|
|
|
题名
|
作者
|
年代
|
出处
|
被引量
|
| 1 | 基于PINNs的单旋翼植保无人机下洗流场预测模型显示文摘植保无人机(unmanned aerial vehicle,UAV)进行喷施作业时,旋翼高速旋转所产生的下洗流场是影响雾滴飘移的重要因素。为了快速准确地预测单旋翼植保无人机下洗流场的速度等流场参数,提升无人机精准施药效果,该研究基于物理信息神经网络(physics-informed neural networks,PINNs)构建了单旋翼植保无人机下洗流场的预测模型。在全连接神经网络结构的基础上,嵌入纳维-斯托克斯(Navier-Stokes,N-S)方程作为物理学损失项来参与训练,减轻网络模型对数据依赖性的同时增强了模型的可解释性。通过最小化损失函数,使得该模型学习到流场中流体的运动规律,得到时空坐标与速度信息等物理量之间的映射关系,从而实现对单旋翼无人机下洗流场的速度等参数的快速预测。最后通过风洞试验验证该预测模型的可行性和准确性。结果表明:没有侧风的情况下,预测模型在旋翼下方0.3、0.7、1.1以及1.5 m共4个不同高度处各向速度的预测值和试验值的误差均小于0.6 m/s,具有较小的差异性;不同侧风风速情况下,水平和竖直方向速度的预测值与试验值的总体拟合优度R2分别为0.941和0.936,表明所提出的模型在单旋翼植保无人机下洗流场预测方面具有良好的应用效果,能够快速准确地预测下洗流场的速度信息。研究结果可为进一步研究旋翼风场对雾滴沉积分布特性的影响机理提供数据支撑。 | 王涛 文晟 兰玉彬 张海艳 尹选春 张建桃 | 2023 | 农业工程学报2023，39，6： | 1 |
| 2 | Mechanistic Machine Learning:Theory,Methods,and Applications显示文摘Recent advances in machine learning are currently influencing the way we gather data,recognize patterns,and build predictive models across a wide range of scientific disciplines.Noticeable successes include solutions in image and voice recognition that have already become part of our everyday lives,mainly enabled by algorithmic developments,hardware advances,and,of course,the availability of massive data-sets.Many of such predictive tasks are currently being tackled using over-parameterized,black-box discriminative models such as deep neural networks,in which theoretical rigor,interpretability and adherence to first physical principles are often sacrificed in favor of flexibility in representation and scalability in computation. | Paris Perdikaris Shaoqiang Tang | 2020 | Theoretical & Applied Mechanics Letters2020，10，3： | 1 |
| 3 | Derivation of the Orthotropic Nonlinear Elastic Material Law Driven by Low-Cost Data(DDONE)显示文摘Orthotropic nonlinear elastic materials are common in nature and widely used by various industries.However,there are only limited constitutive models available in today's commercial software(e.g.,ABAQUS,ANSYS,etc.)that adequately describe their mechanical behavior.Moreover,the material parameters in these constitutive models are also difficult to calibrate through low-cost,widely available experimental setups.Therefore,it is paramount to develop new ways to model orthotropic nonlinear elastic materials.In this work,a data-driven orthotropic nonlinear elastic(DDONE)approach is proposed,which builds the constitutive response from stress–strain data sets obtained from three designed uniaxial tensile experiments.The DDONE approach is then embedded into a finite element(FE)analysis framework to solve boundary-value problems(BVPs).Illustrative examples(e.g.,structures with an orthotropic nonlinear elastic material)are presented,which agree well with the simulation results based on the reference material model.The DDONE approach generally makes accurate predictions,but it may lose accuracy when certain stress–strain states that appear in the engineering structure depart significantly from those covered in the data sets.Our DDONE approach is thus further strengthened by a mapping function,which is verified by additional numerical examples that demonstrate the effectiveness of our modified approach.Moreover,artificial neural networks(ANNs)are employed to further improve the computational efficiency and stability of the proposed DDONE approach. | Qian Xiang Hang Yang K.I.Elkhodary Zhi Sun Shan Tang Xu Guo | 2022 | Acta Mechanica Solida Sinica2022，35，5： | 0 |
| 4 | 基于数据同化的气动压力稀疏重构方法显示文摘风洞实验中获取模型高精度压力分布至关重要,但现有测量方法仍然存在一些缺陷。为获得风洞模型的全域压力分布,本文通过集合变换卡尔曼滤波(ETKF)对风洞实验的稀疏实测数据和数值计算数据进行同化,实现了基于模型物面有限测点的全空间流场高精度重构。分别使用二维翼型RAE 2822和NACA 0012进行实验验证,RAE 2822的压力稀疏重构结果比线性理论修正更加接近实测结果,此效果在激波位置体现得尤其明显,压力系数的预测误差降低了约3%,使用ETKF修正后的迎角及马赫数集合均值计算得到的机翼升力系数和力矩系数与实验值的误差均小于1%;NACA 0012实验面向风洞测量的全场感知应用,探讨了基于少量测点进行压力重构的可行性。实验结果表明:采用机翼物面6个测点重构的压力系数,相对误差可达2.42%,且同化效果与数据点位置密切相关。 | 黄俊 郭雨欣 冀晶晶 黄永安 | 2023 | 实验流体力学2023，37，5： | 0 |
| 5 | Neuroevolution-enabled adaptation of the Jacobi method for Poisson’s equation with density discontinuities显示文摘Lacking labeled examples of working numerical strategies,adapting an iterative solver to accommodate a numerical issue,e.g.,density discontinuities in the pressure Poisson equation,is non-trivial and usually involves a lot of trial and error.Here,we resort to evolutionary neural network.A evolutionary neural network observes the outcome of an action and adapts its strategy accordingly.The process requires no labeled data but only a measure of a network’s performance at a task.Applying neuro-evolution and adapting the Jacobi iterative method for the pressure Poisson equation with density discontinuities,we show that the adapted Jacobi method is able to accommodate density discontinuities. | T.-R.Xiang X.I.A.Yang Y.-P.Shi | 2021 | Theoretical & Applied Mechanics Letters2021，11，3： | 0 |
您所在的机构暂未开通此部分的功能，请联系管理员