Polymerization Reactor Research Articles

Long Short-Term Memory Neural Networks for Modeling Dynamical Processes and Predictive Control: A Hybrid Physics-Informed Approach.

This work has two objectives. Firstly, it describes a novel physics-informed hybrid neural network (PIHNN) model based on the long short-term memory (LSTM) neural network. The presented model structure combines the first-principle process description and data-driven neural sub-models using a specialized data fusion block that relies on fuzzy logic. The second objective of this work is to detail a computationally efficient model predictive control (MPC) algorithm that employs the PIHNN model. The validity of the presented modeling and MPC approaches is demonstrated for a simulated polymerization reactor. It is shown that the PIHNN structure gives very good modeling results, while the MPC controller results in excellent control quality.

Hybrid-modeling for PTFE polymerization reaction with deep learning-based reaction rate model

Abstract The tetrafluoroethylene (TFE) polymerization process is an essential industrial process to produce polytetrafluoroethylene (PTFE), which is extensively utilized in aerospace and medical domains. A precise mechanism model is a prerequisite for comprehensively understanding this process. However, significant uncertainties in the kinetic model parameters may hinder attaining an optimal reaction rate. This study proposes a hybrid polymerization reaction model that integrates process mechanism modeling and data-driven modeling to address this challenge. In the hybrid modeling approach, the mechanism model for the polymerization reaction is developed based on reaction kinetic and thermodynamic assumptions. Additionally, a long short-term memory (LSTM) neural network is employed to predict the reaction rate for chain initiation by leveraging temporal relationships derived from archived measurements. The proposed methodology is implemented using a PTFE polymer reactor system, and experimental comparisons affirm its superior performance and effectiveness.

A robustly stabilizing fault‐tolerantMPC: A performance improvement‐based approach

AbstractIn terms of model predictive control (MPC) performance degradation caused by operational faults, in this article, a robust MPC strategy with active fault tolerance properties is proposed. The proposed strategy incorporates a fault supervision layer into the structure of conventional cost‐contracting formulation‐based robust MPC for the online update of the nominal controller model in the event of faults. The robust MPC is based on multiplant uncertainty, while the supervisory layer consists of a bank of unknown input observers and a decision‐making algorithm. Simulation results in a nonlinear polymerization reactor subject to process faults demonstrate that the proposed approach offers superior performance compared to the conventional strategy.

Methodology for designing solution polymerization reactors based on multiscale model construction and simulation

This work aims to develop a reliable methodology for designing solution polymerization reactors based on multiscale model construction and simulation. Firstly, accurate polymerization kinetic model, thermodynamic property model, transport property model, fluid dynamic model and simplified equipment model are established. Secondly, process and reactor scale simulations are employed in systematically studying the impact of various factors (e.g., operating variables, reactor type, combination mode, structure) through coupling the established models. The design of reactor system is conducted by deeply analyzing simulated results and comprehensively considering various influence factors. The research scheme for polymerization reactor design, along with the proposed methods for model construction, multiscale modeling, simulation cases design and data analysis, are explained in detail. Furthermore, taking a pilot-scale copolymerization of ethylene and 1-octene as an example, 248 simulation cases are carried out to provide a comprehensive description on the process of its model construction and multi-scale simulations.

Development and characterization of a continuous ultrasound emulsification and nano-emulsion polymerization process

Typically, nano-emulsions are prepared in batch high-speed homogenization or ultrasound processes and polymerized afterwards in batch reactors. However, fully continuous processes have the potential to decrease production costs and energy consumption compared to batch processes. This research thus focuses on implementing ultrasound into a fully continuous emulsification and nano-emulsion polymerization process to obtain latex nano-particles from butylmethacrylate and ethylene glycol dimethacrylate. The goal of this research is to produce nano-emulsions and nano-particles with the smallest obtainable size and polydispersity in a fully continuous two-stage process. In the first stage or emulsification stage, parameters like flowrate, residence time and acoustic power are varied to influence and determine the optimal energy density. In second stage or reaction stage, residence time and reaction temperature are studied to obtain particles of monodisperse sizes. Samples produced in both stages are analysed with Dynamic Light Scattering to measure the average size and polydispersity (PdI) of the emulsion droplets and particles. Emulsification results indicate that droplet size and PdI decreases at increasing energy densities (J/ml) until 60 J/ml after which a constant droplet size of 150 nm and PdI of 0.230 are reached. Furthermore, a particle size of 50 nm and PdI of 0.080 were achieved in both batch and continuous polymerization reactors after 5 min at 85 °C. By implementing the most optimal process parameters, small emulsion droplets and particles were obtained more energy efficient in a shorter emulsification and reaction time compared to literature ultrasound assisted nano-emulsion polymerization processes.

Mixing effects on the gas flow and mass transfer in a vertical single helical ribbon agitated reactor for propylene polymerization

The gas–solid vertical single helical ribbon agitated reactor is widely used in the polypropylene industry. The gas flow and mass transfer determine the heat transfer capacity and product quality. In this work, the pressure drop was measured below and above minimum fluidization velocity (umf). Then, the axial and radial dispersion coefficients (DL and DR) were estimated below umf by measuring the step residence time distribution and the stationary radial tracer profiles, respectively. Pressure analysis show the agitation significantly reduces the bed pressure drop, changing the fixed bed below umf and bubbling bed above umf to an agitated bed with a periodic voidage. The DL was proportional to the gas velocity, with little influences of the stirring speed. The radial tracer profiles show the bed is divided by the inner diameter of the helical ribbon into the peripheral ascending region and the central descending region, without gas exchange between them. Furthermore. the tracer profiles in the descending region are classified into two types. For type I, the DR, calculated by the plug flow model with radial dispersion, is proportional to the gas velocity, with little effects of stirring speed. Type II shows uniform profiles due to the deflection of the gas flow toward the low pressure area at the inner diameter of helical ribbon. The causes of the two types are explained by the relative size of particle circulation and gas flux. These two flow regimes are quantitatively clarified with an empirical correlation of Re and Fr as flow regime transition boundary, namely dispersed plug flow and perfect mixing flow. With lower Re and higher Fr, it changes from dispersed plug flow to perfect mixing flow.

Investigation of a dynamic operation mode of ethylene polymerization process with a liquid-containing fluidized bed

Fluidized bed reactors are widely applied to the production of polyolefin resins. A novel dynamic operation mode of a liquid-containing fluidized bed reactor for ethylene polymerization is proposed. The spatial and temporal scale effects of the condensed liquid on the polymerization reaction conditions and chain structure of polyethylene are investigated by periodically introducing the condensed liquid into the fluidized bed reactor. It is found that the instantaneous yield of polyethylene, instantaneous molecular weight and instantaneous comonomer composition of polyethylene varied along with time in the same direction. The simulation results indicate that the dynamic operation mode of the liquid-containing fluidized bed reactor provides an effective approach to regulating the chain branching distribution of polyethylene.

Evolutionary artificial neural network for temperature control in a batch polymerization reactor

The integration of artificial intelligence techniques introduces fresh perspectives in the implementation of these methods. This paper presents the combination of neural networks and evolutionary strategies to create what is known as evolutionary artificial neural networks (EANNs). In the process, the excitation function of neurons was modified to allow asexual reproduction. As a result, neurons evolved and developed significantly. The technique of a batch polymerization reactor temperature controller to produce polymethylmethacrylate (PMMA) by free radicals was compared with two different controls, such as PID and GMC, demonstrating that artificial intelligence-based controllers can be applied. These controllers provide better results than conventional controllers without creating transfer functions to the control process represented.

Intelligent Modeling for Batch Polymerization Reactors with Unknown Inputs

While system identification methods have developed rapidly, modeling the process of batch polymerization reactors still poses challenges. Therefore, designing an intelligent modeling approach for these reactors is important. This paper focuses on identifying actual models for batch polymerization reactors, proposing a novel recursive approach based on the expectation-maximization algorithm. The proposed method pays special attention to unknown inputs (UIs), which may represent modeling errors or process faults. To estimate the UIs of the model, the recursive expectation-maximization (EM) technique is used. The proposed algorithm consists of two steps: the E-step and the M-step. In the E-step, a Q-function is recursively computed based on the maximum likelihood framework, using the UI estimates from the previous time step. The Kalman filter is utilized to calculate the estimates of the states using the measurements from sensor data. In the M-step, analytical solutions for the UIs are found through local optimization of the recursive Q-function. To demonstrate the effectiveness of the proposed algorithm, a practical application of modeling batch polymerization reactors is presented. The performance of the proposed recursive EM algorithm is compared to that of the augmented state Kalman filter (ASKF) using root mean squared errors (RMSEs). The RMSEs obtained from the proposed method are at least 6.52% lower than those from the ASKF method, indicating superior performance.

Gray‐Box Modeling of the Molecular Weight Distribution in a Batch Polymerization Reactor

AbstractWe introduce a gray‐box approach for modeling the molecular weight distribution in step‐growth polymerization reactions using the aggregation population balance equation. The approach is based on extracting a data‐based kernel function from in‐process measurements of the molecular weight distribution. The method is applied to historical data from an industrial batch polymerization reactor. The resulting model is used for decision support in production by predicting the reaction endpoint corresponding to a target molecular weight. The accuracy of the predictions proved to be sufficient for the deployment of the method.

Modeling of Particle Size Distributions in Industrial Poly(vinyl chloride) Suspension Polymerization Reactors

In the present paper, a mathematical model is built and implemented to describe the trajectories of mass inventories, pressures and polymer properties with emphasis on final particle size distributions of industrial scale poly(vinyl chloride) suspension polymerization reactors. The model comprises the mass balances, statistical moment balances, equilibrium relationships and population balance equations. A discretization scheme is employed to transform the integro-differential equations resulting from the population balance model into a system of differential equations. The obtained results show, for the first time, that classical breakage and coalescence kernels described in the literature can provide very good fittings of actual industrial scale data when coupled with proper parameter estimation procedures, so that the proposed model is able to represent the available operation data with good accuracy at distinct conditions. Particularly, it is also shown that the use of a top condenser for control of the reactor temperature can lead to changes of parameters that control the particle size distributions.

ПОСТРОЕНИЕ МОДЕЛИ ПРОЦЕССА ПОЛИМЕРИЗАЦИИ ЭТИЛЕНА В АВТОКЛАВНОМ РЕАКТОРЕ С МЕШАЛКОЙ

A mathematical model of a polymerization reactor as a cascade of ideal mixing apparatuses is proposed

Математическое моделирование процесса полимеризации этилена в автоклавном реакторе с мешалкой

Building a mathematical model of the polymerization reactor is necessary to determine the dependencies between the input variables, the technological mode and the output variables of the ethylene polymerization process in an autoclave reactor with a stirrer, as well as to analyze the effect of the reactor design parameters on efficiency. A deterministic mathematical model developed on the basis of analysis of the physico-chemical regularities of the ethylene polymerization process is proposed to solve these problems. There is shown a schematic diagram of ethylene polymerization indicating the flows of substances and energy, a description of the process, the mechanism of chemical kinetics of ethylene polymerization, a system of assumptions simplifying the construction of a mathematical model, equations of a mathematical model of an ethylene polymerization reactor in an autoclave reactor with a stirrer. The autoclave ethylene polymerization reactor is presented as a cascade of ideal mixing apparatus to describe each of the reactor zones. The rates of chemical reactions of initiation, growth and chain breakage are expressed in accordance with the law of acting masses. The mathematical model in its final form is a system of ordinary differential equations for describing each of the reactor zones. The initial numerical values of variables and parameters for modeling ethylene polymerization in the autoclave reactor are given. The results of numerical experiments are shown, and the influence of various factors on the output variables of the ethylene polymerization process is investigated. The proposed mathematical model can be used to optimize the technological regime of the ethylene polymerization process in an autoclave reactor with a stirrer, as well as to analyze the influence of the design parameters of the reactor on its efficiency.

Scale-up and thermal stability analysis of fluidized bed reactors for ethylene polymerization

A set of hydrodynamic similarity laws is applied to the scale-up of ethylene polymerization fluidized bed reactors (FBRs) under the condensed mode operation. The thermal stability of open-loop controlled FBRs is investigated by the homotopy continuation method. And the Hopf bifurcation point is selected as an index of the thermal stability similarity. The simulation results show the similarity in state variables, operation parameters, the space–time yield (STY), and the thermal stability of FBRs with different scales. Furthermore, the thermal stability behaviors and similarity of the closed-loop controlled FBRs with different scales are analyzed. The observed similar trend of Hopf bifurcation curves reveals the similarity in the thermal stability of closed-loop controlled FBRs with different scaling ratios. In general, the results of the thermal stability similarity confirm that the hydrodynamics scaling laws proposed in the work are applicable to the scale-up of FBRs under the condensed mode operation.

Thermal stability and decomposition mechanism analysis of 1, 1’-Azobis(cyclohexanecarbonitrile) by STA, DSC, ARC and TG-FTIR

1, 1′-Azobis(cyclohexanecarbonitrile) (ABCN) is a kind of azo compound, which is widely used as an initiator in the polymer industry. Plenty of heat will be released during the thermal decomposition, and fire or explosive accidents could happen when the cooling system of the polymerization reactor fails. In this manuscript, the thermal decomposition behavior of ABCN under a dynamic temperature environment was investigated by simultaneous thermogravimetric analyzer (STA). Differential scanning calorimeter (DSC) was used to detect the thermal effects of different stages of mass loss. The apparent activation energy was calculated by Kissinger-Akahira-Sunose methods (K-A-S methods). The thermal runaway behavior was investigated by accelerating rate calorimeter (ARC), which was used to predict time to maximum rate under adiabatic condition (TMRad). The results show that the temperatures at TMRad = 8 h and 24 h are 93.0 °C and 89.7 °C, respectively. The self-accelerating decomposition temperature (SADT) is 86.6 °C, which is calculated based on the thermal theory of Semenov. Since ABCN is used at 60–90 °C, a major accident could happen when the heat accumulated in the reactor is not removed timely. The risk of thermal runaway reactions for ABCN was defined against the risk matrix as level Ⅲ, which is an unacceptable risk. The carbon radicals and nitrogen were released during the decomposition of ABCN based on the experiment result of thermal gravimetric analyzer-Fourier transform infrared spectrometer (TG-FTIR) system. Several safety assessment parameters could supply references for the industry to alleviate thermal hazards and prevent critical incidents.

Modelling of a continuous kneader reactor for the polymerization of partially neutralized acrylic acid

AbstractA mathematical model and analysis of the continuous polymerization of partially neutralized acrylic acid (AA) in a continuous kneader reactor is presented here as an initial attempt to simulate the synthesis of a superabsorbent polymer. A detailed kinetic model has been used to describe the copolymerization of AA and sodium acrylate (NaA) in aqueous medium. This model is used to describe batch and continuous operations. The polymerization is initiated by a mixture of potassium persulphate (K2S2O8, KPS) and hydrogen peroxide (H2O2) as oxidizing agent and ascorbic acid (AsA) as reducing agent. A novel set of kinetic parameters has been estimated by fitting experimental data from different literature sources. The operation of a continuous kneader reactor modelled as a plug‐flow reactor with axial dispersion is theoretically investigated to predict temperature profile, total and individual monomer conversion, consumption of KPS, H2O2, and AsA, and polymer average molecular weights. The simulation results show the presence of a hot spot close to the reactor entrance that could be potentially severe during startup and could have a detrimental impact on polymer quality. This model is a first step in the direction of achieving optimal operating protocols and exploring improved polymerization reactor designs.

Mixing in High‐Pressure Polymerization Reactors: A Combined Experimental and Modeling Approach

AbstractMixing in low‐density polyethylene (LDPE) reactors is crucial regarding process safety as well as process efficiency. The extreme process conditions of pressures up to 3000 bar make it difficult to get an insight into LDPE reactors. This problem is tackled by a combined experimental and modeling approach. A poly(methylmethacrylate) duplicate of a laboratory high‐pressure reactor is set up to measure residence times and visualize flow lines. Experimental results are then compared to a computational fluid dynamics model.

Sustainable Production of Polyamide 6 Fibers: Direct Melt Spinning and Efficient Reuse of Residual Oligomers during Polymerization

Direct melt spinning of PA6 from a polymerization reactor is currently unachievable in the industrial production of polyamide 6 (PA6) fibers because the hydrolysis polymerization of its monomer (caprolactam) is a reversible reaction which requires a boiling water extraction of oligomers from the polymer before melt spinning. Here, we proposed a polymerization method where the content of the cyclic dimer is controlled by reducing the temperature at the monomer ring-opening stage and the content of oligomers is monitored by rapid devolatilization through the polycondensation stage. As a result, the content of oligomers in PA6 is less than 1.5 wt % and the cyclic dimer is at 0.1–0.3 wt %, which are 80 and 48% lower than the values for the conventional PA6 production process. This allows for direct melt spinning of PA6 fibers at the industrial scale (spinning speed: 4000 m/min and a continuous production line of 10 kg/h), which reduces the production time from over 40 h to within 10 h, and the resource consumption reduces by over 30%. The resulting PA6 fibers show comparable mechanical and dyeing properties to those spun with the conventional multistep spinning method. In addition, the unreacted oligomers extracted during the polymerization stage can be directly reused in the polymerization, with no further treatment. This work provides an effective route to the sustainable production of PA6 fibers at the industrial scale, which significantly decreases the consumption of resources and the generation of chemical wastes while improving the production efficiency.

On‐line control of emulsion polymerization reactors: A perspective

AbstractEmulsion polymerization is the method of choice to produce a wide range of specialty waterborne polymer dispersions. These are multicharacteristic products that should meet conflicting application properties and are produced through a very complex process prone to suffer run‐to‐run irreproducibilities. In a scenario of increasing competition and smaller margins, achieving even more efficient production of better materials in a consistent, cost effective, safe, and environmentally friendly way will only be possible by implementing on‐line control. However, on‐line control is hindered by the lack of sensors able to monitor most of the characteristics of the emulsion polymers. This perspective discusses the state‐of‐the‐art of the on‐line control of emulsion polymerization reactors highlighting the achievements, challenges, and opportunities.

Roger K. West

Roger K. West, 84, died July 30, 2022, in Montclair, New Jersey. “Roger’s high school career assessment determined he should be a plumber. His family often joked that despite his PhD, working in a polymer reactor plant connecting pipes and actuating valves was not so far from that early prediction. Roger had a productive 30-year career at Exxon’s polymer division, eventually being named an ACS Hero of Chemistry for his work on viscosity modifiers for engine oil. He enjoyed a second career as a patent searcher. Outside of work, Roger was a steadfast family man who also loved to sing, taste wine, and study languages—including French and Mandarin. He is greatly missed.”—Sharon West, daughter Most recent title: Research and Development Engineer, Exxon Chemical Education: BS, chemical engineering, Cornell University, 1961; PhD, chemical engineering, Cornell University, 1965 Survivors: Wife, Mary Lou; children, David and Sharon; four grandchildren To recognize your late