Effects of additions of 0.00064, 0.001 and 0.0042 wt.% Bi on the graphite structure in the section thicknesses of 3, 12, 25, 38, 50, 75 and 100 mm of spheroidal graphite cast iron castings containing 2.11 wt.% Si and rare earth (RE) elements (Ce + La + Nd + Pr + Sm + Gd) in the range from 0.00297 to 0.00337 wt.% were analyzed in this paper. Addition of Bi was not necessary for obtaining high nodule count and nodularity higher than 80% in section thicknesses of 3, 12 and 25 mm. RE elements showed a beneficial effect on the nodule count and nodularity in these sections. Nodularity was below 80% in section thicknesses of 38, 50, 75 and 100 mm when Bi was not added. Detrimental effect of RE elements on graphite morphology in these sections was neutralized by adequate addition of Bi. Addition of 0.001 wt.% Bi (ratio of RE/Bi = 3.27) was enough to achieve nodularity above 80% in the section thickness of 38 mm. Nodularity was increased above 80% in section thicknesses of 50, 75 and 100 mm by addition of 0.0042 wt.% Bi (ratio of RE/Bi = 0.78). At the same time, Bi significantly increased the nodule count. Nodularity above 80% and the high nodule count in the section thicknesses of 75 and 100 mm were also achieved by using an external metallic chill in the mold. In this case, addition of Bi was not required.
2
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
This paper presents the application of articial neural networks in the production process of spheroidal graphite cast iron. Backpropagation neural networks have been established to predict the microstructure constituents (ferrite content, pearlite conent, nodule count and nodularity) of speroidal geaphite cast iron using the thermal analysis parameters as inputs. General-ization properties of the developed artificial neural netyworks are very good, which id=s confirmed by a very good accordance between the predicted and the targeted values of the microstructure constituents on a new data that was not included in the training data set.
PL
Praca przedstawia zastosowanie sieci sztucznych neuronów do procesu wytwarzania sferoidalnego żeliwa grafityzowanego. Celem przewidzenia składowych m ikrostruktury (składowej ferrytycznej, składowej perlitycznej, ilość grafitowych wtrąceń kulkowych, kulistość wtrąceń) sferoidalnego żeliwa grafityzowanego, ustalona została jako danych wejściowych. Ogólne własności zbudowanej sieci neuronowej są bardzo dobre, co zostało potwierdzone poprzez bardzo dobrą zgodność przewidywanych i uzyskanych wartości składowych mikrostruktury.
This paper deals with the waste foundry molding sand which originally comes from the casting production. Adsorption of Cu (II) ions on the waste foundry molding sand was studied. Experimental data were processed using adsorption isotherms. Obtained results show that the experimental data are best described by the Langmuir isotherm. The following adsorption capacities are obtained: 7.153 mg/g to 293 K, 8.403 mg/g at 333 K and 9.208 mg/g at 343 K. The kinetics and thermodynamics of the process were analysed. The obtained results indicate that the adsorption process takes place according to the pseudo second order kinetic model with the following constants: 0.438 g/mg min at 293 K, 0.550 g/mg min at 333 K and 1.872 g/mg min at 343 K. The following values of ΔG° were obtained: − 95.49 J/mol at 293 K, − 736.99 J/mol at 333 K and − 1183.46 J/mol at 343 K. The value of ΔH° is − 4.16 kJ/mol and the value of ΔS° is 15.17 J/molK. These results were confirmed by microscopic examinations. The results indicate that the adsorption process of Cu (II) ions on waste foundry molding sand is possible. Results of microscopic examinations show the homogeneity of the surface, which is proof of the chemisorption. Cu (II) ions on the surface of the waste foundry molding sand were detected after adsorption by EDS analysis, which proves the existence of the adsorption process.