基于粗糙集与人工神经网络的变压器故障诊断

根据电力变压器故障诊断问题,提出了基于粗糙集与人工神经网络的变压器故障诊断模型,分析了该模型的实现步骤.采用Kohonen网络对连续属性值进行离散化,应用粗糙集理论对特征参数进行属性约简,并把约简结果生成规则作为BP网络的输入.仿真结果表明,把经过粗糙集理论预处理过的数据送入BP网络训练,提高了学习速度和故障诊断正确率,减少了训练时间.     

融合scSE模块的改进Mask R-CNN海洋锋检测方法

海洋锋是重要的中尺度海洋现象,具有数据量小、目标小、弱边缘等特性。针对实际检测任务中弱边缘、小目标海洋锋的检测精度低、错检及漏检率高等问题,融合scSE (spatial and channel Squeeze&Excitation)空间注意力模块构建了一种改进的Mask R-CNN海洋锋检测模型。该方法首先对Mask R-CNN骨干网络结构进行改进,采用scSE模块引导的ResNet-50网络作为特征提取网络,通过加权策略对图像通道和空间位置进行特征突出,提升网络对重要特征的提取能力;其次,针对海洋锋目标边缘定位不准确的问题,引入IoU boundary loss构建新的Mask损失函数,提高边界检测精度。最后,为验证方法的有效性,从训练数据和实验模型上,分别设计多组对比实验。实验结果表明,相比传统Mask R-CNN、YOLOv3神经网络及现有Mask R-CNN改进网络,本文方法对SST梯度影像数据集上的强、弱海洋锋检测效果最好,定位准确率（IoU,Intersection-over-union））及检测精度（mAP,Mean Average Precision）均达0....     

基于权重调整的BP神经网络在Nino区海温预报中的应用

传统BP神经网络在训练完之后,其权重是固定不变的,加上神经网络的样本的标准化处理,将使得网络不易描绘样本峰值.因此,本文考虑变权的方法,以调节训练后的BP网络权重,基于变权次数,建立不同网络模型,并利用不同网络输出值与相应实测值进行比较.结果表明:变权BP网络预报效果有较大提升,同时,降低了对因子相关性的要求.     

基于贝叶斯网的象山港网箱养殖水环境指标建模

贝叶斯网络具有强大的推理能力,能与先验知识和数据结合,进行定性和定量分析,提供了1条有效的处理预测问题的途径,本文首先介绍了贝叶斯网络基本理论及其特点,并讨论如何学习贝叶斯网络结构,然后由专家知识和给定数据,采用基于依赖分析的贝叶斯网络学习算法构造了海底网箱养殖水环境指标间的贝叶斯网结构模型.该模型能有效的表达网箱养殖环境各个指标之间的因果关系和影响程度,实验结果表明,试验数据显示准确性是92.3%,kappa指数是0.882.以上证明该方法是有效可行的,表明贝叶斯网络是一种很有前途的预测评价方法.     

利用Kmeans-RBF神经网络提高重力格网化数据精度

针对重力测量数据在格网化过程中精度会被降低的问题,顾及空间重力异常和地形的强相关性,提出了三维Kmeans-RBF神经网络方法,该方法利用神经网络的复杂非线性映射学习能力进行推估建模,并在模型训练和推估时加入地形数据作为物理控制。最后基于美国爱达荷州地区的实测重力数据进行验证,实验结果表明:该方法相对于二维Kmeans-RBF神经网络方法和直接进行拟合推估的Kriging方法,实验区内精度分别提高了24.85%和44.84%。     

一种基于人工免疫网络的遥感图像分类算法

提出了一种基于人工免疫网络的遥感图像分类算法。该算法通过借鉴生物免疫网络的分类和泛化能力,训练出能反映训练数据分布特性的网络细胞,然后使用这些网络细胞进行分类。实验结果表明,基于人工免疫网络的遥感图像分类算法具有较好的分类性能,其分类总精度、kappa系数均优于一些传统分类算法。     

无标注的含噪中文长篇幅语音文本的文语对齐研究

文语对齐技术是语音识别领域中的一项关键技术。传统文语对齐方法利用语音识别器将文语对齐问题转换成了文本与文本的对齐问题,但是该方法依赖于大量有标注数据训练的声学模型。本文提出一种利用开放识别引擎和基于有限状态自动机的语言模型来得到语音与文本一一对齐数据的算法,来摆脱对于大量标注数据的依赖。实验表明利用该算法得到语音文本数据的准确率为99%,可以用于识别器的训练。接着利用该部分数据训练一个面向要识别领域的声学模型,来对文本和语音进行迭代的,自适应的文语对齐。     

基于改进生成式对抗网络的水下图像增强

由于水介质的吸收和散射特性会导致雾化、低对比度、颜色退化等各种水下成像失真,严重影响了水下图像的后续利用。为了恢复清晰的水下图像,提出一种基于改进生成对抗网络的深度学习模型。借助图像质量评价技术,将生成的过程样本与高质量样本进行拟合,并将拟合得到的差值信息用于优化网络中的生成器。改进的生成式对抗网络有效改善了由真假训练逻辑带来的图像质量提升限制的问题。实验结果显示：该方法有效的恢复了水下图像的色彩,并改善了图像的清晰度和对比度;相比其他方法,提出的方法在SSIM、UCIQE和UIQM指标上分别提升了2.9%、6.2%和14.3%。     

缺失风浪数据补足的神经网络模型

基于波浪数据的完备性对于海岸海洋工程设计而言非常关键,详细阐述了风浪观测数据补足神经网络模型的建立方法,构建了两个网络模型,以已有观测资料为样本进行了验证.结果表明,两个网络的训练效果均很好,且单输出目标的分层模拟要优于多输出目标的单层模拟.表明了利用人工神经网络推导缺失波浪条件的可行性.     

智能船舶航行场景复杂度感知方法研究

为了解决智能船舶测试场景构建中的长尾问题并提高其测试效率,提出一种基于视觉图像的航行场景复杂度感知方法。本文从图像纹理特征分析出发,构建航行场景复杂度与特征指标之间的数学模型。首先,采用灰度共生矩阵对待测试图像信息进行特征提取,并利用能量、熵、对比度、逆差矩和相关性等多个参数组成特征向量。接着,提出利用集成学习AdaBoost网络模型进行船舶航行场景复杂度感知,即利用大量的图片对所提模型进行训练和学习,获得场景复杂度与各指标之间的非线性数据感知模型。通过在不同数据集上的不同方法进行对比,实验结果表明该感知模型能够真实的反应船舶航行场景的复杂程度,获得结果与人类视觉感知的结果基本一致,其对智能船舶自主航行场景设计与构建都具有的参考价值。     

基于BP神经网络的长江口北支河槽容积分析

根据实测水文及泥沙等资料,采用现在较成熟的且应用广泛的BP人工神经网络建立了北支0m以下河槽容积与大通流量、大通输沙量及北支分流比3个因子问的神经网络模型,网络结构为3.1-7-1,通过选择合适的参数,模型训练较好,预测结果与线性回归模型预测结果相近,说明BP神经网络模型能够广泛应用于河口水文等方面的预报.     

Prediction and assimilation of surf-zone processes using a Bayesian network: Part II: Inverse models

A Bayesian network model has been developed to simulate a relatively simple problem of wave propagation in the surf zone (detailed in Part I). Here, we demonstrate that this Bayesian model can provide both inverse modeling and data-assimilation solutions for predicting offshore wave heights and depth estimates given limited wave-height and depth information from an onshore location. The inverse method is extended to allow data assimilation using observational inputs that are not compatible with deterministic solutions of the problem. These inputs include sand bar positions (instead of bathymetry) and estimates of the intensity of wave breaking (instead of wave-height observations). Our results indicate that wave breaking information is essential to reduce prediction errors. In many practical situations, this information could be provided from a shore-based observer or from remote-sensing systems. We show that various combinations of the assimilated inputs significantly reduce the uncertainty in the estimates of water depths and wave heights in the model domain. Application of the Bayesian network model to new field data demonstrated significant predictive skill (R² = 0.7) for the inverse estimate of a month-long time series of offshore wave heights. The Bayesian inverse results include uncertainty estimates that were shown to be most accurate when given uncertainty in the inputs (e.g., depth and tuning parameters). Furthermore, the inverse modeling was extended to directly estimate tuning parameters associated with the underlying wave-process model. The inverse estimates of the model parameters not only showed an offshore wave height dependence consistent with results of previous studies but the uncertainty estimates of the tuning parameters also explain previously reported variations in the model parameters.     

System identification for nonlinear maneuvering of large tankers using artificial neural network

This paper deals with the application of nonparametric system identification to a nonlinear maneuvering model for large tankers using artificial neural network method. The three coupled maneuvering equations in this model for large tankers contain linear and nonlinear terms and instead of attempting to determine the parameters (i.e. hydrodynamic derivatives) associated with nonlinear terms, all nonlinear terms are clubbed together to form one unknown time function per equation which are sought to be represented by the neural network coefficients. The time series used in training the network are obtained from simulated data of zigzag maneuvers and the proposed method has been applied to these data. The neural network scheme adopted in this work has one middle or hidden layer of neurons and it employs the Levenberg–Marquardt algorithm. Using the best choices for the number of hidden layer neurons, length of training data, convergence tolerance etc., the performance of the proposed neural network model has been investigated and conclusions drawn.     

基于多种神经网络的风暴潮增水预测方法的比较分析

简要介绍了利用BP神经网络、小波神经网络、递归神经网络进行风暴潮增水值预测的原理。选取广东省珠江口以南的阳江站2017年风暴潮增水数据进行测试。结果表明,三种神经网络方法针对阳江地区风暴潮增水的预测均具有可靠性和实用性。以当前增水值为输入量的单因子模型更能反映真实风暴潮增水趋势,而从增水极值预测的准确性来看,以台风风力、气压、风向等相关参数为输入量的多因子模型优于单因子模型。BP神经网络更适用于多因子长时间预测,小波神经网络在单因子短时间预测上准确性更高,递归神经网络预测值与实测值相关性更强。在工程运用中,需根据地域时空特点、数据资料的丰富度与预测值评估指标选择合适的方法。     

基于决策树与二分分割算法的BP神经网络在赤潮等级预测中的应用研究

针对赤潮灾害等级预测难的现状,提出了一种基于C4.5决策树与二分分割算法优化的BP(反向传播)神经网络赤潮等级预测模型。该模型针对传统BP神经网络输入参数难以选择和隐含层节点数量难以确定的问题,通过决策树分类获取最优的属性组合,来解决输入参数难以选择的问题;通过"二分分割算法",来解决隐含层节点数难以确定的问题。实验结果表明,该模型在青岛近海海域赤潮灾害等级预测中,预测结果的均方根误差(RMSE)小于传统BP神经网络的预测误差,并且在网络训练时间上有所缩短,预测精度上有所提高,能够获得良好的预测结果,可为赤潮等级预测提供新的解决方法。     

时滞网络控制系统的动态输出鲁棒控制

针对网络控制系统中存在的不确定时滞,将网络控制系统模型转化为不确定系数的离散时间模型。利用Lya-punov稳定性理论和线性矩阵不等式技术,提出动态输出鲁棒控制器的设计方法,同时给出闭环系统鲁棒渐近稳定的充分条件。仿真例子验证了所提方法的有效性。     

The influence of seasonality on estimating return values of significant wave height

A time-dependent generalized extreme value (GEV) model for monthly significant wave heights maxima is developed. The model is applied to several 3-hour time series from the Spanish buoy network. Monthly maxima show a clear non-stationary behavior within a year, suggesting that the location, scale and shape parameters of the GEV distribution can be parameterized using harmonic functions. To avoid a possible over-parameterization, an automatic selection model, based on the Akaike Information Criterion, is carried out. Results show that the non-stationary behavior of monthly maxima significant wave height is adequately modeled, drastically increasing the significance of the parameters involved and reducing the uncertainty in the return level estimation. The model provides new information to analyze the seasonal behavior of wave height extremes affecting different natural coastal processes.     

Sea surface wind speed retrieval from Sentinel-1 HH polarization data using conventional and neural network methods

Conventional retrieval and neural network methods are used simultaneously to retrieve sea surface wind speed(SSWS) from HH-polarized Sentinel-1(S1) SAR images. The Polarization Ratio(PR) models combined with the CMOD5.N Geophysical Model Function(GMF) is used for SSWS retrieval from the HH-polarized SAR data. We compared different PR models developed based on previous C-band SAR data in HH-polarization for their applications to the S1 SAR data. The recently proposed CMODH, i.e., retrieving SSWS directly from the HHpolarized S1 data is also validated. The results indicate that the CMODH model performs better than results achieved using the PR models. We proposed a neural network method based on the backward propagation(BP)neural network to retrieve SSWS from the S1 HH-polarized data. The SSWS retrieved using the BP neural network model agrees better with the buoy measurements and ASCAT dataset than the results achieved using the conventional methods. Compared to the buoy measurements, the bias, root mean square error(RMSE) and scatter index(SI) of wind speed retrieved by the BP neural network model are 0.10 m/s, 1.38 m/s and 19.85%,respectively, while compared to the ASCAT dataset the three parameters of training set are –0.01 m/s, 1.33 m/s and 15.10%, respectively. It is suggested that the BP neural network model has a potential application in retrieving SSWS from Sentinel-1 images acquired at HH-polarization.     

Application of artificial neural networks in tide-forecasting

An accurate tidal forecast is an important task in determining constructions and human activities in ocean environments. Conventional tidal forecasting has been based on harmonic analysis using the least squares method to determine harmonic parameters. However, a large number of parameters are required for the prediction of a long-term tidal level with harmonic analysis. Unlike conventional harmonic analysis, this paper presents an artificial neural network (ANN) model for forecasting the tidal-level using the short term measuring data. The ANN model can easily decide the unknown parameters by learning the input–output interrelation of the short-term tidal records. Three field data with three types of tides will be used to test the performance of the proposed ANN model. The numerical results indicate that the hourly tidal levels over a long duration can be predicted using a short-term hourly tidal record.