Abstract: A method includes obtaining a plurality of master sensor responses with a master sensor in a set of training fluids and obtaining node sensor responses in the set of training fluids. A linear correlation between a compensated master data set and a node data set is then found for a set of training fluids and generating node sensor responses in a tool parameter space from the compensated master data set on a set of application fluids. A reverse transformation is obtained based on the node sensor responses in a complete set of calibration fluids. The reverse transformation converts each node sensor response from a tool parameter space to the synthetic parameter space and uses transformed data as inputs of various fluid predictive models to obtain fluid characteristics. The method includes modifying operation parameters of a drilling or a well testing and sampling system according to the fluid characteristics.

Abstract: A model can be trained for discriminant analysis for substance classification and/or measuring calibration. One method includes interacting at least one sensor with one or more known substances, each sensor element being configured to detect a characteristic of the one or more known substances, generating an sensor response from each sensor element corresponding to each known substance, wherein each known substance corresponds to a known response stored in a database, and training a neural network to provide a discriminant analysis classification model for an unknown substance, the neural network using each sensor response as inputs and one or more substance types as outputs, and the outputs corresponding to the one or more known substances.

Abstract: Mutual-complementary modeling and testing methods are disclosed that can enable validated mapping from external oil and gas information sources to existing fluid optical databases through the use of forward and inverse neural networks. The forward neural networks use fluid compositional inputs to produce fluid principal spectroscopy components (PSC). The inverse neural networks apply PSC inputs to estimate fluid compositional outputs. The fluid compositional data from external sources can be tested through forward models first. The produced PSC outputs are then entered as inputs to inverse models to generate fluid compositional data. The degree of matching between reconstructed fluid compositions and the original testing data suggests which part of the new data can be integrated directly into the existing database as validated mapping. The applications of using PSC inputs to reconstruct infrared spectra and estimate oil-based-mud (OBM) contamination with endmember spectral fingerprints are also included.

Abstract: A method may include collecting measurement data using a first operational sensor and a second operational sensor of a downhole tool, standardizing optical responses of each operational sensor to a master sensor in a tool parameter space to obtain a standardized master sensor response, transforming the standardized master sensor response to a synthetic parameter space response of the master sensor, applying a fluid model with the synthetic parameter space response of the master sensor to predict a fluid characteristic, comparing a first prediction obtained with the fluid model from the first operational sensor with a second prediction obtained with the fluid model from the second operational sensor, determining a fluid characteristic from the first prediction and the second prediction, and optimizing a well testing and sampling operation according to the fluid characteristic.

Abstract: A method includes obtaining a plurality of master sensor responses with a master sensor in a set of training fluids and obtaining node sensor responses in the set of training fluids. A linear correlation between a compensated master data set and a node data set is then found for a set of training fluids and generating node sensor responses in a tool parameter space from the compensated master data set on a set of application fluids. A reverse transformation is obtained based on the node sensor responses in a complete set of calibration fluids. The reverse transformation converts each node sensor response from a tool parameter space to the synthetic parameter space, and uses transformed data as inputs of various fluid predictive models to obtain fluid characteristics. The method includes modifying operation parameters of a drilling or a well testing and sampling system according to the fluid characteristics.

Abstract: Disclosed are systems and methods that use discriminant analysis techniques and processing in order to reduce the time required to determine chemical and/or physical properties of a substance. One method includes optically interacting a plurality of optical elements with one or more known substances, each optical element being configured to detect a particular characteristic of the one or more known substances, generating an optical response from each optical element corresponding to each known substance, wherein each known substance corresponds to a known spectrum stored in an optical database, and training a neural network to provide a discriminant analysis classification model for an unknown substance, the neural network using each optical response as inputs and one or more fluid types as outputs, and the outputs corresponding to the one or more known substances.

Abstract: This disclosure includes methods for designing a simplified Integrated Computational Element (ICE) and for optimizing a selection of a combination of ICE designs. A method for fabricating a simplified ICE having one or more film layers includes predicting an optimal thickness of each of the one or more film layers of the simplified ICE using a neural network. A method for recalibrating the fabricated ICE elements for system implementation is also disclosed. The disclosure also includes the simplified ICE designed by and the ICE combination selected by the disclosed methods. The disclosure also includes an information handling system with machine-readable instructions to perform the methods disclosed herein.

Abstract: System and methods for downhole fluid classification are provided. Measurements are obtained from one or more downhole sensors located along a current section of wellbore within a subsurface formation. The measurements obtained from the one or more downhole sensors are transformed into principal spectroscopy component (PSC) data. One or more fluid types are identified for the current section of the wellbore within the subsurface formation, based on the PSC data and a fluid classification model. The fluid classification model is refined for one or more subsequent sections of the wellbore within the subsurface formation, based at least partly on the one or more fluid types identified for the current section of the wellbore.

Abstract: A method including selecting candidate sensors for pooled calibration and determining a multivariate fluid characterization model with the pooled sensors is provided. The method includes determining a virtual master kernel standardization model with the pooled sensors, implementing a calibration result into a processor circuit and determining a value of a fluid characteristic by applying the multivariate fluid composition model to a plurality of responses obtained from a plurality of sensor responses to the fluid sample. The plurality of responses may be obtained from the plurality of sensor responses using the virtual master kernel standardization model. The method includes optimizing a wellbore operation based on the value of the fluid characteristic. A device for implementing the above method is also provided.

Abstract: A photometric system design methodology employs genetic algorithms to optimize the selection of optical elements for inclusion in the photometric system in order to improve system performance with respect to environmental conditions (i.e., to “ruggedize” the photometric system). The genetic algorithms utilize a multi-objective fitness function to evolve simulated optical element selection, which may be a combination of optical filters and integrated computational elements. The system may also output a size reduced database that serve as simulated candidate optical elements through global optimization, or may output a fixed number of simulated optical elements through conditional optimization for actual tool implementation and calibration analysis.

Abstract: A method of cross-tool optical fluid model validation includes selecting verified field data measured with a first sensor of an existing tool as validation fluids and selecting a second sensor for a new tool or on a different existing tool. The method may also include applying cross-tool optical data transformation to the validation fluids in a tool parameter space from the first sensor to the second sensor, and calculating the synthetic optical responses of the second sensor on the validation fluids through cross-space data transformation. The method may further include determining a new or adjusting an existing operational fluid model of the second sensor in a synthetic parameter space according to the candidate model performance evaluated on the validation fluids, and optimizing well testing and sampling operation based on real-time estimated formation fluid characteristics using the validated fluid models of the second sensor in an operating tool.

Abstract: A system includes an optical computing device having an optical multiplexer that receives a sample light generated by an optical interaction between a sample and an illumination light is provided. The system includes sensing elements that optically interact with the sample light to generate modified lights, and a detector that measures a property of the modified lights separately. Linear and nonlinear models for processing data collected with the above system to form high-resolution spectra are also provided. Methods for designing optimal optical multiplexers for optimal reconstruction of high-resolution spectra are also provided.

Abstract: An apparatus to determine fluid viscosities downhole in real-time includes a housing and an excitation element positioned therein. Electrical circuitry provides a drive signal that excites an excitation element into rotational oscillations. A detector produces a response signal correlating to the detected oscillating movement of the excitation element. Circuitry onboard the apparatus utilizes the drive and response signals to determine the fluid viscosity.

Abstract: A system includes an optical computing device having an optical multiplexer that receives a sample light generated by an optical interaction between a sample and an illumination light is provided. The system includes sensing elements that optically interact with the sample light to generate modified lights, and a detector that measures a property of the modified lights separately. Linear and nonlinear models for processing data collected with the above system to form high-resolution spectra are also provided. Methods for designing optimal optical multiplexers for optimal reconstruction of high-resolution spectra are also provided.

Abstract: Two or more Integrated Computational Element (“ICE”) structures are designed and utilized in an optical computing device to combinatorily reconstruct spectral patterns of a sample. To design the ICE structures, principal component analysis (“PCA”) loading vectors are derived from training spectra. Thereafter, two or more ICE structures having spectral patterns that match the PCA loading vectors are selected. The selected ICE structures may then be fabricated and integrated into an optical computing device. During operation, the ICE structures are used to reconstruct high resolution spectral data of the samples which is utilized to determine a variety of sample characteristics.

Abstract: A system for integrated computational element (“ICE”) design optimization and analysis utilizes a genetic algorithm to evolve layer thickness of each fixed ICE structure using a constrained multi-objective merit function. The system outputs a ranked representative group of ICE design candidates that may be used for further fabricability study, ICE combination selection, efficient statistical analysis and/or feature characterization.

Abstract: A method for shale fracturing includes determining dynamic-elastic properties of a shale deposit in a geological formation. A training database is generated by three-dimensional fracture modeling. A neural network is generated in response to output results of the training database. The shale fracturing may then be performed based on the neural network.

Abstract: Dimensionality reduction systems and methods facilitate visualization, understanding, and interpretation of high-dimensionality data sets, so long as the essential information of the data set is preserved during the dimensionality reduction process. In some of the disclosed embodiments, dimensionality reduction is accomplished using clustering, evolutionary computation of low-dimensionality coordinates for cluster kernels, particle swarm optimization of kernel positions, and training of neural networks based on the kernel mapping. The fitness function chosen for the evolutionary computation and particle swarm optimization is designed to preserve kernel distances and any other information deemed useful to the current application of the disclosed techniques, such as linear correlation with a variable that is to be predicted from future measurements. Various error measures are suitable and can be used.

Abstract: A method includes obtaining a measurement of one or more properties of a downhole fluid using a fluid analysis tool. The fluid analysis tool includes fluid sensors and one or more processors coupled with the fluid sensors. A first prediction is generated using the measurement from the fluid sensors. A second prediction is generated using an adaptive neuro-fuzzy inference system based on the first prediction of the properties.

Abstract: System and methods for downhole fluid analysis are provided. Measurements are obtained from one or more downhole sensors along a current section of wellbore within a subsurface formation. The measurements obtained from the one or more downhole sensors are transformed into principal spectroscopy component (PSC) data. At least one fluid composition or property is estimated for the current section of the wellbore, based on the PSC data and a fluid analysis model. The fluid analysis model is refined for one or more subsequent sections of the wellbore within the subsurface formation, based at least partly on the fluid composition or property estimated for the current section of the wellbore.