agency of Argentina. Suipacha 636 - 4p. (1008) Cap. Fed. - Argentina. FAX: (5411) 4348 -0532/0535-e-mail: sgil@df.uba.ar International Gas Research Conference (IGRC) Vancouver – November 2004 ABSTRACT: We present a novel model intended to predict mainly the residential and commercial natural gas consumption in urban areas, for the short and intermediate ranges of time. In the short range, the model has been successfully used to forecast the daily gas consumption of major cities of Argentina. It is able to predict the consumption 1 to 5 days in advance with 10% of uncertainty. In the intermediate range (1 to 5 years), the model 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 derived from the monthly billing. I.- INTRODUCTION The prediction of natural gas consumption is crucial for the gas distribution and transportation companies as well as for the government agencies associated to the natural gas sector. In particular, the short range prediction, 1 to 5 days, is important to ensure the normal supply of natural gas to a given city or community. This type of prediction is particularly important for countries like Argentina, where the production sites are far from the major centers of consumption. In the case of Argentina, these distances are of about 2000 km. This circumstance and the lack of large reservoirs make it necessary to develop reliable models to predict the gas consumption a few days in advance. The present study addresses the issue of residential, commercial and Natural Gas Vehicle Service Stations (NGV-SS), that have non-interruptible gas services (firm contract). Most large industries and electric plants in Argentina have interruptible contracts. Therefore their consumption patterns are different to the non-interruptible component and will not be discussed in this paper. We will use the terms firm and non interruptible to designate this type of component of consumption. There is also a need to predict the firm component of consumption in the intermediate range of time, within 1 to 4 years, in order to adapt and upgrade the infrastructure of transportation and distribution. This type of prediction is also useful for all the sectors of the gas industry that need to plan their production and optimize their anticipated purchase. The most important factors that affect the gas consumption of residential and commercial users are temperature, day-of-the-week (holiday or working day) and prevailing scenario of consumption. Other factors that may also influence the consumption are: wind speed and its direction, humidity, etc. Due to the lack of reliable information on these parameters, we have not included them in our model.