Technical system for mud loss analysis and diagnosis in drilling engineering to prevent reservoir damage

Mud loss is the most serious formation damage in oil and gas well drilling engineering and is an unsolved technical problem. To prevent mud loss, it is necessary to accurately understand and identify three key factors of mud loss, including the location of the loss, the time of occurrence, and the severity of the loss. The diagnosis of mud loss is a prerequisite for the proper formulation of mud loss control techniques. It emphasizes the integration of predrilling, drilling, and postanalysis information to describe and characterize loss zones and predict potential loss zones. On the basis of the theory of engineering fuzzy mathematics, we develop a mathematical model for loss probability evaluation that combines logging anomaly features and engineering data to predict the location of losses from drilling mud and develop a loss formation identification method. The study of hydraulic fracture deformation through stress‐sensitive experiments and numerical simulations can predict the deformation and severity of loss channels, which can help optimize the loss of circulating material by adding drilling fluid. Loss pressure models have been developed based on mud loss mechanisms, and loss pressure for hydraulic fracture creation, connection, and extension has been studied to help identify loss mechanisms and types. Mud losses can be identified by unusual engineering characteristics, including sudden changes in drilling times, cuttings, and mud logging. Real‐time logging parameters can be used to monitor the loss process and hence predict the loss trend. The framework of loss diagnosis techniques is established, which helps in successful mud loss controlling.

Mud loss, a problem relevant to geology, rock mechanics, non-Newtonian and multiphase flow mechanics, physical chemistry, material science, systems science, and other courses, remains an engineering problem of fundamental importance that has not yet been solved.[9][10][11] Though lost circulation control depends on engineering experiences, lacking enough theoretical foundation for mud loss.Shafer et al. 12 discussed seven kinds of flow mete that were used to monitor the drilling fluid, providing some advice to choose and using reasonable flowmeters to monitor the lost circulation.Dyke et al. 13 found that the different kinds of lost circulation have different loss curves by comparison by analyzing the well site lost circulation date, and proved that it is possible to distinguish the types of lost circulation by the characteristics of the lost circulation curve.Oliver 14 researched MWD of fracture width by well real-time lost circulation record and provided the creed of choosing drilling fluid and lost circulation material, it is worth learning to solve the practical problems of lost circulation.Beda and Carugo 15 have recorded the mud loss process by precision electromagnetic current material-time mud loss is helpful to identify some characteristics like the location of the loss zone, and the mud loss types can be identified by real-time lost circulation curve initially.Majidi et al. 16 quantitatively analyzed the lost circulation of naturally fractured reservoirs, established the flow model of Herschel-Bulkley drilling fluid flowing in a fracture, and obtained that drilling fluids' rheological characteristics (flow stress, flow index, etc.) influence mud loss.Jin et al. 17 analyzed mud loss parameters acquired from oil fields, used the theory of fuzzy mathematics, and the relationship between lost circulation extent and approaching extent was determined, as well as a model used to identify lost circulation formation and predict predrilling risk was established.The technology has perfect performance in predicting the risk of lost circulation formation.
Sanders et al. 18 detailed the Lost Circulation Assessment and Planning process that has been employed to explore and evaluate specific lost circulation problems and link them to existing products, systems, and services.The integrative preplanning process analyzes offset histories and formation data not only to identify risk zones but also to gather information on the exact fracture and pore size as well as fracture density.Direct measurement of loss zones includes image logging, nuclear magnetic resonance (NMR), and microseismic monitoring.However, image logging and NMR suffer from practical difficulties in finding loss zones.Microseismic monitoring does not work well with narrow single fracture planes, as the width of the fracture is too narrow to be assessed by microseismicity. 19Therefore, an alternative method is needed to fill this gap and locate the loss of circulation zones.Majidi et al. 20 incorporated the effects of formation fluid into the model and compared the results with the models neglecting formation fluids.The comparison indicated that the models neglecting the effects of the formation fluids overestimated the fluid losses.They also concluded that, if the ratio of the viscosity of formation fluid to the rheology of drilling fluid is less than or equal to 0.01 the effects of formation fluids can be neglected.The model also strengthened the fact that the shear thinning of the fluid increases the mud losses while the ultimate loss volume is controlled by the yield stress.The shortcomings of the model were that it was considered a nondeformable fracture with no leak-off from the walls of the fracture which leads to the underestimation of fluid losses.Shahri et al. 21extended the model presented by Majidi et al., 20 by replacing the linear deformation of rock with an exponential deformation function.The resulting equations were numerically solved for both fracture ballooning and breathing phenomenon and the results were compared with the model considering linear deformation of rocks.The results show that the fluid-loss rate drops faster in the case of exponential deformation and the cumulative mud losses are less as compared with the model considering linear deformation.This is because the pressure builds up faster in case of exponential deformation resulting in a lower pressure differential which is responsible for fluid losses.This model tried to capture the reality and to a certain extent it was successful, however, it does not account for the fluid losses due to the permeability of the fracture walls.Tan et al. 22 performed the mechanism of hydraulic fracture propagation and proppants migration by several groups of large-scale true triaxial fracturing tests with an ingenious method of sand adding, and the interaction behavior between vertical hydraulic fracture and bedding plane was discussed.Chen et al. 23 presented a transient wellbore thermal model to estimate the location of mud loss by interpreting distributed temperature measurement which is facilitated by a recently developed drilling microchip technology.The model is to predict the effect of mud loss on the change in circulating mud temperature profile for tubular fluids, which is developed based on heat balance equations with variable local flow rates along the wellbore due to the mud loss.Razavi et al. 24 investigated the effect of fluid rheology on the mud loss prediction for both radial and linear flow types in natural fractures assuming constant fracture width.They have shown how incorporating the effect of leak-off and flow types can affect our prediction of the mud loss volume.Then they presented a theoretical model to simulate the invasion of the drilling fluid into natural fractures.Mathematical modeling was performed for two types of rheology constitutive models which are commonly used for drilling fluids, namely, Bingham Plastic and Herschel-Bulkley. 25an et al. 26 established a numerical model for transversely isotropic layered shale with a transition zone by utilizing the extended finite element method based on the cohesive zone model, and the effects of in situ stress, dip angle, anisotropy, and tensile strength of transition zone, and anisotropy of shale matrix, and injection rate on fracture vertical propagation behavior were investigated in their studies.
Bychina et al. 27 presented a new comprehensive analytical model that can be used for various fluid rheologies and thus, eliminating the need for different models for different fluids.The model also takes into consideration fracture deformation with constant and variable leak-off rates into the formation.Addagalla et al. 28 designed the new phase-transforming loss circulation material to pump easily and achieve thixotropic behavior under downhole conditions, resisting losses in the thief zone before setting it as a rigid plug with high compressive strength.Dong et al. 29 developed a three-dimensional (3D) coupled thermalhydro-chemical model to investigate the drilling fluid invasion process and dynamic responses of gas hydrate reservoirs.This model deals with the fluid-loss properties and flow field characteristics as well as wellformation interactions considering the effect of hydrate dissociation.Li et al. 30 investigated the mud loss in naturally fractured oil layers with a two-phase flow model and a discrete fracture model.They presented that the mud loss rate of the single-phase flow model is higher than that of the two-phase flow model.In the single-phase flow model, the rock matrix permeability cannot be directly used to calculate the mud loss rate, and the influence of oil saturation on water phase permeability must be considered.Besides, for the oilwet rock, the capillary pressure will resist the entrance of drilling fluids.They also found that once the natural fractures are connected to the wellbore, the mud loss rate goes high at the early stage, and it decreases quickly as the loss time goes on.If the natural fractures are not connected to the wellbore, the mud loss rate is very small at the beginning (less than 0.01 m 3 /min), however, the mud loss rate will increase if the drilling time is long enough.And we can judge when the fracture is connected to the wellbore.
Lost circulation not only costs large volumes of valuable drilling fluid but also results in large amounts of nonproductive time, as when circulation occurs the drilling crew cannot continue to perform most of their functions.In addition, if there is a drop in the mud level, this could cause additional drilling problems such as wellbore instability, jammed pipes due to poor cuttings removal, and extensive formation damage due to loss into the producing zone.Mud loss diagnostic requires analysis of the geologic information and neighboring wells' drilling information as much as possible, uniting geology, dynamics, drilling, well logging, and another branch of learning, abstracting the characteristic parameter of geology, from the factor that occurs mud loss, predicting and avoiding the potential loss zone, but also analyzing the abnormal engineering specifications and phenomenon real-time during mud loss, quantitative analyzes, and diagnoses and identifies the types of mud loss, predicted the loss tendency, declined the loss severity, scientifically optimized the technology of lost circulation control.

| PREDICTING POTENTIAL LOSS ZONE
A loss zone generally possesses a variety of geologic features that can cause mud loss, which is inevitable under certain drilling techniques or certain drilling conditions, a specific form of which has not yet been explored and has not yet resulted in a well loss.Analyzing the geologic information of neighboring wells, the drill hole information, and the excellent log information, the anomalous hole log data can unambiguously reflect the characteristics of mud loss and predict the potential loss zones in the same block.It directs the primordial well to prevent or avoid mud loss.

| The conventional log responses to mud loss
Regular logging methods to identify loss regions include resistive logging, acoustic logging, and kernel logging.
The conventional logarithmic response concerning the fracture-pore reservoir is characterized by low gammaray values.The gamma-ray values are highly intense when the large break is followed by a large aperture filled with argillaceous fraction; the logarithm of the dual is low, with a positive separation; the density is low, neutrons are abundant, and the acoustic travel time is large; the natural gamma-ray values are low and moderate for narrow regular pore spaces, limestone caves, and the difference in values for binary transverse depths is less pronounced; the neutron and density acoustic logarithmic corrections are large.For cavities with large cavity diameters, the typical response of the conventional logarithm is a large amplitude positive separation for binary transverse depth values; the amplitude causes severe glitches and the acoustic travel time becomes significantly longer; natural gamma-ray values are strong, but there is some disagreement as to whether or not the limestone caves were filled in; the neutron density becomes abnormally large; the density is lower.
In an oil well with a depth of 6518-6541 m in the oil field, the value of the dual lateral resistivity log curve declines from 317.0 Ω m to around 1.0 Ω m obviously; the acoustic travel time value becomes large obviously, changing from 54 to 84 µs/ft; the change of the natural gamma-ray value is small.The drilling time log declined in 1-5 min/m, the characteristic is all the characteristics of drilling karst cave.The process of force drilling occurs to travel empty, and the relief section is a depth of 6518.00-6541.00m.The test illustrates that this section is a limestone cave, and the limestone cave has not been filled (Figure 1).It proves that the conventional log can identify the loss zone.

| Characteristics of imagery log response
The imagery log of the borehole wall can describe directly and carefully the complex pore configuration, like, fracture, dissolved pore, limestone cave, stratification, borehole wall sloughing, and so on.The formation information from the date of the imagery log is reflected by the color and form of the imagery picture.The picture color can be separated into three kinds light, dark colors, and variegated colors, the lithological characters, fractures, and others can be identified by the picture colors.
(1) Fractures identified by imagery log The imagery log can provide a full hole scanning image, which can provide fractures' angle of depression and angle of orientation, and also distinguishes the gaping fissure and closed fracture, natural fracture, and induced fracture.The response of fractures on the date of the imagery log is that the scope of the pingerproof echo waves in fracture-cave is small, the interval travel time is long; the dimensions of fracture and cave can be reflected by the abnormal area in the amplitude picture, the parameter of poid reflect the occurrence of fractures; high-filtering-flow-fracture display as a darkcolored poid in the formation micro imager (FMI) pictures (Fullbore formation microimage).
(2) The cavern identified by the imagery log The limestone cave is the result of the formation of water attacking the block mass.The existence of the limestone caves must result in the difference in the reflected situation between limestone caves and the block mass around the limestone cave.The existence of the limestone caves also adds the interval travel The empty zones' log responses to mud loss.AC, acoustictime; CAL, caliper logging; DRTM, drilling time; GR, gamma ray logging.
time, so it is easy to be identified.The color of the limestone cave in the FMI picture is dark, Figure 2 is like a small bubble rock, mostly like the separated star point or bead-like (Figure 2).

| AHP steps
The application of AHP can be divided into the following five steps: (1) Establishing a hierarchy First, the subordination of each component of the problem is analyzed, and the structure model of AHP is established.In this model, the elements of the same layer are dominated by the corresponding elements of the upper layer, and each element also dominates the elements of the next layer.The number of elements in the next layer dominated by each layer should not exceed 9, otherwise, it will be difficult to compare and judge the following pairs.
(2) Constructing a judgment matrix First, compare the importance of factors in the same structural layer in pairs under the unified standard.For example, the important judgment values of factors i and j can be expressed as a ij .The comparison results of each element will be expressed in the form of the matrix, which is the judgment matrix.(1) Any judgment matrix satisfies The value of the judgment matrix directly reflects the understanding of the relative importance of each element.Generally, a 1-9 scale is used to assign importance.The scale and its meaning are shown in Table 1. (

3) Weight allocation
The weight is calculated by the square root method.
Modify the number of indicators to n, and the detailed calculation steps are as follows: ① First calculate the geometric mean value of all elements in each line of the judgment matrix to obtain the vector where ② Normalize the vector M to obtain the relative weight vector where .

③ Consistency inspection
To evaluate the effectiveness of the above ranking, it is necessary to check the consistency of the evaluation results obtained from the judgment matrix.This parameter can be recorded as CR, when | 1341 CR < 0.1, it is considered to meet the consistency; when CR ≥ 0.1, the consistency condition is not met, and the judgment matrix shall be modified until the weight obtained meets the consistency condition.The calculation formula of CR is where RI is the average random consistency index and CI is the coefficient of consistency, which is related to the matrix order n and the maximum eigenvalue.
The calculation formula of the consistency coefficient is where λ max is the maximum eigenvalue of the judgment matrix; n is the order of the judgment matrix (Table 2).The weight of the individual factor subfactors can be calculated according to the above method.Because each evaluation factor is affected by subfactors, the relative weight of the corresponding subfactors is calculated by the weight of each subactor.

| Loss layer evaluation system
The complexity and randomness of porous media would lead all factors of the objective geological factors about lost circulation and the seismic attribute which is sensitive to judge the porous media to layered, using the fuzzy mathematics index membership degree of characterization methods, establishing the subordinating degree function with the easily quantitative indicators, using the subjective experience judgment method, expert-investigation method (1-9 scaling), AHP to confirm the weight of each factor using the method of weighting synthesis to obtain the evaluating value of lost circulation.Using the engineering fuzzy mathematics theory, establishing the mathematical model of evaluating the probability of lost circulation, combining the lithological properties, geological factors, neighbor wells' information that influence the lost circulation, and the seismic attribute which is sensitive to the porous media, establishing the subordinating degree function (Figure 3), to use the nine scaling and the method root means square to calculate the weight and distribute, having the comprehensive assessment about the probability of formation loss.This approach brings together the goodness of various other methods for predicting loss cycles into a comprehensive analytical and quantitative solution, providing a new way to identify potential loss regions.

| Model calculation (1) Basic principle
In case of partial loss during drilling, the drilling fluid returning from the wellbore annulus is divided into two parts: one part of the drilling fluid leaks into the formation, and the other part continues to return to the surface from the annulus above the loss layer.In the well section below the loss horizon, the upward flow back in the wellbore before and after the loss of drilling fluid is the same, and its pressure loss is also the same.The drilling fluid flow in the well section above the loss layer is reduced compared with the upward return flow before the loss, the annulus pressure loss above the loss layer is also reduced, and the riser pressure is also reduced.Therefore, the location of the loss layer can be determined by calculating the change of the annulus pressure loss before and after the loss: where ΔP s is the standpipe pressure change before and after lost circulation, MPa; P a and P a ′ is the annulus pressure loss of the well section above the lost circulation layer before and after lost circulation, MPa.The reason why the riser pressure changes due to the loss of drilling fluid is that the annulus pressure consumption of the drilling fluid in the well section above the loss layer changes due to the change of the upward return flow.The change range of standpipe pressure can indirectly reflect the magnitude of loss and the approximate location of the loss layer.The larger the loss of drilling fluid, the greater the change of annulus pressure consumption, and the more significant the decrease of riser pressure; the deeper the location of the loss layer, the greater the pressure drop of the riser.If the loss occurs, accurately measure the return flow of drilling fluid at the ( ) where V ai is the flow rate of drilling fluid in annulus outside the drill string of section i, m/s; Q 1 is the pumping displacement of drilling fluid, L/s; D is the diameter of wellbore, cm; d i is the outer diameter of section i drill string, cm.Flow pattern index n ai and consistency coefficient K ai of drilling fluid in the annulus ai n 300 ai (10)   where n ai is the flow pattern index of drilling fluid in the outer annulus of the ith section of the drill string; K ai is the consistency coefficient of drilling fluid in the outer annulus of the ith section of the drill string; Φ 600 and Φ 300 are the corresponding readings of the fan rotational viscometer when the speed is 600 and 300 r/ min, respectively.Effective viscosity of drilling fluid in the casing annulus where ρ is the density of drilling fluid, g/cm 3 .Reynolds number of drilling fluid in casing annulus Re ai and flow pattern discrimination where Re ai is the Reynolds number of drilling fluid outside the drill string of section ith; Friction coefficient f ai outside the drill string of section ith Laminar: ai ai (13)   Turbulence: ( ) where The formula for calculating the annular pressure loss of oil casing is When Re n < 3470 − 1370 ai ai , the flow pattern of drilling fluid is laminar: When Re n > 4270 − 1370 ai ai , the flow pattern of drilling fluid is turbulence: .
where P ai is the pressure loss outside the drill string of section ith, MPa.
(3) Computing method Determination of the location of the loss zone when the drilling fluid is lost without loss and has returned.For the convenience of explanation, if the loss layer is located at the well depth of the first section of the drill pipe, then ∆ ( ) ∆ ( ) where i = 1, Q 1 is the displacement the of drilling fluid pump, L/s; Q 2 is the up flow of drilling fluid after loss, L/s.The circulating loss pressure above the lost circulation layer before the loss of circulation is  Formula ( 22) can be simplified as Circulating pressure loss above the lost circulation layer after loss of circulation where For simultaneous formulas ( 23) and ( 24), the change of circulating pressure loss before and after the loss of circulation is ∆ ( ) According to formula (25), the first approximation value of the well depth of the loss zone can be calculated H loss1 : ∆ ( ) Check whether the first approximation value of the well depth of the loss layer verified and calculated can meet the condition:

2
, to determine whether to carry out the second approximation calculation and whether the circular approximation calculation can meet the condition:

2
, then cut-off.It is finally calculated that the H lossi meeting the conditions at this time are the well depth of the real loss layer, which is generally ) where B and B' are the sum of the annular pressure loss of the jth section drill pipe length L j and the above sections of drilling tools before and after the drilling fluid loss Substitute f ′ ai , Q′ p and, respectively, for B, and to get B′: open or expand to become leaky fractures, which aggravates the degree of drilling fluid loss, as shown in Figure 4. 31 The drilling fluid leaks into the fractures connected with the wellbore, and at the same time, the fluid column pressure of the wellbore is transferred to the fracture wall.The fractures are expanded and deformed by the wellbore fluid column pressure to overcome the minimum horizontal principal stress of the formation, resulting in the expansion of the loss channel and further aggravation of the loss.This section evaluates the stress sensitivity and width change of fractures in fractured vuggy carbonate rocks through laboratory experiments, depicts the change of microfracture width of fractured vuggy rock samples under variable confining pressure and the impact of pore development on fracture deformation, and evaluates the width change of formation microfractures.
The fracture width cannot be directly and accurately measured in the experiment because it is small.You can try to indirectly reflect the change in the fracture width by measuring the permeability change of the fracture rock sample.Therefore, based on Darcy's law and the parallel plate seepage theory, the measured permeability value of the fracture rock sample is converted into the fracture width according to the formula, avoiding the error caused by the direct measurement of the microfracture width.According to Parsons' description of the whole rock matrix and fracture system by combining Darcy's law with the theoretical model of parallel plates, 32 the total flow formula characterizing fluid in porous media is where K fr is the total permeability of fractured and unbroken rock systems, 10 −3 µm 2 ; W is the fracture width, µm; D is the fracture spacing, µm; K r is the permeability of the unbroken rock matrix, 10 −3 µm 2 ; Α is the angle between the pressure gradient axis and fracture surface, °.
F I G U R E 4 Deformation and propagation of wellbore fractures induced by pressure fluctuation.
ZHANG ET AL.
| 1345 The carbonate rock matrix is relatively dense and has low permeability, so the matrix permeability in formula (24) can be ignored when considering the relationship between fracture rock sample permeability and fracture width.When measuring permeability, it is generally considered that the included angle between the pressure gradient axis of fluid flow and the fracture wall is 0°, which is considered as parallel fracture wall flow.Therefore, formula (30) can be simplified as According to formula (25), the fracture width value of a rock sample can be calculated by using the permeability value of the fractured rock sample, which can be expressed by the following formula: where K f is the permeability of the rock, 10 −3 µm 2 ; W ji is the fracture width, µm; D is the fracture spacing, µm.
The experimental device is used to measure the permeability value of fracture rock samples corresponding to different effective stress points of confining pressure.Through the theoretical formula proposed by Parsons, the permeability value of fracture-cave rock samples measured in the experiment is converted into the fracture width value of rock samples.This research method overcomes the disadvantage of not being able to directly measure the changing fracture width of rock samples, and avoids the error caused by visual measurement.The experiment sets different confining pressure values to simulate the effective stress changes caused by wellbore fluid column pressure fluctuations in drilling engineering, which leads to stress sensitivity of reservoir rock fractures and dynamic changes in fracture width.Assuming that the in situ effective stress is 30 MPa, which is the initial stress state of the formation fracture, the corresponding fracture width can be assumed to be the initial static width of the fracture.By changing the confining pressure value to simulate the wellbore fluid column pressure disturbance in drilling engineering, the effective stress is reduced, resulting in changes in the reservoir rock fracture width.
According to the above analysis of drilling-before date of geology, logging, and so on, defining the situation of fracture development and the date of crustal stress parameters, providing the stress sensitivity of formation rock fracture, as is shown in Figure 5. On the basis of the multiscale properties of formation rock fracture, by carrying out the stress-sensitive experiment about formation samples, we could analyze the dynamic width changes of formation microfractures while drilling.Considering the formation microfractures cannot represent the width change of large macrofractures in the formation, so using the geological parameters accessed, carrying out the numerical simulation study about the width changes of the large fracture, to predict the width change of large fracture.

| Large-scale fracture deformation simulation
Due to the limitation of the scale of instruments and equipment, the dynamic width change of fractures under conventional indoor experiment simulation engineering operation only stays on the standard core column size, which is very different from the actual formation fracture size.Therefore, it is necessary to apply numerical simulation to study the dynamic width change of fractures under the influence of large-scale fractures and cavern properties.The finite element method of elastic fracture mechanics is applied to simulate the change of fracture width with the fluctuation of wellbore fluid column pressure, taking into account the influence of karst cave development, cave diameter property, and fracture length on fracture deformation.The purpose is to use the numerical simulation method to characterize the changed behavior of fracture width under the influence of wellbore pressure difference, karst cave property, and fracture length in the fractured cave rock mass, to reasonably optimize the size to provide a theoretical basis for loss control scientifically.(1) Joint hole combination finite element solid model Figure 6 shows the combined entity model of the hole fracture system.The vertical fractures on the wellbore are parallel to the direction of the vertical wellbore.In the process of drilling, the force around the wellbore is the wellbore fluid column pressure P w , the two horizontal principal stresses σ H and σ h and the vertical principal stress σ v , the formation pore pressure P 0 , and the fracture tip critical strength factor K I .The rock mechanics parameters selected for the model can be obtained from the rock mechanics property experiment, geological characteristics, logging, and well-testing data. (

2) Establishing a mechanical model
Given the model assumptions and elastic mechanics theory, the fracture cavity rock mass model (Figure 7a) in this section is considered a plane strain problem, so the mechanical model established by the model can also be regarded as a finite element mechanical model in a 2D plane.According to the symmetry principle of the wellbore model, considering the simplicity and straightness of the calculation, 1/4 part of the wellbore model is selected (Figure 7b).AB arc segment represents 1/4 of the wellbore, the BC segment represents fractures, and the CD segment represents caves, the fracture cavity system is connected with the wellbore, so it is assumed that the pressure exerted on the fracture BC section and the cavity CD section is the wellbore fluid column pressure P, and DE and GA are both symmetrical constrained sections.The EF section applies the maximum effective horizontal principal stress P 1 , which is parallel to the fracture extension direction.The minimum effective horizontal principal stress P 2 is applied to the FG section.Maximum effective horizontal principal stress: Minimum effective horizontal principal stress: According to this, the maximum horizontal principal stress is σ H ' is 75 MPa, minimum horizontal principal stress σ h ' is 50 MPa, and different wellbore differential pressures can be set by changing wellbore fluid column pressure P The deformation of fracture width at borehole wall under different drilling positive pressure difference is simulated.| 1347 (4) Analysis of model results Figures 8 and 9, respectively, show the simulation of the fracture and fluid pressure distribution of the fissured sample and the situation along with the fracture in the lab.It also can stimulate each kind of complex formation situation by changing the fissured sample.On the basis of the numerical simulation study about the dynamic width changes of the large fracture, obtained from the prediction results of numerical simulation, the change of fracture width along with the influence of borehole pressure, fracture length, the development of limestone caves, and other factors increase greatly (Figure 10).A given specific engineering parameter can simulate to predict the range of the fracture width change, thus the lost circulation material which matches the fracture dynamic width can be added while drilling, and sealing the dynamic fracture because of stress sensitivity in time, forbidding the happen of the diffusible lost circulation.This method has been applied in the well site, reducing the happen of lost circulation obviously, reducing the degree of borehole wall sloughing at the same time, getting a good anticipated effect.

| TECHNOLOGIES FOR DIAGNOSING MUD LOSS
While drilling the potential loss zone where fractures and caves are developed, using the technology of good logging to monitor the parameter of drilling engineering, drilling fluid parameter, gas parameter, formation pressure monitor parameter, and drilling-rig-shakeparameter real-time, finding the change of the related parameter because of lost circulation in time, diagnosing the lost circulation rapidly, is hopeful to take countermeasures to control the lost circulation in time, reducing the degree of lost circulation.

| Drilling time log response to mud loss
The drill time log can judge and monitor the lost circulation by drawing the drilling time log, and the drilling time date will be combined by using the plane rectangular coordinate method according to a certain proportion, comparing the order of good depth.The drilling time log can identify the drilling fluid-no-out or the loss-large-fracture-cave when drilling in the fracture, the interval where the caves are developed, the drilling time has a sharp drop, and the  blowdown phenomenon.Through the compound logging equipment or the geodata auto-recording the drilling time at the well site, according to the abrupt change of the good depth and the abnormal variation of drilling displayed in the equipment, the fractures, and caves under the earth can be identified clearly.The large fractures or the limestone cave can result in the change of well depth at some meters, creating the blowdown phenomenon, however, the corresponding drilling time is litter or close to zero.The variation of the drilling time curve can judge and identify the mud loss, which is helpful to identify the loss zone with fractured-cavity development and control the mud loss.Because the formation contains big fractures, a pore-cave network system, leads to a sharp drop in drilling time, and then the loss blowdown zone can be identified according to the drilling time curve (Figure 11).
For permeable loss by microfracture and it is not obviously in the drilling time curve.Without observing the fluid level change of the drilling fluid pool, the change of gas potentiometric hydrocarbon resistivity value can be monitored, and the microfractures and the micropores contain a definite hydrocarbon compound or the nonhydrocarbon, with the drilling fluid in the wellbore leaking into the pore space, the hydrocarbon compound or the nonhydrocarbon existing in the pore space return to the ground following drilling fluid, making the gas value of the hydrocarbon compound or the nonhydrocarbon monitored by the logging equipment has an abrupt increasing.For example, a well in the oilfield happened mud loss when drilling to the depth of 6257.0 m at 8:10, losing 3.0 m 3 at all until 9:30, equivalent to the average loss rate of 2.03 m 3 /h, it is drilling to 6269.4 m at 21:00, the loss drilling fluid is 77.0 m 3 , equivalent to the average loss rate of 6.70 m 3 /h.When the compound logging equipment monitored drilling to 6256 m, the drilling time did not change, but the hydrocarbon content of the returning fluid increased by 0.066 from 0.03 at 6225.0 m and reduced by 0.038 when drilling at 6258.0 m, in combination with the level change of drilling fluid pool the permeable loss can be identified, and the corresponding treatment measures can be advanced: if the loss rate increase, it should be pulled out and quiescence to seal, if the loss rate is no changing or changing little, continue to drill.

| Sample-log-monitor
Through the sample log, the stratigraphic sequence and lithological character can be controlled, the situation of the formation containing oil, gas, and water can be understood preliminarily, and the underground situation can be acquired in time.According to the cuttings taken by the returning fluid, the rough opening size of the loss passageway can be realized indirectly, it is helpful to optimize lost circulation material and control the lost circulation targeted.The fractures in carbonate rock almost have been filled by other minerals, including calcspar, bitter spar, anhydrous gypsum, and another carbonate mineral.Generally, the more epigenetic mineral-filled in the cutting, it indicates that the fracture-cave in the rock stratum is more developed.So selecting the epigenetic mineral in the returning cutting, and calculating the secondary mineral percentage of the cuttings, to check the interval where the fracture-pore developed.The secondary mineral filled in the fractures of carbonate rock can be divided into the idiomorphic crystal and nonidiomorphic crystal according to the crystal morphology.
The idiomorphic crystal generally develops along with the formation of a cavern or the sides of fracture to the freedom space, the surface of the crystal has a certain geometric polyhedron shape, generally clear or translucent, and many become crystal clusters because of clumped together.The nonidiomorphic crystal filled the total space of fracture and pore, the surface of the crystal does not have a certain geometric polyhedron shape, generally clear or translucent, like, calcspar, quartz, and another mineral.Thus it can be seen, that the content of the idiomorphic crystal is high, which indicates the development degree of the fracture-pore is good; the size of the grain diameter of the idiomorphic crystal, reflects the size of the fracture-pore to a certain extent, and the size of the crystal, indicates the bigger of the fracture-pore space containing the crystal, it is easier to induce lost circulation of the development of fracture-pore, so, through the sample log observing the content of the idiomorphic crystal the nature of the development of the fracture-pore which is in the formation drilled can be acquired indirectly, it is helpful to predict and make definite the loss zone, and provide the foundation to optimize lost circulation material.When the lost circulation is serious, the bottom hole cutting cannot return to the ground, and the sample log has to end, at this time it should monitor and diagnose the lost circulation by mud log.

| Mud log monitoring
The drilling fluid parameters contain the flow in and out, density, temperature, specific conductance, the drilling fluid volume, and so on.change of the of the drilling fluid reflects directly the exercise of the downhole fluid and the balance situation between the wellbore fluid column pressure and the formation pressure, paying attention to the abnormal changes in the drilling fluid parameters, can avoid major accidents, like, lost circulation, blowing, and so on.The lost circulation is different during different well drilling processes, but the reduction of the drilling fluid is a direct phenomenon, so the happening of lost circulation and the process of it can be diagnosed in time according to the parameters of the drilling fluid log.

| Predicting mud loss while drilling
During the process of faster penetration, because of the drilling fluid filtration, and building up the filter cake to brace the wellbore face, the consumption of the drilling fluid is large, it needs supply the drilling fluid in time, the change of the drilling fluids' volume is big, so mud loss cannot be judged by fluid level change of the drilling fluid pool only, it should be judged by comprehensive analyzing.Through the drilling time and the returning cutting the lithological character can be understood generally if the permeability of the drilling formation is good, if the pump pressure fluctuates, the lost circulation will be created, and the fluid level change of the drilling fluid pool, the rate of in and out, pump pressure should be observed consanguineously at this time, if the pump pressure has a drop drilling the process of faster drilling, the export delivery rate of the drilling fluid reduces, the fluid level of the drilling fluid pool declines, lost circulation is likely to happen.

| Predicting mud loss during tripping
Mud loss is mainly judged by observing the situation of cement grouting during the trip, after stripping 3-5 drill pipes generally, if the volume of insertion is bigger than the volume of the stripped drill pipe or more, it is possible to induce lost circulation.It is likely to induce lost circulation because of the whipping stress during the process of going down.Monitoring the rate of going down, can prevent the generating whipping stress from becoming so big to fracture the formation.It can judge whether lost circulation happens according to the amount of returning drilling fluid during going down.After tripping in 3-5 drill pipes, if the volume of the returning drilling fluid is smaller than the volume of the tripping drilling rig even with no returning drilling fluid it is possible to induce lost circulation.The conductivity sensor or the temperature sensor installed in the export of the drilling fluid can monitor whether the drilling fluid is returning during going down, the specific conductance, and whether the temperature has changed when the drilling fluid returns.If there is some drilling fluid returning during going down, the specific conductance increases from zero, and the change is obvious, if there is no drilling fluid returning, the specific conductance remains unchanged, thus the specific conductance can be used to judge the lost circulation.The temperature judges whether there is drilling fluid returning, influenced by the external environment seriously, for example, the temperature of the returning drilling fluid in winter is largely different from the environment temperature, the abnormal change of the temperature is easy to be judged, however, the range of temperature is littler at summer, the abnormal change is not obvious.There is still some difference between the fluid level of the drilling fluid pool and the in and out rate of flow of different kinds of lost circulation.Through monitoring the change in the fluid level of the drilling fluid pool and the in and out rate of flow at the wellhead, and then the characteristics of lost circulation and fracture should be analyzed (Figure 12 and Table 3).

| Real-time mud loss diagnosis by synthetic mud logging technology
Compound logging devices are used to log while drilling, directly monitors numerous parameters such as drilling, drilling fluid, gas, formation, and so forth, continuously monitor the entire drilling process, continuously monitor drilling accidents, and quantitatively analyze and judge.During the drilling, the compound logging equipment can acquire several parameters, like, drilling time, drilling stress, hook load, SPP, table rotation speed, rotary table torque, hydrodynamic force, and the drilling fluid capability (density, temperature, and specific conductance).Parameters that should have anomalous changes response to any kind of loss of circulation before the loss of circulation are roughly summarized in Table 4.

| PREDICTING LOSS PRESSURE
The corresponding loss pressure is also different due to the different mechanisms of loss circulation.By monitoring the ground pressure values, such as SPP, slush pump stress, and so forth, to realize the changes in the bottom hole pressure, we can indirectly obtain the cause of the loss circulation and analyze the loss pressure to diagnose the type of mud loss.

| Fractured mud loss
Artificial-induced fracture is commonly considered to be the loss channel at this time, and the value of the loss pressure is considered to be close to the formation rock break pressure.If the fracture pressure of the strata becomes larger, the value of the loss pressure also becomes larger and it is more difficult to induce the loss circulation and vice versa.For perfect strata where cracks are not developed, the loss pressure is the sum of the circumferential stress of the bore and the tensile strength of the rock.The wellbore fluid column pressure when the rocks are destroyed in tension (formation breakdown pressure) If the mud cakes formed near the wellbore face are considered, the filter intensity σ t should be considered, and the loss pressure is The point of formation rupture pressure reflects the wellbore fluid column pressure overcoming the extended formation rock strength, leading to rock rupture, formation of new fractures, and loss of circulation.The fracture expansion stress P f the point at which the stress flattens out, represents the pressure responsible for the propagation of the new fracture deep into the formation.After the fracture is induced, the drilling fluid is injected into the fracture and then the column pressure of the wellbore fluid is passed into the fracture.The fracture propagation pressure is reached when the transverse pressure is larger than the minimum horizontal principal stress on the rock, and the fracture generally propagates from the borehole to the formation.If the filtration strength is taken into account before considering the existing mud cake, then the filtration strength should be first conquered and the structure of the mud cake destroyed, and the corresponding curve should rise on this basis.

| Mud loss caused by fracture propagation
For the fracture propagation lost circulation, there is a critical fracture width.When the fracture width is bigger than the critical fracture, the drilling fluid filtration rate increases remarkably, changing the lost circulation.Assuming the fracture deformation accords with the relationship of power function, the flow of the drilling fluid in the fracture accords with the cubed law, and the relationship between fracture width and actual stress can be concluded, which means: In formula (37) w is the dynamic fracture width, mm; who is the fracture width as the wellbore stress is equal to the formation pressure; σ is the effective direct stress which is perpendicular to the fracture plane, MPa; σ 0 is the effective direct stress which is perpendicular to the fracture plane when the wellbore stress is equal to the formation pressure, MPa; A and are the undetermined coefficient.For example, the simple vertical-induced fracture, ignoring the accumulation of stress around the wellbore, can be deduced In formula (38), P f is the effective wellbore fluid column, MPa.
Formulas (37) and (38) are simultaneous, and the relationship between the dynamic fracture width and the drilling fluid volume stress can be concluded.
According to formula (39), with the wellbore effective fluid volume stress increasing, the fracture width increases also.When the fracture width is larger than the critical fracture width, the drilling fluid loss and loss pressure are calculated as the fracture width becomes close to the critical fracture width.
In formula (40), p l 2 is the fracture propagation pressure, MPa; w c is the critical fracture width, mm.
As the bottom hole pressure increases, when it reaches the smallest horizon major stress which the fracture surface can stand, the wellbore rack fracture opens, and the loss increases abruptly, with the dynamic wellbore fluid volume stress transmitting to the sharp fracture, leading to fracture propagation, the loss pressure is smaller than the formation fracture break down pressure at this time, and it is the fracture propagation pressure.If one takes into account the strength of the mud cake at the bore, the loss pressure should be (41)

| Mud loss caused by large-scale fracture-caved
Because the dimensions of the lost channel are too huge, the drilling fluid does not need positive differential pressure or because the density of the working fluid is bigger than the density of the formation fluid and then leading to displacement and lost circulation, a large amount of drilling fluid losses into the formation at a large rate, leading to the sharp drop of the bottom hole pressure (Figure 13).At this point, the missing circulation only needs to overcome the formation of fluid pressure and the deterioration of the flow in the missing channel.For permeable mud loss, the loss pressure is close to the formation pore pressure, with a similar degree of dependence on the permeability of the formation and filter cakes.The formation and filter cakes are better permeable and the loss pressure is close to the formation pore pressure.As a result, the loss pressure increases with the formation of pore pressure, and the growth of the pore pressure is minor as the pore pressure decreases.The loss pressure is slightly larger than the pore pressure because it has to overcome the filter cake strength as the fluid flows through the pore.The natural loss pressure is equal to the pore pressure only if the pore dimension of the loss zone is so large that the filter cake is suitably permeable.

| FRAMEWORK OF MUD LOSS SYNTHETIC DIAGNOSTIC
Mud loss synthetic diagnostic technology should make full use of the data of geology and engineering, the loss time should be monitored in detail by drilling through the potential leak layers, the lost circulation situations of each step have been diagnostically analyzed, and the frame of lost circulation synthetic diagnostic technology has been established preliminarily (Figure 14), optimizing the lost circulation control technology.Incubation period: realizing the characteristics of the drilling formation, analyzing the date of neighbor wells, including the date of drilling, logging, mud logging, acquiring the characteristics of drilling formation, including the developing situation of the fracture and pore-cave belonging to the drilling formation, the formation breaks down pressure and so on, determining the potential loss zone, the mechanism of lost circulation and the induced parameter.Induced period: relying on the synthetic log technology, combining the geological data, through analyzing the change of the mud logging parameter (drilling time, SPP, returning cutting, etc.) to predict the lost circulation in time.Development period: the process of lost circulation is monitored by using synthetic log techniques, combining previous dates to analyze the cause of lost circulation and draw up relevant countermeasures for the control of lost circulation.

| CONCLUSIONS
(1) The abnormal response of the loss layer is analyzed, by combining it with geological characteristics, adjacent well drilling, and logging.On the basis of the theory of engineering fuzzy mathematics, the mathematical model of loss probability evaluation is developed to comprehensively evaluate the formation loss probability and predict the potential loss layer in the same block before drilling.
(2) Through stress-sensitive laboratory experiments and numerical simulation, the deformation law of multiscale fractures can be predicted, which is helpful to optimize the plugging material.The method has been tested on-site and the sealing effect is evident.

F 1 Factors i and j are equally important 3 Factor i is slightly more important than factor j 5 Factor i is significantly more important than factor j 7 Factor i is more important than factor j 9 Factor i is extremely more important than factor j 2 , 4 , 6 , 8
I G U R E 2 Fracture and caves in FMI.T A B L E 1 The 1-9 scale judgment matrix scale meaning.ScaleMeaning Judgment median of two adjacent factors ZHANG ET AL.
and the change value of standpipe pressure, and roughly determine the location of the loss layer by measuring and comparing the changes of annulus pressure consumption above the loss point before and after.(2)Calculating the relevant parameters of the location of the loss layer The annulus flow rate of drilling fluid.
flow pattern of drilling fluid is laminar; Re n > 4270-1370 ai ai , the flow pattern of drilling fluid is turbulence; n Re n 3470-1370 < < 4270-1370 ai ai ai , the flow pattern of drilling fluid is transitional flow.
flow pattern of drilling fluid is the transitional flow:

2 |
Predicting fracture width variationThe bottom hole pressure fluctuation caused by the drilling engineering disturbance causes the formation of multistage fractures to deform and expand, and the closed fractures

F I G U R E 5
Different samples' fracture width versus effective stress in experiments.

( 3 )
Mechanical parameters of formation rock Mechanical parameters of formation rock at the well depth of 6000 m: Elastic modulus of rock E = 3.06 × 10 4 MPa, Poisson's ratio v = 0.32, Biot coefficient α = 0.74, formation pore pressure P 0 = 60 MPa, maximum horizontal principal stress σ H = 118.5 MPa, minimum horizontal principal stress σ h = 92 MPa.According to the void elastic effect, the payload in the finite element model is calculated as follows:

F
I G U R E 6 Stress analysis of borehole wall fracture.F I G U R E 7 (A) Numerical models of fracture-caved rock mass.(B) Stress analysis of fracture at the quarter wellbore.ZHANG ET AL.

F
I G U R E 8 Simulation results of fracture deformation (D = 50 mm).F I G U R E 9 Fluid pressure distribution in the fracture at the borehole wall.

F
I G U R 10 Numerical simulation fracture deformation with different caves.

F
I G U R E 11 Identifying mud loss intervals through the drilling-time curve.

F
I G U R E 12 Different flows and the fluid level of drilling fluid pool versus different loss types. 3T A B L E 3 Drilling fluid logging monitor parameters responses to the different loss types.Loss types The monitor parameters response of the drilling fluid logging The change in the fluid level of the drilling fluid pool The change of differential flow of in/out of drilling fluid Permeable lost circulation Down slowly, tending towards the normal stabilization with the solid phase in the drilling fluid sealing Increase slowly at the original time, down with the sealing, and get right at last Natural fracture lost circulation Down abruptly at the original time, down slowly with the fracture are sealed Increase abruptly at the original time, then down slowly gradually Fracture propagation lost circulation Down slowly, tending towards normal with the sealing is successful Increase abruptly, then down abruptly, and recover the original data at last Large-scale fracture-cave lost circulation Down abruptly, and no tending towards slowing down Increase abruptly, remain unchanged with the occurrence of the lost circulation

T A B L E 4
Abbreviation: SEP, self-elevating platform.
R E 13 Permeable circulation and large-scale fracture-cave mud loss.The loss pressure model is established based on the loss mechanism, and the loss pressure generated by induced hydraulic fracture is calculated, which is helpful to determine the loss mechanism and type.The technical framework for loss diagnosis is preliminarily established, abnormal features such as drilling time, cuttings, drilling fluid, and loss pressure are analyzed, the type of loss is identified, and the process of loss is described, thus providing a theoretical basis for scientific control of loss.AUTHOR CONTRIBUTIONS Heng Zhang and Mingwei Wang: Conceptualization, funding acquisition, project administration, resources, writing-original draft, and software.Yong Gao and Xiaofei Wang: Investigation, methodology, software, data curation, formal analysis, and methodology.Song Li: Writing-original draft, writing-review and editing.Wancai Nie: Project administration and resources.All authors have read and agreed to the published version of the manuscript.