Data-driven realizations of kernel and image representations and their application to fault detection and control system design

This paper deals with the data-driven design of observer-based fault detection and control systems. We first introduce the definitions of the data-driven forms of kernel and image representations. It is followed by the study of their identification. In the context of a fault-tolerant architecture, the design of observer-based fault detection, feed-forward and feedback control systems are addressed based on the data-driven realization of the kernel and image representations. Finally, the main results are demonstrated on the laboratory continuous stirred tank heater (CSTH) system.     

故障检测中核参数优化方法性能评估

近年来,基于核主元分析与核偏最小二乘的方法经常被应用于过程监控与故障检测领域以克服工业过程的非线性.研究发现此类方法的检测性能很大程度上受核参数的影响,而目前学术界对该参数的优化方法研究较少.因此,本文以最常用的高斯核方法为例,首先总结了3类常用的核参数优化方法:二分法、基于BP神经网络的重构法和基于样本分类的重构法;其次重点分析每个方法的特点和它们之间的联系,并评估它们的性能;最后将上述方法设计成一个核参数优化系统应用于热连轧过程的故障检测中.应用结果表明,优化后的核参数能显著提高故障检测性能.     

Syllabification rules versus data-driven methods in a language with low syllabic complexity: The case of Italian

Linguistic rules have been assumed to be the best technique for determining the syllabification of unknown words. This has recently been challenged for the English language where data-driven algorithms have been shown to outperform rule-based methods. It may be possible, however, that data-driven methods are only better for languages with complex syllable structures. In this study, three rule-based automatic syllabification systems and two data-driven automatic syllabification systems (Syllabification by Analogy and the Look-Up Procedure) are compared on a language with lower syllabic complexity – Italian. Comparing the performance using a lexicon containing 44,720 words, the best data-driven algorithm (Syllabification by Analogy) achieved 97.70% word accuracy while the best rule set correctly syllabified 89.77% words. These results show that data-driven methods can also outperform rule-based methods on Italian syllabification, a language of low syllabic complexity.     

Fuzzy systems design: direct and indirect approaches

A systematic classification of the data-driven approaches for design of fuzzy systems is given in the paper. The possible ways to solve this modelling and identification problem are classified on the basis of the optimisation techniques used for this purpose. One algorithm for each of the two basic categories of design methods is presented and its advantages and disadvantages are discussed. Both types of algorithms are self-learning and do not require interaction during the process of fuzzy model design. They perform adaptation of both the fuzzy model structure (rule-base) and the parameters. The indirect approach exploits the dual nature of Takagi-Sugeno (TS) models and is based on recently introduced recursive clustering combined with Kalman filtering-based procedure for recursive estimation of the parameter of the local sub-models. Both algorithms result in finding compact and transparent fuzzy models. The direct approach solves the optimisation problem directly, while the indirect one decomposes the original problem into on-line clustering and recursive estimation problems and finds a sub-optimal solution in real-time. The later one is computationally very efficient and has a range of potential applications in real-time process control, moving images recognition, autonomous systems design etc. It is extended in this paper for the case of multi-input–multi-output (MIMO systems). Both approaches have been tested with real data from an engineering process.     

行程时间预测方法研究

行程时间预测,有助于实施高级旅行者信息系统.自20世纪90年代起,已经有多种行程时间预测方法被研发出来.将行程时间预测方法分为模型驱动方法和数据驱动方法两大类.介绍了两种常见的模型驱动方法,即排队论模型和细胞传输模型.数据驱动方法被分类为参数方法和非参数方法:参数方法包括线性回归、自回归集成移动平均和卡尔曼滤波,非参数方法包括神经网络、支持向量回归、最近邻和集成学习方法.对现有行程时间预测方法从源数据、预测范围、准确率、优缺点和适用范围等方面进行了分析总结.针对现有方法的一些缺点,提出了可能的解决方案.给出了一种新颖的数据预处理框架和一个行程时间预测模型,最后指出了未来的研究方向.     

Data-driven output-feedback fault-tolerant control for unknown dynamic systems with faults changing system dynamics

This paper studies the data-driven output-feedback fault-tolerant control (FTC) problem for unknown dynamic systems with faults changing system dynamics. In a framework of active FTC, two basic issues are addressed: the fault detection employing only the measured input–output information; the controller reconfiguration to achieve optimal output-feedback control in the presence of multiple faults. To detect faults and write the system state via the input–output data, an approach to data-driven design of a residual generator with a full-rank transformation matrix is presented. An output-feedback approximate dynamic programming method is developed to solve the optimal control problem under the condition that the unknown linear time-invariant discrete-time plant has multiple outputs. According to the above results and the proposed input–output data-based value function approximation structure of time-varying plants, a model-free output-feedback FTC scheme considering optimal performance is given. Finally, two numerical examples and a practical example of a DC motor control system are used to demonstrate the effectiveness of the proposed methods.     

Data-driven parallel production systems

Much effort has been expended on developing special architectures dedicated to the efficient execution of production systems. While data-flow principles of execution offer the promise of high programmability for numerical computations, it is shown that the data-driven principles can also be applied to symbolic computations. In particular, a mapping of the RETE match algorithm along the line of production systems is considered. Bottlenecks of the RETE match algorithm in a multiprocessor environment are identified and possible solutions are suggested. The modifications to the actor set as well as the program graph design are shown for execution on the tagged data-flow computer. The results of a deterministic simulation of this multiprocessor architecture demonstrate that artificial intelligence production systems can be efficiently mapped on data-driven architectures     

Data-driven optimal switching and control of switched systems

In this paper, optimal switching and control approaches are investigated for switched systems with infinite-horizon cost functions and unknown continuous-time subsystems. At first, for switched systems with autonomous subsystems, the optimal solution based on the finite-horizon HJB equation is proposed and a data-driven optimal switching algorithm is designed. Then, for the switched systems with subsystem inputs, a data-driven optimal control approach based on the finite-horizon HJB equation is proposed. The data-driven approaches approximate the optimal solutions online by means of the system state data instead of the subsystem models. Moreover, the convergence of the two approaches is analyzed. Finally, the validity of the two approaches is demonstrated by simulation examples.     

Understanding Data-Driven Decision Support Systems

It is important for managers and Information Technology professionals to understand data-driven decision support systems and how such systems can provide business intelligence and performance monitoring. Data-driven DSS is one of five major types of computerized decision support systems and the features of such systems vary across specific implementations. Different development packages also impact the capabilities of data-driven DSS and hence criteria for evaluating data-driven DSS development software are important to understand. Overall, this article builds on an historic foundation of prior decision support systems theory.     

Data-driven predictive control for continuous-time linear parameter varying systems with application to wind turbine

A new data-driven predictive control method based on subspace identification for continuous-time linear parameter varying (LPV) systems is presented in this paper. It is developed by reformulating the continuous-time LPV system which utilizes Laguerre filters to obtain the subspace prediction of output. The subspace predictors are derived by QR decomposition of input-output and Laguerre matrices obtained by input-output data. The predictors are then applied to design the model predictive controller. It is shown that the integrated action is incorporated in the control effect to eliminate the steady-state offset. We control the continuous-time LPV systems to obtain the attractive performance with the proposed data-driven predictive control method. The proposed controller is applied to a wind turbine to verify its effectiveness and feasibility.     

Data-driven shuttle service design for sustainable last mile transportation

Last mile transportation is important in both freight and passenger transport as it accounts for a large portion of the costs and emissions in the transportation industry. In urban transport, the continuously growing travel demands and the rapid development of mass transit systems place a high stress on last mile transportation, which is a vital but underdeveloped part of urban transportation systems. This underdevelopment greatly impedes the further improvement of bus sharing rates and the realisation of sustainable transportation. Therefore, this research proposes a data-driven method to design shuttle services to improve the efficiency and convenience of last mile transportation. Specifically, a unified tool is developed to identify the last mile travel demands from various data sources. Based on these demands, the locations of bus stop are planned through an improved clustering algorithm, and the routing and scheduling of shuttle services are designed using a data-driven method. In addition, a simulation-based cost-benefit analysis is conducted to evaluate the performances of shuttle services in different areas. Finally, a case study using bicycle-sharing data in Shanghai is presented to demonstrate the working process of the proposed method and verify its performance.     

A hierarchical expert system for computer process control

This paper reports on the design and implementation of an expert system for computer process control (HESCPC). The complexity of the expertise necessary for computer process control applications requires that the expert system architecture be structured into a hierarchy of classes of specialized experts. The architecture of HESCPC integrates four classes of expert systems: operator/manager companion expert class, control system algorithm design expert class, hardware expert class, and software expert class. The paper is concerned with the design and implementation of the general system architecture, an operator adviser expert for a nuclear power plant and a control system designer expert using a state space feedback approach. Although the design and implementation aspects of all experts are discussed, the emphasis is on the latter.

At this stage of the HESCPC development, the declarative knowledge represented by 423 metarules and 1261 rules is distributed on a hierarchical structure among 20 experts on different levels of the hierarchy which are able to communicate among themselves to solve difficult control problems.

Examples of control system design sessions of linear mono and multivariable systems using feedback state space approach are given. A run time of an operator-adviser data-driven expert system for a nuclear plant is also presented.     

A probabilistic principal component analysis-based approach in process monitoring and fault diagnosis with application in wastewater treatment plant

Probabilistic principal component analysis (PPCA) based approaches have been widely used in the field of process monitoring. However, the traditional PPCA approach is still limited to linear dimensionality reduction. Although the nonlinear projection model of PPCA can be obtained by Gaussian process mapping, the model still lacks robustness and is susceptible to process noise. Therefore, this paper proposes a new nonlinear process monitoring and fault diagnosis approach based on the Bayesian Gaussian latent variable model (Bay-GPLVM). Bay-GPLVM can obtain the posterior distribution rather than point estimation for latent variables, so the model is more robust. Two monitoring statistics corresponding to latent space and residual space are constructed for PM-FD purpose. Further, the cause of fault is analyzed by calculating the gradient value of the variable at the fault point. Compared with several PPCA-based monitoring approaches in theory and practical application, the Bay-GPLVM-based process monitoring approach can better deal with nonlinear processes and show high efficiency in process monitoring.     

Data-driven,using design of dynamic experiments,versus model-driven optimization of batch crystallization processes

A new data-driven experimental design methodology, design of dynamic experiments (DoDE), is proposed as a means of developing a response surface model that can be used to effectively optimize batch crystallization processes. This data-driven approach is especially useful for complex processes for which it is difficult or impossible to develop a knowledge-driven model in a timely fashion for the optimization of an industrial process. Design of dynamic experiments [1] generalizes the formulation of time-invariant design variables from design of experiments, allowing for consideration of time-variant design variables in the experimental design. When combined with response surface modeling and an appropriate optimization algorithm, a data-driven optimization methodology is produced, which we call DoDE optimization. The method is used here to determine the optimal cooling rate profile, which integrates to give the optimum temperature profile, for a batch crystallization process. To examine the effectiveness of the DoDE optimization method, the data-driven optimum temperature profile is compared to the optimum temperature profile obtained using a model-based optimization technique for the potassium nitrate–water batch crystallization model developed by Miller and Rawlings [2]. The temperature profiles calculated using DoDE optimization yield response values within a few percent of the true model-based optimum values. A sensitivity analysis is performed on one case study to evaluate the distribution of the response variable from each method in the presence of parameter and initial seed distribution variability. It is demonstrated that there is partial overlap in the distributions when only variability in the model parameters is evaluated and there is substantial overlap when variability is included in both the model and initial seed distribution parameters. From this evidence, it can be concluded that the DoDE optimization method has the potential to be a useful data-driven optimization tool for batch crystallization processes where a first-principles model is not available or cannot be developed due to time and/or cost constraints.     

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.     

Discovering and Using Design Patterns in the WWW

In this paper, we discuss how the idea of design patterns can be used in the context of the World Wide Web, for both designing and implementing web sites or more complex information systems. We first motivate our work by discussing which are the most outstanding problems in designing Web-based information systems. Then we briefly introduce design patterns and show how they are used to record and reuse design information. We next present some simple though powerful design patterns and show known uses in the WWW. Finally, we outline a process for building applications by combining a design methodology (OOHDM) with design patterns.     

Data-driven safe gain-scheduling control

Data-based safe gain-scheduling controllers are presented for discrete-time linear parameter-varying systems (LPV) with polytopic models. First, $\lambda$-contractivity conditions are provided under which the safety and stability of the LPV systems are unified through Minkowski functions of the safe sets. Then, a data-based representation of the closed-loop LPV system is provided, which requires less restrictive data richness conditions than identifying the system dynamics. This sample-efficient closed-loop data-based representation is leveraged to design data-driven gain-scheduling controllers that guarantee $\lambda$-contractivity and, thus, invariance of the safe sets. It is also shown that the problem of designing a data-driven gain-scheduling controller for a polyhedral (ellipsoidal) safe set amounts to a linear program (a semi-definite program). The motivation behind direct learning of a safe controller is that identifying an LPV system requires satisfying the persistence of the excitation (PE) condition. It is shown in this paper, however, that directly learning a safe controller and bypassing the system identification can be achieved without satisfying the PE condition. This data-richness reduction is of vital importance, especially for LPV systems that are open-loop unstable, and collecting rich samples to satisfy the PE condition can jeopardize their safety. A simulation example is provided to show the effectiveness of the presented approach.     

基于感知推理方法的语言动力系统性质及数据驱动设计研究(英文)

The linguistic dynamic systems(LDSs) based on type-1 fuzzy sets can provide a powerful tool for modeling, analysis,evaluation and control of complex systems. However, as pointed out in earlier studies, it is much more reasonable to take type-2fuzzy sets to model the existing uncertainties of linguistic words. In this paper, the LDS based on type-2 fuzzy sets is studied, and its reasoning process is realized through the perceptual reasoning method. The properties of the perceptual reasoning method based LDS(PR-LDS) are explored. These properties demonstrated that the output of PR-LDS is intuitive and the computation complexity can be reduced when the consequent type-2 fuzzy numbers in the rule base satisfy some conditions. Further, a data driven method for the design of the PR-LDS is provided. At last, the effectiveness and rationality of the proposed data-driven method are verified by an example.     

复杂产品协同设计中数据建模与驱动方法

针对在传统基于工作流的协同设计中,不同专业设计人员交流和任务协调困难导致产品设计效率低下的问题,提出复杂产品"一元三层"数据模型和基于数据驱动的复杂产品协同设计技术。首先采用多维多粒度的数据建模和本体描述方法完成了对复杂产品的信息建模,然后采用基于本体的语义检索技术完成协同设计过程任务的数据订阅,最后实现基于数据订阅/发布的复杂产品任务协同技术。实验结果表明,基于数据驱动的复杂产品协同设计技术解决了传统协同设计过程中不同专业设计人员之间交流与任务协调的困难,实现复杂产品协同设计过程的螺旋式上升,从而提高了产品设计效率。     

Control loop diagnosis with ambiguous historical operating modes: Part 2, information synthesis based on proportional parametrization

Control loop diagnosis has become an increasingly important tool for improving the efficiency, reliability and safety for a variety of processes. While a number of model-based diagnosis methods have been proposed, constructing models may be a difficult task. An alternative approach is to use data-driven control-loop diagnosis, a family of diagnosis methods that make use of historical data for training the diagnostic models. Bayesian methods have been applied to data-driven control loop diagnosis to combine prior process knowledge with historical data, and can be used to assign probabilities to different modes (or operation statuses) after combination. However, one difficulty with Bayesian methods is that there must be exact knowledge of the underlying mode so that the corresponding monitor readings in the historical data can be used. If there is uncertainty about the underlying mode, the mode becomes ambiguous, which Bayesian methods do not deal with. An alternative method is proposed in this paper that exploits the properties of data-driven Bayesian methods, and can be applied for diagnosis in the presence of ambiguity. The proposed method is evaluated through simulation examples as well as applied to industrial process data.