The Dynamic Monitoring Technique For The Production Allocation Of The Hybrid Recovery Wells In Oilfield

There are a variety of techniques employed to geochemical research of oil production allocation, including gas chromatographic fingerprinting, liquid chromatographic, ultraviolet spectrum, stable isotope analysis, GC-MS method, etc. The geochemical techniques for research of oil production allocation are generally characterized by quick, low cost and low sampling requirements, which can be easily combined with conventional logging. Therefore, the geochemical techniques have a very bright prospect in terms of its application.Gas chromatographic fingerprinting is one of important techniques for oil reservoir geochemistry, which have been proved to be significant for the horizontal and vertical connectivity of fluids, division of fluid units, determination of oil production allocation, monitoring of the remnant oil distribution and so on. With the exploration and development of an oilfield, many issues will be encountered in oil production, such as decrease in production potential, water-contained oil increasing, contradictions of production allocation between/within producing zones, the unknown distribution of the remanent oil, as well as the contributions of the oil-producing zones in hybrid production well. In particular, the research for the production allocation in multiple-zone production well has been one of the core issues and has paid widespread attention from the managers of oil reservoirs, consequently an in-depth study is required on the application of GC fingerprinting technique in dynamic monitor of oilfield development.In the thesis the mathematic models to compute the production allocation of hybrid producing wells are established and the dynamic monitoring technique is provided for multiple-zone production in an oilfield by using non-linear artificial neural networks and support vector regression, which are made based on both technical and theoretical study of total hydrocarbons GC fingerprint, by means of compound-specific peak height ratio, absolute quantitative analysis with internal standards, and a great number of matching experiments of crude oils from individual zone.The models and technique are applied to Wenchang 13-1/2 oilfield to monitor the dynamic production allocation of hybrid recovery wells and calculate contribution of individual producing zone in hybrid recovery wells.First of all, the study was focused on geological investigation of the current status of the oilfield. Before calculation of production allocation of different producing zones in a block was performed, a detail investigation must be made, including the status of connectivity between zones in the oilfield, the geology and geochemistry of the oil reservoirs and the design/development history of the oilfield in order to well understand such factors as thickness, porosity and permeability of the reservoirs that may affect the oil-producing zones in the oilfield. In addition, GC fingerprint features of crude oil are affected by the organic facies and maturity of source rocks, which will result in the difference in crude oil compositions and GC fingerprint features. Therefore, GC fingerprint features represent the inhomogeneity of the crude oil from individual producing zone.Secondly, that modern instrument is employed is very important for the stability and the reliability of the analytical results. This is a key for the accuracy calculation of production allocation from different oil-producing zones in the hybrid co-recovery wells. Thus, a state-of-the-art gas chromatograph, chemical workstation system and an automatic sample-feeding machine are employed. In this way the precision for detection of the instruments has been greatly improved to enhance the accuracy and reliability of the analytical data.Finally, different fitting models and geochemical fingerprint features are used to predict the contribution of individual producing zone according to hybrid recovery conditions. For the recovery from two producing zones, the linear fitting and the peak height ratios of geochemical fingerprints are employed and the satisfied results can be obtained. For the hybrid recovery of multiple (3 and more than 3) producing zones, however, the fingerprints of the crude oils from individual producing zone show a complex, non-linear relationships, it can't be fit by a simple, linear function. Thus, absolute quantification fingerprints and non-linear fitting have to use for calculation of the production allocation from multiple producing zones.The BP nueral networks, SVR and software are here developed. Together with the absolute quantification fingerprints, BP and SVR are applied to monitoring the dynamic production allocation of the hybrid recovery wells in Wenchang 13-1/2 oilfield. The production allocations of 5 hybrid recovery wells are computed, the results showing that the predicted values are basically consistent with those measured by stop production. This provides a theoretical support and applied technique for adjusting production allocation in this oilfield.