This paper presents results to date from on-going analytical and experimental investigations of passive and active control methods for suppression of stall in axial flow compressors. The passive control method involve...
详细信息
This paper presents results to date from on-going analytical and experimental investigations of passive and active control methods for suppression of stall in axial flow compressors. The passive control method involved inclusion of flow separators in the inlet and outlet ducts of the compressor. The active control methods investigated included bleed valve modulations, flow recirculation from plenum bleed to inlet of the compressor and fuel flow modulations. It was found that in contrast to theoretical expectations, the use of separators in the inlet and/or outlet ducts did not prevent the onset of rotating stall. However, it was found that with proper number of separators, the onset manner of the rotating stall was tempered. Experimental results to date indicate that it is possible to significantly suppress the onset of rotating stall through a combination of passive and active control schemes, and thus, extend the useful operational range of a compressor.
Stringent emission regulations combined with customer demands for improved fuel economy and performance have forced the automotive industry to consider more advanced powertrain configurations than standard port-fuel i...
详细信息
Stringent emission regulations combined with customer demands for improved fuel economy and performance have forced the automotive industry to consider more advanced powertrain configurations than standard port-fuel injected gasoline engines. Modern state-of-the-art powertrain systems may combine several power sources (internal combustion engines, electric motors, fuel cells, etc.) and various exhaust aftertreatment devices (catalytic converters, lean NOx traps, particulate filters, etc.) in addition to conventional engine subsystems such as turbochargers and exhaust gas recirculation. The determination of the way in which these systems need be operated to meet driver's torque demand, performance and fuel economy expectations while satisfying federal emission regulations is a complex and a multiobjective optimal control problem. This paper reviews some of the approaches to this problem in the context of two case studies.
A model-based recurrent neural network (MBRNN) is introduced for modeling nonlinear dynamic systems. The topology of MBRNN as well as its initial weights are defined according to the linearized state-space model of th...
详细信息
A model-based recurrent neural network (MBRNN) is introduced for modeling nonlinear dynamic systems. The topology of MBRNN as well as its initial weights are defined according to the linearized state-space model of the plant. As such, the MBRNN has the ability to incorporate the analytical knowledge of the plant in its formulation. With its original topology intact, the MBRNN can then be trained to represent the plant nonlinearities through modifying its nodes' activation functions, which consist of contours of Gaussian radial basis functions (RBFs). Training involves adjusting the weights of the RBFs so as to modify the contours representing the activation functions. The performance of the MBRNN is demonstrated via several examples. The results indicate that it requires much shorter training than needed by ordinary recurrent networks. This training efficiency is attributed to the MBRNN's fixed topology, which is independent of training.
An optimal control based algorithm, using a linear quadratic state-feedback regulator (LQR) on a linear model, has been developed which identifies viable proportional-integral with feedforward (PI/FF) control strategi...
详细信息
An optimal control based algorithm, using a linear quadratic state-feedback regulator (LQR) on a linear model, has been developed which identifies viable proportional-integral with feedforward (PI/FF) control strategies. In this approach, two matrices, SFB and SFF, are calculated from the linear state space matrices and the optimal control gain. The SFB and SFF reveal the respective dominant feedback and ratio/feedforward paths of a system. Three screening and four performance heuristics have been developed to interpret these matrices to reveal the dominant control pathways of a system. Information gained from these matrices is used to formulate PI/FF control architecture(s). These architectures are tested on a fixed dynamic simulation to judge their performance. A continuous polymerization process was used as test case to evaluate the results of the optimal control based algorithm and to test the heuristics presented here. The algorithm identified weak manipulated variables and eliminated them from the list of degrees of freedom. It also identified viable potential control architectures that were tested with the fixed dynamic simulations. The algorithm provided information on each candidate architecture's simplicity and degree of decoupled, feedback-path dominance. These results, for each case, were confirmed through simulations.
A new learning algorithm suitable for pattern classification in machine condition health monitoring based on fuzzy neural networks called an `Incremental Learning Fuzzy Neuron Network'(ILFN) has been developed. Th...
详细信息
A new learning algorithm suitable for pattern classification in machine condition health monitoring based on fuzzy neural networks called an `Incremental Learning Fuzzy Neuron Network'(ILFN) has been developed. The ILFN, using Gaussian neurons to represent the distributions of the input space, is an on-line, one-pass, and incremental learning algorithm. The network is a self-organized classifier with the ability to adaptively learn new information without forgetting old knowledge. To prove the concept, the simulations have been performed with the vibration data known as Westland vibration data set. Furthermore, the classification performance of the network has been tested on other benchmark data sets, such as, the Fisher's Iris data, and a vowel data set. For the generalization capability, comparison studies among other well-known classifiers were performed and the ILFN was found competitive with or even superior to many existing classifiers. Additionally, the ILFN uses far less training time than conventional classifiers.
Condition monitoring of dynamic systems based on vibration signatures has generally relied upon Fourier based analysis as a means of translating vibration signals in time domain into the frequency domain. However, Fou...
详细信息
Condition monitoring of dynamic systems based on vibration signatures has generally relied upon Fourier based analysis as a means of translating vibration signals in time domain into the frequency domain. However, Fourier analysis provided a poor representation of signals well localized in time. In this case, it is difficult to detect and identify the signal pattern from their expansion coefficients because the information is diluted across the whole basis. The wavelet packet transform is introduced as an alternative means of extracting time-frequency information from vibration signature. Moreover, with the aid of statistical based feature selection criteria, a lot of feature components containing little discriminant information could be discarded resulting in a feature subset with reduced number of parameters. This significantly reduces the long training time that is often associated with neural network classifier and increases the generalization ability of the neural network classifier.
A reduced order output feedback control algorithm based on a low order modeling methodology for suppressing the vibration and obtaining the high speed, high precision positioning of the flexible robot arm is proposed ...
详细信息
A reduced order output feedback control algorithm based on a low order modeling methodology for suppressing the vibration and obtaining the high speed, high precision positioning of the flexible robot arm is proposed and discussed. The proposed low order modeling methodology gives no modal truncation error and makes it possible to vary the mode shapes of the link according to the change of the boundary conditions due to the feedback gains and the payload at the tip of the arm. Furthermore, this methodology can take into account not only the flexibility of links but also the flexibility of joints. The reduced order output feedback control algorithm is designed by the combination of sliding mode controller and suboptimal output feedback controller so as to achieve both the damping of the system vibration and the stabilization at the same time. The results of numerical simulation and of experiment for a two-link robot arm model prove that the proposed modeling methodology is able to describe the dynamic behaviors of the arm, and the controller designed applying the suboptimal output feedback sliding mode control is able to control the arm quite well.
This paper presents an integrated approach to magnetic suspension of a tubular beam for non-contact processing. We describe noncontact sensors and actuators, structure modelling/identification, and control methods for...
详细信息
This paper presents an integrated approach to magnetic suspension of a tubular beam for non-contact processing. We describe noncontact sensors and actuators, structure modelling/identification, and control methods for such systems. We have also designed an experiment that uses 8 sensors and 8 actuators to suspend a 3 m long, 6.35 mm diameter steel tube. Results from this experiment are presented herein. To facilitate the experiment, we designed a novel two-dimensional position sensor by utilizing the concept of an x-y variable differential transformer. We also designed two types of actuators that apply magnetic force on the tube in two axes. The system dynamics are modelled theoretically, and identified experimentally from within the closed-loop system. In order to control the system under varying boundary conditions, we propose a slow roll-up lead compensator. We also introduce sensor interpolation and sensor averaging methods to make the control loop more robust. Our proposed ideas are verified in the experimental results.
We consider run-to-run feedforward control from lithography to etch in wafer patterning. Under feedforward control, measured deviations in photoresist linewidth are corrected automatically by adjustment of etch bias. ...
详细信息
We consider run-to-run feedforward control from lithography to etch in wafer patterning. Under feedforward control, measured deviations in photoresist linewidth are corrected automatically by adjustment of etch bias. Two potential problems with feedforward control are listed: (1) the possibility of feeding measurement noise forward exists and could result in increased final CD variability, and (2) it may be difficult to adjust etch bias without disturbing other important etch characteristics. First, a controller design that addresses concern (1) is considered. Models of the lithography and measurement errors along with minimum mean square error estimation are employed to reduce the chances of acting on erroneous measurements. Building on the first, a second controller is then designed to address concern (2). The controller selects a recipe from a pre-qualified set rather than making arbitrary etch adjustments. It is assumed the recipes in the set provide a means for adjusting etch bias while guaranteeing acceptable results for other characteristics. Both controllers are simulated using industrial data.
A high performance, high power dynamometer system is proposed. By using a field controlled DC generator (rather than a DC motor) 10-200 HP, high bandwidth loads can be accurately emulated. The implementation is both e...
详细信息
A high performance, high power dynamometer system is proposed. By using a field controlled DC generator (rather than a DC motor) 10-200 HP, high bandwidth loads can be accurately emulated. The implementation is both economical to build and to operate. Because power loads are sometimes needed (for instance, they are required when testing motor efficiency), the approach emulates power loads, rather than the commonly emulated torque loads. By emphasizing only the pertinent dynamics, a simplified model appropriate for control is found, but the model is still nonlinear and parameter dependent. A linearizing control which also measures parameter changes without the need for specialized instruments is proposed. The control applies to machines which maintain nearly a constant velocity over large loads, so it is appropriate for induction motors, synchronous motors, etc. Experimental tests using both standardized motor efficiency trajectories and power measurements from an oilfield pump jack confirm the system's ability to respond as quickly and as accurately as the best methods previously proposed, but at a fraction of the implementation and operational costs.
暂无评论