Thermal power plant NOx emission optimization control method and system

Abstract

The invention provides a thermal power plant NOx emission optimization control method and system. The method comprises the steps of: determining input variables, carrying out dimension reduction on the determined input variables, and enabling the variables after dimension reduction to serve as final input variables; performing a test according to the final input variables, selecting a certain rectangular flue section near each actual NOx measuring point position of a power plant as a measuring plane, and taking the average value of the data of the measuring points as the accurate NOx emissionvalue under the working condition; and optimizing the NOx accurate emission value data under the working condition by using a least square support vector machine to obtain an LS-SVM model, and evaluating the model by using a test sample to obtain a NOx generation rule. Based on NOx data which is accurately measured and has a causal relationship, dimension reduction is carried out on variables in advance, optimization is carried out by adopting a least square support vector machine algorithm, an accurate NOx emission model is obtained, and then an optimal mode is searched for to control NOx generation and guide operation.

Description

technical field [0001] The disclosure belongs to the technical field of thermal measurement and artificial intelligence, and in particular relates to a method and system for optimal control of NOx emissions in a thermal power plant. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] Pollutant emission standards for thermal power plants are becoming increasingly stringent, and NOx is one of the main pollutants, so controlling its emission has become a top priority. Although SCR denitrification technology can be used to control NOx emissions, it is undoubtedly cost-effective to control the amount of NOx generated from the source. [0004] The formation mechanism of NOx when pulverized coal is burned in the boiler is very complicated, not only related to coal type and coal quality, but also related to the air distribution mode of primary air, ...

AI Technical Summary

Problems solved by technology

[0007] (1) NOx as an output variable is not easy to measure accurately

At present, as the capacity of the unit increases to 600MW and 1000MW, the cross-sectional area of ​​the flue gas increases accordingly. The distribution of NOx on the cross section is extremely uneven, and sometimes the overall NOx emission increases while the DCS measuring point data becomes smaller. At present, some studies only use a single NOx measuring point data on the DCS data as an important output variable for fitting Regression will definitely cause distortion, and the correct internal laws cannot be obtained, let alone accurate prediction and control

[0008] (2) NOx is a combustion product, and all input variables interact in the furnace, while the NOx measurement point is located in the boiler tail flue, that is to say, the causal relationship with all input variables at a time point is actually certain after the time point NOx data at the same time point, rather than the NOx data on the DCS at the same time point, the current research does not take this time difference into account

[0009] (3) There are many input variables, and there may be multiple correlations between input variables. The model is extremely complex and may easily cause calculation divergence or fall into a local minimum value, and the correct generation law cannot be obtained.

Images