Batch Chemical Processes Research Articles

Meta-learning-based continuous state-space models for rapid monitoring using heterogeneous grade sources with uneven sampling

In an industrial batch, it is common to produce products with different grades, so only a limited number of data are available for each grade of product. The data are also easily contaminated by external disturbances and noise. The sampling interval would be messy and irregular if contaminated samples were removed. Thus, it is difficult to build models for monitoring the corresponding process. This paper proposes a meta-learning-based functional continuous state space model for fast modeling and monitoring of multi-grade batch processes. It takes advantage of wavelet function properties to effectively handle the irregular sampling issue. In addition, the meta-learning method uses diverse historical multi-grade datasets for modeling. It rapidly establishes an efficient and reliable monitoring model for new running batches with a small number of batch data. The effectiveness of the proposed method is shown through a numerical case and an industrial case in the chemical batch process.

Compensation of accuracy by increased data “thickness” for high timeliness in fault diagnosis of chemical process

BackgroundFault diagnosis rate (FDR) and fault diagnosis timeliness (FDT) are critical to a fault diagnosis model applied to industrial processes. Generally, high FDR requires process data of longer time, which means a low timeliness; while, the time-length reduction of data will result in a decreased FDR as learning features from less data is more difficult. MethodsTo address this problem, an idea is proposed that the time-length reduction of process data could be compensated by enhancing the data “thickness”, i.e., the shortage of feature information of data could be compensated by investigating more patterns from multiple different aspects of the data. Based on this, the multiple pattern representation-convolutional neural network (MPR-CNN) is proposed, in which multiple pattern representation algorithms are used first to pre-extract features from data. The resulting three-dimensional matrix contains richer features, which could be learned more efficiently by the CNN. Significant FindingsThe diagnosis results from a continuous and a batch chemical process show that MPR-CNN significantly improves the fault diagnosis performance, which achieved 96.1% of FDR in the Tennessee Eastman process and 98.1% of FDR in a semi-batch crystallization process, and outperforms other CNN structure-based models.

High‐order iterative learning model predictive control for batch chemical processes

AbstractIn order to address two‐dimensional (2D) control issue for a class of batch chemical processes, we propose a novel high‐order iterative learning model predictive control (HILMPC) method in this paper. A set of local state‐space models are first constructed to represent the batch chemical processes by adopting the just‐in‐time learning (JITL) technique. Meanwhile, a pre‐clustered strategy is used to lessen the computational burden of the modelling process and improve the modelling efficiency. Then, a two‐stage 2D controller is designed to achieve integrated control by combining high‐order iterative learning control (HILC) on the batch domain with model predictive control (MPC) on the time domain. The resulting HILMPC controller can not only guarantee the convergence of the system on the batch domain, but also guarantee the closed‐loop stability of the system on the time domain. The convergence of the HILMPC method is ensured by rigorous analysis. Two examples are presented in the end to demonstrate that the developed method provides better control performance than its previous counterpart.

An inherently safer design approach based on process safety time for batch chemical reaction processes

The economic attractiveness of an industrial process depends a large extent on its safe and reliable operation. The inherently safer design can help reduce the potential hazard on a fundamental level and eliminate escalation at the early stage of process development. In this work, an inherently safer design method for batch chemical reactors is proposed, which maximizes the yield of target product while minimizes Dow’s fire and explosion index. Process safety time is introduced as criterion to respond to the thermal runaway risk risen by unexpected failures. Finally, two examples for specific reaction systems with given kinetics are performed to demonstrate the feasibility and validity of the proposed method. The results indicate that the product yield optimized with the proposed method is significantly higher than that obtained from literature approach involving temperature limit or divergence criterion under the same PST. With the same yield, the PST optimized with the proposed method is longer.

Monitoring of Mean Processes with Mixed Moving Average – Modified Exponentially Weighted Moving Average Control Charts

In Statistical Process Control, a control chart is the most effective equipment for monitoring and improving processes. Classic control charts were created in the past and were effective at detecting both small and large changes. However, the mixed control chart has been presented to improve the performance of the traditional control chart. This research introduces a new mixed control chart, MA-MEWMA, which combines the moving average (MA) and the modified exponentially weighted moving average (MEWMA) charts to detect the tiny changes in the procedures both of symmetric and asymmetric distributions. The average run length (ARL) can also be used to measure progress in the MA-MEWMA chart with Shewhart, MA, and MEWMA charts that employ Monte Carlo simulation. The experiments demonstrated that the proposed chart had a greater impact compared to all other control charts with the parameter level ±0.05, ±0.10, ±0.25, ±0.50, ±0.75, ±1.00, ±1.50 through discovering a change in the average of the method in the control where ARL0 = 370. On the other hand, when the parameter level was set to 2.00, ±3.00, ±4.00, the MA control chart performed admirably. An excellent example is data set on viscosity from a batch chemical process. Environmental information data were provided to explain how the suggested chart and MA-MEWMA charts are implemented, demonstrating that the MA-MEWMA chart was more successful than other charts in detecting changes.

A Multi-objective optimization for batch chemical reaction Processes: The trade-off between economy and safety

he thermal safety is always a challenging topic in the case of batch reactors carrying exothermic reactions. To reduce the risk of thermal runaway from both perspectives of likelihood and severity, it is desirable and practicable to develop a multi-objective optimization framework which can achieve trade-off between economy and safety. Therefore, this work provided an optimization framework for batch chemical reactor design with inherent safer operating and equipment parameters through maximizing yield of the target product while minimizing degree of Dow’s fire and explosion index along with thermal runaway constraints. Thermal runaway criteria, including temperature limit and divergence criterion, are employed as safety constraints. The response time is calculated with the optimal schemes selected from the Pareto solutions by a multi-attribute decision-making method (MADM). Finally, two examples for specific reaction systems with given kinetics are performed to demonstrate the feasibility and validity of the proposed method.

Process improvement using value stream mapping and lean methodology: a case study application in batch chemical process industry

Intensifying competition in the global chemicals market is increasing the pressure on companies to adopt lean manufacturing and improve their operational efficiency to remain competitive. Companies operating in emerging economies are no exception. This study looks at the application of value stream mapping (VSM) in the production process of a batch chemical manufacturing company based in Indonesia. The application of value stream method is a multistep approach. A value stream scope is determined in the beginning, followed by a data collection step to construct a current state map. The current state is analysed to gain a deep understanding of how things currently operate and provides the foundation for the ideal future state of the production process. The production process is then redesigned following lean manufacturing principles, and the future state map is drawn. Analysis is then conducted to estimate the potential cost reduction from implementing the proposed future state. The results indicate a more efficient production system as shown by decreased production lead time and reduced production costs.

Lyapunov-based economic model predictive control for nonlinear descriptor systems

A large number of physical applications such as electrical circuits, chemical processes and multibody systems can be accurately described by a set of differential algebraic equations (DAEs). For instance, sustainable water desalination techniques such as membrane distillation (MD) is an example of a chemical process that is described by DAEs. Optimal control strategies such as economic model predictive control (EMPC) has many operational advantages (e.g., maximizing process economics and enhancing operational safety) for chemical applications. However, constructing EMPC schemes for a general class of nonlinear DAEs systems while ensuring closed-loop stability and recursive feasibility is an open research challenge that has not been considered. Motivated by the above considerations, this note introduces a Lyapunov-based economic model predictive control (LEMPC) design that can economically operate nonlinear descriptor systems while satisfying input and states constraints. To guarantee closed-loop stability and recursive feasibility of the proposed control design, a Lyapunov-based control law is introduced to characterize the stability region at which the LEMPC can maximize the process economics. Finally, a chemical batch process modeled by nonlinear differential algebraic equations (DAEs) is utilized to demonstrate the applicability of the proposed framework.

Application of a grey-box modelling approach for the online monitoring of batch production in the chemical industry

Abstract Model-based solutions for monitoring and control of chemical batch processes have been of interest in research for many decades. However, unlike in continuous processes, in which model-based tools such as Model Predictive Control (MPC) have become a standard in the industry, the reported use of models for batch processes, either for monitoring or control, is rather scarce. This limited use is attributed partly to the inherent complexity of the batch processes (e. g., dynamic, nonlinear, multipurpose) and partly to the lack of appropriate commercial tools in the past. In recent years, algorithms and commercial tools for model-based monitoring and control of batch processes have become more mature and in the era of Industry 4.0 and digitalization they are slowly but steadily gaining more interest in real-word batch applications. This contribution provides a practical example in this application field. Specifically, the use of a grey-box modeling approach, in which a multiway Projection to Latent Structure (PLS) model is combined with a first-principles model, to monitor the evolution of a batch polymerization process and predict in real-time the final batch quality is reported. The modeling approach is described, and the experimental results obtained from an industrial batch laboratory reactor are presented.

Reinforcement learning of adaptive online rescheduling timing and computing time allocation

Mathematical optimization methods have been developed to a vast variety of complex problems in the field of process systems engineering (e.g., the scheduling of chemical batch processes). However, the use of these methods in online scheduling is hindered by the stochastic nature of the processes and prohibitively long solution times when optimized over long time horizons. The following questions are raised: When to trigger a rescheduling, how much computing resources to allocate, what optimization strategy to use, and how far ahead to schedule? We propose an approach where a reinforcement learning agent is trained to make the first two decisions (i.e., rescheduling timing and computing time allocation). Using neuroevolution of augmenting topologies (NEAT) as the reinforcement learning algorithm, the approach yields, on average, better closed-loop solutions than conventional rescheduling methods on three out of four studied routing problems. We also reflect on expanding the agent’s decision-making to all four decisions.

Direct Metal Layer Forming with Good Adhesion on Porous AAO Films by Electroless Cu Plating

Porous anodic aluminum oxide (AAO) films have great potential for micro and nano fabrication. Development of metallization techniques for AAO films is necessary for AAO device fabrication. In this work, a process for electroless Cu plating on AAO films was investigated. A quick Pd2+ removal process after activation improved the adhesion between plated copper (Cu) layer and AAO films. Since the Pd catalyst at the AAO top surface was effectively removed by a complex agent, the plating reaction started approximately 15–20 μm deep in the nanopores. The Cu was continuously deposited to the top of the nanopores and finally covered the AAO surface. The adhesion strength between the Cu films and AAO films measured by the 90° peeling test method was over 0.5 kN/m. Successful electroless Cu plating on AAO is an all-wet chemical batch process that does not require sophisticated fabrication facilities, and can be used as a low-cost metallization process for producing AAO microdevices.

The design and scheduling of chemical batch processes: Computational complexity studies

Summary The pharmaceutical industry is quite restrictive concerning quality and safety, the manufacturing disruptions often lead to drug shortages in despite of the high costs involved. Due to the minimization of equipment costs, the design and scheduling of chemical batch processes (DSCBP) is a well-known problem, and various sub-problems are selected from the literature because they were successively enlarging the design and schedule policies: single machine or multiple machines (S or M) in each stage; and single product campaigns or multiple products campaigns (SPC or MPC). In this paper, four problems are studied (by combinatorics: SS, MS, SM, and MM) and it is shown that they are all NP-hard in strong sense through polynomial reduction. This study can support innovative algorithms and methodologies for solving DSCBP problems, in a way to improve equipments sizing and configurations design, and thereby contributing to curb disruptions within pharmaceutical supply chains (PharmSC).

Parameters controlling chemical deposits in micro-irrigation with treated wastewater

Micro-irrigation with treated wastewater has the potential to be the most efficient irrigation technique, especially in water scarce areas. Its main disadvantage is the high sensitivity of the drippers to clog. This study focused only on the chemical precipitation mechanisms. In a batch chemical process in parallel with PHREEQC software, two temperatures (22 and 55°C), four pH (8, 8.5, 9 and 9.5) and CO2 partial pressure were tested. The aim was to analyze the quantity of precipitates and their crystalline nature and calculate the effects of these factors on the behavior of dissolved chemical elements in treated wastewater to be able to validate and calibrate a geochemical software in order to predict chemical precipitation. The amount of precipitate increases by increasing pH and temperature. Precipitates were analyzed using thermo gravimetric analysis (TGA) and X-ray diffraction (XRD). Calcium carbonate (CaCO3) in the form of calcite was found to be the predominant precipitate. Experimental and model results showed that the saturation index (SI) of calcite was found to be the factor that most frequently affected calcite precipitation. Calcite SI is pH, temperature and CO2 partial pressure dependent. In the case of irrigation, water equilibrium with atmospheric CO2 minimizes precipitation of calcite.

Computationally‐Light Non‐Lifted Data‐Driven Norm‐Optimal Iterative Learning Control

AbstractComputational complexity and model dependence are two significant limitations on lifted norm optimal iterative learning control (NOILC). To overcome these two issues and retain monotonic convergence in iteration, this paper proposes a computationally‐efficient non‐lifted NOILC strategy for nonlinear discrete‐time systems via a data‐driven approach. First, an iteration‐dependent linear representation of the controlled nonlinear process is introduced by using a dynamical linearization method in the iteration direction. The non‐lifted NOILC is then proposed by utilizing the input and output measurements only, instead of relying on an explicit model of the plant. The computational complexity is reduced by avoiding matrix operation in the learning law. This greatly facilitates its practical application potential. The proposed control law executes in real‐time and utilizes more control information at previous time instants within the same iteration, which can help improve the control performance. The effectiveness of the non‐lifted data‐driven NOILC is demonstrated by rigorous analysis along with a simulation on a batch chemical reaction process.

Inkjet Printing of Single‐Crystalline Bi2Te3 Thermoelectric Nanowire Networks

Large‐scale and low‐cost fabrication techniques are needed to commercialize highly efficient, nanomaterial‐based thermoelectric generators (TEGs) for use in small‐scale, flexible applications (e.g., wearable energy harvesters). This study presents the first demonstration of inkjet‐printed networks of phase‐pure, single‐crystalline Bi2Te3 thermoelectric nanowires (BTNWs) that are amenable to large‐scale production. The BTNWs are synthesized via chemical batch processing and formulated into a jettable ink that is printed onto glass substrates and subsequently annealed in nitrogen and forming gas environments. The inkjet‐printed BTNWs annealed in forming gas provide the most favorable results with comparable thermoelectric performances to bulk Bi2Te3 materials (Seebeck coefficient up to 140 µV K−1) while approximately utilizing only 1% to 3% of the telluride materials found in their bulk counterparts. Thus, these printed BTNWs help pave the way for the development of low‐cost and scalable TEGs.

Systematic retrofit methodology for chemical batch processes to improve economic and production performance

Systematic retrofit methodology for chemical batch processes to improve economic and production performance

Systematic retrofit methodology for chemical batch processes to improve economic and production performance

In the process industry, whenever there is new product introduced then generally it is manufactured at any of the existing plant utilities that are being used for any other previous product. Now with certain modifications in plant setup, it is then utilised and made compatible with formulations process of new product. This process is termed as retrofitting of plant utilities. For example, the product 'A' is being manufactured at a plant that is having five reactors and various other equipments. It is decided that some new product 'B' would be manufactured in the same plant by taking the change over from 'A'. The mechanical engineer section then would do some modifications in the existing plant as they could inter connect only three rectors or more based on the requirement of manufacturing process of new product 'B' and also with certain extra fittings it is made change over to new product 'B'. The objective is not restricted to economic process improvement as ecological aspects related to the process energy consumption are also investigated and Pareto-front analysis is conducted. By retrofitting it has been concluded with a massive change in the efficiency in both the aspects.

Recursive Gaussian Process Regression Model for Adaptive Quality Monitoring in Batch Processes

In chemical batch processes with slow responses and a long duration, it is time-consuming and expensive to obtain sufficient normal data for statistical analysis. With the persistent accumulation of the newly evolving data, the modelling becomes adequate gradually and the subsequent batches will change slightly owing to the slow time-varying behavior. To efficiently make use of the small amount of initial data and the newly evolving data sets, an adaptive monitoring scheme based on the recursive Gaussian process (RGP) model is designed in this paper. Based on the initial data, a Gaussian process model and the corresponding SPE statistic are constructed at first. When the new batches of data are included, a strategy based on the RGP model is used to choose the proper data for model updating. The performance of the proposed method is finally demonstrated by a penicillin fermentation batch process and the result indicates that the proposed monitoring scheme is effective for adaptive modelling and online monitoring.

Optimal Sensor Scheduling in Batch Processes Using Convex Relaxations and Tchebycheff Systems Theory

In several industrial processes, measurements are limited due to time or cost constraints. Hence, scheduling of these measurements to obtain maximally informative and accurate estimates of the states becomes important. The focus of this contribution is to determine a schedule of measurements that maximizes the quality of the estimates at the end of a finite time horizon. This work finds applications in batch chemical processes. The estimate covariance matrix as obtained by applying standard Kalman filtering theory is approximated and the original problem is relaxed and reformulated as a cone program. For a certain class of systems, the approximated covariance matrix is parameterized in terms of a point belonging to a moment space induced by a Tchebycheff system. The theory of Tchebycheff systems is used to determine an improved upper bound on the guaranteed minimum number of measurements. A tractable solution methodology to obtain the optimal cost and a parsimonious discrete schedule using the theory of cone programming, generalized barrier functions and Tchebycheff systems is described.

Automatic steady state identification for batch processes by nonparametric signal decomposition and statistical hypothesis test

In chemical batch processes, online identification of the batch-to-batch steady state is important for ensuring consistency of final product quality and satisfactory process control. In this paper, an automatic steady state identification (SSID) method is developed for batch processes, which utilizes a nonparametric signal decomposition technique named ensemble empirical mode decomposition (EEMD) to extract related information contained in variable trajectories and then conducts a statistical hypothesis test. In the proposed method, EEMD is combined with a moving window procedure to decompose the signal of each variable trajectory into a finite number of intrinsic mode functions (IMFs) in real-time. Then, the inter-batch trend information is extracted by computing the instantaneous frequencies of each IMF. Using the variance ratio test, batch-to-batch steady state can be identified from the inter-batch trend of each process variable. Since most of the disturbance and noise information have been removed through EEMD, robust SSID result can be expected. To deal with the multiple process variables, a multivariate SSID algorithm is proposed based on the statistical test for the equality of covariance matrices. The effectiveness of the proposed method is demonstrated with an injection molding process.