外文翻译---电站锅炉热效率与NOx排放响应特性建模方法

《外文翻译---电站锅炉热效率与NOx排放响应特性建模方法》由会员分享，可在线阅读，更多相关《外文翻译---电站锅炉热效率与NOx排放响应特性建模方法（18页珍藏版）》请在毕设资料网上搜索。

1、PDF外文：http:/ 中文3748字 英文原文  Modeling and optimization of efficiency and NOx emission at a coal 一 fired utility boiler Huan Zhao,  Pei-bong Wang School of Energy and Environment, SoutheastUniversity, NanjingJiangsu, China , Abstract-In order to improve boiler efficiency and to reduce the NO

2、x emission of a coal-fired utility boiler using combustion optimization,  a  hybrid  model,  by  combining  support  vector regression (SVR) with simplified boiler efficiency model, was proposed to express the relation between operational parameters of the utility

3、boiler and both NOx emission and boiler efficiency. SVR' parameters were determined by the grid search method and 5-fold cross validation method. The predicted NOx emission and boiler efficiency from the hybrid model, compared with that of the BPNN-based hybrid model, shows better agreement with

4、 the measured. Then, based on the hybrid model, the modified center particle swarm optimization (CenterPSO) was employed to optimize the two objectives, the one is minimization of NOx emission and maximization of boiler efficiency and the other one is  maximization  of  boiler  e

5、fficiency  under  NOx  emission constraint. The optimized  results indicate that the proposed method can effectively control NOx emission and improve boiler efficiency. Keywords-NOx；boiler efficiency； combustion optimizationSVR； CenterPSO  I. INTRODUCTION Because of the harm

6、ful effects on environment and human health, nitrogen oxides (NOx) emissions, produced by combustors and engines, have been the subject of restrictive regulations in many countries in recently years1. For instance, in the People s Republic of China, the current NOx emission limit for dry bottom boil

7、ers with a capacity of 300MW and larger is 650 mg/Nm3 (at 6 vol% O2 dry), and it will decrease in the future 2. As a consequence, the control of NOx emission is widely concern as the utilization of fossil fuels continues to increase, especially in coal-fired power plants.  Recently, combustion

8、optimization has been proved to be an effective way to realize low NOx combustion in coal-fired utility setting 2-8. In view of on-line plant data from distributed control system (DCS) and continuous emissions monitoring system (CEMS), the relation between NOx emissions and various operational param

9、eters of the boiler is modeled using artificial intelligence such as neural-network, expert system, fuzzy logic, generalized regression and support vector   regression. And then, by combining the optimization approach such as genetic algorithm, information analysis and ant colony optimization,

10、the NOx model is applied to search the optimal operational parameters to obtain low NOx  combustion. Combustion optimization avoid or postpone large capital expenditures while meeting environmental compliance requirements compared with the relatively expensive flue gasNOx reduction technologies

11、 4.  However, when the operational parameters are regulated to reduce NOx emission, the boiler efficiency is also affected and even decreased 9-12. It is not desirable. Achieving NOx emission reduction target does not necessarily mean poor boiler performance. The boiler efficiency should be mai

12、ntained or improved while the NOx emission is controlled, which is the requirement of real production and management.  In this study, the boiler efficiency is also considered as oneof targets in combustion optimization. The hybrid model,composed of support vector regression (SVR) and simplified

13、boiler efficiency model, was presented to predict NOx emission and boiler efficiency. Subsequently, the model was incorporated into the modified center particle swarm optimization approach to find the optimal operational parameters of the boiler which result in high boiler efficiency and low NOx emi

14、ssion. The aim of this study is to propose an alternative way to meet the needs of high boiler performance and low NOx emission behavior by using combustion optimization.   II. MODELINGFORNOX EMISSIONANDBOILER EFFICIENCY A. Experimental Data The model was established based on the experiment dat

15、a which are from 2. The experiments have been carried on in a 600-MW capacity pulverized coal-fired, dry bottom boiler. The tilting fuel and combustion air nozzles including six primary air burners and seven secondary air burners and two over fire air (OFA) ports are located in each corner of the fu

16、rnace. The burners on A-E levels are put into operation under the rated load and the burners on F elevation are out of service. Total 12 tests have been carried out on this boiler, changing the boiler load, OFA distribution pattern, secondary air distribution pattern, coal quality, nozzles tilting a

17、ngle, respectively, to analyze the emission characteristics of the tangentially fired system. For a more detailed description of experimental setup and operating conditions, please refer to 2. The NOx emission, the unburned carbon and oxygen content in flue gas are illustrated in Table 1.   &nb

18、sp; B. Support Vector Regression Support vector regression (SVR) has been used to solve a nonlinear regression estimation problem by introducing the alternative loss function. Its basic idea is to map the input data x into a high-dimensional feature space  F by nonlinear mapping?, to yield and

19、solve a linear regression problem in this feature space. By this method, the unknown function with a tolerance epsilon band between inputs and output can be obtained. Considering the complete SVR theory and equations, please refer to 13.  C. Hybrid Model for NOx Emission and Boiler Efficiency T

20、he prediction process of NOx emission and boiler efficiency by SVR and analytical model is schematically shown in Fig.1.In hybrid model, three SVR models having 30 inputs in the view of physical analysis were employed to model the relationships between operational parameters and combustion products

21、such as NOx emission, unburned carbon and oxygen content in flue gas.These inputs were the total air flow rate, six elevations of secondary air opening value, two OFA damper opening values, wind box pressure-measuring point, five coal feeders opening value, five air flow rate of mills, nozzle tilt, outlet air temperature of air preheater, the boiler load