基于机器学习聚类算法的水平井流型预测

随着水平井钻井技术的日趋完善其规模不断扩大,井下流体流型判断也成为测井作业中的首要工作,通过利用聚类算法对现有测井数据进行预测分析并将结果与实验流体流型进行对比,算法整体预测准确度能够达到80%以上,验证了此方法在流型预测方面的有效性并为井下流体流型预测提供了一种新思路。

基于机器学习BP算法和树模型算法的井筒流体流型预测

近年来随着石油行业的不断发展,为适应油藏在地下不同分布情况,除了传统采用的垂直井以外,大斜度井、水平井也不断被利用于对油藏的探测与开发。判断不同斜度井在不同流量、含水率下井中流体的流型就成了一个亟需解决的问题。本文通过分别使用BP神经网络和树模型,对不同流量、含水率的各斜度井井下流体流型进行预测,与实验所获取的数据进行对比,两种算法预测准确度分别能达到75%和91.7%,验证了BP神经网络算法和树模型算法在流型预测中的有效性,为后续井下流型研究提供了一种新的方案。

基于PCA-GA-SVM的竖直下降两相流流型预测

为提高小样本条件下的流型识别精度和时效性,提出了一种融合小波包分解(WPD)、主元分析(PCA)、遗传算法(GA)和支持向量机(SVM)的优化识别模型,并成功应用在竖直下降两相流流型辨识工作中。利用WPD对非平稳电导波动信号进行分解、重构,提取小波包能量构造特征向量;通过PCA对特征向量进行降维,降低特征输入的复杂性;同时采取GA全局迭代寻优的方式确定SVM的关键参数惩罚因子(C)和核函数参数(g)。对PCA-GA-SVM识别效果进行验证后与SVM、PCA-SVM、GASVM网络进行对比。结果表明,经过PCA和GA优化后的SVM网络在流型识别精度和时效性方面均提升显著,对泡状流、弹状流、搅拌流和环状流的总体预测精度达到了94.87%,耗时仅3.95 s,可满足在线识别需求。

基于AFSA-RF的两相流型图扩展技术

为预测高流速条件下的流型并建立流型图,提出一种基于人工鱼群算法(artificial fish swarm algorithm,AFSA)优化的随机森林(random forest,RF)的机器学习模型,基于最优、简化参数出发,进行流型的智能识别。该模型成功地应用于竖直下降两相流流型的识别,通过不同分类模型以及优化方法对实验数据进行计算,发现AFSA-RF模型的流型识别精度与稳定性高于未优化的RF模型以及其他主流优化方法,对高流速区域的流型的识别成功率达到了90.91%,进一步验证了该预测模型的有效性。依托建立的模型,对现有流型图的适应范围进行了扩展,获得了适用于高流速条件下的流型图。

不同黏度液体的垂直气液两相管流流型判别

为了得到不同黏度下垂直管气-液两相流搅动流与环状流流型判别方法,整理了多相流实验室不同黏度的流体样本实验数据,对现有的搅动流与环状流流型判别方法进行了评估,分析了现有方法的不足和潜在的改进。评估结果表明,现有方法未对流体的黏度进行考虑,导致预测结果不佳。基于此,提出了一种不同黏度范围下的搅动流与环状流流型判别方法,即向现有方法添加了黏度的无因次量,适用范围为5~150 mPa·s。并将实验数据与现有方法进行对比,发现所提方法能够更精确地预测搅动流与环状流边界。

海上油气水多相管流中沿程摩阻因数分析

沿程摩阻因数是制约多相管流发展的关键,摩阻因数是压降计算的重要内容之一,其计算的准确性直接决定了压降计算的准确性。基于Xiao等人的流型判别法,将流型分为分层流、段塞流、环状流和分散泡状流等4种流型。分散泡状流仅需用到气体或液体与管壁之间的相互作用;分层流、环状流和段塞流不仅用到气体或液体与管壁之间的相互作用,还要用到气液界面之间的相互作用。气体或液体与管壁之间的沿程摩阻因数可以采用单相流体的沿程摩阻因数方法计算。对于不同的流型,气液界面的摩阻因数计算方法也不同。

重力热管喷涌不稳定性调控判据

重力热管内的两相流型对其工作特性有着至关重要的影响。本文借鉴垂直圆管内两相流型转变的实验研究结果和近饱和自由降膜流始发沸腾的观测结论,通过简化分析,给出重力热管蒸发段内由泡状流向弹状流过渡的流型转变判据。在低饱和蒸汽压和高过冷度下,这种流型转变将导致周期性振荡的两相流不稳定性,这就是喷涌不稳定性;因此,流型转变判据也就是喷涌不稳定性调控判据。文中还结合数值计算实例,指出影响流型转变的重要因素。

Prediction of Pressure Drop of Slurry Flow in Pipeline by Hybrid Support Vector Regression and Genetic Algorithm Model

This paper describes a robust support vector regression （SVR） methodology, which can offer superior performance for important process engineering problems. The method incorporates hybrid support vector regression and genetic algorithm technique （SVR-GA） for efficient tuning of SVR meta-parameters. The algorithm has been applied for prediction of pressure drop of solid liquid slurry flow. A comparison with selected correlations in the lit- erature showed that the developed SVR correlation noticeably improved the prediction of pressure drop over a wide range of operating conditions, physical properties, and pipe diameters.

A Regional-Scale Method of Forecasting Debris Flow Events Based on Water-Soil Coupling Mechanism

A debris flow forecast model based on a water-soil coupling mechanism that takes the debrisflow watershed as a basic forecast unit was established here for the prediction of disasters at the watershed scale.This was achieved through advances in our understanding of the formation mechanism of debris flow.To expand the applicable spatial scale of this forecasting model,a method of identifying potential debris flow watersheds was used to locate areas vulnerable to debris flow within a forecast region.Using these watersheds as forecasting units and a prediction method based on the water-soil coupling mechanism,a new forecasting method of debris flow at the regional scale was established.In order to test the prediction ability of this new forecasting method,the Sichuan province,China was selected as a study zone and the large-scale debris flow disasters attributable to heavy rainfall in this region on July 9,2013 were taken as the study case.According to debris flow disaster data on July 9,2013 which were provided by the geo-environmental monitoring station of Sichuan province,there were 252 watersheds in which debris flow events actually occurred.The current model predicted that 265 watersheds were likely to experience a debris flow event.Among these,43 towns including 204 debrisflow watersheds were successfully forecasted and 24 towns including 48 watersheds failed.The false prediction rate and failure prediction rate of thisforecast model were 23% and 19%,respectively.The results show that this method is more accurate and more applicable than traditional methods.

Connection Between Liquid Distribution and Gas-Liquid Mass Transfer in Monolithic Bed

With a particular focus on the connection between liquid flow distribution and gas-liquid mass transfer in monolithic beds in the Taylor flow regime, hydrodynamic and gas-liquid mass transfer experiments were carriedout in a column with a monolithic bed of cell density of 50 cpsi with trio different distributors （nozzle and packed bed distributors）. Liquid saturation in individual channels was measured by using self-made micro-conductivity probes. A mal-distribution factor was used to evaluate uniform degree of phase distribution in monoliths. Overall bed pressure drop and mass transfer coefficients were measured. For liquid flow distribution and gas-liquid masstransfer, it is found that the superficial liquid velocity is a crucial factor and the packed bed distributor is better than the nozzle distributor. A semi-theoretical analysis using single channel models shows that the packed bed distributor always yields shorter and uniformly distributed liquid slugs compared to the nozzle distributor, which in turn ensures a better mass transfer performance. A bed scale mass transfer model is proposed by employing the single channel models in individual channels and incorporating effects of non-uniform liquid distribution along the bedcross-section. The model predicts the overall gas-liquid mass transfer coefficient wig a relative error within ＋30%.

A new mathematical model for predicting flow stress of X70HD under hot deformation

To realize numerical simulation of rolling and obtain the hot forming process parameters for X70 HD steel, the flow stress behaviors of X70 HD steel were investigated under different temperatures（820-1100 ℃ and strain rates（0.01-10 s-1） on a Gleeble-3500 thermo-simulation machine. A new flow stress model was established. The linear and exponential relationship methods were applied to the parameters with respect to temperature and deformation rates. The rise of curve ends under certain conditions was analyzed. The flow stress of X70 HD steel predicted by the proposed model agrees well with the experimental results. So, it greatly improves the precision of the metal thermoplastic processing through finite element method and practical application of engineering.

Traffic flow prediction of urban road network based on LSTM-RF model

Traffic flow prediction,as the basis of signal coordination and travel time prediction,has become a research point in the field of transportation.For traffic flow prediction,researchers have proposed a variety of methods,but most of these methods only use the time domain information of traffic flow data to predict the traffic flow,ignoring the impact of spatial correlation on the prediction of target road segment flow,which leads to poor prediction accuracy.In this paper,a traffic flow prediction model called as long short time memory and random forest(LSTM-RF)was proposed based on the combination model.In the process of traffic flow prediction,the long short time memory(LSTM)model was used to extract the time sequence features of the predicted target road segment.Then,the predicted value of LSTM and the collected information of adjacent upstream and downstream sections were simultaneously used as the input features of the random forest model to analyze the spatial-temporal correlation of traffic flow,so as to obtain the final prediction results.The traffic flow data of 132 urban road sections collected by the license plate recognition system in Guiyang City were tested and verified.The results show that the method is better than the single model in prediction accuracy,and the prediction error is obviously reduced compared with the single model.

Pedestrian environment prediction with different types of on-shore building distribution

The aim of this work is to evaluate how the building distribution influences the cooling effect of water bodies. Different turbulence models, including the S-A, SKE, RNG, Realizable, Low-KE and RSM model, were evaluated, and the CFD results were compared with wind tunnel experiment. The effects of the water body were detected by analyzing the water vapor distribution around it. It is found that the RNG model is the most effective model in terms of accuracy and computational economy. Next, the RNG model was used to simulate four waterfront planning cases to predict the wind, thermal and moisture environment in urban areas around urban water bodies. The results indicate that the building distribution, especially the height of the frontal building, has a larger effect on the water vapor dispersion, and indicate that the column-type distribution has a better performance than the enclosed-type distribution.

Predictive current control system of PMSM based on LADRC

For a permanent magnet synchronous motor(PMSM)model predictive current control(MPCC)system,when the speed loop adopts proportional-integral(PI)control,speed regulation is easily affected by motor parameters,resulting in the inability to balance the system robustness and dynamic performance.A PMSM optimal control strategy combining linear active disturbance rejection control(LADRC)and two-vector MPCC(TV-MPCC)is proposed.Firstly,a mathematical model of a PMSM is presented,and the PMSM TV-MPCC model is developed in the synchronous rotation coordinate system.Secondly,a first-order LADRC controller composed of a linear extended state observer and linear state error feedback is designed to reduce the complexity of parameter tuning while linearly simplifying the traditional active disturbance rejection control(ADRC)structure.Finally,the conventional PI speed regulator in the motor speed control system is replaced by the designed LADRC controller.The simulation results show that the speed control system using LADRC can effectively deal with the changes in motor parameters and has better robustness and dynamic performance than PI control and similar methods.The system has a fast motor speed response,small overshoot,strong anti-interference,and no steady-state error,and the total harmonic distortion is reduced.

Model predictive current control for PMSM driven by three-level inverter based on fractional sliding mode speed observer

Based on the fractional order theory and sliding mode control theory,a model prediction current control(MPCC)strategy based on fractional observer is proposed for the permanent magnet synchronous motor(PMSM)driven by three-level inverter.Compared with the traditional sliding mode speed observer,the observer is very simple and eases to implement.Moreover,the observer reduces the ripple of the motor speed in high frequency range in an efficient way.To reduce the stator current ripple and improve the control performance of the torque and speed,the MPCC strategy is put forward,which can make PMSM MPCC system have better control performance,stronger robustness and good dynamic performance.The simulation results validate the feasibility and effectiveness of the proposed scheme.

Research on traffic flow forecasting model based on cusp catastrophe theory

This paper intends to describe the relationship between traffic parameters by using cusp catastrophe theory and to deduce highway capacity and corresponding speed forecasting value through suitable transformation of catastrophe model. The five properties of a catastrophe system are outlined briefly, and then the data collected on freeways of Zhujiang River Delta, Guangdong province, China are examined to ascertain whether they exhibit qualitative properties and attributes of the catastrophe model. The forecasting value of speed and capacity for freeway segments are given based on the catastrophe model. Furthermore, speed-flow curve on freeway is drawn by plotting out congested and uncongested traffic flow and the capacity value for the same freeway segment is also obtained from speed-flow curve to test the feasibility of the application of cusp catastrophe theory in traffic flow analysis. The calculating results of catastrophe model coincide with those of traditional traffic flow models regressed from field observed data, which indicates that the deficiency of traditional analysis of relationship between speed, flow and occupancy in two-dimension can be compensated by analysis of the relationship among speed, flow and occupancy based on catastrophe model in three-dimension. Finally, the prospects and problems of its application in traffic flow research in China are discussed.

Regression model for daily passenger volume of high-speed railway line under capacity constraint

A non-linear regression model is proposed to forecast the aggregated passenger volume of Beijing-Shanghai high-speed railway(HSR) line in China. Train services and temporal features of passenger volume are studied to have a prior knowledge about this high-speed railway line. Then, based on a theoretical curve that depicts the relationship among passenger demand, transportation capacity and passenger volume, a non-linear regression model is established with consideration of the effect of capacity constraint. Through experiments, it is found that the proposed model can perform better in both forecasting accuracy and stability compared with linear regression models and back-propagation neural networks. In addition to the forecasting ability, with a definite formation, the proposed model can be further used to forecast the effects of train planning policies.

A prediction method of operation trend for large axial-flow fan based on vibration-electric information fusion

As the critical equipment,large axial-flow fan(LAF)is used widely in highway tunnels for ventilating.Note that any malfunction of LAF can cause severe consequences for traffic.Specifically,fault deterioration is suppressed tremendously when an abnormal state is detected in the stage of early fault.Thus,the monitoring of the early fault characteristics is very difficult because of the low signal amplitude and system disturbance(or noise).In order to overcome this problem,a novel early fault judgment method to predict the operation trend is proposed in this paper.The vibration-electric information fusion,the support vector machine(SVM)with particle swarm optimization(PSO),and the cross-validation(CV)for predicting LAF operation states are proposed and discussed.Finally,the results of the experimental study verify that the performance of the proposed method is superior to that of the contrast models.

Novel predictive model for metallic structure corrosion status in presence of stray current in DC mass transit systems

The novel method to analyze metallic structure corrosion status was proposed in the presence of stray current in DC mass transit systems. Firstly, the characteristic parameter and the influence parameters for the corrosion status were determined. Secondly, an experimental system was established for simulating the corrosion process within the stray current interference. Then, a predictive model for the corrosion status was built, using a support vector machine(SVM) method and experimental data. The data were divided into two sets, including training set and testing set. The training set was used to generate the SVM model and the testing set was used to evaluate the predictive performance of the SVM model. The results show that the relationship between the characteristic parameter and the influence parameters is nonlinear and the SVM model is suitable for predicting the corrosion status.

Dynamics and Predictive Control of Gas Phase Propylene Polymerization in Fluidized Bed Reactors

A two-phase dynamic model, describing gas phase propylene polymerization in a fluidized bed reactor, was used to explore the dynamic behavior and process control of the polypropylene production rate and reactor temperature. The open loop analysis revealed the nonlinear behavior of the polypropylene fluidized bed reactor, jus- tifying the use of an advanced control algorithm for efficient control of the process variables. In this case, a central- ized model predictive control （MPC） technique was implemented to control the polypropylene production rate and reactor temperature by manipulating the catalyst feed rate and cooling water flow rate respectively. The corre- sponding MPC controller was able to track changes in the setpoint smoothly for the reactor temperature and pro- duction rate while the setpoint tracking of the conventional proportional-integral （PI） controller was oscillatory with overshoots and obvious interaction between the reactor temperature and production rate loops. The MPC was able to produce controller moves which not only were well within the specified input constraints for both control vari- ables, but also non-aggressive and sufficiently smooth for practical implementations. Furthermore, the closed loop dynamic simulations indicated that the speed of rejecting the process disturbances for the MPC controller were also acceotable for both controlled variables.