Predicting Mud Toxicity
- Roger Bleier (M-I Drilling Fluids Co.)
- Document ID
- Society of Petroleum Engineers
- Journal of Petroleum Technology
- Publication Date
- October 1991
- Document Type
- Journal Paper
- 1,192 - 1,193
- 1991. Society of Petroleum Engineers
- 1.11 Drilling Fluids and Materials, 1.11.2 Drilling Fluid Selection and Formulation (Chemistry, Properties), 4.2.3 Materials and Corrosion
- 5 in the last 30 days
- 234 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 1.00|
|SPE Non-Member Price:||USD 2.00|
Acute toxicity of drilling muds is measured in the U.S. by the mysid shrimptest. Drilling muds that fail the test cannot be discharged into the Gulf ofMexico, and such muds and their cuttings must he brought onshore for disposal.Discharge of water-based muds that pass the test is permitted in mostinstances. Because of the economic implications associated with haulingcuttings and fluids, a model that predicts test results on the basis of mudcomposition is clearly desirable. This paper focuses on the modeling of mysidshrimp test data. European laboratories use different test species andprocedures. It seems plausible to expect, procedures. It seems plausible toexpect, however, that the line of reasoning used here could apply to themodeling of aquatic data on other test species once a sufficient quantity ofsuch data becomes available. Fig. 1 illustrates the mysid shrimp testprocedure.1 The mud sample to be tested procedure.1 The mud sample to be testedis mixed with seawater at a ratio of one part mud to nine parts seawater. Themixture is stirred for 5 minutes and then allowed to settle for 1 hour. At thatpoint, the liquid (and whatever particles are suspended in it) is poured off,and this "suspended-particulate poured off, and this"suspended-particulate phase" is the toxicant tested. The toxicantphase" is the toxicant tested. The toxicant is added at variousconcentrations to test dishes containing mysid shrimp in seawater to determinethe LC50, the concentration of toxicant lethal in 96 hours to 50% of the mysidpopulation. In this paper, we follow the common practice of calling theconcentration the LC50 of the mud, even though it is the LC50 of the fluidobtained from the 1:9 mixture of mud and seawater. A high LC50 means that themud is not very toxic to mysid shrimp; conversely, the lower the LC50, the moretoxic the mud. The U.S. Environmental Protection Agency (EPA) prohibits thedischarge of mud and cuttings in the Gulf of Mexico if the mud's LC50 is<30,000 ppm. Because the test period is 96 hours, 1 week is a goodturnaround time for a test laboratory; however, 3 weeks is not uncommon. As apractical matter, an operator discharging cuttings into the Gulf of Mexico doesso on the basis of estimates that the mud will almost surely pass when tested.Mistakes can lead to stiff fines. On the other hand, overcautiousness leads tounnecessary hauling of mud and cuttings to shore or, worse, to underdesign ofdrilling-fluid performance characteristics. performance characteristics.Sujective Predictions
Early attempts to predict mud toxicity were necessarily qualitative andsubjective. Operators restricted themselves to mud formulations that hadpreviously tested satisfactorily and avoided specialty products wheneverpossible. (Note that in this paper "mud toxicity" means the toxicitymeasured by the mysid test.) As more data became available, better distinctionscould be made. In 1986, Jones et aL 2 listed products by generic name, tradenames, and toxicity ranges. Leuterman et aL 3 updated the list in 1989.Conventional weighting agents, clays, and long-chain polymers were determinedto be nontoxic polymers were determined to be nontoxic in the mysid test. Othermaterials, such as chrome lignosulfonates, were slightly toxic to the shrimpbut posed no threat at normal concentration ranges. Some small organicmolecules were identified as very toxic to the mysid shrimp. Many lubricants,spotting fluids, corrosion inhibitors, and defoamers fell into that category.(More satisfactory alternatives have since been developed.) Surprisingly, somebactericides were not very toxic to the shrimp at normal concentrations. On theother hand, potassium chloride was a problem at concentrations potassiumchloride was a problem at concentrations preferred for inhibiting the swellingof some preferred for inhibiting the swelling of some shale formations.
Large quantities of data eventually were accumulated, and it became possibleto attempt analysis with the goal of a quantitative model. Such a model wouldpredict, to a reasonably close approximation, a mud's LC50 from itscomposition. There are two reasons to approach the use of any quantitativemodel with caution. First, the mud may pick up unforeseen contaminants, such ascrude oil, that can drastically change its toxicity. This is not really a faultof the model so much as it is our inability to foresee the contaminant as partof the composition. (Crude or diesel oil present at more than a trace can causeany mud, no matter how carefully designed, to fail the mysid test.) Second,test results can very among laboratories. A good bioassay laboraratory shoulddeliver results that vary by no more than 20% on splits of the same sampleVariability between two different labortories may be higher, however,apparently because of subtle differences in technique. The prudent course is tobase any model on data from a single laboratory, preferably one that tends topredict LC50 numbers in the lower range among laboratories. By 1988, we hadacquired sufficient data to attempt analysis. A significant body of data hadbeen accumulated on the effects of single additives on one particular referencemud, a seawater lignosulfonate laboratory mud. This body of data allowed theisolation of variables (mud additives). The remaining data, on more complicatedmuds, provided a screening test for the provided a screening test for theeffectiveness of any proposed model that Would be used for recombining thevariables.
Development of One Quantitative Model
Preliminary analysis of the reference-mud Preliminary analysis of thereference-mud data suggested that water-based mud components fall into threecategories. Type 1 additives do not affect LC50 at normal concentrations inmuds. Type 2 additives reduce LC50, but the effect of increasing concentrationsis sight. Type 3 additives reduce LC50 in an inverse linear concentration(e.g., doubling concentration cuts the LC50 in half) Type 1 and 3 behaviors areeasy to explain. Type 1 additives are simply too insert or large in molecularsize to affect the mysids. Type 3 behavior is expected of additives with atoxic component that, at concentrations of normal use, is totally soluble inseawater. Type 2 additives are mildly toxic additives for which the rate ofreduction in LC50 decreases with increasing additive concentration. Carefulinspection suggests that the reduction can be adequately described by asquare-foot function.
|File Size||342 KB||Number of Pages||2|