Neural network inverse model-based controller for the control of a steel pickling process

Daosud, Wachira; Thitiyasook, Piyanuch; Arpornwichanop, Amornchai; Kittisupakorn, Paisan; Hussain, Mohd Azlan

doi:10.1016/j.compchemeng.2005.06.007

Cited by 47 publications

(24 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A range of new and innovative applications of neural networks in different technological processes of the chemical, metallurgical, mechatronics and biochemical fields, include model reference control, inverse control, knowledge-based supervisory control and hybrid model control (Daosud et al, 2005;Galvanauskas et al, 2004;Huang, 2003;Hunt et al, 1992;Lina et al, 2001;Nagy, 2007;Ng and Hussain, 2004;Peres et al, 2001;Salman, 2005;Uraikul et al, 2007;Wang et al, 2008). Neural networks based hybrid modeling is one of the attractive choices and has been used effectively to design control strategies for bioprocesses (Alford, 2006;Azevedo et al, 1997;Gadkar et al, 2005;James et al, 2002;Komives and Parker, 2003;Lith et al, 2002;Lubbert and Simutis, 1994;Oliveira, 2004;Peres et al, 2001).…”

Section: Introductionmentioning

confidence: 98%

Sequential adaptive networks: An ensemble of neural networks for feed forward control of L-methionine production

Nayak

Gomes

2009

Chemical Engineering Science

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 98%

Sequential adaptive networks: An ensemble of neural networks for feed forward control of L-methionine production

Nayak

Gomes

2009

Chemical Engineering Science

View full text Add to dashboard Cite

“…Mathematically, the inverse models are expressed as the function of inputs to the model as shown below: (17) The defined neural networks are trained with the Levenberg-Marquardt method where the common objective is to reduce the error between the neural network predicted value and the actual targeted value. The detail of procedure for obtaining the inverse neural network models are define in research [29].The optimum structures is selected by the minimum MSE method [29]. Based on the minimizing MSE error values, it is found that 5 hidden nodes appear to be the best to be applied for the inverse models which will be used as controller in the control strategy.…”

Section: Inverse Neural Network Control (Innc)mentioning

confidence: 99%

Improving of Crystal Size Distribution Control Based on Neural Network-Based Hybrid Model for Purified Terephthalic Acid Batch Crystallizer

Kittisupakorn¹,

Somsong²,

Hussain³

et al. 2017

Self Cite

View full text Add to dashboard Cite

Abstract. A main difficult task in batch crystallization is to control the size distribution of crystal products. Complexity and highly nonlinear dynamic behavior directly affect to model-based control strategies which heavily depend on the rigorous knowledge of crystallization. In this work, neural network-based model predictive control and inverse neural network control strategies are proposed and integrated with an optimization based on neural network-based hybrid model to control temperatures of a purified terephthalic acid batch crystallizer. A neural network-based hybrid model of the batch crystallizer is developed to provide nonlinear dynamic responses used in optimization algorithm for finding an optimal temperature profile related to the quality of a crystal product. Then, the obtained optimal profile is used as set points of the proposed control strategies for improving the crystal product quality. The performances and robustness of the proposed controllers are evaluated in several cases such as for set point tracking and plant/model mismatches. Simulation results show that the neural network-based model predictive control gives the best control performance among the inverse neural network control and a conventional PID controller in all cases.

show abstract

“…The removal of contaminants such as sulphides, inorganic oxide, silicone oil and carbon black can be carried out using mechanical methods, chemical cleaning methods, dry ice blasting, ultrasonic cleaning methods, laser cleaning methods and plasma cleaning methods. Although the above conventional methods [1][2][3][4][5] are applicable to clean moulds, it is impossible to reduce the adhesion between the rubber or plastic products and the mould surfaces. Due to the physical adsorption, interface reaction, and mould shrinkage, adhesion may occur between rubber or plastic parts and moulds.…”

Section: Introductionmentioning

confidence: 99%

Fabrication of anti-adhesion surfaces on aluminium substrates of rubber plastic moulds using electrolysis plasma treatment

et al. 2014

View full text Add to dashboard Cite

An anti-adhesion surface with a water contact angle of 167° was fabricated on aluminium samples of rubber plastic moulds by electrolysis plasma treatment using mixed electrolytes of C6H5O7(NH4)3 and Na2SO4, followed by fluorination. To optimise the fabrication conditions, several important processing parameters such as the discharge voltage, discharge time, concentrations of supporting electrolyte and stearic acid ethanol solution were examined systematically. Using scanning electron microscopy (SEM) to analyse surfaces morphology, micrometer scale pits, and protrusions were found on the surface, with numerous nanometer mastoids contained in the protrusions. These binary micro/nano-scale structures, which are similar to the micro-structures of soil-burrowing animals, play a critical role in achieving low adhesion properties. Otherwise, the anti-adhesion behaviours of the resulting samples were analysed by the atomic force microscope (AFM), Fourier-transform infrared spectrophotometer (FTIR), electrons probe micro-analyzer (EPMA), optical contact angle meter, digital Vickers microhardness (Hv) tester, and electronic universal testing. The results show that the electrolysis plasma treatment does not require complex processing parameters, using a simple device, and is an environment-friendly and effective method. Under the optimised conditions, the contact angle (CA) for the modified anti-adhesion surface is up to 167°, the sliding angle (SA) is less than 2°, roughness of the sample surface is only 0.409μm. Moreover, the adhesion force and Hv are 0. 9KN and 385, respectively.

show abstract

Neural network inverse model-based controller for the control of a steel pickling process

Cited by 47 publications

References 9 publications

Sequential adaptive networks: An ensemble of neural networks for feed forward control of L-methionine production

Sequential adaptive networks: An ensemble of neural networks for feed forward control of L-methionine production

Improving of Crystal Size Distribution Control Based on Neural Network-Based Hybrid Model for Purified Terephthalic Acid Batch Crystallizer

Fabrication of anti-adhesion surfaces on aluminium substrates of rubber plastic moulds using electrolysis plasma treatment

Contact Info

Product

Resources

About