Intelligent fuzzy modeling of heavy-duty gas turbine for smart power generation

Natural gas-fired combined cycle unit is appropriate alternative of coal-fired unit for clean power generation. To address the dramatic nonlinearity, strong coupling and observable uncertainty of heavy-duty gas turbine, which is deemed as core device in combined cycle unit, a data-driven intelligent fuzzy modeling strategy is proposed for smart power generation. Firstly, the collected on-side data used for model identification is preprocessed through wavelet denoising for data cleaning. Then, an intelligent T-S fuzzy identification method is adopted for plant modeling via identifications of antecedent part and consequence part. In the antecedent part identification, an automatic fuzzy C-means clustering approach is presented to adaptively divide data space under different operation conditions. Besides, the cluster centers modification, and clusters merging are innovatively used to promote rationality of the clustering result for simpler modeling. Furthermore, a simultaneous flower pollination algorithm is proposed to acquire consequence parameters in each sub-model. Finally, the presented intelligent modeling strategy is applied to a heavy-duty gas turbine system in combined cycle unit for simulation validation. In the comparative experiments, fitting errors of the proposed intelligent fuzzy modeling approach are almost half of that of other seven counterparts while corresponding modeling times are at least 1s faster than others. Therefore, the intelligent fuzzy modeling method reveals competitive precise and rapidity.