Dong-Feng Wang

Modeling the circulating fluidized bed boiler using RBF-NN based on immune genetic algorithm

Combining the excellence of immune algorithm and Genetic Algorithm(GA), a modeling. method using parallel Radial Basis Function Neural Network(RBF-NN) is proposed based on immune GA optimization. And this method is applied to modeling of a Circulating Fluidized Bed(CFB) boiler so as to realize neural network modeling of CFB boiler. The established neural network model is very useful for both characteristic research and advanced control strategy development of CFB boiler.

Selective ensemble using discrete differential evolution algorithm for short-term load forecasting

In order to further improve the accuracy of short-term load forecasting, a selective neural network ensemble method using discrete differential evolution algorithm is proposed. Firstly, the individual vectors in differential evolution algorithm are dispersed. Secondly, a group of RBF neural networks with larger difference are trained independently and a binary bit string in multi-dimensional space with the value of 0 or 1 is used to describe all the possible neural network integrations. Lastly, part of individual networks is optimized selected to ensemble and an entropy method is used to determine the integrated weighted coefficient of component neural networks according to the variability of prediction error sequences. The experiments show that the proposed approach has higher accuracy and stability.

Wind speed conformal prediction in wind farm based on algorithmic randomness theory

An approach of a mean hourly wind speed conformal prediction in wind farm is proposed. Conformal prediction is a new prediction methodology. It can be used not just to make predictions but also to estimate the confidence under the usual independent and identically distributed assumption. Based on support vector regression, wind speed regions are predicted by inductive confidence machine. Wind speed regionpsilas width and confidence reflect the accuracy and reliability of the prediction. Compared to bare wind speed forecasting, the accuracy and reliability of the prediction can be used to reduce the risk of decision-making. Experimental results are given by using wind mean hourly speed measured in wind farm, and the application of the method carried out a detailed analysis and verification.

Thermal Process System Identification Using Particle Swarm Optimization

System identification adopting an open loop step response curve is a feasible way to obtain the mathematic model of the control object. Due to the satisfying performance in global optimization, evolution computing (EC) methods such as genetic algorithm have been applied to the open loop step response curve analysis and achieved effective results. In this paper, particle swarm optimization (PSO) algorithm which is considered as a new relative addition to the EC methods is introduced to solve the system identification problem for thermal process control objects. Typical forms of transfer functions for the thermal process are adopted, utilizing PSO algorithm to estimate the parameters, for the convenient application of which, a set of software is also developed. With these softwares, some characters of the experimental data are specified by the user. And then the initial values for the model parameters are deduced from these characters. Around these initial values, a smaller search space is determined, within which the PSO algorithm searches the optima for the model parameters. Thus the search efficiency can be improved remarkably. The software has been applied in some power plants, the results of which prove the effectiveness of the method

Identification of fractional order system using Particle Swarm Optimization

This paper presents the identification of fractional order system in frequency domain by using Particle Swarm Optimization (PSO) algorithm. PSO is extended to estimate the fractional derivative order. Meanwhile, recursive least squares algorithm is associated to calculate the denominator and numerator coefficients of transfer function. Simulation examples with noise-free and noisy data are given to verify the effectiveness of the method proposed in this paper.

Application of grey self-tuning fuzzy immune PID control for main steam temperature control system

The prominent virtues of grey prediction are small information, few data and a little of computation. The conventional PID controller has strong robustness, sophisticated technology and simple structure, so it has been used widely in the process control. With the adjustable role of feedback response of biological immune systems, the capability of simulating non-linear functions with fuzzy rationalizing logic, and the advantage of gray prediction, an application of fuzzy immune PID control with gray prediction has been proposed. In view of the grey-step having great effect to the results of forecast, the fuzzy rules are used to adjust the forecast step, which is a very good supplement to the grey prediction. The method mentioned in this paper is applied to the main steam temperature control system of power plant, the simulation shows that this strategy has strong capability of anti-disturbance.

FCM clustering algorithm for T-S fuzzy model identification

An approach for building T-S fuzzy model is proposed based on fuzzy c-mean clustering algorithm on the basis of nonlinear modeling experience. An alternative T-S fuzzy model is adapted, which has the uniformed premise structure, the premise parameter is decided by fuzzy c-mean clustering algorithm and the consequence parameters is calculated by least square algorithm, and the identification precision is enhanced. Finally the effectiveness and practicability of this method is demonstrated by the simulation result of Box-Jenkins gas furnace data and Mackey-Glass chaos time series.

A dynamic selective neural network ensemble method for fault diagnosis of steam turbine

A new dynamic selective neural network ensemble method for fault diagnosis of steam turbine is proposed. Firstly, a great number of diverse BP neural network models are produced. Secondly, the error matrix is calculated and the K-nearest neighbor algorithm is used to predict the generalization errors of different neural networks on each testing sample. Thirdly, the individual networks whose generalization errors are in a threshold will be dynamically selected and a conditional generalized variance minimization method is used to choose the most suitable ensemble members again. Finally, the predictions of the selected neural networks with weak correlations are combined through majority voting. The practical applications in fault diagnosis of steam turbine show the proposed approach gives promising results on performance even with smaller learning samples, and it has higher accuracy and efficiency compared with other methods.

Application of Fuzzy Predictive Control in Superheated Steam Temperature Control

In order to achieve effectiveness and high control performance of nonlinear, time-variable processes, a fuzzy predictive scheme, combing fuzzy modeling and simplified increment model algorithmic control (SIMAC) is proposed. In this algorithmic, impulse response for different operating conditions are extracted online from the nonlinear fuzzy model. The method is better than PID in robustness and dynamic performance. Simulation for the superheated system is carried out. The results show the proposed method has perfect load adaptability

Kernel independent component analysis and dynamic selective neural network ensemble for fault diagnosis of steam turbine

A new method for fault diagnosis of steam turbine based on kernel independent component analysis (KICA) and dynamic selective neural network ensemble is proposed Firstly, the fault data of steam turbine is analyzed using KICA to extract main features from high dimensional patterns Not only is the diagnosing efficiency improved but also the diagnosing accuracy is ensured Then, the generalization errors of different neural networks to each validating sample are calculated and the information is collected into a performance matrix, according to which the K-nearest neighbor algorithm is used to predict the generalization errors of different neural networks to each testing sample Lastly, the individual networks whose generalization errors are in a threshold λ will be dynamically selected and the predictions of the component neural networks are combined through majority voting The practical applications in fault diagnosis of steam turbine show that the proposed approach gives promising results on performance even with smaller learning samples, and it has higher accuracy and stability.