Non-spherical solid-non-Newtonian liquid fluidization and ANN modelling: Minimum fluidization velocity

A novel pore-array intensified tube-in-tube microchannel (PA) with high throughput was developed to couple with a high-shear mixer (PAHSM) for improving its mixing performance. Effects of structural and operating parameters were intensively investigated with the Villermaux/Dushman system. The results indicated that the mixing performance in PAHSM was significantly impacted by the structural parameters, and it increased with both the rotor speed and flow rate but not with the flow ratio. The micromixing time estimated with the incorporation model could reach 10–4 s. Compared with other mixing devices, the PA liquid distributor can significantly enhance the mixing performance of HSM with high throughput but low pressure drop. An artificial neural network (ANN) was applied to correlate the micromixing time with the investigated parameters, and the comparison between the model and experimental data indicated that it was an efficient method to fit experimental data in the PAHSM.

Section: Artificial Neural Network (Ann) Modelmentioning

confidence: 99%

Mixing Performance of an Inline High-Shear Mixer with a Novel Pore-Array Liquid Distributor

Xia

Zhao

et al. 2019

“…What is more, the correlation focusing on span has not been proposed, so Back Propagation Neural Network (BPNN) is conducted to fit the nonlinear relationship of parameters with d 32 and span . ,, In this ANN method, the input layer contains five parameters N , Q C , Φ, h r , and the number of stator openings, and all 60 sets of experimental data of the blind-side rotor-circular opening stator assembly are divided into training sets (80%) and testing sets (20%) and normalized before use. The Sigmoid function is used as the activation function and Levenberg–Marquardt as the training algorithm. , The obtained ANN structure is shown in Figure S8 in the Supporting Information. The prediction results in Figure show that the coefficient of determination ( R 2 ) for the estimation of d 32 is 0.9991 and 0.9988 for the training and testing sets, respectively, and R 2 for estimating span is 0.9993 and 0.9907, respectively, which indicates that the ANN structure allows for relatively accurate estimation in this study.…”

Section: Resultsmentioning

confidence: 99%

Novel Rotor-Stator Assembly Promotes the Emulsification Performance in an Inline High-Shear Mixer

Zhao

Guo

et al. 2022

For industrial emulsification processes, it remains a challenge to obtain desirable droplet size distribution (DSD) as well as a smaller span via facile and efficient routes. In this work, a novel blind-side rotor is designed to assemble with the stator of a laboratory-scale inline high-shear mixer equipped with an annular cylinder inside the mixing chamber (bRS HSM). The effects of operating parameters, structural parameters on the average droplet size, DSD, and span are experimentally investigated and compared with those of commercially used toothed-type rotor-stator assembly. In addition, computational fluid dynamics calculations are performed to investigate the influences on the flow field of the fluid within the HSM. The results show that the structure of the bRS HSM not only generates an optimal span but also reduces the energy consumption of the mixer, which is more advantageous for production of emulsions with narrower DSD. In the bRS HSM, Sauter mean diameter (d 32) and span decrease with increasing rotor speed but decreasing flow rate of the continuous phase, while span decreases and then increases with increasing volume fraction of the dispersed phase. Meanwhile, the effect of structural parameters on span is nonmonotonic. Furthermore, the artificial neural network (ANN) model is used to perform the correlation of d 32 and span, providing an estimation method to optimize the operating and structural parameters of the inline HSM in the emulsification process.

“…It is difficult to establish a simple mathematical relationship between the micromixing time and various operating conditions and structural parameters of the in-line HSM due to the uncertainty and complexity of the reaction system. Here, we used the feed-forward backpropagation neural network (BPNN) , to establish the artificial neural network (ANN) model for predicting the micromixing time under different operating conditions and structural parameters of the in-line HSM. The input layer was designed to include four input parameters, involving rotor tip speed ( u t ), solid concentration ( C M ), rotor teeth number ( N t ), and shear gap (δ).…”

Section: Resultsmentioning

confidence: 99%

Evaluation on Micromixing of a Continuous Solid Particle Flow in In-Line HSMs by Experiments and Artificial Intelligence Approaches

Liu

Zhang

Guo

et al. 2022

The intensification of micromixing in an in-line HSM assisted with a continuous solid particle flow was investigated by the Villermaux/Dushman method. The ANN model was established to predict the micromixing time under different operating conditions and structural parameters. The results suggested that increasing the rotor tip speed can effectively improve micromixing in the in-line HSM. The continuous solid particle flow used as the buffer and acid solutions and increasing the proportion of large microparticles significantly intensified the micromixing. In the aspect of structural parameters, increasing the rotor teeth number improved the micromixing at low rotor tip speeds, and increasing the shear gap improved the micromixing at high rotor tip speeds. Then, the optimal structural parameters corresponding to the minimum micromixing time of 0.139 ms were determined by the incorporation model. This work can provide useful guidance for the application of in-line HSMs in multiphase processes involving solid particles.