人工智能在注射成型参数设置及优化中的研究进展

对专家系统及案例推理、进化计算和机器学习这3类注射成型工艺参数设置及优化的人工智能技术发展现状进行了综述,并对今后的研究方向提出了建议。     

人工智能在注射成型参数设置及优化中的研究进展

对专家系统及案例推理、进化计算和机器学习这3类注射成型工艺参数设置及优化的人工智能技术发展现状进行了综述,并对今后的研究方向提出了建议。     

Genetic-based multi-objective optimization of alkylation process by a hybrid model of statistical and artificial intelligence approaches

The purpose of this research is to find the optimal operating point in the production process of the cumene. Therefore, the production process was optimized through statistical and genetic algorithm-based methods. The performance of an alkylation reactor was optimized through maximizing the yield of cumene production. Response surface methodology (RSM) with design type of central composite was applied for design of experiment, modelling, and optimizing the process. The analysis of variance (ANOVA) was performed for finding the important operative parameters as well as their effects. The effects of three parameters including temperature, reactor length, and pressure on the alkylation process were investigated. Further, two types of feed-forward neural network were applied to model the alkylation reactor. To develop the neural network model, leave-one-out method was used. The best prediction performance belonged to a fitting network with 2 and 8 neurons in the hidden layer, respectively. This model was used for optimizing the performance of the alkylation reactor. The statistical and artificial intelligence systems were capable of prediction of cumene production yield in different conditions with R² of 0.9098 and 0.9986, respectively. Genetic algorithm-based optimization was performed by the developed neural network model. The maximum accessible value of cumene production yield was 0.7771, which can be achieved when the temperature, length of reactor, and column pressure are 160°C, 2 m, and 4000 kPa, respectively. By finding the optimal operating point in the cumene production process, capital cost, energy consumption, and other operating costs can be significantly reduced.     

回转窑人工智能专家系统的应用实践

根据水泥熟料煅烧过程操作控制原理研发的"回转窑人工智能专家系统"采用了人工智能模拟技术、实时数据清洗技术、控制数据关联匹配技术。该系统由分解炉出口温度自动控制系统、篦速自动控制系统和氨水用量自动控制系统三个模块组成。该系统投运后,分解炉出口温度控制有效率由之前人工控制的平均60%~70%提高至平均80%~95%;篦下压力和二次风温稳定性提高;氨水消耗量下降显著;窑系统投料量增加10 t/h左右,熟料标准煤耗降低2.85 kg/t,熟料强度均有提高。     

Reconciling deep learning and first-principle modelling for the investigation of transport phenomena in chemical engineering

The use of machine learning in chemical engineering has the potential to greatly improve the design and analysis of complex systems. However, there are also risks associated with its adoption, such as the potential for bias in algorithms and the need for careful oversight to ensure the safety and reliability of machine learning-powered systems. This paper explores the opportunities and risks of using machine learning in chemical engineering and provides a perspective on how it may be integrated into engineering practices in a responsible and effective manner. We generated the text of this abstract with GPT-3, OpenAI's large-scale language-generation model. Upon generating the draft, we ensured that the language was to our liking, and we take ultimate responsibility for the content of this publication.     

酶工程：从人工设计到人工智能

计算机在酶工程中的应用使得酶的序列空间探索度不断被扩大。随着不同分子力场参数的建立,涌现出诸多以计算分子能量为基础的算法,并被用于酶的催化活性、稳定性、底物特异性等的改造与筛选。伴随计算机硬件的提升与算法的优化,从头设计全新功能的人工酶取得成功并得以发展。近年来,人工智能在蛋白质结构预测上不断获得突破,同时也被应用到酶的设计中。介绍了分子力场基础和酶设计与筛选的算法,重点阐述了从头设计的方法和成功案例,以及机器学习设计酶的流程和最新的研究进展,展望了人工智能在酶工程领域的未来发展,为酶的改造与全新功能的生物催化剂的设计助力。     

基于人工智能技术的轮胎磨损性能优化

利用三维有限元分析软件,在正交试验设计、神经网络和遗传算法的基础上,提出一种新的结构优化方法,以12.00R20子午线轮胎的带束层结构为设计变量,以轮胎接地压力分布的偏度值为目标函数进行优化分析。数值和试验分析表明,该方法可以优化轮胎接地压力分布,改善轮胎磨损性能。     

Uncertainty quantification in machine learning and nonlinear least squares regression models

Machine learning (ML) models are valuable research tools for making accurate predictions. However, ML models often unreliably extrapolate outside their training data. The multiparameter delta method quantifies uncertainty for ML models (and generally for other nonlinear models) with parameters trained by least squares regression. The uncertainty measure requires the gradient of the model prediction and the Hessian of the loss function, both with respect to model parameters. Both the gradient and Hessian can be readily obtained from most ML software frameworks by automatic differentiation. We show examples of the uncertainty method in applications of molecular simulations and neural networks. We further show that the uncertainty measure is larger for input space regions that are not part of the training data. Therefore, this method can be used to identify extrapolation and to aid in selecting training data or assessing model reliability.     

智能诊断在旋转机械故障研究中的应用

旋转机械是广泛应用于石化等行业的关键设备,随着计算机技术和人工智能技术的迅速发展,智能诊断技术为旋转机械的故障分析提供了先进手段。论文概述了旋转机械故障智能诊断最重要的九种方法及其工程运用,并对每种方法列举其优缺点。从大量的科研和实例中得出结论,指出了今后的发展方向。     

化工过程的智能混合建模方法及应用

随着人工智能技术和配套数据系统的快速发展,化工过程建模技术达到了新的高度,将多个机理模型和数据驱动模型以合理的结构加以组合的智能混合建模方法,可以综合利用化工过程的第一性原理及过程数据,结合人工智能算法以串联、并联或者混联的形式解决化工过程中的模拟、监测、优化和预测等问题,建模目的明确,过程灵活,形成的混合模型有着更好的整体性能,是近年来过程建模技术的重要发展趋势。本文围绕近年来针对化工过程的智能混合建模工作进行了总结,包括应用的机器学习算法、混合结构设计、结构选择等关键问题,重点论述了混合模型在不同任务场景下的应用。指出混合建模的关键在于问题和模型结构的匹配,而提高机理子模型性能,获取高质量宽范围的数据,深化对过程机理的理解,形成更有效率的混合建模范式,这些都是现阶段提高混合建模性能的研究方向。     

人工智能在电化学水处理过程中的应用

近年来,通过人工智能建立的模型可以对工业过程进行精确调控,人工智能应用于电化学水处理技术过程得到了广泛关注。在电化学水处理过程中,人工智能模型可以降低电化学过程的能耗,获取最优能效比。本文对人工智能在电化学水处理的应用进行了综述、分类和归纳,并介绍了其应用方法,概述了人工智能应用于电化学水处理过程的特点、优势以及局限性,比较了用于电化学水处理的人工智能建模与响应面模型、回归模型和经验动力学模型的优劣。进而提出了人工智能在工程应用上的改进思路,为相关研究提供了参考。     

Advances in artificial intelligence for the colour industry

The use of artificial intelligence in finding solutions to traditional and emergent colorimetric problems is reviewed. Developments in artificial intelligence are briefly discussed and some of the more prominent ideas are described. Rather than give the complete background to specific technologies that are being used, the author considers the application of these technologies. Applications relevant to practical colour measurement are considered that complement the CMC(2:1) equation.     

人工神经网络在橡胶工业中的应用进展

简要介绍了人工神经网络的原理、特点、结构、发展历史及现状,着重总结了该方法在橡胶工业的机械设计、性能预测、配方优化等方面的应用情况,指出了前人研究成果的创新之处及现阶段存在的问题,并对人工神经网络在橡胶工业领域的发展趋势做了预测。     

典型湿法脱硫系统存在的问题及人工智能在优化运行中的应用

石灰石-石膏湿法烟气脱硫是我国最主要的脱硫方式,随着该法的不断改进与简化,脱硫系统运行更加稳定,造价也在不断降低,但运行过程中依然存在诸多问题。本文概括总结了石灰石-石膏湿法烟气脱硫系统运行过程中发生的典型问题,诸如结垢堵塞、腐蚀磨损、起泡溢流和石膏品质差等,指出了这些问题的具体分类类型以及主要影响因素;继而从系统运行机理角度阐述了这些问题的相互关系和交互影响,并将这些问题归因于脱硫系统的紊乱。随后结合电厂目前条件、当前研究基础以及可以实现的技术手段从四个方面论证了人工智能应用于脱硫系统的可行性,建议借助人工智能建立系统运行的诊断模型解决已有问题,并提出了基于人工智能的脱硫系统概念模型及具体实施思路。     

人工神经网络的发展及其在控制中的应用

本文综述了人工神经网络的发展历程,介绍了神经网络的基本结构和训练原理,并着重阐述了在化工过程控制领域中的应用现状和前景。     

Towards design guidelines for virtual reality training for the chemical industry

Operator training in the chemical industry is important because of the potentially hazardous nature of procedures and the way operators' mistakes can have serious consequences on process operation and safety. Currently, operator training is facing some challenges, such as high costs, safety limitations and time constraints. Also, there have been some indications of a lack of engagement of employees during mandatory training. Immersive technologies can provide solutions to these challenges. Specifically, virtual reality (VR) has the potential to improve the way chemical operators experience training sessions, increasing motivation, virtually exposing operators to unsafe situations, and reducing classroom training time. In this paper, we present research being conducted to develop a virtual reality training solution as part of the EU Horizon 2020 CHARMING Project, a project focusing on the education of current and future chemical industry stakeholders. This paper includes the design principles for a virtual reality training environment including the features that enhance the effectiveness of virtual reality training such as game-based learning elements, learning analytics, and assessment methods. This work can assist those interested in exploring the potential of virtual reality training environments in the chemical industry from a multidisciplinary perspective.     

TeCSMART: A hierarchical framework for modeling and analyzing systemic risk in sociotechnical systems

Recent systemic failures in different domains continue to remind us of the fragility of complex sociotechnical systems. Although these failures occurred in different domains, there are common failure mechanisms that often underlie such events. Hence, it is important to study these disasters from a unifying systems engineering perspective so that one can understand the commonalities as well as the differences to prevent or mitigate future events. A new conceptual framework that systematically identifies the failure mechanisms in a sociotechnical system, across different domains is proposed. Our analysis includes multiple levels of a system, both social and technical, and identifies the potential failure modes of equipment, humans, policies, and institutions. With the aid of three major recent disasters, how this framework could help us compare systemic failures in different domains and identify the common failure mechanisms at all levels of the system is demonstrated. © 2016 American Institute of Chemical Engineers AIChE J, 62: 3065–3084, 2016     

Catalyst design using artificial intelligence: SO2 to SO3 case study

Catalyst design is key to the improvement of chemical process efficiency. The required work for the development of new catalysts can be supported through the proper application of artificial intelligence to identify optimal compositions. A generic methodology for the application of machine learning to catalysis research is therefore outlined in this work. The catalytic oxidation of SO₂ was used to exemplify the first iteration of this methodology. 1784 data points from 31 published papers were compiled into a databank. The inlet SO₂ concentration ranged from 0 to 66 mol%. An artificial neural network (ANN) was trained on the databank in order to predict SO₂ conversion based on the catalyst composition and the reactor operating conditions (temperature, pressure, catalyst mass: volumetric flowrate ratio (w/v), and feed composition). The model achieved a root-mean-square error of 6.6%. A preliminary screening step identified 3:1 V-Mg/SiO₂ catalysts as exhibiting high conversion at 648 K. A multi-objective optimization was then performed on a single catalyst to identify solutions exhibiting high conversion and high productivity at 648 K while minimizing the catalyst cost. The optimal solution was predicted to be a 2.9 wt% V/0.2 wt% Mg/SiO₂ catalyst operating at a w/v of 7.49 kg-cat · s/m³ STP, achieving 100% SO₂ conversion with a material cost among the bottom third of cost values. Artificial intelligence can then be employed to extract useful knowledge from published catalytic data and orient future search for novel catalyst development.     

Calculating porosity and permeability from synthetic micro-CT scan images based on a hybrid artificial intelligence

Nowadays, lattice Boltzmann is one of the standard and exact methods of simulation in micro-CT images of rock. However, it has a high weakness in run time. Therefore, the effort in this article is to reach a comprehensive substitute method for permeability calculation with less run time than the lattice Boltzmann method. The other purposes are the automation of processing operations, preparation of images, and in the end, the calculation of porosity. The best way to achieve these outcomes is to use hybrid artificial intelligence. In this research work, comprehensive model architecture has been used to design a hybrid artificial intelligence to be able to calculate permeability and porosity in complex images. A thousand images were randomly generated with high complexity, which makes the model comprehensive and extensible, and image processing was applied. After that, the lattice Boltzmann method as the direct simulation was selected. Finally, the convolutional neural network and multilayer perceptron based on a new and comprehensive model were evaluated for the first time; the mean squared error resulting from the evaluation of training data is 0.01, and the test data is 0.03. Expert systems have been used as a subset of artificial intelligence for automated image processing and porosity calculation. In this way, problems related to the direct implementation of classical algorithms for image processing, models, and patterns related to machine learning and needing an expert were solved to an acceptable extent, and an error of less than 5% was achieved.     

Biosorption of Cr(VI) from aqueous solution using agricultural wastes,with artificial intelligence approach

Removal of Cr(VI) from aqueous solution by date-palm-leaves (DPL) and broad-bean-shoots (BBS) was investigated. FTIR, SEM, and EDAX showed that DPL has higher ability for ion-exchange to remove Cr(VI). Langmuir and Freundlich adsorption isotherms and kinetics revealed that DPL exhibited higher biosorption capacity. At Cr(VI) 100 mg/L, biosorbent-dose 5 g/L and 60 min contact-time, maximum Cr(VI) removal for DPL (98%) and BBS (95%) was achieved at pH 2 and 1, respectively. Adaptive-neuro fuzzy inference system determined the most important factor affecting Cr(VI) removal. The model indicated that DPL is more tolerant to pH levels, while BBS is a pH-sensitive adsorbent.