A grey box model-based method for fault diagnosis is proposed in this paper. The method is based on a first principle model of the process unit: a heat exchanger, and on a grey box model of the faults: the deteriorati...
详细信息
A grey box model-based method for fault diagnosis is proposed in this paper. The method is based on a first principle model of the process unit: a heat exchanger, and on a grey box model of the faults: the deterioration of the heat transfer surface by aging and the leaking of the outer container. The deterioration of the heat transfer surface is due to material settling, and in old heat exchangers pieces of this material can break off and cause damage in the equipment. A recursive least squares estimator with forgetting factor is used to track both the heat transfer coefficients and the cold side volume. The heat transfer coefficients are estimated both from the hot and cold side equations. The settled material breakage fault is detected via detection of abrupt positive jumps in the estimated heat transfer coefficients using a detector based on a cumulative sum (CUSUM) test. The proposed method enables simultaneous detection of the two fault types considered. Fault localization along the equipment length is also possible when temperature measurements are available along the length.
A novel approach is proposed to the state estimation of a class of nonlinear systems which consist of known linear part and unknown nonlinear part. A linear observer is first designed then a nonlinear compensation ter...
详细信息
A novel approach is proposed to the state estimation of a class of nonlinear systems which consist of known linear part and unknown nonlinear part. A linear observer is first designed then a nonlinear compensation term in the nonlinear observer is determined using the proposed "deconvolution method". The B-spline neural network is used to model the estimated compensation term. Three simulation examples are given to compare the effectiveness of the proposed approach and some analytical approaches.
In the training of B-spline networks, iterative gradient method with a constant learning rate are often used. It is well-known that the training speed depends on the choice of the learning rate, yet few guidelines in ...
In the training of B-spline networks, iterative gradient method with a constant learning rate are often used. It is well-known that the training speed depends on the choice of the learning rate, yet few guidelines in the selection of a suitable learning rate are available in the literature. In this paper, an adaptive learning rate to update the weights of a B-spline network with a scalar or multi-output is proposed. It is shown that under certain conditions, the performance index for a training algorithm using the proposed adaptive learning rate converges to a constant as the number of iterations increases. Also, a method for computing the criterion for terminating the training is presented. Simulation examples are presented, showing that training of the networks using the adaptive training is much faster than that using a constant learning rate.
Discrete-time control schemes are often used for realization of positional control strategies for continuous time control problems (Krasovskii and Subbotin, 1974; Osipov, 1971) . However for general systems with delay...
详细信息
Discrete-time control schemes are often used for realization of positional control strategies for continuous time control problems (Krasovskii and Subbotin, 1974; Osipov, 1971) . However for general systems with delays (whicb are often called functional differential equations (FDE)) exact solutions in both continuous and discrete schemes can be found in exceptional cases. So, as rule, it is necessary to solve corresponding control problems using suitable numerical algorithms . In surveys (Cryer and Tavernini, 1972; Hall and watt, 1976; Bellen, 1985) different numerical methods for FDE are discussed. In this paper positional implicit Runge-Kutta-like numerical methods of solving general FDE are presented. The elaborated methods are direct analogs of the corresponding numerical methods of ODE case. Also we present new theorem (which generalizes the results of (Kim and Pimenov, 1997; Kim and Pimenov, 1998)) on convergence order. The obtained results are based on distinguishing of finite dimensional and infinite dimensional components in the structure of FDE, on interpolation and extrapolation of the discrete prehistory of the model, on application of constructions of i- smooth analysis (Kim, 1996) . The notion of an approximation order for an optimal strategy in the discrete scheme of realization is introduced. The conditions, which guarantee required approximation order of the optimal strategy, are presented.
This paper considers the abstract features of human/machine interaction systems that are required for the production of intelligent behaviour. A conceptual architecture is proposed for a subset of intelligent systems ...
详细信息
This paper considers the abstract features of human/machine interaction systems that are required for the production of intelligent behaviour. A conceptual architecture is proposed for a subset of intelligent systems called human-centred intelligent systems (HCISs) and it is argued that such systems must be autonomous, robust and adaptive in order to be intelligent. Soft computing is proposed as a promising new technique that can be used to build HCISs, and examples are presented where this is already being done. Finally, flexibility is defined to be a combination of the often-conflicting requirements of robustness and adaptability, and it is argued that the right balance between these two features is necessary to achieve intelligent behaviour.
A prototype large electrical machine running on active magnetic bearings is described. This rig is controlled by a digital signal processor connected by a custom interface to MATLAB/Simulink hosted by a PC. The on-lin...
详细信息
A prototype large electrical machine running on active magnetic bearings is described. This rig is controlled by a digital signal processor connected by a custom interface to MATLAB/Simulink hosted by a PC. The on-line tuning of a PID controller is set up as an optimisation problem from MATLAB and a multiobjective genetic algorithm is used to drive the optimisation. The results of an optimisation are presented and analysed.
Research in human/computer interaction has mainly focused on natural language, text, speech and vision primarily in isolation. Recently there have been a number of research projects that have concentrated on the integ...
详细信息
Research in human/computer interaction has mainly focused on natural language, text, speech and vision primarily in isolation. Recently there have been a number of research projects that have concentrated on the integration of such modalities using intelligent reasoners. The rationale is that many inherent ambiguities in single modes of communication can be resolved if extra information is available. This paper describes an intelligent multi-modal system called the Smart Work Manager. The main characteristics of the Smart Work Manager are that it can process speech, text, face images, gaze information and simulated gestures rising the mouse as input modalities, and its output is in the form of speech, text or graphics. The main components of the system are the reasoner, a speech system, a vision system, an integration platform and the application interface. The overall architecture of the system will be described together with the integration platform and the components of the system which include a non-intrusive neural network based gaze-tracking system. The paper concludes with a discussion on the applicability of such systems to intelligent human/computer interaction and lessons learnt in terms of reliability and efficiency.
In this paper we present a new classification and image segmentation system based on the addition of a variational method to a classic clustering algorithm. This system constitutes an improvement respect traditional s...
详细信息
A block diagonal form about a nonlinear system is defined. Based on the de finition, a design method ca1led block diagonal controller (BDC) is proPOsed bo feedbacklinearization. Thus, a linearization design of a class...
详细信息
A block diagonal form about a nonlinear system is defined. Based on the de finition, a design method ca1led block diagonal controller (BDC) is proPOsed bo feedbacklinearization. Thus, a linearization design of a class of nonlinear system can be simply re-alized. The result of design has been proved by mathematical simulation of a certain anti-ship missile.
暂无评论