Papers by Keyword: Statistical Modeling

Abstract: Проаналізовано вітчизняні та зарубіжні вогнегасні покриття та особливості їх використання як елементів пасивного та активного вогнезахисту з матеріалами, що спучуються, а також оцінено доцільність їх застосування для будівельних конструкцій з металевими елементами для підвищення вогнестійкості будівлі та споруди промислового сектору. Отримано залежності, які дають змогу прогнозувати безпечний час перебування людей всередині об’єкта до обвалення будівельних конструкцій у разі виникнення пожежі. Розглядалися можливості виробництва з використанням ряду недорогих композицій, які набухають і підвищують вогнестійкість будівель з металевими елементами за рахунок набухання та адгезії до горючих поверхонь металоконструкцій. Були проведені експерименти для отримання даних про можливість безпечного перебування людей всередині будівлі або споруди до їх руйнування. Встановлено, що набухання та зчеплення з поверхнями металевих конструкцій дозволяють підвищити пожежну безпеку особового складу пожежно-рятувальних та інших служб, залучених до ліквідації таких надзвичайних ситуацій.

Abstract: The process of wire electrical discharge machining (WEDM) of welded joints applied to stainless steel parts has raised important questions due to the effects on the machined surface and even the deep substrates. The paper proposes the possibility of improving working conditions by multiple activation of the effective processing space. The experimental research was performed under the conditions of simultaneous activation, both ultrasonic of the wire electrode and by magnetic activation of the dielectric liquid. The selection of the optimal parameters of magnetic activation was made using an experimental statistical modeling method, with a statistical analysis of the objective functions, using specialized software. The additional introduction of ultrasonic energy into the workspace has led to the rapid refresh of the dielectric fluid and the accelerated evacuation of eroded particles.

Abstract: This paper presents a theoretical yet practical device targeting method to extract typical model parameters of MOSFET devices on wafer for RF analog integrated circuit design. This method employs skewing algorithms with model parameters of existing typical device which are selected by using inter-lot process electrical test parameters. Although this technique can be applied for both n-channel and p-channel MOSFETs, only n-channel devices could be prepared for our experiments in this research. To demonstrate the plausibility of this method, a cascode amplifier is designed to simulate frequency characteristic of S₂₁ with this method.

Abstract: By considering several applications of aluminum based particle reinforced composites especially in automobile, aerospace and electronic industries, in this work, prediction of machinability responses of A356 alloy-SiC particles (5, 10, 15 and 20 vol%) reinforced metal matrix composites is described. Composites were synthesized by vacuum hot pressing (VHP) assisted powder metallurgy (P/M) process. Effect of cutting speed (Vc), feed (f), depth of cut (d) and quantity of SiC (vol %) on machinability of composites in terms of material removal rate (MRR) and resultant cutting forces (F_R) during end milling were investigated. Milling experiments were carried in dry condition based on central composite design and KISTLER dynamometer was used to measure cutting forces. Resultant cutting force values were increased from 21 to 105 N with an increase in ‘f’ and ‘d’, but decreased with increase in ‘Vc’. Increase in machining parameters increased the MRR from 2.3 to 8.6 × 10³ mm³/min and increase in SiC reduced the MRR. Statistical modeling with cubic response equations were used to predict the results and predicted results were closely matching with experimental values.

Abstract: This paper presents a statistical model based on PSP MOSFET model in 40nm CMOS technology. Nine parameters are modified in PSP model to establish the statistical model. The simulation result verifies the accuracy of the model, the error of σ value is controlled within 5% and the error of mean value within 2%. The proposed model could optimize the circuit design in nanometer CMOS technology node.

Abstract: Surface roughness is one of the most important requirements in the cutting process. In this paper, advantages of Taguchi’s experimental design technique, experimental measurements, and analysis of variance (ANOVA) are integrated to statistically investigate surface roughness in hard milling H13 steel (HRc50±1) with coated round inserts under dry cutting conditions. A 12-parameter cubic regression model for surface roughness is developed, which simultaneously takes into consideration the main effects and the interactions of four cutting parameters. A good correlation is observed between the predictive values and the experimental data.

Abstract: In this paper, a novel method for the modeling and characterization of surface structures is presented. The method is based on the nonparametric statistical modeling of the surface structure, represented as a two dimensional field. Using a sample of a surface structure in question, the model reproduces the structure of the surface and estimates the corresponding machining parameters. The applicability of the method is demonstrated on the modeling and characterization of the functional surfaces generated by intentionally invoked chatter during the milling process. Based on the model of the generated surface structures, the corresponding diameter of the toroidal cutter and spindle speed are estimated using the surface structure data as the input for the model.

Abstract: This paper provides the results of the MD simulations of the interaction between a screw dislocation and a copper precipitate in iron. From the results, the screw dislocation has an attractive interaction with the copper precipitate. Also, the dependence of the Critical Resolved Shear Stress (CRSS) for the screw dislocation to break away from the copper precipitate on the size of the precipitate and temperature is studied. Finally, the CRSS obtained by the MD simulations is modeled statistically using a Russel-Brown model. Then we found that an addition of the Peierls stress, which is calculated by the MD simulations, to the Russel-Brown model gives a good prediction of the CRSS.