Development scheme screening considering reservoir damage based on the AHP model

Abstract

The development process of oil and gas fields is affected by engineering and geology. Reservoir damage is significant as a bridge connecting these two disciplines. The life cycle of oil and gas fields can be prolonged by making development plans that consider reservoir damage. Therefore, this study performed experiments and discusses a Cretaceous reservoir development in the Chun17 block of the Chunguang Oilfield. We have carried out five sensitivity experiments on typical cores from the block, and analyzed lithology, physical properties, and crude oil. A mathematical model based on the Analytic Hierarchy Process (AHP) was then established based on the core data.Overall, the reservoir in the study area has ultra-high porosity and high permeability. The reservoir has moderately strong water sensitivity (damage rate is 38 ~ 90%) and moderately strong alkali sensitivity (damage rate is 22 ~ 75%). The crude oil belongs to extra-heavy oil reservoirs (containing 2.43% wax, 0.96% sulfur, 20.25 m% gum, 7.35 m% asphalt, 38.26 m% saturated hydrocarbon, and 27.51 m%). The crude oil is sensitive to temperature changes. Based on the AHP model analysis, the development scheme chooses the third (weight 0.1757) of the conventional analytic results of AHP; the Thermal oil production C1 scheme has the most stable future trend and appears to be the best option. In the next step, using fire flooding technology to develop the reservoir will benefit the most economically. The workflow consisting of "conclusion of reservoir experiment → AHP mathematical modeling → verification of relative permeability of high-temperature oil and water → verification of trial production" is simple and effective. The findings of this study can help to better understand the idea and process of making development plans for sensitive oil and gas fields through reservoir evaluation. Combining experimental data with mathematical modeling can find a balanced combination point in qualitative and quantitative analysis. Whether this combination point is correct can be verified by field engineering practice.

Article highlights

1. Heavy oil development schemes in the study area are discussed.

2. Cost-benefit analysis is considered by combining mathematical and experimental data.

3. Verification of the development method by experiments and trial production.

1 Introduction

The development of oil and gas fields should follow reservoir characteristics and geological trends. Economic exploitation is done based on reasonable development methods, surface engineering, and string matching. Optimizing and adjusting the development scheme based on reservoir damage couldmaximize the development life of the oil and gas fields and achieve sustainable development. Reservoir damage mainly includes the damage caused by external fluid entering oil and gas reservoirs and reservoir sensitivity[1,2,3]. Reservoir sensitivity refers to the property that the pore structure and permeability of oil and gas reservoirs change due to various physical or chemical interactions with exogenous fluids [4]. If the reservoir interacts with exogenous foreign fluids, the permeability of the reservoir deteriorates, damaging the reservoir to varying degrees, eventually leading to productivity loss or production decline.

Theoretically, research on reservoir damage mainly focuses on the simulation of actual working conditions by mathematical and physical models[5]. It is embodied, for example, in the study performed by Chengyuan Xu (2018), who established a microscopic model to characterize the particle-suspension migration law in fractured media. This model is used to explain the connectivity change of fracture network[6]; L smeraglia (2021) used remote sensing images of unmanned aerial vehicles to build a multi-scale model for faults and fracture development areas of carbonate rocks in central Apennine, Italy [7].

Research on reservoir damage mainly focuses on measuring data under different assumptions using the presupposition of instruments, equipment, and experimental means[8]. It is embodied, for instance, in the study performed by Kamal (2019), who studied the damage of clay minerals to reservoirs by the injection experiments of different media. With the help of X-ray diffraction (XRD), scanning electron microscope (SEM), nuclear magnetic resonance (NMR) analysis and other means, quantitative analysis and graphical display were carried out. Finally, the damage degree and probability of different media to reservoirs are obtained [9]; Wang (2021) constructed the core displacement experiment of two media, and obtained the particle migration law with the help of scanning (CT) images[10]. Boyang Wang (2019) conducted the water sensitivity (salt sensitivity) experiment under the preset boundary conditions, and analyzed the mineral content and occurrence in the sampling area by means of scanning electron microscope (SEM) and infrared spectroscopy [11].

As the results-oriented, the research on reservoir damage mainly focuses on the phenomenon characterization and engineering problem-solving in specific blocks under certain working conditions. ** on the capillary pressure trend line. The analysis shows that the capillary pressure curve platform is prominent, and the size and distribution of particles in the rock mainly determine the pressure. The curve is characterized by platform development, indicating that the pore radius is medium and concentrated, and the obvious degree of the platform shows that the pore throat has good sorting properties.

4 materials and methodology

Reservoir damage is defined as the vicious changes in pore structure and permeability caused by various physicochemical interactions between reservoirs and foreign fluids. Therefore, it is necessary to analyze the sensitivity of various formation damage types before oil and gas development and determine the reasonable development mode. The main causes of reservoir damage are invasion and blockage of foreign particles, the interaction between foreign fluids and rocks, incompatibility between foreign fluids and reservoir fluids, and microbial action. However, the interaction between foreign fluids and rocks can cause hydration expansion of clay minerals, migration of particles in the formation, and chemical precipitation during acidification, which should be considered in the primary oil recovery process. According to the SY/T5358-2010 evaluation method of reservoir sensitivity flow experiment, the experimental analyses of velocity sensitivity, water sensitivity, acid sensitivity, alkali sensitivity, and salt sensitivity were carried out, as described below [19].

The general experimental structure is as follows: 1. To ensure the mineral composition and pore structure of the original core remain unchanged, the samples of the original core at five depths of 952.30 m, 952.88 m, 955.14 m, 957.08 m, and 958.58 m were prepared. Rock samples were prepared (2.54 cm in diameter and 3.81 cm in length) with flat end faces and cylindrical surfaces; 2. The rock sample was washed with alcohol for 10 min and then desalted with methanol. The original fluid was cleaned and the rock sample was ensured to be hydrophilic; 3. The reaction kettle was used to dry the rock sample. The drying temperature was ≤ 60 °C for ≤ 48 h with relative humidity within 40 ~ 50%; 4. Saltwater with a salinity of 8% was prepared, and an analog liquid similar to formation crude oil was on standby; 5. According to the requirements of the sensitivity test, the rock samples were placed on the test bench (Fig. 6) in turn, and the experiments were carried out at a constant speed or pressure.

Fig. 6

Core sensitivity evaluation test bench setup

The general calculation formula for permeability is as follows:

$${\text{K}}_{1} = \frac{\mu \times L \times Q}{{\Delta p \times A}} \times 10^{2}$$

(1)

where: $K_{1}$ is the liquid permeability of the rock, ($10^{ - 3} \mu m^{2}$); $\mu$ is the fluid viscosity under test conditions, ($mPa \cdot s$); $L$ is the length of the rock sample, ($cm$); $A$ is the cross section area of the rock sample, ($cm^{2}$); $\Delta p$ is the pressure difference at both ends of the rock sample, ($MPa$); $Q$ is the volume of fluid passing through the rock sample in unit time, ($cm^{3} /s$);

$$V_{p} = \frac{{m_{1} - m_{0} }}{{\rho_{t} }}$$

(2)

$$\phi = \frac{{V_{p} }}{{V_{t} }} \times 100\%$$

(3)

where: $m_{0}$ is the mass of the dry rock sample, ($g$); $m_{1}$ is the mass of the rock sample after saturated with liquid, ($g$); $\rho$ is the density of the saturated liquid at the measured temperature, ($g/cm^{3}$); $V_{p}$ is the effective pore volume of the rock sample, ($cm^{3}$); $V_{p}$ is the total volume of the rock sample, ($cm^{3}$); $\phi$ is the porosity of rock sample;

$$D_{w} = \frac{{\left| {K_{w} - K_{i} } \right|}}{{K_{i} }} \times 100\%$$

(4)

where: $D_{w}$ is the permeability change rate of the rock sample corresponding to different flow velocities; $K_{w}$ is the permeability of the rock sample (measured at different flow velocities in the experiment), ($10^{ - 3} \mu m^{2}$); $K_{i}$ is the initial permeability of the rock sample (measured at the minimum flow velocity in the experiment), ($10^{ - 3} \mu m^{2}$).

Velocity sensitivity is the phenomenon that particles in reservoir rocks migrate and block the roar channel due to the change in fluid flow velocity, which leads to a change in reservoir rock permeability [20]. A high-pressure displacement pump was used to pump the simulation fluid through the sample, and the confining pressure was slowly adjusted to 2.0 MPa. The test bench valves were gradually opened for a flow rate of 0 ~ 90 m / D. Permeability of the rock sample was then recorded. Finally, after each flow state stabilized, the ratio of the permeability of the experimental rock sample to the original permeability was recorded. The diagram between the permeability change rate of the rock sample (ordinate) and the flow velocity (abscissa) was drawn.

Water sensitivity occurs when injected water with lower salinity enters the reservoir; clay minerals expand, disperse and/or migrate, causing changes in seepage channels and permeability[21]. Salt water with a salinity of 8% was flooded into the samples. The same initial speed as the speed-sensitive experiment to measure the liquid permeability was used, and the injection pore volume was gradually increased, ensuring the same displacement speed. Under the premise of kee** the confining pressure and temperature unchanged, the core permeability was measured when 10 ~ 15 times the pore volume of the rock sample was displaced. Finally, the diagram between the permeability rate of change (ordinate) and injected pore volume multiple (abscissa) was drawn after data calculation.

Salinity sensitivity occurs after a series of brine with different salinity enters the reservoir; clay minerals expand, disperse and/or migrate due to changes in fluid salinity, resulting in changes in reservoir rock permeability[22]. The salinity of the experimental fluid ranged between 0 ~ 60000 mg / L according to the results of the water sensitivity experiment. The other parameters were set according to the speed sensitivity experiment. The confining pressure and temperature were kept constant. Core permeability was measured when 10 ~ 15 times the pore volume of the rock sample was displaced. Finally, the diagram between permeability change rate (ordinate) and mineralization (abscissa) of rock samples was drawn after data calculation.

Alkaline sensitivity occurs when an alkaline liquid reacts with reservoir minerals to cause precipitation or induce clay minerals to expand, disperse and/or migrate, resulting in changes in permeability[23]. The prefabricated simulation liquid was first introduced into the test bench to measure the original permeability of the core. NaOH was mixed with the prefabricated simulation liquid, and the flow velocity was set according to the speed-sensitivity experiment. The pore volume of the rock sample was measured when 10–15 times the fluid was displaced. The pH was gradually increased by 1–1.5. The alkali liquid fully reacted with the rock minerals for over 12 h. The pH experiments ended at 13 h. Finally, the diagram between permeability change rate (ordinate) and pH (abscissa) was drawn after data calculation.

Acid sensitivity is when an acid reacts with reservoir minerals to produce precipitation or release particles, which changes reservoir rock permeability[24]. The prefabricated simulation liquid was first introduced into the test bench to measure the original permeability of the core. Then, 15% HCl was mixed with the prefabricated simulation liquid, and 0.5–1.5 times the pore volume of acid was reversely injected into the experimental bench. Finally, the displacement was stopped, and the inlet and outlet valves of the gripper were closed. The reaction was allowed to proceed for 1 h, followed by the forward displacement of the prefabricated simulation liquid. Permeability was then measured. A diagram between the permeability change of the rock sample (ordinate) and the displacement pore volume multiple (abscissa) was then drawn.

The study of crude oil characteristics mainly refers to fluid viscosity, density and gas saturation characteristics. Because the research object is heavy oil, the relationship between crude oil viscosity and the temperature was mainly studied. Crude oils from five representative wells were collected, and the six-speed rotary viscometer, thermostatic water bath heater, densitometer, and corresponding recording instruments were selected to construct the experimental bench. The viscosity of crude oil at formation temperature of 25 ~ 85 °C was measured by simulation. The measured liquid was placed in a beaker container with a diameter ≥ 70 mm and a height ≥ 125 mm. The water bath heater was ensured a constant temperature before starting the measurement. Also, water was required to check a horizontal bench accurately. The viscosity range of crude oil was checked to select the appropriate rotor and speed. To ensure the authenticity of the experimental conclusion, the No. 1 to No. 4 rotors of the equipment were selected and matched with 8-speed regulating gears of 0.3rmp, 0.6rmp, 1.5rmp, 3rmp, 6rmp, 12rmp, 30rmp, and 60rmp. The diagram between viscosity and temperature was drawn after normalizing the data.

5 Experimental investigation

5.1 Reservoir damage

The velocity sensitivity characteristic curve is shown in Fig. 7. When the main factor causing permeability change is velocity, with the increase in injection velocity, fine particles on the surface of the rock particles are washed off by the fluid. These fine particles often transfer to the bellowing part with the fluid. The permeability is reduced when the volume of fine particles accumulates above the bellowing radius. The permeability of this rock sample shows a slight downward trend as a whole, indicating that when the injection velocity increases, a certain degree of particle migration occurs in rock samples. f the migration channel is not smooth, it can directly lead to changes in the overall seepage capacity of rock samples. When describing the relationship between velocity and permeability in days, the critical velocity is 5 ~ 42 m/d. The damage rate of velocity sensitivity is 13 ~ 67%, with an average of about 30%. Overall, there is a weak velocity sensitivity.

Fig. 7

Velocity sensitivity characteristic curve of the Cretaceous reservoir in the Chun17 block

The water sensitivity characteristic curve is shown in Fig. 8. The surface fluid injection amount and salinity are significant in controlling the change of reservoir rock samples. With the increase in injection times and the decrease in salinity of injected water, the change in permeability caused by injected water always decreases. The sensitivity of distinguishing strong and weak in the study area is limited by a 50% damage rate. The water sensitivity damage rate of the Cretaceous reservoir is about 38 ~ 90%, with an average of about 65%. The average damage rate exceeds a 50% damage rate limit, thus showing moderately strong water sensitivity characteristics.

Fig. 8

Water sensitivity characteristic curve of the Cretaceous reservoir in the Chun17 block

The salinity sensitivity characteristic curve is shown in Fig. 9. The greater the difference in the salinity of fluid in the reservoir, the more severe the damage to the physical properties of the reservoir. As the salinity of injected water decreases, the permeability decreases. The greater the salinity of water injection in the process of water injection development, the more obvious the permeability change caused by changing the salinity. Overall, the physical properties change greatly under high salinity, with a critical salinity of 40,000 mg/L.

Fig. 9

Salt sensitivity characteristic curve of the Cretaceous reservoir in the Chun17 block

Based on the analysis of the alkali sensitivity characteristic curve (Fig. 10), it is easier to form sediment to disperse and migrate when the salinity of such injection fluid is higher. With the increase in the pH of the injection water, the permeability the rock samples show a downward trend as a whole. The sensitivity of divided reservoirs to alkaline fluids is limited to 50%. Through targeted sampling tests, the alkali sensitivity damage rate is 22 ~ 75%, and the average damage rate is about 52%, showing moderate to strong alkali sensitivity.

Fig. 10

Alkali sensitivity characteristic curve of the Cretaceous reservoir in the Chun17 block

From the analysis of the acid sensitivity characteristic curve (Fig. 11), with an increase in the volume multiple of injected pores, the permeability of injected water of rock samples shows a downward trend. The sensitivity is divided into strong and weak by the 50% damage rate point. The sampling test results show that the acid sensitivity damage rate of the reservoir in the study area is 16 ~ 65%, with an average damage rate of 33%, indicating that the acid injection body has a weak damage effect on the reservoir physical properties. The main influencing factor is the released particles, which block the pore channels along the directional transfer of fluid and show moderately weak acid sensitivity characteristics.

Fig. 11

Acid sensitivity characteristic curve of the Cretaceous reservoir in the Chun17 block

A comprehensive analysis of the potential sensitivity and dynamic sensitivity characteristics of the above reservoirs was conducted. Due to the difference between salinity and acidity of formation water, comprehensive analysis of all samples, elimination of special circumstances for comprehensive evaluation of reservoir sensitivity, and summary of similarities and differences of different types of sensitivity greatly influence the accuracy of evaluation results. The evaluation results are shown in Table 4.

Table 4 Comprehensive analysis of cretaceous reservoir sensitivity characteristics in the Chun 17 block 

5.2 Characteristics of crude oil

Crude oil samples were taken from wells Chun2-200, Chun2-9, Chun101e, and Chun10-7 for the degassing tests. The analysis results show that the density of ground-degassed crude oil in the KIV 2–3 sublayer of the Chun17 block is 0.9658 g/m³. Viscosity test results show that the viscosity of crude oil at reservoir temperature is lower than that after degassing at ground temperature, which limits the formation temperature range to a certain extent. The viscosity of degassed crude oil at 50 °C is 26,350.8 mPa.s through weighted average viscosity data statistics, and the degassed viscosity under the oil layer is 34,424.68–62,055 mPa.s when the crude oil temperature is controlled close to the formation temperature by technical means. It contains 2.43% wax, 0.96% sulfur, 20.25 m% gum, 7.35 m% asphaltene, 38.26 m% saturated hydrocarbon, and 27.51 m% aromatic hydrocarbon. The viscosity changes significantly at reservoir temperature, which indicates that crude oil is sensitive to temperature changes (Fig. 12). The viscosity changes of crude oil in the degassed state are not obvious compared with those in the non-degassed state [25]. These crude oils are classified as extra and super heavy oil reservoirs.

Fig. 12

Viscosity-temperature curve of crude oil in the KIV 2–3 sublayer of the Chun17 block

The crude oil is super heavy oil, denuded to some extent in the upper part of the Cretaceous reservoir west of the study area. The viscosity of the Cretaceous reservoir near the denuded surface is higher due to the influence of surface water. According to the total analysis results of crude oil, the viscosity of crude oil in the Chun 2–5 well located near the denudation surface is 61,005 mPa.s at the oil layer temperature, indicating that the crude oil in this area is a super heavy oil.

The viscosity of crude oil in the Chun 2–200 at reservoir temperature is about 312,433 MPa.s, and the crude oil is super heavy oil. The oil test and production test analysis, combined with the relative position of oil-bearing horizon and denudation surface and sampling analysis, shows a significant difference in oil test occurrence between the two oil wells. Overall, the recovery capacity of extra heavy oil is poor. The main oil-bearing horizon of the oil well near the high part in the northwest is greatly affected by the denudation surface because the plane position of the oil well is closely related to the denudation surface; so the geographical position should be noted when summarizing the influencing factors of crude oil viscosity [26]. The viscosity distribution of the Cretaceous crude oil in the Chun17 block has the following trends: The area affected by the denudation surface is closer to the provenance in the horizontal direction, and the oil-bearing horizon is relatively close to the denudation surface. The oil well near the denudation surface has a high viscosity, mainly super heavy oil. The Chun2-5 well near the denudation surface was analyzed emphatically. The oil property characteristics of the Chun2-5 well are symbolic. The oil viscosity of well Chun2-5 is 60,000 MPa.s.

In the area where the oil-bearing horizon is far away from the denudation surface, well Chun2-200 was taken as the research object. The oil viscosity of K436-34H corresponding to the oil-bearing horizon far away from the denudation surface is relatively low. The laboratory data show that the viscosity of crude oil from well K436-34H to Chun2-200 decreased from 41,300 MPa.s to about 30,000 MPa.s.

5.3 AHP modeling

Five schemes, including thermal oil recovery, oil-based fracturing, chemical flooding, heavy oil fire flooding, and cold sand production, were selected for evaluation. The analytic hierarchy process (AHP) was selected to establish a structural model with reservoir damage, production cost, recovery factor, and technical maturity as evaluation criteria to maximize reservoir protection. The best development scheme was finally determined by comparing the results of conventional, interactive, and sensitivity analyses [27].

The analytic hierarchy process (AHP) decomposes abstract complex problems into various components through decision-making, then sets the factors and problem logic into a target, criterion, and scheme layer with a hierarchical grou** structure. Finally, matrix construction compares the mutual influence relationships between the factors, thus quantifying the problem components and calculating their respective weights in the system [28].

Firstly, five selected schemes were summarized:

• (1) Thermal oil recovery mainly adopted three methods: steam huff-puff, steam drive, and SAGD, all of which inject high-temperature steam into the ground, with an average cost of RMB 140 yuan per ton of steam injection, and the steam injection cost accounts for more than 60% of the production cost of thermal oil recovery. The average cost of thermal recovery of heavy oil is 1.5 times that of thin oil, and the selling price is 85% of that of thin oil. The price and cost are too high. Furthermore, the recovery ratio is 30 ~ 53% [29].

• (2) Oil-engine fracturing: this technology uses sodium acetate and a low-temperature activator to prepare a compound gel breaker in proportion, which can completely break the gel in a controllable time at 30 ~ 40 °C and is suitable for reservoirs with strong water sensitivity, small pore throats, and easily water locked. Low temperatures and pressures are not conducive to gel breaking and flowback of fracturing fluid. This technology was successfully applied to well Pai 2–401 in the Chunguang Oilfield in 2012, with a total of 138m3 fracturing fluid and 11.3m3 sand injected in the interval of 772.6 ~ 777.6 m. The highest sand ratio was 30% and the average sand ratio was 14%. The daily oil production increased from 0.6t to 2.8t after fracturing [30].

• (3) Chemical flooding technology includes alkali flooding, polymer flooding and surfactant flooding, and was successfully applied to conventional crude oil production. Alkali flooding and surfactant flooding improve production efficiency by reducing oil–water interfacial tension. Polymer flooding increases production by reducing the water–oil mobility ratio and increasing sweep efficiency. According to adaptability research, polymer flooding is still useful in crude oil reservoirs with a viscosity as high as 10,000cp. At the application level, Canada has achieved more than 60,000 barrels per day in the Pelican Lake project. This method is predicted to increase oil recovery to 20 ~ 30% of OOIP. The ultimate recovery ratio of the polymer flooding project in the Mooney Oilfield can reach 17 ~ 25%. By March 2013, the recovery ratio of the central well in the test area was close to 12% of OOIP [31].

• (4) The recovery mechanism of heavy oil in situ combustion consists of air being injected into the oil layer, and oxygen in the air reacts with crude oil to generate heat, thus driving the recovery of crude oil. Fire flooding has the advantages of a high recovery ratio, low cost, and wide application range. Under the same conditions, the heat energy loss accounts for only 25% of the steam drive. Compared with steam flooding and SAGD technology, fire flooding has a broader application range, and the recovery ratio is 50 ~ 80% [32].

• (5) Cold Heavy Oil Production with Sand (CHOPS) does not use heat injection or sand control, however, the crude oil and sand are produced together by a screw pump. There are two crucial mining mechanisms: one is the formation of a "wormhole," where a large amount of sand is produced and the porosity and permeability are greatly improved. Secondly, foam oil is formed, produced simultaneously as the dissolved gas and crude oil become foam oil. This maintains the stability of the earthworm hole, avoids degassing of bottom crude oil and prolongs the stable production time. Dissolved gas provides the internal driving force for crude oil, which reduces the viscosity of crude oil and is more conducive to the flow of crude oil. The thickness of the cold sand production reservoir should be more than 3 m. The burial should be more than 300 m. The best viscosity is about 1,000 ~ 50,000 Mpa.s, and the density of degassed crude oil is 0.92 ~ 0.98 g/cm³. Too thin of a reservoir is not conducive to the maximization of economic benefits and the formation of a wormhole network. Too shallow of a reservoir leads to an energy shortage. This technology is best applied to undeveloped new areas or new strata in old areas [33].

Finally, the structural model diagram of scheme evaluation was developed (Fig. 13). For the convenience of matrix operation, A, B, and C labels are used instead of the target layer, criterion layer, and scheme layer. The calculation process is as follows:

Fig. 13

Structural evaluation and production diagram

• (1) Determine the objectives of development plan evaluation tasks and build a hierarchical structure model.

• (2) According to the calculation results of the comprehensive correlation degree of the X1 ~ X12 sequence with respect to the × 0 sequence and engineering experience, the score of the lower to upper layers is determined. In the criterion layer, the proportion of each factor is different under the decision maker's measurement standard. The numbers 1 ~ 9 and their reciprocals are used as scales to define and judge the matrix [34].

• (3) Hierarchical synthesis calculation and consistency are checked using formula (5) to calculate consistency index CI:

  $${\text{CI}} = \frac{{\lambda_{\max } - n}}{n - 1}$$

  (5)

$\lambda_{\max }$ In the formula: ф is the maximum eigenvalue of the judgment matrix; N is the column vector.

Formula (6) is used to calculate the consistency ratio CR:

$${\text{CR}} = \frac{{{\text{CI}}}}{{{\text{RI}}}}$$

(6)

where: CI is the consistency index; RI is a random consistency index.

When CR < 0.10, it indicates that it has passed a one-time inspection, otherwise, it will not have satisfactory consistency.

(4) The weight vector is then calculated. The geometric average method is used, consisting of the root of the continuous product of each variable value calculated several times. The geometric average is used to calculate the development speed of the predicted target [35]:

$$W_{i} = \frac{{\left( {\prod\nolimits_{j = 1}^{n} {a_{ij} } } \right)^{\frac{1}{n}} }}{{\sum {_{i = 1}^{n} } \left( {\prod\nolimits_{j = 1}^{n} {a_{ij} } } \right)^{\frac{1}{n}} }}\,i\, = \,1,2,3 \ldots , \, n$$

(7)

The failure analysis of the pneumatic control valve obtains each judgment matrix and weight vector W.

$$\begin{gathered} \, \,A \, \,\,\,\,B1\,\, \, B2 \, \,\,B3 \, \,\,B4 \hfill \\ \begin{array}{*{20}c} {B1} \\ {B2} \\ {B3} \\ {B4} \\ \end{array} \left[ \begin{gathered} 1\,\,\, 4\,\,\,\,\,\, 2\,\,\,\,\,\,\,\,\,4 \, \hfill \\ \,\,\,\,\,\,\,1 \,\,\,\,\,1/3 \,\,\,1 \hfill \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,1\,\,\,\,\,\,\,\,3 \hfill \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\, \,\,\,\,\,\,\,\,\,\,\,\,\,1 \hfill \\ \end{gathered} \right] \, \hfill \\ \end{gathered}$$

Weight vector W = (0.4852, 0.1090, 0.2968, 0.1090); CR = 0.0077

$$\begin{gathered} B1 \, \,\,\,\,\,\,\,C1 \,\,\,C2 \,\,C3 \,\,\, C4 \,\,\, C5 \, \hfill \\ \begin{array}{*{20}c} {C1} \\ {C2} \\ {C3} \\ \begin{gathered} C4 \hfill \\ C5 \hfill \\ \end{gathered} \\ \end{array} \left[ {\begin{array}{*{20}c} { \, 1} & { \, 2} & 1 & {1/3} & {1/3} \\ {} & { \, 1} & { \, 1/2} & {1/3} & {1/4} \\ {} & {} & 1 & {1/2} & {1/3} \\ {} & {} & {} & 1 & {1/2} \\ {} & {} & {} & {} & 1 \\ \end{array} } \right] \hfill \\ \end{gathered}$$

Weight vector W = (0.1255, 0.0780, 0.1342, 0.2646, 0.3978); CR = 0.0161

$$\begin{gathered} B2 \,\,\,\,\,\,\,\, C1 \,\,\, C2 \, \,\,C3 \, \,\,C4 \,\,\, C5 \, \hfill \\ \begin{array}{*{20}c} {C1} \\ {C2} \\ {C3} \\ \begin{gathered} C4 \hfill \\ C5 \hfill \\ \end{gathered} \\ \end{array} \left[ {\begin{array}{*{20}c} { \, 1} & { \, 3} & { \, 2} & { \, 1/3} & {1/2} \\ {} & { \, 1} & { \, 2} & { \, 1/2} & {1/2} \\ {} & {} & { \, 1} & { \, 1/4} & {1/3} \\ {} & {} & {} & { \, 1} & {1/2} \\ {} & {} & {} & {} & 1 \\ \end{array} } \right] \hfill \\ \end{gathered}$$

Weight vector W = (0.1808, 0.1216, 0.0773, 0.2916, 0.3288); CR = 0.0710

$$\begin{gathered} B3 \,\,\,\,\,\,\,\,\, C1 \,\,\, C2 \,\,\, C3 \,\,\, C4 \,\,\, C5 \, \hfill \\ \begin{array}{*{20}c} {C1} \\ {C2} \\ {C3} \\ \begin{gathered} C4 \hfill \\ C5 \hfill \\ \end{gathered} \\ \end{array} \left[ {\begin{array}{*{20}c} { \, 1} & { \, 2} & 3 & {1/2 \, } & 1 \\ {} & { \, 1} & { \, 1/3} & {1/5 \, } & 1 \\ {} & {} & 1 & {1/3 \, } & 1 \\ {} & {} & {} & 1 & 1 \\ {} & {} & {} & {} & 1 \\ \end{array} } \right] \hfill \\ \end{gathered}$$

Weight vector W = (0.2240, 0.0939, 0.1484, 0.3494, 0.1843); CR = 0.0954

$$\begin{gathered} B4 \,\,\,\,\,\,\,\,C1 \,\,\, C2 \,\,\, C3 \,\,\, C4 \,\,\, C5 \, \hfill \\ \begin{array}{*{20}c} {C1} \\ {C2} \\ {C3} \\ \begin{gathered} C4 \hfill \\ C5 \hfill \\ \end{gathered} \\ \end{array} \left[ {\begin{array}{*{20}c} { \, 1} & { \, 2} & {1/2} & {3 \, } & 6 \\ {} & { \, 1} & { \, 1/3 \, } & {5 \, } & 6 \\ {} & {} & 1 & {4 \, } & 5 \\ {} & {} & {} & {1 \, } & 3 \\ {} & {} & {} & {} & 1 \\ \end{array} } \right] \hfill \\ \end{gathered}$$

Weight vector W = (0.2621, 0.2158, 0.3963, 0.0828, 0.0430); CR = 0.0746.

Different engineers have differences in weight selection and matrix construction, and the final data of modeling results even have significant deviations. However, here, this method is only used to detect the development trend of factor ranking and sensitivity analysis, and it weakens the interference caused by the weight results. The accuracy of the AHP modeling conclusion can be further demonstrated through field production data and compensatory experiments.

The above CR calculation results are all less than 0.1 and pass the one-time inspection. The weight matrix W of each criterion element of different schemes and the relative weight matrix (B1-B4) W of each criterion element are respectively multiplied to obtain the sample weights of the criterion layer and the scheme layer (Fig. 14).

Fig. 14

Sample weights of the criterion layer and scheme layer

The order of scheme layer weights is C5 > C4 > C1 > C3 > C2. The order of the weight of the criterion layer is B1 > B3 > B4 = B2, which accords with the original research intention that reservoir damage should be considered first in the evaluation of the development scheme. The weight of the other criterion layers prioritizes the recovery factor, and it has been determined by default that the mining cost is consistent with the weight of technical maturity.

Data interaction analysis was performed based on the stable model and one-time compliance inspection. Assuming that the weights of B1, B2, B3, and B4 in the criterion layer are the same, the calculation results in Fig. 15 show that the order of scheme layer weights is C4 > C5 > C1 > C3 > C2.

Fig. 15

Assumptions of the scheme layer sample weights with the same criteria layer weights

The sensitivity analysis aimed to determine the influence of changes in system parameters or surrounding conditions on the current model state. Generally, when one factor at the criterion level changes, each factor at the scheme level changes accordingly[36]. The engineering significance was based on the premise of ensuring the stability of the model. By assuming the weight development of the factors in the criterion layer, the future change trend of each sample in the scheme layer is predicted. Assuming that the weights of B1, B2, and B3 in the criterion layer increase, the variation trend of each sample in the scheme layer was obtained (Figs. 16, 17, 18 and 19).

Fig. 16

Sensitivity analysis curve of condition B1

Fig. 17

Sensitivity analysis curve of condition B2

Fig. 18

Sensitivity analysis curve of condition B3

Fig. 19

Sensitivity analysis curve of condition B4

According to the engineering experience, the evaluation of the development plan needs to consider whether the long-term implementation process has a significant impact on the factors of the criterion layer, so it is necessary to optimize the development plan with a stable development trend under the framework of the elements of each criterion layer, and the insensitive sample is determined as the best plan here [37].

Engineering significance: sensitive samples may have an unexpected influence on development results in the future. It is necessary to adjust the cost and technical maturity in advance. To visually show the operation process, determine whether the sample is sensitive according to the changing relationships of the slope of a straight line in Figs. 16, 17, 18 and 19. The results show that C5 is considered sensitive under the B1 hypothesis, all samples are stable under the B2 hypothesis, and C1 is the most stable; C4 and C5 are considered sensitive under the B3 hypothesis. All samples are determined as sensitive under the B3 hypothesis.

Routine analysis results show that the order of scheme level is C5 > C4 > C1 > C3 > C2. The weights of C5 and C4 are all greater than 0.2, which is 0.1 order of magnitude higher than those of C1, C3, and C2. Hypothetical analysis results: assuming that the weight of the criterion layer is the same. The ranking of the scheme layer changes to C4 > C5 > C1 > C3 > C2. Only the ranking of C4 and C5 weights changes, and other parameters are similar to the conventional analysis results. Therefore, the conventional and interactive analysis results prove that C1, C3, and C2 samples are robust. The sensitivity analysis results show that C4 and C5 are sensitive, and C1 is the most stable under different assumptions. Therefore, based on the above three analysis results, C4 and C5 samples with poor robustness and sensitive judgment are preferentially eliminated. Thermal oil recovery C1, which has the third ranking in weight and the most stable trend in the future, is the best development scheme in the study area.

6 Results and discussion

6.1 Characteristics of relative permeability of high-temperature oil and water

According to the optimization results of the mining scheme in the previous section, thermal mining is the best way to develop the study area at present. This section further demonstrates the feasibility through experiments.

The core of the KIV2-3 layer of Chun 17–10 well in the study area has a lithology of of mainly fine-grained particles with a small amount of medium-grained sandstone interbedded. Fine-grained sandstone with uniform color is mainly selected during sampling. The high-temperature oil–water relative permeability experiments were carried out at 100 °C, 150 °C and 200 °C, with 100 °C as the initial temperature. The other two methods were adopted respectively. Experimental data and oil–water relative permeability curves are shown in Table 5, Table 6, and Fig. 20. The experimental results show that the high-temperature oil–water relative permeability of the target layer has the following characteristics.

Table 5 Characteristic data table of high-temperature oil–water relative permeability curve of the Cretaceous reservoir in the Chun17 block

Table 6 Characteristic data table of comprehensive relative permeability curve of the Cretaceous high-temperature oil–water in the Chun17 block

Fig. 20

Comprehensive relative permeability curve of high-temperature oil and water of the Cretaceous reservoirs at different temperatures in the Chun17 block

When the temperature increases from 100 °C to 200 °C, the irreducible water saturation of the reservoir increases from 28.3 to 36.2%, indicating that the rock particles attach more to the fluid in the temperature rise process. Or, the temperature can dredge the communication between pores and throats, which leads to the expansion of the fluid volume in the unit rock sample and the increase of the irreducible water saturation.

When the temperature increases from 100 °C to 200 °C, the residual oil saturation decreases from 36.2 to 25.5%, and irreducible water occupies the original residual oil space to a certain extent. In water injection development, a small part of the injected body replaces the residual oil and turns it into irreducible water. The water flooding efficiency increased from 50.5 to 60% during this heating process. In the process of increasing temperature, the production degree of the tight oil layer is improved, and the oil and gas recovery ratio is significantly increased. It shows that with the increase in temperature, the residual oil saturation gradually decreases, which fully proves that the thermal recovery technology should be effectively used to improve the oil recovery in the oilfield heavy oil development process.

When the temperature increases from 100 °C to 200 °C, the fluid saturation increases, and the proportion of the total volume of the rock sample occupied by the homogeneous fluid increases. The results of the sampling test show that the saturation of two-phase co-permeability increases from 50% to about 60%. When the temperature increases, the activity of oil and gas particles on the core surface is more robust than that of irreducible water, indicating that the core surface tends to be more hydrophilic.

In summary, the high-temperature oil–water relative permeability from 100 °C to 200°C shows that with an increase in temperature, the irreducible water saturation of the reservoir increases from 28.3 to 36.2%. A high temperature can strengthen pore throat connectivity. During the heating process, the water flooding efficiency increases from 50.5 to 60%, and the production degree of the tight oil layer increases. With the increase in temperature, the saturation of two-phase co-permeability increases from 50% to about 60%, and the core surface tends to be more hydrophilic. Through experiments, it is shown that the thermal recovery method is feasible. It is recommended to use steam stimulation-assisted gravity drainage (SAGD) to fully realize downhole heat communication and use horizontal wells and steam chambers to enhance oil recovery. At the same time, attention is needed to sand production caused by casing deformation in thermal recovery wells, and completion methods should be optimized. Sand control by gravel packing with a wire-wound screen. If economic conditions permit, other schemes, such as adding a viscosity reducer, can be tried for compound mining.

6.2 Trial production

This section further demonstrates the optimization of the AHP development scheme through the trial production of characteristic wells in the study area. Three vertical wells in the Chun17 block were tested by steam stimulation. The average daily output of the Chun 2–200 well in the first cycle was 5.8t, with a cycle production time of 114 days, a cycle oil production of 658t, and a stage oil–gas ratio of 0.64. The average daily output of the Chun 101E well in the first cycle was 5.2t, with a cycle production time of 53 days, a cycle oil production of 278t, and a stage oil–gas ratio of 0.19. The average daily output of the Chun 17–4 well in the first cycle was 3.4t, with a cycle production time of 34 days, a cycle oil production of 114t, and a stage oil–gas ratio of 0.14. The first-cycle production time of the Chun 2–200 well with relatively high steam injection intensity and matching viscosity reduction measures is longer. The daily output, periodic output, oil–gas ratio and other indexes are better than the Chun 17–4 well without matching viscosity reduction technology. The Chun2-200 with matched viscosity reduction measures had a relatively good production test effect (Fig. 21), and superior to the other two production test wells without matching technology.

Fig. 21

Comparison of production results in the first 3 cycles of the Chun 2-200 well

In summary, the trial production effect is consistent with the conclusion of AHP scheme optimization. The Chun2-200, with thermal recovery and matching viscosity reduction measures, has a good production effect. After matching viscosity reduction technology, the oil–water mobility ratio is reduced, and the three links of crude oil from formation to the wellbore, from the wellbore to the pump, and from the pump to the ground are smoother. These reduce the rapid decline in crude oil production due to lifting difficulties, thus ensuring the continuous expansion of steam thermal efficiency. The oil well can obtain longer production times, higher daily output, and an oil–gas ratio. Therefore, it is further demonstrated that using AHP to optimize the development scheme is feasible based on thoroughly studying reservoir sensitivity.

6.3 Results Summary

The reservoir sensitivity experiment shows that:

• 1 The acid sensitivity damage rate of the reservoir in the study area is 16 ~ 65%, and the average damage rate is 33%, which is determined as moderately weak acid sensitivity.

• 2 When describing the relationship between speed and permeability in units of days, the analysis shows that when the critical speed is 5 ~ 42 m/d, the damage rate of speed sensitivity is 13  ~ 67%, with an average of about 30%, indicating weak speed sensitivity on the whole.

• 3 The water sensitivity damage rate of the Cretaceous reservoirs is about 38 ~ 90%, with an average of about 65%, and the average damage rate exceeds the limit of 50% damage rate; these reservoirs have moderately strong water sensitivity.

• 4 Through targeted sampling tests, the damage rate of alkali sensitivity is 22 ~ 75%, and the average damage rate is about 52%, showing moderately strong alkali sensitivity.

Comparison of modeling analysis results so that:

• 1 AHP conventional analysis results show that prioritizing reservoir damage B1 in the criterion layer has the most considerable weight, and it is concluded that the optimal order of development schemes is C5 > C4 > C1 > C3 > C2. The weight of cold heavy oil production with sand C5 is 0.2882, and that of in situ combustion C4 is 0.2729, much larger than other samples in other scheme layers.

• 2 The results of hypothesis analysis results show that assuming all samples' weights in the criterion layer are the same, the optimal order of development schemes is C4 > C5 > C1 > C3 > C2. The weight of the cold heavy oil production in sand C5 is 0.2382, and the weight of in situ combustion C4 is 0.2469, much heavier than other samples in other scheme layers.

• 3 Sensitivity analysis results: In situ combustion C4 and cold heavy oil production with sand C5 are sensitive, and thermal oil production C1 is the most stable under different assumptions. The conventional and interactive analysis results show that C1, C3, and C2 samples are robust. Therefore, C4 and C5 samples with poor robustness and sensitivity are preferentially rejected by comparing the above three analysis results. Thermal oil production C1, which ranks third in weight and has the most stable trend in the future, is the best development scheme.

The relative permeability of oil and water at high temperatures shows that with the increase in temperature, the irreducible water saturation of the reservoir increases from 28.3 to 36.2%, and high temperature can strengthen the connectivity of the pore throat. During the heating process, the water displacing oil efficiency increases from 50.5 to 60%, and the production degree of the tight oil layer increases. With the temperature increase, the two-phase co-permeability saturation increases from 50% to about 60%, and the core surface tends to be more hydrophilic. Through experiments, it is proved that the thermal recovery method is feasible.

It should be noted that the AHP model can combine fuzzy and qualitative conclusions simultaneously. Samples of different dimensions can also be characterized by weight quantification. The normalized conclusion is drawn to guide the on-site production. Also, because there is some distortion in the process of data weight transformation, it is necessary to compare the results of conventional analysis, hypothesis analysis, and sensitivity analysis and determine the final result under the practical engineering framework to improve the simulation accuracy.

From an engineering prospect, for ultra-deep heavy oil (burial depth greater than 1,500 m), bottom hole dryness cannot be guaranteed during steam injection because of the deeply buried reservoir. Therefore, the wellbore heat loss is significant, so using steam injection for development is challenging. Fire flooding development is less affected by the buried depth of the reservoir. If the performance of the surface air compressor can be satisfied, the development of ultra-deep heavy oil by fire flooding will have more technical and economic advantages than steam injection. For thin-layered and thin-interbedded heavy oil, steam injection also faces the problems of significant heat loss (heat transfer to the cap, bottom layer, and interlayer), low thermal efficiency, and low economic benefit. With the improvement of technology maturity, develo** this kind of oil reservoir with fire flooding technology is expected to reduce production costs and improve economic benefits.

There are the following differences between this case study and previous research:

• 1 Our research focuses on proposing an engineering scheme that can better serve the site. Other examples pay attention to the history and degree of reservoir damage, such as in [1].

• 2 Our research experiment configuration is equal to the proportion of mathematical analysis, and it can describe the whole picture of the research case comprehensively and simply. Other examples only describe geological and engineering data, such as in [2].

• 3 Our research introduces the workflow details and verifies the experiment and trial production. However, the workflow given by other examples is primarily a novel hypothesis, which lacks field verification, such as in [5].

The advantages of this study are that it gives a workflow that has passed the field verification. This workflow can determine the development scheme's determination under the premise of a limited workload. The AHP model used in this workflow can integrate data and better transform quantitative and qualitative analyses. This study has several disadvantages, such as the limitation of relying too much on reservoir sensitivity experiments. The whole workflow cannot be carried out without experimental data, and other experimental and engineering data cannot be used instead. Secondly, this study is unsuitable for the initial stage of oil and gas field development, and it cannot be applied on the premise of fewer data to ensure the accuracy of the results. Finally, the workflow proposed in this study has a single function and can only make decisions. It does not have the function of prediction and probability statistics.

The novelty of this work consists of a research method to determine reservoir damage that combines experimental and mathematical modeling. This method can realize the functions of formulating the development plan and verifying the correctness of the research area. The mathematical modeling part of this method can realize the random conversion between quantitative and qualitative analysis and esnures accuracy of the results.

7 Summary and conclusions

Formulating a development plan based on the angle of reservoir damage is significant in heavy oil recovery with complex matching technology and a long development cycle. Given the limited geological data of the Cretaceous reservoir in the Chun17 Block of the Chunguang Oilfield, this study aims to scientifically evaluate the development plan of the study area with the primary consideration of reservoir damage. Research was performed on reservoir lithology, reservoir physical properties, reservoir damage, and crude oil characteristics in the study area. Using the AHP model to analyze data, routine analysis, hypothesis analysis, and sensitivity analysis were performed. A development plan was established by comparing the results. With the help of high-temperature oil–water relative permeability experiments and pilot production, reverse verification was carried out to guide the next large-scale exploitation. The engineering significance of this study consisted of suggesting to mine and use the existing geological feature data. This study jointly determined the preset development scheme and evaluated the matching. The interaction between geological and engineering factors in the dynamic development process was described, which intuitively reflects the nature of oil and gas field development, and guides sustainable development. The research results are as follows:

• 1. The reservoir sensitivity in the study area is weak velocity sensitivity, medium to strong water sensitivity, weak salt sensitivity (critical salinity 40,000 mg/l), medium to strong alkali sensitivity, and medium to weak acid sensitivity.

• 2. The thermal oil production C1 scheme, with the third largest weight ranking of AHP analysis results and the most stable trend in the future, is preferred for the development scheme.

• 3. In the next step, for ultra-deep heavy oil (more than 1,500 m burial depth), due to the deep burial depth of the reservoir, the dryness at the bottom of the well cannot be guaranteed during steam injection, and the wellbore heat loss is immense. Using fire drive technology to develop such reservoirs can reduce production costs and improve economic benefits.