Modeling And Research Of Large Complex Equipment Based On Data Driving

Simulators are playing an increasingly important role in new power plants or already in operation.It can train employees to familiarize themselves with the operation interface and operation process before commissioning;it can also reflect the logical correspondence of various equipment.In the various simulators currently used,the basic modeling of various simple or complex equipment using mechanism modeling principles is realized.A data-driven modeling method is proposed.For the system under study,it is not necessary to understand its detailed internal structure and working principle.This method depends only on the input or output data involved in the system operation.Aiming at the difficult modeling process of some complex equipment,this article takes coal mill modeling as an example,and uses data-driven algorithms to build a multivariate linear regression dual-state model structure to realize the modeling of coal mills in online systems.This paper selects the complex equipment of coal grinder based on data to realize mathematical modeling,and mainly completes the following problems:(1)The research status and development of modeling methods for large and complex equipment are studied.Based on the understanding of the main problems facing this research direction at home and abroad,this paper proposes the problems that need to be addressed in this article: due to the complicated mechanism and mathematical principles of complex equipment Makes its modeling process difficult.(2)The source of the data set for the coal mill modeling experiment and the preprocessing of the data before modeling are pointed out.In order to make the modeling results more accurate,the vacant data part and the abnormal data part in the data set are determined to a certain extent.Degree processing,and continue to use K-means clustering algorithm to classify the preprocessed data.Data sets for different operating types of coal mills are modeled offline.(3)The basic theory of least squares and linear regression is studied in detail,and the basic characteristics of the least squares algorithm based on data-driven multiple linear regression based on the complex characteristics of coal mill equipment are derived in this paper.The multivariate linear regression-driven least-squares method is used to model the coal pulverizer in the offline state,and save the weights of the groups obtained by modeling the coal pulverizer in the offline state.(4)Bring the appropriate weight values saved into the simulation system of the online coal mill to participate in the system operation to replace the coal mill model obtained by the previous mechanism modeling.(5)In order to verify the effectiveness of the multivariate linear regression least squares algorithm in the modeling of complex equipment,this paper selects a complex equipment such as a coal grinder to model it through data sets under different working conditions.Parameters to measure the feasibility and effectiveness of the algorithm used,and feedback to the offline model by changing the running status of the online model,and updated to different sets of weight values saved in the establishment of the offline model to adapt to the actual situation of the online simulator Operating conditions.For the modeling of complex equipment in the simulator,the data-driven dual model modeling method has better accuracy and generalization than the mechanism modeling.