Mao Zhi-zhong

Feed-forward and inferential control and its application

Inferential control system has many excellent performances such as disturbance resisting and set-point tracking, however, the application is restricted when strong load disturbance exists or stable control accuracy and response speed are highly required in the system. Feed-forward control system responds quickly to the system with measurable disturbance, but the control accuracy is easily affected by disturbance, so there will be a big error if the disturbance model is not accurate enough. Compound control algorithm which combines the feed-forward and inferential control algorithm is proposed in this paper. Feed-forward control algorithm is used to cancel the load- torque disturbance. Inferential control algorithm is used to cancel the error caused by soft sensing method. It is used to control the speed control of guide disc. The accuracy, robustness and fast response of the control system are proved by simulation.

An improved multi-objective particle swarm optimization algorithm and its application in EAF steelmaking process

An efficient improved multi-objective particle swarm optimization algorithm based weighted pheromone sharing mechanism (PM-MOPSO) approach for solving the power supply curve of electric arc furnace(EAF) steelmaking process is presented in this paper. In PM-MOPSO algorithm, the weighted pheromone sharing mechanism coordinates specific gravity among the optimal solutions; the position migration accelerates algorithm convergence speed; the clustering population compression maintains population diversity. Finally, the application shows that it reduces the electric energy consumption, shortens smelting time and improves lifetime of the furnace lining and cover. The result expresses that the algorithm is effective.

A predictive adaptive controller and its application to the control of the wet autogenous mill

A simple predictive adaptive controller is proposed. Since the incremental impulse response predictive model is adopted, so the control algorithm is more suitable for industrial process control. The running results used to apply it to the control of a mineral processing wet autogenous mill show that the control algorithm has some nice properties; it is fairly robust to delay mismatch, load disturbances and stochastic disturbance.<<ETX>>

Study on hybrid modeling approach for gold cyanidation leaching process

Gold cyanidation leaching process is a complicated chemical process. Establishing an accurate process model is of important significance for control and optimization of leaching process. By analyzing the inherent characteristics of leaching process, two types of dynamic hybrid models are designed in this paper, which consist of mass conservation equations as the dynamic mechanistic model and BP neural networks that are used to compensate for the difference between the model output and the real output or to estimate the unknown parameters. For the serial hybrid model, because of the immeasurability of kinetic reaction rates, the Tikhonov regularization method is used to estimate the kinetics reaction rates of gold and cyanide ion, which reduces the propagation of errors in the measured data effectively. The simulation results show that the hybrid models have superior predictive performance than that of the pure mechanistic model. Among the three models, the serial hybrid model has the best predictive precision. The serial hybrid model can predict the concentration of both components accurately, which is an important model basis for subsequent control and optimization of the leaching process.

Endpoint prediction of electric arc furnace based on T-S fuzzy system

The endpoint parameters are very important to the process of electric arc furnace steel-making, but they are difficult to be measured on line. The soft sensor technology is widely used on the prediction of endpoint parameters. Based on the analysis of the smelting process and the advantages of support vector machines, a soft sensor model for predicting the endpoint parameters is established by T-S fuzzy system. A hybrid modeling method is proposed to construct the structure and to tune the parameters of T-S fuzzy model in this paper. Two steps were carried out: the establishment of an initial T-S fuzzy system by extracting rules in the total input space uniformly, and addition of new fuzzy rules to the system according to the Absolute Error index. Both the Levenberg-Marquardt method for nonlinear parameter optimization and the least squares method for linear parameter estimation were used to accelerate the computational convergence. The accuracy of the soft sensor model is perfectly improved. The simulation result demonstrates the practicability and efficiency of the T-S fuzzy system in the endpoint prediction.

On-line adaptation algorithm for RBF kernel based FS-SVM

The performance of Fixed-Size least squares support vector machines (FS-SVM) has been illustrates on the large-scale modeling problem. This paper presents an adaptive RBF kernel based FS-SVM and an on-line adaptation algorithm for time-varying nonlinear systems. The key feature of this algorithm method is the direct approach used for formulating the training target. Based on the feature of RBF kernel, the error (objective) function between actual active model and target model is formulated and can be minimized by Gradient descent algorithm. The proposed algorithm is capable of maintaining the accuracy of learned patterns even when a large number of aged patterns are replaced by new ones through the adaptation process. The simulation results show the effectiveness of this architecture for adaptive modeling.

An adaptive neural network controller for nonlinear MIMO discrete-time systems

In this paper, a multivariable adaptive control approach is proposed for a class of unknown nonlinear multi-variable discrete-time systems. An equivalent model in affine-like form is first derived for the original nonaffine discrete-time systems. Then, feedback linearization adaptive control is implemented based on the affine-like equivalent model identified with neural networks. A new NNs weight updating method is proposed based on idea of

A Genetic Algorithm Based Approach For Mixed-Media Advertising Budgeting

Mixed-media advertising which has become more effective and popular attracts the attention of advertising agencies, media planners and advertisers. How to decide and allocate the budget for mixed-media advertising becomes a fundamental part of a companys advertising campaign.In this paper, an optimization approach based on Genetic Algorithm is proposed to deal with this advertising budget problem which is formulated as a constrained nonlinear programming problem. Then an experiment example is given to explain the process of our approach. The result of this example supports the feasibility and efficiency of our method.

The Research of the Credit Evaluation Based on Projection Pursuit-Fuzzy Clustering System Model

This paper presents an approach based on Projection Pursuit and fuzzy rule extraction combining new hybrid method of classification system. This method is the first to use projection pursuit technology to deal with training set of sample dimensionality reduction and in accordance with the sample classification. According to the results of the classification and the best value projection, using trapezoid distribution method extraction fuzzy rules, producing three types of fuzzy membership function. Finally, in accordance with Fuzzy nearness value to determined the sample under the credit level. The results of experiment showed this method of classification accuracy than the traditional credit evaluation techniques have markedly improved, the reliability is better, has a good application value.

On coprime factorization for linear system with finite discrete jumps

This paper deals with the problem of active fault-tolerant control design for a class of linear systems with time-delays. For a class of linear time delay system, firstly, adaptive observer was designed to estimate fault, then a new active fault-tolerant controller is designed which can guarantee that the controlled system keep stable in faulty status, thus the active fault-tolerant control design for this systems is realized. Finally, A numerical example is given to illustrate the efficiency of the proposed design method.