Field Tests of a New Optical Sensor Based on Integrated Computational Elements for Downhole Fluid Analysis

Abstract Optical fluid analyzers have been used in wireline formation tests for real-time downhole fluid analysis during pumpout tests for over a decade. Based on conventional optical spectroscopic methods, they separate broadband light into its constituent wavelengths via notch filters or multichannel grating spectrometers. A few narrow wavelength constituents are then detected and mathematically recombined to yield an answer product. Complex hydrocarbon fluids have many overlapping spectra and are optically active over a wide range of optical wavelengths. Consequently, accurate analyte detection in hydrocarbon fluids generally requires analysis of a large number of wavelengths over a large spectral region. The performance and range of detectable analytes of conventional optical fluid analyzers is band-limited. Of all available spectra from visible to near- and mid-infrared, only a small fraction of spectral data is used. In addition, the splitting of the optical beam into its wavelength constituents typically decreases signal-to-noise ratios (SNR) by orders of magnitude thereby limiting the accuracy, sensitivity, and viable ranges of the answer product. A new downhole optical sensor platform has been developed for downhole in-situ fluid analysis based on multivariate optical computing (MOC) technology. Historically developed for other markets, MOC is a well-established tool that combines chemometrics and pattern recognition with the power of optical computing. The heart of this new optical sensor platform is an optical component called the ICE CORE™ integrated computational element. The ICE Core sensor is analogous to the processing chip of a PC and performs calculations literally at the speed of light within the multivariate optical computer. Each ICE Core special multilayer optical element is encoded with a pre-designed multivariate regression vector specific to an analyte or property of interest. These optical elements are typically very broadband and may have a response that extends from 400 nm to 5000 nm. The wide bandwidth of these optical elements combined with their intrinsic, high etendue SNR advantage enables laboratory-grade optical analyses downhole. The compact and passive nature of the ICE Core sensors results in high reliability. A multivariate optical computer may consist of many different ICE Core sensors designed to detect many different analytes or properties. The downhole optical sensor platform in this study can have 20 (or more) different ICE Core sensors per MOC. This platform has been under field trial qualification for over a year using a step-by-step process with methane (C1) ICE Core sensors as the primary validation analyte. In this paper, the results of three different downhole pumpout field trials are presented. After reviewing the fundamental principles of MOC and ICE Core technology, the field trial validation process is described. Results showing downhole ICE Core measured methane analyte concentrations vs. time are presented and compared with other downhole sensor data. ICE Core measured concentrations are also compared with independent laboratory test results. The results demonstrate the downhole viability of the optical platform and the sensitivity associated with the ICE Core detection method.