Summary
Determination of the type and quality of hydrocarbon fluid that can be produced from a formation prior to construction of production facilities is of equal economic importance to predicting the fluid rate and flowing pressure. We have become adept at making such estimates for formations drilled with water-based muds, using open-hole formation evaluation procedures. However, these standard open-hole methods are somewhat handicapped in wells drilled with synthetic oil-based mud because of the chemical and physical similarity between the synthetic oil-based filtrate and any producible oil that may be present. Also complicating the prediction is that in situ hydrocarbons will be miscibly displaced away from the wellbore by the invading oil-based mud filtrate, leaving little or no trace of the original hydrocarbon in the invaded zone. Thus, normal methods that sample fluids in the invaded zone will be of little use in predicting the in situ type and quality of hydrocarbons deeper in the formation. Only when we can pump significant volume of filtrate from the invaded zone to reconnect and sample the virgin fluids are we successful. However, since the in situ oil and filtrate are miscible, diffusion mixes the materials and blurs the interface; as mud filtrate is pumped from the formation into the borehole, the degree of contamination is greater than one might expect, and it is difficult to know when to stop pumping and start sampling. What level of filtrate contamination in the in situ fluid is tolerable?
We propose a procedure for enhancing the value of the data derived from a particular open-hole wireline formation tester by quantitatively evaluating in real time the quality of the fluid being collected. The approach focuses on expanding the display of the spectroscopic data as a function of time on a more sensitive scale than has been used previously. This enhanced sensitivity allows one to confidently decide when in the pumping cycle to begin the sampling procedure. The study also utilizes laboratory determined PVT information on collected samples to form a data set that we use to correlate to the wireline derived spectroscopic data. The accuracy of these correlations has been verified with subsequent predictions and corroborated with laboratory measurements. Lastly, we provide a guideline for predicting the pump-out time needed to obtain a fluid sample of a pre-determined level of contamination when sampling conditions fall within our range of empirical data.
Conclusions
This empirical study validates that PVT quality hydrocarbon samples can be obtained from boreholes drilled with synthetic oil-based mud utilizing wireline formation testers deployed with downhole pump-out and optical analyzer modules. The data set for this study has the following boundary conditions: samples were obtained in the Gulf of Mexico area; the rock formations are unconsolidated to slightly consolidated, clean to slightly shaly sandstones; the in situ hydrocarbons and the synthetic oil-based mud filtrate have measurable differences in their visible and/or near infrared spectra. Specifically, this study demonstrates that during the pump-out phase of operations we can use the optical analyzer response to predict the API gravity and gas/oil ratio of the reservoir hydrocarbons prior to securing a downhole sample. Additionally, we can predict the pump out time required to obtain a reservoir sample with less than 10% mud filtrate contamination if we know or can estimate reservoir fluid viscosity and formation permeability. Extension of this method to other formations and locales should be possible using similar empirical correlation methodology.
Introduction
The high cost of offshore production facilities construction and deployment require accurate prediction of hydrocarbon PVT properties prior to fabrication. In the offshore Gulf of Mexico, one method to obtain a PVT quality hydrocarbon sample is to use a cased hole drill stem test. However, this procedure is usually quite costly due to the need for sand control. Shell has been an advocate of eliminating this costly step by utilizing openhole wireline test tools to obtain the PVT quality sample of the reservoir hydrocarbon. The success of this approach depends upon the availability of a wireline tool with a downhole pump that permits removal of the mud filtrate contamination prior to sampling the reservoir fluids, and a downhole fluid analyzer that can distinguish reservoir fluid from filtrate. One such tool is the Modular Formation Dynamics Tester (MDT).1 The optical fluid analyzer module of the MDT functions by subjecting the fluids being pumped to absorption spectroscopy in the visible and near-infrared (NIR) ranges. Interpretation of these spectra is the subject of this paper. Tool descriptions and basic theory of operations were presented in an earlier text.2
The concept of using visible and/or NIR spectroscopy to characterize the fluids being sampled while pumping is straightforward when there are measurable differences in the spectra of the mud filtrate and the reservoir hydrocarbons. As shown in Fig. 1, there are well known areas3,4 of the NIR spectrum (800-2000 nm) that are diagnostic of water and oil. The optical fluid analyzer module (OFA) of the MDT has channels tuned at 10 locations as indicated in Fig. 1, and thus the response in channels 6, 8, and 9 can be used to discern water from hydrocarbon. Another section of the OFA is designed to detect gas by measuring reflected polarized light from the pumped fluids, but we do not discuss its operation further except to say that it is a reliable gas indicator.
Some indicators of water quality of the Tamis river
