Stochastic analysis of natural gas consumption in residential and commercial buildings

Consumer consumption characteristic is an important asset for safe design and management of gas distribution networks. Different characteristics of natural gas consumption in residential and commercial buildings are studied from statistical and stochastic points of view. The technique is applied during 2008 and 2009 to a densely populated district in Tehran, Iran, with relatively large number of buildings (67,655 residential and 13,286 commercial buildings). There are different trends in the histograms of gas consumption, but there is a general trend in diagrams of probability index (the probability of gas consumption exceeding a specific value) and their regressions. The most frequent amount of gas consumption for all 45-day periods is 100 m³ as compared with the annual average of 320 m³ for residential buildings. The latter reduces to 80 m³ for the averaged periodic consumption per unit in a building. Also it seems that the most frequent amount of periodic gas consumption of residential buildings is about 31% of their respective annual average during the warm months of the year, and 150% during the cold months. Periodic consumptions less than 1500 m³ and average consumptions less than 1400 m³ are more probable in residential buildings, which are larger than that of commercial ones, but this trend reverses at higher consumption values. If actual consumption is normalized by the average consumption, the number of units in the building or the floor area, the probability index of commercial buildings is generally higher than residential ones. The binomial distribution is analytically used to predict the probability of average gas consumption exceeding 320 and 2000 m³ in two example cases of 500 and 1000 buildings.