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

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.     

Influence of Curing Agents Molecular Structures on Interfacial Characteristics of Graphene/Epoxy Nanocomposites: A Molecular Dynamics Framework

This paper presents a comprehensive molecular dynamics study on the effects of the stoichiometric ratio of epoxy:hardener, hardener's linear and cyclic structure, and number of aromatic rings on the interfacial characteristics of graphene/epoxy nanocomposite. The van der Waals gap and polymer peak density as a function of the type of the hardener is calculated by analyzing the local mass density profile. Additionally, steered molecular dynamics are used to conduct normal pull-out of graphene to study the effect of the mentioned features of hardeners on the interfacial mechanical properties of nanocomposites, including traction force, separation distance, and distribution quality of reacted epoxide rings in the epoxy. Influence of the hardeners on the damage mechanism and its initiation point are also studied by analyzing the evolution of local mass density profile during the normal pull-out simulation. It is seen that stoichiometric ratio and geometrical structure of the hardeners affect the interfacial strength. It is also revealed that the hardener type can change the epoxy damage initiation point. The damage occurs in the interphase region for a higher stoichiometric ratio or cyclic structure of hardener. In comparison, for hardener's lower stoichiometric ratio and non-cyclic structure, failure begins in the epoxy near graphene layers.     

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

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.     

Release of bitumen from mature fine tailings in the presence of a polymeric dispersant

The effect of the addition of a polymeric dispersant, sodium lignosulphonate (LS), on the dispersion of mature fine tailings (MFT) was studied using zeta potential and total organic carbon measurements. Three different types of LSs were investigated to determine the importance of molecular weight and level of anionicity of LSs on the treatment of MFT. The presence of two fractions of bitumen was identified in the tested MFT sample. A small portion of bitumen was found to occur in the form of weakly held aggregates between bitumen and fine solids. This fraction was easily dispersed by small dosages of LS, resulting in bitumen liberation to the tailings surface. A much larger amount of bitumen was found to be strongly attached to the solids, and only very high dosages of LS were capable of partly liberating this fraction. The zeta potential promoted an understanding of the mechanism of adsorption of LS on the particles and liberation of bitumen from MFT. Carbon measurements facilitated determination of the adsorption density of the selected types of LS on solid particles.     

Studying the Role of Gelation Agents in Gelcasting Non-porous Si3N4 Bodies by Pressureless Sintering

Silicon - The monomer content in the gelcasting process affects the kinetics of cross-linking reactions which determines the quality of the gel network structure and the final properties of the...     

Rational Ligand Design of Heteroleptic Iridium (III) Complexes toward Nearly Perfect Horizontal Dipole Orientation for Highly Efficient Red-Emitting Phosphorescent Organic Light-Emitting Diodes

The horizontal orientation of the transition dipole moment of the phosphorescent emitters is understood to be an important factor to enhance the external quantum efficiency (EQE) of organic light-emitting diodes by improving light out-coupling in optical microcavity structures. Here, red-emitting heteroleptic iridium (III) complexes exhibiting an extremely high horizontal ratio of emitting dipole orientation (EDO) and photoluminescence quantum yield (PLQY), as well as longer device operational lifetime, without scarifying any other photophysical properties are reported. The systematic molecular design of main and ancillary ligands in heteroleptic iridium complexes leads to the achievement of both a horizontal EDO of 92% and a PLQY of 98% in the red-emitting phosphorescent devices along with a shorter exciton decay time of 0.71 µs. Accordingly, the red-emitting devices show excellent performances of maximum EQE of 32% and low-efficiency roll-off with the 1931 Commission Internationale de L′Eclariage coordinates of (0.66, 0.34). Therefore, this approach opens the way for further development of new red-emitting iridium complexes pushing the device efficiency toward the theoretical limits.