Model Reduction in Emulsion Polymerization Using Hybrid First‐Principles/Artificial Neural Network Models

A first‐principles mathematical model for emulsion polymerization was reduced by using a hybrid mathematical model composed by artificial neural networks (ANN) and material balances. The goal was to have an accurate model that may be integrated fast enough to be used for online optimization purposes. In the reduced model the polymerization rate and the instantaneous weight‐average molecular weight were calculated by means of artificial neural networks. These ANNs were incorporated to first‐principles material balances. The accuracy of the reduced model under a wide range of conditions was assessed. Savings in computer time were achieved by using the reduced model, which makes it suitable for online optimization purposes.

Effect of the temperature on the cumulative weight‐average molecular weight: first principles mathematical model (—); (ANN2) and hybrid model predictions: (▵) 50 °C, (▪) 60 °C^(training), (▿) 70 °C^(validation), (•) 80 °C, (○) 90 °C.     

New Approach in Modeling of Metallocene‐Catalyzed Olefin Polymerization Using Artificial Neural Networks

A new approach for the estimation of kinetic rate constants in olefin polymerization using metallocene catalysts is presented. The polymerization rate has been modeled using the method of moments. An ANN has been used and trained to behave like the mathematical model developed before, so that it gets polymerization rate at different reaction times and predicts reaction rate constants. The network was trained using modeling results in desired operational window. The polymerization rates were normalized to make the network work independent of operational conditions. The model has also been applied to real polymerization rate data and the predictions were satisfactory. This model is specially useful in comparing different new metallocene catalysts.

     

化学中的人工神经网络法

反向传输人工神经网络是应用最为广泛的一种方法, 本文较详细地介绍了该种方法及其相关的问题, 同时给出了Kohonen 模型和Hopfield 网络的简单算法。关于神经网络在化学中的应用, 该文介绍了6 个方面: 定量结构2活性性质相关性(QSAR/QSPR )研究, 有机化合物结构解析, 光谱的数据处理, 化学反应性, 流程优化, 故障诊断及控制, 蛋白质结构。     

Block Copolymers From Living Emulsion Polymerization: Reactor Operating Strategies and Blocking Efficiency

Diblock copolymers are generated using xanthate‐based RAFT agents in conjunction with emulsion polymerization via stage‐wise operations. First, emulsion polymerization is conducted for styrene, methyl acrylate, and butyl acrylate monomers to obtain polymers of specified molar mass. At the second stage, polymers undergo chain extension to produce block copolymers. Linear growth of molecular weight with respect to conversion establishes the living characteristics of the process. Under batch conditions, partly homopolymers are produced. Semi‐batch operation produces copolymers of higher purity with low polydispersity. The choice of blocking sequence is crucial for reducing the influence of the terminated chains on the distribution sequence of copolymers produced.

     

Estimation of Polymerization Conditions Needed to Make Ethylene/1‐olefin Copolymers with Specific Microstructures Using Artificial Neural Networks

Two artificial neural network models (forward and inverse) are developed to describe ethylene/1‐olefin copolymerization with a catalyst having two site types using training and testing datasets obtained from a polymerization kinetic model. The forward model is applied to predict the molecular weight and chemical composition distributions of the polymer from a set of polymerization conditions, such as ethylene concentration, 1‐olefin concentration, cocatalyst concentration, hydrogen concentration, and polymerization temperature. The results of the forward model agree well with those from the kinetic model. The inverse model is applied to determine the polymerization conditions to produce polymers with desired microstructures. Although the inverse model generates multiple solutions for the general case, unique solutions are obtained when one of the three key process parameters (ethylene concentration, 1‐olefin concentration, and polymerization temperature) is kept constant. The proposed model can be used as an efficient tool to design materials from a set of polymerization conditions.

     

Rate Optimization in Controlled Radical Emulsion Polymerization Using RAFT

Summary: Means of improving rates in RAFT‐mediated radical emulsion polymerizations are developed, by setting out strategies to minimize the inhibition and retardation that always are present in these systems. These effects arise from the RAFT‐induced exit of radicals, the desorption of the RAFT‐reinitiating radical from the particles, and the specificity of the reinitiating radical to the RAFT agent. Methods for reducing the inhibition period such as using a more hydrophobic reinitiating radical are predicted to show a significant improvement in the inhibition periods. The time‐dependent behavior of the RAFT adduct to the entering radical and the RAFT‐induced exit (loss) of radicals from particles are studied using a previously described Monte Carlo model of RAFT/emulsion particles. It is shown that an effective way of reducing the rate coefficient for the exit of radicals from the particles is to use a less active RAFT agent. Techniques for improving the rate of polymerization of RAFT/emulsion systems are suggested based upon the coherent understanding contained in these models: the use of an oligomeric adduct to the RAFT agent, a less water‐soluble RAFT re‐initiating group, and a less active RAFT agent.

Populations of the different types of particles (left axis) along with the concentration of the initial RAFT agent, D_R (right axis), as a function of time.     

人工神经元网络法提高毛细管电泳药物控制分析准确度的研究

研究了人工神经元网络法在毛细管电泳定量测定memantine中提高测定准确度 的可行性。在毛细管电泳法定量测定memantine的过程中，其浓度与峰高或峰面积 以及与二者和内标的比值均没有良好的线性关系。人工神经元网络具有很强的非线 性校正能力，其最大优点是无须对分离体系及组分的迁移行为预先予以了解。人工 神经元网络的输为memantine的峰高和峰面积，输出为memantine的浓度。通过实验 确定的网络结构为2：1：1型。由于人工神经元网络的通用性，该法也可用于毛细 管电泳在其他药物控制分析中改善定量分析的准确度。     

Predictive Models of Phytosterol Degradation in Rapeseeds Stored in Bulk Based on Artificial Neural Networks and Response Surface Regression

The need to maintain the highest possible levels of bioactive components contained in raw materials requires the elaboration of tools supporting their processing operations, starting from the first stages of the food production chain. In this study, artificial neural networks (ANNs) and response surface regression (RSR) were used to develop models of phytosterol degradation in bulks of rapeseed stored under various temperatures and water activity conditions (T = 12–30 °C and a_w = 0.75–0.90). Among ANNs, networks based on a multilayer perceptron (MLP) and a radial basis function (RBF) were tested. The model input constituted a_w, temperature and storage time, whilst the model output was the phytosterol level in seeds. The ANN-based modeling turned out to be more effective in estimating phytosterol levels than the RSR, while MLP-ANNs proved to be more satisfactory than RBF-ANNs. The approximation quality of the ANNs models depended on the number of neurons and the type of activation functions in the hidden layer. The best model was provided by the MLP-ANN containing nine neurons in the hidden layer equipped with the logistic activation function. The model performance evaluation showed its high prediction accuracy and generalization capability (R² = 0.978; RMSE = 0.140). Its accuracy was also confirmed by the elliptical joint confidence region (EJCR) test. The results show the high usefulness of ANNs in predictive modeling of phytosterol degradation in rapeseeds. The elaborated MLP-ANN model may be used as a support tool in modern postharvest management systems.     

Evaluation of the Chain Length Distribution in Free‐Radical Polymerization, 1. Bulk Polymerization

A method for the direct computation of the chain length distribution in a bulk polymerization is developed, based on the discretization procedure introduced by Kumar and Ramkrishna (Chem. Eng. Sci. 1996 , 51, 1311) in the context of particle size distribution. The overall distribution of chain lengths is partitioned into a finite number of classes which are supposed to be concentrated at some appropriate pivotal chain lengths. Several of the involved reactions lead to the formation of chain whose length differs from the pivotal values. Rules have been introduced in order to share chains between two contiguous classes, which have been designed so as to preserve two well‐defined properties of the distribution, such as, for example, two of its moments. The method has been applied to a polymerization system including propagation, bimolecular terminations and two different chain branching mechanisms: chain transfer to polymer and crosslinking. In addition, complex systems such as one with chain length‐dependent kinetic constants or a two‐dimensional distribution of chain length and number of branches have been considered.     

Modeling Interfacial Tension of n-Alkane/Water-Salt System Using Artificial Neural Networks

In this paper, a feed forward neural network is built and trained using experimental data reported in the literature to model interfacial tension of n-alkane/water-salt systems. Temperature, pressure, molecular weight of n-alkane, and ionic strength of electrolyte solution are used as input to the neural network. The model succeeded to predict interfacial tension of liquid n-alkane/water system with or without the presence of electrolyte and yielded average absolute deviation of 0.58% over all data points. The performance of the model is analyzed and compared against the performance of the other alternative models. It was found out that the proposed model outperforms the other alternatives.     

Well‐Defined High Molecular Weight Polystyrene with High Rates and High Livingness Synthesized via Two‐Stage RAFT Emulsion Polymerization

A highly living polymer with over 100 kg mol⁻¹ molecular weight is very difficult to achieve by controlled radical polymerization since the unavoidable side reactions of irreversible radical termination and radical chain transfer to monomer reaction become significant. It is reported that over 500 kg mol⁻¹ polystyrene with high livingness and low dispersity could be synthesized by a facile two‐stage reversible addition–fragmentation transfer emulsion polymerization. The monomer conversion reaches 90% within 10 h. High livingness of the product is ascribed to the extremely low initiator concentration and the chain transfer constant for monomer unexpectedly much lower than the well‐accepted values in the conventional radical polymerization. The two‐stage monomer feeding policy much decreases the dispersity of the product.

     

Industrial SBR Process: Computer Simulation Study for Online Estimation of Steady‐State Variables Using Neural Networks

This work investigates the industrial production of styrene‐butadiene rubber in a continuous reactor train, and proposes a soft sensor for online monitoring of several processes and polymer quality variables in each reactor. The soft sensor includes two independent artificial neural networks (ANN). The first ANN estimates monomer conversion, solid content, polymer production, average particle diameter, and average copolymer composition; the second ANN estimates average molecular weights and average branching degrees. The required ANN inputs are: (i) the reagent feed rates into the first reactor and (ii) the reaction heat rate in each reactor. The proposed ANN‐based soft sensor proved robust to several measurement errors, and is suitable for online estimation and closed‐loop control strategies.

     

Molecular Weight Control in Emulsion Polymerization by Catalytic Chain Transfer: A Reaction Engineering Approach

Summary: For the application of catalytic chain transfer in (mini)emulsion polymerization, catalyst partitioning and deactivation are key parameters that govern the actual catalyst concentration at the locus of polymerization and consequently the final molecular weight distribution. A global model, based on the Mayo equation, catalyst partitioning and deactivation was developed. The influence of several reaction parameters on the instantaneous number average molecular weight was quantified.     

The Influence of Xanthate‐Based Transfer Agents on Styrene Emulsion Polymerization: Mathematical Modeling and Model Validation

A mathematical model was developed to account for the evolution of polymer product attributes in the emulsion polymerization of styrene. The effects of transfer agent, surfactant, initiator and temperature were investigated. Polymerization rate, and particle size decreased with increasing concentration of the transfer agent. The polymerization rate increased with increasing surfactant and initiator concentrations, while an increase in temperature led to a decrease of molecular weight but an increase of polymerization rate and particle size. Chain extension was successfully achieved in the presence of our RAFT agent. The model predictions compared well with our experimental results.

     

Preparation of CO2/N2‐Triggered Reversibly Coagulatable and Redispersible Polyacrylate Latexes by Emulsion Polymerization Using a Polymeric Surfactant

We report here a novel approach for making reversibly coagulatable and redispersible polyacrylate latexes by emulsion (co)polymerization of methyl methacrylate (MMA) using a polymeric surfactant, poly(2‐(dimethylamino)ethyl methacrylate)₁₀‐block‐poly(methyl methacrylate)₁₄. The surfactant was protonated with HCl prior to use. The resulted PMMA latexes were readily coagulated with trace amount of caustic soda. The coagulated latex particles, after washing with deionized water, could be redispersed into fresh water to form stable latexes again by CO₂ bubbling with ultrasonication. The recovered latexes could then be coagulated by N₂ bubbling with gentle heating. These coagulation and redispersion processes were repeatable by the CO₂/N₂ bubbling.