Machine learning and Artificial Intelligence Techniques to Expect Drilling Fluid Density , Rate Infiltration and Loss Circulation Anticipation

Machine learning and Artificial Intelligence Techniques to Expect Drilling Fluid Density , Rate Infiltration and Loss Circulation Anticipation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine learning and Artificial Intelligence Techniques to Expect Drilling Fluid Density , Rate Infiltration and Loss Circulation Anticipation Abdoulaye Seyni Mahamadou, Gu Jun This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3857471/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The success of drilling operations depends on several factors, including the drilling properties, and environmental and financial constraints. However, relying solely on lab testing is often insufficient due to limited resources and time. Drilling operations can result in significant costs due to loss of circulation. To address this issue, we propose using AI and machine learning techniques to predict drilling fluid density and prevent circulation loss using an intelligent approach. We conducted scientific evaluations of the rheological properties and mud components of polyacrylamide/polyethyleneimine (PAM) mud. Four distinct ML algorithms (the adaptive neuro-fuzzy inference system, particle swarm optimization-based adaptive neuro-fuzzy inference system, least squares support vector machine with a genetic algorithm, and radial basis function) were used to investigate the rheological qualities of various mud components at different concentrations and test conditions. In the LSSVM-GA model, we found that the linear equation for predicting fluid density was "y = 1.0041x + 0.0019", with a correlation coefficient (R2) of 0.9966. The RBF model was used to predict fluid density due to its superior performance over other conventional models. The linear equation for predicting fluid density was "y = 1.0009x + 0.0034", with a correlation coefficient (R2) of 0.9999. Based on our experience, we have found that by using an appropriate combination of materials, we can achieve satisfactory rheological properties, thereby avoiding circulation loss incidents Machine learning Artificial Intelligence lost circulation Drilling Fluid Rate of Penetration Prediction Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Lately, there have been tremendous degrees of progress in the oil and gas industry; The expulsion of cuttings and the energy of placing mud into the borehole are two of the main advantages of the relentless interaction. Indeed, even at low shear rates, penetrating liquids require an expansion in consistency to suspend and ship particles. Li, M.-C.; Wu, Q.; 2018); It is essential to keep in mind that a significantly developed entering mud has the potential to reduce total working costs by 5 to 15%2 (Bloys, B.; Davis 1994), specialists have organized changed getting procedures to find the key parts impacting the oil business (Hornik et al. 1989 ). All along, it was in 1990 that Arehart (1990) proposed an arrangement for closing the ROP looking at the strategy for counterfeit brain affiliations (ANN). Water-driven power per square inch (HSI), WOB, T, and RPM were the information components in this work. Exhausting exercises require the use of entering fluid, and going with some inadmissible or undesirable choice of this part can emphatically influence communication. Kutassov et al. ( 1988) and Babu et al. ( Taking into account the thermodynamic properties of a wellbore, Kutasov ( 1988 ; 1993) introduced a preliminary relationship for concluding entering liquid thickness (Kutasov, 1988 ;). Effective models for identifying and anticipating significant boundaries are essential for the oil and gas industry. In this way, mechanized thinking methods have truly been used in different areas of the oil and gas industry, as shown under. Supplies (Ghorbani et al., 2017b); oil well wandering (Abdali et al., 2021 ); harm to the game plan (Mohammadian and Ghorbani, 2015), wellbore disparity (Darvishpour et al., 2019), rheology and filtration (Davoodi et al., 2019, 2020 ), and development (Ghorbani and Moghadasi, 2014 a; Ghorbani and other, 2014b; et al., 2017a) and 2019 entry fluid). A geomechanical strategy and shear modulus-based check of bundling breakdown (Moradi, H.,2012). A solid and practical method for imagining impairing liquid thickness, the type of invading liquid (Newtonian or non-Newtonian), ROP, WOB, and so on, ought to be viewed as in the made plan (Li et al., 2017; Guo and co., 2021). ). As of late, extraordinary man-made knowledge procedures, for instance, soft reasoning, FL, support vector machine, SVM, genetic computation fake cerebrum association, ANN, have been participated in oil mining, especially in the field of entering fluid planning. A piece of these applications consolidates predicting fluid spread plans in the wellbore annulus [. Mahmoud, A.A., Elkatatny 2020). The organization is prepared to utilize different calculations to limit blunders to propel predisposition and mass. Levenberg-Marquardt (LM), an iterative bend fitting calculation, is one of these calculations. In tackling nonlinear least squares problems, this calculation has demonstrated its uncommon success (Rashidi, S., 2020). The leftover ML to added because of late progressions in calculations and further enlisting working environments (Dimiduk et al. 2018 ; LeCun and Co. 2015). Using ML techniques, mistakes made during directional boring tasks can be derived. The methods planned work by contrasting ongoing information and mishaps that happened before (Gurina et al. 2020 ). Penetrating muds are divided into three rules: water-based, manufactured, and oil-based muds. Due to concerns about the environment and the cost, water-based mud is a better and more appealing option for drilling (Salih, A.; 2016 Elshehabi). Diminished dissemination is a typical boring issue, especially in exceptionally penetrable developments, exhausted repositories, and cracked or huge developments (Moosavi, S. R 2019 ); According to Duan, J., Zhao, J., Xiao ), the range of loss of circulation issues ranges from shallow, unreinforced formations to well-reinforced formations that are fractured by drilling mud's hydrostatic loading. Loss of flow episodes, alluded to as the deficiency of penetrating liquids in the development, is known to be one of the most troublesome issues to anticipate or alleviate during the boring stage. The seriousness of the results changes depending upon the reality of the calamity; it could begin with a straightforward loss of penetrating liquid and end with a victory (Whittaker, A 2011 ). 2. Modeling lost Circulation The examination of data has become a popular topic in recent times due to the vast amounts of information that are stored and available. Using data analysis techniques can help to evaluate and understand the performance of a particular process, and it can help to predict future outcomes. Many companies have used data analysis, especially for fuel plans and supply plans. However, there is a lack of data analysis in drilling, as most drilling data is private and protected by organizations. To address this, researchers have used artificial intelligence techniques to predict issues and select the best course of action before entering the occurrences zone. There are several data mining techniques used in this regard, such as Innocent Bayes, Event-Based, and Brain Relationships. Researchers have also developed new models to focus on volume loss, crude streaming density (ECD), and rate of penetration (ROP) for lost flow events. These new models have been compared to previous models developed by other researchers. 3. Artificial intelligence Algorithm Various computer-based data analyses have been carried out in different areas of the oil and gas industry, including materials, flooding of oil wells, damage to the philosophy, imbalance at the breaking point, rheology and filtration, and development and moving into liquid. Geomechanical and shear modulus-based estimates of packaging breakdown have also been conducted. The readiness model is a model that changes the dependence model based on a chosen computation. The test is conducted to examine the reliability of the used model. Figure 1 shows the flowchart for the Firefly algorithm. The following parameters were used as input data: pressure at the standpipe, tank rank 02, input flow, hook load, torque of the rotary table, penetration rate, weight on the bit, and gas rate 4. Rheology of drilling and hydraulic fluids Rheology and liquid water power are fundamental variables in penetrating strategies. Rheology is depicted as the assessment of the stream and curving of materials, especially entering liquids, and power through pressure recommends the strategy for overseeing acting of the development of liquids, including mud, in the entering framework, which utilizations lines, valves, and experts for pass entering mud, discard stays, cool woods, and oil up the drill string. The capacity of the siphon, the harmony of the wellbore walls, the financial implications of bundling, and the fluid's capacity to exert pressure on the wellbore walls are all affected by the rheology of the infiltrating fluid. Rheological pieces of entering fluids, for instance, PV, YP, gel strength, and thickness can change concerning temperature. As a depiction, if you increase the HP of fluid, this will have explicit effects. This is especially an expansion in ECD, flood and swab strain, and the likelihood of differential adhering because of additional solids in the liquid. Moreover, the expansion in the consistency of the plastic will prompt a decrease in the entrance rate because of the unfortunate disinfection of the openings (. Yavari, H., Sabah, M 2018 ). The investigation of liquid tension, stream, and speed in the drill table framework is planned in liquid water power. The water-driven elements of penetrating liquid affect the efficiency of different tasks, for example, mud unloading, boring and course, an appropriately requested water-driven strategy can keep away from blockages and guarantee proficient waste expulsion. Consequently, the rheology and hydrodynamics of the mud are vital parts of penetrating procedures that should be considered while arranging and completing boring activities (Elkatatny, S.; Tariq, Z.; 2017). 5. Rheology and Mechanization Measurement System The use of artificial intelligence (AI) techniques is essential for a complete system that can provide the rheological characteristics of different types of drilling fluids. Current devices are examined using AI approaches, but AI prediction methods are still necessary. Pipe viscosity measures the time required to flow a fixed volume of liquid (930 cm3) through the open hole of the pipe. According to Igwilo, K.C.; Okoro, E.E.; and Ohia, the impact of fluid properties on well control and drilling operations requires a different measurement rate for each property. Therefore, AI is considered an alternative research horizon for predicting the rheological properties of mud. The automatic instruction computer coding process is used to determine how far mathematical correlations can go by utilizing the relationships between elements in a set of data to provide an initiation capability. This procedure is evolved based on the factual breakdown of information and calculations to accomplish the learning objective for arrangement or expectation purposes (Agwu, O.E.; Akpabio, 2018). 6. Support vector machine learning method Establishing AI devices is one of the numerical machines that are suitable for providing a sufficient allowance for oil penetration. Because of the different parts of AI devices in boring activities and the exactness of their expectations for field applications, they depend on the investigation of traits from a gathered informational index. This will empower us to foster a satisfactory framework in the boring area in light of relapse examination. De Moor and Suykens (2002) 7. ANFIS model To find this parameter, we first used the ANFIS algorithm to specify the drilling fluid's density. The legitimate straight condition to deduce exhausting fluid thickness is "y = 0.9647x + 0.0412" with the relationship variable of (R2 = 0.8072). To be even more clear on the entering liquid thickness gauge by ANFIS, we add a groundbreaking selection of the fundamental information. It is gotten from the aphorism "y = 0.4547e0.742x" and the association coefficient R 2 of 0.8139. Along these lines, a remarkable derivation would be good for working out entering fluid thickness rather than an immediate condition. To better understand the prediction, the model must be developed using power derivation or longitude. 8. PSO-ANFIS model A PSO-ANFIS computation was added to represent the improvement of the ANFIS system to give additional exact speculations. For this, it is focal that the agreeable direct condition to expect the thickness of the attacking fluid is "y = 0.9877x + 0.0288" with the relationship part of (R2 = 0.9225). We add an amazing derivation based on the exploratory data to make the incapacitating fluid thickness measurement by PSO-ANFIS significantly clearer. It is obtained from "y = 0.4719e0.7314x" and an R 2 = 0.9068 relationship component. The exceptional derivation would be more effective at imaging the thickness of the entering fluid than the straight condition. Man-made keen limits and changed heading contraptions should be executed to equip engineers with extra data to sort out an exquisite and harmless standard design method ( Davarpanah and Mirshekari, 2019 ). 9. ANN model ANN is one of the most widely recognized artificial intelligence techniques, important to address particular designing issues of colossal intricacy that outperform the computational capacity of traditional math and activities (Salehi, S 2016). The construction of the ANN comprises three all-encompassing sorts of layers. The principal concerns the info boundaries. The neurons that total the trade capacities among information sources and results are arranged in the ensuing layer, which is known as the mystery layer. The third sort concerns exits. These layers, with the fitting preparation calculation, uncover the idea of the issue ( Gamal, H.; Abdelaal 2021). 10. Volume loss model The system creation procedure requires a section on the number of latent factors. Score plots help sort the ideal number of latent factors that will be used in the model. Unlike Principal Constituent Analysis (PCA), PLS score plots are calculated to explain the variation in x and y and to maximize the relationship between the x and y variables. Choosing the optimal number of latent factors is a complicated process and requires trial and error until the optimal number of latent factors is reached. Using too many latent factors will lead to over-adaptation of the system, which in turn will reverse the sign of certain variables and make the model unrealistic seen in Fig. 2 . On the other hand, the use of a very small number of latent factors will not explain The limit among x and y. Two measures are utilized to wrap up the ideal number of languid parts. The significant differentiations diminish to the mean supporting of the typical extra level of squares (PRESS). The second permits you to control the score plots for the x and y factors, and each dormant part will show a score plot of x versus y. The system ought to remember the inactive variable if the score outline has an example. The latent part ought to be ignored regardless of this, provided that no model is displayed in the score frame (Dong, S., Zhu, H 2011 ). The focal mean of PRESS as a piece of how many lazy variables is portrayed in Fig. 6 . The figure is given out for ten lazy components to investigate the best number of idle parts, believe it or not. Applying the fundamental measure, it is easy to see that having someplace close to two lethargic components will reduce the main mean of PRESS. Regardless, an undefined quantity of inactive variables should be picked. This is where score plots come in. The score plots for six dormant parts are depicted in Fig. 3 . Applying the accompanying standard, having different idle parts won't add any huge data to the turn of events, since there are on a very basic level no dormant parts. The proportion of variation explained by each factor for x-variables and y, respectively, is depicted in Fig. 4 . decipted Scores plots for the ECD model. 11. Field Data Description Utilizing man-made consciousness (simulated intelligence), this study expects to foresee constant mud mass (MW) and comparable flowing thickness (ECD) during penetrating activities. By conveying results that are straightforward and trust, this approach desires to support penetrating effectiveness. Due to its abundance of moderate and crucial data, the stretch of a few spots in the degree X3000 and X4200 ft was chosen for the use of the two new models and replicated data. We joined 4,371 records for ECD and 33,588 records for MW achieved by meandering toward the sea gas well A, then wandered toward the sea oil well B, and then floated away from the seaward gas well A. level oil C. This reach was viewed as urgent because of the way that it needed practically no missing information and contained data of unlimited importance. By distinguishing this specific region, the review had the option to guarantee the evaluation's comfort and worth by smoothing out the utilization of assets and time. Facilitators were able to focus on the most significant and concrete information thanks to this procedure, improving the overall quality and authenticity of their results. The rheological properties, like plastic consistency and yield strength, were studied utilizing a rheometer at 48°C and at standard barometrical strain in this review, which was on an exceptionally major level more enormous in picking the thickness of the entering fluid. Besides, a mud counter was used to dissipate the liquid stage and gather the additional solids, and Swamp channel consistency was assessed using a Bog line at room temperature and including strain. What's more, other vitals like the entry rate, mud siphon stream rate, and were recorded to wellspring pressure. The enlightening rundown was divided into a status set (80%) and a test set (20%), with the unpredictable state set at 42, to guarantee that the planning and test sets would remain connected throughout the various runs. Because of how hyper limit blocking was completed, a help set was not viewed as central, so the 80:20 split degree was picked because the standard degree is utilized the most. To anticipate MW and ECD, the initial approach made use of three designs: express Choice Tree (DT), Counterfeit Brain Affiliation (ANN), and Sponsorship Vector Machine (SVM), all of which were modified in Python. The portrayal of the three structures for expecting ECD and MW is depicted as follows: More about this source texture text expected for extra grasping data Equivalent circulating density (ECD) prediction The relationship between ECD-PWD and various elements, including SPP (psi), LSYP, and GPM, has been examined for the suspicion for ECD. ECD-PWD (PCF) revealed the most grounded positive relationship with GPM (0,951728), while various endpoints equivalently uncovered a positive relationship with ECD-PWD (PCF) with values more clear than 0,8. 12. Mud weight (MW) Prediction The relationship between ECD-PWD and several sections, such as SPP (psi), LSYP, and GPM, was considered for ECD doubt. ECD-PWD (PCF) had the most grounded positive relationship with GPM (0.951727), yet unique endpoints with values above 0.8 additionally had a positive relationship with ECD-PWD (PCF). Rheology information Rheological information was made pondering the particular mud recipes, material aggregate, and test temperature. In this part, the impact of PAM, bentonite, and The liberality of the mud on the rheological properties (like plastic consistency, and clear thickness) is incorporated. We took a gander at different PAM fixations going from 6 to 13 weight percent to perceive what they meant for boring liquid power through pressure and, thus, the same traffic thickness (EPD) depicts the PV got from viscometer assessments of various social affairs of non-crosslinked PAM in refined water. Figure 4 depicts the AV values for various PAM obsessions at 1021, 10, and 5 s shear speeds, as they were. The qualities similarly show the impact of temperature, which was significantly more displayed at low shear rates. It was seen that the Accomplice rate went with an unending move in VA and PV. Sensible qualities were acquired with polymer levels of 7.5 to 10% by weight. Past With these convergences of PAM, the viscosities would in general increment thoroughly. The polymer will in general show huge viscosities at this high convergence of 10 to 12.5 weight percent; Notwithstanding this, at high temperatures (200F), a 70–75% decline in thickness was noticed contrasted with the qualities estimated at surface temperature. 13. Feat and Accuracy of Predictions For other PAM/PEI and mud-added substance structures, intriguing rheological data speculations can be made by utilizing various evaluations. Figure 12 secludes the commonplace attributes from the test dataset with the guaranteed credits utilizing the four assessments attempted in this review for high most likely gains of shear rate. It might be estimated that the entire sagacious imperative of point helping was more splendid than another evaluation for plastic and clear consistency. Dismissing this, the presentation of Ada Lift in foreseeing a further degree of plastics and clear consistency defeated the accomplishments of different assessments. Counting a similar framework for the low shear rate information was excruciating. As depicted in Fig. 5 , the expected characteristics and the actual characteristics diverged, and the gauge capacity was substantially reduced. In pursuit to gauge the viability of these calculations, the dataset was recklessly picked for preparation and tests in a degree of 20:80. For both consistency respects, every calculation was run on different events, and quantifiable appraisal was achieved. The ideal outcome for every calculation utilizing the high shear rate data is displayed. Table 1 Mud Component and Concentrations. Component/Tem Concentrations lb/bbl Weight % PAM 6 6 7 17–44 5-12.5 PEI wt.% (lb/bbl) 0 0.25 0.5 9.8-5 0.25–1.5 Bentonite wt.% (lb/bbl) 1 2 2.5 0–5 0–15 Caustic soda (lb/bbl) 0 0.5 2 0-1.5 0-0.4 Mud weight (ppg)/ 9.5 10 11.5 Oct-90 2 to 30 Barite (lb/bbl) 10 54 90 0–15 3.5 Fiber (lb/bbl) 0 5 10 0.5 0.8 Temperature ( F) 70 120 160 180 200 Table 2 Mud and well Information used for the hydraulic Measurements Component Parameter Information Mud information Mud Density 9.5 ppg Water-based mud AV = 32, PV = 22 mPa.s PAM/PEI-based mud AV = 27, PV = 26 mPa.s Well information Well Depth 6000 Ft Surface Hole Diameter 7 7/8 In Main Hole Diameter 9 5/8 In Casing Depth 7 7/8 In DrillCollar Information ID 3500 Ft ID 2 1/4 In OD 6 1/4 In Drill Pipe Information OD 270 Ft Total Length 3 5/6 In Flow Rate Information ID 4 1/2 In OD 6430 Ft Total Length 275 GPM Surface Pressure loss 100 Psi Table 3 Composition of PAM/PEI-based Mud and the water-Mud Component PAM/PEI-Based Mud lb/bbl Water-Based Mud lb/bbl Water 312.3 318 Caustic soda 0.5 0.5 Lignite 0 4 0.4 Bentonite 3.5 20 Mud deflocculant 0 4 Calcium Carbonate 0 55 PAM 25 0 PEI 3.4 0 Barite 56,5 0.85 14. Discussion Various samples were used to determine the extent of the penetration rate based on different inputs. For the testing stage, 80% of the dataset was used for training, while the remaining part was used. The MLP-ABC model produced an extraordinarily high prediction accuracy for all four sets of samples evaluated, with RMSE values of less than 0.009812 m/h and R 2 values of 1.000 for the test subset, RMSE values of less than 0.006561 m/h and R2 values of 1.000 for the training subset, and RMSE values of less than 0.007411 m/h and R2 values of 1.000 for the total subset. Figure 6 displays the predicted and balanced ROP values for each data archive in each subset estimated by the four models. Figure 6 shows the accuracy based on R2 for the training, testing, and total datasets. The presented data allows for a comparison of the accuracy of four hybrid machine learning methods. These methods combine the multi-layer perceptron (MLP) algorithm with optimizers such as the firefly algorithm, the gravitational search algorithm, the artificial bee colony algorithm (ABC), and the independent component analysis algorithm optimizers to predict ROP. The advantage of these hybrid models is that they allow optimizers to quickly and cost-effectively adjust the control principles of machine learning models. The hybrid of algorithms includes MLP-FF, MLP-GSA, MLP-ABC, and MLP-ICA. Conclusion 1. The rheology of PAM/PEI-based mud is determined by the centralization of PAM, not by the PEI rate. The most desirable characteristics went from 7 to 10% by weight. Consistency was in a general sense impacted by different materials; regardless, the PAM rate ought to be loosened up to achieve allowed rheology, for high solid things like barite. 2. for the instance of lopsided rheological portrayal information getting together, inclination supporting was more fitting than different calculations, especially k-Closest Neighbor, Arbitrary Timberland, and AdaBoosting 3. Rheological information at low shear rates was difficult, albeit the expectation accomplishments were extremely low; in any case, great expectations were accomplished at low consistency rates where low paces of mud-added substances were utilized. Expanding the size of the dataset ought to build the presentation of the model. In light of the different writing surveys, a significant use of artificial intelligence in boring was introduced, from the send-off of the preparation plan to the expectation of the entrance rate. 4. During the introductory information examination, wells with difficulties experienced during boring were uncovered. To assign the uncovered PD, a PC method was introduced. 5. While breaking down the boring reports, a rundown of the principal components was ordered, which added to the model: tension at standpipes; tank volume; input stream rate; snare load; force of the rotational table; entrance rate; weight on the piece; gas content 6. A list of the main parameters and who contributed to the model was compiled during the analysis of the drilling reports: tension at standpipes; tank level; input stream rate; snare load; force of the rotational table; infiltration rate; weight on the piece; gas volume. 7. Estimating mud losses, ECD, and ROP is much easier with the newer models than with the older ones. 8. TFA is a vital component of the ROP model. The ROP model suffers as a result of this. Subsequently, the decision on spout size should be made cautiously. Declarations Ethics approval and consent to participate Funding No funding for this paper, for the APC option we will be in charge of the publication fee Data Availability Statement: All datasets generated and analyzed from the various experiments conducted for this work are available. The data was obtained through experiments conducted in the laboratories. The datasets used and analyzed during the current study are available in the main document. The authors, Abdoulaye Seyni Mahamadou and Professor GU JUN can easily share them without any major issues. All data generated or analyzed during this study are included in the published article to provide readers and novices in the field with more explanations and to bring innovation to the field of oil drilling. We are willing and order the publication of all data without any restrictions to reach the greatest number of readers and the public. It is necessary to explain how the data was obtained for clarity. We worked with Professor Gu Jun of China University of Geosciences Wuhan and together released ideas on the work of strengthening oil wells through our laboratory experiments. After the release of the various conclusions presented in the main document, we concluded that the data is self-produced as an innovation publication with the support of the articles we cited in the document. All data was collected according to our laboratory experiments. We declare that Abdoulaye Seyni Mahamadou and Issa Moctar Maimouna, as authors, contributed to the realization of the different experiments to be carried out, and all others contributed to the revision of this document. As students of China University of Geosciences, we share the same laboratory for the realization of our different works. Consent for publication All authors agree with the publication and consent and agree to the Publication References Li, M.-C.; Wu, Q.; Song, K.; De Hoop, C.F.; Lee, S.; Qing, Y.; Wu, Y. Cellulose Nanocrystals and Polyanionic Cellulose as Additives in Bentonite Water-Based Drilling Fluids: Rheological Modeling and Filtration Mechanisms. Ind. Eng. Chem. Res. 2016, 55, 133–143. [CrossRef] Bloys B, Davis N, Smolen B, et al. Designing and managing drilling fluid[J]. Oilfield Review, 1994, 6(2): 33-43. Hornik, K., Stinchcombe, M., White, H., 1989. Multilayer feedforward networks are universal approximators. Neural Network. 2 (5), 359e366. https://www.cs.cmu. edu/~bhiksha/courses/deeplearning/Fall.2016/notes/Sonia_Hornik.pdf. Hossain, M.E., Al-Majed, A.A., 2015. Fundamentals of Sustainable Drilling Engineering. John Wiley & Sons. Kutasov, I.M., 1988. Empirical correlation determines downhole mud density. Oil Gas J.. Li, Yuwei, Ghorbani, H., Moghadasi, J., 2014. Development of a new comprehensive model for choke performance correlation in Iranian oil wells. Adv. Environ. Biol. 8 (17), 877e882. http://www.aensiweb.net/AENSIWEB/aeb/aeb/September%202014/877-882.pdf. Ghorbani, H., Moghadasi, J., Mohamadian, N., Mansouri Zadeh, M., Hezarvand Zangeneh, M., Molayi, O., Kamali, A., 2014. Dev. New Comprehen. Model Choke Perform. Correlat. Iran. Gas Condensate Wells 8 (17), 308e313. Abdali, M.R., Mohamadian, N., Ghorbani, H., Wood, D.A., 2021. Petroleum well blowouts as a threat to drilling operation and wellbore sustainability: causes, prevention, safety and emergency response. J. Construct. Mater. Special Issue Sustain. Petrol. Eng. ISSN 2652, 3752. https://iconsmat.com.au/wp-content/ uploads/2021/02/si1.1r.pdf. Davoodi, S., Sa, A.R., Rukavishnikov, V., Minaev, K., 2020. Insights into application of acorn shell powder in drilling fluid as environmentally friendly additive: filtration and rheology. Int. J. Environ. Sci. Technol. 1e14. https://link.springer. com/article/10.1007/s13762-020-02880-0. Davoodi, S., Sa, A.R., Soleimanian, A., Jahromi, A.F., 2019. Application of a novel acrylamide copolymer containing highly hydrophobic comonomer as filtration control and rheology modifier additive in water-based drilling mud. J. Petrol. Sci. Eng. 180, 747e755. https://doi.org/10.1016/j.petrol.2019.04.069. Moradi, H., Bahari, M.H., Bagher, M., Sistani, N., 2012. Drilling rate prediction using an innovative soft computing approach. Sci. Res. Essays 5, 1583e1588. https:// lirias.kuleuven.be/retrieve/289930. Li, Yuwei, Jia, Dan, Rui, Zhenhua, Peng, Jiyong, Fu, Chunkai, 2017. Evaluation method of rock brittleness based on statistical constitutive relations for rock damage. J. Petroleum Sci. Eng. 153, 123–132. http://dx.doi.org/10.1016/j.petrol.2017.03.041. Guo, X., Liu, J., Dai, L., Liu, Q., Fang, D., Wei, A., Wang, J., 2021. Friction-wear failure mechanism of tubing strings used in high-pressure, high-temperature and high-yield gas wells. Wear http://dx.doi.org/10.1016/j.wear.2020.203576 Mahmoud, A.A., Elkatatny, S., Al-AbdulJabbar, A., Moussa, T., Gamal, H., Shehri, D.A., Artificial neural networks model for prediction of the rate of penetration while horizontally drilling carbonate formations. In: Paper Presented at the 54th US Rock Mechanics/Geomechanics Symposium. https://www.onepetro. org/conference-paper/ARMA-2020-1694. Rashidi, S., Mohamadian, N., Ghorbani, H., Wood, D.A., Shahbazi, K., Ahmadi Alvar, M., 2020. Shear modulus prediction of embedded pressurize salt layers and pinpointing zones at risk of casing collapse in oil and gas wells. J. Appl.Geophys. https://doi.org/10.1016/j.jappgeo.2020.104205, 104205 Dimiduk DM, Holm EA, Niezgoda SR (2018) Perspectives on the impact of machine learning, deep learning, and artificial intelligence on materials, processes, and structures engineering. Integr Mater Manuf Innov 7(3):157–172 LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521(7553):436–444 Gurina E, Klyuchnikov N, Zaytsev A, Romanenkova E, Antipova K, Simon I, Makarov V, Koroteev D (2020) Application of machine learning to accidents detection at directional drilling. J Pet SciEng 184:106519 Salih A H, Elshehabi T A, Bilgesu H I. Impact of nanomaterials on the rheological and filtration properties of water-based drilling fluids[C]//SPE Eastern regional meeting. SPE, 2016: SPE-184067-MS. Moosavi, S. R.; Wood, D. A.; Ahmadi, M. A.; Choubineh, A.ANN-Based Prediction of Laboratory-Scale Performance of CO2-Foam Flooding for Improving Oil Recovery. Nat. Resour. Res. 2019, 28,1619−1637. Duan, J., Zhao, J., Xiao, L., Yang, C., Li, C., 2015. A crop optimization approach based on improved bp neural network pso. In: Advances in Swarm and Computational Intelligence. Springer International Publishingpp, pp. 11e18. https://doi.org/ 10.1007/978-3-319-20469-7_2. Whittaker, A. Theory and Applications of Drilling Fluid Hydraulics; Springer: Berlin/Heidelberg, Germany, 2011; ISBN 978-9401088428 Al-Dhafeeri A M, Seright R S, Nasr-El-Din H A, et al. High-permeability carbonate zones (Super-K) in Ghawar Field (Saudi Arabia): identified, characterized, and evaluated for gel treatments[C]//SPE International Improved Oil Recovery Conference in Asia Pacific. SPE, 2005: SPE-97542-MS. Drucker, H.; Burges, C.J.; Kaufman, L.; Smola, A.; Vapnik, V. Support Vector Regression Machines. In Advances in Neural Information Processing Systems; The MIT Press: Cambridge, MA, USA, 1996; Volume 9, pp. 155–161. Jimmy, D.;Wami, E.; Ogba, M.I. Cuttings Lifting Coefficient Model: A Criteria for Cuttings Lifting and Hole Cleaning Quality of Mud in Drilling Optimization. In Proceedings of the SPE Nigeria Annual International Conference and Exhibition, Lagos, Nigeria, 1–3 August 2022. [CrossRef] Al-Hameedi, A.T.T., Alkinani, H.H., Dunn-Norman, S., Al-Alwani, M.A., lkhamis, M.M., Al-Bazzaz, W.H., 2019. Application of artificial intelligence in the petroleum industry: Volume loss prediction for naturally fractured formations. In: Soc. Pet. Eng. - SPE/IATMI Asia Pacific Oil Gas Conf. Exhib. 2019, APOG 2019. http://dx.doi.org/10.2118/196243-ms. Mohamadian, N., Ghorbani, H., Wood, D.A., Hormozi, H.K., 2018. Rheological and filtration characteristics of drilling fluids enhanced by nanoparticles with selected additives: an experimental study. Adv. Geo-Energy Res. 2 (3), 228e236. https://core.ac.uk/download/pdf/229382392.pdf. Ranaee, E., Ghorbani, H., Keshavarzian, S., Ghazaeipour Abarghoei, P., Riva, M., Inzoli, F., Guadagnini, A., 2021. Analysis of the performance of a crude-oil desalting system based on historical data. Fuel. https://doi.org/10.1016/j.fuel.2020.120046 Yavari, H., Sabah, M., Khosravanian, R., Wood, D.A., 2018. Application of an adaptive neuro-fuzzy inference system and mathematical rate of penetration models to pre- dicting drilling rate. Iran. J. Oil Gas Sci. Technol. 7, 73e100. https://doi.org/10.22050/IJOGST.2018.83374.1391. Elkatatny S M, Tariq Z, Mahmoud M A, et al. Optimization of rate of penetration using artificial intelligent techniques[C]//ARMA US Rock Mechanics/Geomechanics Symposium. ARMA, 2017: ARMA-2017-0429. Alsabaa A, Gamal H A, Elkatatny S M, et al. Real-time prediction of rheological properties of all-oil mud using artificial intelligence[C]//ARMA US Rock Mechanics/Geomechanics Symposium. ARMA, 2020: ARMA-2020-1645. Haklidir, Fusan Tut, and Mehmet Haklidir. 2019. "Geothermal reservoir temperatures prediction with a deep learning model: a case study western Anatolia." Stanford Geothermal Workshop. Stanford: Stanford University. Accessed January 18, 2019. Igwilo, K.C.; Okoro, E.E.; Ohia, P.N.; Adenubi, S.A.; Okoli, N.; Adebayo, T. Effect of MudWeight on Hole Cleaning During Oil and Gas Drilling Operations Effective Drilling Approach. Open Pet. Eng. J. 2019, 12, 14–22. [CrossRef] Agwu O E, Akpabio J U, Alabi S B, et al. Artificial intelligence techniques and their applications in drilling fluid engineering: A review[J]. Journal of Petroleum Science and Engineering, 2018, 167: 300-315. Suykens J A K, De Brabanter J, Lukas L, et al. Weighted least squares support vector machines: robustness and sparse approximation[J]. Neurocomputing, 2002, 48(1-4): 85-105. Davarpanah A, Mirshekari B. Effect of formate fluids on the shale stabilization of shale layers[J]. Energy Reports, 2019, 5: 987-992. Salehi, S.; Madani, S.A.; Kiran, R. Characterization of drilling fluids filtration through integrated laboratory ex-periments and CFD modeling. J. Nat. Gas Sci. Eng. 2016, 29, 462–468. [CrossRef] Gamal, H.; Abdelaal, A.; Elkatatny, S. Machine Learning Models for Equivalent Circulating Density Prediction from Drilling Data. ACS Omega 2021, 6, 27430–27442. [CrossRef] Dong, S., Zhu, H., Zhong, S., Shi, K., Liu, Y., 2021. New study on fixed-time synchronization control of delayed inertial memristive neural networks. Appl. Math. Comput. 399, 126035. http://dx.doi.org/10.1016/j.amc.2021.126035. Brown I, Mues C. An experimental comparison of classification algorithms for imbalanced credit scoring data sets[J]. Expert systems with applications, 2012, 39(3): 3446-3453. Naganawa, S.; Okatsu, K. Fluctuation of Equivalent Circulating Density in Extended Reach Drilling with Repeated Formation and Erosion of Cuttings Bed. In Proceedings of the IADC/SPE Asia Pacific Drilling Technology Conference and Exhibition, Jakarta, Indonesia, 25–27 August 2008; p Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3857471","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":267735159,"identity":"b5fe52f8-4869-4bb9-96d6-4b8bd83b5f0d","order_by":0,"name":"Abdoulaye Seyni Mahamadou","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9UlEQVRIiWNgGAWjYDADfmbG9g8fGBgSQJwPRGmRbG8+xjgDogVEEwEMzhxLY+YhRot8+9mHj3lzbBIbbuSYPbZts8vjZ29gbK7Ao4WxJ93YmHdbWmLjjBxz49y25GLJngOMjWfwaGFmSGOT5t12OLFZIsdAOreNOXHDjQT2hw14tLDxPwNp+Z/YBtJi2VYP0sLYiE8LjwTYlgOJPTzH0qQZ2w4T1iIh8YzZcO62ZOMZ7M2HDXvOHU+c2XOwEa8W+f40xgdvt9nJ7j/M2PjgR1l1Yj9780G8WmDAEayIkQ1MEqOBgcEeQv0hSvEoGAWjYBSMMAAAPVZSLfx9NIIAAAAASUVORK5CYII=","orcid":"","institution":"China University of Geosciences","correspondingAuthor":true,"prefix":"","firstName":"Abdoulaye","middleName":"Seyni","lastName":"Mahamadou","suffix":""},{"id":267735160,"identity":"e6006ad5-e12e-46a6-81f0-0478c0172bad","order_by":1,"name":"Gu Jun","email":"","orcid":"","institution":"China University of Geosciences","correspondingAuthor":false,"prefix":"","firstName":"Gu","middleName":"","lastName":"Jun","suffix":""}],"badges":[],"createdAt":"2024-01-12 15:44:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3857471/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3857471/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":49877897,"identity":"15efed0d-9cb3-4030-9e2b-e691b16bcb43","added_by":"auto","created_at":"2024-01-19 14:30:27","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":58906,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart for the Firefly algorithm.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-3857471/v1/4ee579b8e33e8db68ef5024d.jpeg"},{"id":49878663,"identity":"c115d187-5371-4836-9d53-39e5a73ace46","added_by":"auto","created_at":"2024-01-19 14:46:27","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":66501,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePLS algorithm summary\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-3857471/v1/4a62da0dd2090b81a97705ab.jpeg"},{"id":49877899,"identity":"7b202888-3315-43d2-9654-ea15481574f4","added_by":"auto","created_at":"2024-01-19 14:30:27","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":185963,"visible":true,"origin":"","legend":"\u003cp\u003eScores plots for Volume loss Model.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-3857471/v1/8b6971767def7c5461d8f707.jpeg"},{"id":49877900,"identity":"3ddad578-6882-4f3f-b536-aefb6b370079","added_by":"auto","created_at":"2024-01-19 14:30:27","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":57083,"visible":true,"origin":"","legend":"\u003cp\u003ePercent of Variation Explained by each Factor for X-variables\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-3857471/v1/d7313dae021d68cf8cb2ba3d.jpeg"},{"id":49878344,"identity":"7c56e584-4c09-4179-8caf-9e8719185cbb","added_by":"auto","created_at":"2024-01-19 14:38:27","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":13872,"visible":true,"origin":"","legend":"\u003cp\u003ePredicted vs. actual values of (a) PV and (b) AV of the testing dataset for different algorithms (GB: Gradient Boosting, RFR: Random Forest, AB: AdaBoosting, kNN: K-Nearest Neighbor). \u0026nbsp;for low shear rate The rheological information's feedback character type makes the informational index lopsided. For imbalanced datasets, a few scientists have estimated that Tendency Supporting is better than different computations (Brown and Mues, 2012 Absolutely, it shows from the outcomes mentioned in Table 1 that the Point Supporting experience was greater than different assessments. Regularly, it means a lot to fight against extending the instructive record's size so the evaluations' wide capacities can be gathered., With the best degree of materials, PAM/PEI-based mud can be changed to get tremendous rheological properties. In case of a stream misfortune, PAM/BEI-based mud can really and monetarily supplant water-based mud. As displayed in our past assessment (Naganawa, S.; Okatsu 2008), the normal gelling of cross-related PAM/PEI mud will assist with recovering the trouble scattering when mature gel enters the improvement after the fluid goes through the difficulty zone. The inconsistency among ECD and improvement pressure, especially in wells with a slight window between the break slant and the pore pressure slant, presents a test for the entering structure. Recipes for choosing materials with the proper rheological attributes to more readily impact DPE are constrained by the instrument we are creating here. To accept this result, a genuine evaluation was used. Using water-based mud, a well containing the things kept in Table 2 was exhausted to 6,000 feet. the ECD picked utilizing water-based mud and PAM/PEI-based mud. The disclosures make a general well can be created with mud containing 9.5 ppg PAM/PEI. Moreover, the PAM/PEI-based mud's streaming strain was unimportant in contrast with the Wbm's. 7.5% PAM and 1% PEI were utilized to make the cross-related polymer slurry used to figure the ECD. The two mud designs' outright parts and improvements are shown in Table 3.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-3857471/v1/abf16a8fc6f6cca7dccdfe1d.jpeg"},{"id":49877901,"identity":"99a056af-d703-404d-b56f-6ba06e8559a3","added_by":"auto","created_at":"2024-01-19 14:30:27","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":129653,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution function (CDF) of the Sorush field dataset variables.\u003c/p\u003e","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-3857471/v1/cf89045b295281ece7207c21.png"},{"id":49934331,"identity":"7f0d8929-3f22-41a4-ab9d-815ed14c8ca4","added_by":"auto","created_at":"2024-01-21 17:07:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":924829,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3857471/v1/3d6d00a5-91ce-4cd9-ad16-11522a83a8e8.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine learning and Artificial Intelligence Techniques to Expect Drilling Fluid Density , Rate Infiltration and Loss Circulation Anticipation","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eLately, there have been tremendous degrees of progress in the oil and gas industry; The expulsion of cuttings and the energy of placing mud into the borehole are two of the main advantages of the relentless interaction. Indeed, even at low shear rates, penetrating liquids require an expansion in consistency to suspend and ship particles. Li, M.-C.; Wu, Q.; 2018); It is essential to keep in mind that a significantly developed entering mud has the potential to reduce total working costs by 5 to 15%2 (Bloys, B.; Davis 1994), specialists have organized changed getting procedures to find the key parts impacting the oil business (Hornik et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e1989\u003c/span\u003e). All along, it was in 1990 that Arehart (1990) proposed an arrangement for closing the ROP looking at the strategy for counterfeit brain affiliations (ANN). Water-driven power per square inch (HSI), WOB, T, and RPM were the information components in this work. Exhausting exercises require the use of entering fluid, and going with some inadmissible or undesirable choice of this part can emphatically influence communication. Kutassov et al. ( 1988) and Babu et al. ( Taking into account the thermodynamic properties of a wellbore, Kutasov (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; 1993) introduced a preliminary relationship for concluding entering liquid thickness (Kutasov, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1988\u003c/span\u003e;). Effective models for identifying and anticipating significant boundaries are essential for the oil and gas industry. In this way, mechanized thinking methods have truly been used in different areas of the oil and gas industry, as shown under. Supplies (Ghorbani et al., 2017b); oil well wandering (Abdali et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e); harm to the game plan (Mohammadian and Ghorbani, 2015), wellbore disparity (Darvishpour et al., 2019), rheology and filtration (Davoodi et al., 2019, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and development (Ghorbani and Moghadasi, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2014\u003c/span\u003ea; Ghorbani and other, 2014b; et al., 2017a) and 2019 entry fluid). A geomechanical strategy and shear modulus-based check of bundling breakdown (Moradi, H.,2012). A solid and practical method for imagining impairing liquid thickness, the type of invading liquid (Newtonian or non-Newtonian), ROP, WOB, and so on, ought to be viewed as in the made plan (Li et al., 2017; Guo and co., 2021). ). As of late, extraordinary man-made knowledge procedures, for instance, soft reasoning, FL, support vector machine, SVM, genetic computation fake cerebrum association, ANN, have been participated in oil mining, especially in the field of entering fluid planning. A piece of these applications consolidates predicting fluid spread plans in the wellbore annulus [. Mahmoud, A.A., Elkatatny 2020). The organization is prepared to utilize different calculations to limit blunders to propel predisposition and mass. Levenberg-Marquardt (LM), an iterative bend fitting calculation, is one of these calculations. In tackling nonlinear least squares problems, this calculation has demonstrated its uncommon success (Rashidi, S., 2020). The leftover ML to added because of late progressions in calculations and further enlisting working environments (Dimiduk et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; LeCun and Co. 2015). Using ML techniques, mistakes made during directional boring tasks can be derived. The methods planned work by contrasting ongoing information and mishaps that happened before (Gurina et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Penetrating muds are divided into three rules: water-based, manufactured, and oil-based muds. Due to concerns about the environment and the cost, water-based mud is a better and more appealing option for drilling (Salih, A.; 2016 Elshehabi). Diminished dissemination is a typical boring issue, especially in exceptionally penetrable developments, exhausted repositories, and cracked or huge developments (Moosavi, S. R 2019 ); According to Duan, J., Zhao, J., Xiao ), the range of loss of circulation issues ranges from shallow, unreinforced formations to well-reinforced formations that are fractured by drilling mud's hydrostatic loading. Loss of flow episodes, alluded to as the deficiency of penetrating liquids in the development, is known to be one of the most troublesome issues to anticipate or alleviate during the boring stage. The seriousness of the results changes depending upon the reality of the calamity; it could begin with a straightforward loss of penetrating liquid and end with a victory (Whittaker, A \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2011\u003c/span\u003e ).\u003c/p\u003e"},{"header":"2. Modeling lost Circulation","content":"\u003cp\u003eThe examination of data has become a popular topic in recent times due to the vast amounts of information that are stored and available. Using data analysis techniques can help to evaluate and understand the performance of a particular process, and it can help to predict future outcomes. Many companies have used data analysis, especially for fuel plans and supply plans. However, there is a lack of data analysis in drilling, as most drilling data is private and protected by organizations. To address this, researchers have used artificial intelligence techniques to predict issues and select the best course of action before entering the occurrences zone. There are several data mining techniques used in this regard, such as Innocent Bayes, Event-Based, and Brain Relationships. Researchers have also developed new models to focus on volume loss, crude streaming density (ECD), and rate of penetration (ROP) for lost flow events. These new models have been compared to previous models developed by other researchers.\u003c/p\u003e"},{"header":"3. Artificial intelligence Algorithm","content":"\u003cp\u003eVarious computer-based data analyses have been carried out in different areas of the oil and gas industry, including materials, flooding of oil wells, damage to the philosophy, imbalance at the breaking point, rheology and filtration, and development and moving into liquid. Geomechanical and shear modulus-based estimates of packaging breakdown have also been conducted. The readiness model is a model that changes the dependence model based on a chosen computation. The test is conducted to examine the reliability of the used model. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the flowchart for the Firefly algorithm. The following parameters were used as input data: pressure at the standpipe, tank rank 02, input flow, hook load, torque of the rotary table, penetration rate, weight on the bit, and gas rate\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"4. Rheology of drilling and hydraulic fluids","content":"\u003cp\u003eRheology and liquid water power are fundamental variables in penetrating strategies. Rheology is depicted as the assessment of the stream and curving of materials, especially entering liquids, and power through pressure recommends the strategy for overseeing acting of the development of liquids, including mud, in the entering framework, which utilizations lines, valves, and experts for pass entering mud, discard stays, cool woods, and oil up the drill string. The capacity of the siphon, the harmony of the wellbore walls, the financial implications of bundling, and the fluid's capacity to exert pressure on the wellbore walls are all affected by the rheology of the infiltrating fluid. Rheological pieces of entering fluids, for instance, PV, YP, gel strength, and thickness can change concerning temperature. As a depiction, if you increase the HP of fluid, this will have explicit effects. This is especially an expansion in ECD, flood and swab strain, and the likelihood of differential adhering because of additional solids in the liquid. Moreover, the expansion in the consistency of the plastic will prompt a decrease in the entrance rate because of the unfortunate disinfection of the openings (. Yavari, H., Sabah, M 2018 ). The investigation of liquid tension, stream, and speed in the drill table framework is planned in liquid water power. The water-driven elements of penetrating liquid affect the efficiency of different tasks, for example, mud unloading, boring and course, an appropriately requested water-driven strategy can keep away from blockages and guarantee proficient waste expulsion. Consequently, the rheology and hydrodynamics of the mud are vital parts of penetrating procedures that should be considered while arranging and completing boring activities (Elkatatny, S.; Tariq, Z.; 2017).\u003c/p\u003e"},{"header":"5. Rheology and Mechanization Measurement System","content":"\u003cp\u003eThe use of artificial intelligence (AI) techniques is essential for a complete system that can provide the rheological characteristics of different types of drilling fluids. Current devices are examined using AI approaches, but AI prediction methods are still necessary. Pipe viscosity measures the time required to flow a fixed volume of liquid (930 cm3) through the open hole of the pipe. According to Igwilo, K.C.; Okoro, E.E.; and Ohia, the impact of fluid properties on well control and drilling operations requires a different measurement rate for each property. Therefore, AI is considered an alternative research horizon for predicting the rheological properties of mud. The automatic instruction computer coding process is used to determine how far mathematical correlations can go by utilizing the relationships between elements in a set of data to provide an initiation capability. This procedure is evolved based on the factual breakdown of information and calculations to accomplish the learning objective for arrangement or expectation purposes (Agwu, O.E.; Akpabio, 2018).\u003c/p\u003e"},{"header":"6. Support vector machine learning method","content":"\u003cp\u003eEstablishing AI devices is one of the numerical machines that are suitable for providing a sufficient allowance for oil penetration. Because of the different parts of AI devices in boring activities and the exactness of their expectations for field applications, they depend on the investigation of traits from a gathered informational index. This will empower us to foster a satisfactory framework in the boring area in light of relapse examination. De Moor and Suykens (2002)\u003c/p\u003e"},{"header":"7. ANFIS model","content":"\u003cp\u003eTo find this parameter, we first used the ANFIS algorithm to specify the drilling fluid's density. The legitimate straight condition to deduce exhausting fluid thickness is \"y\u0026thinsp;=\u0026thinsp;0.9647x\u0026thinsp;+\u0026thinsp;0.0412\" with the relationship variable of (R2\u0026thinsp;=\u0026thinsp;0.8072). To be even more clear on the entering liquid thickness gauge by ANFIS, we add a groundbreaking selection of the fundamental information. It is gotten from the aphorism \"y\u0026thinsp;=\u0026thinsp;0.4547e0.742x\" and the association coefficient R\u003csup\u003e2\u003c/sup\u003e of 0.8139. Along these lines, a remarkable derivation would be good for working out entering fluid thickness rather than an immediate condition. To better understand the prediction, the model must be developed using power derivation or longitude.\u003c/p\u003e"},{"header":"8. PSO-ANFIS model","content":"\u003cp\u003eA PSO-ANFIS computation was added to represent the improvement of the ANFIS system to give additional exact speculations. For this, it is focal that the agreeable direct condition to expect the thickness of the attacking fluid is \"y\u0026thinsp;=\u0026thinsp;0.9877x\u0026thinsp;+\u0026thinsp;0.0288\" with the relationship part of (R2\u0026thinsp;=\u0026thinsp;0.9225). We add an amazing derivation based on the exploratory data to make the incapacitating fluid thickness measurement by PSO-ANFIS significantly clearer. It is obtained from \"y\u0026thinsp;=\u0026thinsp;0.4719e0.7314x\" and an R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.9068 relationship component. The exceptional derivation would be more effective at imaging the thickness of the entering fluid than the straight condition. Man-made keen limits and changed heading contraptions should be executed to equip engineers with extra data to sort out an exquisite and harmless standard design method ( Davarpanah and Mirshekari, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e"},{"header":"9. ANN model","content":"\u003cp\u003eANN is one of the most widely recognized artificial intelligence techniques, important to address particular designing issues of colossal intricacy that outperform the computational capacity of traditional math and activities (Salehi, S 2016). The construction of the ANN comprises three all-encompassing sorts of layers. The principal concerns the info boundaries. The neurons that total the trade capacities among information sources and results are arranged in the ensuing layer, which is known as the mystery layer. The third sort concerns exits. These layers, with the fitting preparation calculation, uncover the idea of the issue ( Gamal, H.; Abdelaal 2021).\u003c/p\u003e"},{"header":"10. Volume loss model","content":"\u003cp\u003eThe system creation procedure requires a section on the number of latent factors. Score plots help sort the ideal number of latent factors that will be used in the model. Unlike Principal Constituent Analysis (PCA), PLS score plots are calculated to explain the variation in x and y and to maximize the relationship between the x and y variables. Choosing the optimal number of latent factors is a complicated process and requires trial and error until the optimal number of latent factors is reached. Using too many latent factors will lead to over-adaptation of the system, which in turn will reverse the sign of certain variables and make the model unrealistic seen in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. On the other hand, the use of a very small number of latent factors will not explain\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe limit among x and y. Two measures are utilized to wrap up the ideal number of languid parts. The significant differentiations diminish to the mean supporting of the typical extra level of squares (PRESS). The second permits you to control the score plots for the x and y factors, and each dormant part will show a score plot of x versus y. The system ought to remember the inactive variable if the score outline has an example. The latent part ought to be ignored regardless of this, provided that no model is displayed in the score frame (Dong, S., Zhu, H 2011 ). The focal mean of PRESS as a piece of how many lazy variables is portrayed in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. The figure is given out for ten lazy components to investigate the best number of idle parts, believe it or not. Applying the fundamental measure, it is easy to see that having someplace close to two lethargic components will reduce the main mean of PRESS. Regardless, an undefined quantity of inactive variables should be picked. This is where score plots come in. The score plots for six dormant parts are depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Applying the accompanying standard, having different idle parts won't add any huge data to the turn of events, since there are on a very basic level no dormant parts.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe proportion of variation explained by each factor for x-variables and y, respectively, is depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. decipted Scores plots for the ECD model.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"11. Field Data Description","content":"\u003cp\u003eUtilizing man-made consciousness (simulated intelligence), this study expects to foresee constant mud mass (MW) and comparable flowing thickness (ECD) during penetrating activities. By conveying results that are straightforward and trust, this approach desires to support penetrating effectiveness. Due to its abundance of moderate and crucial data, the stretch of a few spots in the degree X3000 and X4200 ft was chosen for the use of the two new models and replicated data. We joined 4,371 records for ECD and 33,588 records for MW achieved by meandering toward the sea gas well A, then wandered toward the sea oil well B, and then floated away from the seaward gas well A. level oil C. This reach was viewed as urgent because of the way that it needed practically no missing information and contained data of unlimited importance. By distinguishing this specific region, the review had the option to guarantee the evaluation's comfort and worth by smoothing out the utilization of assets and time. Facilitators were able to focus on the most significant and concrete information thanks to this procedure, improving the overall quality and authenticity of their results. The rheological properties, like plastic consistency and yield strength, were studied utilizing a rheometer at 48\u0026deg;C and at standard barometrical strain in this review, which was on an exceptionally major level more enormous in picking the thickness of the entering fluid. Besides, a mud counter was used to dissipate the liquid stage and gather the additional solids, and Swamp channel consistency was assessed using a Bog line at room temperature and including strain.\u003c/p\u003e \u003cp\u003eWhat's more, other vitals like the entry rate, mud siphon stream rate, and were recorded to wellspring pressure. The enlightening rundown was divided into a status set (80%) and a test set (20%), with the unpredictable state set at 42, to guarantee that the planning and test sets would remain connected throughout the various runs. Because of how hyper limit blocking was completed, a help set was not viewed as central, so the 80:20 split degree was picked because the standard degree is utilized the most. To anticipate MW and ECD, the initial approach made use of three designs: express Choice Tree (DT), Counterfeit Brain Affiliation (ANN), and Sponsorship Vector Machine (SVM), all of which were modified in Python. The portrayal of the three structures for expecting ECD and MW is depicted as follows: More about this source texture text expected for extra grasping data Equivalent circulating density (ECD) prediction The relationship between ECD-PWD and various elements, including SPP (psi), LSYP, and GPM, has been examined for the suspicion for ECD. ECD-PWD (PCF) revealed the most grounded positive relationship with GPM (0,951728), while various endpoints equivalently uncovered a positive relationship with ECD-PWD (PCF) with values more clear than 0,8.\u003c/p\u003e"},{"header":"12. Mud weight (MW) Prediction","content":"\u003cp\u003eThe relationship between ECD-PWD and several sections, such as SPP (psi), LSYP, and GPM, was considered for ECD doubt. ECD-PWD (PCF) had the most grounded positive relationship with GPM (0.951727), yet unique endpoints with values above 0.8 additionally had a positive relationship with ECD-PWD (PCF).\u003c/p\u003e \u003cp\u003e \u003cb\u003eRheology information\u003c/b\u003e \u003c/p\u003e \u003cp\u003eRheological information was made pondering the particular mud recipes, material aggregate, and test temperature. In this part, the impact of PAM, bentonite, and The liberality of the mud on the rheological properties (like plastic consistency, and clear thickness) is incorporated. We took a gander at different PAM fixations going from 6 to 13 weight percent to perceive what they meant for boring liquid power through pressure and, thus, the same traffic thickness (EPD) depicts the PV got from viscometer assessments of various social affairs of non-crosslinked PAM in refined water. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e depicts the AV values for various PAM obsessions at 1021, 10, and 5 s shear speeds, as they were. The qualities similarly show the impact of temperature, which was significantly more displayed at low shear rates. It was seen that the Accomplice rate went with an unending move in VA and PV. Sensible qualities were acquired with polymer levels of 7.5 to 10% by weight. Past With these convergences of PAM, the viscosities would in general increment thoroughly. The polymer will in general show huge viscosities at this high convergence of 10 to 12.5 weight percent; Notwithstanding this, at high temperatures (200F), a 70\u0026ndash;75% decline in thickness was noticed contrasted with the qualities estimated at surface temperature.\u003c/p\u003e"},{"header":"13. Feat and Accuracy of Predictions","content":"\u003cp\u003eFor other PAM/PEI and mud-added substance structures, intriguing rheological data speculations can be made by utilizing various evaluations. Figure\u0026nbsp;12 secludes the commonplace attributes from the test dataset with the guaranteed credits utilizing the four assessments attempted in this review for high most likely gains of shear rate. It might be estimated that the entire sagacious imperative of point helping was more splendid than another evaluation for plastic and clear consistency. Dismissing this, the presentation of Ada Lift in foreseeing a further degree of plastics and clear consistency defeated the accomplishments of different assessments. Counting a similar framework for the low shear rate information was excruciating. As depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, the expected characteristics and the actual characteristics diverged, and the gauge capacity was substantially reduced. In pursuit to gauge the viability of these calculations, the dataset was recklessly picked for preparation and tests in a degree of 20:80. For both consistency respects, every calculation was run on different events, and quantifiable appraisal was achieved. The ideal outcome for every calculation utilizing the high shear rate data is displayed.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMud Component and Concentrations.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eComponent/Tem\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003eConcentrations\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003elb/bbl\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eWeight %\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePAM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e17\u0026ndash;44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5-12.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePEI wt.% (lb/bbl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9.8-5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.25\u0026ndash;1.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBentonite wt.% (lb/bbl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u0026ndash;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u0026ndash;15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCaustic soda (lb/bbl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0-1.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0-0.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMud weight (ppg)/\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOct-90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2 to 30\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBarite (lb/bbl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u0026ndash;15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFiber (lb/bbl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTemperature (\u003c/p\u003e \u003cp\u003eF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e120\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e160\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e180\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e200\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMud and well Information used for the hydraulic Measurements\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eComponent\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eParameter\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInformation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eMud information\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMud Density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eppg\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWater-based mud\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAV\u0026thinsp;=\u0026thinsp;32, PV\u0026thinsp;=\u0026thinsp;22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003emPa.s\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePAM/PEI-based mud\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAV\u0026thinsp;=\u0026thinsp;27,\u003c/p\u003e \u003cp\u003ePV\u0026thinsp;=\u0026thinsp;26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003emPa.s\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eWell information\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWell Depth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFt\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSurface Hole Diameter\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7 7/8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIn\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMain Hole Diameter\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 5/8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIn\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCasing Depth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7 7/8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIn\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eDrillCollar Information\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eID\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFt\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eID\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 1/4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIn\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6 1/4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIn\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eDrill Pipe Information\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e270\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFt\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTotal Length\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 5/6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIn\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eFlow Rate Information\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eID\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 1/2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIn\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6430\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFt\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTotal Length\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e275\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGPM\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSurface Pressure loss\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePsi\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComposition of PAM/PEI-based Mud and the water-Mud\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eComponent\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePAM/PEI-Based Mud\u003c/p\u003e \u003cp\u003elb/bbl\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWater-Based Mud\u003c/p\u003e \u003cp\u003elb/bbl\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWater\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e312.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e318\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCaustic soda\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLignite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBentonite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMud deflocculant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCalcium Carbonate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e55\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePAM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePEI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBarite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e56,5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"14. Discussion","content":"\u003cp\u003eVarious samples were used to determine the extent of the penetration rate based on different inputs. For the testing stage, 80% of the dataset was used for training, while the remaining part was used. The MLP-ABC model produced an extraordinarily high prediction accuracy for all four sets of samples evaluated, with RMSE values of less than 0.009812 m/h and R\u003csup\u003e2\u003c/sup\u003e values of 1.000 for the test subset, RMSE values of less than 0.006561 m/h and R2 values of 1.000 for the training subset, and RMSE values of less than 0.007411 m/h and R2 values of 1.000 for the total subset. Figure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e displays the predicted and balanced ROP values for each data archive in each subset estimated by the four models. Figure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e shows the accuracy based on R2 for the training, testing, and total datasets. The presented data allows for a comparison of the accuracy of four hybrid machine learning methods. These methods combine the multi-layer perceptron (MLP) algorithm with optimizers such as the firefly algorithm, the gravitational search algorithm, the artificial bee colony algorithm (ABC), and the independent component analysis algorithm optimizers to predict ROP. The advantage of these hybrid models is that they allow optimizers to quickly and cost-effectively adjust the control principles of machine learning models. The hybrid of algorithms includes MLP-FF, MLP-GSA, MLP-ABC, and MLP-ICA.\u003c/p\u003e "},{"header":"Conclusion","content":"\u003cp\u003e\u003cspan\u003e\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e1. The rheology of PAM/PEI-based mud is determined by the centralization of PAM, not by the PEI rate. The most desirable characteristics went from 7 to 10% by weight. Consistency was in a general sense impacted by different materials; regardless, the PAM rate ought to be loosened up to achieve allowed rheology, for high solid things like barite.\u003c/p\u003e\u003cspan\u003e\n \u003cp\u003e2. for the instance of lopsided rheological portrayal information getting together, inclination supporting was more fitting than different calculations, especially k-Closest Neighbor, Arbitrary Timberland, and AdaBoosting\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003e3. Rheological information at low shear rates was difficult, albeit the expectation accomplishments were extremely low; in any case, great expectations were accomplished at low consistency rates where low paces of mud-added substances were utilized. Expanding the size of the dataset ought to build the presentation of the model. In light of the different writing surveys, a significant use of artificial intelligence in boring was introduced, from the send-off of the preparation plan to the expectation of the entrance rate.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003e4. During the introductory information examination, wells with difficulties experienced during boring were uncovered. To assign the uncovered PD, a PC method was introduced.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003e5. While breaking down the boring reports, a rundown of the principal components was ordered, which added to the model: tension at standpipes; tank volume; input stream rate; snare load; force of the rotational table; entrance rate; weight on the piece; gas content\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003e6. A list of the main parameters and who contributed to the model was compiled during the analysis of the drilling reports: tension at standpipes; tank level; input stream rate; snare load; force of the rotational table; infiltration rate; weight on the piece; gas volume.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003e7. Estimating mud losses, ECD, and ROP is much easier with the newer models than with the older ones.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003e8. TFA is a vital component of the ROP model. The ROP model suffers as a result of this. Subsequently, the decision on spout size should be made cautiously.\u003c/p\u003e\n\u003c/span\u003e\n\u003cp\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo funding for this paper, for the APC option we will be in charge of the publication fee\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll datasets generated and analyzed from the various experiments conducted for this work are available. The data was obtained through experiments conducted in the laboratories. The datasets used and analyzed during the current study are available in the main document. The authors, Abdoulaye Seyni Mahamadou and Professor GU JUN can easily share them without any major issues. All data generated or analyzed during this study are included in the published article to provide readers and novices in the field with more explanations and to bring innovation to the field of oil drilling.\u003c/p\u003e\n\u003cp\u003eWe are willing and order the publication of all data without any restrictions to reach the greatest number of readers and the public. It is necessary to explain how the data was obtained for clarity. We worked with Professor Gu Jun of China University of Geosciences Wuhan and together released ideas on the work of strengthening oil wells through our laboratory experiments.\u003c/p\u003e\n\u003cp\u003eAfter the release of the various conclusions presented in the main document, we concluded that the data is self-produced as an innovation publication with the support of the articles we cited in the document. All data was collected according to our laboratory experiments. We declare that Abdoulaye Seyni Mahamadou and Issa Moctar Maimouna, as authors, contributed to the realization of the different experiments to be carried out, and all others contributed to the revision of this document. As students of China University of Geosciences, we share the same laboratory for the realization of our different works. \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors agree with the publication and consent and agree to the Publication\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eLi, M.-C.; Wu, Q.; Song, K.; De Hoop, C.F.; Lee, S.; Qing, Y.; Wu, Y. Cellulose Nanocrystals and Polyanionic Cellulose as Additives in Bentonite Water-Based Drilling Fluids: Rheological Modeling and Filtration Mechanisms. Ind. Eng. Chem. Res. 2016, 55, 133\u0026ndash;143. [CrossRef]\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"2\"\u003e\n\u003cli\u003eBloys B, Davis N, Smolen B, et al. Designing and managing drilling fluid[J]. Oilfield Review, 1994, 6(2): 33-43.\u003c/li\u003e\n\u003cli\u003eHornik, K., Stinchcombe, M., White, H., 1989. Multilayer feedforward networks are universal approximators. Neural Network. 2 (5), 359e366. https://www.cs.cmu. edu/~bhiksha/courses/deeplearning/Fall.2016/notes/Sonia_Hornik.pdf. Hossain, M.E., Al-Majed, A.A., 2015. Fundamentals of Sustainable Drilling Engineering. John Wiley \u0026amp; Sons.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"4\"\u003e\n\u003cli\u003eKutasov, I.M., 1988. Empirical correlation determines downhole mud density. Oil Gas J.. Li, Yuwei,\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"5\"\u003e\n\u003cli\u003eGhorbani, H., Moghadasi, J., 2014. Development of a new comprehensive model for choke performance correlation in Iranian oil wells. Adv. Environ. Biol. 8 (17), 877e882. http://www.aensiweb.net/AENSIWEB/aeb/aeb/September%202014/877-882.pdf.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"6\"\u003e\n\u003cli\u003eGhorbani, H., Moghadasi, J., Mohamadian, N., Mansouri Zadeh, M., Hezarvand Zangeneh, M., Molayi, O., Kamali, A., 2014. Dev. New Comprehen. Model Choke Perform. Correlat. Iran. Gas Condensate Wells 8 (17), 308e313.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"7\"\u003e\n\u003cli\u003eAbdali, M.R., Mohamadian, N., Ghorbani, H., Wood, D.A., 2021. Petroleum well blowouts as a threat to drilling operation and wellbore sustainability: causes, prevention, safety and emergency response. J. Construct. Mater. Special Issue Sustain. Petrol. Eng. ISSN 2652, 3752. https://iconsmat.com.au/wp-content/ uploads/2021/02/si1.1r.pdf.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"8\"\u003e\n\u003cli\u003eDavoodi, S., Sa, A.R., Rukavishnikov, V., Minaev, K., 2020. Insights into application of acorn shell powder in drilling fluid as environmentally friendly additive: filtration and rheology. Int. J. Environ. Sci. Technol. 1e14. https://link.springer. com/article/10.1007/s13762-020-02880-0.\u003c/li\u003e\n\u003cli\u003eDavoodi, S., Sa, A.R., Soleimanian, A., Jahromi, A.F., 2019. Application of a novel acrylamide copolymer containing highly hydrophobic comonomer as filtration control and rheology modifier additive in water-based drilling mud. J. Petrol. Sci. Eng. 180, 747e755. https://doi.org/10.1016/j.petrol.2019.04.069.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"10\"\u003e\n\u003cli\u003eMoradi, H., Bahari, M.H., Bagher, M., Sistani, N., 2012. Drilling rate prediction using an innovative soft computing approach. Sci. Res. Essays 5, 1583e1588. https:// lirias.kuleuven.be/retrieve/289930.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"11\"\u003e\n\u003cli\u003eLi, Yuwei, Jia, Dan, Rui, Zhenhua, Peng, Jiyong, Fu, Chunkai, 2017. Evaluation method of rock brittleness based on statistical constitutive relations for rock damage. J. Petroleum Sci. Eng. 153, 123\u0026ndash;132. http://dx.doi.org/10.1016/j.petrol.2017.03.041.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"12\"\u003e\n\u003cli\u003eGuo, X., Liu, J., Dai, L., Liu, Q., Fang, D., Wei, A., Wang, J., 2021. Friction-wear failure mechanism of tubing strings used in high-pressure, high-temperature and high-yield gas wells. Wear http://dx.doi.org/10.1016/j.wear.2020.203576\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"13\"\u003e\n\u003cli\u003eMahmoud, A.A., Elkatatny, S., Al-AbdulJabbar, A., Moussa, T., Gamal, H., Shehri, D.A.,\u003c/li\u003e\n\u003cli\u003eArtificial neural networks model for prediction of the rate of penetration while horizontally drilling carbonate formations. In: Paper Presented at the 54th US Rock Mechanics/Geomechanics Symposium. https://www.onepetro. org/conference-paper/ARMA-2020-1694.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"14\"\u003e\n\u003cli\u003eRashidi, S., Mohamadian, N., Ghorbani, H., Wood, D.A., Shahbazi, K., Ahmadi Alvar, M., 2020. Shear modulus prediction of embedded pressurize salt layers and pinpointing zones at risk of casing collapse in oil and gas wells. J. Appl.Geophys. https://doi.org/10.1016/j.jappgeo.2020.104205, 104205\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"15\"\u003e\n\u003cli\u003eDimiduk DM, Holm EA, Niezgoda SR (2018) Perspectives on the impact of machine learning, deep learning, and artificial intelligence on materials, processes, and structures engineering. Integr Mater Manuf Innov 7(3):157\u0026ndash;172\u003c/li\u003e\n\u003cli\u003eLeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521(7553):436\u0026ndash;444\u003c/li\u003e\n\u003cli\u003eGurina E, Klyuchnikov N, Zaytsev A, Romanenkova E, Antipova K, Simon I, Makarov V, Koroteev D (2020) Application of machine learning to accidents detection at directional drilling. J Pet SciEng 184:106519\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"18\"\u003e\n\u003cli\u003eSalih A H, Elshehabi T A, Bilgesu H I. Impact of nanomaterials on the rheological and filtration properties of water-based drilling fluids[C]//SPE Eastern regional meeting. SPE, 2016: SPE-184067-MS.\u003c/li\u003e\n\u003cli\u003eMoosavi, S. R.; Wood, D. A.; Ahmadi, M. A.; Choubineh, A.ANN-Based Prediction of Laboratory-Scale Performance of CO2-Foam Flooding for Improving Oil Recovery. Nat. Resour. Res. 2019, 28,1619\u0026minus;1637.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"20\"\u003e\n\u003cli\u003eDuan, J., Zhao, J., Xiao, L., Yang, C., Li, C., 2015. A crop optimization approach based on improved bp neural network pso. In: Advances in Swarm and Computational Intelligence. Springer International Publishingpp, pp. 11e18. https://doi.org/ 10.1007/978-3-319-20469-7_2.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"21\"\u003e\n\u003cli\u003eWhittaker, A. Theory and Applications of Drilling Fluid Hydraulics; Springer: Berlin/Heidelberg, Germany, 2011; ISBN 978-9401088428\u003c/li\u003e\n\u003cli\u003eAl-Dhafeeri A M, Seright R S, Nasr-El-Din H A, et al. High-permeability carbonate zones (Super-K) in Ghawar Field (Saudi Arabia): identified, characterized, and evaluated for gel treatments[C]//SPE International Improved Oil Recovery Conference in Asia Pacific. SPE, 2005: SPE-97542-MS.\u003c/li\u003e\n\u003cli\u003eDrucker, H.; Burges, C.J.; Kaufman, L.; Smola, A.; Vapnik, V. Support Vector Regression Machines. In Advances in Neural Information Processing Systems; The MIT Press: Cambridge, MA, USA, 1996; Volume 9, pp. 155\u0026ndash;161.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"24\"\u003e\n\u003cli\u003eJimmy, D.;Wami, E.; Ogba, M.I. Cuttings Lifting Coefficient Model: A Criteria for Cuttings Lifting and Hole Cleaning Quality of Mud in Drilling Optimization. In Proceedings of the SPE Nigeria Annual International Conference and Exhibition, Lagos, Nigeria, 1\u0026ndash;3 August 2022. [CrossRef]\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"25\"\u003e\n\u003cli\u003eAl-Hameedi, A.T.T., Alkinani, H.H., Dunn-Norman, S., Al-Alwani, M.A., lkhamis, M.M., Al-Bazzaz, W.H., 2019. Application of artificial intelligence in the petroleum industry: Volume loss prediction for naturally fractured formations. In: Soc. Pet. Eng. - SPE/IATMI Asia Pacific Oil Gas Conf. Exhib. 2019, APOG 2019. http://dx.doi.org/10.2118/196243-ms.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"26\"\u003e\n\u003cli\u003eMohamadian, N., Ghorbani, H., Wood, D.A., Hormozi, H.K., 2018. Rheological and filtration characteristics of drilling fluids enhanced by nanoparticles with selected additives: an experimental study. Adv. Geo-Energy Res. 2 (3), 228e236. https://core.ac.uk/download/pdf/229382392.pdf.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"27\"\u003e\n\u003cli\u003eRanaee, E., Ghorbani, H., Keshavarzian, S., Ghazaeipour Abarghoei, P., Riva, M., Inzoli, F., Guadagnini, A., 2021. Analysis of the performance of a crude-oil desalting system based on historical data. Fuel. https://doi.org/10.1016/j.fuel.2020.120046\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"28\"\u003e\n\u003cli\u003eYavari, H., Sabah, M., Khosravanian, R., Wood, D.A., 2018. Application of an adaptive neuro-fuzzy inference system and mathematical rate of penetration models to pre- dicting drilling rate. Iran. J. Oil Gas Sci. Technol. 7, 73e100. https://doi.org/10.22050/IJOGST.2018.83374.1391.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"29\"\u003e\n\u003cli\u003eElkatatny S M, Tariq Z, Mahmoud M A, et al. Optimization of rate of penetration using artificial intelligent techniques[C]//ARMA US Rock Mechanics/Geomechanics Symposium. ARMA, 2017: ARMA-2017-0429.\u003c/li\u003e\n\u003cli\u003eAlsabaa A, Gamal H A, Elkatatny S M, et al. Real-time prediction of rheological properties of all-oil mud using artificial intelligence[C]//ARMA US Rock Mechanics/Geomechanics Symposium. ARMA, 2020: ARMA-2020-1645.\u003c/li\u003e\n\u003cli\u003eHaklidir, Fusan Tut, and Mehmet Haklidir. 2019. \"Geothermal reservoir temperatures prediction with a deep learning model: a case study western Anatolia.\" Stanford Geothermal Workshop. Stanford: Stanford University. Accessed January 18, 2019.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"32\"\u003e\n\u003cli\u003eIgwilo, K.C.; Okoro, E.E.; Ohia, P.N.; Adenubi, S.A.; Okoli, N.; Adebayo, T. Effect of MudWeight on Hole Cleaning During Oil and Gas Drilling Operations Effective Drilling Approach. Open Pet. Eng. J. 2019, 12, 14\u0026ndash;22. [CrossRef]\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"33\"\u003e\n\u003cli\u003eAgwu O E, Akpabio J U, Alabi S B, et al. Artificial intelligence techniques and their applications in drilling fluid engineering: A review[J]. Journal of Petroleum Science and Engineering, 2018, 167: 300-315.\u003c/li\u003e\n\u003cli\u003eSuykens J A K, De Brabanter J, Lukas L, et al. Weighted least squares support vector machines: robustness and sparse approximation[J]. Neurocomputing, 2002, 48(1-4): 85-105.\u003c/li\u003e\n\u003cli\u003eDavarpanah A, Mirshekari B. Effect of formate fluids on the shale stabilization of shale layers[J]. Energy Reports, 2019, 5: 987-992.\u003c/li\u003e\n\u003cli\u003eSalehi, S.; Madani, S.A.; Kiran, R. Characterization of drilling fluids filtration through integrated laboratory ex-periments and CFD modeling. J. Nat. Gas Sci. Eng. 2016, 29, 462\u0026ndash;468. [CrossRef]\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"37\"\u003e\n\u003cli\u003eGamal, H.; Abdelaal, A.; Elkatatny, S. Machine Learning Models for Equivalent Circulating Density Prediction from Drilling Data. ACS Omega 2021, 6, 27430\u0026ndash;27442. [CrossRef]\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"38\"\u003e\n\u003cli\u003eDong, S., Zhu, H., Zhong, S., Shi, K., Liu, Y., 2021. New study on fixed-time synchronization control of delayed inertial memristive neural networks. Appl. Math. Comput. 399, 126035. http://dx.doi.org/10.1016/j.amc.2021.126035.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"39\"\u003e\n\u003cli\u003eBrown I, Mues C. An experimental comparison of classification algorithms for imbalanced credit scoring data sets[J]. Expert systems with applications, 2012, 39(3): 3446-3453.\u003c/li\u003e\n\u003cli\u003eNaganawa, S.; Okatsu, K. Fluctuation of Equivalent Circulating Density in Extended Reach Drilling with Repeated Formation and Erosion of Cuttings Bed. In Proceedings of the IADC/SPE Asia Pacific Drilling Technology Conference and Exhibition, Jakarta, Indonesia, 25\u0026ndash;27 August 2008; p\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Machine learning, Artificial Intelligence, lost circulation, Drilling Fluid, Rate of Penetration, Prediction","lastPublishedDoi":"10.21203/rs.3.rs-3857471/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3857471/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe success of drilling operations depends on several factors, including the drilling properties, and environmental and financial constraints. However, relying solely on lab testing is often insufficient due to limited resources and time. Drilling operations can result in significant costs due to loss of circulation. To address this issue, we propose using AI and machine learning techniques to predict drilling fluid density and prevent circulation loss using an intelligent approach. We conducted scientific evaluations of the rheological properties and mud components of polyacrylamide/polyethyleneimine (PAM) mud. Four distinct ML algorithms (the adaptive neuro-fuzzy inference system, particle swarm optimization-based adaptive neuro-fuzzy inference system, least squares support vector machine with a genetic algorithm, and radial basis function) were used to investigate the rheological qualities of various mud components at different concentrations and test conditions. In the LSSVM-GA model, we found that the linear equation for predicting fluid density was \"y = 1.0041x + 0.0019\", with a correlation coefficient (R2) of 0.9966. The RBF model was used to predict fluid density due to its superior performance over other conventional models. The linear equation for predicting fluid density was \"y = 1.0009x + 0.0034\", with a correlation coefficient (R2) of 0.9999. Based on our experience, we have found that by using an appropriate combination of materials, we can achieve satisfactory rheological properties, thereby avoiding circulation loss incidents\u003c/p\u003e","manuscriptTitle":"Machine learning and Artificial Intelligence Techniques to Expect Drilling Fluid Density , Rate Infiltration and Loss Circulation Anticipation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-19 14:30:22","doi":"10.21203/rs.3.rs-3857471/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"55907088-655c-426a-8cc6-5363a7c34baa","owner":[],"postedDate":"January 19th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-01-21T16:59:11+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-19 14:30:22","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3857471","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3857471","identity":"rs-3857471","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}