Design and Operating Factors That Affect Waterflood Performance in Michigan

Abstract

Tinker, Gordon E.; SPE; Shell OilCo. Summary Water flooding started in the carbonate oil reservoirs of the northern Michigan Salina Niagaran reef trend in 1978 with Shell's Chester 18waterflood. Ten projects had been installed by the end of 1982, so operational results are available to expand and to reinforce the reservoir-simulation-based design and operating program. The small areal size of these pinnacle reef fields. variations in rock quality, and uncertain reservoir continuity have made successful water flood design difficult to achieve. Some existing wells are being, redrilled at the stall of a project and others may have to be redrilled later in the life of a water flood to allow the various porosity zones to be fully exploited within the flood pattern and to maintain adequate well spacing for oil bank formation. The operating strategies for these projects are based on a reservoir simulation study that stressed increased oil recovery and project economics. The operating procedures planned for these projects have been proved successful as evaluated by 4.5 year, of actual water flood performance in the Chester 18 water flood and from preliminary results from other projects. One of the most successful of these operational techniques has been the use of high-volume submersible pumps to maintain oil production response with increasing volumes of water, to give flexibility to injection pattern design, and to increase ultimate recovery. The mobility ratio in these reservoirs is very favorable, so water production results from channeling and, in some cases, bottom water coning rather than fingering. Well-completion strategy, a part of fluid-production and water-injection control, depends on knowing the location of secondary gas caps and the water zone. Water injection into even thin gas caps should be avoided to prevent gas trapping at the top of the reservoir and excessive early water production. Coning of gas and water, potential problems in many reservoirs, can result in reduced oil-recovery. Project monitoring procedures were carefully planned to facilitate project evaluation and changes in operating policy. It is believed that the design and operating policies as developed in this study have continued application for the many water floods planned in northern Michigan and to some extent for projects in other areas, but they may not be optimal for other types of reservoirs. Introduction Waterflooding for Shell started in Michigan during 1978 with the Chester 18project. Ten projects had been installed by the end of 1982, so operational results are available to expand and reinforce the early design and operating program based on reservoir simulation. The various projects located in northern Michigan (Fig. 1) range in size from two wells on 50acres to 17 wells covering 560 acres. This wide size range has led to some variation in results and problems, but it has been possible to develop a system of design and operating factors that pertain to the program as a whole. Shell's initial success in water flooding is expected to develop into a program that could eventually include as many as 50 water flood projects in northern Michigan. JPT P. 1884^