Correlation Analysis Of Complex Thermal System Based On Data Fusion Technology

As the renewable energy represented by the wind incorporated into the grid with large-scale integration,thermal power unit load frequency control task is increasingly difficult.Under the new energy structure,higher and higher requirements are placed on the control quality of the thermal power unit control system.The traditional thermal power unit optimization theory relies on the establishment of the system model and the optimization algorithm.However,the difficulty of modeling the complex thermal power system and the calculation cost of the optimization process often lead to poor practical application and low efficiency.In response to this problem,this paper mainly studies the method of thermal power unit optimization based on data fusion.At present,with the continuous improvement of power plant informatization level,Safety instrumented system is widely used in thermal power plants.The SIS system stores the historical data of the long-term operation of the unit,which including all measuring points of the unit and all working conditions.It provides massive data information and experimental verification platform for the study of thermal power unit optimization based on data fusion.This article focuses on the thermal signal multi-scale related issues as follows:1.The correlation analysis of thermal signals based on wavelet multi-scale decomposition is proposed.The effectiveness of this method is verified by taking the extraction of feed-forward signals of superheated steam temperature control system as an example.The traditional method of selecting feed-forward signal mainly depends on the mechanism analysis and human experience.The effect is not ideal and inefficient taking the complexity of the steam temperature object into account.This article uses correlation analysis method based on multi-scale filter calculating and analyzing the correlation of the mass steam temperature data.The analysis shows that the inlet temperature of the first desuperheater is related to the high temperature of the outlet steam temperature of the superheater and the correlation coefficient comes to 0.89.It is finally selected as a feed-forward signal to the steam temperature control system which avoiding the complexity and difficulty of modeling mechanisms.2.The thermal target pattern recognition method based on wavelet transform and D-S evidence theory is proposed.The effectiveness of the proposed method is verified by the identification of the combustion disturbance in thermal power plant.In the past,the thermal engineering target recognition often ignored the multi-scale correlation between the thermal signals and directly used the field data as the evidence signal to fuse the data,which often resulted in the recognition being inaccurate or unrecognizable.In this paper,the thermal target pattern recognition method using genetic wavelet transform and D-S evidence theory.The specific frequency information of the affected signal is extracted and reconstructed in the time domain and dynamically aligned before being used as evidence for data fusion.The evidence support for the target model is 20% more than the direct evidence theory,so as to effectively solve inaccurate target recognition or can not identify the problem caused by multi-scale and non-linearity of thermal signals.