The Gas Resources Assessment Expert System of the Songliao Basin

666Vol. 74 No. 3CTA GEOLOGICA SINICASept. 2000The Gas Resources Assessment Expert Systemof the Songliao basinLONG Shengxiang, WANG ChuanResearch Institute of Petroleum Geology, China National StarPetroleum Corportation CNSPC, 3/ Xueyuan Rd, Beijing 100083andi YuzhanExploration Department, CNSPC, 3/ Xueyuan Rd, Beijing 100083Abstract The gas resources assessment expert system is one of the advanced methods for appraising oil and gasresources. The establishment of a knowledge base is the focal task in developing the expert system. This paper pre-sents a summary of the mechanism and the major controlling factors in the formation of gas pools in the southeastuplift of the Songliao basin. Then an appropriate assessment model is established for trapping the gas resources andknowledge base built in the expert system to realize the model. By using the expert system to appraise the gasbearing probability of 25 major traps of the Quantou and denglouku Formations in the Shiwu-Dehui area, the authorshave proved that the expert system is suitable for appraising traps in the Songliao basin and similar basinsKey words: gas resources, assessment model, expert system, trap, pool-forming mechanism1 Introductionlimitation of the assessment model. In recent yearsmany experts have realized that an expert system canThe expert system, one of the three main branches of analyze all the geological conditions for formingartificial intelligence, is a computer software which pools based on the specific pool-forming mechanismuses the experts knowledge to solve some kinds of and greatly improve the reliability of the assessmentprofessional problems. Since the first expert system results. Up to the present there are 6 expert systemsDENDRAL was set up at Stanford University by the the expert system for classifying sedimentary basinsend of the 1960s, the expert systems have quickly the knowledge base for predicting source rocks, thecome into many fields, e. g. biology, medicine, chemi- PLAYMAKER expert system, the expert system forcal industry education. layt,engineering, researching exploration of the San Juan basin, thebusiness, military affairs, meteorology and geology, expert system PRES(Wu et al., 1987; Wang Weiyuanand displayed a brilliant prospect of applications and Wang Yuwen, 1989) and the gas resources a-( Denham, 1985; Krystinik and Clifton, 1985; Charles sessment expert system (Li et al., 1994; Wang Chuanand Stout, 1985; Liu, 1991). In China, the expert sys- et al., 1996), to be used directly or indirectly in petroems have been designed and applied in oil trap leum resources assessment. The above-mentioned gasevaluation(Tong Zhaoqi et al., 1993), hydrogeology resources assessment expert system is the first perfect(Wu Heng, 1994), crustal stability evaluation (Yin expert system aimed at gas resource. Based on theYueping et al., 1996), mineral deposit prognosis (Xiao experiences gained in partaking in the developmentShujian et al., 1997)and engineering(Yan Changhong and application of the system, combined with re-etal,1999)searches on basin analysis, petroleum system analysisetroleum resources assessment is a complex sys- and pool-forming mechanisms, the authors have etems engineering. Traditional statistical methods and tablished an assessment model for the gas resources inbasin modelling are often unable to integrate all the traps of the Songliao basin and used it to appraise thegeological factors to give reliable results owing to the traps in the shiy中国煤化工CNMHGThe Gas Resources Assessment Expert System of the Songliao BasinLor2 Briefing on the Gas Resources Assessment Expert systemInterfaceThe system can be divided into the main body and theAcquiring knowledgeInterpretationAs shown in Fig. I the main body consists of aknowledge base, a module for upholding the know l-edge base, a module for acquiring knowledge, a com-prehensive data base, an inference machine and anKnowledge basData baseInfereinterpretation module. The knowledge base is composed of five parts:(1)A rule base which restoresrules (or expert s experience). These rules are indicat-ed by sidelines in which the nodes are geological Fig. 1. Framework of the main body of the expert systemevents and the sidelines are relationships between thegeological events. All the sidelines are linked to formThe peripheral part consists of programmes fora complex network (Fig. 2).(2)An original knowl- calculating the amount of resources, the amount of gasdge base. which restorknowledge controlling in different phases, the potential energy of fluids, thethe operation of the system. (3)a data base restoring amount of gas supply and the gas column height,he static parameters and data and the dynamic data in which serve to supplement and perfect the assessmentrunning the system. (4)A dictionary containing names methods and provide parameters for improvement ofand codes of all the elements in the rule base. (5)a the assessment accuracycalculation programme package, which is used tocalculate parameters such as the content of cement, 3 The Model for Assessing Gas Resourcesthe intensity of generation and expulsion of hydrocar- in Traps of the Songliao Basinbons etc to classify the organic matter etc. and tocalculate the confidence. The module for upholding In petroleum resources assessment the most difficultthe knowledge base and the one for acquiring knowl- task is to analyze the petroleum-bearing probability ofdge are used to establish, expand and automatically the prospects(areas or traps). Establishing a gas prob-dmodify the knowledge base, and to check the rules ability assessment model and a corresponding knowl-that remain to be confirmed. The comprehensive data edge base is a bottle-neck problem in establishing thebase restores the information needed for running the system. Different assessment models must be estabsystem. The inference machine uses uncertain infer- lished for different basins because they have differentence models to fit the mechanism of forming gas geological conditions and gas pool-forming mechapools. It infers mainly forward and secondarilybackward. similar to the thinking models of himan experts, and calculate and judge the validityEnd nodeof the middle or end nodes(or conclusions) from8he starting or other middle nodes(or presupposeMiddle nodeMiddle node 2to ask questions and answers the users questions,arting node IStarting node 2Starting node 3and waits for the users to add new data. Of themhe knowledge base. the inference machine andthe interpretation module are the core of the sys- Fig. 2. Schematic diagram showing relationships between the nodesThe figures denote the confidence of the sidelinesV-or, A-andtem, which combine to realize the interpretation中国煤化工CNMHGVol. 74 NoCTA GEOLOGICA SINICASept. 2000nisms. Based on the results of basin analyses, petro- late in the fault-anticline traps, fault-block traps, faultleum system analyses and pool-forming mechanism nose traps of structures formed at the end of the denstudies, this paper presents a summary of the mecha- glouku period as well as in their overlying drape foldolling factors of pool fcnd traps, updipping-pinch out sandstone traps and lenestablishes a model for assessing the gas potential in ticular sandstone traps. At the end of the Nenjiangtraps of the Songliao basinperiod many inversion structures and compressiveanticlines formed and tectonic activities, especially3.1 Mechanism of pool formationfault activities, on the one hand, caused the gas toWang Pujun et al. (1995), Fu Guang et al.(1998), Liu readjust and accumulate in the Quantou FormationDelai and Ma li( 1998), Lou Zhanghua et al.(1998, Denglouku Formation; and on the other hand, promot1999)have studied the stratigraphy and others of the ed the gas at depths of the basin to vertically migratSongliao basin. According to the results of dissecting and accumulate in various traps of the Quantou Forthe Houwujiahu and Xiaowujiazi gas fields and basin mation-Denglouku Formation, forming primary oranalysis and petroleum system analysis, the authors secondary gas pools. After the Nenjiang period ero-consider that the main mechanisms of forming gas sion destroyed the caprocks of the Yaojia and Nenpools in the southeast uplift of the songliao basin can jiang Formations and even the first member of thebe described as follows. The source rocks are black Qingshankou Formation in some areas, which worsmudstone and coal of the Shahezi Formation- ened the preservation conditions so that shallow gasYingcheng Formation at depths of the basin. The re- pools were destroyed by diffusing, leaking and peneservoir rocks are developed in the upper Denglouku trating of fresh water. At the same time, the gas in theFormation-Quantou Formation and the caprocks are formation water which was released because of lowdeveloped at the top of the Denglouku Formation, the ering of temperature and pressure and the gas migratupper part of each member of the Quantou Formation ing upwards from depths of the basin replenished theand the first member of the Qingshankou Formation. existing gas pools or formed new gas pools. As theThese source rocks, reservoir rocks and caprocks are replenishment of gas did not make up for its evaporavertically united to form a combination which spans tion, the gas pools are often underpressureddifferentTo sum up, the major controlling factors in the for-glouku Formation a large quantity of oil and gas was mation and preservation of gas pools include: (1) Thegenerated in the source rocks and migrated laterally development degree and type of the traps and thethrough permeable beds to the steep slope zones and match between the traps' formation time and the mainthe gentle slope zones of the fault basins. Part of the petroleum migration and accumulation period. generoil and gas migrated vertically along the faults into the ally speaking, traps are mainly developed in stronDenglouku Formation and soon lost by leaking and inversion or compressional structures, most of whichdiffusing, the remaining retained in permeable beds of are fault-anticline traps, fault-block traps and fault-Kish-Kiyc. By the end of the Denglouku period a few nose traps related to faults. Their formation time coininversion structures were formed. As a result, oil and cides with the migration time of plenty of gas, whichgas accumulated in K sh-Kyc to form primary oil and is favourable for the formation of gas pools. (2) Thegas pools(mainly oil pools) on the one hand, and on relationship between the source area and the trapsthe other hand the active faults caused large quantities The traps within or near the main source area are mostof oil and gas to migrate vertically and evaporate to favourable for forming gas pools. (3) The relationshipthe earths surface. During the Quantou-Nenjiang between the traps and the migration direction of theperiod some good caprocks were deposited, which gas. Gentle slopes of fault basins are the major destigreatly improved the preservation condition. At thattime a large quantity of gas was generated in the sour- advantageous pool-forming conditions than the otherce rocks and migrated vertically along the faults, un- traps. (4) Faults are the most important passages foconformity interfaces and permeable beds to accumugas migrationTYHEal he gas pools中国煤化工CNMHGThe Gas Resources Assessment Expert System of the Songliao BasinLor669and gas-bearing structures had early normal faults judge if the traps can accumulate gas and form poolinverted later on) developed in them, which (3) Gas supply conditions of early oil and gas pools inenabled the gas to migrate laterally into the traps of the depth: The model first makes a judgment if thethe Quantou Formation-Denglouku Formation. (5) early oil and gas pools in the depths supplied gas toThe sealing of caprocks and their destruction by ero- the shallow parts and then analyzes the supply modesion are the key to preservation of gas.(6) TectonicReservoir conditionsactivities. which resulted in reformation or even dThe first step is to analyze the development of restruction of the primary gas pools and regeneration of servoir rocks based on the sedimentary facies andsecondary gas pmeasured data. The second step is to appraise the reservoir rocks according to the composition of theent modelrocks, the composition and content of the cements, theAn assessment model consists of direct identification type of cementation, the grain size and characteristicsand pool-forming condition analysis, and the assess- of structures. And the third step is to determine thement is accordingly divided into two partsdevelopment of pores and crevices from variations of3.2.1 Direct identificationthe diagenetic environment, type of diagenesis andDirect indicators(or shows)are picked up from the tectonic position; to determine the porosity and per-drilling data, log data, seismic data and geochemical meability conditions and classify the reservoir typesexploration data to determine the petroleum potential according to the measured data or regional correlaof the traps. Relatively speaking, drilling data have a tions; and finally to evaluate the pore textures on thehigh reliability, of which encountering gas flows is the basis of the explusion pressure, the median pressure,most reliable. next are oil sands and oil contained in the ratio of minimum unsaturated pore volume and thecrevices, whereas anomalies measured in gas logging radius of maximum linking-up pore throatand engineer have lower reliabilities Shows in wellTrapping conditionslogging also have a high reliability, so the system usesFirst, the model classifies the traps and analyzesthe log data to calculate the saturation of water and their macroscopic conditions; second, it infers the sizeidentify the gas-bearing beds on the basis of an inte- of the traps based on their area, closure and height ofgration of multiple logging data. Reliable indicators the hydrocarbon column; and third, studies if the formay also be found from seismic data, e.g., velocity mation time of the traps is favourable for forminganomalies, frequency variations, phrase variations and pobright-spots. However, these shows have multipleAccumulation conditionsinterpretations, therefore the system makes a compre-The analysis of the accumulation conditions ishensive study of the various data. Geochemical data most important part of the model, the dominant icare used to determine the gas potential of the traps, but of which are to make a time-space analysis of thethe results are not very reliablerelationships among the various geological elements3.2.2 Analysis of the formation conditions of gas and geological processes(1) Types of source-reservoir association: on theGas source conditionsbasis of comparisons of oil and gas with the sourceThe analysis deals with the following three cases: rocks the model analyzes the spatial relationship(1)Self-generating gas. The model firstly determines between reservoir rocks and source rocks, and dividesthe expulsion intensity of gas and identifies the gas them into four typessupply conditions, then analyzes if the maturity of the(2) Modes of migration: the main mode of migsource rock, the main gas expulsion period, the pre- tion is vertical migration or lateral migration or bothsent gas expulsion state and phase are favourable for The analysis of the lateral migration includes(a)forming gas pools. (2) Gas supply conditions of the studying the geochemical signatures;(b) emphasizingmain source area: It mainly analyzes the spatial relathe position of traps in a fault basin and the character-tionship between traps and the main source area toistics of the flu中国煤化工CNMHG670Vol. 74 NoCTA GEOLOGICA SINICAept.2000tion distances; and(c) studying the migration passages. certain influence on the preservation conditions. OfIn the analysis of the vertical migration, the model them the normal faults have the greatest influence, thelays emphasis on the role of the faults as passages, and strike-Slip faults come next, and the reversed faultsit mainly studies the size of faults and the match be- have the slightest influence. The more active the faultstween the active period and the pool formation period. the greater the influence. When the throw of a fault isAt the same time, it also studies the role of unconfor- larger than the converted thickness of the caprock itmity interfaces and fissures on hydrocarbon migration. would strongly destroy the preservation condition, and(3)Phases of migration: The migration and accu- vice versa. Moreover, when the fault plane is filled upmulation of oil or free gas are relatively simple, but with fault clay or bitumen the fault would have a verythe accumulation of water-dissolved gas requires a good sealing capability. But if there is no filling, consubstantial fall in temperature and pressure so that gas tact of mudstones of both walls will be favourable forty be dissolved from water. Therefore, a judgment petroleum preservation, contact of mudstone withindex on the migration of water-dissolved gas and a sandstone has a certain possibility of sealing, butcriterion for classification of regional tectonic uplifts contact of sandstones of both walls is very unfavorare set up in the modelable for petroleum preservation. Finally, the model(4) Time of migration and accumulation: The model studies the influence of faults on the preservation conmainly considers if the match between the migration ditions based on the location of the faults in oil andand accumulation of oil and gas and the formation of gas reservoirstraps is favourable for forming poolIn considering other influences of the post-poolPreservation conditionsformation tectonic activity the model suggests that(1)Caprocks: The model presents a macroscopical subsidence or slight uplift of the plate will not haveanalysis of the positional relation of caprocks and destructive influences on the preservation of gas poolreservoir rocks, the thickness and lateral variation of but a substantial and drastic uplift may result in erocaprocks, the influence of faults and erosion on the sion of the overburden rock so that the gas pools willcontinuity of caprocks, the covering of caprocks over be destroyed because they are exposed on the earthsthe traps, the absolute sealing capability of caprocks surface or located in the atmosphere-fresh water aon the basis of the breakthrough pressure, the domi- tion zone. Secondly, it suggests that the lateral movenant pore radius, the breakthrough time and the diffu- ment may cause the original traps to change theirsion coefficient the relatiapability of forms and vanishcaprocks from the difference between breakthroughpressure and median pressure; the evolution of 3.3 Realization of the assessment model in thecaprocks and the formation time of effective caprocks expert systemin view of the diagenesis etc.According to the production rule and framework, the(2)Hydrogeology: The model examines the types various rules in the above-mentioned assessmentand mineralization degrees of the groundwater, ana- model are linked together by the characteristic tablelyzes if it is in a reduced environment and the location and 6 logical relations, namely," and,, "or,"exchof traps in relation to the groundwaterSive or”,"¨ weighted", new and'and¨ independent,(3)Succeeding reformation: It is to study if the gas and a corresponding rule base is established in thepools are seriously destroyed after their formation and expert system. The base consists of a general modulethe evaporation of oil and gas, and to study if oil and and 6 submodules, which are the gas-bearing indic agas are continually provided to form dynamic pools. tion, source condition, reservoir condition, trapping(4)Tectonic conditions: Emphasis is laid on ana- condition, accumulation condition, preservation corlyzing the sealing of the faults. If the early faults are ditionnot active after the pool formation period, they usuallyThe assessment by the expert system is virtuallyhave good sealing, but if the faults continue to be a- conducting inference and calculation according to thetive after the formation of pools they would have a assessment mole base. the中国煤化工CNMHGThe Gas Resources Assessment Expert System of the Songliao BasinLor671main logic relationships for the inference are as fol-If there are some different relationships among allthe e, which are presuppositions of the conclusion H(1)The case with only one presuppositionthen calculation nodes(imaginary nodes)are used toSuppose h is the conclusion and e is the presuppo- reflect the priority of all the relationshipssition, if the confidence of E, cf(E)and the confidencefrom e to h. a and b are known then4 Application of the Gas Resources AsO cf(H)=acf(E)when cf(E)>osessment Expert System(2 cf(H)=bcf()when cf(E)∑w;cf(Hmeasured data and analytical data. The assignments ofwhere wi is the weight. In a special case, all the parameters are traditional methods, and thecf(H)>>W;cf (H) is stipulated to show the ""sudden following is a brief introduction to the assignment ofchangethe confidence4 If the relationships among all the E s are"exch-The assignment of the confidence values mainlysive or, then one of the e must satisfy cf(H)>t >0(t depends on the analysis of the reliability of the data.is the threshold value) and cf(H=cf, (h)such as maps, relationship curves, distribution curves5 If relationships among all the e are"independ- figures and qualitative descriptions, from which theent, first let us suppose n=2 for the purpose of sim- fact is obtained. The specific assignment principlesplicity, thenare as followscf, (H)+cf,( H)-cf, (H)xcf(H) when cf, (H),(I) For structures and traps which are very reliable,f2(H)≥0;the confidence values of the nodes are 1.0: for thosecf(H)= cf,(H)+cf,(H)-cf, (H) xcf,(H) when cf,(H)which are reliable. the confidence values are.81.0cf2(H)<0and 0.6-0.8 respectively for large structures and smalcf, (H)+cf,2() when cf, (H)xcf2 (H)<0structures: for those which are unreliable. the confiWhen n>2, all the cf (H) with the same sign are dence values are 0.2-0.6erged according to the above-mentioned formula,(2) The confidence values of the nodes confirmedand then the sum of them is calculated to get cf(H)by drill test and sample analysis are 1.0; those interB If the relationships among all the E, are"new preted by logging are 0.8-1.0; and those interpretedand,, and the e, are main nodes that play key roles, by geophysical and geochemical data are 0.5-0.8(3) For the nodes which are fixed by maps or seccf, (H) when cf,(H), cf, (H), " 'cf, ()tions. the confidence values are 0.7-0.9 when thereall equal to 0;are data-points in the adjacent area; 0.6-0.7 whencf(H=cf1(H)×(1+0.3)+∑cf( H) when thethere are no data-points but the maps or the sectionsabsolute value of the expression<1;basically conform to reality: 0.3-0.6 when the maps orI when cf (H)x(1+0.3)+2cf, (H)>1the sections are unreliable1 when cf(H)×(1+0.3)+2c(H<-1(4) For the ne confidence中国煤化工CNMHGVol. 74 No. 3CTA GEOLOGICA SINICASept. 2000values are 0.-0.8 when the unknown bodies are simi- the exploration practice: The gas-bearing probabilitieslar to the sample bodies, and <0.5 if notare all 1.00 for the Xiaowujiazi, Houwujiahu, Nongan(5)The confidence values for the inferred nodes are Wanjinta and Bawu gas pools: the structural traps0.4-0.7 when there are obvious geological regularities such as Shibei-I are of grade Ill, in which dry holesin the area and the geological models are reliable, and have been drilled and the gas-bearing probabilities are<0. 4 if notthe lowest; the traps located at the hydrocarbon(6)The values are 1.0 for industrial gas flows, generating centre and the adjacent gentle slope are0.7-1.0 for the shows found in testing, 0.5-0.9 for basically grade-I ones, while all those which havelogging interpretation, and 0.2 when no drilling has unfavorable preservation conditions or are far awaybeen donee source area areI or lil. it(7) The confidence values of the nodes are 0 when thus be seen that this system is suitable for assessinthe fact has been confirmed to be falsethe gas resources of the traps in the Songliao basiand similar basins with very reliable results, and can4.2 Results of assessmentbe used as a basis for planning the explorationThe gas-bearing probabilities (or the confidence values)of the traps which are inferred by the system areManuscript received Feb 2000listed in Table 1. The authors divide the traps intoaccepted May 2000different grades according to the gas-bearing probedited by Zhu Xilingabilities: among the 17 traps in the Quantou Forma-tion, there are 10 grade-I ones(or the most favourable Referencestraps, with >0.7), 4 grade-Il ones(or favourable traps, Charles F I, and Stout, J. L, 1985. Knowledge-based geologiith gas-bearing probability 0.7-0.55), 3 grade-III cal information systems. AAPG Bulletin, 69(11): 2048ones (or general traps, with gas-bearing probabilityDenham L R, 1985. Expert system in seismic exploratioAAPG Bulletin, 69(2): 249.0.55-0.4; while among the 8 traps in the DengloukuFormation, there are 4 grade-I ones, 2 grade-Il onesFu Guang, Chen Zhangming, Wang Pengyan and Lii Yanfang,1998. The formation period of the sealing ability of mud-stone caprocks of the Qingshankou formation in the SanzThe above assessment results basically tally withhao depression of the Songliao Basin. Geological ReviewTable 1 Results of assessment of the gas-bearing probability of trapFormationName of trap probabilityGradeFormationName of trapWanjinaNonganedong-lDedong-2Dedong. 3Denan-ⅡDenan-3IIShibei-2IShibe1-30.65ⅡShidong. 20.70Shidong-lXiaowujiaziShixi-1DengloukuShinan -1IShinan-2I中国煤化工CNMHGThe Gas Resources Assessment Expert System of the Songliao Basin67344(3): 295-301 (in Chinese with English abstract).The chronostratigraphy and stratigraphic classification ofKrystinik K B, and Clifton, H. E, 1985. Expert system forthe Cretaceous of the Songliao basin. Acta Geologica Sinicacomputer interpretation of beach and nearshore faciesEng. ed.).9(2):207-217AAPG Bulletin, 69(2): 276Wu Heng, 1994. Some problems on the establishment of theLi Shuseng, Wang Chuan and Fan Mingxing, 1994. Our undermine hydrogeology expert system. Geological Reviewstanding of trap's oil and gas resources assessment expert40(2): 131-139(in Chinese with English abstract).system. Oil Gas Geology, 15(3): 187-192 (in Chinese,Xiao Shujian, Wu Runting and Xiao renyue, 1997. The expertLiu Chengzuo, 1991. Geological Expert System. Beijingsystem for mineral deposit prognosis and its application inOcean Press, 1-10 (in Chinese)exploration of copper-nickle-sulphide deposits in XinjiangLiu Delai and Ma Li, 1998. Relation between prerift volcan-Geological Review, 43(2): 210-216(in Chinese with Eng-ics and the rift basin and geodynamic processes in Songliaolish abstract)basin. Geological Review, 44(2): 130-135(in Chinese with Yan Changhong, Luo Guoyu, Chen Zhaoqian Xu Yingwu, LiuEnglish abstract)Jianfang, Zhang Lianshen and Liu Huaipu, 1999. Study onLou Zhanghua, 1998. Diagenetic reactions, geochemical propthe suspended river of the Yellow River and its expert sy S-erties and origin of pore fluid in reservoirs of the Songliatem. Geological Review, 45(2 ): 218-224(in Chinese withBasin. Acta Geologica Sinica(Eng. ed. ) 72(2): 144-152English abstract)Lou Zhanghua, Lan Xiang, Lu Qingmei and Cai Xiyuan, 1999. Yin Yueping, Hu Haitao and Kang Hongda, 1996. An expertControls of the topography, climate and lake level fluctuastem of regional crustal stability evaluation of the siting oftion on the depositional environment of a shallow-waterkey engineering works. Geological Review, 42(2): 174-186deltar--A case study of the Cretaceous Putaohua reservoir(in Chinese with English abstract)in the northern part of Songliao basin. 1999. Acta GeologicaSinica, 73(1): 83-92 (in Chinese with English abstractabout the first authorTong Zhaoqi, Hu Changjun and Ma Yushu, 1993. The design Long ShengxiangBorn in 1959: now seniorand implementation of an expert system for trap evaluationGeological Review, 39(4): 292-296(in Chinese with Eng- geologist and deputy director of the Institute of Fieldlish abstractDevelopment, Academy of Petroleum GeologyWang Chuan, Huang Zheng, Fan Mingxing et al., 1996. Estab- CNSPC, in Beijing, China. He obtained his B.A. dlishment of accumulation model evaluation system in gas gree in petroleum geology from Chengdu College ofresources. Oil Gas Geology, 17(2): 102-109(in Chinese). Geology in 1982 and Ph. D in geological sciences fromWang Weiyuan and Wang Yuwen, 1989. Newest progress inChina university of geosciences in 1996. His re-geological expert system technique. China Ocean Oil andGas, 3(6): 10(in Chinese)searches are focused on basin analysis, resources aWu Lizhen,Wang Weiyuan, Rong Quanling et al., 1987. Ex- sessment and reservoir characterization. His public a-pert System for source condition assessment of PRES. Chi- tions include more than 10 study reports, a monographna Ocean Oil and Gas, 1(1): 43-54(in Chinese)and more than 10 papersWang Pujun, Du Xiaodi, Wang Jun and Wang Dongpo, 1995中国煤化工CNMHG