Effect of Drilling Fluid Properties on Rate of Penetration

A. M. Paiaman, M. K. Ghassem Al-Askari, B. Salmani, B. D. Al-Anazi and M. Masihi

ORIGINAL SCIENTIFIC PAPER

The scope of this work is to determine the effect of drilling fluid properties on penetration rate in a field using daily drilling reports. Based on our field studies the following drilling fluid properties affect penetration rate to varying degrees: mud weight, plastic viscosity and solid content. The result of this study shows that the net effect of drilling fluid properties on penetration rate is less than what it thought to be. Penetration rate is decreased by increasing plastic viscosity, solid content and mud weight. Decreasing the penetration rate is more attributed to increase of depth, because by increasing the depth, rock strength increases and porosity decreases. Key words: Mud weight, plastic viscosity and solid content, penetration rate, regression

1. INTRODUCTION
The factors which affect rate of penetration are exceedingly numerous and perhaps important variables exist which are unrecognized up to this time. A rigorous analysis of drilling rate is complicated by difficulty of completely isolating the variable under study. For example, interpretation of field data may involve uncertainties due to the possibility of undetected changes in rock properties. Studies of drilling fluid effects are always plagued by difficulty of preparing two muds having all properties identical except one which is under observation. While it is generally desirable to increase penetration rate, such gains must not be made at the expense of overcompensating, detrimental effects. The fastest on-bottom drilling rate does not necessarily result in the lowest cost per foot of drilled hole. Other factors such as accelerated bit wear, equipment failure, etc., may raise cost. These restrictions should be kept in mind during the following discussion. In any engineering study of rotary drilling it is convenient to divide the factors which affect the rate of penetration into the following list: 1. Personal efficiency 2. Rig efficiency 3. Formation characteristics (e.g. strength, hardness and / or abrasiveness, state of underground formations stress, elasticity, stickiness or balling tendency, fluid content and interstitial pressure, porosity and permeability) 4. Mechanical factors (e.g. weight on bit, bit type and rotary speed) 5. Hydraulic factor (e.g. jet velocity, bottom- hole cleaning ) 6. Drilling fluid properties (e.g. mud weight, viscosity, filtrate loss and solid content) Formation as nearly an independent or uncontrollable variable is influenced to a certain extent by hydrostatic pressure. Laboratory experiments indicate that in some
NAFTA 60 (3) 129-134 (2009)

formations increased hydrostatic pressure increases the formation hardness or reduces its drill-ability.7 The bit type selected, i.e., whether a drag bit, diamond bit, or roller cutter bit is used and the various tooth structures affect to some extent the drilling rate obtainable in a given formation. The mechanical factors of weight on the bit and rotary speed are then linearly related to the drilling rate, provided the hydraulic factors are in proper balance to insure proper cleaning of the hole. The hydraulic factors affect drilling rate only when they influence the rate of penetration or the efficiency of the drilling. It is known that drilling fluid properties can affect drilling rate. This fact was established early in the drilling literature, and confirmed by numerous laboratory studies.4 Several early studies were focused directly on mud properties, clearly demonstrating the effect of mud properties on drilling rate.4 Inasmuch as it is impossible to change one property of drilling fluid without affecting the others, it is difficult to evaluate the true effect of an individual parameter on penetration rate.6 Therefore in this article it is tried to eliminate effects of other parameters (such as weight on bit-WOB, rotary speed-N, bit hydraulic, and depth) that are affecting the rate of penetration-ROP , for identifying effects of drilling fluid properties. For this purpose field data were collected and by using multiple regression1,5,2, Bourgoyne-Yong equations constant were identified, then by using this equation penetration rate was normalized to identify the effect of drilling fluid properties on penetration rate.5

2. PROCEDURE
To accomplish the objectives of this project the data of an oilfield were used.1 At first stage, data of wells which had been drilled from date of 2002 through 2006 were gathered.1 At the second stage required information were selected from these data, which data selection was limited

129

A. M. PAIAMAN, M. K. GHASSEM AL-ASKARI, B. SALMANI ...

EFFECT OF DRILLING FLUID ...

Table 1. Data for Multiple Regression Analysis1

Data Entry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Depth (ft) 3 231 3 592 3 608 4 021 4 156 4 156 4 251 4 267 4 500 4 749 4 789 4 808 4 858 4 979 4 982 4 999 5 005 5 196 5 225 5 241 5 261 5 287 5 405 5 451 5 674 5 911 6 180 6 754 6 760 7 032 ROP (ft/hr) 31.3 25.3 45.0 16.5 32.2 35.6 15.9 29.5 22.0 17.6 26.2 13.9 17.1 10.6 18.2 21.4 18.3 7.3 21.7 16.1 7.1 13.6 20.8 13.4 14.8 9.8 6.9 8.7 3.3 8.0 h 0.25 0.25 0.50 0.50 0.44 0.38 0.50 0.38 0.38 0.63 0.00 0.50 0.38 0.50 0.38 0.63 0.31 0.25 0.63 0.50 0.38 0.25 0.38 0.50 0.50 0.63 0.50 0.75 0.63 0.31 Jet Impact Force (1 000 lb) 1.86 1.82 2.35 1.77 1.85 2.16 2.00 1.96 2.22 2.24 1.88 2.18 1.69 2.15 2.37 2.40 2.16 1.91 2.46 2.36 1.68 1.64 2.24 2.11 2.16 1.71 1.58 1.73 1.65 1.91 WOB (klb/in) 1.57 2.86 3.29 1.86 4.00 3.14 2.57 3.00 2.86 2.57 2.43 2.57 2.33 2.14 2.86 3.14 2.14 1.07 2.29 3.00 0.97 2.14 2.57 2.71 2.86 2.53 1.86 2.57 1.00 2.29 RPM 155 160 165 165 190 155 160 180 165 155 145 163 193 163 160 148 160 155 180 150 195 165 155 167 180 160 160 155 145 180 ECD (lb/gal) 9.29 9.02 8.78 9.25 8.57 9.11 8.89 8.73 9.53 8.84 9.35 9.72 9.02 9.60 9.42 9.64 9.00 9.66 9.02 9.93 9.09 10.10 9.05 9.75 9.12 9.76 10.70 10.58 11.15 10.25 Pore gradient (lb/gal) 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0

to formation in 17.5 inch borehole and rolling cutter bit. Table 1 is a summary of data which consists of bit data, drilling parameter, hydraulic, dull bit grading, and depth out of bit. If rate of penetration is normalized to weight on bit, rotary speed, bit hydraulic, bit tooth wear, and depth, the result is penetration rate that is not affected by the variables that listed above, except by drilling fluid properties. For this purpose a drilling model equation should be selected for predicting the effect of drilling parameters on penetration rate. In 1974, Bourgoyne and Young proposed a relation using a complex drilling model to compensate mathematically for changes in the various drilling parameters.4 They proposed it using eight functions to model the effect of most of drilling variables discussed above. The equation form is:

8 dD = exp a1 + aj x j dt j= 2

(1)

Where exp (x) is used to indicate the exponential function e x . The terms a2 x 2 and a3 x 3 model the effect of compaction on penetration rate and x2 and x3 are respectively:
x2= 10 000 - depth x3= D
0.69 ( g
p

(2) (3)

- rn )

Thus an exponential increase in penetration rate with pore pressure gradient can be assumed. The exponential nature of the effect of compaction on penetration rate was suggested by compaction theory but has not yet been verified experimentally. Therefore an exponential decrease in penetration rate with depth in normally comNAFTA 60 (3) 129-134 (2009)

130

EFFECT OF DRILLING FLUID ...

A. M. PAIAMAN, M. K. GHASSEM AL-ASKARI, B. SALMANI ...

pacted formations may be assumed. The effect of compaction on penetration rate has been normalized to equal 1.0 for normally compacted formation at 10 000 ft and pressure gradient equal nine (gp =9.0 lb/gal). The term a4 x 4 models the effect of differential pressure across the bottom-hole on penetration rate. Moreover, x4 is defined by:
x 4 = D (gp - rc ) (4)

Therefore an exponential decrease in penetration rate with excess bottom-hole pressure may be assumed. The term a5 x 5 models the effect of bit weight and bit diameter on penetration rate. Also x5 is defined by:
w / d -( w / d ) t x 5 = ln 4.0 -(w / d ) t (5) Fig. 1. Observed penetration rate vs. plastic viscosity Sl. 1. Uoena brzina buenja u ovisnosti o plastinom viskozitetu

Thus it is assumed that penetration rate is directly proportional to ( w / d)a5 . The term e a5 x 5 is normalized to equal 1.0 for 4 000 lb per inch of bit diameter. The threshold bit weight, ( w / d)t must be estimated with drill-off tests. The term a6 x 6 represents the effect of rotary speed on penetration rate. Also x6 is defined by:
N x 6 = ln 100 (6)

In above relation, it is assumed that penetration rate is directly proportional to N a6 . The term a7 x 7 models the effect of tooth wear on penetration rate. Also x7 is defined by:
x 7 = -h (7) Fig. 2. NROP vs. plastic viscosity Sl. 2. NROP u ovisnosti o plastinom viskozitetu

Where h is the fractional tooth height that has been worn away. The value of a 7 depends primarily on bit type and on formation type. The term e a7 x 7 is equal 1 when either h or a7 is zero. The term a8x 8 models the effect of bit hydraulics on penetration rate. x8 is defined by:
Fj 8 x 8 = ln 1 000
a

(8)

Where Fj is the hydraulic jet impact force beneath the bit. Hence, it is suggested that penetration rate is directly proportional to (Fj )a8 based on experimentally microbits studies performed by Eckel.4 To calculate the best values of the regression constant a1 through a8 using data shown in Table 1, the parameters x2 through x8 must be calculated using Equations 2 through 8 for each data entry. Eight
NAFTA 60 (3) 129-134 (2009)

Fig. 3. ROP vs. solid content Sl. 3. ROP u ovisnosti od koncentracije krutih estica

131

A. M. PAIAMAN, M. K. GHASSEM AL-ASKARI, B. SALMANI ...

EFFECT OF DRILLING FLUID ...

equations with eight unknowns a1 through a8 can then be obtained from x2 through x8 using procedure described in Appendix. When the resulting system of eight equations is solved for the eight unknowns, the constants a1 through a8 for data shown in Table 1 are obtained. The constants a1 through a8 are obtained as:
a1= 1.522 9, a2=0.000 34, a3=0.000 7, a4=0.000 02, a5=0.81, a6=0.64, a7=0.561, a8=0.49

3. REGRESSION ANALYSIS AND RESULTS

By normalizing rate of penetration into the weight on bit, rotary speed, bit hydraulic, and depth, result is penetration rate that is not affected by bit weight, rotary speed or hydraulics and depth, which are significant variables in controlling rate of penetration.
dD = exp (a1 + a2 x 2 + a3 x 3 +......+a8 x 8 ) dt (dD / dt ) n = (dD / dt ) * exp[ a2 (x 2 n - x 2 ) +a3 (x 3 n - x 3 ) +a4 (x 4 - x 4 ) + a5 (x 5 n - x 5 ) +a6 (x 6 n - x 6 ) + a7 (x 7 n - x 7 ) +a8 (x 8 n - x 8 )] Fig. 4. NROP vs. solid content Sl. 4. NROP u ovisnosti od koncentracije krutih estica

Effect of plastic viscosity on drilling rate

For investigating effect of plastic viscosity on rate of penetration (ROP) two cases would be considered. In case one ROP is plotted with respect to plastic viscosity. As it can be seen in Fig.1, penetration rate is decreased by normal slope with increasing plastic viscosity. In the second case, normalized rate of penetration (NROP) plotted against plastic viscosity. As it shown in Fig. 2, by increasing plastic viscosity, NROP is decreased. But this reduction is less than the ROP reduction.

Fig. 5. Effect of solid content on penetration rate at constant plastic viscosity Sl. 5. Djelovanje koncentracije krutih estica na brzinu buenja kod konstantne plastine viskoznosti

Effect of solid contents on penetration rate

For investigating effect of solid contents on penetration rate two cases would be considered. Case one is when plastic viscosity is not constant. In other word, all solid contents are variable. As it shown in Figs. 3 and 4, ROP and NROP are decreased by increasing solid content. Case two is when plastic viscosity is constant. As it is shown in Figs. 5 and 6, when plastic viscosity is constant, by increasing solid content, ROP and NROP are decreased.
Fig. 6. Effect of solid content on NROP at constant plastic viscosity Sl. 6. Djelovanje koncentracije krutih estica NROP kod konstantne plastine viskoznosti
NAFTA 60 (3) 129-134 (2009)

132

EFFECT OF DRILLING FLUID ...

A. M. PAIAMAN, M. K. GHASSEM AL-ASKARI, B. SALMANI ...

Fig. 7. Effect of mud weight on penetration rate Sl. 7. Djelovanje te ine isplake na brzinu buenja

D= Depth, ft gp= Formation pressure gradient, lb/gal rn= Normal fluid pressure gradient, psi/ft W= Weight on bit, klb/in d= Bit diameter, in N= Rotary speed, rpm h= Fractional tooth height Fj= Hydraulic jet impact force beneath the bit, lb NROP= Normalized rate of penetration, ft/hr rc = Equivalent circulating density gradient, psi/ft t= Threshold w= Bit weight, lb a= Coefficient r= Residual error G= Regression index of correlation ob= Constant c= Constant

Fig. 8. Effect of mud weight on NROP Sl. 8. Djelovanje te ine isplake na NROP

Effect of mud weight on penetration rate

Effect of mud weight on penetration rate normalized penetration rate is shown in Figs. 7 and 8.

4. CONCLUSIONS
1. Effect of drilling fluid on penetration rate in Aghajari formation is less than it that is thought should to be. 2. By increasing plastic viscosity, rate of penetration and normalized rate of penetration are decreased, which is confirmed by other authors3, but decrease in rate of penetration is more than normalized penetration rate. 3. Rate of penetration is decreased by increasing mud weight which is confirmed by other authors.7 4. At constant plastic viscosity, penetration rate is decreased by increasing solid content.

5. NOMENCLATURE
ROP= Rate of penetration, ft/hr
NAFTA 60 (3) 129-134 (2009)

133

A. M. PAIAMAN, M. K. GHASSEM AL-ASKARI, B. SALMANI ...

EFFECT OF DRILLING FLUID ...

ACKNOWLEDGMENTS
We would like to express our special thanks to Mr. Mohammadian (drilling department of NISOC) for his consulting. Thanks all friends which aided us in this work.

APPENDIX
Bourgoyne and Young model is:
8 dD a + a x = exp 1 j j dt j= 2

(A-1)

7. REFERENCES
1. Afsari M., Petroleum University of Technology, Investigation the Effect of Drilling Fluid on Rate of Penetration. A thesis submitted to the Graduate Office in partial fulfillment of the requirements for the B.Sc. Degree in Petroleum Engineering, June, 2003. 2. Albetari E. A., Moussa M. M., Alotaibi S., Multiple Regression Approach to Optimize Drilling Operation in the Persian Gulf Area, SPE 13494. 3. Beck F. E., Arco Alaska, Powell J. W., and Zamora M., The Effect of Rheology on Rate of Penetration, SPE/IADC 29368. 4. Bourgoyne A. T., Keith K. Millheim, Martin E. Chenevert, F. S. Yong, Applied Drilling Engineering, SPE Textbook series, 1984. 5. Bourgoyne A. T., Yong F. S., A Multiple Regression Approach to Optimal Drilling and Abnormal Pressure Detection, SPE 4238. 6. James L. Lummus J. L., Azar J. J., Drilling Fluid Optimization, A Practical Field Approach, Penn Well Publishing Company, Tulsa, Oklahoma. 7. Rupert J. P., Prado C. W., and Blatted S. R., The Effect of Weight Material Type and Mud Formulation on Penetration Rate Using Invert Oil System SPE 10102. 8. Stevev R. Blatted and J. Paul Rupert, The Effect of Weight Material Type on Rate of Penetration Using Dispersed and Non-dispersed Water-Base Muds, SPE 10961

Taking the logarithm of both side of equation (A-1) yields

8 dD = a1 + aj x j ln dt j= 2

(A-2)

Equation (A-2) can be checked for validity in given formation at each depth at which data have been collected. If the residual error of the i the data point, ri is defined by:
dD ri = a1 + aj x j - ln dt j= 2
8

(A-3)

Now for n data points, where n is any number greater than 8, the sum of the square of residuals, ri 2 is minimum when
i=1 n

its derivations is zero.

ri 2
i= 1 n

aj

= 2ri
i= 1

r i aj

= 2ri x j = 0
i= 1

(A-4)

For j =1, 2, 3 8. Thus the constant a1 through a8 can be obtained by simultaneously solving the system of equations obtained by expanding of ri xi
i=1 n

For j =1, 2, 3 8. Expanding of ri xi yields

i=1
n n n n dD na1 + a2 x 2 + a3 x 3 +...+a8 x 8 = ln dt i= 1 i= 1 i= 1 i= 1

(A-5)
n

a1 x 2 + a2 x 2 2 + a3 x 2 x 3 +...+a8 x 2 x 8 = x 2 ln
i= 1 i= 1 i= 1 i= 1 i= 1

dD dt

(A-6)

a1 x 3 + a2 x 3 x 2 + a3 x 3 2 +...+a8 x 3 x 8 = x 3 ln
i= 1 i= 1 i= 1 i= 1 i= 1

dD dt

(A-7)

n n n n n dD 2 a1 x 8 + a2 x 8 x 2 + a3 x 8 x 3 +...+a8 x 8 = x 8 ln dt i= 1 i= 1 i= 1 i= 1 i= 1

(A-8)

The final correlation is checked for accuracy using the regression index of correlation G, given by:
dD dD ln - ln dt ob dt c ln dt dD dD - ln dt ob
2 2

G = 10 . -

v
Authors:

(A-9)

Abouzar Mirzaei Paiaman, Sharif University of Technology Coressponding Author: Iran, Lorestan, Kuhdasht, Khaghani Ave., Postal Code: 68417-88368, Tell: +989361642229 M. K. Ghassem Al-Askari, Petroleum University of Technology Barat Salmani, Petroleum University of Technology Bandar Duraya Alanazi, King Abdulaziz City for Science & Technology Mohsen Masihi, Dr., Iran, Tehran, Azadi Ave., Sharif University of Technology, PO Box: 11365-9465 , Department of Chemical and Petroleum Engineering

When any of the regression constants are assumed. To be known, the corresponding terms aj xj can be moved to the left side of Eq. A-1 and the previous analysis applied to the remaining terms.

134

NAFTA 60 (3) 129-134 (2009)