A new analysis framework based on probability density evolution method(PDEM)and its Chebyshev collocation solution are introduced to predict the dynamic response and short-term extreme load of offshore wind turbine(OW...A new analysis framework based on probability density evolution method(PDEM)and its Chebyshev collocation solution are introduced to predict the dynamic response and short-term extreme load of offshore wind turbine(OWT)towers subjected to random sea state.With regard to the stochastic responses,random function method is employed to generate samples of sea elevation,the probability density evolution equation(PDEE)is solved to calculate time-variant probability density functions of structural responses.For the probabilistic load estimation,a FAST model of NREL 5MW offshore turbine is established to obtain samples of bending moment at the tower base.The equivalent extreme event theory is used to construct a virtual stochastic process(VSP)to assess the short-term extreme load.The results indicate that the proposed approach can predict time-variant probability density functions of the structural responses,and shows good agreement with Monte Carlo simulations.Additionally,the predicted short-term extreme load can capture the fluctuation at the tail of the extreme value distribution,thus is more rational than results from the typical distribution models.Overall,the proposed method shows good adaptation,precision and efficiency for the dynamic response analysis and load estimation of OWT towers.展开更多
The current measurement was exploited in a more efficient way. Firstly, the system equation was updated by introducing a correction term, which depends on the current measurement and can be obtained by running a subop...The current measurement was exploited in a more efficient way. Firstly, the system equation was updated by introducing a correction term, which depends on the current measurement and can be obtained by running a suboptimal filter. Then, a new importance density function(IDF) was defined by the updated system equation. Particles drawn from the new IDF are more likely to be in the significant region of state space and the estimation accuracy can be improved. By using different suboptimal filter, different particle filters(PFs) can be developed in this framework. Extensions of this idea were also proposed by iteratively updating the system equation using particle filter itself, resulting in the iterated particle filter. Simulation results demonstrate the effectiveness of the proposed IDF.展开更多
概率密度演化方法(probability density evolution equation,PDEM)为非线性随机结构的动力响应分析提供了新的途径.通过PDEM获得结构响应概率密度函数(probability density function,PDF)的关键步骤是求解广义概率密度演化方程(generali...概率密度演化方法(probability density evolution equation,PDEM)为非线性随机结构的动力响应分析提供了新的途径.通过PDEM获得结构响应概率密度函数(probability density function,PDF)的关键步骤是求解广义概率密度演化方程(generalized probability density evolution equation,GDEE).对于GDEE的求解通常采用有限差分法,然而,由于GDEE是初始条件间断的变系数一阶双曲偏微分方程,通过有限差分法求解GDEE可能会面临网格敏感性问题、数值色散和数值耗散现象.文章从全局逼近的角度出发,基于Chebyshev拟谱法为GDEE构造了全局插值格式,解决了数值色散、数值耗散以及网格敏感性问题.考虑GDEE的系数在每个时间步长均为常数,推导了GDEE在每一个时间步长内时域上的序列矩阵指数解.由于序列矩阵指数解形式上是解析的,从而很好地克服了数值稳定性问题.两个数值算例表明,通过Chebyshev拟谱法结合时域的序列矩阵指数解求解GDEE得到的结果与精确解以及Monte Carlo模拟的结果非常吻合,且数值耗散和数值色散现象几乎可以忽略.此外,拟谱法具有高效的收敛性且序列矩阵指数解不受CFL (Courant-Friedrichs-Lewy)条件的限制,因此该方法具有良好的数值稳定性和计算效率.展开更多
This paper studies the dynamic estimation problem for multitarget tracking. A novel gat- ing strategy that is based on the measurement likelihood of the target state space is proposed to improve the overall effectiven...This paper studies the dynamic estimation problem for multitarget tracking. A novel gat- ing strategy that is based on the measurement likelihood of the target state space is proposed to improve the overall effectiveness of the probability hypothesis density （PHD） filter. Firstly, a measurement-driven mechanism based on this gating technique is designed to classify the measure- ments. In this mechanism, only the measurements for the existing targets are considered in the update step of the existing targets while the measurements of newborn targets are used for exploring newborn targets. Secondly, the gating strategy enables the development of a heuristic state estima- tion algorithm when sequential Monte Carlo （SMC） implementation of the PHD filter is investi- gated, where the measurements are used to drive the particle clustering within the space gate. The resulting PHD filter can achieve a more robust and accurate estimation of the existing targets by reducing the interference from clutter. Moreover, the target birth intensity can be adaptive to detect newborn targets, which is in accordance with the birth measurements. Simulation results demonstrate the computational efficiency and tracking performance of the proposed algorithm.展开更多
In Bayesian multi-target fltering,knowledge of measurement noise variance is very important.Signifcant mismatches in noise parameters will result in biased estimates.In this paper,a new particle flter for a probabilit...In Bayesian multi-target fltering,knowledge of measurement noise variance is very important.Signifcant mismatches in noise parameters will result in biased estimates.In this paper,a new particle flter for a probability hypothesis density（PHD）flter handling unknown measurement noise variances is proposed.The approach is based on marginalizing the unknown parameters out of the posterior distribution by using variational Bayesian（VB）methods.Moreover,the sequential Monte Carlo method is used to approximate the posterior intensity considering non-linear and non-Gaussian conditions.Unlike other particle flters for this challenging class of PHD flters,the proposed method can adaptively learn the unknown and time-varying noise variances while fltering.Simulation results show that the proposed method improves estimation accuracy in terms of both the number of targets and their states.展开更多
提出一种基于粒子概率假设密度滤波器(Sequential Monte Carlo probability hypothesis density filter,SMC-PHDF)的部分可分辨的群目标跟踪算法.该算法可直接获得群而非个体的个数和状态估计.这里群的状态包括群的质心状态和形状.为了...提出一种基于粒子概率假设密度滤波器(Sequential Monte Carlo probability hypothesis density filter,SMC-PHDF)的部分可分辨的群目标跟踪算法.该算法可直接获得群而非个体的个数和状态估计.这里群的状态包括群的质心状态和形状.为了估计群的个数和状态,该算法利用高斯混合模型(Gaussian mixture models,GMM)拟合SMC-PHDF中经重采样后的粒子分布,这里混合模型的元素个数和参数分别对应于群的个数和状态.期望最大化(Expectation maximum,EM)算法和马尔科夫链蒙特卡洛(Markov chain Monte Carlo,MCMC)算法分别被用于估计混合模型的参数.混合模型的元素个数可通过删除、合并及分裂算法得到.100次蒙特卡洛(Monte Carlo,MC)仿真实验表明该算法可有效跟踪部分可分辨的群目标.相比EM算法,MCMC算法能够更好地提取群的个数和状态,但它的计算量要大于EM算法.展开更多
考虑到存活目标与新生目标在动态演化特性上的差异性,提出了面向快速多目标跟踪的协同概率假设密度(collaborative probability hypothesis density,CoPHD)滤波框架。该框架利用存活目标的状态信息,将量测动态划分为存活目标量测集与新...考虑到存活目标与新生目标在动态演化特性上的差异性,提出了面向快速多目标跟踪的协同概率假设密度(collaborative probability hypothesis density,CoPHD)滤波框架。该框架利用存活目标的状态信息,将量测动态划分为存活目标量测集与新生目标量测集,在两个量测集分别运用PHD组处理更新基础上建立了处理模块的交互与协同机制,力图在保证跟踪精度的同时提高计算效率。该框架由于采用PHD组处理方式而具有状态自动提取功能。进一步给出了该框架的序贯蒙特卡罗算法实现。仿真结果表明,该算法在计算效率以及状态提取精度上具有明显优势。展开更多
基金This research is supported by the National Natural Science Foundation of China(Grant No.51578444)Key Science Research Program of Education Department of Shaanxi Province(Grant No.20JY032).
文摘A new analysis framework based on probability density evolution method(PDEM)and its Chebyshev collocation solution are introduced to predict the dynamic response and short-term extreme load of offshore wind turbine(OWT)towers subjected to random sea state.With regard to the stochastic responses,random function method is employed to generate samples of sea elevation,the probability density evolution equation(PDEE)is solved to calculate time-variant probability density functions of structural responses.For the probabilistic load estimation,a FAST model of NREL 5MW offshore turbine is established to obtain samples of bending moment at the tower base.The equivalent extreme event theory is used to construct a virtual stochastic process(VSP)to assess the short-term extreme load.The results indicate that the proposed approach can predict time-variant probability density functions of the structural responses,and shows good agreement with Monte Carlo simulations.Additionally,the predicted short-term extreme load can capture the fluctuation at the tail of the extreme value distribution,thus is more rational than results from the typical distribution models.Overall,the proposed method shows good adaptation,precision and efficiency for the dynamic response analysis and load estimation of OWT towers.
基金Project(61271296) supported by the National Natural Science Foundation of China
文摘The current measurement was exploited in a more efficient way. Firstly, the system equation was updated by introducing a correction term, which depends on the current measurement and can be obtained by running a suboptimal filter. Then, a new importance density function(IDF) was defined by the updated system equation. Particles drawn from the new IDF are more likely to be in the significant region of state space and the estimation accuracy can be improved. By using different suboptimal filter, different particle filters(PFs) can be developed in this framework. Extensions of this idea were also proposed by iteratively updating the system equation using particle filter itself, resulting in the iterated particle filter. Simulation results demonstrate the effectiveness of the proposed IDF.
文摘概率密度演化方法(probability density evolution equation,PDEM)为非线性随机结构的动力响应分析提供了新的途径.通过PDEM获得结构响应概率密度函数(probability density function,PDF)的关键步骤是求解广义概率密度演化方程(generalized probability density evolution equation,GDEE).对于GDEE的求解通常采用有限差分法,然而,由于GDEE是初始条件间断的变系数一阶双曲偏微分方程,通过有限差分法求解GDEE可能会面临网格敏感性问题、数值色散和数值耗散现象.文章从全局逼近的角度出发,基于Chebyshev拟谱法为GDEE构造了全局插值格式,解决了数值色散、数值耗散以及网格敏感性问题.考虑GDEE的系数在每个时间步长均为常数,推导了GDEE在每一个时间步长内时域上的序列矩阵指数解.由于序列矩阵指数解形式上是解析的,从而很好地克服了数值稳定性问题.两个数值算例表明,通过Chebyshev拟谱法结合时域的序列矩阵指数解求解GDEE得到的结果与精确解以及Monte Carlo模拟的结果非常吻合,且数值耗散和数值色散现象几乎可以忽略.此外,拟谱法具有高效的收敛性且序列矩阵指数解不受CFL (Courant-Friedrichs-Lewy)条件的限制,因此该方法具有良好的数值稳定性和计算效率.
基金supported by the Aeronautical Science Foundation of China(No.201401P6001)
文摘This paper studies the dynamic estimation problem for multitarget tracking. A novel gat- ing strategy that is based on the measurement likelihood of the target state space is proposed to improve the overall effectiveness of the probability hypothesis density （PHD） filter. Firstly, a measurement-driven mechanism based on this gating technique is designed to classify the measure- ments. In this mechanism, only the measurements for the existing targets are considered in the update step of the existing targets while the measurements of newborn targets are used for exploring newborn targets. Secondly, the gating strategy enables the development of a heuristic state estima- tion algorithm when sequential Monte Carlo （SMC） implementation of the PHD filter is investi- gated, where the measurements are used to drive the particle clustering within the space gate. The resulting PHD filter can achieve a more robust and accurate estimation of the existing targets by reducing the interference from clutter. Moreover, the target birth intensity can be adaptive to detect newborn targets, which is in accordance with the birth measurements. Simulation results demonstrate the computational efficiency and tracking performance of the proposed algorithm.
基金supported by National High-tech Research and Development Program of China (No.2011AA7014061)
文摘In Bayesian multi-target fltering,knowledge of measurement noise variance is very important.Signifcant mismatches in noise parameters will result in biased estimates.In this paper,a new particle flter for a probability hypothesis density（PHD）flter handling unknown measurement noise variances is proposed.The approach is based on marginalizing the unknown parameters out of the posterior distribution by using variational Bayesian（VB）methods.Moreover,the sequential Monte Carlo method is used to approximate the posterior intensity considering non-linear and non-Gaussian conditions.Unlike other particle flters for this challenging class of PHD flters,the proposed method can adaptively learn the unknown and time-varying noise variances while fltering.Simulation results show that the proposed method improves estimation accuracy in terms of both the number of targets and their states.
文摘提出一种基于粒子概率假设密度滤波器(Sequential Monte Carlo probability hypothesis density filter,SMC-PHDF)的部分可分辨的群目标跟踪算法.该算法可直接获得群而非个体的个数和状态估计.这里群的状态包括群的质心状态和形状.为了估计群的个数和状态,该算法利用高斯混合模型(Gaussian mixture models,GMM)拟合SMC-PHDF中经重采样后的粒子分布,这里混合模型的元素个数和参数分别对应于群的个数和状态.期望最大化(Expectation maximum,EM)算法和马尔科夫链蒙特卡洛(Markov chain Monte Carlo,MCMC)算法分别被用于估计混合模型的参数.混合模型的元素个数可通过删除、合并及分裂算法得到.100次蒙特卡洛(Monte Carlo,MC)仿真实验表明该算法可有效跟踪部分可分辨的群目标.相比EM算法,MCMC算法能够更好地提取群的个数和状态,但它的计算量要大于EM算法.
文摘考虑到存活目标与新生目标在动态演化特性上的差异性,提出了面向快速多目标跟踪的协同概率假设密度(collaborative probability hypothesis density,CoPHD)滤波框架。该框架利用存活目标的状态信息,将量测动态划分为存活目标量测集与新生目标量测集,在两个量测集分别运用PHD组处理更新基础上建立了处理模块的交互与协同机制,力图在保证跟踪精度的同时提高计算效率。该框架由于采用PHD组处理方式而具有状态自动提取功能。进一步给出了该框架的序贯蒙特卡罗算法实现。仿真结果表明,该算法在计算效率以及状态提取精度上具有明显优势。
