Why Cement Fluid Loss Additives Are Necessary

Abstract

SPE Member Abstract With respect to the fluid loss behaviour of a cement slurry basically two stages need be considered:a dynamic one corresponding to the placement and thena static one, the waiting on cement. During the first period the slurry flow is eroding the filter cake as it is growing thus rapidly a steady state is reached where the filtration occurs through a cake of constant thickness; at the same time, since the slurry is losing water but no solid particle, its density is increasing in line with the fluid loss rate. During the second period, referred to as "static", the cake grows due to the absence of flow. It may possibly grow up to a point where it locally but completely fills the annulus: bridging takes place and the hydrostatic pressure is no longer transmitted to the deeper zones. Dynamic: A maximum acceptable value of the total volume of fluid lost during the cement placement can be easily calculated from an upper density limit. Since basic slurry properties, like thickening time and rheology are greatly dependent on density (at least in the low water content domain), they may be used to define this limit. To convert the total amount of fluid loss into an API fluid loss value, three parameters are needed: the permeable formation area, the mudcake thickness and its permeability. Static: A maximum acceptable value of the thickness of the cake built-up during the waiting on cement period is deduced from the annular gap. The time taken by the filter cake to completely fill up the annulus, i.e., the bridging time, depends on the cement and mud cake properties (mudcake thickness and permeability and cement cake permeability). Therefore, by comparing thickening time and bridging time, a maximum value for the cement cake permeability can be deduced which is then expressed as an API fluid loss value. From typical mudcake resistances it can be estimated that, both in dynamic and in static conditions, the fluid loss could, in some conditions, have to be reduced to an API value one order of magnitude lower than what is generally considered as a fair control of fluid loss. Some examples are given. However, still very little is known about the effect of spacers, washes, mechanical aids and cement itself on the mudcake state and more data have to be gathered on mudcake thickness and permeability under various conditions, before definite fluid loss limits can be asserted. Introduction For more than 20 years, fluid loss control agents have been added to oil-well cement slurries and it is now recognized in the industry that the quality of cementing jobs has significantly improved. Indeed, it is generally clearly acknowledged that a lack of fluid loss control may be responsible for primary cementing failures, due to excessive density increase or annulus bridging and that formation invasion by cement filtrate may be deleterious to the production. With respect to squeeze-cementing, the problem is to adjust the level of fluid-loss to perforation size and formation nature. However, both for primary and remedial cementing very little has been written to justify the level of fluid loss control really required to achieve a good cement job. To address properly the quantitative evaluation of fluid loss limits compatible with successful cementing operations, two different stages have to be considered, first, placement or dynamic stage and then, waiting-on-cement or static stage. During the first stage, the slurry is flowing and eroding the cement cake which, after a short transient period, stops growing. In contrast, when the pumping is stopped the cake can grow freely. P. 853^