Multi-objective optimization of thermophysical properties of multiwalled carbon nanotubes based nanofluids

Accurate prediction of dynamic viscosity of polyalpha-olefin boron nitride nanofluids using machine learning

This study focuses on predicting the dynamic viscosity of nanofluids, specifically Polyalpha-Olefin-hexagonal boron nitride (PAO-hBN) using machine learning models. The primary goal of this research is to assess and contrast the effectiveness of three distinct machine learning models: Support Vector Regression (SVR), Artificial Neural Networks (ANN), and Adaptive Neuro-Fuzzy Inference System (ANFIS). The main objective is the identification of a model that demonstrates the highest level of accuracy in predicting a nanofluid's viscosity namely, PAO-hBN nanofluids. The models were trained and validated using 540 experimental data points, where the mean square error (MSE) and the coefficient of determination R² were utilized for performance evaluation. The results demonstrated that all three models could predict the viscosity of PAO-hBN nanofluids accurately, but the ANFIS and ANN models outperformed the SVR model. The ANFIS and ANN models had similar performance, but the ANN model was preferred due to its faster training and computation time. The optimized ANN model had an R² of 0.99994, which indicates a high level of accuracy in predicting the viscosity of PAO-hBN nanofluids. The elimination of the shear rate parameter from the input layer improved the accuracy of the ANN model to an absolute relative error of less than 1.89% over the full temperature range (-19.7 °C-70 °C) compared to 11% in the traditional correlation-based model. These results suggest that the use of machine learning models can significantly improve the accuracy of predicting the viscosity of PAO-hBN nanofluids. Overall, this study demonstrated that the use of machine learning models, specifically ANN, can be effective in predicting PAO-hBN nanofluids' dynamic viscosity. The findings provide a new perspective on how to predict the thermodynamic properties of nanofluids with high accuracy, which could have important applications in various industries.

Automated assessment of human engineered heart tissues using deep learning and template matching for segmentation and tracking

The high rate of drug withdrawal from the market due to cardiovascular toxicity or lack of efficacy, the economic burden, and extremely long time before a compound reaches the market, have increased the relevance of human in vitro models like human (patient-derived) pluripotent stem cell (hPSC)-derived engineered heart tissues (EHTs) for the evaluation of the efficacy and toxicity of compounds at the early phase in the drug development pipeline. Consequently, the EHT contractile properties are highly relevant parameters for the analysis of cardiotoxicity, disease phenotype, and longitudinal measurements of cardiac function over time. In this study, we developed and validated the software HAARTA (Highly Accurate, Automatic and Robust Tracking Algorithm), which automatically analyzes contractile properties of EHTs by segmenting and tracking brightfield videos, using deep learning and template matching with sub-pixel precision. We demonstrate the robustness, accuracy, and computational efficiency of the software by comparing it to the state-of-the-art method (MUSCLEMOTION), and by testing it with a data set of EHTs from three different hPSC lines. HAARTA will facilitate standardized analysis of contractile properties of EHTs, which will be beneficial for in vitro drug screening and longitudinal measurements of cardiac function.

Modelling and optimization study to improve the filtration performance of fibrous filter

Fibrous filter made up of non-woven material was utilized in many industrial applications for increasing the collection efficiency and the quality factor. But there exists a competing effect among the fibre diameter, filtration efficiency, pressure drop, and sometime type of aerosol (liquid or solid) plays a crucial role in the performance of the fibrous filter. To avoid overdesigning of the filter along with better performance, optimum set of parameters are to be decided before the manufacturing process. In the current effort, the desirability approach and along with the "Response Surface Methodology (RSM)" were considered to optimize filtration efficiency and pressure drop simultaneously. In this perspective, the impact of Filtration velocity (v), Basis weight (φ), Particle diameter (dp), and Packing fraction (α) on filtration efficiency (η) and pressure drop (Pd) was studied. Based on the outcome, the predicted values lie within experimental data through smart agreement. The maximum percentage (%) error was only 3% and 6% filtration efficiency (η) and pressure drop (Pd), which determine the effectiveness of this useful model. The most dominant factor which affects the filtration efficiency (η) was found to be the Basis weight (φ), followed by packing fraction. However, in the case of pressure drop, the most dominant factors were filtration speed followed by the pachining fraction. Moreover, artificial neural network (ANN) models are developed for the prediction of filtration efficiency and pressure drop. The model accuracy has been estimated by calculating "Mean Square Error (MSE), Mean Absolute Error (MAE), and coefficient of determination (R2)". Both models show promising results when compared with experimental data with the R2 value of 98.50-99.86. The optimized values of the maximum filtration efficiency and minimum pressure drop simultaneously were obtained for v = 5, φ = 59.60, dp = 52.23, α = 0.24 according to desirability approach.