基于Deform软件二次开发和BP神经网络的TA15多向锻造微观组织预报

通过热压缩实验构建了TA15合金的热变形真应力-应变曲线,以此为基础分别建立了合金双相区及单相区温度区间的热变形本构方程;基于热压缩试样动态再结晶的统计数据建立了TA15合金的动态再结晶模型。借助Deform提供的二次开发功能实现相关数学模型的程序化,制定正交实验方案,实现了TA15合金多向锻造变形的微观组织仿真。通过正交实验分析得出各项因子的影响对象及强弱差异,提出了双相区及单相区温度区间内的多向锻造最佳因子组合。建立了TA15合金多向锻造变形微观组织的BP （Back Propagation）神经网络预测模型,将预测结果与有限元仿真结果进行比较,结果表明两种方法的预测结果基本一致,但神经网络具备有限元仿真难以实现的良好细节预测能力,能更为细致地实现对微观组织分布状态的划分。     

Strong geomagnetic activity forecast by neural networks under dominant southern orientation of the interplanetary magnetic field

The paper deals with the relation of the southern orientation of the north–south component _Bz$B_z$ of the interplanetary magnetic field to geomagnetic activity (GA) and subsequently a method is suggested of using the found facts to forecast potentially dangerous high GA. We have found that on a day with very high GA hourly averages of _Bz$B_z$ with a negative sign occur at least 16 times in typical cases. Since it is very difficult to estimate the orientation of _Bz$B_z$ in the immediate vicinity of the Earth one day or even a few days in advance, we have suggested using a neural-network model, which assumes the worse of the possibilities to forecast the danger of high GA – the dominant southern orientation of the interplanetary magnetic field. The input quantities of the proposed model were information about X-ray flares, type II and IV radio bursts as well as information about coronal mass ejections (CME). In comparing the GA forecasts with observations, we obtain values of the Hanssen–Kuiper skill score ranging from 0.463 to 0.727, which are usual values for similar forecasts of space weather. The proposed model provides forecasts of potentially dangerous high geomagnetic activity should the interplanetary CME (ICME), the originator of geomagnetic storms, hit the Earth under the most unfavorable configuration of cosmic magnetic fields. We cannot know in advance whether the unfavorable configuration is going to occur or not; we just know that it will occur with the probability of 31%.     

综合路由的无线传感器网络覆盖调度

利用目前公认为理论上最好的无线传感器网络集中式覆盖控制算法MCLC(Most-Constrained Least-Constraining),结合DD(Directed Diffusion)和LEACH(Low Energy Adaptive Clustering Hierarchy)路由协议,基于节点最小剩余能量调度覆盖集,观察最小剩余能量不同选取对节点能耗和均衡性以及网络生存期的影响.研究能耗均衡性和网络生存期的关系,提出了一种新的一般性能耗均衡的覆盖集调度的方法.仿真结果表明:所提方法可以有效提高网络生存期,同时为设计覆盖控制算法和覆盖集调度策略提供理论指导.      

利用神经元拓展正则极端学习机预测时间序列

为实现对于时间序列预测数据的准确预测,提出一种神经元拓展正则极端学习机(NERELM,Neuron-Expanding Regularized Extreme Learning Machine),并研究了其在时间序列预测中的应用.NERELM根据结构风险最小化原理权衡经验风险与结构风险,以逐次拓展隐层神经元的方式自动确定最佳的网络结构,以避免传统神经网络训练过程中需人为确定网络结构的弊端.应用于时间序列的仿真结果表明:NERELM可有效实现对于RELM最佳网络结构的自动确定,具有预测精度高与计算速度快的优点.      

液体火箭发动机传感器故障检测与数据恢复算法研究

以某型液体火箭发动机为研究对象,根据其传感器的故障特性,提出了基于BP神经网络的传感器故障检测与数据恢复算法。通过定义的传感器置信度来判断传感器是否发生故障,以及确定故障传感器,利用已训练好的神经网络结构对故障传感器进行数据恢复。研究内容能够实现传感器的故障检测、定位与补偿,能够有效提高发动机故障检测方法的可靠性和鲁棒性。     

导弹发射装置随动系统模糊神经PID控制器

在有些系统性能指标要求很高的导弹发射装置的随动系统中,传统的PID控制往往难以满足要求。文章以某型导弹发射装置的随动系统为模型,结合了模糊控制、神经网络及PID等控制算法,设计了模糊神经PID控制器。通过与传统的PID控制器仿真实验对比,可以看出应用模糊神经PID控制器能够有效地提高该随动系统的动态性能和鲁棒稳定性。     

Aerodynamic Modeling and Parameter Estimation from QAR Data of an Airplane Approaching a High-altitude Airport

Aerodynamic modeling and parameter estimation from quick accesses recorder (QAR) data is an important technical way to analyze the effects of highland weather conditions upon aerodynamic characteristics of airplane. It is also an essential content of flight accident analysis. The related techniques are developed in the present paper, including the geometric method for angle of attack and sideslip angle estimation, the extended Kalman filter associated with modified Bryson-Frazier smoother (EKF-MBF) method for aerodynamic coefficient identification, the radial basis function (RBF) neural network method for aerodynamic modeling, and the Delta method for stability/control derivative estimation. As an application example, the QAR data of a civil airplane approaching a high-altitude airport are processed and the aerodynamic coefficient and derivative estimates are obtained. The estimation results are reasonable, which shows that the developed techniques are feasible. The causes for the distribution of aerodynamic derivative estimates are analyzed. Accordingly, several measures to improve estimation accuracy are put forward.     

具有饱和补偿的碟形飞行器神经滑模控制

针对执行机构饱和非线性未知的碟形飞行器,在考虑执行机构动态的基础上,设计了基于神经网络的滑模控制器,给出了相应的控制律和参数选择方法.神经网络用来估计执行机构的饱和量,从而在设计控制器时,可以对执行机构的饱和进行相应补偿.仿真结果表明了该方法的正确性和有效性     

基于BP神经网络的高精度陀螺恒流源补偿方法

为了提高陀螺恒流源精度,提出一种基于BP神经网络的陀螺恒流源补偿方法.采用BP神经网络训练恒流源控制指令与恒流源输出之间的非线性映射稳态模型,以实时估计恒流源输出偏差.设计恒流源控制指令补偿判据,当输出偏差超出设定裕度时,按比例对恒流源控制指令值进行实时补偿,使得恒流源输出更接近控制目标值,以实现更优的精度.通过实物在回路仿真验证了上述方案的有效性,并通过与传统对控制指令进行分段线性标定方法相比较,显示了上述方案的恒流控制优势.     

BP神经网络反演核事故源项中重要参数的研究

核事故发生时,可靠、准确的源项信息能为应急防护行动措施决策提供数据支持。采用Matlab软件神经网络工具箱可以实现基于BP神经网络的核事故源项反演,为提高核事故源项反演计算的准确度,针对反演时的几个重要参数进行研究,包括隐含层节点数、训练函数、学习率和隐含层数。研究结果表明,在单隐含层神经网络结构情况下,存在着最优隐含层节点数,综合考虑训练时间和误差,本文选取隐含层节点数为50来对其他参数影响进行进一步研究;在相同参数设置条件下,训练函数Trainlm比Traingdm更适合数据量较小时的核事故源项反演,反演计算准确度更高,在节点数为50时训练时间缩短了近35%;高学习率以及双隐含层能有效地提高核事故源项反演的精度,但训练时间相对增加。