Some Technical and Economic Aspects Of Underground Gas Storage

Abstract This article deals with comparative technical and economic aspects of conventional and some nonconventional methods of storing gas. Conventional gas storage was first begun by injection and subsequent production of gas in a depleted gas field in Ontario, Canada in 1915. Conventional methods also include storage in depleted in oil fields and aquifers. Aquifer storage was first introduced into the United States with the injection of gas into the Galesville aquifer at Herscher, Ill. in 1953. Nonconventional methods include storage of gas in coal mines, mined salt caverns steel pipe and earth strata with artificial caprock and lateral confinement created by impermeable chemical grouts. Another method is storage of liquified gas in frozen earth or mined caverns. The growth and status of gas storage in the U.S. and Western Europe is summarized and technical and economic factors are related to the probable future direction and growth of storage in these areas. Introduction Major markets for natural gas in the U. S. and Western Europe often consume more gas during the four coldest winter months than during the remainder of the year. Peak winter demand usually exceeds three times the average summer consumption rate. Unless some form of near-market gas storage is used, large enough pipelines must be installed from producing fields to handle this peak winter demand. The resulting pipeline load factor, defined as average yearly flow rate divided by maximum or design rate, is then low and gas transmission costs are high. Near-market storage of gas serves as a buffer to allow a high pipeline load factor. Experience shows that the savings in transmissions costs are generally two to three times the cost of storage. Technical Aspects of Underground Gas Storage In addition to the basic requirements of size and proximity to market, a gas storage reservoir must possess an impervious roof and lateral confinement. Depleted reservoirs offer a caprock of guarantied integrity and sufficient structural closure or other lateral confinement to contain the gas. Partly for these reasons, we prefer to store gas whenever possible in depleted fields rather than in aquifers. Abandoned or poorly cemented wells are sources of gas leakage in depleted fields. In many cases, considerable time and expense are necessary to locate and recondition or plug such wells. In general, however, this is cheaper than the initial drilling and completion of wells in developing aquifer storage. In developing aquifer storage, extensive geological and hydrological work is performed to investigate the adequacy of caprock integrity and structural or lateral confinement. In spite of this effort, many of the aquifer storage reservoirs in the U. S. leak gas to shallower formations. Extensive efforts failed to locate a source of the leak at the Galesville aquifer project in Herscher, Ill., and in 1960 over 13 MMcf/D were circulated from shallower formations back into the Galesville aquifer.' This amounted to 4.6 Bcf/year,* a significant fraction of the 34.2 Bcf stored at the end of that year. Delivery capacity is one of the most important considerations in designing a storage reservoir. For a given number of wells, the delivery rate is proportional to reservoir pressure which, in turn, is proportional to gas in place. This presents a problem since the largest required delivery rates often occur in the latter part of the winter when gas reserves are lowest. In the case of a dry gas reservoir this problem can be solved rather simply since the known, constant reservoir pore volume allows easy prediction of pressure from a given gas withdrawal schedule. From the predicted pressure behavior during the season, delivery capacity can be calculated for a given number of wells or the number of wells necessary to ensure a given delivery capacity. Water movement in aquifer and water drive fields considerably complicates the calculation of pressure as a function of gas withdrawn over the winter season. In this case, reservoir pore volume can vary considerably, growing with spring and summer injections and shrinking with winter withdrawals. Methods of calculating this water movement and relating it to reservoir pressure and withdrawals have been extensively studied and are described in the literature.