基于特征空间k最近邻的批次过程监视

针对具有非高斯、非线性及多工况特性的批次过程,提出一种基于特征量最近邻统计指标的过程监视方法. 首先,将批次过程正常工况原始数据投影到其特征空间,提取主元T和平方预测误差SPE,并进行特征量k最近邻距离平方和的求解. 然后,采用核密度估计法获得概率密度分布函数,确定统计监视控制限. 特征空间的主元T和SPE特征量能全面代表原始数据的有用信息. 采用特征量k最近邻建立监视模型将会节省存储空间,提高建模样本数量与变量之比以及检测异常工况的速度. 另外,利用局部近邻数据建模可以解决过程具有的非线性和多工况问题,而应用核密度估计法可以解决过程数据具有的非高斯分布问题. 最后,在半导体生产过程的成功应用表明了所提方法的有效性.     

Constrained MIMO dynamic discrete-time modeling exploiting optimal experimental design

This article presents a new multiple input, multiple output (MIMO) constrained discrete-time modeling (DTM) approach for dynamic block-oriented processes that does not require the nonlinear steady state characteristics to be known prior to model development. This approach uses an efficient statistical experimental design to provide design points for sequential step tests. The DTM is developed from this data in two stages. In the first stage, the ultimate response (steady state) model is determined from just the ultimate response data of the sequential step tests. In the second stage, the dynamic parameters are estimated under the constraint of the fitted ultimate response model obtained in the first stage. The constrained formulation is given for MIMO Hammerstein and Wiener block-oriented systems. Comparison of the proposed constrained DTM method is made with unconstrained DTM and constrained continuous-time modeling (CTM). Prediction accuracy of the proposed method is significantly better than unconstrained DTM and comparable to constrained CTM for the process studied.     

Computationally feasible narrow‐band antenna modeling using response features

Accurate and low‐cost models of input characteristics are of primary importance from the point of view of efficient design of antenna structures. Yet, the modeling problem is difficult because reflection responses are highly nonlinear functions of frequency and change considerably when adjusting antenna dimensions. Conventional approximation‐based models require massive datasets and often fail to provide required accuracy. This work demonstrates a possibility of dramatic reduction of the number of training samples, which is achieved by reformulating the modeling problem in a space of appropriately defined response features. The key factor is that dependence of feature point coordinates (both frequency and level) on antenna dimensions is less nonlinear than for the standard responses (S‐parameters vs. frequency). Our methodology permits construction of reliable surrogates using much smaller datasets than those required by conventional approaches. Experimental validation indicates that our models provide accuracy that is sufficient for practical antenna design.     

Robust <Emphasis Type="Italic">μ</Emphasis>-synthesis controllers for suppressing stick-slip induced vibrations in oil well drill strings

Stick-slip friction is a major cause of drill-string failure. This paper addresses the problem of suppressing stick-slip induced oscillations in oil well drill strings using a control design technique known as μ-synthesis. This technique allows for the inclusion of modeling errors in the control design process in terms of uncertainty weights. The dynamic model of the drill string with stick-slip friction is highly nonlinear and has to be linearized around an operating point in order to use μ-synthesis. The difference between the linear and nonlinear models is characterized in terms of uncertainty weights and included in the control design process. The designed controllers are robust to uncertainty in the dynamic model, spillover, actuator uncertainty, and noise. Two controllers were designed using μ-synthesis and the simulation results are presented and discussed here. The first controller assumes no measurement delay; however, the second controller includes a sensor time delay in the measurements. Both controllers are robust and performed well.     

多自由度非线性系统动态参数化模型建模方法研究

针对多自由度非线性系统的动态模型辨识问题,基于NARX(Non-linear Autoregressive with Exogenous inputs)模型的建模方法,考虑系统的物理设计参数,建立非线性系统动态参数化模型.首先,根据系统输入、输出数据建立系统不同参数下的NARX模型,并通过EFOR(Extended Forward Orthogonal Regression)算法对不同参数下NARX模型进行修正,以统一辨识得到的系统模型结构.随后,建立NARX模型系数与物理设计参数间的函数关系,得到多自由度非线性系统的动态参数化模型.以单输入、单输出两自由度非线性系统为例,根据数值仿真结果,对系统的动态参数化模型建模过程进行说明.最后,以带非线性涂层阻尼的悬臂梁作为试验对象,建立其动态参数化模型以反映其动力学特性.试验结果表明,非线性系统动态参数化模型能准确预测多自由度非线性系统的输出响应,为非线性系统的分析与优化设计提供了理论基础.     

基于有限元分析的重型装备空投系统匹配和评估

针对采用实装试验研究空投装备着陆缓冲过程的方法难以得到有效结果的问题,采用非线性有限元分析方法,结合气体热力学理论,建立重型装备空投系统解析模型和非线性有限元模型,结合试验数据验证模型的有效性和精度.在分析重型装备空投着陆缓冲过程的响应特性基础上,采用拉丁超立方试验设计方法和移动最小二乘法构建响应面模型,结合环境条件的统计规律研究空投系统的匹配和评估方法.分析高海拔条件下空投过程中存在的问题,给出其发生机理并提出解决方案.     

导弹多学科设计优化中响应面方法研究

使用响应面方法取代高精度工程模拟可以减少设计周期和成本。目前,有多种响应面模型应用到导弹多学科设计优化中,每种响应面模型有不同优点和不足。本文的目的是通过比较六种响应面模型来帮助设计者选择合适的响应面模型。七种测试函数用来比较响应面模型的三个方面：近似精度、鲁棒性和应用难易度。结果表明,Kriging响应面和增强径向基函数响应面对线性响应、二次响应和高阶非线性响应都有很好的近似效果,而二次多项式响应面和移动最小二乘响应面适合于线性和二次响应,径向基函数响应面适合于高阶非线性响应。神经网络响应面在使用更多的采样点时得到更精确的模型。     

A systematic approach on computational analysis and optimization design: for a nonlinear coupling shock absorber

The aim of this article is to provide a systematic approach to perform computational simulation and optimization design of parameters matching selection for a nonlinear coupling shock absorber. A theoretical mathematical model with nonlinear coupling for shock absorber is induced based on relative literature. The model considers the coupling of quadratic damping, viscosity damping, coulomb damping and nonlinear spring. Approximate computational solution is deduced by introducing harmonic balance method and Fourier transform method. These approximate theoretical solutions include output response of the system, absolute acceleration transmissibility in vibration or impact, and the maximum relative displacement in impact process, etc. The approximate computational results are compared with those obtained by numerical integration to confirm the validity of the mathematical model. In the meantime, an optimization design model for parameters is built. The design example is illustrated to confirm the validity of the modeling method and the theoretical solution.     

Fuzzy-logic-based process modeling using limited experimental data

Process modeling with limited experimental data is always a difficult task. It becomes even more difficult if the process is highly nonlinear and is characterized by multiple inputs and outputs. Under these circumstances, fuzzy logic may show its capabilities for model development. In this paper, an efficient fuzzy modeling methodology is introduced. The resulting fuzzy model consists of a number of fuzzy implications, each of which is of an IF–THEN form. The IF part consists of a set of logically related antecedents, while the THEN part contains a consequent expressed as a set of linear models. To ensure model simplicity and to accelerate the modeling process, an effective model-development route has been developed. To guarantee the model’s reliability, a t-test-based non-linearity analysis is proposed when each fuzzy implication is developed. The efficacy of the methodology is demonstrated by modeling two nonlinear industrial processes.     

Neural Network Modeling of the Level of Observation Decision in an Acute Psychiatric Ward

Patients in an acute psychiatric ward need to be observed with varying levels of closeness. We report a series of experiments in which neural networks were trained to model this “level of observation” decision. One hundred eighty-seven such clinical decisions were used to train and test the networks which were evaluated by a multitrialv-fold cross-validation procedure. One neural network modeling approach was to break down the decision process into four subproblems, each of which was solved by a perceptron unit. This resulted in a hierarchical perceptron network having a structure that was equivalent to a sparsely connected two-layer perceptron. Neural network approaches were compared with nearest neighbor, linear regression, and naive Bayes classifiers. The hierarchical and sparse neural networks were the most accurate classifiers. This shows that the decision process is nonlinear, that neural nets can be more accurate than other statistical approaches, and that hierarchical decomposition is a useful methodology for neural network design.     

Design of an auto-associative neural network by using design of experiments approach

Advanced monitoring systems enable integration of data-driven algorithms for various tasks, for e.g., control, decision support, fault detection and isolation (FDI), etc. Due to improvement of monitoring systems, statistical or other computational methods can be implemented to real industrial systems. Algorithms which rely on process history data sets are promising for real-time operation especially for online process monitoring tasks, e.g., FDI. However, a reliable FDI system should be robust to uncertainties and small process deviations, thus, false alarms can be avoided. To achieve this, a good model for comparison between process and model is needed and for easier FDI implementation, the model has to be derived directly from process history data. In such cases, model-based FDI approaches are not very practical. In this paper a nonlinear statistical multivariate method (nonlinear principal component analysis) was used for modeling, and realized with auto-associative artificial neural network (AANN). A Taguchi design of experiments (DoE) technique was used and compared with a classic approach, where according to the analysis best AANN model structure was chosen for nonlinear model. Parameters that are important for neural network’s performance have been included into a joint orthogonal array to consider interactions between noise and control process variables. Results are compared to AANN design recommendations by other authors, where obtained nonlinear model was designed for reliable fault detection of very small faults under closed-loop conditions. By using Taguchi DoE robust design on AANN, an improved and reliable FDI scheme was achieved even in case of small faults introduced to the system. The accuracy and performance of AANN and FDI scheme were tested by experiments carried out on a real laboratory hydraulic system, to validate the proposed design for industrial cases.     

Implementation of a human-friendly systems methodology for intelligent control system modelling and simulation

We have been developing an interactive computer software for the systematic support to modeling and simulation of intelligent control systems, based on a human-friendy systems methodology. The support system has a universal application in data analysis, system structuring, statistical and fuzzy modeling, and simulation, with the aid of human-computer interfaces to acquire knowledge or judgment of the domain experts. This paper presents our soft systems methodology and its implementation into the computer to develop intelligent process control systems. New technical proposals include a modeling method of fuzzy implication inference models and a design method of model predictive controllers.     

Constrained multisine input signals for plant-friendly identification of chemical process systems

This paper considers the use of constrained minimum crest factor multisine signals as inputs for plant-friendly identification testing of chemical process systems. The methodology presented here effectively integrates operating restrictions, information-theoretic requirements, and state-of-the-art optimization techniques to design minimum crest factor multisine signals meeting important user-specified time and frequency domain properties. A series of optimization problem formulations relevant to problems in linear, nonlinear, and multivariable system identification are presented; these culminate with their application to the modeling of the Weischedel–McAvoy high-purity distillation column problem, a demanding nonlinear and highly interactive system. The effectiveness of these signals for modeling for control purposes and the ability to incorporate a priori nonlinear models in the signal design procedure are demonstrated in this distillation system case study.     

Simultaneous design and control of processes under uncertainty: A robust modelling approach

This paper presents a new methodology to integrate process design and control. The key idea in this method is to represent the system’s closed-loop nonlinear behaviour as a linear state space model complemented with uncertain model parameters. Then, robust control tools are applied to calculate bounds on the process stability, the process feasibility and the worst-case scenario. The new methodology was applied to the simultaneous design and control of a mixing tank process. The resulting design avoids the solution of computationally intensive dynamic optimizations since the integration of design and control problem is reduced to a nonlinear constrained optimization problem.     

Determining model accuracy of network traces

Accurate network modeling is critical to the design of network protocols. Traditional modeling approaches, such as Discrete Time Markov Chains (DTMC) are limited in their ability to model time-varying characteristics. This problem is exacerbated in the wireless domain, where fading events create extreme burstiness of delays, losses, and errors on wireless links. In this paper, we describe the data preconditioning modeling technique that is capable of capturing the statistical characteristics of wired and wireless network traces. We revise our previous developed data preconditioning modeling algorithm, the Markov-based Trace Analysis (MTA), and present the Multiple states MTA (MMTA) algorithm. Our main contributions are methodologies created to quantify the accuracy of network models, methodology to choose the most accurate model for a given network and characteristic of interest (e.g., delay, loss, or error process), and the validation of our data preconditioning modeling algorithms.     

Linear control of nonlinear systems based on nonlinearity measures

In reality, virtually every process is a nonlinear system. Nevertheless, linear controller design methods have proved to be adequate in many applications. In practice, the linear controller design is usually done disregarding a possible nonlinear plant/linear model mismatch. In this work we introduce a general framework for the development of linear controllers for nonlinear systems based on nonlinearity measures. Nonlinearity measures are tools to assess the extent of a system’s inherent nonlinearity instead of just recognizing a system as being linear or nonlinear. Recent work shows that nonlinearity measures characterize the magnitude of the modeling error when an optimal linear model is used for the nonlinear system. The best linear model can then be used to design a linear controller that robustly stabilizes the linear system in presence of the nonlinear modeling error. A crucial point is that both, the best linear model and the modeling error, are determined for a specified region of operation, thus significantly increasing the class of applicable nonlinear systems. Examples demonstrate the (necessity and) effectiveness of the proposed approach.     

Design of robust gain-scheduled PI controllers for nonlinear processes

Gain-scheduling has proven to be a successful design methodology in many engineering applications. However, in the absence of a sound theoretical analysis, these designs come with no guarantees of robust stability, performance or even nominal stability of the overall gain-scheduled deign.This paper presents such an analysis for one type of nonlinear gain-scheduled control system based on the process input for nonlinear chemical processes. A methodology is also proposed for the design and optimization of the robust gain-scheduled PI controller. Conditions which guarantee robust stability and performance are formulated as a finite set of linear matrix inequalities (LMIs) and hence, the resulting problem is numerically tractable. Issues of modeling error and input-saturation are explicitly incorporated into the analysis. A simulation study of a nonlinear continuous stirred tank reactor (CSTR) process indicates that this approach can produce efficient sub-optimal robust gain-scheduled controllers.     

Large‐signal distributed millimeter‐wave multifinger pHEMT modeling using time‐domain technique

The finite‐difference time‐domain (FDTD) method is used for the large‐signal modeling of a multifinger pHEMT, which is considered as five nonlinear coupled distributed transmission lines. The developed model, which is based on the exact physical layout of multifinger pHEMT, not only accurately describes the propagation effects along the electrodes at higher frequencies but it also includes major nonlinearities of the I–V and Q–V characteristics. Using the transmission line theory, a proper nonlinear equivalent lumped circuit model is allocated for the differential length of the quintuple‐line transistor and the nonlinear active multiconductor transmission line (NAMCTL) equations are derived. These nonlinear, coupled differential equations are numerically solved using the FDTD method. The proposed model is applied to a 100 nm GaAs pHEMT and the simulation results are compared with the results of conventional sliced model in Keysight ADS simulator. The developed transient nonlinear model accurately predicts both the S‐parameters (1–150 GHz) and large‐signal power performances especially at millimeter wave frequency range. The proposed model can be useful in design and analysis of various types of high‐frequency nonlinear integrated circuits.     

一种基于工况分解的热工过程非线性控制模型建立方法及应用

充分考虑大多数复杂热工控制对象非线性特性与运行工况密切相关的实际特点,采用基于工况分解的多模型建模思路,提出一种面向控制的非线性过程建模方法.将该方法应用于某电厂300MW机组锅炉过热汽温对象,实际考核结果表明采用该方法建立的模型,即使在运行工况大范围变化时也具有满意的动态预测效果,验证了提出的方法的有效性.