Comparison of Field and Laboratory-Simulated Drill-Off Tests

Abstract

Summary. Field drill-off test results are compared with data from laboratory simulations. A simple theory for analyzing drill-off tests is developed. The weight-on-bit (WOB) decay with time is close to exponential, but large threshold WOB'S, resulting from poor weight transmission downhole, are sometimes observed in field tests. Introduction The drill-off test was devised nearly 30 years ago as a practical means of estimating and optimizing drilling performance, but its use in the field has been limited. Two factors seem to limit its value: the quality of data recorded with conventional rig equipment and uncertainty in data interpretation. A joint program was undertaken by Sedco-Forex and Schlumberger Cambridge Research to address these limitations and hence to increase the information actually obtained from a drill-off test. The work involved two aspects: the development and implementation of a rig-site drill-off test data acquisition system and the laboratory simulation of drill-off tests. This paper describes the progress to date, concentrating on the comparison of field and laboratory drill-off test data and on the resultant improvements in our interpretation of drill-off tests. Procedure The drill-off test is a simple, practical procedure, proposed by Lubinski, for determining the relationship between rate of penetration (ROP) and surface WOB. The driller builds WOB to a predetermined maximum and then sets the brake. As the bit drills, the WOB decays at a rate determined by the ROP and the drillstring compliance. The surface WOB is recorded as a function of time. The drillstring compliance is determined from a knowledge of the drillstring composition and is used with the recorded WOB to compute the distance drilled as a function of time. Differentiation yields the ROP as a function of time, and crossplotting gives the dependence of ROP on WOB. Ref. 2 describes an alternative procedure for the analysis of drill-off test data based on Vidder's procedure. The times taken to drill off successive increments of 4,000 lbf [ 18 kN] WOB are recorded, and the mean ROP over each increment is computed from the drillpipe compliance. (Note that length changes in tool joints and drill collars can be neglected.) A log-log plot of ROP against WOB reveals the exponent relating ROP to WOB. If there is a significant threshold WOB-i.e., a WOB below which the bit will not drill-then the exponent can be determined by subtracting different threshold WOB's from the recorded values until a straight-line fit is obtained. Regardless of which analysis is adopted, variations in rock drillability and oscillations in the hook load can combine to restrict the meaning of the derived relationship between ROP and WOB unless some method of smoothing the raw data is used. The next section presents one way of overcoming the problem. When a drill-off test is designed, both the distance drilled and the rate of change of hook load must be considered if the test is to generate results that are representative of the formation being drilled. Assuming that no change occurs in hydraulics during the drill-off, the change in length of the drillstring is related to the change in WOB by ..........................................(1) With an 8 1/2 -in. [21.6-cm] roller-cone bit, the maximum change in WOB is restricted to at)out 45,000 lbf [200 kN], and the bit would normally be run with 5-in. [ 12.7-cm] drillpipe. The compliance of the bottomhole assembly (BHA) is negligible compared with that of the drillpipe. Therefore, the following equation gives the length of formation drilled: ..........................................(2) The rate of change of WOB with depth is ..........................................(3) Initially regarding 5,700 lbf/in. [1 kn/mml as a reasonable upper limit for the rate of change of WOB and assuming 600 ft [200 m] of drill collars, we calculated a minimum drillstring length of 3,000 ft [ 1,000 m]. Thus, the maximum length of formation drilled during the drill-off will be 8 in. [20 cm]. As a result of these calculations, we decided that it was unlikely that valuable information would be derived from drill-off tests performed at depths of less than 3,000 ft [1000 m], and we restricted our investigations accordingly. Interpretation Theory Laboratory and field studies have shown that it is reasonable, in many circumstances, to consider the ROP, R, proportional to the WOB, W. Thus, ..........................................(4) During a drill-off test the WOB is linked to the bit ROP by the elasticity of the drillstring. ..........................................(5) ..........................................(6) Integration gives the following expression for WOB as a function of time during a drill-off test. ..........................................(7) We found in some cases that the WOB decays not to zero, but to a finite offset or threshold WOB, W,. If this threshold WOB is introduced into the expression relating ROP to WOB, the dependence of WOB on time becomes ..........................................(8) Three valuable pieces of information can be obtained by fitting this expression to WOB data recorded during a drill-off test: the quality of fit, which indicates whether the assumption of linearity between ROP and WOB is valid the value of W, whose significance is discussed below; and the value of K. which characterizes the ROP. Note that the curve-fitting scheme adopted to determine W, is equivalent to the trial-and-error procedure previously mentioned because both rely on finding the value of W, that gives the best fit between the data and the assumed relationship between ROP and WOB. Here, however, we work with data recorded at a much higher frequency and consequently gain enhanced resolution. Experimental Details More than 50 drill-off tests were performed in vertical wells and in wells deviated up to 40deg.