Fault detection and identification in dynamic systems with noisy data and parameter/modeling uncertainties

A probabilistic approach is presented which can be used for the estimation of system parameters and unmonitored state variables towards model-based fault diagnosis in dynamic systems. The method can be used with any type of input–output model and can accommodate noisy data and/or parameter/modeling uncertainties. The methodology is based on Markovian representation of system dynamics in discretized state space. The example system used for the illustration of the methodology focuses on the intake, fueling, combustion and exhaust components of internal combustion engines. The results show that the methodology is capable of estimating the system parameters and tracking the unmonitored dynamic variables within user-specified magnitude intervals (which may reflect noise in the monitored data, random changes in the parameters or modeling uncertainties in general) within data collection time and hence has potential for on-line implementation.     

Non-local modelling of cyclic thermal shock damage including parameter estimation

In this paper, rate dependent evolution laws are identified and characterized to model the mechanical (elasticity-based) and thermal damage occurring in coarse grain refractory material subject to cyclic thermal shock. The interacting mechanisms for elastic deformation driven damage induced by temperature gradients and thermal damage induced by isotropic thermal expansion are combined and represented by a single variable for the total damage. The constitutive model includes the shielding of micro-structural thermal damage by the non-local elasticity-based damage developed at the macroscopic and microscopic scale. Quasi-stationary thermal experiments are used to identify the parameters used in the evolution law for thermal damage. The remaining model parameters, including a micro-structural length scale, are quantified by inverse modelling of cyclic thermal shock experiments. Longitudinal wave propagation measurements through damaged material are simulated, enabling the identification on the basis of the first and second thermal shock cycle. A third thermal shock cycle enabled the evaluation of the quality of the obtained parameter set. The set-up of the thermal shock experiments has been optimized through a parameter identifiability analysis. The damage evolution in three consecutive thermal shock cycles is investigated numerically with the optimized model.     

Fault diagnostics of dynamic system operation using a fault tree based method

For conventional systems, their availability can be considerably improved by reducing the time taken to restore the system to the working state when faults occur. Fault identification can be a significant proportion of the time taken in the repair process. Having diagnosed the problem the restoration of the system back to its fully functioning condition can then take place. This paper expands the capability of previous approaches to fault detection and identification using fault trees for application to dynamically changing systems. The technique has two phases. The first phase is modelling and preparation carried out offline. This gathers information on the effects that sub-system failure will have on the system performance. Causes of the sub-system failures are developed in the form of fault trees. The second phase is application. Sensors are installed on the system to provide information about current system performance from which the potential causes can be deduced. A simple system example is used to demonstrate the features of the method. To illustrate the potential for the method to deal with additional system complexity and redundancy, a section from an aircraft fuel system is used. A discussion of the results is provided.     

Online parameter estimation and run-to-run process adjustment using categorical observations

Categorical observations are frequently observed in run-to-run processes where obtaining accurate measurements of quality characteristics is difficult. In such circumstances, the use of categorical observations to estimate a process model and generate an adjustment recipe becomes inevitable. However, most conventional run-to-run controllers cannot be applied if no continuous observations are available; some parameter estimation methods that can handle categorical data only use historical dataset in an offline manner. In practice, it is common to see observations collected following a time sequence in a run-to-run process. Taking the lapping process in semiconductor manufacturing as an example, this paper develops an online approach for parameters estimation and run-to-run process adjustment using categorical observations. The proposed method optimises a penalised Maximum Likelihood (ML) function and updates parameters step by step when new categorical observations become available. A control strategy is also derived to generate receipts for process update between runs. The computational results of performance evaluation show that the proposed method is capable of estimating unknown parameters and control output quality online when initial bias exists.     

主动声纳宽带信号的一种高效参数估计算法

讨论在宽带信号体制下,对水中高速目标的检测与估计问题,给出了高斯包络线性调频信号必须采用宽带模糊函数进行检测时的带宽和目标速度范围。在仿真分析了信号宽带、窄带模糊函数主峰特性的基础上,提出了一种联合利用宽带模糊函数和窄带模糊函数的高效目标参数估计算法,该算法可以将二维相关运算简化到一维相关运算,降低了运算量,同时具有良好的估计精度。     

Bayesian state and parameter estimation of uncertain dynamical systems

The focus of this paper is Bayesian state and parameter estimation using nonlinear models. A recently developed method, the particle filter, is studied that is based on stochastic simulation. Unlike the well-known extended Kalman filter, the particle filter is applicable to highly nonlinear models with non-Gaussian uncertainties. Recently developed techniques that improve the convergence of the particle filter simulations are introduced and discussed. Comparisons between the particle filter and the extended Kalman filter are made using several numerical examples of nonlinear systems. The results indicate that the particle filter provides consistent state and parameter estimates for highly nonlinear models, while the extended Kalman filter does not.     

Estimation of mixed Weibull distribution parameters using the SCEM-UA algorithm: Application and comparison with MLE in automotive reliability analysis

The shuffled complex-evolution metropolis algorithm (SCEM-UA) is used to estimate mixed Weibull distribution parameters in automotive reliability analysis. The results are compared with maximum likelihood estimation (MLE) results. The comparison shows that, in the examples given, SCEM-UA can deliver more accurate results than MLE overall.     

联合检测估计及其性能分析

二元假设检验问题可视作对一个取值0、1的离散随机变量的估计,于是原先分离的信号检测和参数估计可以纳入统一的框架进行处理,这就是联合检测估计。介绍了该方法的一般形式,并就不同的情况讨论它的性能及其与传统顺序方法的比较。发现在线性模型下,联合方法和顺序方法具有内在的等价性;但对更一般的情况,则要略逊色于顺序方法。联合方法的优势在于只需计算一个广义的MLE而不是多重积分,因而在应用中更为可行。特别在高信噪比条件下,两种方法的差别微乎其微,联合方法是一种更为可取的准最佳方法。     

Parameter estimation with correlated outputs using fidelity maps

This paper introduces a new approach for parameter estimation and model update based on the notion of fidelity maps. Fidelity maps refer to the regions of the parameter space within which the discrepancy between computational and experimental data is below a user-defined threshold. It is shown that fidelity maps provide an efficient and rigorous approach to approximate likelihoods in the context of Bayesian update or maximum likelihood estimation. Fidelity maps are constructed explicitly in terms of the parameters and aleatory uncertainties using a Support Vector Machine (SVM) classifier. The approach has the advantage of handling numerous correlated responses, possibly discontinuous, without any assumption on the correlation structure. The construction of accurate fidelity map boundaries at a moderate computational cost is made possible through a dedicated adaptive sampling scheme. A simply supported plate with uncertainties in the boundary conditions is used to demonstrate the methodology. In this example, the construction of the fidelity map is based on several natural frequencies and mode shapes to be matched simultaneously. Various statistical estimators are derived from the map.     

一种新颖的故障检测滤波器设计

研究了含有未知输入线性系统的故障检测问题,针对传统方法丢失信息或繁琐复杂的缺点,设计了一种新颖的故障检测滤波器。该方法使未知输入在故障发生前衰减到几乎处处为零,系统近似为无未知输入的标准系统,残差只敏感于故障,动态的实现了未知输入与故障的分离。仿真结果表明,该方法效果良好。     

Parameter estimation of an aeroelastic aircraft using neural networks

Application of neural networks to the problem of aerodynamic modelling and parameter estimation for aeroelastic aircraft is addressed. A neural model capable of predicting generalized force and moment coefficients using measured motion and control variables only, without any need for conventional normal elastic variables or their time derivatives, is proposed. Furthermore, it is shown that such a neural model can be used to extract equivalent stability and control derivatives of a flexible aircraft. Results are presented for aircraft with different levels of flexibility to demonstrate the utility of the neural approach for both modelling and estimation of parameters.     

Fault detection using dynamic time warping (DTW) algorithm and discriminant analysis for swine wastewater treatment

This paper proposes a diagnosis system using dynamic time warping (DTW) and discriminant analysis with oxidation-reduction potential (ORP) and dissolved oxygen (DO) values for swine wastewater treatment. A full-scale sequencing batch reactor (SBR), which has an effective volume of 20 m(3), was auto-controlled, and the reaction phase was performed by a sub-cycle operation consisting of a repeated short cycle of the anoxic-aerobic step. Using ORP and DO profiles, SBR status was divided into four categories of normal and abnormal cases; these were influent disturbance, aeration controller fault, instrument trouble and inadequate raw wastewater feeding. Through the DTW process, difference values (D) were determined and classified into seven cases. In spite of the misclassification of high loading rates, the ORP profile provided good diagnosis results. However, the DO profiles detected five misclassifications that indicated different statuses. After the DTW process, several statistical values, including maximum value, minimum value, average value, standard deviation value and three quartile values, were extracted and applied to establish the discriminant function. The discriminant analysis allows one to classify seven cases with a percentage of 100% and 92.7% for ORP and DO profiles, respectively. Consequently, the study showed that ORP profiles are more efficient than DO profiles as diagnosis parameters and DTW diagnosis algorithms and discriminants.     

Photoacoustic determination of thermophysical properties of thin metallic plates using parameter estimation

The phase and the amplitude of the photoacoustic signal were measured as a function of chopping frequency for several kinds of widely used thin metallic plates (stainless steel 304, brass, aluminum, and copper) attached to plexiglass backing. The experimental data have been analyzed systematically by parameter estimation technique based on the two-layer model developed from Rosencwaig-Gersho (R-G) theory. Using this analysis, the values of thermal diffusivity and thermal effusivity of the materials have been determined.     

A fuzzy logic based approach to reliability improvement estimation during product development

During early stages of product development process, a vast amount of knowledge and information is generated. However, most of it is subjective (imprecise) in nature and remains unutilized. This paper presents a formal structure for capturing this information and knowledge and utilizing it in reliability improvement estimation. The information is extracted as improvement indices from various design tools, experiments, and design review records and treated as fuzzy numbers or linguistic variables. Fuzzy reasoning method is used to combine and quantify the subjective information to map their impact on product reliability. The crisp output of the fuzzy reasoning process is treated as new evidence and incorporated into a Bayesian framework to update the reliability estimates. A case example is presented to demonstrate the proposed approach.     

基于慢度估计的次声台阵信号自动检测算法及应用

针对次声监测研究领域广泛使用的短时间与长时间信号功率比检测方法性能不能满足要求,渐进多通道互相关检测算法虚警率较高的缺点,利用信号的时、频、波数域特性和次声台阵各子台信号的互相关性特性,提出了一种基于次声台阵信号慢度估计的自动检测算法。次声信号慢度估计结果的合理性、一致性是该检测算法的依据。考虑到该检测算法计算量较大,在具体应用中加入其他计算量较小的传统算法,进行逐次检测,最后应用的慢度估计方法既是一种检测方法,也是对其他检测算法结果的最终审核,一方面可以降低单项检测方法对检测参数的敏感性,另一方面也可提高检测的正确率(虚警率和漏警率同时降低)。该复合检测算法应用于某实验次声台站信号的检测(信噪比大于1.1),结果虚警率为4.0%,漏警率为5.0%,表明了该方法的优良性能。从检测应用处理结果还可以推测,结合一个区域现有密集的地震台站数据和2~3个次声台阵的数据,可以方便地识别、监控爆破方式日益复杂的人工爆破事件和真正的地震(余震)事件,为地震灾害减灾防灾、采矿监控等多领域服务,具有重大推广价值。     

谐波信号的闭环检测和估计算法及其应用

从联合检测估计的角度讨论了检测器和估计器之间的耦合机制,并就水声目标线谱特征提取问题研究了一种谐波信号的闭环检测和估计算法.该算法通过迭代形式在检测器和估计器之间不断交换信息,用以改善未知参数的先验分布,从而得到较好的检测和估计性能.对海试数据的仿真结果表明,该算法可以有效地提取目标的线谱特征.     

Reliability estimation by Bayesian method: definition of prior distribution using dependability study

In this article, we present an evaluation method of the system reliability using Bayesian statistics and data feedback. In particular, we deal with the problem of the prior definition. A method is proposed to build the prior distribution from the dependability study (risk evaluation study).     

Bayes estimation of the general hazard rate model

In reliability theory and life testing models, the life time distributions are often specified by choosing a relevant hazard rate function. Here a general hazard rate function h(t)=a+bt^c−1, where c, a, b are constants greater than zero, is considered. The parameter c is assumed to be known. The Bayes estimators of (a,b) based on the data of type II/item-censored testing without replacement are obtained. A large simulation study using Monte Carlo Method is done to compare the performance of Bayes with regression estimators of (a,b). The criterion for comparison is made based on the Bayes risk associated with the respective estimator. Also, the influence of the number of failed items on the accuracy of the estimators (Bayes and regression) is investigated. Estimations for the parameters (a,b) of the linearly increasing hazard rate model h(t)=a+bt, where a, b are greater than zero, can be obtained as the special case, letting c=2.     

Self-sensing magnetic bearings using parameter estimation

A signal processing technique is presented by which the position of a rotor supported in magnetic bearings may be deduced from the bearing current waveform. The bearing currents are presumed to be developed with a two-state switching amplifier which produces a substantial high-frequency switching ripple. This ripple is demodulated using a parameter estimation technique which extracts the length of the bearing air gap while rejecting the influence of amplifier voltage and duty-cycle variation. The performance of the estimator is evaluated both by simulation and experiment. The method is demonstrated to produce a sensor with bandwidth of at least 1 kHz and acceptably low feed-through of the bearing force     

Fault detection via nonlinear profile monitoring using artificial neural networks

Fault detection is the characterization of a normal behavior of a system using a response function or profile of interest and the identification of any deviation from such normal behavior. As system complexity grows, predicting the underlying structure or form of response function becomes challenging if not impossible. This article presents a data‐driven approach for fault detection of complex systems using multivariate statistical process control based on artificial neural network (ANN) characterization. In this approach, the quality of a system is characterized where one explanatory variable is adequately explained as a function of the other variables using an ANN model. The vector of weights and biases of the ANN model is monitored by using Hotelling T ² through control charts. The proposed method is tested and compared with existing methods such as polynomial and sum of sine function regression for 3 cases from the literature. Moreover, it is applied to a 4‐story reinforced concrete building that uses continuous monitoring to avoid potentially catastrophic failures. The proposed ANN approach outperforms the existing methods for small shifts (deviations) from healthy states. For large and medium shifts, it provides comparable results that are on the conservative side.