**4.4. Hydrogen ion potential results**

Drilling muds are always treated to be alkaline (i.e., a pH > 7). The pH will affect viscosity, bentonite is least affected if the pH is in the range of 7 to 9.5. Above this, the viscosity will increase and may give viscosities that are out of proportion for good drilling properties. For minimizing shale problems, a pH of 8.5 to 9.5 appears to give the best hole stability and control over mud properties. A high pH (10+) appears to cause shale problems.

The corrosion of metal is increased if it comes into contact with an acidic fluid. From this point of view, the higher pH would be desirable to protect pipe and casing (Baker Hughes, 1995).

The pH values of all the samples meet a few of the requirements stated but Diesel OBM with a pH of less than 8.5 does not meet with specification. Algae, Jatropha, Moringa and Canola OBM's show better results since their pH values fall within this range.


**Table 6.** pH Values

64 New Technologies in the Oil and Gas Industry

Properties **DIESEL ALGAE JATROPHA MORINGA Canola** 

**Volume** 6.9ml 6.2ml 6.3ml 7.2ml 6.0 ml

**Oil volume** 2.3ml 1.1ml 1.1ml 2.5ml 1.0 ml **Water Volume** 4.6ml 5.1ml 4.2ml 4.7ml 4.3 ml **Cake Thickness** 1.0mm 0.9mm 0.8mm 0.9mm 0.78mm

iii. Differential sticking of the drillstring due to increased contact area and rapid

i. Formation damage due to filtrate and solids invasion. Damaged zone too deep to be remedied by perforation or acidization. Damage may be precipitation of insoluble compounds, changes in wettability, and changes in relative permeability to oil or gas,

ii. Invalid formation-fluid sampling test. Formation-fluid flow tests may give results for

iii. Formation-evaluation difficulties caused by excessive filtrate invasion, poor transmission of electrical properties through thick cakes, and potential mechanical

iv. Erroneous properties measured by logging tools (measuring filtrate altered properties

v. Oil and gas zones may be overlooked because the filtrate is flushing hydrocarbons

Drilling muds are always treated to be alkaline (i.e., a pH > 7). The pH will affect viscosity, bentonite is least affected if the pH is in the range of 7 to 9.5. Above this, the viscosity will increase and may give viscosities that are out of proportion for good drilling properties. For minimizing shale problems, a pH of 8.5 to 9.5 appears to give the best hole stability and

The corrosion of metal is increased if it comes into contact with an acidic fluid. From this point of view, the higher pH would be desirable to protect pipe and casing (Baker Hughes, 1995).

control over mud properties. A high pH (10+) appears to cause shale problems.

Filtration

**Total Fluid** 

**Table 5.** Mud Filtration Results

Problems caused as a result of excessive thickness include4:

ii. Increased surges and swabbing due to reduced annular clearance.

development of sticking forces caused by higher filtration rate.

The problems as a result of excessive filtration volumes include4:

iv. Primary cementing difficulties due to inadequate displacement of filter cake.

formation plugging with fines or solids, and swelling of in-situ clays.

i. Tight spots in the hole that cause excessive drag.

the filtrate rather than for the reservoir fluids.

problems running and retrieving logging tools.

away from the wellbore, making detection more difficult.

rather than reservoir fluid properties).

**4.4. Hydrogen ion potential results** 

v. Increased difficulty in running casing.

#### **4.5. Results of cuttings carrying index**

Only three drilling-fluid parameters are controllable to enhance moving drilled solids from the wellbore:Apparent Viscosity (AV) density (mud weight [MW]), and viscosity. Cuttings Carrying Index (CCI) is a measure of a drilling fluid's ability to conduct drilled cuttings in the hole. Higher CCI's, mean better hole cleaning capacities.

From the Table, we can see that Jatropha OBM showed best results for CCI iterations.


**Table 7.** Cuttings Carrying Indices (CCI's)

#### **4.6. Pressure loss modeling results**

The Bingham plastic model is the standard viscosity model used throughout the industry, and it can be made to fit high shear- rate viscosity data reasonably well, and is generally associated with the viscosity of the base fluid and the number, size, and shape of solids in the slurry, while yield stress is associated with the tendency of components to build a shearresistant.


**Table 8.** Bingham Plastic Pressure Losses in Psi

It can be seen from the table that Jatropha and Canola OBM's gave better pressure loss results than Diesel OBM as a result of lower plastic viscosities, and hence should be encouraged for use during drilling activities.

#### **4.7. Result of the toxicity measurements**

Samples of 100ml of each of the selected oils were exposed to both corn seeds and bean seed and the no of days which the crop survived are as indicated in Figure 16. The growth rate was also measured i.e the new length of the plant was measured at regular time intervals. For the graph of toxicity of diesel based mud the reduced growth rate indicates when the leaves began to yellow, and the zero static values indicate when the plant died.

From the results indicated by the figure 16, it can be concluded that jatropha oil has less harmful effect on plant growth compared to canola and diesel. Bean seeds were planted and after one week, they were both exposed to 100ml of both jathropha formulated mud and diesel formulated mud. The seeds exposed to jatropha survived for 18 days, while that exposed to diesel mud survived for 6 days and then withered. When the soil was checked, there was no sign of any living organisms in diesel mud sample while that of the jatropha mud, there were signs of some living organisms such as earth worms, and other little insects. This shows that jatropha mud sample is environmentally safer for both plants and micro animals than diesel mud sample.

From the figure 17, it can be seen that the seeds exposed to jatropha had the highest number of days of survival which indicates its lower toxicity while that of diesel had the lowest days of survival which indicates its high toxicity. The toxicity of diesel can be traced to high aromatic hydrocarbon content. Therefore, replacements for diesel should either eliminate or minimize the aromatic contents thereby making the material non toxic or less toxic. Biodegradation and bioaccumulation however depend on the chemistry of the molecular character of the base fluids used. In general, green material i.e plant materials containing oxygen within their structure degrade easier.

**Figure 16.** Comparison of Growth Rate Curve of Different Mud Types

#### **4.8. Results of density variation with temperature**

Densities were measured for the various samples at temperatures ranging from 30OC to 80OC and are summarized in Table 9.

Novel Formulation of Environmentally Friendly Oil Based Drilling Mud 67

**Figure 17.** Toxicity of different mud types

**4.7. Result of the toxicity measurements** 

micro animals than diesel mud sample.

oxygen within their structure degrade easier.

**Figure 16.** Comparison of Growth Rate Curve of Different Mud Types

**4.8. Results of density variation with temperature** 

80OC and are summarized in Table 9.

Samples of 100ml of each of the selected oils were exposed to both corn seeds and bean seed and the no of days which the crop survived are as indicated in Figure 16. The growth rate was also measured i.e the new length of the plant was measured at regular time intervals. For the graph of toxicity of diesel based mud the reduced growth rate indicates when the

From the results indicated by the figure 16, it can be concluded that jatropha oil has less harmful effect on plant growth compared to canola and diesel. Bean seeds were planted and after one week, they were both exposed to 100ml of both jathropha formulated mud and diesel formulated mud. The seeds exposed to jatropha survived for 18 days, while that exposed to diesel mud survived for 6 days and then withered. When the soil was checked, there was no sign of any living organisms in diesel mud sample while that of the jatropha mud, there were signs of some living organisms such as earth worms, and other little insects. This shows that jatropha mud sample is environmentally safer for both plants and

From the figure 17, it can be seen that the seeds exposed to jatropha had the highest number of days of survival which indicates its lower toxicity while that of diesel had the lowest days of survival which indicates its high toxicity. The toxicity of diesel can be traced to high aromatic hydrocarbon content. Therefore, replacements for diesel should either eliminate or minimize the aromatic contents thereby making the material non toxic or less toxic. Biodegradation and bioaccumulation however depend on the chemistry of the molecular character of the base fluids used. In general, green material i.e plant materials containing

Densities were measured for the various samples at temperatures ranging from 30OC to

leaves began to yellow, and the zero static values indicate when the plant died.


**Table 9.** Density Changes in ppg at Varying Temperatures.

The mud samples were heated at constant pressure, and in an open system, hence the density increment.

At temperatures of 60OC and 70OC, the densities of Diesel and Jatropha OBM's were constant, while that happened with Canola OBM at a lower range of 40OC and 50OC. This is shown in Figure 18. This could be due to the differences in temperature and heat energy required to dissipate bonds, which vary with fluid properties (i.e the continuous phases).

**Figure 18.** Density against Temperature (Diesel, Jatropha and Canola OBM's)

After the results were recorded, extrapolations were made and hypothetical values were derived for temperatures as high as 320OC, to enhance the prediction using Artificial Neural Network (ANN).

These values are summarized Tables 10 to 12


**Table 10.** Hypothetical Temperature-Density Values (extrapolated from regression analysis).

#### **4.9. Results of neural networking**

68 New Technologies in the Oil and Gas Industry

These values are summarized Tables 10 to 12

Network (ANN).

After the results were recorded, extrapolations were made and hypothetical values were derived for temperatures as high as 320OC, to enhance the prediction using Artificial Neural

 **Diesel Jatropha Canola 30OC** 10 10 10 **40OC** 10.1 10.05 10.05 **50OC** 10.17 10.1 10.05 **60OC** 10.2 10.15 10.1 **70OC** 10.2 10.15 10.15 **80OC** 10.25 10.2 10.17 **90OC** 10.31133 10.24333 10.20667 **100OC** 10.35648 10.2819 10.24095 **110OC** 10.40162 10.32048 10.27524 **120OC** 10.44676 10.35905 10.30952 **130OC** 10.4919 10.39762 10.34381 **140OC** 10.53705 10.43619 10.3781 **150OC** 10.58219 10.47476 10.41238 **160OC** 10.62733 10.51333 10.44667 **170OC** 10.67248 10.5519 10.48095 **180OC** 10.71762 10.59048 10.51524 **190OC** 10.76276 10.62905 10.54952 **200OC** 10.8079 10.66762 10.58381 **210OC** 10.85305 10.70619 10.6181 **220OC** 10.89819 10.74476 10.65238 **230OC** 10.94333 10.78333 10.68667 **240OC** 10.98848 10.8219 10.72095 **250OC** 11.03362 10.86048 10.75524 **260OC** 11.07876 10.89905 10.78952 **270OC** 11.1239 10.93762 10.82381 **280OC** 11.16905 10.97619 10.8581 **290OC** 11.21419 11.01476 10.89238 **300OC** 11.25933 11.05333 10.92667 **310OC** 11.30448 11.0919 10.96095 **320OC** 11.34962 11.13048 10.99524

**Table 10.** Hypothetical Temperature-Density Values (extrapolated from regression analysis).

From the Artificial Neural Network Toolbox in the MATLAB 2008a, the following results were obtained:

60% of the data were used for training the network, 20% for testing, and another 20% for validation.


On training the regression values, returned values are summarized in Table 11


Since all regression values are close to unity, this means that the network prediction is a successful one.

The graphs of training, testing and validation are presented below:

The values were returned after performing five iterations for each network. This also goes to say that the Artificial Neural Network, after being trained and simulated, is a viable and feasible instrument for prediction.

Figures 19 to 31 present the plots of Experimental data against Estimated (predicted) data for training, testing and validation processes from MATLAB 2008.

**Figure 19.** Diesel OBM Validation values

**Figure 20.** Diesel OBM Test values

**Figure 21.** Diesel OBM Training values

**Figure 22.** Diesel OBM Overall values

Novel Formulation of Environmentally Friendly Oil Based Drilling Mud 71

**Figure 23.** Diesel OBM Overall values

70 New Technologies in the Oil and Gas Industry

**Figure 20.** Diesel OBM Test values

**Figure 21.** Diesel OBM Training values

**Figure 22.** Diesel OBM Overall values

**Figure 24.** Jatropha OBM Validation values

**Figure 25.** Jatropha OBM Test values

**Figure 26.** Jatropha OBM Training values

**Figure 27.** Jatropha OBM Overall values

**Figure 28.** Canola OBM Validation values

Novel Formulation of Environmentally Friendly Oil Based Drilling Mud 73

**Figure 29.** Canola OBM Test values

72 New Technologies in the Oil and Gas Industry

**Figure 26.** Jatropha OBM Training values

**Figure 27.** Jatropha OBM Overall values

**Figure 28.** Canola OBM Validation values

**Figure 30.** Canola OBM Training values

**Figure 31.** Canola OBM Overall values

We can see from the Figures 19 to 31 that the data points all align closely with the imaginary/arbitrary straight line drawn across. This validates the accuracy of the network predictions and this also gives rise to the high regression values (tending towards unity) presented in Table 11



**Table 12.** Errors, Experimental Values, and Estimated Values for Diesel OBM


Novel Formulation of Environmentally Friendly Oil Based Drilling Mud 75


**Table 13.** Errors, Experimental Values, and Estimated Values for Jatropha OBM

74 New Technologies in the Oil and Gas Industry

presented in Table 11

We can see from the Figures 19 to 31 that the data points all align closely with the imaginary/arbitrary straight line drawn across. This validates the accuracy of the network predictions and this also gives rise to the high regression values (tending towards unity)

Errors, estimated values and experimental values are summarized in Tables 12 to 14

**Table 12.** Errors, Experimental Values, and Estimated Values for Diesel OBM

**Temperature oC Exp Values Est Values Errors**  30 10 10 0 40 10.05 10.05 0 50 10.1 10.0998 -0.0002 60 10.15 10.1485 -0.0015

**Temperature oC Exp Values Est Values Errors**  30 10 10.049 0.049 40 10.1 10.1407 0.0407 50 10.17 10.1794 0.0094 60 10.2 10.2022 0.0022 70 10.2 10.2236 0.0236 80 10.25 10.24 -0.01 90 10.31133 10.287 -0.02433 100 10.35648 10.3579 0.001424 110 10.40162 10.3904 -0.01122 120 10.44676 10.4222 -0.02456 130 10.4919 10.4835 -0.0084 140 10.53705 10.5204 -0.01665 150 10.58219 10.5455 -0.03669 160 10.62733 10.6133 -0.01403 170 10.67248 10.687 0.014524 180 10.71762 10.7202 0.002581 190 10.76276 10.7714 0.008638 200 10.8079 10.8335 0.025595 210 10.85305 10.8611 0.008052 220 10.89819 10.8991 0.00091 230 10.94333 10.9623 0.018967 240 10.98848 10.9955 0.007024 250 11.03362 11.0273 -0.00632 260 11.07876 11.085 0.006238 270 11.1239 11.1195 -0.0044 280 11.16905 11.1474 -0.02165 290 11.21419 11.2049 -0.00929 300 11.25933 11.2432 -0.01613 310 11.30448 11.2545 -0.04998 320 11.34962 11.2674 -0.08222



**Table 14.** Errors, Experimental Values, and Estimated Values for Canola OBM

The minute errors encountered in the predictions further justify the claim that the ANN is a trust worthy prediction tool.

The Experimental outputs were then plotted against their corresponding temperature values, and also fitted into the polynomial trend line of order 2.

The Equations derived are7:

Diesel OBM:

$$
\rho = -4 \times 10^{-7} T^2 + 0.004T + 9.915 \tag{1}
$$

Jatropha OBM:

$$
\rho = 7 \times 10^{-7} T^2 + 0.003T + 9.994 \tag{2}
$$

Canola OBM:

$$
\rho = -2 \times 10^{-6} T^2 + 0.004T + 9.827 \tag{3}
$$

Also by comparing the networks created with that of Osman and Aggour12 (2003), we can see that this work is technically viable in predicting mud densities at varying temperatures as the network developed in the course of this project showed regression values close to those proposed by Osman and Aggour12.

Errors, percentage errors and average errors as compared with Osman and Aggour12 are relatively lower, thus guaranteeing the accuracy of the newly modeled network.

Table 15 shows the regression values of Osman and Aggour for oil based mud density variations with temperature and pressure12.



**Table 15.** Table Showing the Regression Values from Osman and Aggour12

76 New Technologies in the Oil and Gas Industry

trust worthy prediction tool.

The Equations derived are7:

Diesel OBM:

Jatropha OBM:

Canola OBM:

**Temperature oC Exp Values Est Values Errors**  170 10.48095 10.4994 0.018448 180 10.51524 10.519 0.003762 190 10.54952 10.5537 0.004176 200 10.58381 10.5952 0.01139 210 10.6181 10.6145 -0.0036 220 10.65238 10.6444 -0.00798 230 10.68667 10.6888 0.002133 240 10.72095 10.7105 -0.01045 250 10.75524 10.7365 -0.01874 260 10.78952 10.7895 -2.4E-05 270 10.82381 10.8224 -0.00141 280 10.8581 10.8465 -0.0116 290 10.89238 10.8971 0.004719 300 10.92667 10.9337 0.007033 310 10.96095 10.945 -0.01595 320 10.99524 10.9562 -0.03904

**Table 14.** Errors, Experimental Values, and Estimated Values for Canola OBM

values, and also fitted into the polynomial trend line of order 2.

those proposed by Osman and Aggour12.

The minute errors encountered in the predictions further justify the claim that the ANN is a

The Experimental outputs were then plotted against their corresponding temperature

7 2

7 2

6 2

Also by comparing the networks created with that of Osman and Aggour12 (2003), we can see that this work is technically viable in predicting mud densities at varying temperatures as the network developed in the course of this project showed regression values close to

Errors, percentage errors and average errors as compared with Osman and Aggour12 are

relatively lower, thus guaranteeing the accuracy of the newly modeled network.

4 10 0.004 9.915 *T T* (1)

7 10 0.003 9.994 *T T* (2)

2 10 0.004 9.827 *T T* (3)

**Table 16.** Table of the Relative Deviations

Table 17 compares the Average Absolute Percent Error abbreviation (AAPE), Maximum Average relative deviation (Ei) and Minimum Ei for Diesel, Jatropha and Canola OBM's as well as the values from Osman and Aggour.


**Table 17.** Table Comparing Maximum Ei, Minimum Ei, and AAPE

## **5. Conclusion**

The lower viscosities of jatropha, moringa and canola oil based mud (OBM's) make them very attractive prospects in drilling activities.

The results of the tests carried out indicate that jatropha, moringa and canola OBM's have great chances of being among the technically viable replacements of diesel OBM's. The results also show that additive chemistry must be employed in the mud formulation, to make them more technically feasible. In addition, the following conclusions were drawn:

