Universiti Sains Malaysia
In the presence of stiction, the control valves may present an oscillatory behaviour that affects the regulatory control performance, thereby causing a loss of product quality and increasing energy consumption. Detection of stiction in the early phase is a crucial key for process control to avoid major disruptions to the plant operations. In this paper, a novel technique based on a well-developed fuzzy clustering approach is proposed. Based on a dramatic change of the slope of the lines obtained from successive cluster centres in the presence of stiction, a new performance index to distinguish the cause of oscillation is proposed. The simulation, experimental and industrial results are provided.
Many researchers focus on detecting and modelling the valve stiction because it has undesirable effects on the control loop performance, which consequently results in poor product quality and increased energy consumption. It is difficult to model a process with a sticky valve using the mathematical definition because of its nonlinear properties such as stiction, hysteresis, dead band and dead zone. This work aims to develop and determine the appropriate model of a process with stiction, which can be used in controller design to mitigate the undesirable effect of the stiction. To achieve this goal by mapping the process with valve stiction to a fuzzy system, a dynamic fuzzy model of the plant is derived through an iterative well-developed fuzzy clustering algorithm, which generates suitable antecedent parameters from a set of input–output measurements that are obtained from the control output (OP) and the process output (PV). To determine the consequent parameters, the least square (LS) estimation is applied. The results reveal that the obtained data-driven Takagi–Sugeno-type (TS) fuzzy rule-based model can effectively represent an appropriate model of the process with stiction for different amounts of stiction that are obtained from the simulation and different industrial loops.
Presence of dead-band in engineering process decreases the system performance. Modeling of systems with such nonlinear properties is a key factor in model-based control and in fact a challenging task by conventional mathematic methods. In this paper, application of radial basis neural networks in such systems is investigated. The nonlinear static part of the system can be decoupled first from linear dynamic part and then modeled using Radial Basis Function (RBF) network; the dynamic linear part of the system can be identified using linear models. Results show that RBF can capture well, the key model of the systems with dead band.
The study of static friction in control engineering is the subject of many researches due to its impact on degradation of performance of the control loops. Mathematical model of systems with static friction is not straight forward. Precise and proper model of this phenomenon is a key factor in model-based control to mitigate its effect. By increasing number of smart valve in industry, demand for identification of such valves is rising. In these valves, identification of process is limited to control signal (OP) and valve position (MV). By taking advantage of Hammerstein approach, identification is divided in two parts, linear dynamic part and nonlinear static part. In this paper, adaptive neuro-fuzzy inference system (ANFIS) is used for identification of nonlinear static part of the plant. The linear dynamic part can be identified using linear identification methods. Results reveal that ANFIS which integrates both neural networks and fuzzy logic principles and has potential to capture the benefits of both in a single framework can capture well the key model of the systems with smart valves involved in static friction
Robot manipulators have become increasingly important in the field of flexible automation. So modeling and control of robots in automation will be very important. But Robots, as complex systems, must detect and isolate faults with high probabilities while doing their tasks with humans or other robots with high precision and they should tolerate the fault with the controller.This paper introduces a Neuro-Fuzzy Controller (NFC) for position control of robot arm. A five layer neural network is used to adjust input and output parameters of membership function in a fuzzy logic controller. The hybrid learning algorithm is used for training this network. In this algorithm, the least square estimation method is applied for the tuning of linear output membership function parameters and the error backpropagation method is used to tune the nonlinear input membership function parameters. The simulation results show that NFC is better and more robust than the PID controller for robot trajectory control.
In many industrial manufacturers, the quality of product heavily depends on control loop performance. Oscillations in control loop decreases performance of their systems. This is made worse with the increased number of control loops in a control process. There are several reasons for oscillations such as poor controller tuning, oscillating load or disturbances with a high-frequency. However, the most common reason for oscillations is friction in the valve. This paper presents a brief summary of some effective methods of detection of stiction in control loops and the compensation techniques to mitigate this effect. The aim is to determine the appropriate model of stiction in order to design a proper controller to prevent oscillations in the control loops. Based on the literature, Data Driven model with fewer parameters provides simpler solution and is the most common method used in comparison with physical based model which presents a number of unknown physical parameters. Moreover, previous methods (even in data-driven models) were based on some assumptions and thus their judgments are not reliable when the assumptions were not satisfied. From this study too, it is seen that numerous methods of stiction detection are available. Hence, it is suggested that the future research emphasis should be more on compensation of stiction especially for nonlinear processes.
This paper proposes a digital fuzzifier with 7 bit resolution. To achieve a high performance, the A/D conversion and fuzzification operation are done simultaneously on the same functional block, through a programmable membership function. To implement the idea we have used flash method for the 4 LSBs which results in high speed conversion. Although the 4 bit converting is adequate for most practical applications, we take advantage from successive approximation technique to achieve better precision, if needed. The proposed converter allows a significant amount of silicon area to be saved compared to fully parallel A/D. Result of extracting of the 7 bit A/D converter shows that the size of converter’s layout is less than 0.02 mm2 in 0.35µm CMOS standard technology. Moreover, the Hspice simulation showed that it can achieve 25 MHz speed and total power dissipation is 5mw in the aforementioned technology.
Presented in this paper is design of a new FLC chip with mixed-signal (analog and digital) inputs and digital outputs. This work is based on a new strategy in which analog advantages such as low die area, high speed and simplicity are added to the system advantages, whose output is digital considering unchanged digital system properties. For implementing this idea, a new programmable Fuzzifier circuit based on mixed-signal input and generation of three different current membership functions including Gaussian, Trapezoidal and Triangular shapes are proposed. To contribute antecedents in inference block, three new integrated circuits for implementing Min–Max operators are proposed. We improved and designed a Multiplier/Divider circuit and a current mode Analog to Digital (A/D) converter with 7bit resolution to complete and implement Defuzzifier block. The proposed controller circuit which consists of two inputs, four membership functions for each one, sixteen rules and one output designed less than 0.1 mm2 area in 0.35 μm CMOS standard technology. The systematical simulation results of MATLAB software was compared to the HSPICE simulation results using extracted circuit layout. The inference speed of the controller is about 16.6 MFLIPS.
Master of Science, Control Engineering
Master of Science, Control Engineering
Master of Science, Control Engineering
Master of Science, Electrical Engineering