Search

This study proposes a practical approach to online control performance monitoring. In this approach, three key issues are primarily studied, namely: the online method of trend extraction, the method of online state recognition based on the slopes of the adjacent sections, and the online delay estimation method based on recursive least squares. The performance assessment procedure is then presented. On this basis, an online performance monitoring program is developed with the use of Matlab GUI toolbox. A reheat steam system controller for a 1000 MW power unit is used to test the effectiveness of the proposed approach and program.

The data-based parameter identification is an efficient way for the modeling of Wind Turbine Generation System (WTGS) which only reveals the external characteristic. However the inner characteristic is unknown. In practice the parameters of the mechanism model are changing with time. Thus in the paper the grey-box modeling approach which utilizes the operation data and operating mechanism together are proposed to realize the online parameter identification of WTGS under closed-loop. Considering the essential difference between aerodynamic system and the other systems the driven-train system and the electrical system are identified together which also simplify the identification process and the switching mechanism is proposed for the parameters update of mechanism model which is much efficient and applicable in practice. Besides the identification of driven-train system and electrical system in combination could ensure the global approximation ability to the dynamics of WTGS while the local approximation ability to the subsystems could also be validated when the identification parameters are obtained. In the paper the combined state space model including the driven-train model and Doubly-Fed Induction Generator (DFIG) model is deduced. The GH Bladed Software would be used to generated the operation data for identification and the genetic algorithm is adopted to identify the model parameters. As a result the nonlinear state space model could be obtained and its approximation ability to the dynamic characteristic of WTGS is validated through simulation. Then the practical characteristic of wind turbine could be approximated more accurately. Note that the paper provides meaningful research basis for the control design and model-based fault diagnosis of large-scale WTGS.

To improve the security and reliability of equipment and reduce their failure rate, a data-driven state detection algorithm was proposed. The concepts of multi-scale system, multi-scale entropy and multi-scale exergy were defined. The algorithm is used for multi-scale systems whose state parameters change over time and have the characteristic of increasing monotonically on a dominant scale. An abrasion index for the middle speed roller ring mill was constructed, which was used to monitor the states of the instruments. Noise that affected the accuracy of the results was analyzed. The results of simulation experiments demonstrate the effectiveness of the algorithm, which can provide a technical basis for condition maintenance.

The least squares support vector machine (LS-SVM) is sensitive to noises or outliers. To address the drawback, a new robust least squares support vector machine (RLS-SVM) is introduced to solve the regression problem with outliers. A fuzzy membership function, which is determined by heuristic method, is assigned to each training sample as a weight. For each data point, firstly a deleted input neighborhood is found when the high-dimension feature space of input is focused on. Then the new field is reformulated after the output is brought in the neighborhood which we have found. The fuzzy membership function (weight) is set according to the distance from the data point to the center of its neighborhood and the radius of the neighborhood, which implies the probability to be an outlier. Two benchmark simulation experiments and analysis are presented to verify that the performance is improved.

A solution to real-time optimal load dispatching problem of power plant is introduced in this paper. Based on analysis of amounts of operating data, it is proved that the characteristics curve of unit coal consumption rate (CCR) against load are constructed by discrete points and it can't be deduced by linear regression. At the default condition of accurate thermodynamic calculation, the factors impacting unit CCR for power supply are studied at different operating conditions. Data mining method is employed to predict CCR for power supply at different operating points and environmental condition based on history and real-time data. Thus, discrete mathematical model between load and CCR is build up for optimal load dispatching (OLD) among the units. Global optimal solution can be approached by dynamic programming method for the reason of the low dimension (number of units in one plant) of this issue now. The method proposed in this paper is applied in a power plant of Datang group and it is proved to be practical and effective.

A rule extraction method for thermal power unit operation optimization which based on data mining technology is proposed in this paper. Data mining technology is employed to analyze history data stored in real-time/history database of the unit. Stability index, economic index, environmental index and synthetic index are constructed to evaluate the performance of current operating condition. Through comparison analysis on similar operating condition, the best operating parameters and operating mode can be found out at the current condition. Accompanying with data accumulation proceeding in database, nonlinear static model between the unit operation performance index and controllable parameters is drawn over all operating conditions. Effectiveness of this method is illustrated by a practical operation optimization project, in which the realization scheme of operation guidance is also given out.

The determination of operation optimization value is very important for economic analysis and operation optimization in power plants. The operation optimization value determined by traditional methods usually cannot reflect the actual running states correctly in power plants with the increase in running time. Based on a large amount of history data stored in power plants, a way of operation optimization by applying data mining technique is proposed. The structure of operation optimi-zation based on data mining is established and the fuzzy association rule mining algorithm is introduced to find the operation optimization target value to guide the operation in power plants. Based on the actual local data in a 300 MW unit, the operation optimization value in typical load ranges is found out by data mining to provide better adjustment guidance in industry process. Experimental results show that the determination of operation optimization value based on data mining can improve the efficiency and decrease the emission of pollutants.

In this paper, a robust model predictive control algorithm based on infinite norm optimization is applied to a nonlinear boiler-turbine unit. Simulation results show that the designed controller has good tracking, disturbance rejection and robustness against the variations of the operation points due to the plant nonlinearity. Compared with conventional model predictive controller, this control system functions better when the operating point varies.

Against the serious coupling among system variables, long time delay of outputs response and time changing of model of the low speed pulverizer of power plant, a kind of self-tuning fuzzy control algorithm is presented. In the algorithm, the coupling among system variables has been eliminated, the proportion of the deviation and its differential can be adjusted automatically according to the control process, and accurate object model is not needed when setting the controller parameters. The fuzzy control algorithm in this paper has been used into power plant and satisfactory control result acquired.

A data driven state detection algorithm was proposed to improve the security and reliability of equipment. The algorithm is used for one class of objects whose state parameters change slowly and cumulatively in the long term. The concepts of multi-scale system, multi-scale entropy and multi-scale exergy were used to describe these processes. With the help of the algorithm, an ash deposition index for the radiant heating surface of a 600 MW unit was constructed to monitor the states of the instruments. Noise was analyzed. The results of simulation experiments demonstrate the effectiveness of the algorithm, which can provide a technical basis for condition maintenance.

There are some load areas of thermal power units, around which some coal mills need frequent starting and stopping. In this paper, those areas are considered as Prohibited Operating Zones (POZs), and a new strategy is proposed to solve this issue. PPLD with POZs can be treated as a special non-convex space optimization problem and new method is needed. An improved GA-TS algorithm is proposed intend to search optimal solution. There are three major improvements on the algorithm, including application of Elite Solution List (ESL), design of solution repair strategy, and stop criterion of Tabu Search. In final part of this paper, simulation on a thermal plant with 8 units show effectiveness of the proposed strategy and improved GA-TS algorithm.

An improved genetic-tabu search algorithm is proposed in this paper, which combines large scale search ability of genetic algorithm (GA) and anti-premature ability of tabu search (TS). A long term memory list is designed to record elite solutions searched by GA operation (GA-LTM), and those solutions are induced to each crossover procedure, and drive offspring close to the optimum value. After certain iterations of GA operation, local optimum solutions in solution space are recorded in the GA-LTM. Then, GA operations stop, and solutions in GA-LTM are adopted as start points of TS algorithm. For each solution in GA-LTM, TS search procedure initializes with a trace-back strategy. If TS cannot find better solutions after several steps, the best solution found in previous steps would be recorded as best solution for the start point. After TS search for solutions in GA-LTM accomplish, a comparison is made among those TS found solutions, and the best one is recognized as the optimum solution. At end of this paper, a simulation study on 5 thermal units is done. Validity of the proposed algorithm is approved.

Aiming at multi-disturbance characteristic of drum water level system, control system uses three impulse cascade control strategy. This paper designs three impulse cascade control combined of incomplete differential PID and immune P control, fully using the advantages of artificial immune feedback control and incomplete differential PID. Boiler drum level simulation result shows the cascade control based on immune feedback mechanism, can improve the system dynamic performance. The simulation results also prove the validity of the control strategy.

In this paper a novel self-tuning controller based on the concept of process supervision and gain adjustment are documented. The features of the design include leading a time varing factor λ (k) to controller output and adopting estimater switch technique. The new design known as ISTC was extended to MIMO system under a particular process model assumption. Simulation study was done in a 2x2 load/pressure system of a turbine - boiler unit. Experiment results shown the efficiency of new design and possibility of industry application in practice.

The problem of fuzzy modeling for state and input time-varying delay systems with a class of nonlinear uncertainties by fuzzy T-S model is addressed. Utilizing LMI (linear matrix inequality), the problem of fuzzy robust Hinfin controller design for the systems is studied. Assuming that the nonlinear uncertain functions in the model considered are gain-bounded, we obtain a sufficient condition for the robustly asymptotic stability of the closed-loop system via Lyapunov stability theory. By solving the LMI, we construct a feedback control law which guarantees the robustly asymptotic stability of the closed-loop system and reduce the effect of the disturbance input on the controlled output to a prescribed level.

A state observer based on classical dynamics and a closed loop control system consisted of the state observer are constructed. A general controller based on classical dynamics is constructed, which is explained in classical dynamics: the controlled plant output locus is controlled and led to the desired output locus using feedback control of position, speed and acceleration of the general controller. The cases study show the general controller can improve performances and qualities of control system. At the same time, the paper analyses that PID controller, inner model controller and robust controller are identical in classical dynamics, and gives some improved methods for designing these controllers.

Combustion instability is a phenomenon that happens in a gas turbine when the lean premixed combustion mode is used to reduce the NOx emissions. The process is unstable and thus difficult to control. This paper applies the internal model control (IMC) method in the control of the combustion instability, with the aim of finding effective control structures and methods for this kind of system, and preparing for the study of wide-range robust control. Design results show that the IMC controller has good disturbance rejection capability, but the method should be improved to achieve better performance for the active control of combustion instability.

Multi-model control of a boiler-turbine unit is discussed in this paper. Based on the nonlinear model of the unit, this paper applies the gap metric concept to analyze the problem whether a multi-model control is necessary for the unit, and select the appropriate operating points accordingly. The multi-model control is then designed based on local coordinated PID controllers. Simulation and the experimental results at Daqi 4# unit show that the multi-model controller can guarantee the desired performance under wide range of load variations.

On the basis of analysis to the boiler models that have exist, the modeling method and simulation environment for large-scale coal-fired power unit are discussed. The power unit is divided into many parts; a set of dynamic model library of thermal equipment and subsystem is setup on with modularization modeling method. The simulation environment is selected as MATLAB/simulink and S-function is used. The integrated dynamic model of 300 MW coal-fired power unit is set up in simulink.

The main steam temperature control is a typical time-variable dead time and time-variable parameter process. The relation between the dead time of the main steam temperature controlled process and steam flow is presented in this paper. A kind of adaptive control arithmetic based on single neuron is designed to make parameters of the internal model to approach the object. Parameters of the controller can be tuned based on parameters of the internal model through a parameter modifying arithmetic. Internal control strategy is adopted in the controller, and the control system designed is applied through a distributed control system. Operation result shows that the control system designed in this paper has good performance to overcome disturbance and deviations of model parameters.

A boiler dynamic model is a relatively complicated model. In order to construct a more practical control system, we must analyse the more detailed dynamic process. This paper introduces a new analysis method named load/pressure increment prediction model. This increment prediction model can fit the practical signal well. Simulation illustrates the introduced method.

A kind of improved genetic algorithm for identifying transfer function of thermal process in power plant is introduced. In the algorithm, floating-point coding, rank-based selection, elitist reservation and grouping method are used, the premature convergence is restrained, the global and local searching ability is improved. The genetic algorithm-based model identification MATLAB program is designed, the transfer functions of thermal process can be got with it according to the operating data log files. The identification results to topical thermal process is given. It is shown by simulation research that accurate identification results can be got no matter what kind of input signal is used, such as step signal, random operating signal, even if there is a strong noise in the input signal.