Method and system for detecting NOx concentration at scr inlet under multiple boundary conditions

Abstract

The invention discloses a method and system for detecting NOx concentration at an SCR inlet under multiple boundary conditions. The method includes the following steps: determining influencing factors related to the NOx at the SCR inlet according to system mechanism analysis; collecting historical operation data and determining input variables; Cluster the input variables under multiple boundary conditions by the KMeans algorithm to obtain a training data set under multiple boundary conditions; use the SVR method to establish the optimal SCR inlet NOx dynamic prediction model under each of the various boundary conditions; obtain the current moment and The operating data related to NOx at the SCR inlet determines the predicted value of NOx at the SCR inlet corresponding to the boundary conditions at the current moment. The invention respectively establishes inlet NOx prediction models under various boundary conditions such as stable load, lifting load, mill start-stop, and purging, and can effectively predict the SCR inlet NOx concentration under variable working conditions.

Description

technical field [0001] The invention relates to the technical field of flue gas denitrification in coal-fired power plants, in particular to a method and system for detecting NOx concentration at an SCR inlet under multiple boundary conditions. Background technique [0002] With the promulgation of a series of national policies and regulations, such as "Air Pollution Prevention and Control Law", "Emission Standards of Air Pollutants for Thermal Power Plants", "Regulations on the Administration of Collection and Use of Sewage Discharge Fees", "Comprehensive Implementation of Ultra-low Emissions and Energy-saving Transformation of Coal-fired Power Plants" The Air Pollutant Emissions of Coal-fired Power Plants has been brought into strict supervision, and each power plant has successively carried out retrofits for ultra-low emission of flue gas. After ultra-low emission transformation, the NOx emission concentration of coal-fired power plants must be lower than 50mg / m 3 , to d...

AI Technical Summary

Problems solved by technology

However, this measurement method has disadvantages such as high investment cost, long time-consuming analysis of data, and serious lag in the feedback of measured values ​​in the measurement process, which will inevitably make it difficult to accurately display NOx at the inlet of the denitrification system in real time, and ultimately lead to the control effect of the denitrification system on NOx not ideal

When the load of the unit changes, the NOx at the inlet of the denitrification reactor will fluctuate greatly. If the amount of ammonia injection is insufficient, the NOx emission will increase or even exceed the standard; 3 The reaction produces ammonium bisulfate and ammonium sulfate, which reduces catalyst sparks, causes dust accumulation and corrosion in the air preheater, and affects the safe operation of the boiler. At the same time, ammonia escape increases operating cost waste and secondary environmental pollution

[0004] In order to solve the problem of inaccurate measurement and measurement lag of the NOx value at the reactor inlet due to the nonlinearity, large delay, and large inertia of the existing denitrification system, intelligent methods such as the least square method, neural network, and fuzzy method are usually used to carry out the denitrification system. The modeling of inlet NOx, but these methods often only establish a unified model, which can predict the inlet NOx concentration under stable conditions, but for coal-fired power plants, there are problems such as load reduction, mill start-stop, and purge These working conditions make the operation process of the denitrification system very complicated, so a unified model cannot solve the inlet NOx prediction problem under various working conditions

Images