Determination of Density and Viscosity of Crude Oil Samples from FTIR Data using Multivariate Regression, Variable Selection and Classification

Alizadeh, Shahla; Ta, Souvik; Ray, Ajay K.; Lakshminarayanan, S.

doi:10.1016/j.ifacol.2022.07.550

Cited by 3 publications

(7 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For this property, the ANN model with three layers and 5000 epochs provided the best model with mean absolute error (MAE) and MAPE of the testing set being 0.008% and 0.901%, respectively, and with R 2 values of 0.968 and 0.963, respectively, for training and testing datasets. As shown by Alizadeh et al, [ 5 ] linear regression models such as principal component regression (PCR) and PLS were also providing very reliable results for predicting density with RMSECV of 4.5 × 10 −5 . The relationship between viscosity and infrared spectroscopy appears to be the most difficult property to predict from FTIR as compared to other properties.…”

Section: Resultsmentioning

confidence: 98%

“…Hence, determining their physicochemical properties is valuable to production specialists and reservoir engineers. [ 5 ] Over the last few decades, several correlations have been developed to estimate crude oil properties. However, these correlations may be helpful only in regional geological provinces and may not provide satisfactory results when applied to crude oils from other regions.…”

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…As stated earlier, rapid and significant advances in multivariate statistics and machine learning methods have made it possible to predict many properties of interest whose measurement is costly, difficult-to-measure, or eco-unfriendly. [5] Furthermore, analytical procedures which are less dependent on sample size [6] are now available for mapping out the relationship between the easily available measurements, such as spectra, and the difficult-to-obtain measurements, such as the viscosity of crude oil. [5] In recent years, infrared spectroscopy has shown to be a promising tool in the quantitative analysis of petroleum, diesel, and biodiesel.…”

Section: Introductionmentioning

confidence: 99%

“…[5] Furthermore, analytical procedures which are less dependent on sample size [6] are now available for mapping out the relationship between the easily available measurements, such as spectra, and the difficult-to-obtain measurements, such as the viscosity of crude oil. [5] In recent years, infrared spectroscopy has shown to be a promising tool in the quantitative analysis of petroleum, diesel, and biodiesel. Santos et al [7] analyzed diesel samples by Fourier transform infrared spectroscopy (FTIR) and FTIR-Raman spectroscopy using partial least squares regression (PLS) and artificial neural networks (ANNs).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Application of artificial neural network for prediction of 10 crude oil properties

Alizadeh

Samavedham

et al. 2023

Can J Chem Eng

Self Cite

View full text Add to dashboard Cite

This study aims to develop an industrially reliable and accurate method to estimate crude oil properties from their Fourier transform infrared spectroscopy (FTIR) spectra. We used the complete FTIR spectral data of selected crude oil samples from seven different Canadian oil fields to predict 10 important crude oil properties using artificial neural networks (ANNs). The predicted properties include specific gravity, kinematic viscosity, total acid number, micro carbon content, and production of light and heavy naphtha, Kero, and distillate in oil refineries. The 107 different (65 light oil and 42 heavy/medium oil samples) crude oil samples used in this study came from seven oil fields and reservoirs across Canada. In line with standard practice, we used 80% of the dataset for training the ANN models and used the remaining 20% of the crude oil samples to test the models. In the ANN analysis, the mean squared error (MSE) was used as the loss function in models, and the mean absolute prediction error (MAPE) was used as a reference to compare the performance of different neural networks constructed with different numbers of layers. This work demonstrates that FTIR spectroscopy is a promising technique that provides rapid and accurate estimates for the oil properties of interest to the industry. A comparison of the values predicted by the validated ANN models and their corresponding measured (actual) values showed excellent prediction with the acceptable range of error (below 15%) aimed for by our industry partner for all properties except viscosity, for which building models based on the natural logarithmic values of measured viscosities significantly improved the results.

show abstract

Section: Resultsmentioning

confidence: 98%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Application of artificial neural network for prediction of 10 crude oil properties

Alizadeh

Samavedham

et al. 2023

Can J Chem Eng

Self Cite

View full text Add to dashboard Cite

show abstract

Deep chemometrics using one‐dimensional convolutional neural networks for predicting crude oil properties from FTIR spectral data

Ta,

Alizadeh,

Samavedham

et al. 2023

Can J Chem Eng

View full text Add to dashboard Cite

The determination of physicochemical properties of crude oils is a very important and time‐intensive process that needs elaborate laboratory procedures. Over the last few decades, several correlations have been developed to estimate these properties, but they have been very limited in their scope and range. In recent years, methods based on spectral data analysis have been shown to be very promising in characterizing petroleum crude. In this work, the physicochemical properties of crude oils using Fourier transform infrared (FTIR) spectrums are predicted. A total of 107 samples of FTIR spectral data consisting of 6840 wavenumbers is used. One dimensional convolutional neural networks (CNNs) were used employing FTIR spectral data as the one‐dimensional input and Keras and TensorFlow were used for model building. The Root Mean Square Error decreased from 160 to around 60 for viscosity when compared to previous machine learning methods like partial least squares (PLS), principal component regression (PCR), and partial least squares regression with genetic algorithm (PLS‐GA) on the same data. The important hyper‐parameters of the CNN were optimized. In addition, a comparison of results obtained with different neural network architectures is presented. Some common preprocessing techniques were also tested on the spectral data to determine their impact on model performance. To increase interpretability, the intermediate neural network layers were analyzed to reveal what the convolutions represented, and sensitivity analysis was done to gather key insights about the wavenumbers that were the most important for prediction of the crude oil properties using the neural network.

show abstract

Multiple linear regression and gene expression programming to predict fracture density from conventional well logs of basement metamorphic rocks

Hasan,

M. Tóth

2024

J Petrol Explor Prod Technol

View full text Add to dashboard Cite

Fracture identification and evaluation requires data from various resources, such as image logs, core samples, seismic data, and conventional well logs for a meaningful interpretation. However, several wells have some missing data; for instance, expensive cost run for image logs, cost concern for core samples, and occasionally unsuccessful core retrieving process. Thus, a majority of the current research is focused on predicting fracture based on conventional well log data. Interpreting fractures information is very important especially to develop reservoir model and to plan for drilling and field development. This study employed statistical methods such as multiple linear regression (MLR), principal component analysis (PCA), and gene expression programming (GEP) to predict fracture density from conventional well log data. This study explored three wells from a basement metamorphic rock with ten conventional logs of gamma rays, thorium, potassium, uranium, deep resistivity, flushed zone resistivity, bulk density, neutron porosity, sonic porosity, and photoelectric effect. Four different methods were used to predict the fracture density, and the results show that predicting fracture density is possible using MLR, PCA, and GEP. However, GEP predicted the best fracture density with R2 > 0.86 for all investigated wells, although it had limited use in predicting fracture density. All methods used highlighted that flushed zone resistivity and uranium content are the two most significant well log parameters to predict fracture density. GEP was efficient for use in metamorphic rocks as it works well for conventional well log data as the data is nonlinear, and GEP uses nonlinear algorithms.

show abstract

Determination of Density and Viscosity of Crude Oil Samples from FTIR Data using Multivariate Regression, Variable Selection and Classification

Cited by 3 publications

References 11 publications

Application of artificial neural network for prediction of 10 crude oil properties

Application of artificial neural network for prediction of 10 crude oil properties

Deep chemometrics using one‐dimensional convolutional neural networks for predicting crude oil properties from FTIR spectral data

Multiple linear regression and gene expression programming to predict fracture density from conventional well logs of basement metamorphic rocks

Contact Info

Product

Resources

About