بررسي عوامل كنترل‌كننده كيفيت مخزني با استفاده از داده‌‌‌‌هاي لرزه‌‌اي در ميدان نفتي شادگان

Study of reservoir quality controller factors using seismic data

در اين تحقيق به‌منظور برآورد مشخصات مخزن و بررسي‌‌‌هاي رخساره‌‌اي در ميدان شادگان، ابتدا سرسازند سروك و بخش‌‌‌‌هاي بالايي آن به‌منزله لايه‌‌‌‌هاي مخزني هدف، تفسير ساختماني شد و سپس با اِعمال روش‌‌‌‌هاي وارون‌سازي لرزه‌‌اي، نشانگر مقاومت صوتي لايه‌‌‌‌هاي پيش‌گفته استخراج شد. با به‌كارگيري روش‌‌‌‌‌‌هاي آماري و شبكه‌‌‌‌هاي عصبي مصنوعي پارامتر‌‌‌‌هاي مخزني تخلخل و سيري (اشباع) شاره‌‌ در لايه‌‌‌‌هاي مخزني سازند سروك مورد بررسي قرار گرفت و نقشه توزيع آنها تهيه شد. تحليل نتايج و بررسي نقشه‌‌‌‌هاي مقاومت صوتي، تخلخل مفيد، سيري شاره‌‌ و مقدار هيدروكربور نشان داد كه زون‌‌‌هاي مخزني گروه بنگستان محدود به بخش‌‌‌‌هاي سروك 2 و سروك 4 است كه داراي تغييرات رخساره‌‌اي هستند؛ به‌طوري‌كه اين تغييرات در گستره مكاني باعث تفكيك جانبي زون‌‌‌‌هاي مخزني به دو بخش كاملاً مجزاي رسوبي با AIو ميزان تخلخل كاملاً متفاوتي شده است. وقتي كه اين تغييرات رخساره‌‌اي و مقدار تخلخل در كنار سيري شاره‌‌ مورد بررسي و ارزيابي قرار گرفت مشخص شد كه مخازن گروه بنگستان در ميدان شادگان تحت‌تاثير تغييرات رخساره‌‌اي‌اند و به دو بخش شمال غربي و جنوب شرقي تقسيم مي‌شوند كه بخش شمال غربي، داراي پتانسيل مخزني بيشتري است. با توجه به نتايج اين طرح، مخازن سازند سروك در اين ميدان از نوع مخازن چينه‌‌اي ارزيابي مي‌شوند و تغييرات رخساره‌‌اي درحكم عامل اصلي كنترل‌كننده كيفيت مخزني آنها در نظر گفته شده است. لذا براي توسعه و حفاري‌‌‌هاي جديد، بخش شمال غربي ميدان و زون‌‌‌‌هاي مخزني 2 و 4 پيشنهاد مي‌شود.

Sarvak Formation is deposited between overlaying Ilam Formation and underlying Daryian Formation spanning in time from Albian to Turonian in Dezful Embayment. This formation has an overall thickness of 500 m consisting of intermittent limestone and inter-bedding shaly limestone in Shadegan Oilfield and the upper parts have reservoir qualities. In current reservoir characterization practice, after structural interpretation, correlation between reservoir properties and seismic attributes at well locations were firstly derived by statistical and neural network technics. Then this is and applied to seismic data accordingly and the reservoir parameters such as porosity and fluid saturation were expanded to seismic data and analyzed lateraly and vertically. In reservoir parameters estimation workflow, aside from internal seismic attributes, seismic acoustic impedance was also utilized that was indirectly constructed by seismic inversion from seismic amplitudes and well log data. Different seismic inversion technics were tested in Strata software in order to achieve the most credible seismic acoustic impedance volume. According to this analysis and considering conformity of the resulted values at well locations, Model Based inversion technic was finally selected as better correlation was observed between well log values and inverted results using this method. As the result, Model Based Inversion was chosen for computing acoustic impedance volume. Estimation of effective porosity as one of important reservoir properties was carried out from seismic data and well log data of well 2 and 13 using multi-regression and artificial neural network technics. Sensitivity analysis which was conducted to obtain the optimum combination of seismic attributes showed that adding any other attributes to acoustic impedance is not conducive to better results in this project and will increase estimation error. The final correlation between effective porosity and seismic attribute is about 60 percent. For fluid saturation estimation, another sensitivity analysis was carried out after extracting different seismic attributes. Finally, the most correlated seismic attribute and fluid saturation at well location is the Average Frequency. After this single attribute analysis, multi-attribute analysis was conducted to find the optimum combination of seismic attributes in estimation of fluid saturation. The result shows a maximum correlation of 52 percent between values of 1/AI and Quadrature Trace. Artificial neural network methods are recently the center of notice for many researchers with regard to their non-linear nature of finding weights of estimation equation. Conducting several neural network techniques in this project showed that PNN technic contains the most effective training and estimation of water saturation. Using this method showed better correlation of 57 percent and error of 0.175 which is a slightly better estimation in comparison with the previous method. Finally this method was used for the estimation of water saturation throughout the reservoir volume. Visualization and analysis of maps and volumes generated for reservoir parameters showed that reservoir zones pertaining to Bangestan Group are limited to Sarvak zone 2 and 4 in this field. Moreover, study of seismic attributes especially AI suggested that aforementioned reservoir zones have experienced facies variation in a way that there is two distinct facies zones with completely different AI values. Meaningful and conspicuous reduction in effective porosity in reservoir zone from north-west to south-east is also solid evidence ascribed to facies variation in this field. Importance of this issue was even more accentuated when these variations was juxtaposed to fluid saturation. Simultaneous analysis of these three pieces of information showed that firstly Bangestan reservoir group in Shadegan field is under control of spatial faceis variation and secondly this resulted into relatively distinct separation of north-west/south-east reservoir parts with better reservoir quality in the north-west part. Sarvak zone 2 has good reservoir quality and accumulation of hydrocarbon in north-east and also central regions but poor reservoir quality and no considerable hydrocarbon accumulation in south-west parts. Finally, this study shows that the Sarvak reservoir zones in this field should be deemed stratigraphic as facies variations is a major factor in the reservoirs quality, so in the field development it is suggested that drilling of layers Sarvak 2 and 4 in the north-west part of field should be a priority.