Study on prediction technology and method of carbonate reservoir in Tahe area

(Planning and Design Institute of Northwest Petroleum Bureau of Urumqi Xinxing Company 8300 1 1)

The author analyzes the characteristics and difficulties of carbonate reservoir prediction in Tahe Oilfield in northern Tarim Basin, Xinjiang, and initially explores a set of methods and techniques for carbonate reservoir prediction using seismic data. The application in Tahe Oilfield has achieved good results. These technical methods mainly include: paleogeomorphology research, paleodrainage research, amplitude extraction, coherent calculation, wave impedance inversion technology, pattern recognition technology, multi-parameter histogram, two-dimensional crossplot, three-dimensional crossplot analysis technology and so on.

Paleogeomorphology research; Ancient water system research; Amplitude extraction; Coherent calculation; Wave impedance inversion; Pattern recognition; Multi-parameter analysis

1 Introduction

Ordovician carbonate reservoir in Tahe area is one of the main target layers for oil and gas exploration in northern Tarim. The reservoir has two remarkable characteristics: first, the target layer is deeply buried (mostly below 5000 meters) and the seismic reflection signal is weak; Secondly, the reservoir space is a system composed of structural fractures, dissolved pores, karst caves and fractures, with strong vertical and horizontal heterogeneity.

With the deepening of exploration and development, there are three main problems in the study of carbonate reservoir prediction:

Geophysical methods and technical exploration of (1) carbonate reservoir prediction.

(2) Distinguish the nature of the fillings in the carbonate reservoir space, that is, the identification of fillings such as oil, gas, water, siliceous, sandy argillaceous and calcite, that is, the identification of oil and gas.

(3) Establish the geological and geophysical model of carbonate reservoir.

In view of the above problems, the following carbonate reservoir prediction techniques are mainly adopted:

(1) Using 3D seismic data to study ancient landforms and ancient water systems.

(2) Using special seismic processing technology to carry out subsequent reservoir prediction research. Extraction of seismic attribute parameters (amplitude, frequency, etc. ), seismic feature calculation (coherence value), seismic inversion (logging constraint inversion), logging inversion (logging inversion using seismic constraint).

(3) Oil and gas identification technologies such as pattern recognition.

Using the above geophysical parameters, histogram analysis, two-dimensional and three-dimensional crossplot analysis and other technologies are introduced to carry out multi-parameter comprehensive analysis and evaluation, which has achieved remarkable results in predicting reservoir distribution and made the success rate of drilling carbonate reservoirs reach over 80%, which proves that the above technical methods have the value of popularization and application.

2 carbonate reservoir prediction research method

Through the analysis of the characteristics and difficulties of carbonate reservoir prediction in Tahe area, as well as long-term exploration and practice, the research flow of carbonate reservoir prediction as shown in figure 1 is determined, and a set of effective technical methods is preliminarily explored. The basic principles and application conditions of various methods will be introduced below.

2. 1 Study on Ancient Geomorphology and Ancient Water System

Figure 1 Block diagram of carbonate reservoir prediction using seismic data 1 Workflow of carbonate reservoir prediction using seismology.

The structural position of Tahe area belongs to Akkule Uplift of Shaya Uplift in northern Tarim Basin, Xinjiang. The paleokarst development period of Ordovician carbonate rocks in Akkule uplift is mainly in the early Hercynian, and some areas overlap with the karst in the late Hercynian.

The development degree of ancient karst is controlled by many factors, including lithology, structure, climate and karst duration. Among them, structure is one of the important external factors that control the development of paleokarst, which mainly shows as follows: ① tectonic background is the basis of paleokarst development; ② The tectonic framework controls the zoning of karst landforms; (3) Faults and fractures are important channels of groundwater and play an important role in controlling the development of ancient karst. Therefore, the study of ancient geomorphology and ancient water system is an important link in Ordovician reservoir prediction.

The Bachu Formation at the bottom of Carboniferous is a filling deposit on the weathering crust of the Middle and Lower Ordovician, and its thickness can indirectly reflect the geomorphological characteristics of the weathering crust karst of the Middle and Lower Ordovician. The bimodal limestone at the top of Carboniferous Bachu Formation is a regional marker bed, representing a relatively stable sedimentary environment. We use seismic data flattening technology to flatten the seismic data volume or the top structural map of Ordovician along the top surface of bimodal limestone. The top structural features of Ordovician after flattening basically represent the geomorphological features of karst development in the early Hercynian period. In this way, karst highlands, karst slopes and karst depressions can be determined according to the characteristics of paleogeomorphology, so as to predict the favorable distribution range of carbonate reservoirs.

The study on the development of ancient water system is another important link in the study of ancient karst. On the basis of fine structural interpretation, the amplitude attribute is adjusted by combining seismic leveling technology with three-dimensional visual interpretation technology, and calibrated by drilling core and logging interpretation data. The development of ancient surface water system and underground water system is directly analyzed by using three-dimensional seismic data volume, and the favorable distribution area of carbonate reservoir is predicted by using the distribution law of ancient water system.

2.2 Geophysical parameter method for reservoir prediction

(1) coherent volume technique

Coherent volume technology uses seismic information to calculate the correlation between traces and highlight irrelevant anomalies. It is generally believed that when the original strata are deposited, the strata are continuous, even if there are changes in the lateral direction, it is a gradual process. So the seismic waves are basically the same in the horizontal direction. There are many factors that affect the correlation between seismic traces, such as noise in seismic data processing, changes in dip angle of strata, lithological changes, faults and fractures in strata, igneous rocks, reefs, salt mounds and mudstone piercing. In the Ordovician carbonate reservoir in Tahe area, if the seismic data is of good quality and the lateral lithology changes little, and the location of faults can be determined by seismic profile interpretation, then the irrelevant factors affecting seismic traces are mainly fractures, dissolved pores and micro-faults, which are the main storage spaces of carbonate rocks. Therefore, the coherent volume technique can be used to predict the development zones of pores, caves and fractures in carbonate rocks. It is clear here that the coherent body is only a macroscopic prediction of the development zones of holes, caves and fractures in carbonate rocks. As for whether holes, holes and cracks are filled with oil, gas and water, or filled with argillaceous, calcareous and siliceous substances, it is impossible to distinguish. Although the filling of these materials will also cause irrelevance between seismic traces, the degree of irrelevance cannot be judged, so other methods are needed.

(2) amplitude extraction technology

There are many factors that affect the amplitude of seismic reflection wave. In addition to the influence of seismic data acquisition and processing, it is assumed that the amplitude is well preserved in seismic processing, and the development degree of holes, caves and fractures affects the amplitude in Ordovician carbonate reservoirs. It is generally believed that the development of holes and fractures in reservoirs will reduce the amplitude, so amplitude extraction technology is also one of the effective means to predict carbonate reservoirs, which can point out the development zones of holes and fractures in carbonate rocks and cannot distinguish their fillings. The amplitude attenuation of different fillers in the hole is different, and whether it can be distinguished depends on the resolution of seismic waves.

(3) Wave impedance inversion

Inversion of wave impedance data from seismic data is an effective means of lithologic interpretation. According to the different constraints of inversion, it can be divided into non-well constrained inversion, single well constrained inversion and multi-well constrained inversion. There are many inversion algorithms, and the factors that affect the inversion results are as follows:

A. The seismic basic data has good data quality (high signal-to-noise ratio, high resolution and high fidelity) and good inversion effect; Otherwise, it will be bad.

B for carbonate reservoir, whether acoustic logging curve can correctly reflect fracture developed zone and undeveloped zone directly affects the result of logging constrained inversion. If the acoustic time difference curve can't reflect the fracture development zone, it is necessary to establish a velocity model through other logging curves, such as lateral resistivity curves, and make forward simulation comparison with known well-side paths to correct the acoustic time difference curve and improve the accuracy and effect of inversion.

C. The extraction and determination of neutron wave in constrained inversion is also a factor that affects the results of wave impedance inversion. The wavelets vary in time and space, especially those determined by multiple wells. Whether one or two wavelets are used in inversion will affect the accuracy of wave impedance inversion results.

D In constrained inversion, the establishment of initial model, namely fine horizon calibration and interpretation, is the basis of good or bad inversion results.

E. How to correct the acoustic logging curve and make high-precision synthetic seismic records is the key to inversion.

F The more wells participate in logging constrained inversion, the more reliable the inversion results are.

Carbonate formation is a kind of high wave impedance formation. When there are wells, caves and fractures in the formation, the wave impedance value will decrease, and the low impedance zone basically reflects the reservoir development zone. It should also be noted that the low wave impedance zone only reflects the development zone of holes and fractures, so it is necessary to judge whether the filling is oil, gas, water or argillaceous, sandy mud and siliceous according to the resolution of wave impedance inversion and the wave impedance difference between the filling and the matrix and surrounding rock.

If the velocity of limestone seismic wave is 6350m/s and the porosity of limestone is 4%, according to the time average equation, when the pores are filled with gas (gas is taken at 340m/s), the velocity will decrease by 37.5%. When the filler is water (v= 1500m/s), the speed will decrease 10%. When the filler is oil (v= 1200m/s), the velocity decreases to 12.7%. When the fill is muddy (v = 4700m/s), the speed decreases by 65438 0.2%. When cracks and holes develop in limestone, the porosity increases, and the corresponding deceleration will also increase.

(4) Logging inversion technology constrained by seismic data in 4)Jason inversion.

In this method, a seismic data volume is interpolated by the seismic trace near the well, the interpolated seismic data volume is compared with the measured seismic data volume, and then the weight coefficient value of each sampling point is changed until the error between the interpolated data volume and the measured data volume meets the accuracy requirement, so as to find a weight coefficient volume. By using the results of known wells and controlling the weight coefficient, various interpolation and extrapolation results such as wave impedance, porosity and water saturation can be obtained. This method is more suitable for the development stage. It usually needs an area of 65,438+000 km2, in which more than 65,438+00 wells are evenly distributed. If the number of wells in the area is small and the distribution is uneven, its accuracy will be greatly affected.

2.3 Neural Network and Pattern Recognition Method for Oil and Gas Prediction

Neural network and pattern recognition are methods to directly detect oil and gas by using seismic data, which are supplements to previous methods. Using the seismic traces beside the known oil and gas wells and dry wells, the characteristic information is extracted, and the discriminant function is established to discriminate the oil and gas content of unknown samples. These methods have achieved good results in the study of clastic reservoirs. Due to the deep burial of the target layer, weak seismic information and serious lateral heterogeneity of the reservoir, a large number of experimental applications have been made in the selection of characteristic parameters, samples and time windows, and some achievements have been made in the direct detection of oil and gas.

2.4 Geophysical parameter analysis technology

Multi-parameter cluster analysis techniques such as histogram, two-dimensional crossplot and three-dimensional crossplot are used to comprehensively analyze and evaluate various geophysical parameters and predict favorable development zones of carbonate reservoirs.

3 application example of carbonate reservoir prediction technology

3. 1 Example of paleogeomorphology research

Using three-dimensional seismic leveling technology, the top surface t of Ordovician in the north work area of pasture is reduced. The map is flattened along the top surface of the bimodal limestone, and the flattened surface is displayed in three dimensions by three-dimensional visualization technology (Figure 2), which basically reflects the paleogeomorphological characteristics of the area during the paleokarst period.

According to the karst landform of Ordovician carbonate rocks in Akkule uplift, the north area of Muchang is located on the residual hill of karst slope zone, and the work area of the north area of Muchang is located on the karst residual hill where Well Ln27 is located. As can be seen from Figure 2, the north work area of the pasture is not a karst residual hill, but a group of karst residual hills. Compare the ancient karst landform in the north of the pasture with the formation pattern of modern peak cluster and peak forest karst landform in Guilin (Figure 3), and they are very similar.

Fig. 2 Three-dimensional display map (time domain) of paleogeomorphology in the three-dimensional work area in northern Ranchang (Fig. 2 Three-dimensional display of fossil landscape).

Fig. 3 Schematic diagram of karst landform formation of Guilin peak cluster and peak forest Fig. 3 Schematic diagram of karst peak cluster and peak forest formation in Guilin, China.

Through the comprehensive analysis of ancient karst landforms in northern pastoral areas, the following conclusions are drawn:

(1) The ancient landform in the north of the pasture is high in the east and low in the west, which is similar to the present structural form.

(2) The ancient landform karst residual hills in the northern grassland can be divided into three categories. The first type, S48 well area (Tahe No.4 Oilfield) is a first-class residual hill; In the second category, the residual hills in well areas Ln27-S66 and S67 are secondary residual hills; In the third category, the residual hills in the northwest of Well Ln27-S66 are Class III residual hills. There are dissolution ditches between the residual hills. The north, northwest and southwest of the work area are all low-lying areas.

(3) Inferred from the distribution orientation of karst caves in this area, there are mainly two groups of fractures (joints) in this area, one group is NE and the other group is NW. This development model may be related to the fact that this area is located at the intersection of Akkumu structural belt and Akkule structural belt.

(4) Through the analysis of the oil and gas drilling results and paleogeomorphic features in this area, it is found that wells S48, T40 1, T402 and TK408 that have been drilled with industrial oil gas flow are all located on the residual hills with high paleogeomorphic features, indicating that wells S66, S67 and S65 that have been drilled are also located in the high part of karst residual hills or on the slopes of karst residual hills.

(5) According to the above paleogeomorphology analysis, it is predicted that the favorable development zones of carbonate reservoirs in this area are:

The first kind of favorable area, the working class remnant hill in S48 well area; The second kind of favorable area, the second level residual hills in Ln27-S66 and S67 well areas; The third kind of favorable area is the third-grade residual hill in the west of Well Ln27-S66.

Well S7 1 located in the second favorable area encountered industrial oil and gas flow, but its productivity was lower than that of Well S48 located in the first favorable area, which proved that our prediction was correct.

3.2 Ancient water system research examples

Based on seismic data and drilling and logging results, the ancient water system of Tahe No.3 and No.4 oil fields in Aixieke work area is studied.

Firstly, the seismic reflection wave along the top of the bimodal limestone leveled the seismic data volume in the three-dimensional work area of Esek. Load the flattened 3D data volume into the 3D visualization software, and then analyze it downward along the flattened bimodal limestone top surface by adjusting the amplitude attribute. Near the top surface of Ordovician, the black amplitude is connected into a tree, which is similar to the distribution of water system (Figure 4); At about 150 m below the top surface of Ordovician, a group of black amplitudes are observed to be connected in a tree shape, which is also similar to the distribution of water system (Figure 5). According to the drilling data, well S64 encountered breccia limestone near the Ordovician weathering surface, which represented the deposition of fracture fillings. Well S6 1 was vented at 1.29 m below the top surface of Ordovician, and encountered a karst cave. Well S64 and well S6 1 are just near the dendritic black amplitude line, so the dendritic amplitude distribution is probably the seismic reflection of the ancient water system.

The development of surface water system near the top of Ordovician system and groundwater system inside Ordovician system are analyzed. The development of ancient water system in Tahe No.3 and No.4 oilfields has the following characteristics:

(1) The paleostructure at the top of Ordovician in this area is high in the north and low in the south, with developed water system and water flow from north to south.

(2) Four groups of water systems are mainly developed in this area. Mainly in the middle of two groups of water systems. Both surface water and groundwater systems are dendritic. Besides the main river, there are many tributaries. Bifurcated rivers gradually converge downstream.

(3) There is a strong correlation between surface water system and groundwater system, and surface water system is more developed than groundwater system.

(4) The water system in the south of this area flows into the main channel, and the tributary channel is undeveloped, which is consistent with the area covered by the Middle and Upper Ordovician in this area. Surface water and groundwater systems are mainly developed in the missing areas of Middle and Upper Ordovician. Near the middle and upper Ordovician pinchout line is the intersection area of seawater and fresh water, which is a favorable development area of mixed karst.

(5) The development of ancient water system is related to the development of three groups of faults and faults in this area.

(6) Comparison of surface water and groundwater systems between Tahe No.3 Oilfield and Tahe No.4 Oilfield; It can be seen that there are more tributaries in Tahe No.4 Oilfield than in Tahe No.3 Oilfield.

(7) From the analysis of the oil and gas output of the completed wells and the development of surface and underground water systems in this area, the wells located at the confluence of multi-branch rivers, such as well S48, have higher oil and gas productivity. There is little or no oil and gas production in underdeveloped areas, such as well TK303.

Fig. 4 Seismic slice near the top of Ordovician in Esek 3D work area (after leveling) (3448ms) Fig. 4 Seismic slice near the top of Ordovician in IXK3D area (after leveling).

Using seismic data to study ancient water system provides basic data for studying ancient karst development. It is considered that the confluence area of tributaries is a favorable area for the development of karst caves, while the area where the main tributaries are not developed is a unfavorable area for the development of reservoirs. This method provides a reliable geological basis for studying the distribution of underground rivers and caves in Tahe oil region and establishing a geological model of oil reservoirs.

3.3 Application effect of geophysical parameter method in reservoir prediction

Application effect of (1) inversion wave impedance

In order to further study the lateral variation law of Ordovician carbonate reservoirs in Aixieke 3D work area, the 3D seismic amplitude-preserving data volume in this area and the data of 10 wells drilled in this work area, such as T40 1, T402, TK405, TK406, S46, S47, T302, TK303 and S6/kloc-0, are used.

Fig. 6 shows the wave impedance inversion profile of the prediction well TK407. Drilling revealed that TK407 oil and gas reservoir is mainly concentrated in 539 1.5 ~ 5478 meters, that is, within 80 meters (about 30 milliseconds) below the weathered surface. The low wave impedance below the weathering surface of Ordovician in the corresponding wave impedance profile 33ms in good agreement.

By analyzing the wells in this area one by one with the above method, the coincidence rate of constrained wells is about 70% ~ 80%, and the success rate of inspection and prediction wells is 60% ~ 76%.

(2) the application effect of amplitude parameters

In order to analyze the amplitude attribute parameters more objectively, we choose the average amplitude in the range of about 50m around the well as the amplitude value near the well. Analyze the relationship between amplitude value and corresponding oil and gas reservoir, and determine the amplitude threshold. By analyzing the average values of parameters in the window of 20ms below the top surface of Ordovician, the threshold values of amplitude in Aisheke work area are 2400, 7000 in Aisheke north area and 7400 in Muchang north area (amplitude is relative value, dimension 1). The drilling of prediction wells TK407 and TK408 proves that the success rate of this parameter prediction is about 68% ~ 80%.

Fig. 5 Seismic slice (3496ms) of Ordovician in Esek 3D work area (after leveling) Fig. 5 Seismic slice of Ordovician in IXK 3D area (after leveling).

Figure Wave Impedance Inversion Profile of Well TK 407.

(3) the application effect of coherence parameters

According to the same analysis method as the amplitude parameter, the thresholds in the time window of about 20ms below the Ordovician top surface are determined as follows: the three-dimensional contrast threshold of Aixieke North is 9.05% (relative value), the threshold of Aixieke Work Area is 33% (relative value), and the threshold of Ranch North Gate is 95% (relative value). These values are different because the data in each workspace is not standardized. The drilling of T4K 07 and TK408 wells proves that the success rate of coherence parameters in each work area is high.

3.4 Application effect of pattern recognition oil and gas prediction method

In order to further judge the oil-bearing property of reservoir space and indirectly judge the filling properties of favorable reservoir distribution areas, the pattern recognition oil and gas prediction of carbonate reservoirs in Tahe area has been carried out, and certain results have been achieved.

Fig. 7s 7 1 abnormal profile of well pattern recognition (fig. 7)s 7 1 abnormal profile of well pattern recognition.

In the north work area of Ranchang, wells S48 and T40 1 are used as oil well samples, and well LN27 is used as dry well samples for pattern recognition. The time window selected for this pattern recognition is 36 ms, and the main characteristic parameters are Berg spectrum, autocorrelation and autoregressive model parameters. As shown in Figure 7, the abnormal profile of oil and gas prediction identified by S7 1 well pattern shows that S7 1 well is expected to encounter industrial oil and gas flow, and the actual drilling proves that the prediction is correct.

3.5 Application of Geophysical Multi-parameter Analysis Technology

Application of (1) histogram analysis technology

Using histogram analysis technology, the obtained seismic parameters along the layer, such as amplitude, coherence value, wave impedance, pattern recognition anomaly and frequency anomaly, are quantitatively analyzed.

Taking the average wave impedance analysis of the window 20ms below the top surface of the North Ordovician in Aixieke as an example, through the histogram analysis of the average wave impedance, the average wave impedance along the layer in this area is mainly concentrated between 1 1000 and 14500 (the values are relative). Through comprehensive analysis of the relationship between oil and gas reservoir and wave impedance in actual drilling, the threshold is determined as 12800, and the wave impedance distribution range below 12800 is expanded on the plane, and the selection range of wave impedance value is adjusted according to the specific situation, so that the coincidence rate of known drilling in the prediction area can reach over 70%. At this time, the predicted distribution range of favorable reservoirs can be extrapolated.

(2) Application of 2)2D crossplot analysis technology

Through two-dimensional intersection, any two geophysical parameters along the bed are analyzed, such as the average coherence value, amplitude and wave impedance in the 20 ms window below the top surface of Ordovician in the three-dimensional work area of northern pasture, so as to analyze the relationship between the parameters and the relationship between the two parameters and the reservoir. By selecting the parameters of weak amplitude and weak coherence, the favorable distribution range of the selected two parameters can be extended to the plane through the transmitting function of RAVE. You can also adjust the range of the two parameters and analyze the plane distribution range of favorable reservoirs.

(3) Application of 3D crossplot analysis technology

Three-dimensional crossplot is a technique that uses three parameters to analyze crossplot. As shown in Figure 8, the three-dimensional cross plot of average coherence, wave impedance and amplitude below the top surface of Ordovician in Aisheke work area is 20 ms, and the black area in the figure is a favorable parameter area with weak coherence, weak amplitude and low wave impedance. Using the function of RAVE, the favorable distribution range of reservoir represented by favorable parameters can be directly expanded on the plane (Figure 9). This method can be used to evaluate and analyze three parameters, determine the distribution of favorable reservoirs and reduce the limitations of a single parameter.

Fig. 8 Three-dimensional cross plot of average coherence, amplitude and wave impedance in Esek's three-dimensional work area (O 1 20ms under the ceiling) Fig. 8 Three-dimensional cross plot of average coherence coefficient, amplitude and impedance in IXK3D area (O 1 20ms under the ceiling).

In a word, the above analysis methods and techniques provide reliable technical means for quantitative analysis and multi-parameter comprehensive analysis of parameters.

4 conclusion

After several years of practice, a set of methods suitable for carbonate reservoir prediction in northern Tarim has been gradually formed, including logging constrained seismic inversion, seismic constrained logging inversion, coherent volume calculation, amplitude extraction and other methods.

Using pattern recognition method to predict oil and gas provides an effective auxiliary technical means for finding oil and gas reservoirs in carbonate reservoir development areas. Using the leveling technique of three-dimensional seismic data, the ancient landform and ancient water system are studied, which lays the foundation for establishing the geological model of karst development in Tahe area.

The geophysical characteristics of favorable carbonate reservoirs are generally characterized by low wave impedance (low velocity), low amplitude, weak correlation and low frequency. Due to the different seismic data in each work area, the threshold is also different. Geophysical model of carbonate reservoir should be established in different regions. Reservoir prediction cannot rely on a single technical method, each method has its adaptability and limitations, and the principle of multi-parameter comprehensive evaluation and analysis must be adhered to. Histogram analysis and crossplot analysis technology provide effective technical means for multi-parameter analysis.

Fig. 9 Favorable reservoir distribution map (O 1 20ms below the top surface) by analyzing the average coherence coefficient, amplitude and wave impedance in the 3D work area of Aixieke. Impedance in IXK 3D region (20 ms below the top of O 1

The above methods and techniques have achieved obvious results in the study of carbonate reservoir prediction in Tahe, but they need to be summarized, improved, perfected and improved in the future work to play a greater role in carbonate oil and gas exploration.

refer to

[1] N.P. James, P.W., Hu et al. Paleokarst. Beijing: Petroleum Industry Press, 1992, 23 ~ 5 1.

Prediction method of carbonate reservoir in Tahe area

Li Zhongjie Han Huang Zhang Xuguang

Planning and Design Institute of Northwest Petroleum Geology Bureau? Urumqi 8300 1 1)

Abstract: This paper analyzes the characteristics and difficulties of carbonate reservoir prediction, and combining with the seismic data in Tahe area of Tarim Basin, puts forward a set of preliminary techniques for carbonate reservoir prediction. The actual results prove the effectiveness of this method. Technical methods include: paleontological landscape research, paleontological hydrological network research, amplitude acquisition, coherent calculation, acoustic impedance inversion, pattern recognition, multi-parameter block diagram, 2D crossplot and 3D crossplot.

Keywords: paleogeomorphology research paleohydrological network research amplitude acquisition coherent calculation acoustic impedance inversion pattern recognition multi-parameter analysis