Grey Prediction Extension Control Strategy For Generation Units

The rapid development of modern power system brings good economic returns, but at the same time, it makes the stability problem of power system more complicated and serious. As the important measures to improve stable operation ability and to enhance power system stability, excitation system controller and governor controller of generator unit have been widely used. In view of shortcomings and limitations of existing control methods, the study on new control strategy for generator unit is still important and realistic.By the reviewing and analyzing about the study status quo of power system stability control, it is achieved that the conventional linear control methods are disillusionary because of the excessive dimensions, diverse operating modes and intense nonlinearity of modern power system, all kinds of nonlinear control methods based on modern control theory are discouraging because of the complex algorithms and poor real-time operation, and fuzzy control and neural network control are not satisfied because the design of rules or knowledge base is complex and the engineering practicality are poor. Grey prediction control is a pre-control methods, and extension control is a real-time method which does not require accurate mathematical model. The organic integration of two methods will get better control effects.As a base combining grey prediction and extension control method, the basic principles of grey prediction are introduced in brief. For the original raw data containing negatives, a simple and easy-to-realize transform method is used, and an example shows the effectiveness of the method. The extension methods are introduced in brief. On this basis, the widespread duality in nature is brought into matter-element category and an extension method namely analogical-dual based on the duality of matter-element is produced. An example shows that the method can translate the solving of an unknown incompatible problem into the solution of another known problem, so as to provide a possible way to solve the complex problem.In order to combine the advantages of grey prediction method and extension control technology, a grey prediction-based extension control (GPEC) method is proposed. The controller design process is given. A modified genetic algorithm (GA) which can enhance the search speed and accuracy is applied to optimize the controller parameters. The simulation study respectively on linear system, time-lag system and nonlinear system indicate that GPEC control method can improve the dynamic performance compared with PID and extension control method.The GPEC method is applied to design excitation system controller, governing controller and integrated controller of generator set. The terminal voltage of generator is taken as the input of grey predictor of GPEC excitation controller, 5 data are gathered continuously to carry on grey prediction, and the output is taken as the characteristic variable of extension controller. With the duality which exists between governing control system and excitation control system, the GPEC governing controller is designed directly by using the proposed analogical-dual method, which simplified the design process. A multi-input multi-output extension controller is proposed in view of the shortcoming of traditional extension controller which is unable to apply in multi-variable system, and the controller integrating governing and excitation system is designed based on this methodThe time-domain simulation method is applied to examine the control effects of excitation, governing and integrated controller designed by GPEC method. A single-machine infinite-bus power system and a multi-machine system are taken as the object and suitable mathematical models are used to study the controller's effects for large disturbance and small signal under different initial operating modes. The simulation results indicated that the extension controller has good control effect comparing with conventional method, controller's performance is further enhanced after joining the grey prediction and the GPEC control method always has strong robustness in operating condition, model precise degree and the perturbation size.