Abstract: Theoretical test and mathematical calculation methods of quantitative test for contribution of mixed sources rocks of natural gas using chromatogram fingerprints of light hydrocarbon have been proposed, and the related experimental verification for deep source rocks and natural gas samples collected from Xujiaweizi fault-depression in northern Songliao Basin has been made. With light hydrocarbon fingerprint chromatographic detecting method for adsorbed gas and natural gas in hydrocarbon source rocks, using the ratio parameter of light-hydrocarbon molecular compound with adjacent/similar structure and similar feature in absorbed gas in hydrocarbon source rocks, the feature light-hydrocarbon fingerprint parameters are fixed to characterize the four sets of source rocks including Denglouku Formation, Yingcheng Formation, Shahezi Formation and Huoshiling Formtion as well as Carboniferous-Permian systems. According to the different mixed matching models, the geochemical model parameter is obtained and used to establish the test model of source rocks contribution by intelligent neural network algorithm training. So the quantitative contribution of the four sets of deep source rocks in Xujiaweizi fault-depression to 15 deep mixed-source natural gas samples is calculated. Results have indicated that the origin of natural gas from different areas or from different wells and intervals in the same area is different, underlying gas source rock would be as the main source and the gas reservoir is featured by source-controlled type mainly with vertical migration. The contribution of source rocks from Shahezi Formation and Huoshiling Formation, Yingcheng Formation, Permo-Carboniferous System and Denglouku Formation is 69.2%, 16.2%, 9.7% and 4.8%, respectively.

Key words: Xujiaweizi fault-depression, source-mixed natural gas, contribution of source rocks, chromatogram fingerprint of light-hydrocarbon, test model, simulated calculation

摘要： 提出了混源天然气源岩贡献色谱轻烃指纹定量测试的理论测试方法及数学模拟计算方法,选择松辽盆地北部徐家围子断陷深层源岩和天然气样品进行实验验证。采用源岩吸附气和天然气轻烃指纹色谱检测方法,利用源岩吸附气中相邻或相近的结构和性质相似的轻烃分子化合物的比值参数,确定表征深层登娄库组、营城组、沙河子组和火石岭组、石炭-二叠系4套源岩的特征轻烃指纹参数,按不同混合配比模式获得地球化学模型参数,用智能神经网络学习算法训练建立源岩贡献测试模板,模拟计算了徐家围子断陷深层4套源岩分别对15个深层混源天然气样品的定量贡献。结果表明,不同地区或同一地区不同井及层段的天然气来源有差别,呈现主要来源于下伏气源岩和以垂向运移为主的源控型气藏特征,沙河子组和火石岭组、营城组、石炭-二叠系、登娄库组的源岩贡献平均分别为69.2%、16.2%、9.7%、4.8%。

关键词: 徐家围子断陷, 混源天然气, 源岩贡献, 轻烃指纹色谱, 测试模板, 模拟计算