# Dynamic Modeling Using Artificial Neural Network of Bacillus Velezensis Broth Cross-Flow Microfiltration Enhanced by Air-Sparging and Turbulence Promoter

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Materials and Methods

#### 2.1. Preparation of Bacillus Velezensis Cultivation Broth

_{4})

_{2}SO

_{4}, 1 g/L of K

_{2}HPO

_{4}and 0.3 g/L of MgSO

_{4}∙7H

_{2}O. pH value of the cultivation medium was set to 7.0 ± 0.2 before sterilization by autoclaving (121 °C, 2.1 bar, 20 min). Inoculum was prepared using nutrient broth (HiMedia Laboratories, Karnataka, India) under the following conditions: temperature 28 °C, external mixing rate 150 rpm, spontaneous aeration, duration 48 h. Cultivation medium in the Woulff bottles was inoculated with 10% (v/v) of inoculum. Cultivation conditions were as follows: temperature 28 °C, external mixing rate 150 rpm, aeration rate 0.75 vvm (volume of air∙volume of liquid

^{−1}∙min

^{−1}), duration 96 h. The resulting cultivation broth was used as a feed mixture for the microfiltration experiments.

#### 2.2. Microfiltration Experiments

^{2}. Apparatus used in microfiltration experiments was described by Jokić et al. [16]. Recirculation of retentate and permeate in microfiltration experiments had assured constant feed conditions. The temperature was adjusted to 25 °C. Permeate flux (J

_{P}, L∙m

^{−2}∙h

^{−1}) value was calculated using the following equation (Equation (1)):

_{P}(10 mL), and A (m

^{2}) is the effective filtration area.

^{®}F 201AV, Bronkhorst, Ruurlo, The Netherlands). Experimental variables and their values used in the Box–Behnken experimental plan (3

^{3}—three variables varied at three levels) for microfiltration experiments are listed in Table 1.

#### 2.3. Data Compilation

_{normal}and J

_{P}are normalized permeate flux value and measured permeate flux value, respectively, J

_{max}and J

_{min}are maximal and minimal value of permeate flux in the series of experimental data, respectively, and Δ

^{U}and Δ

^{L}are upper and lower values of normalization limit, respectively (0.01 for each limit).

#### 2.4. Artificial Neural Network Modelling

^{−10}. The MSE for each neural network model was calculated for increasing number of neurons in the hidden layer from one to 15 neurons using Equation (3):

_{exp,i}is ith experimental flux value and J

_{pred,i}is ith flux value predicted by the neural network.

^{2}), calculated by Equation (4). The neural networks were trained 30 times, and the average MSE and R

^{2}were calculated to avoid probabilistic weight selection influence:

_{pred,avg}is the average flux value predicted by the neural network.

_{v}and n

_{h}are number of the neurons in the input and the hidden layer, respectively, i

_{j}is the absolute value of connection weights between the input and the hidden layer neurons, and o

_{j}is the absolute value of connection weights between the hidden and the output layer neurons.

## 3. Results and Discussion

#### 3.1. Effect of Learning Algorithm, Transfer Function and Number of Hidden Layer Neurons

^{2}values for this network were 2.69 × 10

^{−4}and 0.99498, respectively. In case of using the Levenberg–Marquardt algorithm and sigmoid logistic function (network type A), minimum MSE value of 2.60 × 10

^{−4}and maximum R

^{2}value of 0.99539 were achieved for a neural network with 15 neurons in the hidden layer.

^{2}were achieved with 15 neurons in the hidden layer, 2.74 × 10

^{−4}and 0.99513, respectively. On the other hand, when using neural network model with Bayesian regularization and hyperbolic logistic function (network type C), minimal value of MSE (2.74 × 10

^{−4}) and maximal value of R

^{2}(0.99515) were achieved with 14 neurons in the hidden layer.

#### 3.2. Verification of the Neural Network Model

^{2}) value of 0.99224 suggested that the linear regression equation for permeate flux could not explain less than 0.8% of the variations in the system. In other words, the majority of the data are close to the line which represents the ideal fitting of the experimental data (full line in the Figure 2, which represents ideal fitting by the linear model). This implies very good prediction consistency of the neural network model. Detailed estimation of neural network capability to predict permeate flux value during cross-flow microfiltration of Bacillus velezensis cultivation broth has been investigated using the analysis of absolute relative error (Table 3). The neural network model was able to predict 85% of the data with error less than 10%.

^{−1}and for air-sparging experiments superficial air velocity was set at 0.2 m∙s

^{−1}. The experiments were undertaken with and without static mixer. The results of the simulation experiments are given in Figure 3.

#### 3.3. Relative Importance of the Input Variables

## 4. Conclusions

## Author Contributions

## Funding

## Conflicts of Interest

## References

- Ruiz-García, C.; Béjar, V.; Martínez-Checa, F.; Llamas, I.; Quesada, E. Bacillus velezensis sp. nov., a surfactant-producing bacterium isolated from the river Vélez in Málaga, southern Spain. Int. J. Syst. Evol. Microbiol.
**2005**, 55, 191–195. [Google Scholar] [CrossRef] [Green Version] - Fan, B.; Wang, C.; Song, X.; Ding, X.; Wu, L.; Wu, H.; Gao, X.; Borriss, R. Bacillus velezensis FZB42 in 2018: The Gram-Positive Model Strain for Plant Growth Promotion and Biocontrol. Front. Microbiol.
**2018**, 9, 2491. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Meena, K.R.; Tandon, T.; Sharma, A.; Kanwar, S.S. Lipopeptide antibiotic production by Bacillus velezensis KLP2016. J. Appl. Pharm. Sci.
**2018**, 8, 91–98. [Google Scholar] [CrossRef] [Green Version] - Wang, C.; Zhao, D.; Qi, G.; Mao, Z.; Hu, X.; Du, B.; Liu, K.; Ding, Y. Effects of Bacillus velezensis FKM10 for Promoting the Growth of Malus hupehensis Rehd. and Inhibiting Fusarium verticillioides. Front. Microbiol.
**2020**, 10, 2889. [Google Scholar] [CrossRef] [PubMed] - Rabbee, M.F.; Ali, S.; Choi, J.; Hwang, B.S.; Jeong, S.C.; Baek, K.-H. Bacillus velezensis: A Valuable Member of Bioactive Molecules within Plant Microbiomes. Molecules
**2019**, 24, 1046. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Pajčin, I.; Vlajkov, V.; Frohme, M.; Grebinyk, S.; Grahovac, M.S.; Mojićević, M.; Grahovac, J.A. Pepper Bacterial Spot Control by Bacillus velezensis: Bioprocess Solution. Microorganisms
**2020**, 8, 1463. [Google Scholar] [CrossRef] - Grahovac, J.; Pajčin, I.; Vlajkov, V.; Rončević, Z.; Dodić, J.; Cvetković, D.; Jokić, A. Xanthomonas campestris biocontrol agent: Selection, medium formulation and bioprocess kinetic analysis. Chem. Ind. Chem. Eng. Q
**2020**, 32. [Google Scholar] [CrossRef] - Tomczak, W.; Gryta, M. Cross-Flow Microfiltration of Glycerol Fermentation Broths with Citrobacter freundii. Membranes
**2020**, 10, 67. [Google Scholar] [CrossRef] [Green Version] - Chew, J.W.; Kilduff, J.; Belfort, G. The behavior of suspensions and macromolecular solutions in crossflow microfiltration: An update. J. Membr. Sci.
**2020**, 601, 117865. [Google Scholar] [CrossRef] - Chang, M.; Zhou, S.; Sun, Q.; Li, T.; Ni, J. Recovery ofBacillus thuringiensisbased biopesticides from fermented sludge by cross-flow microfiltration. Desalination Water Treat.
**2012**, 43, 17–28. [Google Scholar] [CrossRef] - Fan, R.; Ebrahimi, M.; Quitmann, H.; Czermak, P. Lactic acid production in a membrane bioreactor system with thermophilicBacillus coagulans: Fouling analysis of the used ceramic membranes. Sep. Sci. Technol.
**2015**, 50, 2177–2189. [Google Scholar] [CrossRef] - Zhang, Y.; Fu, Q. Algal fouling of microfiltration and ultrafiltration membranes and control strategies: A review. Sep. Purif. Technol.
**2018**, 203, 193–208. [Google Scholar] [CrossRef] - Jokić, A.; Zavargo, Z.; Šereš, Z.; Tekić, M. The effect of turbulence promoter on cross-flow microfiltration of yeast suspensions: A response surface methodology approach. J. Membr. Sci.
**2010**, 350, 269–278. [Google Scholar] [CrossRef] - Jokić, A.; Pajčin, I.; Grahovac, J.A.; Lukić, N.L.; Dodić, J.; Rončević, Z.; Šereš, Z. Energy efficient turbulence promoter flux-enhanced microfiltration for the harvesting of rod-shaped bacteria using tubular ceramic membrane. Chem. Eng. Res. Des.
**2019**, 150, 359–368. [Google Scholar] [CrossRef] - Jokić, A.; Nikolić, N.; Lukić, N.L.; Grahovac, J.A.; Dodić, J.; Rončević, Z.; Šereš, Z. Dynamic Modeling of Streptomyces hygroscopicus Fermentation Broth Microfiltration by Artificial Neural Networks. Period. Polytech. Chem. Eng.
**2019**, 63, 541–547. [Google Scholar] [CrossRef] - Jokić, A.; Pajčin, I.; Grahovac, J.A.; Lukić, N.; Dodić, J.; Rončević, Z.; Šereš, Z. Improving energy efficiency of Bacillus velezensis broth microfiltration in tubular ceramic membrane by air sparging and turbulence promoter. J. Chem. Technol. Biotechnol.
**2019**, 95, 1110–1115. [Google Scholar] [CrossRef] - Hartinger, M.; Napiwotzki, J.; Schmid, E.-M.; Hoffmann, D.; Kurz, F.; Kulozik, U. Influence of Spacer Design and Module Geometry on the Filtration Performance during Skim Milk Microfiltration with Flat Sheet and Spiral-Wound Membranes. Membranes
**2020**, 10, 57. [Google Scholar] [CrossRef] [Green Version] - Krstić, D.M.; Tekić, M.N.; Carić, M.Đ.; Milanović, S.D. The effect of turbulence promoter on cross-flow microfiltration of skim milk. J. Membr. Sci.
**2002**, 208, 303–314. [Google Scholar] [CrossRef] - Ogunbiyi, O.O.; Miles, N.J.; Hilal, N. Comparison of Different Pitch Lengths on Static Promoters for Flux Enhancement in Tubular Ceramic Membrane. Sep. Sci. Technol.
**2007**, 42, 1945–1963. [Google Scholar] [CrossRef] - Liu, Y.; He, G.; Tan, M.; Nie, F.; Li, B. Artificial neural network model for turbulence promoter-assisted crossflow microfiltration of particulate suspensions. Desalination
**2014**, 338, 57–64. [Google Scholar] [CrossRef] - Šereš, L.; Dokić, L.; Ikonić, B.; Šoronja-Simović, D.; Djordjević, M.; Žana, Š.; Maravić, N. Data-driven Modelling of Microfiltration Process with Embedded Static Mixer for Steepwater from Corn Starch Industry. Period. Polytech. Chem. Eng.
**2017**, 62, 114–122. [Google Scholar] [CrossRef] [Green Version] - Pospísil, P. Shear stress-based modelling of steady state permeate flux in microfiltration enhanced by two-phase flows. Chem. Eng. J.
**2004**, 97, 257–263. [Google Scholar] [CrossRef] - Hwang, K.-J.; Hsu, C.-E. Effect of gas–liquid flow pattern on air-sparged cross-flow microfiltration of yeast suspension. Chem. Eng. J.
**2009**, 151, 160–167. [Google Scholar] [CrossRef] - Hwang, K.-J.; Chen, L. Effect of air-sparging on the cross-flow microfiltration of microbe/protein bio-suspension. J. Taiwan Inst. Chem. Eng.
**2010**, 41, 564–569. [Google Scholar] [CrossRef] - Armbruster, S.; Brochard, A.; Lölsberg, J.; Yüce, S.; Wessling, M. Aerating static mixers prevent fouling. J. Membr. Sci.
**2019**, 537–546. [Google Scholar] [CrossRef] - Vatai, G.N.; Krstić, D.M.; Höflinger, W.; Koris, A.K.; Tekic, M.N. Combining air sparging and the use of a static mixer in cross-flow ultrafiltration of oil/water emulsion. Desalination
**2007**, 204, 255–264. [Google Scholar] [CrossRef] - Asghari, M.; Dashti, A.; Rezakazemi, M.; Jokar, E.; Halakoei, H. Application of neural networks in membrane separation. Rev. Chem. Eng.
**2020**, 36, 265–310. [Google Scholar] [CrossRef] - Hermia, J. Blocking Filtration. Application to Non-Newtonian Fluids. In Mathematical Models and Design Methods in Solid-Liquid Separation; Springer Science and Business Media LLC: Berlin/Heidelberg, Germany, 1985; pp. 83–89. [Google Scholar]
- Bowen, W.; Jenner, F. Theoretical descriptions of membrane filtration of colloids and fine particles: An assessment and review. Adv. Colloid Interface Sci.
**1995**, 56, 141–200. [Google Scholar] [CrossRef] - Anis, S.; Hashaikeh, R.; Hilal, N. Microfiltration membrane processes: A review of research trends over the past decade. J. Water Process. Eng.
**2019**, 32, 100941. [Google Scholar] [CrossRef] - Jokić, A.I.; Šereš, L.L.; Milović, N.R.; Šereš, Z.I.; Maravić, N.R.; Šaranovic, Ž.; Dokić, L.P. Modelling of starch industry wastewater microfiltration parameters by neural network. Membr. Water Treat
**2018**, 9, 115–121. [Google Scholar] [CrossRef] - Al-Abri, M.; Hilal, N. Artificial neural network simulation of combined humic substance coagulation and membrane filtration. Chem. Eng. J.
**2008**, 141, 27–34. [Google Scholar] [CrossRef] - Avarzaman, E.M.; Zarafshan, P.; Mirsaeedghazi, H.; Alaeddini, B. Intelligent Modeling of Permeate Flux during Membrane Clarification of Pomegranate Juice. Nutr. Food Sci. Res.
**2017**, 4, 29–38. [Google Scholar] [CrossRef] [Green Version] - Da Silva, I.N.; Flauzino, R.A. An approach based on neural networks for estimation and generalization of crossflow filtration processes. Appl. Soft Comput.
**2008**, 8, 590–598. [Google Scholar] [CrossRef] - Dornier, M.; Decloux, M.; Trystram, G.; Lebert, A. Dynamic modeling of crossflow microfiltration using neural networks. J. Membr. Sci.
**1995**, 98, 263–273. [Google Scholar] [CrossRef] - Hamachi, M.; Cabassud, M.; Davin, A.; Peuchot, M.M. Dynamic modelling of crossflow microfiltration of bentonite suspension using recurrent neural networks. Chem. Eng. Process. Process. Intensif.
**1999**, 38, 203–210. [Google Scholar] [CrossRef] - Chellam, S. Artificial neural network model for transient crossflow microfiltration of polydispersed suspensions. J. Membr. Sci.
**2005**, 258, 35–42. [Google Scholar] [CrossRef] - Aydiner, C.; Demir, I.; Yildiz, E. Modeling of flux decline in crossflow microfiltration using neural networks: The case of phosphate removal. J. Membr. Sci.
**2005**, 248, 53–62. [Google Scholar] [CrossRef] - Fu, R.; Xu, T.; Pan, Z. Modelling of the adsorption of bovine serum albumin on porous polyethylene membrane by back-propagation artificial neural network. J. Membr. Sci.
**2005**, 251, 137–144. [Google Scholar] [CrossRef] - Cheng, L.-H.; Cheng, Y.; Chen, J. Predicting effect of interparticle interactions on permeate flux decline in CMF of colloidal suspensions: An overlapped type of local neural network. J. Membr. Sci.
**2008**, 308, 54–65. [Google Scholar] [CrossRef] - Hilal, N.; Ogunbiyi, O.O.; Al-Abri, M. Neural network modeling for separation of bentonite in tubular ceramic membranes. Desalination
**2008**, 228, 175–182. [Google Scholar] [CrossRef] - Liu, Q.-F.; Kim, S.H.; Lee, S. Prediction of microfiltration membrane fouling using artificial neural network models. Sep. Purif. Technol.
**2009**, 70, 96–102. [Google Scholar] [CrossRef] - Guadix, A.; Zapata, J.E.; Almécija, M.C.; Guadix, E.M. Predicting the flux decline in milk cross-flow ceramic ultrafiltration by artificial neural networks. Desalination
**2010**, 250, 1118–1120. [Google Scholar] [CrossRef] - Hwang, T.-M.; Choi, Y.; Nam, S.-H.; Lee, S.; Oh, H.; Hyun, K.; Choung, Y.-K. Prediction of membrane fouling rate by neural network modeling. Desalination Water Treat.
**2010**, 15, 134–140. [Google Scholar] [CrossRef] [Green Version] - Nandi, B.; Moparthi, A.; Uppaluri, R.; Purkait, M. Treatment of oily wastewater using low cost ceramic membrane: Comparative assessment of pore blocking and artificial neural network models. Chem. Eng. Res. Des.
**2010**, 88, 881–892. [Google Scholar] [CrossRef] - Mhurchú, J.N.; Foley, G.; Havel, J. Modeling process dynamics using a novel neural network architecture: Application to stirred cell microfiltration. Chem. Eng. Commun.
**2010**, 197, 1152–1162. [Google Scholar] [CrossRef] - Soleimani, R.; Shoushtari, N.A.; Mirza, B.; Salahi, A. Experimental investigation, modeling and optimization of membrane separation using artificial neural network and multi-objective optimization using genetic algorithm. Chem. Eng. Res. Des.
**2013**, 91, 883–903. [Google Scholar] [CrossRef] - Nourbakhsh, H.; Emam-Djomeh, Z.; Omid, M.; Mirsaeedghazi, H.; Moini, S. Prediction of red plum juice permeate flux during membrane processing with ANN optimized using RSM. Comput. Electron. Agric.
**2014**, 102, 1–9. [Google Scholar] [CrossRef] - Shahriari, S.; Hakimzadeh, V.; Shahidi, M. Modeling the efficiency of microfiltration process in reducing the hardness, improvement the non-sugar component rejection and purity of raw sugar beet juice. Ukr. Food J.
**2017**, 6, 648–660. [Google Scholar] [CrossRef] - Corbatón-Báguena, M.-J.; Vincent-Vela, M.; Gozálvez-Zafrilla, J.-M.; Álvarez-Blanco, S.; Lora, J.; Catalán-Martínez, D. Comparison between artificial neural networks and Hermia’s models to assess ultrafiltration performance. Sep. Purif. Technol.
**2016**, 170, 434–444. [Google Scholar] [CrossRef] [Green Version] - Demuth, H.; Beale, M. Neural Network Toolbox User’s Guide, 4th ed.; The MathWorks, Inc.: Natick, MA, USA, 2004. [Google Scholar]
- Tanaka, T.; Abe, K.-I.; Asakawa, H.; Yoshida, H.; Nakanishi, K. Filtration characteristics and structure of cake in crossflow filtration of bacterial suspension. J. Ferment. Bioeng.
**1994**, 78, 455–461. [Google Scholar] [CrossRef]

**Figure 1.**The effect of number of hidden neurons, training algorithm and transfer function between the input and the hidden layer to: (

**a**) MSE (mean square error); (

**b**) R

^{2}(coefficient of determination).

**Figure 2.**Normalized experimental permeate flux values plotted against the flux values predicted by the ANN (artificial neural network) model.

Input Variable | Value | |
---|---|---|

Without Static Mixer | With Static Mixer | |

Static mixer (-) | 0 | 1 |

Transmembrane pressure (bar) | 0.2; 0.6; 1.0 | 0.2; 0.6; 1.0 |

Superficial feed velocity (m∙s^{−1}) | 0.43; 0.87; 1.30 | 0.53; 1.06; 1.59 |

Superficial air velocity (m∙s^{−1}) | 0.0; 0.2; 0.4 | 0.0; 0.23; 0.46 |

Filtration time (s) | 0—time to reach stationary flux |

ANN Type | Training Algorithm | Transfer Function | |
---|---|---|---|

Input-Hidden Layer | Hidden-Output Layer | ||

A | trainlm | logsig | puerlin |

B | trainlm | tansig | |

C | trainbr | logsig | |

D | trainbr | tansig |

Absolute Relative Error (%) | <1 | <5 | <10 | <20 | >20 | Sum |
---|---|---|---|---|---|---|

Number of data | 274 | 470 | 199 | 108 | 64 | 1115 |

Percentage of data (%) | 25 | 42 | 18 | 10 | 6 | 100 |

Input | Importance (%) | Rank |
---|---|---|

Static mixer (-) | 13.13 | 3 |

Transmembrane pressure (bar) | 9.44 | 5 |

Superficial air velocity (m∙s^{−1}) | 15.77 | 2 |

Superficial feed velocity (m∙s^{−1}) | 11.36 | 4 |

Filtration time (s) | 50.30 | 1 |

TOTAL: | 100 |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Jokić, A.; Pajčin, I.; Grahovac, J.; Lukić, N.; Ikonić, B.; Nikolić, N.; Vlajkov, V.
Dynamic Modeling Using Artificial Neural Network of *Bacillus Velezensis* Broth Cross-Flow Microfiltration Enhanced by Air-Sparging and Turbulence Promoter. *Membranes* **2020**, *10*, 372.
https://doi.org/10.3390/membranes10120372

**AMA Style**

Jokić A, Pajčin I, Grahovac J, Lukić N, Ikonić B, Nikolić N, Vlajkov V.
Dynamic Modeling Using Artificial Neural Network of *Bacillus Velezensis* Broth Cross-Flow Microfiltration Enhanced by Air-Sparging and Turbulence Promoter. *Membranes*. 2020; 10(12):372.
https://doi.org/10.3390/membranes10120372

**Chicago/Turabian Style**

Jokić, Aleksandar, Ivana Pajčin, Jovana Grahovac, Nataša Lukić, Bojana Ikonić, Nevenka Nikolić, and Vanja Vlajkov.
2020. "Dynamic Modeling Using Artificial Neural Network of *Bacillus Velezensis* Broth Cross-Flow Microfiltration Enhanced by Air-Sparging and Turbulence Promoter" *Membranes* 10, no. 12: 372.
https://doi.org/10.3390/membranes10120372