查看更多>>摘要：In this new era of Industry 4.0, manufacturers tend to store process data from the entire production, regardless of whether they are "normal" or "faulty" for further data analysis. However, almost all the existing monitoring models are constructed based on normal data instead of abnormal data. In fact, the "normal" and some of the "faulty" data originate from the same production line. Thus, not only the normal data but also the abnormal ones can be used to improve the monitoring performance of conventional monitoring performance by simultaneously sharing and extracting common knowledge. In this paper, a fault assistant Gaussian mixture prior variational autoencoder (FA-GMPVAE) is proposed to perform information sharing and enhance the statistic model for the normal operating region. Unlike an ordinary variational autoencoder (VAE) and an ordinary Gaussian mixture prior variational autoencoder (GMPVAE), the structure of FA-GMPVAE is a combination of a "normal" based VAE network and a "normal-relevant" based GMPVAE (NR-GMPVAE) network. FA-GMPVAE can make the shared information non-negative to prevent information loss because only the normal-relevant common information is shared by the one-step transfer learning procedure. In addition, fault diagnosis of NR-GMPVAE can be flexibly updated with the new type of fault. Correspondingly, the probability density estimates of latent variables and residuals instead of point estimates are then given so that distribution-based monitoring indices of the normal data can be designed and the fault detection decisions can be made opportunely. To show the effectiveness of the proposed method, a numerical and a real industrial example are presented. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
Foong, S. Z. Y.Andiappan, VAviso, K. B.Chemmangattuvalappil, N. G....
7页
查看更多>>摘要：Most countries have implemented lockdowns as the main non-pharmaceutical intervention in response to the global Covid-19 pandemic. Such actions have resulted in negative economic impacts on different sectors, which include manufacturing and processing industry. This concern has led to the need to develop post-lockdown "exit strategies" to mitigate the adverse socio-economic impacts. In this work, a systematic approach is developed to quantify the criticality of economic sectors based on a sector criticality index (SCI) and rank them based on their importance relative to the entire national economy. SCI is a weighted composite score based on five criteria which capture relevant dimensions in judging the importance of sectors: economic impact, connectivity, sector size, income multiplier, and employment. The input-output model is then used to evaluate the interaction amongst the economic sectors. The use of the proposed approach is illustrated with a case study focusing on evaluating the impact on oleo-chemical industries and associated agro-industrial sectors in Malaysia. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
查看更多>>摘要：Butanol has the potential as a transportation fuel since its energy density is similar to gasoline and higher than ethanol. The Acetone-Butanol-Ethanol (ABE) fermentation is the current main method to produce biobutanol where the fermentation should be operated in diluted surroundings to avoid toxicity to microorganisms. This article aims to develop efficient separation sequences to purify individual components from the ABE fermentation. The base design comes from the reference proposed by Patrascu et al. [3]. With an updated thermodynamic model, several novel design alternatives are proposed and investigated to reduce energy consumption in the separation sequences. Therein, the proposal of removing acetone at first, and purifying ethanol via integration of distillation and membrace not only reduces total energy consumption and total annual cost by 57% and 52%, but also reaches targeting purity of 99.5% for butanol, acetone and ethanol. Heat integration and column stacking can reduce the overall energy consumption furthermore. This study also carries on plantwide control for the processes with heat integration and column stacking. The superior control performance with +/- 10% changes in throughput and feed compositions is validated for the proposed plantwide process control scheme. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
查看更多>>摘要：The stability of the boiler drum level is important for safe and stable operation of industrial plants. In this study, a two-stage training deep deterministic policy gradient (25-DDPG) comprising offline pretraining and online training was proposed to control the boiler drum level. A comparison of simulation results between the 2S-DDPG, DDPG, and 3E training methods proved that 2S-DDPG can robustly control the boiler drum level. The 2S-DDPG model requires less than half as much interaction with online processes as DDPG does; this ensures stable industrial operation due to the lowered risk of process failures in training. The results indicated the integral absolute error of the three-step 2S-DDPG is the lowest among those of the three control models. Moreover, the three-step 2S-DDPG reduced the overshoot percentage calculated using 3E control from 59% to 0%. For processes with noise and time delay, 2S-DDPG exhibits a faster response and less variation in the control performance with regard to the boiler drum level. The manipulated variable distribution errors of the three-step 2S-DDPG were much less than those of the DDPG model. Therefore, 2S-DDPG can address the shortcomings of the traditional DDPG model. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
查看更多>>摘要：Toxic gas leak accidents can lead to significant casualties. The damage from a leak accident will be greater if toxic gas spreads to places where population density is high or vulnerable facilities with people who have difficulty evacuating. And the damage from leak accidents will also be greater if there is no evacuation warning from the plant despite the significant leak. Therefore, it is necessary to design the fenceline monitoring system through sensor placement optimization that reduces the risk toward the community. In this study, a mixed-integer linear programming formulation is proposed for sensor placement minimizing the risk by reflecting the characteristics of nearby areas. In order to quantitatively estimate the risk toward the community, the risk is calculated using dispersion data from computational fluid dynamics simulations and evacuation vulnerability and population density that define characteristics of a community. A case study is conducted to validate the efficiency of the proposed method. Exposed area and residual risk fraction are compared according to sensor placements from different conditions. Top-3 risk leak scenarios are analyzed quantitatively to manage the worst leak accident. It is validated that the proposed method is more efficient especially for the community that has uneven characteristics distribution. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
查看更多>>摘要：Background: Among various approaches specifically developed for processes with inverse response, a predictor design by augmentation of an inverse response predictor in the conventional feedback control system is an attractive method as it reduces the effect of inverse response and facilitates better PID controller performance. However, the existing methods are predominantly model-based methods and the resulting control performance depends on an accurate model which may not be available in practice. Methods: To alleviate the aforementioned shortcoming, a data-based method for the design of an inverse response predictor and a PID controller is developed under the Virtual-Reference-Feedback-Tuning (VRFT) framework in this paper. Furthermore, a simplified stability criterion is developed to verify the closed-loop stability of the predictor control system. Findings: Examples are used to illustrate the utility of the proposed design and a comparison with model-based benchmarks is made. Simulation results show that the proposed design method gives better control performance compared with the benchmarks. (C) 2021 Published by Elsevier B.V. on behalf of Taiwan Institute of Chemical Engineers.
查看更多>>摘要：Heat integration is one of the most used energy conservation methods in chemical processes, and it typically involves the heat transfer from the vapor stream entering a condenser and the liquid stream entering a reboiler. Nevertheless, this work exploits the idea of heat integration between stages in which heat is supplied at a temperature higher than that at the condenser in a high-pressure column and heat is received at a temperature lower than that at the reboiler in a low-pressure column. Since stages between column are subject to heat integration, a large combinatorial problem arises. Thus, this work proposes a hybrid optimization procedure consisting of two steps: The solution of a deterministic optimization (i.e., MILP problem) and stochastic optimization (e.g., Simulated annealing) to take advantage of the strong points from each optimization technique. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
查看更多>>摘要：The selection and design of suitable synthetic paths are important issues that affect the economics and productivity of chemical processes including reactions and the discovery of new chemicals. However, exploration of reaction information is difficult even with reaction databases, causing path explosion that occurs due to the huge search space and conflicting constrains such as economics, safety, efficiency, etc. In this study, we propose an intelligent system ASICS (Advanced System for Intelligent Chemical Synthesis), which supports synthetic path design at the basic stages of research and process design, based on the hybrid generative exploration and exploitation of reaction knowledge base, encoding big data of patented reactions, and machine learning-based retrosynthetic prediction. Based on the pseudo A* search, ASICS generates optimal synthetic paths minimizing scores of the synthetic reaction value function, composed of the synthetic accessibility score, likelihood score and similarity score. The preference in searching between confirmed reaction spaces and unexplored reaction spaces through prediction can be selected by the user. The suggested hybrid approach, combining the reaction knowledge base with the retrosynthetic prediction model, generates feasible and low-cost synthetic paths beyond the accumulated information in patented reactions. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
查看更多>>摘要：One of the factors that influence the performance of packed bed reactors is the flow pattern, and many flow pattern maps have been reported so far. In this study, a gas-liquid flow pattern map of a packed bed compact reactor, which is being increasingly used in the flow synthesis of functional chemicals, was constructed, and two types of flow patterns, trickle flow and pulse flow, were identified. In addition, an estimation method was developed to distinguish the trickle flow and pulse flow from differential pressure or voltage when the flow inside the reactor cannot be visually recognized from the outside. The effects of the liquid flow rate and the slug formation frequency at the mixer on the pulsation frequency, which is a parameter in reaction optimization, were examined. As a result, it was shown that the pulsation frequency increases as the liquid flow rate increases under a fixed gas flow rate and that the pulsation frequency and the slug formation frequency do not always match due to the possibility of coalescence of the gas and liquid slugs at the reactor inlet. (C) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
查看更多>>摘要：Analysis on the alternative designs on top of the optimal network has proven valuable and meaningful for the decision-makers in determining the most suitable options which fulfill a wide range of objective functions. On the basis of an extension of the P-graph framework, a procedure was developed previously for multiple-solution heat exchanger network (HEN) synthesis. This procedure is capable of generating the n-best HENs depending on predefined structural constraints, for example, the maximum number of heat exchangers used for the entire system, the maximum number of sequential heat exchangers on each stream, the maximum number of stream splittings per stream. Since the choice of a parameter influences the effect of other parameters on the result, it is difficult to find the proper set of parameters for the solver that result in all plausible solutions to the original problem. Naturally, this issue emerges in any HEN synthesis problem, its systematic treatment is essential. The purpose of the current work was to develop a heuristic framework for determining the most reasonable parameters of the HEN generation algorithm and guiding the designer through the optimization process. The outcome of this study not only benefits the researchers and industrial practitioners of this field but may also be extended for educational purposes. (C) 2021 Published by Elsevier B.V. on behalf of Taiwan Institute of Chemical Engineers.
NSTL
Elsevier