Transparency:The Missing Link to Boosting AI Transformations in Chemical Engineering

The issue of opacity within data-driven artificial intelligence(AI)algorithms has become an impediment to these algorithms’extensive utilization,especially within sensitive domains concerning health,safety,and high profitability,such as chemical engineering(CE).In order to promote reliable AI utilization in CE,this review discusses the concept of transparency within AI utilizations,which is defined based on both explainable AI(XAI)concepts and key features from within the CE field.This review also highlights the requirements of reliable AI from the aspects of causality(i.e.,the correlations between the predictions and inputs of an AI),explainability(i.e.,the operational rationales of the workflows),and informativeness(i.e.,the mechanistic insights of the investigating systems).Related techniques are evaluated together with state-of-the-art applications to highlight the significance of establishing reliable AI applications in CE.Furthermore,a comprehensive transparency analysis case study is provided as an example to enhance understanding.Overall,this work provides a thorough discussion of this subject matter in a way that—for the first time—is particularly geared toward chemical engineers in order to raise awareness of responsible AI utilization.With this vital missing link,AI is anticipated to serve as a novel and powerful tool that can tremendously aid chemical engineers in solving bottleneck challenges in CE.

Intelligent Intercommunicating Multiscale Engineering: The Engineering of the Future

1.Introduction Despite being widely known and investigated as a computer science discipline,artificial intelligence(AI)has attracted incomparable interest from researchers in diversified areas[1].In 1950,Alan Turing raised the classic question that has inspired numerous researchers to date:“Can machines think?”[2].The ultimate benchmark of AI was set by Turing’s revised“imitation game.”

Micro-flow structure at regime transition from bubbling to turbulent fluidization in a fluidized bed

Gas-solid fluidized beds have found extensive utilization in frontline manufacturing,in particular as low-velocity beds.The fluidization status,the bubbling or turbulent flow regime and the transition in between,determine the system performance in practical applications.Though the convoluted hydrodynamics are quantitively evaluated through numerous data-processing methodologies,none of them alone can reflect all the critical information to identify the transition from the bubbling to the turbulent regime.Accordingly,this study was to exploit a coupling data processing methodology,in the combination of standard deviation,power spectrum density,probability density function,wavelet transform,and wavelet multiresolution method,to jointly explain the micro-flow structure at the regime transition from bubbling to turbulent fluidization.The transient differential pressure fluctuation was measured for the evaluation in a fluidized bed(0.267 m i.d.×2.5 m height)with FCC catalysts(d_(p)=65μm,ρ_(p)=1780kg/m^(3))at different superficial gas velocities(0.02–1.4 m/s).The results show that the onset of turbulent fluidization starts earlier in the top section of the bed than in the bottom section.The wavelet decomposition displays that the fluctuation of differential pressure mainly concentrates on the sub-signals with an intermediate frequency band.These sub-signals could be synthesized into three types of scales(micro-scale,meso-scale,and macro-scale),representing the multi-scale hydrodynamics in the fluidized bed.The micro-scale signal has the characteristic information of bubbling fluidization,and the characteristic information of turbulent fluidization is mainly represented by the meso-scale signal.This work provides a systematic comprehension of fluidization status assessment and serves as an impetus for more coupling analysis in this sector.

Machine learning accelerated carbon neutrality research using big data-from predictive models to interatomic potentials

Carbon neutrality has been proposed as a solution for the current severe energy and climate crisis caused by the overuse of fossil fuels, and machine learning(ML) has exhibited excellent performance in accelerating related research owing to its powerful capacity for big data processing. This review presents a detailed overview of ML accelerated carbon neutrality research with a focus on energy management, screening of novel energy materials, and ML interatomic potentials(MLIPs), with illustrations of two selected MLIP algorithms: moment tensor potential(MTP) and neural equivariant interatomic potential(NequIP). We conclude by outlining the important role of ML in accelerating the achievement of carbon neutrality from global-scale energy management, unprecedented screening of advanced energy materials in massive chemical space, to the revolution of atomicscale simulations of MLIPs, which has the bright prospect of applications.