Forecasting of COVID-19 using deep layer Recurrent Neural Networks (RNNs) with Gated Recurrent Units (GRUs) and Long Short-Term Memory (LSTM) cells.

In December 2019, first case of the COVID-19 was reported in Wuhan, Hubei province in China. Soon world health organization has declared contagious coronavirus disease (a.k.a. COVID-19) as a global pandemic in the month of March 2020. Over the span of eleven months, it has rapidly spread out all over the world with total confirmed cases of ~ 41.39 M and causing a total fatality of ~1.13 M. At present, the entire mankind is facing serious threat and it is believed that COVID-19 may have been around for quite some time. Therefore, it has become imperative to forecast the global impact of COVID-19 in the near future. The present work proposes state-of-art deep learning Recurrent Neural Networks (RNN) models to predict the country-wise cumulative confirmed cases, cumulative recovered cases and the cumulative fatalities. The Gated Recurrent Units (GRUs) and Long Short-Term Memory (LSTM) cells along with Recurrent Neural Networks (RNN) were developed to predict the future trends of the COVID-19. We have used publicly available data from John Hopkins University's COVID-19 database. In this work, we emphasize the importance of various factors such as age, preventive measures, and healthcare facilities, population density, etc. that play vital role in rapid spread of COVID-19 pandemic. Therefore, our forecasted results are very helpful for countries to better prepare themselves to control the pandemic.

Forecasting the dynamics of cumulative COVID-19 cases (confirmed, recovered and deaths) for top-16 countries using statistical machine learning models: Auto-Regressive Integrated Moving Average (ARIMA) and Seasonal Auto-Regressive Integrated Moving Average (SARIMA).

Most countries are reopening or considering lifting the stringent prevention policies such as lockdowns, consequently, daily coronavirus disease (COVID-19) cases (confirmed, recovered and deaths) are increasing significantly. As of July 25th, there are 16.5 million global cumulative confirmed cases, 9.4 million cumulative recovered cases and 0.65 million deaths. There is a tremendous necessity of supervising and estimating future COVID-19 cases to control the spread and help countries prepare their healthcare systems. In this study, time-series models - Auto-Regressive Integrated Moving Average (ARIMA) and Seasonal Auto-Regressive Integrated Moving Average (SARIMA) are used to forecast the epidemiological trends of the COVID-19 pandemic for top-16 countries where 70%-80% of global cumulative cases are located. Initial combinations of the model parameters were selected using the auto-ARIMA model followed by finding the optimized model parameters based on the best fit between the predictions and test data. Analytical tools Auto-Correlation function (ACF), Partial Auto-Correlation Function (PACF), Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were used to assess the reliability of the models. Evaluation metrics Mean Absolute Error (MAE), Mean Square Error (MSE), Root Mean Square Error (RMSE) and Mean Absolute Percent Error (MAPE) were used as criteria for selecting the best model. A case study was presented where the statistical methodology was discussed in detail for model selection and the procedure for forecasting the COVID-19 cases of the USA. Best model parameters of ARIMA and SARIMA for each country are selected manually and the optimized parameters are then used to forecast the COVID-19 cases. Forecasted trends for confirmed and recovered cases showed an exponential rise for countries such as the United States, Brazil, South Africa, Colombia, Bangladesh, India, Mexico and Pakistan. Similarly, trends for cumulative deaths showed an exponential rise for countries Brazil, South Africa, Chile, Colombia, Bangladesh, India, Mexico, Iran, Peru, and Russia. SARIMA model predictions are more realistic than that of the ARIMA model predictions confirming the existence of seasonality in COVID-19 data. The results of this study not only shed light on the future trends of the COVID-19 outbreak in top-16 countries but also guide these countries to prepare their health care policies for the ongoing pandemic. The data used in this work is obtained from publicly available John Hopkins University's COVID-19 database.

Parametric study of the vibration-induced repulsion or attraction force on a particle in a viscous fluid cell.

Experiments and three-dimensional direct numerical simulations were performed to investigate the effects of physical parameters on the repulsion or attraction force affecting the motion of a particle oscillating near a solid wall of a fluid cell under microgravity. The following physical parameters were investigated: fluid cell amplitude, fluid and particle densities, angular frequency of the cell vibration, initial distance between the particle centroid and the closest cell wall, particle radius, and dynamic viscosity. Based on the simulations, a nondimensional relation was developed to relate those physical parameters to the repulsion or attraction force affecting the particle. The relation shows that the repulsion or attraction force is increased by the increase in the cell vibration amplitude and frequency and also the force direction would change from attraction to repulsion above a threshold fluid viscosity. Relations to other physical parameters were also studied and are reported. This paper follows our previous work on the physical mechanism of observed repulsion force on a particle in a viscous fluid cell [M. Saadatmand and M. Kawaji, Phys. Rev. E 88, 023019 (2013)].

Mechanism of vibration-induced repulsion force on a particle in a viscous fluid cell.

Space platforms such as the Space Shuttle and International Space Station have been considered an ideal environment for production of protein and semiconductor crystals of superior quality due to the negligible gravity-induced convection. Although it was believed that under microgravity environment diffusive mass transport would dominate the growth of the crystals, some related experiments have not shown satisfactory results possibly due to the movement of the growing crystals in fluid cells caused by small vibrations present in the space platforms called g-jitter. In ground-based experiments, there have been clear observations of attraction and repulsion of a solid particle with respect to a nearby wall of the fluid cell due to small vibrations. The present work is a numerical investigation on the physical mechanisms responsible for the repulsion force, which has been predicted to increase with the cell vibration frequency and amplitude, as well as the fluid viscosity. Moreover, the simulations have revealed that the repulsion force occurs mostly due to the increased pressure in the narrow gap between the particle and the nearest wall.

Immiscible displacement of oil by water in a microchannel: asymmetric flow behavior and nonlinear stability analysis of core-annular flow.

The immiscible displacement of oil by water in a circular microchannel was investigated. A fused silica microchannel with an inner diameter of 250 µm and a length of 7 cm was initially filled with a viscous silicone oil. Only water then was injected into the channel. We describe our flow observations based on the two-dimensional images captured in the middle of the channel. The water finger displaced the oil and left an oil film on the channel wall. While the oil was being displaced at the core, the flow resistance decreased, which resulted in increases in water flow rate and inertia. Eventually, the water finger reached the channel exit and formed a core-annular flow pattern. The wavelength of the waves formed at the oil-water interface also increased with the increase in inertia. The initially symmetric interfacial waves became asymmetric with time. Also, the water core shifted from the center of the channel and left a thinner oil film on one side of the microchannel. Under all flow rates tested in this study, as long as the water was continuously injected, the water core was stable and no breakup into droplets was observed. We also discuss the flow stability based on nonlinear and linear stability analyses performed on the core-annular flow. Compared to the linear analysis, which ignores the inertia effects, the nonlinear analysis, which includes the inertia effects, predicts longer interfacial wavelengths by a factor of 1/sqrt[1-a(o)/2(We(w) + We(o)a(o)(2)/1-a(o)(2))] where We(w) and We(o) are the Weber numbers of the water and the oil phases, respectively, and a(o) is the unperturbed water core radius made dimensionless by the channel radius.

A miniature cell for gas solubility measurements in oils and bitumen.

A miniature cell has been designed and constructed to measure gas solubility in crude oils and bitumen. The cell was made of stainless steel with a total internal volume of 1.835 cc and only an oil sample of 0.4 cc was required for one set of measurements at different pressures. By using this small cell, the waiting time for reaching equilibrium was less than 10 min. The technique was validated by measuring CO(2) gas solubility in two bitumen samples. The results were compared and found to be in very good agreement with available data. The apparatus was also used to study the effect of ashphaltene on CO(2) solubility in bitumen. It was shown that ashphaltene had a negligible effect on CO(2) solubility in bitumen.

A new proposed approach to estimate the thermodiffusion coefficients for linear chain hydrocarbon binary mixtures.

Thermodiffusion behaviors in nonassociating mixtures have an important role in separation processes of the oil industry. The variations of composition and temperature may either lessen or enhance the separation in mixtures. A new model regarding the prediction of thermodiffusion coefficients for linear chain hydrocarbon binary mixtures using the thermodynamics of irreversible process is proposed. The model predicts the net amount of heat transported based on available volume for each molecule. This newly proposed model combined with the perturbed chain statistical associating fluid theory equation of state has been applied to predict thermodiffusion coefficients for binary hydrocarbon mixtures of C(10)-nC(i) (i=5,6,7,15,16,17,18), C(12)-nC(i) (i=5,6,7,8,9), and C(18)-nC(i) (i=5,6,7,8,9,12). Comparisons of the calculated theoretical results with the experimental data show good performance of the proposed model. In particular, this model which is based on the kinetic approaches has been found to be the most reliable and represents a significant improvement over the earlier models.

A new approach to evaluate the thermodiffusion factor for associating mixtures.

Diffusion behaviors in associating mixtures present a larger degree of complexity than nonassociating mixtures. The direction of flow in associating mixtures may change with the variation of composition and temperature. In this paper a new viscous energy model is proposed for predicting the ratio of evaporation energy to viscous energy. The new model was implemented for prediction of thermodiffusion for acetone-water, ethanol-water, and isopropanol-water mixtures. In particular, this approach is implemented to predict the sign changes in the thermodiffusion factor for associating mixtures, which has been a major step forward in thermodiffusion studies for associating mixtures.

Theoretical approach to evaluate thermodiffusion in aqueous alkanol solutions.

Flow due to thermodiffusion may change direction in fluid mixtures with the variation of composition and temperature. This occurrence remains an unraveled phenomenon in petroleum research. Using a modified theoretical approach, this paper evaluates the thermal diffusion factor in alkanol water mixtures, including methanol, ethanol, and isopropanol aqueous mixtures. By combining this approach with an equation of state perturbed chain statistical associating fluid theory and using two adjustable parameters calculated from experimental data, the present approach provides a good agreement for the prediction of thermal diffusion in alkanol water mixtures when compared with the available experimental data. This work reveals that the thermodiffusion in infinite dilutions may play an important role in understanding the thermodiffusion phenomena.

Numerical simulations of bubble motion in a vibrated cell under microgravity using level set and VOF algorithms.

Understanding the stability of fluid interfaces subjected to small vibrations under microgravity conditions is important for designing future materials science experiments to be conducted aboard orbiting spacecraft. During the STS-85 mission, experiments investigating the motion of a large bubble resulting from small, controlled vibrations were performed aboard the Space Shuttle Discovery. To better understand the experimental results, two-and three-dimensional simulations of the experiment were performed using level set and volume-of-fluid interface tracking algorithms. The simulations proved capable of predicting accurately the experimentally determined bubble translation behavior. Linear dependence of the bubble translation amplitude on the container translation amplitude was confirmed. In addition, the simulation model was used to confirm predictions of a theoretical inviscid model of bubble motion developed in a previous study.