Integrated Predication Method Research On Volcanic Reservior In Xujiaweizi Rift, Songliao Basin

The deep Mesozoic volcanic rocks in North Songliao Basin has become an important reservoir of deep natural gas, according to the previous exploration activities, and it also has better gas-exploration prospects, the gas prediction and control reserves have been provided. Because of the deep-buried volcanic rocks which lack of outcrops, many drillings can't reach the volcanic rocks, some of drillings can drill through them, but no samples of the volcanic rocks can be gained, or the whole samples of the volcanic rocks can't be collected. All of these lead to the very short of the samples of volcanic rocks. The residual cores of volcanic rocks in the drilling holes can't provide the information about the volcanic lithologic distributions, period-times, circles, attitude, facies, structures, etc, in result with poor correspondence relations between the lithological characters and the physical properties of volcanic rocks, undefined distribution scope and facies division of volcanic rocks, unclear reservoir types of volcanic rocks and unclear origin of gas, etc. These issues will affect badly the gas exploration and the prospective appraisement of volcanic rocks in the deep Songliao basin and make deep volcanic reservoir predication and appraisement a major problem in oil-gas explorations. A few related predication methods have been applied at home and abroad focused on the volcanic reservoir predication, but they are all simple and with multi-solution and low predication accuracy. There are no successful methods to meet the current needs for reservoir explorations of volcanic rocks. In this paper, an all-round research on predication methods for volcanic rocks has been developed by using core, well-log and seismic data, based on the basic principle of volcanic lithology and geophysics, and the comprehensive research idea of multi-disciplines. Xujiaweizi Rift in Songliao Basin will be presented as a case in this paper. 一,Volcanic well log identification techniques The identification of volcanic lithology is the basis of volcanic reservoir research. Core analysis is the most efficient and direct method to identify the volcanic lithology in hydrocarbon-bearing basins. Because of the low coring-rate in drilling, however, well log data have to be used in the identification of volcanic lithology. 1. Well log response characteristics of volcanic rocks In this paper, according to the comparative analysis on many wells'logs in Xujiaweizi areas, the following laws of the response characteristics of volcanic well log in the studying areas have been summarized: (1) Lithology transits from base to acidity, and the radioactivity is more and more strong. In the studying areas, it shows that natural gamma value is low in basalt, medium in andesite and high in rhyolite; (2) Neutron well-log response reduces gradually from base to acidity, through neutrality. And in the studying areas, neutron well-log response shows that it's 18P.U. in basalt, 6P.U. in dacite, and less than 3P.U. in rhyolite. Because of alteration, tuff formed into clay minerals to adsorb the formation water, that leads to a high neutron value, which is up to 24P.U.; (3) Density in density logging reduces gradually with decreasing Fe and Mg, from basic rocks to acidic rocks, through neutral rocks; (4) Tuff is characterized by high acoustic time (80ms/ft) and low resistivity(8-20 Ω.m). 2. Crossplot characteristics of volcanic rocks Logging data crossplot is the most efficient method to identify the volcaniclithology in hydrocarbon-bearing basins. In the crossplot, the division and distribution areas of lithology can be found visually, and volcanic lithology also can be identified clearly. In the research, based on the detail lithologic designation results of deep volcanic rocks from 30 wells (144 blocks) in Xujiaweizi areas, which were analyzed in the labs of Beijing Geology University, the log data from 17 wells drilled newly in the resent years, and the lithologic designation from Daqing Oilfield Exploration & Development Research Institute, a complete analysis on measured logging data response characteristics has been performed, and a crossplot chart of natural gamma-neutron density difference has been also built. The volcanic lithology has been divided into 5 types. The coincidence rate of the chart practiced in 18 wells in deep Xujiaweizi areas, such as Xushen 6#, etc, is up to 81.3%. 3. Volcanic lithology identification techniques by using imaging well-log Volcanic lithology is complex in deep Xujiaweizi, it has developed from acidity to mid-base. At present, more than 10 types of lithology can be found there. Volcanic lithology can be divided in general by using conventional logging data and it can be confirmed and sub-divided accurately further, combining with volcanic texture and structural features reflected from imaging logging data, and it can be also identified for more than 10 types. It's impossible to identify all the lithologic standard geologic models, because imaging logging data are the integrated response of the resistivity and the reflected acoustic wave properties of down-hole geologic bodies, and the data have multi-solvability. In the research, based on the calibration imaging logging data of standard geologic lithologic model-base, 5 types of standard logging model images have been built, focused on the texture and structural features of rocks. 4. Volcanic lithology identification techniques using conventional and imaging logging data At first, lithology is confirmed based on rock acidity, using conventional logging data; and then, the lithology is confirmed based on rock texture and structural features,using imaging logging data; at last, the lithology is confirmed based on the principle which is unified with geologic designation, combining conventional logging data with imaging logging data. At present, 10 types of volcanic lithology can be identified, using the methods presented above. In the volcanic lithology identification performed in 36 wells in the studying areas, the coincidence rate is up to 83%. 二,Volcanic seismic reflection features analysis and identification techniques In this paper, according to the theoretical research by reducing geologic model design, the resolution capability of seismic reflection explorations to the deep volcanic rocks has been concluded with different wavelet frequency, and the resolution capability of seismic data to deep volcanic reservoir has also concluded under the current seismic exploration technique conditions. Meanwhile, in order to predict and characterize the distribution of deep volcanic bodies accurately, more volcanic-targeted processing techniques have been applied, such as pre-stack depth migration processing, non-conventional slicing seismic facies processing, etc., and the host location of volcanic rocks and its space arranging laws have been predicted more accurately. 三,Volcanic seismic features parameters extraction and optimized appraisement techniques This paper presents partial least-square optimization method to volcanic seismic attributes at the first time. The research shows that the seismic attribute optimization results from the partial least-square principal component analysis methods are nearly the same to which from conventional analysis methods, when the factors affected the reservoir physical prediction parameters are less; the partial least-square optimization method shows more advantages than the principal component analysis methods and show more strong adaptability, when there's so much noise in initial seismic data. At the same time, 52 types of extracted seismic attributes have been optimized by using the partial least-square optimization method, and 4 types of seismic attributeparameters which are sensible to volcanic rocks; the method is applied in Xingshan area, Xujiaweizi, based on a processing analysis on the 4 types of attribute, the accuracy of volcanic prediction is improved which provides a stable base to the following volcanic reservoir predictions. 四,Integrated prediction techniques for volcanic reservoirs 1. 'Four-Property'relation research on volcanic rocks In this paper, relations between volcanic lithology, facies, electricity, hydrocarbon-bearing capability and reservoir physical properties in the studying areas have been built, based on the 'Four-Property'relation research on volcanic reservoirs. 'Four-Property'relation shows that physical properties of volcanic rocks are controlled by volcanic lithology evidently, the close relations between the hydrocarbon-bearing capability of volcanic rocks and the physical properties and lithology of volcanic rocks have been also found. Fracture development and fracturing of structural high-points are very important to improve the volcanic reservoirs, and they can link up primary pores, developed fracture is a better reservoir space itself. In the research, the pore permeation parameters have been half-classified quantitatively, based on the practical geological conditions, which provides a geological basis to the building of interpretation model of volcanic reservoir parameter logging next. 2. Prediction techniques for crater developing zones The identification of craters and the prediction of crater developing zones are very important to the exploration and development of volcanic reservoirs. To the oil & gas exploration and production companies at home and abroad, the distribution area of craters is the exploration target for a long time. The crater controls the distribution of facies zones of volcanic rock-bodies, and is always the developing area of favorable volcanic reservoirs and the high producing area of hydrocarbons. The seeking for crater developing zones is significant in volcanic explorations. In this paper, crater developing zones have been identified using structural trend surface analysis techniques at the first time, combining with various seismic volumeslices, and the distribution and the space arranging laws of the craters in the testing areas have been also predicted efficiently. 3. Prediction techniques for the plane distribution of volcanic facies Volcanic rocks of Yingcheng formation in Xingcheng-Fengle area, Xujiaweizi, developed mainly in Yingyi segment. In this paper, based on the analysis of volcanic seismic reflection features and volcanic facies presented above, waveform clustering analysis and spectrum imaging techniques have been applied in the research to predict the plane distribution features of volcanic facies in the studying areas, and better results have been gained. 4. Integrated prediction techniques for volcanic reservoirs In this paper, after figuring out the factors which affect the volcanic reservoirs in the studying areas, 'Four-Property' relation of volcanic rocks, and the arranging laws of crater and volcanic facies, a path-breaking research has been performed focused on the prediction issues for volcanic reservoirs, using model annealing inversion method. The research shows that model annealing is a global optimization technique, and the model annealing method can overcome the disadvantages of conventional optimization methods, and it can gain the global optimization solution. According to model annealing algorithm, the model annealing inversion method constrained by broadband can transform given conditions into concrete constraints to control the inversion. Thus, using model annealing method, the seeking issues for global optimization can be solved, and the accuracy and convergence rate of inversion can be also improved at the same time, through using constraints reasonably. 5. Distribution laws of volcanic reservoirs and volcanic gas reservoirs In this paper, using density inversion to predict volcanic porosity and using two attribute analysis methods to predict the hydrocarbon-bearing capability of volcanic reservoirs, a research on the distribution laws of volcanic reservoirs and volcanic gas reservoirs has been performed in the testing area ( Xingcheng-Fengle area ) based on the research findings presented above. The results show: in the testing area, volcanic reservoirs have developed better in the explosive facies and the eruptive facies near...