A novel investigation using thermal modeling and optimization of waste pyrolysis reactor using finite element analysis and response surface methodology.

The influence of humans on the environment is growing drastically and is pervasive. If this trend continues for a longer time, it can cost humankind, social and economic challenges. Keeping this situation in mind, renewable energy has paved the way as our saviour. This shift will not only help in reducing pollution but will also provide immense opportunities for the youth to work. This work discusses about various waste management strategies and discusses the pyrolysis process in details. Simulations were done keeping pyrolysis as the base process and by varying parameters like feeds and reactor materials. Different feeds were chosen like Low-Density Polyethylene (LDPE), wheat straw, pinewood, and a mixture of Polystyrene (PS), Polyethylene (PE), and Polypropylene (PP). Different reactor materials were considered namely, stainless steel AISI 202, AISI 302, AISI 304, and AISI 405. AISI stands for American Iron and Steel Institute. AISI is used to signify some standard grades of alloy steel bars. Thermal stress and thermal strain values and temperature contours were obtained using simulation software called Fusion 360. These values were plotted against temperature using graphing software called Origin. It was observed that these values increased with increasing temperature. LDPE got the lowest values for stress and stainless steel AISI 304 came out to be the most feasible material for pyrolysis reactor having the ability to withstand high thermal stresses. RSM was effectively used to generate a robust prognostic model with high efficiency, R2 (0.9924-0.9931), and low RMSE (0.236 to 0.347). Optimization based on desirability identified the operating parameters as 354 °C temperature and LDPE feedstock. The best thermal stress and strain responses at these ideal parameters were 1719.67 MPa and 0.0095, respectively.

Data on Vulnerability Detection in Android.

The data in this article have been collaborated from mainly four sources- Google Playstore, Wandoujia (third party app store market), AMD and Androzoo. These data include ~85,000 APKs (Android Package Kit), both malicious and benign from these data sources. Static and dynamic features are extracted from these APK files, and then supervised machines learning algorithms are employed for malware detection in Android. This data article also provides the Python code for data analysis. For feature extraction, a generic algorithm has also been incorporated, thereby, selecting important and relevant feature subset. Conclusive results obtained from this data set are further comprehended and interpreted in our latest research study "A Novel Parallel Classifier Scheme for Vulnerability Detection in Android" (Garg et al., 2018). This proved to be precious contribution for ensembling classifiers in machine learning to detect malware in Android.