酶膜耦合反应器中反冲过程的神经网络模拟

In the enzymatic membrane reactor for separating casein hydrolysate, backflushing technology has been used to decrease the fouling of the membrane. Predication of the backflushing efficiency poses a complex non-linear problem as the system integrates enzymatic hydrolysis, membrane separation and periodic backflushing together. In this paper an alternative artificial neural network approach is developed to predict the backflushing efficiency as a function of duration and interval. A contour plot of backflushing performance is presented to model these effects, and the backflushing conditions have been optimized as duration of 10 s and interval of 10 min using this neural network. Also, simple neural networks are established to predict the time evolution of flux before and after backflushing. The results predicted by the models are in good agreement with the experimental data, and the average deviations for all the cases are well within ±5%. The neural network approach is found to be capable of modeling the backflushing with confidence.     

Prediction of Swelling Behavior of AT-Isopropylacrylamide Hydrogels in Aqueous Solution of Sodium Chloride

In this paper, a model is presented to correlate and predict the swelling behavior of hydrogels in aqueous solutions of electrolytes. The model is a combination of VERS-model, "phantom network" theory and "free-volume" contribution. The VERS-model is used to calculate Gibbs excess energy; "phantom network" theory to describe the elastic properties of polymer network, and "free-volume" contribution to account for additional difference in the size of the species. To test the model, a series of N-isopropylacrylamide based hydrogels are synthesized by free radical polymerization in oxygen-free, deionized water at 25℃ under nitrogen atmosphere. Then, the degree of swelling of all investigated gels as well as the partition of the solute between the gel phase and the surrounding coexisting liquid phase are measured in aqueous solution of sodium chloride. The model test demonstrates that the swelling behavior correlated and predicted by the model agrees with the experimental data within the experimental uncert     

基于小波网络的非线性动态过程建模和在环氧氯丙烷生产过程中的应用

A type of wavelet neural network, in which the scale function is adopted only,is proposed in this paper for non-linear dynamic process modelling.Its network size is decreased significantly and the weight coefficients can be estimated by a linear algorithm.The wavelet neural network holds some advantages supeiior to other types of neural networks.First, its network structure is easy to specify based on its theoretical analysis and intuition.Secondly, network training does not rely on stochastic gradient type techniques and avoidd the problem of poor convergence or undesirable local minima.The excellent statistic properties of the weight parameter estimations can be proven here.Both theoretical analysis and simulation study show that the identification method is robust and reliable. Furthermore,a hybrid network structure incorporating first-principle knowledge and wavelet network is developed to solve a commonly existing problem in chemical production processes.Applications of the hybrid network to a practical production process demonstrates that model generalisation capability is significantly improved.     

非牛顿流体非等温充模过程的Level Set方法模拟(英文)

A non-isothermal injection molding process for a non-Newtonian viscous pseudoplastic fluid is simulated. A conservative interface capturing technique and the flow field solving method are coupled to perform a dynamic simulation. The validity of the numerical method is verified by a benchmark problem. The melt interface evolution versus time is captured and the physical quantities such as temperature, velocity and pressure at each time step are obtained with corresponding analysis. A “frozen skin” layer with the thickness increasing versus time during the injection process is found. The fact that the “frozen skin” layer can be reduced by increasing the injection velocity is numerically verified. The fountain flow phenomenon near the melt interface is also captured. Moreover, comparisons with the non-isothermal Newtonian case show that the curvatures of the interface arcs and the pressure con-tours near the horizontal mid-line of the cavity for the non-Newtonian pseudoplastic case is larger than that for the Newtonian case. The velocity profiles are different at different positions for the non-Newtonian pseudoplastic case, while in the case of Newtonian flow the velocity profiles are parabolic and almost the same at different positions.     

液化石油气储罐在喷射火焰作用下的热响应模型

Abstract This paper describes a mathematical model developed to study the behavior of liquefied petroleum gas (LPG) tanks when subjected to jet fire. The model consists of a number of field and zone sub-models which are used to simulate the various physical phenomena taking place during the tank engulfment period. The model can be used to predict the pressure and temperature of the LPG in the tank, the temperature of the wall of tank, and the time of tank explosion. The comparisons between the model predicted results and the test data show good agreement. The results show that the jet fire partially impinging on tank wall led to higher wall temperature and the time to failure was shorter than that in engulfing pool fire. And the exposure of the upper wall in the vapor zone to the fire is more dangerous than that of the LPG contacted wall.     

N-异丙基丙烯酰胺凝胶在NaCl水溶液中溶胀行为的预测研究

In this paper, a model is presented to correlate and predict the swelling behavior of hydrogels in aqueous solutions of electrolytes. The model is a combination of VERS-model, “phantom network“ theory and “free-volume“ contribution. The VERS-model is used to calculate Gibbs excess energy; “phantom network“ theory to describe the elastic properties of polymer network, and “free-volume“ contribution to account for additional difference in the size of the species. To test the model, a series of N-isopropylacrylamide based hydrogels are synthesized by free radical polymerization in oxygen-free, deionized water at 25~C under nitrogen atmosphere. Then, the degree of swelling of all investigated gels as well as the partition of the solute between the gel phase and the surrounding coexisting liquid phase are measured in aqueous solution of sodium chloride. The model test demonstrates that the swelling behavior correlated and predicted by the model agrees with the experimental data within the experimental uncertainty. The phase transition appeared in the experiment, and the influences of the total mass fraction of polymerizable materials ξgel as well as the mole fraction of the crosslinking agent YCR on the swelling behavior of IPAAm-gels can also be predicted correctly. All these show the potential of such model for correlation and prediction of the swelling behavior of hydrogels in aqueous solutions of electrolytes.     

On-line estimation of vapour pressure of stabilized gasoline via ANN’s

The use of artificial neural network based model for the on-line estimation of the Reid Va-por Pressure of stabilized gasoline in a stabilizer after the stripper-reabsorber in the fluid catalyticcracking unit is investigated.The quadratic basis function network(QBFN)which uses a simplequadratic function instead of sigmoid function typically used in back-propagation network is em-ployed.180 sets of historical operation data have been selected for training and testing the QBFN.To overcome the local minimum point which often occurs during the training phase,a new algorithmcombining the simulated annealing approach with the improved GDR has been applied.Furthermore,the developed model has been installed on-line in a refinery for on-line testing.Thetesting results show that the model is sufficiently accurate and it can be used on site as an on-lineanalyzer.     

基于Tent混沌优化的神经网络预测控制

With the unique erggdicity, i rregularity, and.special ability to avoid being trapped in local optima, chaos optimization has been a novel global optimization technique and has attracted considerable attention for application in various fields, such as nonlinear programming problems. In this article, a novel neural network nonlinear predic-tive control （NNPC） strategy baseed on the new Tent-map chaos optimization algorithm （TCOA） is presented. Thefeedforward neural network＇is used as the multi-step predictive model. In addition, the TCOA is applied to perform the nonlinear rolling optimization to enhance the convergence and accuracy in the NNPC. Simulation on a labora-tory-scale liquid-level system is given to illustrate the effectiveness of the proposed method.     

重力热管振荡传热特性RBF神经网络动态建模

The work address the problem of modeling the dynamical oscillating behavior during both unstable and stable operations, of an experimental thermosyphon. A standard RBF artificial neural network-based prediction model was developed for predicting the oscillating heat transfer of thermosyphon by means of input-output experimental measurements with the characteristics of time series. A comparison of prediction values between the RBF network and the MLP network was giving. The precision of RBF network was higher than that of the other neural networks such as BP-MLP network etc. The dynamical model of RBF network could be used to describe, predict and control the heat transfer process of a thermosyphon or a heat pipe system.     

工业PTA氧化过程4-CBA浓度的模糊神经网络模型

A fuzzy neural network (FNN) model is developed to predict the 4-CBA concentration of the oxidation unit in purified terephthalic acid process. Several technologies are used to deal with the process data before modeling.First,a set of preliminary input variables is selected according to prior knowledge and experience. Secondly,a method based on the maximum correlation coefficient is proposed to detect the dead time between the process variables and response variables. Finally, the fuzzy curve method is used to reduce the unimportant input variables.The simulation results based on industrial data show that the relative error range of the FNN model is narrower than that of the American Oil Company (AMOCO) model. Furthermore, the FNN model can predict the trend of the 4-CBA concentration more accurately.     

基于GRNN-NSGA的带嵌件注塑件多目标优化

带嵌件的注塑产品成型过程相较于传统注塑产品较为复杂,产品成型周期和产品质量难以预测。针对这一问题,以带嵌件的静电检测盒为例,运用广义神经网络(GRNN)和非支配排序遗传算法(NSGA-Ⅱ),对注塑成型过程进行控制与优化。以熔体温度、模具温度、注射时间、冷却时间、保压压力和保压时间为输入层,体积收缩率、X方向翘曲变形、Z方向翘曲变形作为输出层,建立GRNN模型。利用正交试验设计得到的样本对神经网络模型进行训练和测试,运用NSGA-Ⅱ对建立的模型进行优化,最终三个目标值分别降低了30.96%、22.76%、15.62%,表明该方法可以对注塑成型过程进行预测和控制。     

注塑过程熔体充填长度的动态神经网络软测量模型

提出用喷嘴压力，注射速度和螺杆位移等在线可测二次变量来预测充填长度的方法，并建立了基于动态神经网络的充填长度模型，验证结果表明，此模型可准确地在线预测各种不同形状模具中的充填长度曲线。     

复杂壳体类塑料件气体辅助注射成型工艺参数优化研究

以21英寸彩电前壳作为研究对象,将Moldflow 2010作为CAE模拟试验平台,以熔体温度、模具温度、熔体注射时间、气体延迟时间、气体压力为关键工艺因素,考察了复杂壳体类塑料件气体辅助注射成型(GAIM)时制件的翘曲变形量和气体穿透情况。以正交试验设计方法为基础,利用遗传算法并结合径向基神经网络建立GAIM工艺参数优化系统,可用于工艺参数组合的快速确定,为GAIM过程中工艺参数优化提供了一种新的求解思路。     

Morphological development during injection molding of self‐reinforcing composites. II: Numerical simulations and analytical predictions

In this study, combined numerical simulation of injection molding and analytical calculations have been used to determine the velocity and elongational strain in the advancing melt front (AMF) region, during the molding of PET/LCP blends, at various injection molding conditions. A model is proposed that establishes the relationship between the aspect ratio of LCP fibers and elongational strain, based on the assumption of an affine deformation of the LCP domains. This model enables us to predict the processing dependent morphology of injection molded PET/LCP blends. The effect of processing parameters on the morphology development during injection molding were investigated. The studies show that injection speed and mold temperature have significant effects on the morphological development of the blends, compared with the effect of the melt temperature. A good correlation between calculated and scanning electron microscopy results was obtained.     

Melt temperature profile prediction for thermoplastic injection molding

A computer system is developed to quantitatively reveal how the melt temperature is affected by the operating conditions during the plastication, dwell and injection stages of the injection molding process. The variables considered in this study are rotation speed, back pressure, barrel heater temperatures, nozzle heater temperature, dwell time and injection velocity profile. A set of Artificial Neural Networks (ANN) has been developed to predict the effect of the operating conditions on the melt temperature during plastication. The dwell period is treated as a heat conduction problem. A free boundary model for the injection phase is developed to simulate the temperature development and melt flow due to the forward motivation of the screw. The overall prediction of nozzle melt temperature is in good agreement with the experimental measurement, validating the proposed procedure combining ANNs and mathematical modeling. This work enhances the understanding of the process and provides a basis for future work on the optimization and advanced control of the process.     

基于神经网络和遗传算法的薄壳件注塑成型工艺参数优化

建立了基于神经网络和遗传算法并结合正交试验的薄壳件注塑成型工艺参数优化系统。正交试验法用来设计神经网络的训练样本，人工神经网络有效的创建了翘曲预测模型；遗传算法完成了对影响薄壳塑件翘曲变形的工艺参数（模具温度、注射温度、注射压力、保压时间、保压压力和冷却时间等）的优化，并计算出了它们的优化值。按该参数进行试验，效果良好，可以有效地减小薄壳塑件翘曲变形，其试验数值与计算数值基本相符，说明所提出的方法是可行的。     

Investigation of melt manipulation phenomena during injection molding via in situ birefringence observation

This study investigates the effects of melt manipulation on the development of molecular orientation during injection molding processing. Vibration‐assisted injection molding (VAIM), a particular method of melt manipulation, is a variation of conventional injection molding in which oscillatory energy is imparted to the polymer melt by vibrating the injection screw axially during the injection and packing stages of the molding cycle. Previous studies have shown that this process positively affects the tensile strength of polystyrene parts, but that the magnitude of the increase is dependent upon the processing parameters. Observation of birefringence patterns in VAIM processed samples show a significant impact on molecular orientation. A specially designed mold and associated image capture system has been developed and is used in this study to record the birefringence patterns of the polymer melt within the cavity during processing. Observation of birefringence shows that orientation develops primarily during post‐vibration packing of the part and not during the vibration phase as previously thought. The observed effects of process parameters such as melt temperature, packing pressure, and vibration duration are discussed. POLYM. ENG. SCI. 46:1691–1697, 2006. © 2006 Society of Plastics Engineers     

EVA塑料发泡倍率的实验研究与预测

采用田口实验设计方法与BP（Back Propagation）神经网络技术,选取模具温度、模压时间、注射压力和模压压力四因素四水平安排正交实验,分析了工艺参数对EVA塑料发泡倍率的影响程度。结果发现,模具温度对发泡倍率的影响较为显著,模压时间次之,注射压力与模压压力的影响较小。最后选取对倍率影响较大的模具温度、模压时间与注射压力工艺参数,对不同形状与尺寸的标准试样模具安排正交实验,以实验结果作为神经网络的样本数据,经过训练后的神经网络能够较为准确地预测EVA塑料的发泡倍率。     

Finite element-based machine learning approach for optimization of process parameters to produce silicon carbide ceramic complex parts

Design and fabrication of silicon carbide ceramic complex parts introduce considerable difficulties during injection molding. Due to the great importance in processing optimization, an accurate prediction on the stress and displacement is required to obtain the desired final product. In this paper, a conceptual framework on combination of finite element method (FEM) and machine learning (ML) method was developed to optimize the injection molding process, which can be used to manufacture large-aperture silicon carbide mirror. The distribution characteristics of temperature field and stress field were extracted from FEM simulation to understand the injection molding process and construct database for ML modeling. To select the most appropriate model, the predictive performance of three ML models were estimated, including generalized regression neural network (GRNN), back propagation neural network (BPNN) and extreme learning machine (ELM). The results show that the developed ELM model exhibits exceptional predictive performance and can be utilized to predict the stress and displacement of the green body. This work allows us to obtain reasonable technique parameters with particular attention to the loading speed and provides some fundamental guidance for the fabrication of lightweight SiC ceramic optical mirror.     

Correlations between extrusion and injection molding process variables for PVC dry blends

Although thermoplastic extrusion and injection molding have been extensively studied by a large number of authors, very little is known about the correlations between the extrusion and injection molding process variables. This paper describes the various comparable process variables between extrusion and injection molding of PVC dry blends. The Brabender extruder of 19 mm diameter and 25:1 length to diameter ratio and the Szekeley reciprocating single screw injection molding machine were used. PVC dry blends of industrial importance were prepared using a high speed mixer. The four mix formulations based on a commercial grade of PVC were used. Process variables studied during the injection molding were the melt temperature near the nozzle, injection pressure, injection speed, and energy consumption. Process variables studied during extrusion were the melt pressure at the die, power consumption, and the melt temperature at the die orifice. The correlations between the extrusion and injection molding process variables for PVC dry blends have been established.