Influence of the design parameters on the stress state of saddle-supported pipelines: an artificial neural network approach

A neural network methodology is herein applied to numerically treat the sensitivity analysis problem of above-ground pipelines under static loading by taking into account the possibility of development of uplifting phenomena at the pipe–saddle interfaces. Assuming classical frictionless unilateral contact to mathematically describe the pipeline support conditions coupled by an appropriate finite element scheme, the discrete problem is put in the form of an inequality constrained quadratic optimization problem with respect to either displacements or stresses. In order to investigate the structural response and the stress states of the above-ground pipeline at hand with respect to the variation of critical design parameters which are the pipe thickness and the support conditions, the sensitivity analysis problem is formulated as a quadratic programming problem with the design parameters appearing in the quadratic term. The feasibility of using appropriately designed neural networks to model the complicated nonlinear relationship between the several input parameters associated with above-ground pipelines and their support conditions is thus demonstrated.     

经济负荷分配的Hopfield神经网络计算

介绍了Hopfield神经网络(HNN)原理及其在机组经济负荷分配(EconomicLoadDispatch,ELD)中的应用。首先将ELD问题映射到Hopfield神经网络模型,然后利用HNN的动力特性搜索最优分配。仿真结果与分段结构优化方法和模拟退火(SimulatedAnnealing,SA)方法进行比较,表明HNN方法能找到近乎全局最优解,可有效地解决经济负荷分配问题。且易于在计算机上实现,有实际应用价值。     

Modelling a non-stationary single tube heat exchanger using multiple coupled local neural networks

This paper presents the application of an online identification neural technique to a single tube heat exchanger with a constant outer surface temperature. To show the feasibility of such an identification, the response to a sequence of random temperatures at the inlet of the inner fluid is studied. In the first part, the numerical solution is given, showing that the model cannot be a first order model. Then the principles of the neural technique are presented. The standard neural architecture, which normally calculates the output of the system directly from the input, is modified. A large number of local identical networks are used, each of them modelling an elementary module. It is shown that the neural model determined from the study of the first local network is representative of all the local networks (using the actual input data). At last it is shown that, when the networks are coupled, the output of the last network is in good agreement with the values obtained by the numerical model, but in a greatly reduced time.     

基于LS-SVM的基坑变形时间序列预测模型

针对神经网络用于基坑变形预测存在结构难确定、训练易陷入局部最优及易过学习等问题,构造滚动时间窗.,以已有的实测时间序列为样本,利用最小二乘支持向量机(LS-SVM)建立基坑预测模型,应用网格搜索算法优化模型参数,连续滚动地多步预测基坑变形.实例结果表明,该模型预测效果优于BP神经网络,具有所需数据少、推广能力强等优点.     

基于神经网络结合遗传算法的Jiles-Atherton磁滞模型参数计算

根据改进的Jiles-Atherton磁滞数学模型,提出了一种实用的磁滞模型参数的提取方法,即基于神经网络结合遗传算法的方法,通过神经网络训练拟合寻优函数,以及遗传算法极值寻优,而得到Jiles-Atherton磁滞模型的5个常规参数。计算证明,应用本文方法的计算参数得到的磁滞回线与实验磁滞回线相吻合。     

A novel NN based rotor flux MRAS to overcome low speed problems for rotor resistance estimation in vector controlled IM drives

This paper presents a new neural network based model reference adaptive system (MRAS) to solve low speed problems for estimating rotor resistance in vector control of induction motor (IM). The MRAS using rotor flux as the state variable with a two layer online trained neural network rotor flux estimator as the adaptive model (FLUX-MRAS) for rotor resistance estimation is popularly used in vector control. In this scheme, the reference model used is the flux estimator using voltage model equations. The voltage model encounters major drawbacks at low speeds, namely, integrator drift and stator resistance variation problems. These lead to a significant error in the estimation of rotor resistance at low speed. To address these problems, an offline trained NN with data incorporating stator resistance variation is proposed to estimate flux, and used instead of the voltage model. The offline trained NN, modeled using the cascade neural network, is used as a reference model instead of the voltage model to form a new scheme named as “NN-FLUXMRAS.” The NN-FLUX-MRAS uses two neural networks, namely, offline trained NN as the reference model and online trained NN as the adaptive model. The performance of the novel NN-FLUX-MRAS is compared with the FLUX-MRAS for low speed problems in terms of integral square error (ISE), integral time square error (ITSE), integral absolute error (IAE) and integral time absolute error (ITAE). The proposed NN-FLUX-MRAS is shown to overcome the low speed problems in Matlab simulation.     

Usage of neural network for coupled parameter and function specification inverse heat conduction problem

A new method for complicated inverse problems is presented in this paper. The inverse heat conduction task treated here looks simultaneously for boundary conditions and time constant of a temperature sensor on the basis of the knowledge of temperature readings from that sensor. A “non-classical” method based on neural network is used for that problem. The neural network belongs to a group of artificial intelligence methods. An example of numerical tests is included.     

基于改进遗传算法优化的小波神经网络短期负荷预测模型

针对传统小波神经网络在电力系统短期负荷预测中存在预测结果的精确度依赖初始网络参数的问题,提出了一种基于改进遗传算法优化的小波神经网络短期负荷预测模型。为了保证神经网络在训练过程中,各个层的权值和阈值按最优方向变化,将遗传算法引入小波神经网络,利用遗传算法寻优能力指导权值和阈值进行优化。将概率分布策略用于遗传算法的种群交叉和变异过程,解决遗传算法在中后期搜索精度差,收敛速度慢等问题。应用结果表明,与基本的小波神经网络的预测模型相比,在只考虑短期负荷历史数据的情况下,通过均方根误差计算比较,基于改进遗传算法优化的小波神经网络短期负荷预测模型具有更高的预测精度。     

High performance drive of DC brushless motors using neural network

In this paper, a multi-layer neural network (NN) architecture is proposed for the identification and control of DC brushless motors operating in a high performance drives environment. The NN in the proposed structure performs two functions. The first is to identify the nonlinear system dynamics at all times. Hence, detailed and elaborate models for the DC brushless machines are not needed. Furthermore, unknown nonlinear dynamics that are difficult to model such as load disturbances, system noise and parameter variations can be recognized by the trained neural network. The second function of the NN is to control the motor voltage so that the speed and position are made to follow pre-selected tracks (trajectories) at all times. The control action emulated by the NN is based on the indirect model reference adaptive control. A hardware laboratory setup is utilized to test and evaluate the proposed neural network structure. The paper shows, based on the laboratory test results, that the proposed neural network structure performance was good: the tracking accuracy was very high and the system robustness was quite evident even in the presence of random and severe disturbances     

多测点多方向BP网络模型在大坝变形监测中的应用

鉴于大坝变形监测资料分析是大坝结构性态安全评价与预报的重要手段,针对单测点模型存在的缺点,建立了既考虑坝体不同方向的位移又考虑空间多个测点分布的多测点多方向位移模型,并利用BP神经网络较强的非线性映射能力,直接选取了对大坝变形有较大影响的自变量因子,解决了在建立大坝多测点多方向传统模型时自变量因子数众多、计算工作量大等问题。实例应用结果表明,多测点多方向BP网络模型可反映大坝变形的分布及变化规律,可见采用BP神经网络建立大坝多测点多方向变形监测模型具有可行性和有效性。     

基于混合神经网络深度学习的短期负荷预测

针对能源互联网环境下用电用户数据量大、多维度这一特点,提出了一种混合神经网络深度学习的短期电力负荷预测方法。首先,考虑常见的电力系统负荷的影响因素,建立多维数据库,并进行偏相关分析,排除其他变量干扰;其次,将LSTM、GRU两种神经网络作为前端神经网络对多维数据库中数据进行处理;最后,采用随机概率剔除与Adam训练优化函数改进的BP神经网络作为末端神经网络,建立负荷预测模型。通过算例仿真对本文方法与传统BP神经网络、LSTM神经网络、GRU神经网络进行了对比,验证了所提方法的有效性。     

基于长短期记忆神经网络的地表太阳辐照度预测

针对地表太阳辐照度(GHI)短期预测问题,提出一种基于长短期记忆神经网络的短期太阳辐照度预测模型。采用递归结构的训练样本,以保证训练样本内部的时间耦合性。为验证所提模型预测GHI的有效性,采用算例与传统人工神经网络模型预测结果进行对比分析。结果表明:基于长短期记忆神经网络预测模型将均方误差降低88.48%,表明所建模型更适用于GHI预测。     

Online estimation of multicomponent heat flux using a system identification technique

In this work system identification techniques are used to map the two-dimensional heat flux into the temperatures through a linear model supported by theoretical and numerical results. The basis of this analysis is a discrete version of the Burggraf Method saying a single component heat flux is a linear combination of the temperatures around the time of its occurrence. Taking the same approach, a linear model (i.e. a linear artificial neural network (ANN)) is employed to estimate a multicomponent heat flux as a linear function of the temperatures. A known heat flux is imposed to the direct model, then the history of heat flux-temperature data are fit to the linear mathematical model (i.e. a linear ANN) using system identification techniques. The achieved model estimates the heat flux based on a series of past and future temperatures and the estimated heat flux components are in a good agreement with the exact ones. Finally, the effect of some important factors on the results is investigated. The proposed solution to inverse heat conduction problems does not need thermophysical and geometrical parameters of the system and is robust against noises. It merely needs some series of heat flux-temperature data from solution of a reliable direct numerical model or experiment.     

An MILP model for the reformation of natural gas pipeline networks with hydrogen injection

A mixed integer linear programming (MILP) model is proposed for the reformation of natural gas pipelines. The model is based on the topology of existing pipelines, the load and pressure at each node and the design factors of the region and minimizes the annual substitution depreciation cost of pipelines, the annual construction depreciation cost of compressor stations and the operating cost of existing compressor stations. Considering the nonlinear pressure drop equations, the model is linearized by a piecewise method and solved by the Gurobi optimizer. Two cases of natural gas pipeline networks with hydrogen injection are presented. Several adjustments are applied to the original natural gas pipeline network to ensure that our design scheme can satisfy the safety and economic requirements of gas transportation. Thus, this work is likely to serve as a decision-support tool for the reformation of pipeline networks with hydrogen injection.     

Heat exchanger design considering variable overall heat transfer coefficient: An artificial neural network approach

This study presents an artificial neural network approach in combination with numerical methods to calculate the heat transfer area assuming a nonlinear variation of the global heat transfer coefficient as a consequence of the thermophysical properties of the fluids, the geometry of the surfaces, and other factors. The development of the article is presented in two applications. The first application takes up the database described by Allan P. Colburn, four possibilities are proposed using functions from the field of artificial neural networks to create several approaches. The second application is presented to verify the goodness of the proposed methodology, the artificial neural network model is applied in an experimental data set of double-pipe vertical heat exchangers, the comparison between the calculated and experimental heat transfer area shows a relative percentage error smaller than 2.8%. The results in the applications are evidence of the competitiveness of the artificial neural network for the prediction of the heat transfer area considering a variable overall heat transfer coefficient.     

A multi-layer feed forward neural network model for accurate prediction of flue gas sulfuric acid dew points in process industries

Acidic combustion gases can cause rapid corrosion when they condense on pollution control or energy recovery equipments. Since the potential of sulfuric acid condensation from flue gases is of considerable economic significance, a multi-layer feed forward artificial neural network has been presented for accurate prediction of the flue gas sulfuric acid dew points to mitigate the corrosion problems in process and power plants. According to the network’s training, validation and testing results, a three layer neural network with four neurons in the hidden layer is selected as the best architecture for accurate prediction of sulfuric acid dew points. The presented model is very accurate and reliable for predicting the acid dew points over wide ranges of sulfur trioxide and water vapor concentrations. Comparison of the suggested neural network model with the most important existing correlations shows that the proposed neuromorphic model outperforms the other alternatives both in accuracy and generality. The predicted flue gas sulfuric acid dew points are in excellent agreement with experimental data suggesting the accuracy of the proposed neural network model for predicting the sulfuric acid condensation in stacks, pollution control devices, economizers and flue gas recovery systems in process industries.     

Artificial neural network modelling and experimental verification of the operating current of mono-crystalline photovoltaic modules

This article presents the artificial neural network modelling of the operating current of a 120 Wp of mono-crystalline photovoltaic module. As an alternative method to analytical modelling approaches, this study uses the advantages of neural networks such as no required knowledge of internal system parameters, less computational effort and a compact solution for multivariable problems. Generalised regression neural network model is used in the present article to predict the operating current of the photovoltaic module. To show its merit, the current predicted from the artificial neural network modelling is compared to that from the analytical model. The five-parameter analytical model is drawn from the equivalent electrical circuit that includes light-generated current, diode reverse saturation current, and series and shunt resistances. The operating current predicted from both the neural and analytical models are compared to the measured current. Results have shown that the artificial neural network modelling provides a better prediction of the current than the five-parameter analytical model.     

基于神经网络预测控制的单元机组协调控制策略

利用BP神经网络的非线性映射能力对单元机组协调控制系统被控对象进行辨识,从而建立其动态模型;在这一模型的基础上对协调控制系统中的控制器参数优化进行研究,提出基于神经网络预测控制的协调控制策略.该方法很好地解决了协调控制系统中强耦合、非线性等问题.仿真实验表明该系统的跟踪速度加快、调节精度提高、并且具有较好的抗干扰性.图6参7     

蒸汽管网模拟优化技术应用

洛阳分公司蒸汽管网包括10MPa、3．5MPa、1.0MPa和0．3MPa共4个等级．其中3．5MPa和1．0MPa蒸汽管网是主要管网。两套管网均存在供汽结构不合理,管段散热损失大,管网保温材料老化及破损严重,管段外表面温发较高（在50℃以上,局部管段超过80℃）等问题。为此．根据3．5MPa和1．0MPa蒸汽管网平衡数据．作出流量平衡表．利用蒸汽管网模拟分析软件（SNAMER）建立蒸汽管网模型并进行离线模拟分析。根据模拟分析结果,提出增设一条蒸汽跨线,以提高1号汽轮机发电机入口压力和汽轮机输出功率;将热电站至化纤装置3．5MPa蒸汽母管管径改为DN500,以减少压降;将部分管线保温材料改为硅酸铝镁纤维,保护层材质改为镀锌铝皮．以减少散热损火。模拟结果显示,实施上述措施后,1号汽轮机发电机入口压力约提高0．3MPa,在耗汽量不变的情况下,输出功率可提高3％：3．5MPa年1．0MPa蒸汽管网总散热损失将分别下降24％和31％;若对部分管线进行改造,每年将节约费用500万元。     

基于Elman神经网络的短期风电功率预测

为提高风电场输出功率预测精度,提出一种动态基于神经网络的功率预测方法。根据实际运行的风电场相关风速、相关风向和风电功率的历史数据,建立了基于Elman神经元网络的短期风电功率预测模型。运用多层Elman神经网络模型对西北某风电场实际1h和24h的风电输出功率预测,与BP神经网络模型对比,经仿真分析证明前者具有预测精度高的特点,三隐含层Elman神经网络模型预测效果最佳。这表明利用Elman回归神经网络建模对风电功率进行预测是可行的,能有效提高功率预测精度。