Luan-Ying Zhang

The research and application of data mining in power plant operation optimization

As the development of electric industry, more and more real-time data are sent to database by DAS and DCS and large amounts of data are accumulated. Abundantly valuable knowledge exists in the history data and it is hard to find and summarize them in traditional way. This paper proposes the operation optimization based on data mining in power plant. The basic structure of the operation optimization based on data mining is established and the fuzzy association rule mining is introduced to find the optimization target value from quantitative values of the equipments in power plant industry process. Based on the history data of a 300MW unit, the optimal values of the operating parameters are found out by data mining techniques. The optimal values are provided to the operators to guide the operation online and the excellent performance is achieved in the power plant.

Study on mathematical model of coordinated control system for supercritical units

In order to achieve the ability of primary frequency regulation better, the boiler-turbine coordinated control system for the large capacity unit has become an important research in thermal control field, using the mechanism mathematical model for reference, utilizing the method of pressure node, the dynamic simulation mathematic model of the coordinated control system for supercritical units is given, analyzing and computing the important parameters of the model, including the heat storage coefficient of the evaporating node and the superheating node. Lastly the model is simulated with MATLAB software, through the studies of the simulation curves, demonstrating the validity of the dynamic simulation mathematic model.

Application of Internal Model and Self-Adaptive PSD Controller in the Main Steam Temperature System

In view of large time delay and uncertainty of the main steam temperature system in a power plant, combining with internal model controller, a control method based on adaptive PSD is proposed. The advanced control algorithms have the advantages of the simple algorithm and the small calculating amount. Compared with the conventional cascade PID control system, the simulation results show that the self-adaptive PSD control system has better control performance on adaptability, such as to decrease the overshoot, shorten the settling time, accelerate the response, etc.

Research and application of data mining in power plant process control and optimization

As more and more real-time data is sent to databases by DAS, large amounts of data are accumulated in power plants. Abundant knowledge exists in historical data but it is hard to find and summarize this in a traditional way. This paper proposes a method of operation optimization based on data mining in a power plant. The basic structure of the operation optimization based on data mining is established and the improved fuzzy association rule mining is introduced to find the optimization values from the quantitative data in a power plant. Based on the historical data of a 300MW unit, the optimal values of the operating parameters are found by using data mining techniques. The optimal values are provided to guide the operation online and experiment results show that excellent performance is achieved in the power plant.

Single-neuron PSD adaptive control based-on Smith predictor for the main steam temperature

The main steam temperature process is regarded as a controllable process with large dead time and immeasurable disturbances. A new single-neuron PSD self-adaptive algorithm that based on Smith predictor was proposed, i.e., a compound controller was combined with Smith predictor and single-neuron PSD self-adaptive controller. Compared with conventional PID controller, the results of simulation show that the system has favorable dynamic properties and good adaptability. The variety of main steam temperature can be maintained in a very small range and it can manipulate variable changes smoothly under large-scale variety of load. Therefore, the scheme has promising application prospects.

Online learning algorithm for sparse kernel partial least squares

we present an improved online learning algorithm for sparse kernel partial least squares, this algorithm improves current methods to kernel-based regression in two aspects. First, it operates online at each time step when it acquires a new input support vector, performs an update and drop out the old data to adapted process changes. Second, it effectively reduces the dimension of feature space and accelerates the speed of training. The simulation results show the improved algorithm has good predict precision and generalize ability, and particularly useful in applications requiring on-line or real-time operation.

Adaptive inferential feed-forward control algorithm and application with reduced model

In the process of industrial production, the system may be affected by many external factors, which can be equivalent to many measurable or immeasurable disturbances. The adaptive inferential feed-forward control algorithm adopts reduced model to design controller which resolves complexity of the adaptive algorithm when the order of model is unknown or too high. Meanwhile, the regularization technique is used to translate the unknown dynamic process into bounded disturbance and the relative dead zone technique is involved to identify parameters of the system, which guarantees bounded stability of the self-tuning control system. Through combining adaptive control, inferential control feed-forward control and adaptive prediction, the algorithm effectively eliminates the influence of measurable and immeasurable disturbances on the system. Finally, the validity and practicability of this algorithm is substantiates by the simulation result of superheated steam temperature control system.

A New Method’s Application Study on the Main Steam Temperature System

When the temperature of main steam is taken as the plant of control, it will be a typical control plant of large dead-time and parameter-varying process. In this paper a new single-neuron PSD self-adaptive algorithm that based on Smith predictor was proposed, i.e., a compound controller was combined with Smith predictor and single-neuron PSD self-adaptive controller. Compared with conventional PID controller, the variety of main steam temperature can be maintains in a very small range and manipulated variable changes smoothly under large-scale variety of load. The results of simulation show that the system has favorable dynamic properties and good adaptability. Therefore, the scheme has hopeful application prospect.Copyright

Study on Applications of State Feedback Based on NN Observer in Super-Heated Steam Temperature Control

It is difficult for the conventional PID control to adjust with the change of dynamic characteristics of the plants. By combing the state feedback based on Elman neural network state observer and the conventional PID, a new control system was presented in this paper. The unmeasurable states of the system was reconstructed through NN observer, the adaptivabilty of system was improved by state feedback. A simulation for power plant super-heated steam temperature control system using presented method is carried out, and resulting in that the control system performance is better than the conventional cascade control system.Copyright