Dimensionality Reduction Design of Operating Conditions for Utility Boiler Combustion Adjustment Tests

To address the contradiction between test efficiency and data completeness, a structured understanding of the affiliation and hierarchical relationships among the regulating variables of air-staged combustion boilers was established. A method for ranking these regulating variables based on monitoring parameter sensitivity was developed, which further decoupled the variables sharing the same affiliation and hierarchy. Based on the test data of individual regulating variable, a data trend analysis method was proposed. Results show that the dimensionality reduction design of operating conditions provides greater flexibility in the execution timing of combustion adjustment tests. It meets the demand for enhancing operational data quality under the frequent interaction between the "fuel" and "electricity" markets, and holds significant importance for constructing self-learning and self-optimizing intelligent control systems.