PETROLEUM EXPLORATION AND DEVELOPMENT Volume 36, Issue 4, August 2009 Online English edition of the Chinese language journal

Cite this article as: PETROL. EXPLOR. DEVELOP., 2009, 36(4): 448–455. RESEARCH PAPER

Application of high-resolution sequence stratigraphy in improving the precision of reservoir prediction
Li Jinsong1,2,*, Zheng Xiaodong2, Gao Zhiyong2, Li Yonggen2, Liu Ying2, Li Xingli2
1. China University of Geosciences, Beijing 100083, China; 2. PetroChina Research Institute of Petroleum Exploration & Development, Beijing 100083, China

Abstract: On the basis of high-resolution sequence stratigraphy, this paper establishes a sequence stratigraphic frame, and predicts reservoirs by analyzing velocity field, seismic inversion, and seismic attributes. In the case study of the Jurassic Toutunhe Formation braided river delta reservoir of Shinan 21 well block, middle of the Junggar Basin, space variant velocity field inversion and reservoir prediction were carried out based on the sequence stratigraphic frame and the reservoir prediction accuracy was improved. Comparing the prediction results with the real drilling shows that, when the burial depth of bed of interest is 2 500 m, the maximum prediction error of depth is only 10 m, and the average relative prediction error is 0.16%; for appraisal wells with an average effective sandstone of 13 m, the maximum prediction error of sandstone thickness is only 5 m. The structure mapping and reservoir prediction constrained by the sequence stratigraphic frame achieved success. Key words: sequence stratigraphic frame; reservoir prediction; Junggar Basin; braided river delta

As the exploration targets for hydrocarbons are changing from structural pools to various types of lithological ones[1-4], the foundation geology has gone through noticeable changes. As represented by high-resolution sequence stratigraphy, sedimentology has progressively become the primary theoretic basis of exploration for lithologic pools; on the basis of the classification and set-up of sequence stratigraphic frame for the study zone, the study on velocity field, seismic inversion, and analysis of seismic attributes would be more precise, consequently improving the reservoir prediction precision in a effective manner. Based on the case study of Jurassic Toutunhe Formation braided river delta reservoir of Shinan 21 well field, middle of Junggar Basin, this paper illustrates the role of high-resolution sequence stratigraphy in improving the precision of reservoir prediction, hopefully to provide new thought in the exploration for lithologic pools in the future[5].

the structural lines, which was favorable for the forming of lithologic traps[6]. Featuring a low-energy braided river lithosomic body, the target formation, i.e., Jurassic Toutunhe Formation (J2t), could be classified (from top to bottom) into 3 sand members: J2t1, J2t2 and J2t3. Concretely, J2t1 and J2t3 are principally of mudstone, while J2t2 features the primary oil pay zone, consisting of dark-gray mediumand fine-grain sandstone, inequigranular sandstone, and a small quantity of arenaceous conglomerate. Overall, J2t2 member, according to the spontaneous potential log, presents the marked feature of “wrap-up of sandstone in mudstone”, where extensive sand develops laterally, indicating a satisfactory condition for the forming of lithologic pools. As demonstrated by drilling, J2t2 reservoir in Shinan 21 well field is principally controlled by lithologic boundary, except for its southern part which is under the control of oil-water boundary.

1

Geological overview of study area

2 Single well classification and comparison of high-resolution sequence stratigraphy
2.1 Single well classification of short base-level cycle In the study area, sequence classification is conducted to each well according to the periodic retrograding-prograding variation as reflected by its log, as well as the sequence interface identification marks, including mudstone color

Perched on the center of Guerban Tonggute Desert in the middle part of Junggar Basin, the Shinan area was dominated by the braided river delta sedimentation in the mid-Jurassic era. Under the control of the NE fault slope break belt, the sand in the branch channels of the delta front extended in strips parallel to or diagonally intersected with

Received date: 15 December 2008; Revised date: 25 April 2009. * Corresponding author. E-mail: ljs@petrochina.com.cn Copyright © 2009, Research Institute of Petroleum Exploration and Development, PetroChina. Published by Elsevier BV. All rights reserved.

SB3) is predominantly grayish fine-grain sandstone. which. / Petroleum Exploration and Development.Well Shi 116 (reflecting the sedimentary environment). and ending at the bottom of SB1. Whereas the seismic response presents successive wave trough reflection.. below the upper interface (i. For the short base-level cycle sequence SSQ1. and. In the short base-level cycle sequence SSQ3. below the upper interface (i. one after another (the uppermost one first). represents a shifting surface between fluvial and delta facies. Further. the lateral seismic follow-up is difficult.. 1 Cross-well seismic interpretation profile of high-resolution sequence stratigraphy for Well Shi 108 ..2 Set-up of isochronal stratigraphic frame for short base-level cycle Based on the single well classification of short base-level cycle for the study area. its overall logging response indicates a transitional surface between rhythm and counter-rhythm. while above it is grayish fine-grain sandstone. the upper interface SB4 is a non-conformity surface between Qingshuihe Formation and Toutunhe Formation. there is one set of variegated reddish mudstone. SSQ1 . starting from the mid-term flooding surface developed within Toutunhe Formation. mixed with dark-gray. 2009. below which is the prograding superimposition pattern and above which is the retrograding pattern. SSQ1 through SSQ4 from bottom to top. the top of mudstone layer represents the flooding surface of the medium base-level cycle (the fourth order sequence). Shi 106 and Shi 109 wells are represented by the dark-mudstone sedimentation of flood-plain (Fig. . grayish-black mudstone with non-uniform thickness. featuring a lithologic sudden-change surface.e. abrupt change and contact of the log (corresponding to the left-over overburden sediments)[7-10]. nearly extending throughout the entire zone. Irrespective of being above or below the interface. SB2) are pelitic siltstone. and black coal seam. concretely. at which time the sedimentary sand was mainly represented by channel sand and channel-bar sand. irrespective of being above or below the interface. Shinan 21. 4 such sequences (i. isochronal comparison is carried out. in other wells. geographically corresponding to the Yanshan episode 1 tectonic movement. only 3 ones (i. In the sedimentation stage of SSQ3. Fig. there is one combination of sandstone and conglomerate. the lower interface (i. 36(4): 448–455 Fig. it is believed that.. where fine siltstone dominates. channel scour surface. and the seismic response is principally represented by parallel reflection..Li Jinsong et al. Shi 108.e. 1 provides the joint seismic interpretation profile of high-resolution sequence stratigraphy for Well Shi 108 – Well Shinan 21 – Well Shi 106 – Well Shi 113 – Well Shi 121 – Well Shi 116. 3 . calcareous mudstone. In the time span of sedimentation of the short base-level cycle SSQ2. due to which the braided branch channels merely developed in Shinan 3 well field.4 short base-level cycle sequences (the fifth order sequence) may be divided. 2.e. while the seismic response is mainly represented by conformity or the corresponding parallel unconformities.e. while above it are mainly the brownish pelitic siltstone and grayish siltstone. On the well-connecting line along the WE structural strike. In the short base-level cycle sequence SSQ2. SB1) features an unconformity surface between Toutunhe Formation and Xishanyao Formation. Shi 112 and Shinan 3 wells feature the sedimentation of channels in braided river delta. the braided branch channels in Shinan 21 well field changed its route towards Shinan 3 well field. the lowest one first) is conducted to the short cycle sequences falling within the semi falling cycle. above it.e. the branch channels of Shinan 21 well field attained the climax of its great prosperity.SSQ3) are available. Below such an interface. geographically corresponding to Yanshan episode 2 tectonic movement. as for Well Shinan 3 located in the east of the well field. features a long-term erosion surface. due to the high value of gamma ray logging inside Toutunhe Formation. Ji 004. to the short cycle sequences falling within the semi rising cycle of the medium base-level. 2). the seismic response is represented by angular-unconformity contact relation. namely. In the same sequence SSQ1.. likewise. comparison in the opposite direction (i. which. starting from the flooding surface and ending at the ceiling of SB4 (or SB5 in the case of Well Shinan 3). the spreading of braided channels inside the study area was quite limited in the sedimentation stage of the short base-level cycle SSQ1. and. SSQ1 SSQ4) could be observed. while Ji 007.e.

On the well-connecting line along the SN structural strike. Shi 114 and Shi 107 well fields were dominated by the brownish mudstone featuring the oxidation environment of flood-plain.Li Jinsong et al. During the sedimentation of SSQ3. Shinan 21 and Shi 107 wells were apparently represented by the extensive sedimentation of dark-. 2 Illustration of isochronal stratigraphic frame of Toutunhe Formation for Ji 004 . similarly. On the basis of well-connecting isochronal stratigraphic . the braided branch channels of braided-river delta plain nearly extended across the entire zone. together with the lateral flood-plain dark-mudstone. 2009. SSQ4 was retained only in Shinan 3 well field.Shi 107 wells while the rest wells were principally represented by the dark-mudstone sedimentation of inter-braided channel flood-plain. the intra-channel sand reservoirs would also constitute lithologic traps. brownish mudstone of the flood-plain facies. In such a stage. while the remaining well fields by the dark mudstone featuring the weak reducing environment (Fig. could form the sand lateral pinch-out lithologic oil/gas pools. being of certain heterogeneity. 3 Illustration of isochronal stratigraphic frame of Toutunhe Formation for Shi 114 . which. / Petroleum Exploration and Development. During the sedimentation of the short base-level cycle SSQ2. which. and the dark-mudstone sedimentation of inter-braided channel flood-plain developed in a handful of well fields (including Shinan 22.Shinan 3 wells Fig. Shi 114 through Shi 107 wells were represented by the sedimentation of braided-river delta flood-plain in the sedimentary time span of the short base-level cycle SSQ1. 36(4): 448–455 Fig. Shi 114. was represented by the sedimentation of braided branch channel. 3). and Shi 107) in the northern part of Shinan 21 well field. the sedimentary sand was mainly represented by the channel sand and the channel-bar sand.

evidently. SB3. in a more objective manner. 36(4): 448–455 Fig.. 2009. the structural patterns are of high consistency. with the constraint of sequence stratigraphic frame. the average velocity field could overcome partial impacts arising from the seismic data itself and artificial interpretation. respectively. It indicates that. Figs. relative to the reservoir prediction based on the conventional horizon interpretation (SB1 and SB4 to be interpreted only). while the random noise would exist when no such constraint is imposed. 4a (constraint of sequence frame incorporated) and b (no constraint of sequence frame incorporated) show the average velocity field of SB3. / Petroleum Exploration and Development. and the 3D sequence stratigraphic frame has been thus set up. 3 Structural variable speed imaging technique under the constraint of sequence frame Fig. SB1. 5 Structural map for SB3 after well depth correction conducted to 9 wells Fig. 4 Average velocity field illustration for SB3. After the structural graph is plotted for SB3 on the basis of . the average velocity field as constrained by sequence frame presents a more marked regularity. full-zone tracking has been winded up (280 km2) for 3 sequences (i. and SB4). as resulted from the inversion based on velocity spectrum data[11]. Due to the isochronal sedimentation within each sequence. any sedimentation body falling between SB1 and SB3 or between SB3 and SB4 may be studied separately via seismic techniques on the basis of high-resolution sequence stratigraphic researches. as indicated therein.e. the control of high-resolution sequence stratigraphic frame could effectively improve the precision in reservoir prediction. in addition to improving the precision of structural variable speed imaging.Li Jinsong et al. 5 and 6 are the structural graphs after depth correction is conducted to 9 wells and 20 wells. with the sequence frame constraint incorporated (a) and not incorporated (b) Fig. 6 Structural map for SB3 after well depth correction conducted to 20 wells frame. the velocity field characteristics of the target interval. and by leveraging the 3D seismic interpretation. consequently reflecting.

Shi 106.254 7.230 2 443. so as to get the predicted depth and prediction error of each of the other 8.1. Ji 007.000 2 504. 8 shows the logging constraint inversion profile through Well Shinan 21 plotted on the basis of conventional horizon interpretation results.0.0. so that the seismic attribute analysis could be conducted specifically to a certain small sedimentary body. That is. with the relative prediction error averaged at 0. Shinan 21. which is resulted from seismic inversion under the constraint of . the maximum error in predicted depth is merely 10 m (target formation: about 2 500 m in depth).000 2 435. and 18 appraisal wells respectively.590 2 504.110 .836 .101 .000 2 502. and 2 wells (see Table 1).500 2 458.2.150 2 451. the initial geological model approaches much closer to the actual sand spreading (both transverse and longitudinal). the drilling history is simulated for the exploration appraisal wells. with the latter presenting a more clear-cut depiction of the reservoir lateral variation particulars.370 2 429.510 4. Fig.950 2 429.053 . thus improving the reservoir prediction precision in the application of seismic inversion. well-site depth correction is carried out to the 9 appraisal wells in the first batch of drilling (namely.657 5.243 .000 2 581.810 2 506.500 2 430.194 3. a higher precision would be attained for the structural variable speed imaging on the basis of average velocity field resulted from the inversion of velocity spectrum data under the constraint of sequence frame.3. and Shi 109 wells). the overall reservoir development characteristics are consistent.670 2 451. the set-up of high-resolution sequence stratigraphic frame could differentiate relatively small sedimentary bodies. which could be compared with the counterpart introducing high-resolution sequence stratigraphic frame (Fig.450 2 494.840 2 438. / Petroleum Exploration and Development. thereafter.131 .470 2 494.550 2 430.560 2 439. the well-site depths of the other 11 wells are fetched for the comparison against the penetration depth of each well. As indicated by the figure. 2009.704 .0.457 8.550 0. the classification result of seismic waveforms between conventional horizon interpretation interfaces is nearly equivalent to the superimposition of the classification result between SB3 and SB4 interfaces with that between SB1 and SB3 interfaces. Shinan 22. According to the two inversion profiles.140 9.860 2 492.718 .4.500 2 457. According to Table 1. 10 and 11 show the thickness of sandstone. it shall be noted that the formation between sequence interfaces of SB1 and SB3 includes the target reservoir.780 Depth fetched from structural map / m 12 wells 15 wells 18 wells 9 wells 4. Ji 004.843 0.200 2 457.280 2 506. After the introduction of the high-resolution sequence stratigraphic frame.000 2 442.217 Prediction error / m 12 wells 15 wells 18 wells 2 446.469 7. 5. so as to acquire the predicted depth and prediction error of each well.150 2 503. Shi 107.000 2 510.0. 15. Shinan 3.Li Jinsong et al.4. Shi 108.878 0.773 .820 2 427. such a process is then iterated for 12. firstly.830 2 505.250 2 573. 7 Waveform classification of formations between sequence interfaces the velocity field resulted from sequence frame constraining.0.354 6.860 2 504.411 . 7 shows the comparison between the waveform classification graphs on the basis of conventional horizon interpretation and the waveform classification graphs of inter-sequence interface seismic response.610 2 459.760 2 507. as a result.000 2 504.3.321 2.990 2 453. That is. 9). the reservoir prediction precision could be improved in the application of seismic attribute analysis technique.16%.169 4.340 2 504.000 9 wells 2 526. 36(4): 448–455 Table 1 Depth prediction and error for each non-drilled well from the structural map in various rounds Well ID Shi 112 Shi 113 Shi 114 Shi 115 Shi 116 Shi 117 Shi 118 Shi 119 Shi 120 Shi 121 Shi 122 Penetration depth / m 2 531.106 Fig.829 6. Figs.560 . 4 Reservoir prediction techniques under constraint of sequence frame[12-14] Fig.

carrying out logging restrained inversion to the 9 appraisal wells of the first batch of drilling (namely. Ji 007. and 2 wells. Shi 106l. Shinan 22. effective improvement . even when the number of wells used for the prediction varies largely. According to the table. Obviously. still attain the prediction precision equivalent to the case with a bigger number of known wells. no matter whether 9 wells or 12 wells are used for the prediction. 15. the maximum prediction error in sandstone thickness is about 5 m. Ji 004. 8 Conventional horizon interpretation results through Well Shinan 21 Fig. seismic inversion could. The drilling history of the exploration and appraisal wells is simulated as follows: firstly. then. / Petroleum Exploration and Development. respectively. with a less number of known wells. Evidently. for 9 wells and 20 wells.44%. On the resultant sandstone thickness maps. whereas the high-resolution sequence stratigraphic frame has established a rational sedimentation comparison relation. Shinan 21. the average relative error of the sample set arrives at 26. indicating that.87%. as constrained by sequence frame sequence stratigraphic frame. Provided that the Shi 118. Shi 107. and Shi 109 wells). such a process is iterated for the cases of 12. Shi 119. difference exists with the thickness of sand bodies. the predicted sandstone thickness for the other 11 wells is fetched. and 18 appraisal wells. 36(4): 448–455 Fig. Shi 120. given the target layer with the burial depth of about 2 500 m and the appraisal well with effective sandstone thickness averaged at 13 m. so as to obtain the prediction errors of the other 8 wells. Shi 108. 9 Seismic inversion profile through Well Shinan 21. 5 wells. 2009. Shinan 3. which is compared with the actual value of each well to calculate the prediction error concerning the sandstone thickness for respective well (Table 2). In other words. and Shi 121 wells are regarded as a sample set.Li Jinsong et al. respectively. however. with the sequence frame being imposed. the average relative error is 28. the improvement in prediction precision is not evident. the sand patterns in the two figures merely present low inconsistency. when 15 wells are used for the prediction.

000 15.000 2. Shinan 21 well field.000 0.900 1.700 0.200 1.Li Jinsong et al.000 15. / Petroleum Exploration and Development.000 14.000 5.000 18.000 13.000 11.000 19. since the structural variable .300 6. 10 Sandstone thickness predicted by seismic inversion with 9 wells constrained by sequence frame Table 2 Depth prediction and error for each non-drilled well from the structural map in various rounds Well Name Shi 112 Shi 113 Shi 114 Shi 115 Shi 116 Shi 117 Shi 118 Shi 119 Shi 120 Shi 121 Actual thickness / m 21.200 5. the set-up of high-resolution sequence stratigraphic frame would play a significant role in improving the precision in seismic prediction.000 14.300 0.000 12.500 4.500 9 wells Predicted Prediction error thickness / m /m 18.300 3.300 6. as well as the analysis on error.000 18.500 15.000 20.400 1.000 1.200 1.200 could be seen with the precision of both seismic attribute analysis and seismic inversion. Junggar Basin.000 0.300 16.000 2.700 -1. 36(4): 448–455 Fig.400 0.000 1.000 19.000 5.000 -1.000 2.300 1.000 3.000 14.200 -1.700 6. 2009.000 12.200 -1.000 5.200 -1.000 5.000 12. 5 Conclusions As indicated by its application in the braided-river sedimentary reservoir of Jurassic Toutunhe Formation.300 12 wells Predicted Prediction error thickness / m /m 15 wells Predicted Prediction error thickness / m /m 18 wells Predicted Prediction error thickness / m /m 1.000 14.100 12.200 1.

could be highly effective in achieving this goal. 1984. Nevertheless. Zhuang Wenshan. The study area elaborated in this paper concerns a cluster of exploration and appraisal wells (up to 20 wells). Petroleum Exploration and Development.Li Jinsong et al. [7] Cross T A. Bao Zhidong. 31(3): 3-9. Jia Chengzao. Beijing: Petroleum Industry Press. 35(1): 17-22. 1998. [5] Jia Chengzao. Li Benliang. / Petroleum Exploration and Development. Whereas the objects of hydrocarbon exploration become increasingly subtle and complicated. Hou Lianhua. Beijing: Geological Publishing House. Sediment volume partitioning: References [1] Hu Jianyi. 2008. Zou Caineng. Subtle oil reservoirs in China. Zhao Wenzhi. The symposium of subtle oil reservoirs exploration in China. Petroleum Exploration and Development. and abundant data are available for the classification of short base-level cycles. Harbin: Heilongjiang Science and Technology Press. Chi Yingliu. et al. NW China: A case from the Shawan Formation in Well Pai 2. Lessenger M A. the scheme provided herein could also be followed by the lithologic reservoir exploration in the zones with small number of wells. Petroleum Exploration and Development. New hydrocarbon exploration areas in footwall covered structures in northwestern margin of Junggar Basin. a better answer for improving reservoir prediction precision would inevitably rest with the combination of sequence stratigraphy with the reservoir prediction techniques. 2004. together with the reservoir prediction technique under such constraint. 2009. . The symposium of the 3th subtle oil reservoirs international conference. 36(4): 448–455 Fig. 2008. 2004. 11 Sandstone thickness predicted by seismic inversion with 20 wells constrained by sequence frame speed imaging under the constraint of sequence frame. 35(2): 195-200. et al. Petroleum geology and exploration of beach sandbar in Junggar Basin. [6] Guan Shuwei. [3] [4] Pan Yuanlin. Two key technologies about exploration of stratigraphic/lithological reservoirs. [2] Xiang Kui.

Wang Hongliang. Application and restriction of seismic inversion. 16(2):89-97. Petroleum Exploration and Development. 35(4): 397-409. . 2000. Gradstein F M. Doligez B.Li Jinsong et al. [9] Deng Hongwen. Oil & Gas Geology. 2003. et al. Baker M R. 1998. High resolution sequence stratigraphy: Theory and application. Milton N J. A new school of thought in sequence stratigraphic studies in U. 27(2): 53-56. Paris: The 7th IFP Exploration and Production Research Conference. 2009. Sequence stratigraphy concepts and applications. Petroleum Exploration and Development. [13] Liu Wenlin. 1995. Oslo: NPF Special Publication. [12] Li Qingzhong. Applications of high-resolution sequence. Application of multi-component seismic exploration in the exploration and production of lithologic gas reservoirs. [11] Zhao Bangliu. The seismic technology of oilfield development. The road to precision exploration. [14] Yao Pengchang. Beijing: Geological Publishing House. Beijing: Petroleum Industry Press.: High resolution sequence stratigraphy. Eschard R. 171-195.stratigraphy to reservoir analysis. 36(4): 448–455 Rational for stratigraphic model evaluation and high resolution stratigraphic correlation. Subsurface reservoir characterization from outcrop observations. 2008. 2002. 423. Sandvik K O. [8] Cross T A. [10] Deng Hongwen. Gan Lideng. 1996. 1993.S. / Petroleum Exploration and Development. Beijing: Petroleum Industry Press.

Sign up to vote on this title
UsefulNot useful