Data-driven intelligent modeling framework for the steam cracking process

Steam cracking is the dominant technology for producing light olefins,which are believed to be the foundation of the chemical industry.Predictive models of the cracking process can boost production efficiency and profit margin.Rapid advancements in machine learning research have recently enabled data-driven solutions to usher in a new era of process modeling.Meanwhile,its practical application to steam cracking is still hindered by the trade-off between prediction accuracy and computational speed.This research presents a framework for data-driven intelligent modeling of the steam cracking process.Industrial data preparation and feature engineering techniques provide computational-ready datasets for the framework,and feedstock similarities are exploited using k-means clustering.We propose LArge-Residuals-Deletion Multivariate Adaptive Regression Spline(LARD-MARS),a modeling approach that explicitly generates output formulas and eliminates potentially outlying instances.The framework is validated further by the presentation of clustering results,the explanation of variable importance,and the testing and comparison of model performance.

Nystose attenuates bone loss and promotes BMSCs differentiation to osteoblasts through BMP and Wnt/β-catenin pathway in ovariectomized mice

Increasing the osteogenic differentiation ability and decreasing the adipogenic differentiation ability of bone marrow mesenchymal stem cells(BMSCs)is a potential strategy for the treatment of osteoporosis(OP).Naturally derived oligosaccharides have shown significant anti-osteoporotic effects.Nystose(NST),an oligosaccharide,was isolated from the roots of Morinda officinalis How.(MO).The aim of the present study was to investigate the effects of NST on bone loss in ovariectomized mice,and explore the underlying mechanism of NST in promoting differentiation of BMSCs to osteoblasts.Administration of NST(40,80 and 160 mg/kg)and the positive control of estradiol valerate(0.2 mg/kg)for 8 weeks significantly prevented bone loss induced by ovariectomy(OVX),increased the bone mass density(BMD),improved the bone microarchitecture and reduced urine calcium and deoxypyridinoline(DPD)in ovariectomized mice,while inhibited the increase of body weight without significantly affecting the uterus weight.Furthermore,we found that NST increased osteogenic differentiation,inhibited adipogenic differentiation of BMSCs in vitro,and upregulated the expression of the key proteins of BMP and Wnt/β-catenin pathways.In addition,Noggin and Dickkopf-related protein-1(DKK-1)reversed the effect of NST on osteogenic differentiation and expression of the key proteins in BMP and Wnt/β-catenin pathway.The luciferase activities and the molecular docking analysis further supported the mechanism of NST.In conclusion,these results indicating that NST can be clinically used as a potential alternative medicine for the prevention and treatment of postmenopausal osteoporosis.

Predicting power conversion efficiency of binary organic solar cells based on Y6 acceptor by machine learning

Organic solar cells(OSCs)are a promising photovoltaic technology for practical applications.However,the design and synthesis of donor materials molecules based on traditional experimental trial-anderror methods are often complex and expensive in terms of money and time.Machine learning(ML)can effectively learn from data sets and build reliable models to predict the performance of materials with reasonable accuracy.Y6 has become the landmark high-performance OSC acceptor material.We collected the power conversion efficiency(PCE)of small molecular donors and polymer donors based on the Y6 acceptor and calculated their molecule structure descriptors.Then we used six types of algorithms to develop models and compare the predictive performance with the coefficient of determination(R^(2))and Pearson correlation coefficient(r)as the metrics.Among them,decision tree-based algorithms showed excellent predictive capability,especially the Gradient Boosting Regression Tree(GBRT)models based on small molecular donors and polymer donors exhibited that the values of R2are 0.84 and 0.69 for the testing set,respectively.Our work provides a strategy to predict PCEs rapidly,and discovers the influence of the descriptors,thereby being expected to screen high-performance donor material molecules.

Near Optimal Output Feedback Control of Nonlinear Discrete-time Systems Based on Reinforcement Neural Network Learning

In this paper, the output feedback based finitehorizon near optimal regulation of nonlinear affine discretetime systems with unknown system dynamics is considered by using neural networks(NNs) to approximate Hamilton-JacobiBellman(HJB) equation solution. First, a NN-based Luenberger observer is proposed to reconstruct both the system states and the control coefficient matrix. Next, reinforcement learning methodology with actor-critic structure is utilized to approximate the time-varying solution, referred to as the value function, of the HJB equation by using a NN. To properly satisfy the terminal constraint, a new error term is defined and incorporated in the NN update law so that the terminal constraint error is also minimized over time. The NN with constant weights and timedependent activation function is employed to approximate the time-varying value function which is subsequently utilized to generate the finite-horizon near optimal control policy due to NN reconstruction errors. The proposed scheme functions in a forward-in-time manner without offline training phase. Lyapunov analysis is used to investigate the stability of the overall closedloop system. Simulation results are given to show the effectiveness and feasibility of the proposed method.

Saturable absorption and visible pulse modulation of few-layer topological nodal-line semimetal HfGeTe

Topological nodal-line semimetals attract growing research attention in the photonic and optoelectronic fields due to their unique topological energy-level bands and fascinating nonlinear optical responses.Here,to the best of our knowledge,we first report the saturable absorption property of topological nodal-line semimetal HfGeTe and the related pulse modulation in passively Q-switched visible lasers.Few-layer HfGeTe demonstrates outstanding saturable absorption properties in the visible-light band,yielding the saturation intensities of 7.88,12.66,and 6.64μJ/cm^(2)at 515,640,and 720 nm,respectively.Based on an as-prepared few-layer HfGeTe optical switch and a Pr:LiYF_(4)gain medium,Q-switched visible lasers are also successfully achieved at 522,640,and 720 nm.The minimum pulse widths of the green,red,and deep-red pulsed lasers are150,125.5,and 420 ns,respectively.Especially for the green and red pulsed laser,the obtained pulse width is smaller than those of the low-dimensional layered materials.Our work sheds light on the application potential of topological nodal-line semimetals in the generation of visible pulsed lasers.

Optimal regulation of uncertain dynamic systems using adaptive dynamic programming

In this tutorial paper,the finite-horizon optimal adaptive regulation of linear and nonlinear dynamic systems with unknown system dynamics is presented in a forward-in-time manner using adaptive dynamic programming(ADP).An adaptive estimator(AE)is introduced with the idea of Q-learning to relax the requirement of system dynamics in the case of linear system,while neural network-based identifier is utilised for nonlinear systems.The time-varying nature of the solution to the Bellman/Hamilton–Jacobi–Bellman equation is handled by utilising a time-dependent basis function,while the terminal constraint is incorporated as part of the update law of the AE/Identifier in solving the optimal feedback control.Utilising an initial admissible control,the proposed optimal regulation scheme of the uncertain linear and nonlinear system yields a forward-in-time and online solution without using value and/or policy iterations.An adaptive observer is utilised for linear systems in order to relax the need for state availability so that the optimal adaptive control design depends only on the reconstructed states.Finally,the optimal control is covered for nonlinear-networked control systems where in the feedback loop is closed via a communication network.Effectiveness of the proposed approach is verified by simulation results.The end result is a variant of a roll-out scheme in ADP wherein an initial admissible policy is selected as the base policy and the control policy is enhanced using a one-time policy improvement at each sampling interval.