This work presents pattern recognition-based methods for controller adaptation and performance evaluation. These methods comprise a passive model-based adaptive control algorithm that is simple to use, easy to underst...
详细信息
This work presents pattern recognition-based methods for controller adaptation and performance evaluation. These methods comprise a passive model-based adaptive control algorithm that is simple to use, easy to understand, stable, and fairly robust in a wide variety of applications. controller adaptation in this work uses excitation diagnostics to initiate batch-wise regression of a process model to dynamic closed-loop process data. The process model is then employed in model-based controller tuning relations to update the controller's character. controller performance evaluation is used to determine appropriate adjustments to the tuning relations such that an accurate process model will produce desired controller performance. These adaptive techniques are implemented using vector quantizing neural networks as efficient pattern recognition tools. The adaptive algorithm is presented in a structure that allows for the implementation of these advanced techniques without requiring the replacement of an existing feedback controller. This is demonstrated using a simulated nonlinear third order process and an IMC tuned PI controller with Smith Predictor.
A self-adjusting diagnostic system for monitoring the frequency trimming process is described. The SAFUDS (self-adjusting fuzzy diagnostic system) is capable of adjusting its knowledge to the changes with two incorpor...
详细信息
A self-adjusting diagnostic system for monitoring the frequency trimming process is described. The SAFUDS (self-adjusting fuzzy diagnostic system) is capable of adjusting its knowledge to the changes with two incorporated neural networks serving as a convenient online process monitoring system. The development of the SAFUDS not only fulfills the requirement of automating the diagnostics of the trimming process, but it also demonstrates the advantage of fuzzy logic over conventional logic in implementing the system. The use of fuzzy logic avoids the rigidity of conventional reasoning where it is difficult to process information that is definite or imprecise. One of the major advantages of the fuzzy monitoring strategy lies in the intelligibility and flexibility by which the process condition and control actions can be described directly from the experience and advice of an expert. With fuzzy logic, the system is able to heuristically interpret the condition of the process, like human operators. After the diagnostic system had been implemented for six months, the production yield rate of the frequency trimming process improved from 85% to 95%. During the implementation of the system, the knowledge of the production engineer as well as of the operators regarding the process improved.< >
We describe a decentralized control method that augments Boolean logic control to provide greatly enhanced diagnostics and monitoring of systems with discrete-valued sensors and actuators. Unlike conventional Boolean ...
详细信息
We describe a decentralized control method that augments Boolean logic control to provide greatly enhanced diagnostics and monitoring of systems with discrete-valued sensors and actuators. Unlike conventional Boolean logic where logic is associated with the actuators, the new method defines a behavior and timing model for both sensors and actuators (called control Elements). The behavior model is defined by a sequence of events called the event signature, in a control Element's (CE) neighborhood. The event signature encodes a subset of the process state into each CE. We describe a method for the continuous evaluation of this signature to compute a measure of the likelihood of a CE to change state. This likelihood measure called an Expectation Function, is used to check and enforce the correct behavior of a CE. Also, each CE learns a statistical temporal model in real time. The temporal model predicts delays in the states of the CE as a function of the previous delays and the current delays in the evaluation of its event signature. The distributed behavior model and the on-line estimated timing model are applied to a simulated bottle filling station to illustrate the detection and-diagnosis of incorrect behavior and failures of control Elements.
A new approach to detect and identify faults in complex processes is proposed. The approach is based on a hierarchical neural network structure. Other neural network applications in process fault diagnostics provide o...
详细信息
A new approach to detect and identify faults in complex processes is proposed. The approach is based on a hierarchical neural network structure. Other neural network applications in process fault diagnostics provide only fault detection and isolation. Through the proposed scheme, fault detection, isolation, and identification (recognizing the size of fault) can be achieved. This is due to the higher learning ability of the hierarchical structure. The performance of the suggested fault detector and identifier is evaluated via an industrial case study. The results show a satisfactory level of accuracy.< >
A three-level process monitoring and diagnostics system has been developed for a plasma etching cell. The physical and logical layout of this system is designed to co-exist with a proposed hierarchical cell control sy...
详细信息
ISBN:
(纸本)0780301528
A three-level process monitoring and diagnostics system has been developed for a plasma etching cell. The physical and logical layout of this system is designed to co-exist with a proposed hierarchical cell control system. The cell control system design incorporates generic cell controller concepts and some of the SEMI Generic Equipment Model (GEM) specifications. At the lowest level of the system, equipment and process parameter monitoring is achieved. An AT-compatible computer serves as an information collection/router agent as well as an interpreter (gateway) between the SECS/RS-232 formatted messages received from the equipment controller system and the higher-level expert systems commands. At the top of the hierarchy, an expert systems package that resides on a SUN SPARC station and operates on an X-Windows environment collects information on the equipment and process from the AT compatible. The many system capabilities that result include: (1) pseudo real-time equipment and process monitoring, (2) equipment diagnostics and alarm reporting, (3) alarm logging;and (4) real-time equipment control.
This paper will discuss and show applications of optical emission spectroscopy to the monitoring and control of physical vapor deposition. It is the intent of this paper to render a tutorial on the technology of optic...
详细信息
ISBN:
(纸本)1878068113
This paper will discuss and show applications of optical emission spectroscopy to the monitoring and control of physical vapor deposition. It is the intent of this paper to render a tutorial on the technology of optical emission spectroscopy, providing the reader with an overview of the effect as well as application methods. Finally this presentation will provide several examples of application of this technology to the sputtering process.
This paper examines the application of trainable classification systems to the problem of diagnosing faults in engines at the manufacturing plant. Combining standard statistical methods and neural networks offers the ...
详细信息
The detection of instants of change in random process properties, a problem with a variety of applications in technical and medical diagnostics, control, and image processing, is considered. The focus in this study is...
详细信息
ISBN:
(纸本)0780302109
The detection of instants of change in random process properties, a problem with a variety of applications in technical and medical diagnostics, control, and image processing, is considered. The focus in this study is on the detection of a jump change in the mean and/or variance of an observed random sequence. The problem is to detect the jump as quickly as possible after its occurrence while avoiding an excessive false alarm rate (i.e. declaring a change before it occurs). For this purpose, the Neyman-Pearson criterion for the design of a disruption detector is used. An optimal design formula is derived for limiting cases. Using a Neyman-Pearson criterion, the designer can determine an optimal window size and an optimal detector sensitivity.
Extended Proportional Integral Derivative (EPID) is a model based control algorithm which consists of a conventional digital PID controller plus an additional compensator. This novel extension to the PID control algor...
详细信息
ISBN:
(纸本)0780302109
Extended Proportional Integral Derivative (EPID) is a model based control algorithm which consists of a conventional digital PID controller plus an additional compensator. This novel extension to the PID control algorithm offers the same properties as the IMC controller for a broad class of common industrial process models. EPID control simplifies the industrial implementation of IMC and has three important additional advantages. First, since EPID is based on the PID algorithm, it is easily understood by process engineers, control engineers and maintenance personnel. Second, the EPID strategy is easy to implement on industrial distributed control systems and can take full advantage of the functionality built into the available PID algorithms such as set-point tracking, automatic mode switching, etc. Third, since the IMC design procedure for sampled-data systems is used to derive the EPID controller parameters, it results in a controller that eliminates intersample rippling of the manipulated variable and overshoot of the process variable. Additionally, EPID explicitly takes into account model uncertainty, and has only one adjustable parameter which is directly related to the closed-loop speed of response and to control robustness. This paper presents the derivation of the EPID algorithm and IMC-based EPID tuning rules for common industrial process models. The practical and powerful EPID design tools, including performance analysis and robust stability analysis allow the designer to achieve the best compromise between the conflicting objectives of performance and robustness.
For systems with difficult dynamics such as large deadtimes, inverse responses or large time constants, conventional PID controllers have to be detuned significantly to ensure stability due to the limitation of the co...
详细信息
ISBN:
(纸本)1556174063
For systems with difficult dynamics such as large deadtimes, inverse responses or large time constants, conventional PID controllers have to be detuned significantly to ensure stability due to the limitation of the controller structure. Model-based control strategies overcome these limitation, but are seldom implemented on distributed control systems, limiting their industrial applications. This paper presents our experience with a novel implementation of Internal Model control on a distributed control system. For first-order and second-order processes with deadtime, the IMC control algorithm can be implemented as a PID controller plus additional compensator terms. These compensator terms are calculated external to the PID controller, and added to the PID controller as a feedforward input. This approach guarantees IMC performance while retaining the look and feel of a PID controller. This allows processes to be under model-based control without requiring personnel to learn a new user interface. Field application results are presented, including data on control loops such as liquid level and distillation column tray temperature. These results detail the procedures required for implementing model-based control directly in distributed control systems, including the methods of finding process models from plant data and controller performance results. These results indicate that it is entirely feasible to implement model-based control directly in distributed control systems and that the performance improvements over PID control can be substantial.
暂无评论