Han Pu

Controller design for networked control systems based on neural network prediction

To the networked control systems in which network-induced delay is time-vary and less than one sample time, a neural network predictive control method is presented to compensate the random delays in this paper. Genetic algorithm is used to optimize parameters. Neural network is introduced to do prediction and optimization calculation. And the uncertainty of the delay is transformed to parametrical uncertainties. Simulation result shows that this control method can evidently improve the control quality with increasing time delay, and has stronger adaptability.

Predictive functional control of integrating plant

A novel Predictive Functional Control (PFC) algorithm for integrating plant is presented based on step response. Control law of PFC with one base function (step function) and two base functions (step and ramp function) are provided. Simulation results show that the algorithm has good performance.

Fuzzy identification method in nonlinear system based on G-K clustering algorithm

In accordance with the problems that the algorithm is too complex in the past fuzzy modeling methods, this article propose a new method of fuzzy modeling for nonlinear system. The method is simple and powerful. In this method, the premise configuration and parameter of this fuzzy model is decided by G-K fuzzy clustering algorithm, and succedent parameter of fuzzy model is identified by orthogonal least square. Finally the effectiveness and practicability of this method is demonstrated by the simulation result of the Box-Jenkins gas furnace data.

Adaptive generalized functional synchronization of chaotic systems with unknown parameters

A universal adaptive generalized functional synchronization approach to any two different or identical chaotic systems with unknown parameters is proposed, based on a unified mathematical expression of a large class of chaotic system. Self-adaptive parameter law and control law are given in the form of a theorem. The synchronization between the three-dimensional Rossler chaotic system and the four-dimensional Chens hyper-chaotic system is studied as an example for illustration. The computer simulation results demonstrate the feasibility of the method proposed.

Optimal load economic distribution research and engineering realization

Nowadays, the difference between peak and lowest of load in electricity market is increasing. Central command requires the power plant to complete the load demand quickly. For this situation, we use particle swarm optimization to solve load economic distribution. We change the weight between speed and economy within the plant, taking into account the auxiliary running. Make the power plant keep up with the central command and keep the unit running at lower coal consumption. Tests show that: Our algorithm is stable and reliable. And we have developed thermal power plant load distribution software, which has been put into commercial operation.

PSO and RBF network-based Wiener model and its application to system identification

In this paper, a new kind of Wiener model structure is introduced, which is realized by using the mapping function of neural networks. The model uses the linear dynamic neurons and a RBF network to express one Wiener models dynamic linear part and static nonlinear part respectively. The parameter identification for the new Wiener model adopts the unified identification method. The learning of parameters includes two cycles that the inner-cycle is executed by gradient training methods based on the BP thought and the outer-cycle uses the PSO (Particle Swarm Optimization) algorithm. The training method based on unified identification makes the new Wiener model converge to the steady state along the expected direction with a small error in a short time. The Wiener model is applied to the identification of the famous Box and Jenkins gas CO2 density, and the simulation results show that the method proposed in this paper is effective.

Power fault using signal analysis with complex window and pattern recognition approach

In power system network, the voltage and current signal exhibit fluctuations in amplitude, phase, and frequency due to nonlinear devices utilized for power generation, transmission and distribution. The power quality problems can cause electric equipment malfunction and consume great electric energy. Therefore, it is necessary to monitor these disturbances. A novel approach is put forward to detect and analyze voltage stability by combining wavelet transform with pattern recognition technique. In signal denoising process, the statistic rule is proposed to determine the threshold of each order of wavelet space, which can determine the decomposition level adaptively, increasing the signal-noise-ratio. The wavelet transform coefficients as feature vector are presented for extracting disturbance signal. The effectiveness of training algorithm for pattern recognition is described, which can be realized by feature vector acquisition. The method incorporates the advantages of morphological filter and multi-scale wavelet transform to extract signal feature meanwhile restraining various noises. The simulation results prove that the proposed method is correct and effective for voltage stability analysis.