Early identification of process deviation based on convolutional neural network

A novel process monitoring method based on convolutional neural network(CNN)is proposed and applied to detect faults in industrial process.By utilizing the CNN algorithm,cross-correlation and autocorrelation among variables are captured to establish a prediction model for each process variable to approximate the first-principle of physical/chemical relationships among different variables under normal operating conditions.When the process is operated under pre-set operating conditions,prediction residuals can be assumed as noise if a proper model is employed.Once process faults occur,the residuals will increase due to the changes of correlation among variables.A principal component analysis(PCA)model based on the residuals is established to realize process monitoring.By monitoring the changes in main feature of prediction residuals,the faults can be promptly detected.Case studies on a numerical nonlinear example and data from two industrial processes are presented to validate the performance of process monitoring based on CNN.

An innovative flake graphite upgrading process based on HPGR,stirred grinding mill, and nanobubble column flotation

Physical upgrading of graphite is typically achieved with many stages of grinding and flotation to produce a concentrate with approximately 95% carbon grade.An innovative grinding and column flotation process has been developed for efficient graphite upgrading to substantially simplify the process flowsheet and reduce operating costs.In this process,a high-pressure grinding roller(HPGR) and a stirred mill were employed as primary comminution techniques and a nanobubble flotation column as a key separation process.The results obtained with a crystalline flake graphite sample with a carbon grade of 11.15% show that the novel process can produce a concentrate with 94.82% carbon grade and 97.89% recovery from an open circuit of one rougher and two cleaner flotation stages.Scanning electron microscope(SEM)microphotographs indicate that HPGR offers the advantage of more effective protection of graphite flakes during crushing.Grinding test results show that stirred mill could not only protect graphite flakes but also promote the efficient liberation of graphite.Compared with the traditional flotation process,nanobubble flotation can effectively recover ultrafine graphite.The new process possesses a number of important advantages over the traditional method,including substantially higher graphite recovery,greatly simplified process flowsheet,better protection of flake size,reduced reagent consumption and process costs,etc.

Surface nanobubble characterization and its enhancement mechanisms for fine-particle flotation:A review

Froth flotation is often used for fine-particle separation,but its process efficiency rapidly decreases with decreasing particle size.The efficient separation of ultrafine particles(UFPs)has been a major challenge in the mineral processing field for many years.In recent years,the use of surface nanobubbles in the flotation process has been recognized as an effective approach for enhancing the recovery of UFPs.Compared with traditional macrobubbles,nanobubbles possess unique surface and bulk characteristics,and their effects on the UFP flotation behavior have been a topic of intensive research.This review article is focused on the studies on various unique characteristics of nanobubbles and their mechanisms of enhancing the UFP flotation.The purpose of this article is to summarize the major achievements on the two topics and pinpoint future research needs for a better understanding of the fundamentals of surface nanobubble flotation and developing more feasible and efficient processes for fine and UFPs.

Strongly nonlinear topological phases of cascaded topoelectrical circuits

Circuits provide ideal platforms of topological phases and matter,yet the study of topological circuits in the strongly nonlinear regime,has been lacking.We propose and experimentally demonstrate strongly nonlinear topological phases and transitions in one-dimensional electrical circuits composed of nonlinear capacitors.Nonlinear topological interface modes arise on domain walls of the circuit lattices,whose topological phases are controlled by the amplitudes of nonlinear voltage waves.Experimentally measured topological transition amplitudes are in good agreement with those derived from nonlinear topological band theory.Our prototype paves the way towards flexible metamaterials with amplitude-controlled rich topological phases and is readily extendable to two and three-dimensional systems that allow novel applications.

Submicrometer optical frequency combs based on SPPs metallic multi-ring resonators

Optical frequency combs(OFCs)have great potential in communications,especially in dense wavelength-division multiplexing.However,the size of traditional OFCs based on conventional optical microcavities or dispersion fibers is at least tens of micrometers,far larger than that of nanoscale electronic chips.Therefore,reducing the size of OFCs to match electronic chips is of necessity.Here,for the first time to our knowledge,we introduce surface plasmon polaritons(SPPs)to the construction of OFCs to realize a miniature device.The thickness of our device is reduced below 1μm.Though the presence of SPPs may induce ohmic and scattering loss,the threshold of the device is obtained as 9μW,comparable to the conventional device.Interestingly,the response time is 13.2 ps,much faster than the optical counterparts.This work provides a feasible strategy for the miniaturization of OFCs.