基于支持向量机的风电机组运行工况分类方法

采用多个运行参数来描述风电机组复杂多变的与振动相关的运行工况,提出一种基于支持向量机的分类方法划分风电机组的运行工况。我国北方某风电场实际运行数据的计算结果表明:与传统的单参数分类方法相比,基于支持向量机的运行工况分类方法能明显降低振动监测过程中的误报警率。     

基于支持向量机和果蝇优化算法的循环流化床锅炉NO_x排放特性研究

为了控制循环流化床(CFB)锅炉的NOx排放量,以某热电厂300MW CFB锅炉测试数据为样本,应用支持向量机(SVM)建立NOx排放特性预测模型.针对SVM回归预测需要人为确定相关参数的不足,应用果蝇优化算法(FOA)优化SVM参数,采用不同工况下的样本数据检验FOA-SVM模型的预测性能,并将该模型的预测结果与粒子群算法(PSO)、遗传算法(GA)和万有引力搜索算法(GSA)优化的SVM模型预测结果进行了比较.结果表明:FOA-SVM模型的泛化能力较强,预测精度较高,训练时间较短,可以相对快速、准确地预测NOx排放质量浓度.     

基于果蝇优化算法的GRNN水电机组状态趋势预测

针对水电机组振动的非线性、非平稳特性,提出了一种基于果蝇优化算法（FOA）的广义回归神经网络（GRNN）模型（FOAGRNN ）,实现了GRNN分布参数的优化选择,并对四川省新政航电工程3台机组5个不同部位的振动序列峰峰值进行了预测,与BP神经网络预测结果的均方误差（MSE）对比结果表明,FOAGRNN预测精度较高。     

基于布谷鸟算法的风电机组装配序列优化

装配序列规划（ASP）是风电机组制造的重要技术。ASP是一个组合优化问题,风电机组最优装配序列的搜索空间和计算量均很大。提出基于布谷鸟算法的风电机组装配序列优化方法。首先,从三维装配体模型中提取多种装配约束信息并表示成装配约束矩阵,以降低最优装配序列的搜索空间;继而构造装配序列目标函数,建立装配序列规划模型,方便算法计算出最优装配序列;对离散布谷鸟算法（DCA）改进,求解ASP模型,获得风电机组产品的最优装配序列。最后通过实验验证了ASP模型的有效性和DCA的优良性能。     

基于改进粒子群优化支持向量机的汽轮机组故障诊断

基于支持向量机（SVM）在核函数参数和惩罚因子人为选取的盲目性以及传统粒子群算法（PSO）后期易陷于局部最小值的不足,提出了一种改进的粒子群算法（MPSO）,建立了汽轮机组振动故障诊断模型并且利用故障数据进行了模式识别.结果表明：模型能够对SVM相关参数自动寻优,并且能达到较为理想的全局最优解;与PSO-SVM和GA-SVM算法相比,MPSO-SVM算法在收敛速度和准确率方面都有所提高.     

基于人工蜂群算法优化支持向量机的燃气轮机故障诊断

针对基于SVM(支持向量机)的故障诊断方法中支持向量机的参数难以选取导致诊断结果较差的问题,采用ABC(人工蜂群算法)对支持向量机的惩罚因子C和核函数参数σ进行优化;并构建了ABC-SVM(人工蜂群优化支持向量机)对燃机涡轮叶片故障进行诊断。诊断实例表明,该方法诊断准确率达到96. 43%,具有很好的诊断效果,为燃气轮机故障诊断提供了一种新的方法,具有实际应用价值。     

基于SCADA参数关系的风电机组部件重要度分析

在风电机组部件重要度分析中,通常将机组性能状态简单划分为"正常运行"和"停机"两种,忽视其他中间状态,且大多依赖专家打分,主观性较强。为解决上述问题,文章提出一种基于SCADA参数关系的风电机组部件重要度分析方法。首先,对原始数据进行预处理,基于"风速-功率"多项式模型评估整机性能;然后,采用高斯过程回归拟合"风速-部件性能参数"基准关系,以偏离基准的程度来评估部件性能;最后,提出森林优化-K近邻回归方法,建立部件性能与整机性能的特征加权回归模型,根据最优特征权重确定部件重要度。基于公开数据集的实验证明了所提方法的有效性。     

基于果蝇优化算法的配电网状态估计

配电网状态估计的目的是根据获取的配电网的各种量测信息,估计配电网系统的运行状态,是配电管理系统（DMS）的重要核心功能之一。根据配电网特点,以电压幅值和相角为状态变量,以量测值和估计值误差最小为目标函数,以潮流方程和估计值的上、下限为约束条件,构建配电网状态估计的优化模型,采用果蝇优化算法对该模型进行求解。算例表明,基于视觉和嗅觉的觅食行为启发的果蝇优化算法可有效求解该优化模型,具有易于实现、控制参数少和计算精度高的特点,为配电网状态估计提供了新的途径。     

基于改进粒子群算法含双馈风电机组 配网无功优化研究

针对风电机组并网后对配电网无功补偿产生的影响,从双馈风电机组自身的有功、无功输出特性出发,基于场景概率的方法计算风力机组出力情况,以网损最小为目标函数寻求优化求解方法。利用改进粒子群算法来实现系统接入双馈风电机组后的无功优化,在Matlab 2013b软件中构造IEEE33节点模型并利用该算法求解。结果表明,双馈风电机组在参与系统无功优化时具有良好的性能,验证了该改进算法的有效性     

基于PSO优化LS-SVM的短期风速预测

提出了一种基于粒子群(PSO)算法优化最小二乘支持向量机(LS-SVM)的风电场风速预测方法。以相关性较高的历史风速序列作为输入,建立预测模型,并用粒子群算法优化模型参数。在对未来1 h风速进行预测时,文章所提出的模型比最小二乘支持向量机模型及BP神经网络模型具有较高的预测精度和运算速度。算例结果表明,经粒子群优化的最小二乘支持向量机算法是进行短期风速预测的有效方法。     

基于CLSSVM的风电机组齿轮箱故障诊断

针对风电机组齿轮箱传统故障诊断方法以全局误诊断率最小化为目标,忽略了误分类型之间的差别的问题,提出基于代价敏感最小二乘支持向量机(Cost-sensitive Least Squares Support Vector Machine,CLSSVM)的风电机组齿轮箱故障诊断方法。该方法在最小二乘支持向量机原始最优化问题中二次损失函数中嵌入不同样本的误分类代价,建立以误分类代价最小化为目标的CLSSVM故障诊断模型,并同最小二乘支持向量机和代价敏感支持向量机比较。实验结果表明,该方法能提高误分类代价高的故障类样本的诊断正确率,具有代价敏感性,其训练速度也足以满足风电机组齿轮箱故障诊断实时性的需求。     

基于GA优化支持向量机的变压器故障诊断

针对变压器故障诊断中出现的多故障分类问题,为提高支持向量机的多故障分类的准确率,利用遗传算法(GA)对支持向量机的相关参数进行了优化。将利用遗传算法优化的支持向量机(GA-SVM)应用于变压器故障诊断中,并与利用粒子群算法优化的支持向量机(PSO-SVM)的识别结果进行比较。对比试验结果可以看出,GA-SVM算法能够更为有效地选择支持向量机的相关参数,在很大程度上提高了变压器多故障分类的准确性。     

基于ACFOA优化RBF的短期风电功率预测

为了提高短期风电输出功率预测的准确度,在分析研究基本预测方法的基础上,提出采用一种自适应混沌果蝇算法(ACFOA)优化RBF神经网络的预测方法。该方法中引入自适应混沌对果蝇算法的进化机制进行优化,并利用ACFOA算法改善RBF神经网络结构参数以提高网络的泛化能力,同时对某风电场的历史数据进行验证分析。仿真结果表明,相比于PSO-RBF预测方法,采用提出的预测模型能有效减少较大误差出现的频率,大幅度提高风电输出功率预测的准确度。     

基于改进共轭梯度优化BP神经网络的风电机组变桨距控制

根据共轭梯度算法和传统BP神经网络的变桨距控制器的原理,针对兆瓦级风电机组变桨距控制设计了一种改进共轭梯度优化BP神经网络的变桨距PID参数自整定控制器,此控制器采用改进共轭梯度法修正BP神经网络的权值和阈值,实现BP神经网络变桨距PID控制器的在线整定。在Matlab/Simulink中仿真,仿真结果表明,采用此变桨距控制器可以在额定风速之上快速响应,在相同风速状况下使发电机桨距角调节命令更加准确,风轮转速更加平稳,输出功率维持在额定功率附近,取得了很好的变桨距控制效果。     

Wind turbine generation performance monitoring with Jaya algorithm

Wind turbine (WT) power curves effectively reflect the generation performance of WTs and depict the relationship between the wind speed and the WT power output. This paper aims at developing an effective method for learning the intrinsic representations of WT power curves, which are robust to external environmental changes. Based on the obtained representations, WT generation performance is monitored. In the proposed approach, data of the supervisory control and data acquisition (SCADA) system is employed to derive the representations. Parametric models of WT power curves are developed using the two‐parameter and four‐parameter logic models. The parameters of these model are identified via Jaya algorithm. To detect the changes of WT power curve model parameters over different time, multivariate control charts are employed. The effectiveness of the proposed WT generation performance monitoring approach is validated based on SCADA data collected from real commercial WTs.     

基于改进粒子群算法的风力机选型应用研究

风力机的选型是风电场建设的重要内容,它对风电场建设造价、投产后的发电量以及运行维护成本等有直接影响。文章在给定风资源的情况下,综合考虑风电场的容量系数和实际发电量,以风力机性能指数作为选型的依据,针对采用常规方法进行风力机参数线性化求解的缺陷,采用智能化的改进粒子群算法对风力机参数进行寻优。与常规计算方法相比,该方法寻得的风力机性能指数更优。结合具体实例计算候选机型的风速加权标准差,选出最优风力机。该研究结果为风电场的风力机选型提供了一种有效可行的方法,具有一定的应用参考价值。     

Fault diagnosis for a wind turbine transmission system based on manifold learning and Shannon wavelet support vector machine

Fault diagnosis for wind turbine transmission systems is an important task for reducing their maintenance cost. However, the non-stationary dynamic operating conditions of wind turbines pose a challenge to fault diagnosis for wind turbine transmission systems. In this paper, a novel fault diagnosis method based on manifold learning and Shannon wavelet support vector machine is proposed for wind turbine transmission systems. Firstly, mixed-domain features are extracted to construct a high-dimensional feature set characterizing the properties of non-stationary vibration signals from wind turbine transmission systems. Moreover, an effective manifold learning algorithm with non-linear dimensionality reduction capability, orthogonal neighborhood preserving embedding (ONPE), is applied to compress the high-dimensional feature set into low-dimensional eigenvectors. Finally, the low-dimensional eigenvectors are inputted into a Shannon wavelet support vector machine (SWSVM) to recognize faults. The performance of the proposed method was proved by successful fault diagnosis application in a wind turbine's gearbox. The application results indicated that the proposed method improved the accuracy of fault diagnosis.     

Wind turbine condition monitoring based on three fitted performance curves

Based on SCADA data, this study aims at fitting three performance curves (PCs), power curve, pitch angle curve, and rotor speed curve, to accurately describe the normal behaviour of a wind turbine (WT) for performance monitoring and identification of anomalous signals. The fitting accuracy can be undesirably affected by erroneous SCADA data. Hence, outliers generated from raw SCADA data should be removed to mitigate the prediction inaccuracy, so various outlier detection (OD) approaches are compared in terms of area under the curve (AUC) and mean average precision (mAP). Among them, a novel unsupervised SVM-KNN model, integrated by support vector machine (SVM) and k nearest neighbour (KNN), is the optimum detector for PC refinements. Based on the refined data by the SVM-KNN detector, several common nonparametric regressors have largely improved their prediction accuracies on pitch angle and rotor speed curves from roughly 86% and 90.6%, respectively, (raw data) to both 99% (refined data). Noticeably, under the SVM-KNN refinement, the errors have been reduced by roughly five times and 10 times for pitch angle and rotor speed predictions, respectively. Ultimately, bootstrapped prediction interval is applied to conduct the uncertainty analysis of the optimal predictive regression model, reinforcing the performance monitoring and anomaly detection.     

Performance prediction of the PEMFCs based on gate recurrent unit network optimized by improved version of prairie dog optimization algorithm

In this article, the black box dynamic model is presented for forecasting the performance of the PEM (Proton-exchange membrane) fuel cell (FC). An optimized deep artificial neural network has been used to build the experimental nonlinear model of the polymer membrane FC series that functions with hydrogen and oxygen. This research investigates predictability for a gate recurrent unit (GRU) optimized by a modified Prairie Dog Optimizer in PEMFCs. The results obtained have been validated by applying a case study and then a comparison is conducted among the outcomes of the offered technique and 2 other published methods: modified relevance vector machine and Lattice Gated Recurrent Unit (LGRU). The voltage clearly changes significantly, as demonstrated by simulations, even though the FC is handled with a low starting temperature and current. Also, the voltage point distribution has become more concentrated when the current and temperature are high. In both the training and prediction phases, the MAPE is reduced to approximately 0.0043 and 0.0047, respectively, showing that the proposed GRU technique produces superior prediction results when the operational settings approach the optimum operating conditions. According to simulations, the proposed IPDO/GRU with a 0.004 root mean square has the least error, followed by the mRVM and GRU with 0.009 and 0.010 root mean square values. The outcomes show that using the offered procedure does provide the finest verification of the empirical data.