Summary. Field experience suggests and full-scale laboratory test results confirm that mud displacement in high-angle wellbores can be complicated by a channel of hard-to-displace mud forming on the low side of the wellbore. This channel is caused by solids settling while the drilling fluid is circulating. Experimental test results obtained with till-scale, permeable and impermeable deviated wellbores indicate that this channel of solids can be prevented with proper rheological control of the drilling fluid. Results reported also demonstrate the effect of casing centralizers, pipe movement, and preflushes on the removal of this low-side channel of solids. Based on the laboratory results, displacement guidelines to improve deviated-wellbore cementing by eliminating the low-side solids-settling channel are presented. Introduction As part of a continuing investigation of factors affecting drilling-fluid displacement during primary cementing, a study has been carried out to identify factors to improve deviated-wellbore cementing. A previous study identified several potential problems associated with cementing high-angle wells, including the occurrence of mud channels on the low side of the annulus and water channels on the high side of the annulus. The most serious potential problem affecting deviated-well cementing appeared to be the deposition of solids caused by settling of weighting agents or drilled cuttings from the drilling mud. The test data from the previous study suggested that the rheology of the mud, specifically the yield point, determined whether solids would settle. Settling from the drilling mud creates a continuous uncemented channel along the low side of the wellbore. Occurrences of continuous mud channels in the annulus can prevent mud displacement and defeat the purpose of cementing-i.e., to surround the casing with a complete sheath of cement that prevents fluid flow in the annulus. Failure to surround and to protect the casing through incomplete mud displacement from the downhole environment can lead to such problems as annular migration of well fluids, casing corrosion or collapse, loss of well control, and high remedial cementing costs. Previous studies in vertical wellbores have found that the highest mud-displacement efficiency could be attained by lowering the yield point of the mud and maximizing pumping rates. The initial study of deviated-well cementing I suggested, however, that high-yield-point muds were required to prevent solids settling and that lowyield-point muds deposited solids to such an extent that complete mud displacement could not be achieved. The purposes of this study were to investigate further (1) the relationship of drilling-mud yield point and the deposition of drilling mud solids and (2) methods known to improve drilling-mud-displacement efficiency in vertical wellbores and to examine their effectiveness for deviated-well conditions. Factors known to influence displacement in vertical wellbores have previously been identified and studied in a simulated vertical wellbore on a large-scale testing apparatus. These factors were examined on a large-scale deviated-wellbore model at various deviation angles. Factors studied included use of preflushes, pipe centralizers, and pipe movement. Findings reported here suggest methods of preventing the formation of mud channels by solids settling from drilling mud, and if mud channeling does occur, methods of displacing the channel of settled mud solids from the low side of the wellbore. Experimental Apparatus and Procedure The apparatus used in conducting this research was designed and operated to simulate as closely as possible the actual conditions experienced during the cementing of an oil well. Fig. 1 shows the wellhead assembly used to circulate the various fluids under deviated conditions. A schematic of the simulated well is shown in Fig. 2. Each test section consists of manmade, permeable, consolidated sands enclosed in a perforated pipe 15 ft [4.6 m] long with a 6 %-in. [ 16.5-cm] ID. During this investigation, two downhole conditions were simulated:permeable formation, thus allowing for filter-cake buildup, andcompletely impermeable formation. For the impermeable tests, a 6 1/2-in. [16.5-cm] -ID steel casing was used as the test section. 5-in. [ 12.7-cm] -OD casing was installed inside the test section. The casing was centralized at the top and bottom of the test section. The test section was lowered into the filtrate jacket and allowed to become saturated with water. Fluid loss to the permeable formation was monitored through an outlet fixed on the filtrate jacket. The filtrate jacket was enclosed in a hot-water heating jacket. JPT P. 961^