Application of machine learning to process simulation of n-pentane cracking to produce ethylene and propene

Modeling light olefin production was one of the main concerns in chemical engineering field.In this paper,machine learning model based on artificial neural networks(ANN)was established to describe the effects of temperature and catalyst on ethylene and propene formation in n-pentane cracking.The establishment procedure included data pretreatment,model design,training process and testing process,and the mean square error(MSE)and regression coefficient(R2)indexes were employed to evaluate model performance.It was found that the learning algorithm and ANN topology affected the calculation accuracy.GD24223,CGB2423,and LM24223 models were established by optimally matching the learning algorithm with ANN topology,and achieved excellent calculation accuracy.Furthermore,the stability of GD24223,CGB2423 and LM24223 models was investigated by gradually decreasing training data and simultaneously transforming data distribution.Compared with GD24223 and LM24223 models,CGB2423 model was more stable against the variations of training data,and the MSE values were always maintained at the magnitude of 10^-3-10^-4,confirming its applicability for simulating light olefin production in n-pentane cracking.

Refined modeling and experimental verification of a torque motor for an electro-hydraulic servo valve

The Permanent Magnet Torque Motor(PMTM)is the key electro-mechanical conversion device in an Electro-Hydraulic Servo Valve(EHSV).In this work,a refined model of a PMTM is developed,considering the non-working air-gaps between the upper or lower yoke and the armature,the fringing effect at the limiting holes,and the nonlinear permeability of soft magnetic material.Based on the refined model,the influences of various factors on the calculation accuracy of the magnetic flux at the pole surfaces of the armature and the output torque are investigated.For verifying the validity of the refined model,a Finite Element Analysis(FEA)of the PMTM is conducted,and a test platform is constructed.Compared with existing models,the refined model can better reveal the intrinsic mechanism of the PMTM,and its calculations are more consistent with the FEA results.The experimental results of the armature deflection displacement show that the refined model can accurately describe the output characteristics of the PMTM.

Evaluation and comparison of various fast fluid dynamics modeling methods for predicting airflow around buildings

Computational fluid dynamics(CFD)methods are being increasingly used for predicting airflow fields around buildings,but personal computers can still take tens of hours to create a single design using traditional computing models.Considering both accuracy and efficiency,this study compared the performances of the conventional algorithm PIMPLE,fast fluid dynamics(FFD),semi-Lagrangian PISO(SLPISO),and implicit fast fluid dynamics(IFFD)in OpenFOAM for simulating wind flow around buildings.The effects of calculation parameters,including grid resolution,discrete-time step,and calculation time for these methods are analyzed.The results of the simulations are compared with wind tunnel tests.It is found that IFFD and FFD have the fastest calculation speeds,but also have the largest discrepancies with test data.The PIMPLE algorithm has the highest accuracy,but with the slowest calculation speed.The calculation speeds of the FFD,SLPISO,and IFFD models are 6.3,3 and 13.3 times faster than the PIMPLE model,respectively.The calculation accuracy and speed of the SLPISO model are in between those of the IFFD,FFD and PIMPLE models.An appropriate algorithm for a project may be chosen based on the requirements of the project.