Analysis of Fluid-Loss Data

Abstract

Summary Over the last few years, fluid-loss testing has evolved from static to dynamic testing. Data analysis, however, has not progressed as rapidly as the experimental techniques used to gather the data. In this paper, various techniques of analyzing dynamic fluid-loss data are paper, various techniques of analyzing dynamic fluid-loss data are discussed and a simple, yet effective, method of describing dynamic fluid-loss curves is presented. The analysis of dynamic fluid-loss data is usually based on the traditional square-root-of-time treatment. This type of analysis works well for static fluid-loss test data because the dynamic processes are less complex than those in dynamic situations. During a dynamic fluid-loss test, a number of different events influence the fluid loss and the shape of the fluid-loss curve. Not only filter-cake compression, but also dynamic cake deposition and cake erosion must be taken into account. In most cases, Darcy flow does not occur over the whole time span of the test. Several complex methods of treating the data from dynamic fluid-loss tests have been suggested, but most rely on modifications of the static fluid-loss-data analysis technique. These techniques are based on the data as a function of the square root of time. With dynamic fluid-loss test data, the square-root-of-time relationship does not adequately describe the fluid-loss data. A straightforward technique to analyze dynamic fluid-loss data is discussed in this paper. Introduction Estimating loss of fluids to the formation is important in drilling, cementing, fracturing, and almost every other type of downhole treatment design. Although the discussion presented here is confined to fluid loss during fracturing, many concepts are applicable to other downhole activities. Howard and Fast, who discussed fluid loss during hydraulic fracturing, separated fluid-loss control into three different mechanisms. Two mechanisms deal with viscosity and with relative permeability and compressibility effects. These mechanisms are not discussed here. The third involves the deposition of particles from the fluid onto the formation face (often called wall building or cake deposition). Extensive studies done on particle deposition during fluid loss for drilling fluids and cements particle deposition during fluid loss for drilling fluids and cements used static measuring systems where cake buildup can be described by equations presented by Collins. A result of this type of analysis is a square-root-of-time dependency of the volume of fluid lost to the matrix. This analysis also worked for noncrosslinked* fracturing fluids under static measuring conditions. Hall and Dollarhide were the first to study fluid-loss behavior under dynamic conditions using a hollow-core device. Sinha used a similar device to study both static and dynamic fluid loss. Gulbis also used a hollow-core device to study dynamic fluid loss in both crosslinked and noncrosslinked fracturing fluids. Penny et al. published a comprehensive study of several different techniques to study dynamic fluid loss. They suggested analyzing the data as a function of the nth power of time rather than the square root of time. The data are plotted on log-log paper to yield a straight line. Ford and Pennys presented a technique to convert the parameters derived from the log-log analysis to the more familiar parameters derived from the log-log analysis to the more familiar parameters derived from a square-root-of-time analysis. Roodhart parameters derived from a square-root-of-time analysis. Roodhart presented another method of analyzing the data from dynamic presented another method of analyzing the data from dynamic fluid-loss studies. He added a term to the standard square-root-of-time analysis to correct for the dynamic-leakoff component. Harris described a technique using impinging flow on a core face to study the fluid loss by foamed fluids. Several investigators used a slot-flow device to study a number of different fracturing fluids. Settari analyzed McDaniel et al., data with a generalized computer model based on fluid flow through the filter cake and in the fracture. This model is complex and is not used in this study. Most, if not all, of these contributions indicate that the square-root-of-time analysis is inadequate to describe dynamic fluid-loss data. Ford and Pennys showed that large errors can result from the traditional analysis of fluid-loss data. Dynamic flow testing appears to complicate the process by providing a mechanism for removing part of the filter cake that the fluid flows past while at the same time redepositing the cake. This process continues until an equilibrium cake thickness is established. process continues until an equilibrium cake thickness is established. During this time, the cake is also compacting, and Darcy flow conditions do not exist in the filter cake. Considering these factors, it is little wonder that the square-root-of-time relationship is not adequate.