Neural network controller for a permanent magnet generator applied in a wind energy conversion system

In this paper a neural network controller for achieving maximum power tracking as well as output voltage regulation, for a wind energy conversion system (WECS) employing a permanent magnet synchronous generator, is proposed. The permanent magnet generator (PMG) supplies a DC load via a bridge rectifier and two buck–boost converters. Adjusting the switching frequency of the first buck–boost converter achieves maximum power tracking. Adjusting the switching frequency of the second buck–boost converter allows output voltage regulation. The on-times of the switching devices of the two converters are supplied by the developed neural network (NN). The effect of sudden changes in wind speed, and/or in reference voltage on the performance of the NN controller are explored. Simulation results showed the possibility of achieving maximum power tracking and output voltage regulation simultaneously with the developed NN controller. The results proved also the fast response and robustness of the proposed control system.     

Numerical schemes for defining reference stress in components containing defects

In R6 and R5, the standards of integrity assessments of components containing defects established by the former Central Electricity Generating Board (CEGB) in the UK, the basic procedure in calculating the crack driving force parameters, J-integration and C*-integration is to define the reference stress, σ_ref, on the defective section. Three methods of directly defining the reference stress, σ_ref [multiple steady-state creep analysis (MSCA) method, creep module-modified iteration (CMMI) method and elastic module-modified iteration (EMMI) method], are suggested in this paper, differing from the common methods in which the limit loads of the components are first calculated and then the σ_ref is convertibly determined. The basic conceptions and calculational steps of these methods are discussed and compared in the context, and the validity is demonstrated by several examples.     

Neural network approach for food temperature prediction during solar drying

In the present study, the application of artificial neural network (ANN) for prediction of temperature variation of food product during solar drying is investigated. The important climatic variables namely, solar radiation intensity and ambient air temperature are considered as the input parameters for ANN modeling. Experimental data on potato cylinders and slices obtained with mixed mode solar dryer for 9 typical days of different months of the year were used for training and testing the neural network. A methodology is proposed for development of optimal neural network. Results of analysis reveal that the network with 4 neurons and logsig transfer function and trainrp back propagation algorithm is the most appropriate approach for both potato cylinders and slices based on minimum measures of error. In order to test the worthiness of ANN model for prediction of food temperature variation, the analytical heat diffusion model with appropriate boundary conditions and statistical model are also proposed. Based on error analysis results, the prediction capability of ANN model is found to be the best of all the prediction models investigated, irrespective of food sample geometry.     

A finite-volume approach to the stress analysis of pressurized axisymmetric structures

A finite-volume procedure for determining the displacement fields and elastic stress distributions within structures that have axisymmetric geometries is presented. The production of a system of linear algebraic equations from the equilibrium equations of the cells that are used to represent the structure is described in detail. An iterative technique is employed to solve these equations and provide the displacement field from which the strain and stress fields can be found. The procedure is used to model two axisymmetric test problems and the stress distribution predicted by the formulation using cell meshes of increasing refinement is compared to analytical solutions and the variations predicted by various finite element procedures.     

A study on the reliability assessment methodology for pressure piping containing circumferential defects: I: computation method of failure probability of welded joint containing circumferential defects

When the reliability of a pressure piping needs to be assessed, every welded joint containing defects should first be analyzed. This paper describes a computational method of failure probability of a welded joint containing circumferential defects; the uncertainties of variables, relation of random variables and three possible failure modes of the pipe are considered. Then a new probability limit state model is set up, and the Monte-Carlo simulation method is employed to compute the failure probability of welded joint. In the end, a typical example of this approach is presented.     

Failure analysis and life prediction of a set of extra high pressure heat exchangers containing defects

Random sampling for inspection has been applied to an extra-high pressure heat-exchanger. It was found a large number of etching pits and cracks were in existence on the outside surface of the extra-high pressure inner pipe of the double-pipe heat exchanger. These threaten safety in production and constitute a hidden peril for the plant. Failure analysis and life prediction in this work provide proposals for guaranteeing safety. They have been adopted by the plant and have given a significant economic benefit.     

Stress concentration factors for an internally pressurized circular vessel containing a radial U-notch

This paper evaluates the stress concentration factors for an internally pressurized cylinder containing a radial U-notch along its length. This work studies the cases where the external to internal radius ratio (Ψ) is equal to 1.26, 1.52, 2.00, and 3.00 and the notch radius to internal radius ratio (Φ) is fixed and equal to 0.026. The U-notch depth varies from 0.1 to 0.6 of the wall thickness. Results are also presented for a fixed size semi-circular notch. Hoop stresses at the external wall are presented, showing regions where the stress matches the nominal one and the favourable places to install strain sensors. The finite element method is used to determine the stress concentration factors (K_t) for the above described situations and for a special case where a varying semi-circular notch is present with Ψ=3.00. This notch depth varies from 0.013 to 0.3 of the wall thickness. It is pointed out that even relatively small notches introduce large stress concentrations and disrupt the hoop stress distribution all over the cross section. Results are also compared to an example found in the literature for semi-circular notches and K_t curves for both cases present the same shape.     

Numerical prediction of entropy generation due to natural convection from a horizontal cylinder

Entropy generation due to natural convection has been calculated for three radii and a wide range of Rayleigh numbers for an isothermal cylinder. It was predominantly caused by conduction. The viscous contribution was negligible. Locations of high local entropy generation were identified for different configurations. The total entropy generation decreased with increasing cylinder size for a given value of the Rayleigh number.     

Inverse prediction of wall temperature distribution on a cylinder exposed to radiatively active flow

In this study, the unknown wall temperature profile of a cylinder was predicted by applying the inverse method. The temperature profile of the cylinder wall was predicted from the given temperature data at measurement points near the cylinder wall. The cylinder was assumed to represent a typical pipe in a bundle of heat exchange tubes operating in a high temperature system. Radiative heat transfer was incorporated as one of major heat transfer modes to consider a hot gas flow passing over the cylinder. The corresponding inverse problem was solved by minimizing an objective function by applying the iterative conjugate gradient method. A multi-block grid composed of three different blocks was used for better computational accuracy and convenience in locating the measurement points. A new method, which could be applicable to non-symmetric geometry, was adopted to solve the adjoint equation. In this study, the effects of number and location of the measurement points were numerically investigated. When the measurement points were too close to the cylinder, the predicted temperature profile exhibited larger fluctuations. The results have also shown that an appropriate number of measurement points were required to improve the prediction of the boundary temperature profile.     

Calculation of stress intensity factors for semielliptical cracks in a thick-wall cylinder

Weight functions for the surface and the deepest point of an internal semielliptical crack in a thick-wall cylinder were derived from a general weight function and two reference stress intensity factors. For several linear and nonlinear crack face stress fields, the weight functions were validated against finite element data. Stress intensity factors were also calculated for the Lamé through the thickness stress distribution induced by internal pressure. The weight functions appear to be particularly suitable for fatigue and fracture analysis of surface semielliptical cracks in complex stress fields. All stress intensity factor expressions given in the paper are valid for cylinders with an inner radius to wall thickness ratio, R_i/t = 4.     

质子交换膜燃料电池膜电极组件温度分布的神经网络预测模型

质子交换膜燃料电池膜电极组件表面的温度分布会影响质子交换膜燃料电池的性能、寿命和可靠性.为探究质子交换膜燃料电池传热规律,本文提出了一种基于神经网络的质子交换膜燃料电池膜电极组件温度分布的预测模型.本研究选取径向基函数神经网络(RBF)和广义回归神经网络(GRNN)两种神经网络,以电流密度、温度点的位置作为网络输入,不同位置的温度作为网络输出,对平行流道质子交换膜燃料电池、蛇形流道质子交换膜燃料电池分别建立了神经网络预测模型.结果显示,RBF神经网络预测的均方根误差平均为0.464、平均绝对百分误差为1.179％,GRNN神经网络预测的均方根误差平均为0.7155、平均绝对百分误差为2.27％;相较于GRNN神经网络,RBF神经网络精度更高;基于RBF神经网络的平行流道质子交换膜燃料电池膜电极组件温度分布预测模型预测值与96％的实验值的相对误差在5％以内.基于RBF神经网络的蛇形流道质子交换膜燃料电池膜电极组件温度分布预测模型预测值与95％的实验值的相对误差在5％以内.     

A simple procedure to evaluate the stress intensity factor in internally pressurized cylinders

The evaluation of stress intensity factors in internally pressurized cylinders, with both surface and sub-surface flaws, is examined. The method of analysis is based on the equivalent linear representation of the circumferential stress distribution in accordance with ASME rules, the non-linear hoop stress distribution then being conservatively approximated by the membrane and bending stresses. The stress intensity factor for an elliptical crack embedded in an elastic solid and subjected to internal pressure is considered for two conditions of load (tension and bending) and the effects are added.The results are presented in non-dimensional form to evaluate the effect on stress intensity factor of the various parameters (outside and inside radius, crack position, cylinder thickness, form of ellipse).     

Prediction of the failure pressure for complex corrosion defects

A new method for predicting the failure pressure of corrosion defects in pipelines has been developed. The failure pressure of a plain pipe represents an upper limit for the failure pressure of a pipe with a corrosion defect. The failure pressure of a uniform depth, infinitely long groove, where the depth is equal to the maximum depth of the corrosion defect, represents a lower limit for the failure pressure of a pipe with a natural corrosion defect. The predicted failure pressure can be calculated from these limits using the weighted depth difference (WDD) method, which accounts for the defect geometry and any interaction with adjacent defects. The WDD method has been validated using the results of 40 burst tests of pipe sections containing real corrosion defects. The results indicate that this method provides more accurate burst pressure predictions than the currently accepted corrosion defect assessment procedures.     

Numerical predictions for stable buoyant regimes within a square cavity containing a heated horizontal cylinder

Buoyancy-induced flow regimes are investigated numerically for the basic case of a horizontal cylinder centred into a long co-axial square-sectioned cavity. In the frame of the 2D assumption, the threshold for the occurrence of time-dependent behaviour is explored. Stable symmetric and non-symmetric steady-state solutions, as well as unsteady regimes are observed, depending on the Rayleigh number, Ra, and the aspect ratio of the cavity, d. Four d-values are considered (d = 0.2, 0.4, 0.6, and 0.8). Heat transfer results are correlated by a single equation covering the full subcritical region.     

Thermal stress intensity factors for a cracked cylinder under transient thermal loading

Transient thermal stress intensity factors are calculated at the deepest and the surface points of an internal axial semi-elliptical surface crack in a thick-walled cylinder under thermal shock. The method of calculation is based on the weight function and an effective procedure of integrating of the weight integrals. Time-dependent thermal boundary conditions are assumed to act on the inner surface of the pressurized cylinder. Exact analytical solutions are derived for the weight function, and then integrated to compute transient thermal stress intensity factors. In the special case of a steady state problem and also the pressurized cylinder, the results show to be in accordance with those cited in the literature.     

圆柱表面沟槽减阻的大涡模拟研究

为了降低钝体绕流的阻力,以圆柱为研究对象,考虑在圆柱表面布置沟槽来实现减阻。研究采用大涡模拟(Large Eddy Simulation, LES)方法对雷诺数Re为3 900的圆柱绕流进行数值模拟,通过表面沟槽影响近壁区流动来达到流动控制的效果。分析了V型和L型两种沟槽分别以横向和纵向形式布置时对圆柱阻力系数的影响,通过对比粗糙圆柱与光滑圆柱的壁面摩擦力系数、流动分离点及近壁区流场结构,研究了表面沟槽对于钝体流动的减阻机理。数值模拟结果表明：表面布置特定深度与间距的沟槽后,粗糙圆柱的阻力系数与光滑圆柱相比最多可降低12.5%,边界层分离点也更延后;流体在流经沟槽表面时,会在沟槽底部形成稳定的旋转小涡,减少了壁面摩擦阻力。     

Lattice Boltzmann method applied to the laminar natural convection in an enclosure with a heat-generating cylinder conducting body

In this paper, a numerical investigation of laminar convective flows in a differentially heated, square enclosure with a heat-conducting cylinder at its center, is carried out. The flow and the temperature are computed using respectively the lattice Boltzmann equation and finite-difference with suitable coupling to take natural convection into account. The investigation is performed for $\mathit{Pr}=0.71$, Rayleigh numbers of $\mathit{Ra}={10}^3\text{–}{10}^6$ and temperature-difference ratio of $\mathrm{\Delta }{T}^1=0\text{–}50$. The average hot and cold walls Nusselt numbers, the flow and temperature fields are presented and discussed. For a constant Ra, the average Nusselt number at the hot and cold walls (${\mathit{Nu}}_h$ and ${\mathit{Nu}}_c$) vary linearly with $\mathrm{\Delta }{T}^1$: ${\mathit{Nu}}_h$ decreases with $\mathrm{\Delta }{T}^1$ while ${\mathit{Nu}}_c$ increases with $\mathrm{\Delta }{T}^1$.     

Application of artificial neural networks for prediction of natural convection from a heated horizontal cylinder

A generalized neural network analysis for natural convection heat transfer from a horizontal cylinder is developed in this paper. Cylinder diameter, cylinder surface temperature and ambient temperature are selected as the input parameters, while the Nusselt number as the output. A three-layer network is used for predicting the Nusselt number. The number of the neurons in the hidden layer was determined by a trial and error process together with cross-validation of the experimental data evaluating the performance of the network and standard sensitivity analysis. The trained network gives the best values over the correlations with less than 2.5% mean relative error. The experimental data of the average Nusselt number over the horizontal cylinders having different diameters of 4.8 mm–9.45 mm are from Atay?lmaz and Teke [1]. The results from the trained network were compared with the proposed correlation for the average Nusselt number over the cylinder and it is shown that the results are in satisfactory agreement. The Nusselt numbers obtained from the experimental study were seen to be consistent by ± 20% with the well known correlations for natural convection heat transfer from horizontal cylinder developed by Morgan [2], Fand and Brucker [3], and Churchill and Chu [4]. Moreover it is seen that that results from the trained network show absolute agreement with the experimental data in ± 5% deviation band better than the correlations given by Morgan [2], Fand and Brucker [3], and Churchill and Chu [4].     

Effects of geometry and inconstant mass flow rate on temperatures within a pressurized hydrogen cylinder during refueling

Higher refueling rate leads to higher temperature rise within the cylinder. Excessive temperature should be avoided during the refueling progress. In this paper, we studied the effective methods to control the temperature rise by simulations based on the Computational Fluid Dynamics (CFD). Cylinders of different length to diameter ratios and different inlet diameters were simulated. We found that smaller radio of length to diameter can boost for temperature control and temperature distribution. Larger inlet diameter can restrain temperature rise. Comparing the simulation results with constant, increasing and decreasing mass flow rate, the refueling with increasing flow rate obtains the lowest temperature rise.     

Neural network model for a commercial PEM fuel cell system

Performance prediction of a commercial proton exchange membrane (PEM) fuel cell system by using artificial neural networks (ANNs) is investigated. Two artificial neural networks including the back-propagation (BP) and radial basis function (RBF) networks are constructed, tested and compared. Experimental data as well as preprocess data are utilized to determine the accuracy and speed of several prediction algorithms. The performance of the BP network is investigated by varying error goals, number of neurons, number of layers and training algorithms. The prediction performance of RBF network is also presented. The simulation results have shown that both the BP and RBF networks can successfully predict the stack voltage and current of a commercial PEM fuel cell system. Speed and accuracy of the prediction algorithms are quite satisfactory for the real-time control of this particular application.