Hybrid Artificial Neural Network–Genetic Algorithm-Based Technique to Optimize a Steady-State Gas-to-Liquids Plant

Immune Optimization of Welding Sequence for Arc Weld Seams in Ship Medium-Small Assemblies

The arc weld seam is a common form in ship medium-small assemblies. In order to reduce the deformation of the welded parts with an arc weld seam, and then improve the welding quality, research on the optimization of welding sequences based on the artificial immune algorithm is carried out in this paper. First, the formation mechanism of welding deformation is analyzed by the thermo-elastic-plastic finite element method; next, the reduction in the welding deformation is taken as the optimization goal, and the welding sequence optimization model for the arc weld seam is constructed under the condition of boundary constraints; then, an immune clonal optimization algorithm based on similar antibody similarity screening and steady-state adjustment is proposed, and its welding sequence optimization ability is improved through antibody screening and median adjustment. Finally, the welding sequence optimization tests are carried out based on the Ansys platform. Numerical tests of a typical arc weld seam show that different welding sequences will cause different welding deformations, which verifies the importance of welding sequence optimization. Furthermore, the numerical test results of four different types of welds in ship medium-small assemblies demonstrated that the use of distributed optimization algorithms for welding sequence optimization can help reduce the amount of welding deformations, and the immune clonal algorithm, based on antibody similarity screening and steady-state adjustment, achieves the optimal combination of the welding sequence. Compared with the other three optimization algorithms, the maximum welding deformation caused by the welding sequence optimized by the proposed immune clonal algorithm is reduced by 3.1%, 4.0%, and 3.4%, respectively, the average maximum welding deformation is reduced by 3.5%, 5.5%, and 4.7%, respectively, and the convergence generation of the optimization algorithm is reduced by 16.8%, 13.1% and 14.5%, respectively, which further verifies the effectiveness and superiority of the proposed immune clonal algorithm in the optimization of welding sequences.

Kinetic Study of Product Distribution Using Various Data-Driven and Statistical Models for Fischer-Tropsch Synthesis

Three modeling techniques, namely, a radial basis function neural network (RBFNN), a comprehensive kinetic with genetic algorithm (CKGA), and a response surface methodology (RSM), were used to study the kinetics of Fischer-Tropsch (FT) synthesis. Using a 29 × 37 (4 independent process parameters as inputs and corresponding 36 responses as outputs) matrix with total 1073 data sets for data training through RBFNN, the established model is capable of predicting hydrocarbon product distribution i.e., the paraffin formation rate (C₂-C₁₅) and the olefin to paraffin ratio (OPR) within acceptable uncertainties. With additional validation data sets (15 × 36 matrix with total 540 data sets), the uncertainties of using three different models were compared and the outcomes were: RBFNN (±5% uncertainties), RSM (±10% uncertainties), and CKGA (±30% uncertainties), respectively. A new effective strategy for kinetic study of the complex FT synthesis is proposed: RBFNN is used for data matrix generation with a limited number of experimental data sets (due to its fast converge and less computation time), CKGA is used for mechanism selections by the Langmuir-Hinshelwood-Hougen-Watson (LHHW) approach using a genetic algorithm to find out potential reaction pathways, and RSM is used for statistical analysis of the investigated data matrix (generated from RBFNN through central composite design) upon responses and subsequent singular/multiple optimizations. The proposed strategy is a very useful and practical tool in process engineering design and practice for the product distribution during FT synthesis.

In-silico Design of Aryl and Aralkyl Amine-Based Triazolopyrimidine Derivatives with Enhanced Activity Against Resistant Plasmodium falciparum

A blend of genetic algorithm with multiple linear regression (GA-MLR) method was utilized in generating a quantitative structure–activity relationship (QSAR) model on the antimalarial activity of aryl and aralkyl amine-based triazolopyrimidine derivatives. The structures of derivatives were optimized using density functional theory (DFT) DFT/B3LYP/6–31 + G* basis set to generate their molecular descriptors, where two (2) predictive models were developed with the aid of these descriptors. The model with an excellent statistical parameters; high coefficient of determination (R2) = 0.8884, cross-validated R2 (Q2cv) = 0.8317 and highest external validated R2 (R2pred) = 0.7019 was selected as the best model. The model generated was validated through internal (leave-one-out (LOO) cross-validation), external test set, and Y-randomization test. These parameters are indicators of robustness, excellent prediction, and validity of the selected model. The most relevant descriptor to the antimalarial activity in the model was found to be GATS6p (Geary autocorrelation—lag 6/weighted by polarizabilities), in the model due to its highest mean effect. The descriptor (GATS6p) was significant in the in-silico design of sixteen (16) derivatives of aryl and aralkyl amine-based triazolopyrimidine adopting compound DSM191 with the highest activity (pEC50 = 7.1805) as the design template. The design compound D8 was found to be the most active compound due to its superior hypothetical activity (pEC50 = 8.9545).