An Extensive Study and Review of Privacy Preservation Models for the Multi-Institutional Data

The deep learning models hold considerable potential for clinical applications, but there are many challenges to successfully training deep learning models. Large-scale data collection is required, which is frequently only possible through multi-institutional cooperation. Building large central repositories is one strategy for multi-institution studies. However, this is hampered by issues regarding data sharing, including patient privacy, data de-identification, regulation, intellectual property, and data storage. These difficulties have lessened the impracticality of central data storage. In this survey, we will look at 24 research publications that concentrate on machine learning approaches linked to privacy preservation techniques for multi-institutional data, highlighting the multiple shortcomings of the existing methodologies. Researching different approaches will be made simpler in this case based on a number of factors, such as performance measures, year of publication and journals, achievements of the strategies in numerical assessments, and other factors. A technique analysis that considers the benefits and drawbacks of the strategies is additionally provided. The article also looks at some potential areas for future research as well as the challenges associated with increasing the accuracy of privacy protection techniques. The comparative evaluation of the approaches offers a thorough justification for the research’s purpose.

Numerical Simulation and Experimental Study of F-EMS for Continuously Cast Billet of High Carbon Steel

A final electromagnetic stirring model was developed for billet continuous casting of high carbon steel using the commercial software ANSYS and CFX, and the numerical model was validated by the magnetic flux density measured under a Teslameter CT-3. The magnetic flux density and fluid flow in the liquid pool at the location of final electromagnetic stirring(F-EMS) were calculated by the present numerical model. Meanwhile, the plant trials were carried out to determine the optimum current intensity and frequency of F-EMS for the continuously cast billet of high carbon steel. The numerical results show that, through increasing the current intensity by 100 A, the corresponding increases of magnetic induction intensity, tangential electromagnetic force and flow velocity at the solid/liquid interface in the strand are 0.025 T, 1933 N/m3 and 6.9 cm/s, respectively. Moreover, the industrial trial results showed that for the continuously cast billet of 60 steel, the optimum current intensity and frequency of F-EMS, which is 8.2 m from the meniscus, are respectively 380 A and 6 Hz. With the optimum F-EMS parameters, the significant improvement of center segregation of billet is achieved, and the center carbon segregation index in billet reaches 1.04.

Size analysis of slag eye formed by gas blowing in ladle refining

The formation of slag eye in a gas stirred ladle was studied through cold models and industrial trials. In the cold model,water and sodium tungstate solution were employed to simulate liquid steel,and silicon oil was employed to simulate slag. The simulation results revealed that the gas flow rate and bath height had strong effects on the slag eye size. In particular,the thickness of slag layer played a strong role in the slag eye size. In addition,the slag eye could not be formed when the thickness of the top layer was more than 4 cm in water-silicone oil model.Besides,the section area of vessel had a great impact on the slag eye size. Industrial trials results showed a similar trend that the gas flow rate was very significant on the slag eye size. The predictions of the existing models showed larger predictions deviations compared with the experimental data. Moreover,a new model without fitting parameters was developed based on force balance and mathematical derivation,and verified by the experimental data. The new model provides the prediction with small deviations by comparing with the data acquired from cold models and industrial trials.