大型风机主导机械动态的智能灰箱建模及其线性状态空间表征

随着风电发展逐渐从量的扩张过渡为质的提高阶段,风机精细化控制日益受到重视,而风机动态的合理建模是其重要基础.本文研究了风机主导机械动态建模并建立其完整的状态空间表征.首先,分析了风机子系统特性,针对气动特性建立气动转矩的分段仿射模型,用于表征气动系统的静态特性.然后,系统地制定了智能灰箱参数辨识步骤,对于多入多出的传动系统设立加权优化目标进行辨识,以获取其蕴含物理意义的状态空间模型,与气动模型合并得到联合状态空间模型.最后,依托FAST的5 MW风机模型进行仿真,验证了建模策略的有效性,仿真结果展现了构建的联合模型对实际动态特性较好的拟合效果.

Intelligent grey-box modeling and linear state-space representation of dominating mechanical dynamics for large-scale wind turbine

As the development of wind power gradually changes from quantitative expansions to qualitative improvements,more attention has been paid to the fine control of wind turbines, and reasonable modeling of wind turbine dynamicsis an important foundation. The dominating mechanical dynamics modeling of wind turbines is studied and its completestate-space representation is established in this paper. Firstly, the characteristics of wind turbine subsystems are analyzedand the piece-wise affine model of the aerodynamic torque is built in terms of aerodynamic characteristics, which is utilizedto characterize the static characteristics. Then, the intelligent grey-box parameter identification procedure is systematicallyformulated. With regard to the multi-input multi-output drive-train system, the weighted optimization objective is utilizedfor identification to acquire its state-space model with the physical meaning, which is combined with aerodynamic modelto form the joint state-space model. Finally, based on the simulation of 5 MW wind turbine model in FAST, the modelingstrategy is verified, and the results show the constructed joint model can achieve a good fitting effect for the actual dynamiccharacteristics.