| Generating unit in thermal power plant is a complicated multi-variable object under control with great uncertainty. It is hard to design a system with conventional control strategy, which can not only adapt to changes of the load in power network quickly, but also make frontal pressure maintain nearby the rating. Multi-variable generalized predictive control strategy has wide prospect and study value for its beneficial characteristics such as wide application and strong robustness, which can make up the deficiencies quoted above efficiently. Improving the control method of generating unit in thermal power plant can strengthen coordinate control performance of generating unit and enhance control quality of generating unit, so this has very important reality meaning.Because of the adoption of decoupling design and dispersive optimization here, synthetical design of a multi-variable system is converted to separate design of a few single-variable systems. So calculations become very easy, and the method is convenient for wide application. Direct algorithm of generalized predictive control for multi-variable system is also analyzed here. It is much easier to achieve online for its simplification. During the design, system identification is introduced to constitute its corresponding self-tuning algorithm, which is aimed at the instance that system parameters are not known or change slowly with time. This algorithm can improve the stability, speediness and robustness of system, and simulation results demonstrate the effectiveness of GPC, because the algorithm is applied to coordinated control for generating unit in thermal power plant. |
