Research on the steam power system operation condition assessment method based on PCA-AE

The traditional steam power system operation condition assessment method relies on expert knowledge and is difficult to achieve complete objectivity. Most data-driven methods require training models with system anomaly data. In response to the above issues, this article proposes the A Novel PCA-AE Steam Power System Operation Condition Assessment Method. This method utilizes the PCA model to obtain assessment indicator parameters, and uses the Autoencoder to automatically assign weights and reconstruct parameters. Finally, the vector similarity principle is used for quantitative assessment. Based on the simulation data of marine nuclear power system steam turbine and the experimental data of supercharged boiler, the feasibility and generalization ability of this method are verified. The new method can meet the requirements of different steam power systems, including the nuclear power system secondary circuits, and provide accurate quantitative assessment results of operating conditions to guide maintenance work. Meanwhile, this method does not rely on expert knowledge and abnormal data training sets, and has fewer limitations compared to traditional methods.