PDF Publication Title:
Text from PDF Page: 011
Processes 2022, 10, x FOR PEER REVIEW 11 of 19 Processes 2022, 10, 812 11 of 19 Figure 3. A generic optimization framework from Ding et al. [67]. Readapted with permission from Figure 3. A generic optimization framework from Ding et al. [67]. Readapted with permission Ref. [67]. Copyright 2018 Springer Nature. from Ref. [67]. Copyright 2018 Springer Nature. The selection of a suitable and accurate model is critical for a PSA data-driven sur- 3.1.3. Surrogate Model rogate model. A deep-learning method along with an artificial neural network (ANN) can fIunllyorddigeroutot tshime mpolisftyetshsenctoiaml lpawutsaotifopnroocfesrseadcahtainagndCqSuSicaknlydersetdabulcisehtahesucroromgpatuetational model [28]. The backpropagation neural network (BP-NN) is frequently used, due to consumption in the optimization process on the basis of the accuracy of the model, re- its excellent prediction and fitting capabilities. A typical multilayer feedforward neural searchers proposed the surrogate model to replace the detailed model. The surrogate network structure is shown in Figure 4. The basic training process is as follows: first, a model generated from data-driven black-box functions, the central problem of which is number of decision variables and objective functions are selected, and then the training learning to establish an input–output relationship that is as accurate and simple as possi- samples are obtained from the detailed PSA model. Then, the ANN toolbox in MATLAB is ble from data obtained from simulations or experiments, instead of solving the equations. used to take the decision variables of samples as input and the objective function values This model investigates the effect of operation parameters (valve, feed flow rate, adsorp- calculated by the detailed model as output. Finally, the algorithm is allowed to train itself tbioynadtjiumstein)gotnheopbrjoepctoirvtieonfuonfcthtieotnrasin(pinugrsiteyt,,vraelcidoavtieorny,septraondutecstitvsietty.,Ifetnherfigtyobctoaninseudmption) tfhroromugtheatrlairnginegnduomesbenrotofmeexeptethriemreqnutairleamnedntssi,mthuelaptieorcnenretasgueltsso.fthethreesetsare changed and retrained. Lee et al. [60] trained and tested a dynamic-model ANN for H The selection of a suitable and accurate model is critical for a PSA data-drive2n surro- recovery and CO2 capture from the tail gas of hydrogen plants, and they found that the gate model. A deep-learning method along with an artificial neural network (ANN) can dynamic-model-based ANN could precisely predict the dynamic behavior and optimum fully dig out the most essential laws of process data and quickly establish a surrogate performance of an integrated process at a low computational cost. model [28]. The backpropagation neural network (BP-NN) is frequently used, due to its excellent prediction and fitting capabilities. A typical multilayer feedforward neural net- work structure is shown in Figure 4. The basic training process is as follows: first, a num- ber of decision variables and objective functions are selected, and then the training sam- ples are obtained from the detailed PSA model. Then, the ANN toolbox in MATLAB is used to take the decision variables of samples as input and the objective function values calculated by the detailed model as output. Finally, the algorithm is allowed to train itself by adjusting the proportion of the training set, validation set and test set. If the fit obtained from the training does not meet the requirements, the percentages of the three sets arePDF Image | Numerical Research on the Pressure Swing Adsorption Process
PDF Search Title:
Numerical Research on the Pressure Swing Adsorption ProcessOriginal File Name Searched:
processes-10-00812-v2.pdfDIY PDF Search: Google It | Yahoo | Bing
CO2 Organic Rankine Cycle Experimenter Platform The supercritical CO2 phase change system is both a heat pump and organic rankine cycle which can be used for those purposes and as a supercritical extractor for advanced subcritical and supercritical extraction technology. Uses include producing nanoparticles, precious metal CO2 extraction, lithium battery recycling, and other applications... More Info
Heat Pumps CO2 ORC Heat Pump System Platform More Info
| CONTACT TEL: 608-238-6001 Email: greg@infinityturbine.com | RSS | AMP |