Chaung Lin

Fuzzy logic control of steam generator water level in pressurized water reactors

In this paper a fuzzy logic controller is applied to control the steam generator water level in a pressurized water reactor. The method does not require a detailed mathematical mode of the object to be controlled. The design is based on a set of linguistic rules that were adopted from the human operators experience. After off-line fuzzy computation, the controller is a lookup table, and thus, real-time control is achieved. Shrink-and-swell phenomena are considered in the linguistic rules, and the simulation results show that their effect is dramatically reduced. The performance of the control system can also be improved by changing the input and output scaling factors, which is convenient for on-line tuning.

A RULE-BASED EXPERT SYSTEM FOR AUTOMATIC CONTROL ROD PATTERN GENERATION FOR BOILING WATER REACTORS

This paper reports on an expert system for generating control rod patterns that has been developed. The knowledge is transformed into IF-THEN rules. The inference engine uses the Rete pattern matching algorithm to match facts, and rule premises and conflict resolution strategies to make the system function intelligently. A forward-chaining mechanism is adopted in the inference engine. The system is implemented in the Common Lisp programming language. The three-dimensional core simulation model performs the core status and burnup calculations. The system is successfully demonstrated by generating control rod programming for the 2894-MW (thermal) Kuosheng nuclear power plant in Taiwan. The computing time is tremendously reduced compared to programs using mathematical methods.

Hierarchical Fuzzy Logic Water-Level Control in Advanced Boiling Water Reactors

AbstractA hierarchical fuzzy logic controller (FLC) is applied to control the water level in an analytical simulation using a simplified advanced boiling water reactor model. To reduce the control effort of the feedwater pump, the linguistic variable, change in pressure, was adopted. Four linguistic variables were used for the FLC design, and the number of control rules became large if the conventional FLC design method was used. Using a hierarchical rule structure reduces the total number of control rules and thus the final decision tables. To reduce the design effort, two methods were separately developed for fine-tuning. One tunes the scaling factors using an optimization method, and the other tunes the control rules using a method similar to a fuzzy model reference control. The simulation results show that the performance of the hierarchical FLC is comparable to that of the proportional-integral controller, but use of the designed controller results in a shorter settling time.

Real-time control of ion density and ion energy in chlorine inductively coupled plasma etch processing

In this study, we have experimentally demonstrated the real-time closed-loop control of both ion density and ion energy in a chlorine inductively coupled plasma etcher. To measure positive ion density, the trace rare gases-optical emission spectroscopy is used to measure the chlorine positive ion density. An rf voltage probe is adopted to measure the root-mean-square rf voltage on the electrostatic chuck which is linearly dependent on sheath voltage. One actuator is a 13.56 MHz rf generator to drive the inductive coil seated on a ceramic window. The second actuator is also a 13.56 MHz rf generator to power the electrostatic chuck. The closed-loop controller is designed to compensate for process drift, process disturbance, and pilot wafer effect and to minimize steady-state error of plasma parameters. This controller has been used to control the etch process of unpatterned polysilicon. The experimental results showed that the closed-loop control had a better repeatability of plasma parameters compared with open-loop control. The closed-loop control can eliminate the process disturbance resulting from reflected power. In addition, experimental results also demonstrated that closed-loop control has a better reproducibility in etch rate as compared with open-loop control.

Neurocontrol of Pressurized Water Reactors in Load-Follow Operations

The neurocontrol technique was applied to control a pressurized water reactor (PWR) in load-follow operations. Generalized learning or direct inverse control architecture was adopted in which the neural network was trained off-line to learn the inverse model of the PWR. Two neural network controllers were designed: One provided control rod position, which controlled the axial power distribution, and the other provided the change in boron concentration, which adjusted core total power. An additional feedback controller was designed so that power tracking capability was improved. The time duration between control actions was 15 min; thus, the xenon effect is limited and can be neglected. Therefore, the xenon concentration was not considered as a controller input variable, which simplified controller design. Center target strategy and minimum boron strategy were used to operate the reactor, and the simulation results demonstrated the effectiveness and performance of the proposed controller.

Real-time feedback control of electron density in inductively coupled plasmas

The real-time feedback control of electron density was performed in an inductively coupled plasma (ICP). A 36 GHz heterodyne interferometer was adopted as the sensor for electron density measurement. The actuator was rf power which drove the ICP antenna. The results show that the electron density in an ICP system is a type 0 system. Therefore, a proportional-integral controller is necessary to eliminate steady-state error. A prefilter was designed to smooth the desired step change of electron density so that the variation of rf power became mild. A feedforward compensator was added to reduce the disturbance effect resulting from pressure change. The experimental results showed that the control system could quickly track the desired electron density and compensate the electron density variation resulting from pressure disturbance.

Design of a fuzzy logic controller for water level control in an advanced boiling water reactor based on input-output data

A fuzzy logic controller (FLC) has been designed to control the water level in an advanced boiling water reactor (ABWR). The performance was comparable to that of a proportional-integral controller. However, the feedwater flow rate did not change smoothly to the steady state. Therefore, a method based on input-output data was adopted to prevent this problem. The data required for deriving the fuzzy rules were the results of various instances of satisfactory manual control of an ABWR simulation model. To construct the membership functions of the linguistic variable, the data were clustered using the fuzzy C-means method. The fuzzy rules were then generated from the data. Because the control actions were not guaranteed to be proper in all the cases and the data were not complete for all the possible operation conditions, the fuzzy rules were modified and extra rules were added based on human knowledge so that satisfactory performance can be achieved. Nevertheless, the method is helpful in deriving a set of ...

Feedback control of chlorine inductively coupled plasma etch processing

Feedback control has been applied to poly-Si etch processing using a chlorine inductively coupled plasma. Since the positive ion flux and ion energy incident upon the wafer surface are the key factors that influence the etch rate, the ion current and the root mean square (rms) rf voltage on the wafer stage, which are measured using an impedance meter connected to the wafer stage, are adopted as the controlled variables to enhance etch rate. The actuators are two 13.56 MHz rf power generators, which adjust ion density and ion energy, respectively. The results of closed-loop control show that the advantages of feedback control can be achieved. For example, with feedback control, etch rate variation under the transient chamber wall condition is reduced roughly by a factor of 2 as compared to the open-loop case. In addition, the capability of the disturbance rejection was also investigated. For a gas pressure variation of 20%, the largest etch rate variation is about 2.4% with closed-loop control as compared ...

Velocity distribution in the peripheral subchannels of the CANDU-type 19 rod bundle

Abstract Measurements of the detailed flow velocity and turbulent intensity in the peripheral subchannels of a non-wrapped CANDU-type 19-rod test bundle were performed by Laser Doppler Anemometry in a water loop. The bundle consisted of rods, 1.905 cm in diameter and 62.28 cm long, positioned by two 3 mm thick end-plates. The test housing was an 11 cm ID transparent pipe. The transverse velocity was negligible (∼0.002 m/s) and the turbulence intensity uniformly distributed (∼7%) throughout all the channel areas. The axial flow patterns for the peripheral subchannels are given, and show velocity gradients and average and peak velocities higher in the subchannels facing triangular channels than in subchannels facing square channels, even though they have the same cross-sectional areas. Experimental results show excellent agreement with COBRA-IIIC computer code calculation of flow distribution in CANDU-type reactors if the bundle is well-aligned and well-centered. The average and peak velocities of a off-centered bundle are also given.

Automatic search of the power ascension path for a boiling water reactor using genetic algorithm and neural network

A method that includes a genetic algorithm (GA), principal component analysis (PCA), and an artificial neural network (ANN) is adopted in the search for the power ascension path of a boiling water reactor that used to rely solely on an operator’s experiences. The power ascension path is formulated as an optimization problem with thermal limits, e.g., minimum critical power ratio, maximum linear heat generation rate, and maximum average planar linear heat generation rate, and with the stability requirement serving as a constraint. The Simulate-3 code is used to calculate the reactor core status, while the optimization problem is solved through the use of the GA. Since the search domain of the GA is relatively large, the ANN, which models the power ascension path, is developed in order to quickly select the candidate solutions for further Simulate-3 calculations, permitting the algorithm to converge effectively. Meanwhile, PCA is used to reduce the ANN input vector’s dimension, which improves the ANN training efficiency and pattern recognition capability. The results show that this method efficiently obtains the proper power ascension path required for meeting all constraints at different fuel exposure levels.