Fig. 5: Construction of VII viscosity and PVI prediction models using SR. | npj Computational Materials

Fig. 5: Construction of VII viscosity and PVI prediction models using SR.

From: Exploring high-performance viscosity index improver polymers via high-throughput molecular dynamics and explainable AI

Fig. 5

For viscosity as the target property, a frequency of different descriptors appearing in formulas (normalized) at 2000 iterations; b error (MSE) of formulas versus complexity; c comparison between MD-calculated viscosity and SR (Formula 1) predicted viscosity. Symbol-descriptor correspondence: viscosity (η), K_bond_ave (\({K}_{b}\)), AATSC3dv (\({A}_{c}\)), MW_ratio (\({M}_{w}\)), HybRatio (\({H}_{r}\)), see Table S3 for specific meanings. At a concentration of 9.66 wt%, d radius of gyration of VII in PEC7 versus main chain length (main chain length in units of reciprocal C-C bond length); e effect of shear rate on PIB radius of gyration, and shear rate \(1.5\times {10}^{9}1/s\) corresponds to the turning point where the radius of gyration starts to increase rapidly. PIB, PMA, OCP, HSD, PE, and MSC are viscosity index improver polymers, PEC7 is ester base oil, and their structural formulas are shown in Fig. S2.

Back to article page