Article

Modelling and prediction of NOx emission in a coal-fired power generation plant

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

In this paper NOx emissions modelling for real-time operation and control of a coal-fired power generation plant is studied. Three model types are compared. For the first model the fundamentals governing the NOx formation mechanisms and a system identification technique are used to develop a grey-box model. Then a linear AutoRegressive model with eXogenous inputs (ARX) model and a non-linear ARX model (NARX) are built. Operation plant data is used for modelling and validation. Model cross-validation tests show that the developed grey-box model is able to consistently produce better overall long-term prediction performance than the other two models.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... In the literature, the NO x models for coal-fired utility boilers are divided into three categories: the analytical NO x models based on detailed chemistry schemes (Choi & Kim, 2009;Dez et al., 2008;De Soete, 1975;Hill, Douglas Smoot, & Smith, 1985;Nordin, Yunus, & Ani, 2014;Williams, Pourkashanian, Bysh, & Norman, 1994); the data-driven NO x models based on artificial intelligence, such as neural networks (Chu et al., 2003;Smrekar, Assadi, Fast, Kutrin, & De, 2009;Zhou, Cen, & Fan, 2004, 2005, the support vector regression (Lv, Liu, Yang, & Zeng, 2013;Si et al., 2009;Zhou, Zhao, Zheng, Wang, & Cen, 2012, etc.; the grey-box NO x models based on simplified chemical kinetics and available in situ data (Li, Thompson, & Peng, 2004). Most NO x models in the first category are extremely complex and static, so they are not suitable for on-line operation purposes. ...
... As indicted by Li, Thompson, et al. (2004), grey-box NO x models did well in overall long-term prediction performance. The fact that grey-box models of NO x formation are generally superior to most black-box models is due to the embedded fundamental NO x formation mechanisms in the grey-box model structure working as a priori knowledge. ...
... The current value of a model output y τ ( ) is assumed to be linear combinations of its previous output time series (the order of outputs is n y ) as well as highly relevant n forms of fundamental elements (each form with an order of n j ). The grey-box model structure (p n n n = × + j y terms of φ τ ( ) i (i p = 1, 2, … ) including fundamental elements as well as orders) and the parameters θ i are identified by the fast recursive algorithm described in Li, Thompson, et al. (2004). Meanwhile, the model residual is denoted by ζ τ ( ). ...
Article
This paper focuses on developing a control-oriented coal-fired utility boiler model for advanced economical Low-NOx combustion (ELNC) controller design. Two boiler combustion models are proposed in this paper: one is a mathematical model describing the key dynamics of the real-time boiler thermal efficiency and the furnace one-dimensional NOx concentration distribution under conventional fuel and overfire air operations; the other recast from the first model is a control-oriented grey-box model with a data-driven furnace combustion submodel. Simulation studies on static and dynamic properties of the first mathematical model indicate that the model can function as a real-time simulator for both advanced boiler combustion control laws testing and generating training and validation data for the control-oriented grey-box model. At the end of this paper, the control-oriented grey-box modelling procedure as well as an optional discrete time linear state-space model are summarised to facilitate model-based advanced combustion controllers design.
... However, as these intermediate variables are normally not measurable, NO x emissions are often estimated using a model which relates various operational inputs and measurements to the NO x emission outputs. A number of models have been studied for different thermal power plants, including coal-fired (Li, Thompson and Peng, 2004), oil-fired (Li and Thompson, 1996), and both oil-and methane fired (Ferretti and Piroddi, 2001). NO x formation is a highly complex process, which can often be described by a set of partial and ordinary differential equations (PDEs and ODEs), based on chemical and physical laws (De Soete, 1975). ...
... Due to the highly nonlinearity in the combustion process, simple linear models based on the input/output relations are not suitable for plant optimal operation and control (Li, Thompson and Peng, 2004). Nonlinear regression models have been developed so far to improve the model performance as well as the model transparency (Li, Thompson and Peng, 2004), the model complexity has however not been significantly reduced. ...
... Due to the highly nonlinearity in the combustion process, simple linear models based on the input/output relations are not suitable for plant optimal operation and control (Li, Thompson and Peng, 2004). Nonlinear regression models have been developed so far to improve the model performance as well as the model transparency (Li, Thompson and Peng, 2004), the model complexity has however not been significantly reduced. As universal approximators, artificial neural networks (ANN) such as the Multi-Layer Perceptron (MLP) and radial basis functions (RBF) networks have been widely used in the industry, including application in NO x emissions in thermal power plants (Ferretti and Piroddi, 2001). ...
Conference Paper
This paper investigates neural network based estimation of NOx emissions in a thermal power plant, fed with both oil and methane fuels. Two types of neural network namely a novel 'eng-genes' architecture and a Multilayer Perceptron (MLP) have been developed, both being optimised using genetic algorithms. Due to the local nature of the NOx generation process, operational information on the burner cells of the combustion chamber has been considered. Neural networks, with different numbers of hidden nodes have been tested on a set of three-dimensional data of the simulated combustion chamber. It is shown that, the proposed 'eng-genes' neural network can produce accurate estimations with better generalisation performance than MLP.
... The formation of thermal NO is dependent on the temperature of the PC furnace. When the temperature in PC furnace is high (typically greater than 1500 C), free radicals such as O and N from atmospheric oxygen and nitrogen are abundant and start forming NO (Li et al. 2004). Major reactions forming thermal NO are given as below (Li et al. 2004): ...
... When the temperature in PC furnace is high (typically greater than 1500 C), free radicals such as O and N from atmospheric oxygen and nitrogen are abundant and start forming NO (Li et al. 2004). Major reactions forming thermal NO are given as below (Li et al. 2004): ...
... The Arrhenius form can be used to calculate reaction rate, k Ni or k -Ni for each reaction. The rate of thermal NO formation is therefore calculated by the following equation (Li et al. 2004), under the assumption of quasi steady state: ...
Chapter
Many thermal power generation plants rely on combustion of pulverised coal carried out in large furnaces. Design and improvement of these furnaces can be effectively assisted by using numerical modelling with Computational Fluid Dynamics (CFD) techniques to develop a detailed picture of the conditions within the furnace, and the effect of operating conditions, coal type, and furnace design on those conditions. The equations governing CFD models of pulverised coal combustion are described, with a focus on sub-models needed for devolatilisation, combustion and heat transfer. The use of the models is discussed with reference to examples of CFD modelling of brown coal fired furnaces in the Latrobe Valley in Australia and black coal fired furnaces described in the literature. Extensions to the CFD models that are required to tackle specific industrial and environmental issues are also described. These issues include control of NOx and SOx emissions and the effect of slagging and fouling on furnace and boiler operation.
... It is therefore necessary to develop techniques for NOx prediction using more effective inputs, such as characteristic information of the flame, which relates directly to the combustion process. Some research has been devoted to establish CFD models for NOx estimation (Li et al., 2004, Stopford, 2002. CFD modeling is, however, generally very time consuming and not applicable to real-time processes (Li et al., 2004). ...
... Some research has been devoted to establish CFD models for NOx estimation (Li et al., 2004, Stopford, 2002. CFD modeling is, however, generally very time consuming and not applicable to real-time processes (Li et al., 2004). ...
Article
Full-text available
This article presents a methodology for predicting NOx emissions from a biomass combustion process through flame radical imaging and deep learning (DL). The dataset was established experimentally from flame radical images captured on a biomass-gas fired test rig. Morphological component analysis is undertaken to improve the quality of the dataset, and the region-of-interest extraction is introduced to extract the flame radical part and rescale the image size. The developed DL-based prediction model contains three successive stages for implementing the feature extraction, feature fusion, and emission prediction. The fine-tuning based on the prediction is introduced to adjust the process of the feature fusion. The effects of the feature fusion and fine-tuning are discussed in detail. A comparison between various image- and machine-learning-based prediction models show that the proposed DL prediction model outperforms other models in terms of root mean square error criteria. The predicted NOx emissions are in good agreement with the measurement results.
... The same group performed a similar study [17] employing averaged values by the multiple linear regression approach and support vector regression. Kang Li et al. [22] compared three model types for NOx prediction, i.e. the grey-box model, ARX and NARX models. The grey-box model incorporated 'a priori' knowledge in terms of fundamental elements and 'posterior' knowledge acquired from model order identification. ...
... From Fig. 4 it can be noticed that model order and the variable groups have an important impact on the model's performance. Looking at the line Group 1, representing only primary variables, MAE TE,avg and MAE TE,max become less than 20 and 30 mg/m 3 respectively, at around model order 4. Group 1's optimal performance is selected at model order 10, where MAE TE,avg = 11.55mg/m 3 and MAE TE,max = 16.23 mg/m 3 as represented by Table 4. Hence, the satisfactory prediction of NOx emissions with only primary variables is not achieved, as can also be observed from past publications [15,22]. ...
... Data-driven identification offers an alternative approach to model systems with strong nonlinearity and unknown parameters. It has been employed in various engineering applications, such as power plant emission prediction [9], communication system [10], polymer extrusion processes [11], etc. For modeling MTs, different types of black-box models, such as nonlinear autoregressive moving average with exogenous inputs models (ARMAX) [12], nonlinear autoregressive exogenous (NARX) [13], neural network models [14], [15], and adaptive network-based fuzzy inference system [16], have been used to capture gas turbine dynamics. ...
... To quantify the accuracy of different models, two indices are used to illustrate both the absolute error and the relative error. The first index is the MSE defined in (9). The second index is the Mean Absolute Percentage Error (MAPE), which is defined as [29] ...
... The NO x decoupled approach has been adopted in previous studies concerning NO x modeling in PF boilers [2,36,37]. It is considered a widely approved method for predicting nitrogen oxide emissions, but has never been applied to CFBC modeling, in the best of the authors' knowledge. ...
Article
In this study, a 3D CFD model for the formation of NOx and N2O in a lignite fired 1.2 MWth CFB pilot plant is developed. The decoupled approach (decoupled from combustion simulation) is tested for the minimization of computational cost. As combustion simulation is prerequired, this was achieved through a simplified 3-D CFD combustion model. The developed model is then applied to the pilot-scale 1.2 MWth CFB plant and validated against experimental data. As concerns the NOx-N2O model, an extensive literature review is also carried out for the incorporation of the appropriate reactions network and respective reaction rates expressions. Results show that homogenous reactions are favoured on the lower section of the bed, due to the abundance of fuel devolatilization products. On the other hand, on the upper section, heterogeneous reactions govern nitric oxide formation/reduction. It is found that for the lignite examined in this work, HCN is released in negligible amounts during char combustion. The proposed and validated CFD model for NOx and N2O, is capable of examining the effect of different operational parameters and coal properties on the overall nitric oxides emissions from a CFB combustor, with low computational cost and without the additional expenses for pilot-scale experiments.
... The NOx decoupled approach has been adopted in many studies (Li et al. 2004;Nikolopoulos et al. 2011Nikolopoulos et al. , 2014 owing to its computational low cost. The decoupled approach is considered to be a widely approved method that can effectively predict nitrogen oxide emissions. ...
Article
This work simulates the Megalopolis IV power plant boiler operation in terms of NOx emissions for a wide range of thermal loads, and with different operational patterns of the firing system, i.e., different numbers and variation of the open burners' positions. This work provides feedback on thermal loads that are lower than the current technical minimum (∼55%) using numerical tools (CFD software packages) for both 40 and 35% partial-loads. For the lowest investigated load, it also examines contemporary combusting systems, with the simultaneous firing of raw and stored dry lignite to achieve flame stability. The results of the current study primarily focus on the effect of lower than 100% boiler operation on the NOx emissions because the majority of works in literature investigates boiler NOx emissions only under full-load conditions. Furthermore, it focuses on the variation of the NOx emissions among different partial-load conditions and the influence of the firing system operational patterns on these. To reduce the required computational resources, the decoupled approach is adopted, which distinguishes the main combustion mechanisms from the NOx emission mechanisms themselves. The results indicate that the NOx emissions concentration reduced at 6% dry oxygen molar concentration at the main furnace outlet surface increases as the thermal load gets lower. The results also prove the direct relation between temperature values and NOx emissions because the highest NOx concentrations are tracked in the hot spot areas; primarily attributed to thermal NOx production mechanisms. Additionally, the firing system operational pattern for each specific thermal load affects the NOx emissions concentration, as the results regarding the 35% partial-load case imply. In particular, for the 35% partial-load case, the even distribution and injection of stored dry lignite from all six vapor burners result in lower NOx concentrations compared to the case for which the injection of the stored dry lignite happens through the inlets of the two open vapor burners. In conclusion, the results of this work can provide valuable information about the boiler performance in terms of its NOx emissions behavior under partial-load conditions, and provide potential measures that can be implemented for NOx emissions reduction, especially under low-thermal conditions.
... Existing power systems got overstressed to meet the increased load demands. Though the power generation by the conventional fossil fuel-fired generators is flexible, controllable, and dispatchable, the demerits of these sources are not economic, environmentally unfriendly, and non-sustainable [1]. Moreover, the triple bottom line [2] approach suggests the reduction in global emissions, increasing profits, and achieving maximum benefits for the people. ...
Article
Full-text available
Due to the surge in load demand, the scarcity of fossil fuels, and increased concerns about global climate change, researchers have found distributed energy resources (DERs) to be alternatives to large conventional power generation. However, a drastic increase in the installation of distributed generation (DGs) increases the variability, volatility, and poor power quality issues in the microgrid (MG). To avoid prolonged outages in the distribution system, the implementation of energy management strategies (EMS) is necessary within the MG environment. The loads are allowed to participate in the energy management (EM) so as to reduce or shift their demands to non-peak hours such that the maximum peak in the system gets reduced. Therefore, this article addresses the complication of solutions, merits, and demerits that may be encountered in today’s power system and encompassed with demand response (DR) and its impacts in reducing the installation cost, the capital cost of DGs, and total electricity tariff. Moreover, the paper focuses on various communication technologies, load clustering techniques, and sizing methodologies presented.
... More effective and efficient EED methods can be used to optimize the operation of power systems, therefore reducing the energy consumption and emissions in power generation. In addition, the energy consumption and emissions can be reduced by some quantitative economic methods and mathematical models, such as differential electricity pricing[117], modeling and prediction[118], economic evaluation[119], and the quantification of carbon emis- sions[120]. ...
... During the last decades, several models have been proposed to predict and reduce the NOx emissions. Generally, they can be categorized into three major categories, i.e., com-putational fluid dynamics (CFD) models (Fan et al., 2001;Khoshhal et al., 2010;Baek et al., 2014), statistical heuristic models (Adali et al., 1999;Ligang et al., 2010;Li et al., 2013;Ahmed et al., 2015) and gray-box models (Pearson and Pottmann, 2000;Li et al., 2004). Due to the time and effort required and complexity involved in the development of CFD models, statistical heuristic models based on process data history find their places quite frequently in the modeling of NOx emissions (Ahmed et al., 2009). ...
Article
Full-text available
This paper presents a novel Teaching-Learning-Self-Study-Optimization (TLSO) algorithm which is not only fast converging according to the number of iterations, but also relatively consistent in converging with high accuracy to the global minimum in comparison with some other algorithms. The original Teaching-Learning-Based Optimization (TLBO) gives uniformly distributed and randomly selected weight to the amount of knowledge to a learner at each phase, i.e., teacher phase and learner phase. This uniformly distributed and randomly selected weight causes the algorithm to converge the average cost of learners in a moderate number of iterations. Li and his coworkers intensified the teacher and learner phases by introducing weight-parameters in order to improve the convergence speed in terms of iterations in 2013 and called it Ameliorated Teaching-Learning-Based Optimization (ATLBO). The criterion of a good evolutionary optimization algorithm is to be consistent in converging the cost of the objective function. For this, it should include intensification for local search as well as diversification for global search in order to reduce the chances of trapping in a local minimum. Some students naturally tend to study by themselves by the means of a library and internet academic resources in order to enhance their knowledge. This phenomenon is termed as self-study and is introduced in the proposed TLSO’s learner phase as a diversification factor (DF). Various other evolutionary algorithms such as ACO, PSO, TLBO, ATLBO and two variants of TLSO are also developed and compared with TLSO in terms of consistency to converge to the global minimum. Results reveal that the TLSO was found to be consistent not only for a higher number of functions among 20 benchmark functions, but also for NOx prediction application. Results also show that the predicted NOx emissions through LSSVM tuned with TLSO are comparable with the other algorithms considered in this work.
... In order to simulate the NOx formation mechanisms, the NOx decoupled approach was adopted. This approach is widely known and extensively used in many studies due to its low computational cost and its high accuracy [52,53]. According to this methodology, the NOx simulation model can be conducted afterwards the simulation of the combustion phenomena. ...
... NO x emissions from coal-fired power plants exhibit typical nonlinear behavior. To capture this nonlinearity, many nonlinear techniques have been proposed for the prediction of NO x emissions such as Partial Least Squares (PLS) (Lee et al., 2005), the Autoregressive Exogenous model (ARX), Nonlinear Autoregressive Exogenous model (NARX), grey box modeling (Li et al., 2004) and Artificial Neural Networks (Adali et al., 1999;Zhou et al., 2004). NO x emissions from coal-fired power plants have also been predicted and reduced by utilizing support vector framework, e.g., Support Vector Regression (SVR) (Wu et al., 2009), Least Squares Support Vector Regression (LSSVR) (Zheng et al., 2010), a combination of SVR and kernel Principal Component Analysis (kPCA) (Shi, 2012), and Least Squares Support Vector Machine (LSSVM) (Lv et al., 2013). ...
Article
Full-text available
In order to deal with the nonlinear varying behavior of NO x emissions for long term predictions, a real-time recursively updating model is indispensable. In this paper, new recursively updating models are proposed to predict NO x emissions. The proposed real-time models are equipped with an initial LSSVM model and subsequent updating methods to adapt the models with recent changes to process data. The updating methods include solo Least Squares Support Vector Machines (LSSVM) update, solo output bias update, and the combination of these two termed as the LSSVM-Scheme. These models are applied to NO x emission process data from a coal combustion power plant in Korea. Prediction results obtained from the proposed real-time LSSVM models are compared with their counterpart real-time PLS models, which reveal that real-time LSSVM models outperform their counterpart real-time PLS models. Among other models developed in this work, LSSVM-Scheme and solo output bias update based on LSSVM predicts NO x emissions robustly for a long pas-sage of time with the highest accuracy.
... This method is mainly focused on the NOx emission of the exhaust gas. In this regard, numerous algorithms, including statistical regression (Li et al., 2004;Chunlin Wang et al., 2018), support vector machine (Wei et al., 2013;Zhou et al., 2012;Ahmed et al., 2015;Lv et al., 2013), artificial neural network (ANN) (Chu et al., 2003;Ilamathi et al., 2013;Preeti and Sharad, 2013;Jacob and Tuttll, 2019), and deep learning (Li and Hu, 2020;Yang et al., 2020;Tan et al., 2019;Xie et al., 2020;Kang et al., 2017;Wang et al., 2017) are often used to predict the NOx concentration. Although remarkable achievements have been obtained in this area, the time complexity of support vector machines increases exponentially as the sample size increases. ...
Article
Full-text available
A novel method for the prediction of three-dimensional (3D) spatial distribution of NOx in a furnace is proposed and evaluated. Computational fluid dynamics (CFD) simulations are conducted to generate the data sets of 3D NOx spatial distribution. The data sets are partitioned based on NOx generation mechanisms to improve the model accuracy. Combining the Pearson coefficient and mutual information (PMI), the model input variables are optimized by feature selection. The prediction model of 3D NOx spatial distribution in the furnace is established based on extreme learning machine (ELM). The experiments are conducted considering a 350 MW coal-fired boiler with a change in the burner tilt angles under a rated load. The experimental results show that the data-driven method based on PMI-ELM can realize the rapid prediction of the 3D spatial distribution of NOx in the furnace with 12.84% mean absolute percentage error.
... The mechanism-based methods are meaningful for understanding the combustion. A grey-box model based on NOx formation mechanisms (Li et al., 2004) was proposed. A char gasification mechanism model (Liu et al., 2016) was presented for a 600 MW tangentially fired pulverized-coal boiler. ...
Article
NOx emission prediction is important for efficient boiler production and waste control. An adaptive data-driven modeling method is proposed to predict the boiler NOx emissions dynamically. In this method, a linear combination kernel is presented to improve the prediction accuracy of least-square support vector machine. The parameters of the kernel are optimized adaptively by a particle swarm optimization algorithm. Additionally, an adaptive moving time window strategy is presented to maintain model performance. The computational results based on the practical data illustrate that the proposed kernel and the adaptive moving time window strategy are positive and the proposed prediction method is superior to some previous prediction methods.
... In reference (Liukkonen et al. 2012), on the basis of multiple regression analysis combined with rolling window, a soft measurement model of NOx emissions is proposed. In reference (Li et al. 2004), the linear ARX model, the nonlinear ARX model, and the gray box model are also used to predict NOx emission concentration. Although the regression model can resolve the linear and nonlinear problems in NOx concentration forecast, it still requires complex data preprocessing. ...
Article
Full-text available
The concentration of nitrogen oxide (NOx) emissions is an important environmental index in the cement production process. The purpose of predicting NOx emission concentration during cement production is to optimize the denitration process to reduce NOx emission. However, due to the problems of time delay, nonlinearity, uncertainty, and data continuity in the cement production process, it is difficult to establish an accurate NOx concentration prediction model. In order to solve the above problems, a NOx emission concentration prediction model using a deep belief network with clustering and time series features (CT-DBN) is proposed in this paper. Particularly, to improve data sparsity and enhance data characteristics, a clustering algorithm is introduced into the model to process the original data of each variable; the time series containing delay information are introduced into the input layer, which combines previous and current variable data into time series data to eliminate the influence of the time delay on the prediction of NOx emission concentration. In addition, restricted Boltzmann machine (RBM) is used to extract data features, and a gradient descent algorithm is used to reversely adjust network parameters to establish a deep belief network model (DBN). Experiments prove that the method in this paper has higher accuracy, stronger stability, and better generalization ability in predicting NOx emission concentration in cement production. The CT-DBN model realizes the accurate prediction of NOx emission concentration, provides guidance for denitration control, and reduces NOx emissions.
... A variety of mechanism models have been derived for the boiler combustion efficiency using the physical principles. For example, the superiority of a gray-box system identification method to the computational fluid dynamics (CFD) method for NOx emission modeling in a coal-fired power plant has been demonstrated [3]. An energy balancing-based model has been employed in Reference [4] to propose an effective method for the boiler combustion efficiency estimation. ...
Article
Full-text available
A data-driven modeling method with feature selection capability is proposed for the combustion process of a station boiler under multi-working conditions to derive a nonlinear optimization model for the boiler combustion efficiency under various working conditions. In this approach, the principal component analysis method is employed to reconstruct new variables as the input of the predictive model, reduce the over-fitting of data and improve modeling accuracy. Then, a k-nearest neighbors algorithm is used to classify the samples to distinguish the data by the different operating conditions. Based on the classified data, a least square support vector machine optimized by the differential evolution algorithm is established. Based on the boiler key parameter model, the proposed model attempts to maximize the combustion efficiency under the boiler load constraints, the nitrogen oxide (NOx) emissions constraints and the boundary constraints. The experimental results based on the actual production data, as well as the comparative analysis demonstrate: (1) The predictive model can accurately predict the boiler key parameters and meet the demands of boiler combustion process control and optimization; (2) The model predictive control algorithm can effectively control the boiler combustion efficiency, the average errors of simulation are less than 5%. The proposed model predictive control method can improve the quality of production, reduce energy consumption, and lay the foundation for enterprises to achieve high efficiency and low emission.
... The models accounted for thermal and fuel NOx dynamics with simplifying assumptions (Dal Secco et al., 2015). Li et al. (2004) investigated formation mechanisms and system identification methods for NOx emission modeling. Experimental results indicated the good performance of the models under certain constraints although the dynamical models are typically complex, inefficient, and hardly applicable in modern control systems due to high complexity and uncertainty of combustion processes and parameters. ...
Article
Nitrogen oxide (NOx) emissions are major pollutants of coal-fired boilers. An adaptive nonlinear model-predictive control approach is presented to reduce NOx emissions of power plant boilers. Firstly, the boiler load and the NOx emissions are dynamically predicted by a differential evolution-based least-square support vector machine. Subsequently, based on data-driven prediction modeling, a nonlinear optimization model, with load and capacity constraints, is proposed for NOx emission minimization. Finally, a differential evolution algorithm is used to solve this optimization problem and obtain the optimal control variable settings. Experimental results based on practical data indicate that the proposed approach exhibits a promising performance in the prediction of the boiler load and NOx emissions. Compared with that obtained using the normal control strategy, the proposed approach can reduce NOx emissions by 3.2% and 4.3% under increasing and decreasing loads, respectively.
... Air pollution has recently become a critical, global issue [1]. In response, environmental regulations have been tightened to reduce the emissions of chemical impurities (such as NO X , SO x , CO, volatile organic compounds (VOCs), and particulate matter (PM)) from power plants, boilers, and mobile sources [2,3]. Among the numerous air pollutants, nitrogen oxides (NO X : NO, NO 2 , and N 2 O) are extremely dangerous as they can easily disperse over long distances and form secondary PM 2.5 by reacting with water vapor, which causes acid rain and smog, contributes to global warming [4,5], and can deeply penetrate human lungs, causing adverse health effects such as increased cardiovascular and respiratory morbidity [6]. ...
Article
Full-text available
In this study, we synthesized V2O5-WO3/TiO2 catalysts with different crystallinities via one-sided and isotropic heating methods. We then investigated the effects of the catalysts’ crystallinity on their acidity, surface species, and catalytic performance through various analysis techniques and a fixed-bed reactor experiment. The isotropic heating method produced crystalline V2O5 and WO3, increasing the availability of both Brønsted and Lewis acid sites, while the one-sided method produced amorphous V2O5 and WO3. The crystalline structure of the two species significantly enhanced NO2 formation, causing more rapid selective catalytic reduction (SCR) reactions and greater catalyst reducibility for NOX decomposition. This improved NOX removal efficiency and N2 selectivity for a wider temperature range of 200 °C–450 °C. Additionally, the synthesized, crystalline catalysts exhibited good resistance to SO2, which is common in industrial flue gases. Through the results reported herein, this study may contribute to future studies on SCR catalysts and other catalyst systems.
... A number of conventional mechanism techniques for predicting NO x emission of coal-fired utility boilers based on first principles such as heat and mass balances have been studied. For example, Li et al. [6] developed the NO x emission model using the system fundamentals and parameters identification; Chui and Gao [7] applied computational fluid dynamics based combustion technology to estimate NO x emission; Belosevic et al. [8] studied the NO x formation based on the differential mathematical method. However, these models are usually complicated and time consuming to build because a complete fundamental knowledge of the combustion process should be well known and at times a large number of parameters are required [9]. ...
Article
The co-combustion of coal with printing and dyeing sludge (PDS) in a pulverized coal power plant was simulated based on the computational fluid dynamics method. The co-combustion characteristics and NOX emissions were analyzed to determine the optimal PDS content, moisture content, and secondary air (SA) distribution scheme. This study comprised three main parts. The first part was a comparative analysis of blending coal with different amounts of PDS, which indicated that the overall temperature level of the furnace decreased slightly and the NOX emissions increased gradually as the PDS content increased. The second part was a comparative analysis of blending coal with PDS at different moisture contents, which indicated that the furnace temperature was lower and the NOX emissions were higher with as the moisture content increased. Furthermore, after analyzing the combustion characteristics and NOX emissions, we found that blending with 10% PDS at a moisture content of 40% was the most reasonable strategy based on a constant total lower calorific value. In the third part, SA optimization analysis indicated that the SA distribution clearly affected the combustion characteristics and pollutant emission, where an SA ratio from top to bottom of 3:1:2:4 was the best SA distribution scheme.
Article
Polymer extrusion is a complex process and the availability of good dynamic models is key for improved system operation. Previous modelling attempts have failed adequately to capture the non-linearities of the process or prove too complex for control applications. This work presents a novel approach to the problem by the modelling of extrusion viscosity and pressure, adopting a grey box modelling technique that combines mechanistic knowledge with empirical data using a genetic algorithm approach. The models are shown to outperform those of a much higher order generated by a conventional black box technique while providing insight into the underlying processes at work within the extruder.
Article
Full-text available
In this paper, NOx emission prediction was studied. A simple model based on response surface methodology (RSM) was first put forward. Response surface models are multivariate polynomial models. Four RSM models were tried. The predicted NOx emission was compared with the measured ones. The RSM model with quadratic terms showed the best agreement with the measurement, and had the mean relative error of 1.6719%. The frequency of those samples whose relative error is less than 5% was 96.8610%. The RSM model was simpler than the non-analytic models such as generalized regression neural network and support vector regression. The present study will be an alternative to developing predictive emissions monitoring systems (PEMS).
Article
Coal-fired power plant technologies should provide higher efficiency of energy conversion, reduction of pollutants emission, operation of facilities in a wide range of loads and efficient utilization of variable quality fuels. In order to achieve these tasks, mathematical modeling is regularly used worldwide for optimization of boiler operation. Reduction of pollutants emission is the task of greatest concerns. Among the most important pollutants are oxides of nitrogen and sulfur. Combustion process modifications for NOx control and sorbent injection for SO2 control are cost-effective clean coal technologies, used either standalone or with other methods. An in-house developed computer code was applied for simulation of processes in the 350 MWe boiler furnace, tangentially fired by pulverized lignite. Predictions suggested optimal combustion organization providing the NOx emission reduction of up to 20-30%, with the flame position improvement. Boiler thermal calculations showed that the facility was to be controlled within narrow limits of working parameters. SO2 reduction by injection of Ca-based sorbent particles into the furnace was simulated for different operation parameters. Such a complex approach enables effective evaluation of alternative solutions, considering emissions, flame position and efficiency of furnace processes and the boiler unit.
Article
Using Fluent software, numerical simulation was conducted to combustion and pollutant emission characteristics of a tangentially-fired boiler burning coal blended with different ratios and moistures of textile dyeing sludge. Results show that with the rise of blending ratio of textile dyeing sludge, the overal temperature in furnace drops a little, and the NOx emission rises rapidly first and steady later on, and the turning point is at the blending ratio of 10% textile dyeing sludge; with the rise of sludge moisture, the overall temperature in furnace reduces slightly, and the average outlet temperature of furance with 40% moisture is only 8.11 K lower than the case with 10% moisture; the NOx emission rises with the growth of sludge moisture; considering the furnace combustion and NOx emission characteristics, it is thought to be reasonable to blend 10% textile dyeing sludge with 40% moisture into the coal, and the optimum secondary air distribution from top to bottom is 3:1:2:4. ©, 2015, Shanghai Power Equipment Research Institute. All right reserved.
Article
The accurate and reliable real-time estimation of NOx emission is indispensable for the implementation of successful control and optimization of NOx emission from a coal-fired power plant. We apply a real-time update scheme to least squares support vector machines (LSSVM) to build a real-time version for real-time prediction of NOx. Incorporation of LSSVM in the update scheme enhances its generalization ability for long-term predictions. The proposed real-time model based on LSSVM (LSSVM-scheme) is applied to NOx emission process data from a coal-fired power plant in Korea to compare the prediction performance of NOx emission with real-time model based on partial least squares (PLS-scheme). Prediction results show that LSSVM-scheme predicts robustly for a long passage of time with higher accuracy in comparison with PLS-scheme. We also present a user friendly and sophisticated graphical user interface to enhance the convenience to approach the features of real-time LSSVM-scheme.
Article
Melt viscosity is a key indicator of product quality in polymer extrusion processes. However, real time monitoring and control of viscosity is difficult to achieve. In this article, a novel “soft sensor” approach based on dynamic gray-box modeling is proposed. The soft sensor involves a nonlinear finite impulse response model with adaptable linear parameters for real-time prediction of the melt viscosity based on the process inputs; the model output is then used as an input of a model with a simple-fixed structure to predict the barrel pressure which can be measured online. Finally, the predicted pressure is compared to the measured value and the corresponding error is used as a feedback signal to correct the viscosity estimate. This novel feedback structure enables the online adaptability of the viscosity model in response to modeling errors and disturbances, hence producing a reliable viscosity estimate. The experimental results on different material/die/extruder confirm the effectiveness of the proposed “soft sensor” method based on dynamic gray-box modeling for real-time monitoring and control of polymer extrusion processes. POLYM. ENG. SCI., 2012. © 2012 Society of Plastics Engineers
Article
This study presents a soft sensing model of coal moisture for utility boilers. The model is based on the energy and mass balance of matter in the inlet and outlet of a positive-pressure, direct-firing, MPS-type mill. Compared with the results obtained with a proximate analysis method, the results calculated with the soft sensing method are consistent with proximate analysis data. Meanwhile, the accuracy and reliability of the soft sensing method can be improved by data fusion of the calculated coal moisture from other running mills.
Conference Paper
This paper presents a method for the prediction of NOx emissions in a biomass combustion process through the combination of flame radical imaging, contourlets transform, and radial basis function network techniques. The images of four flame radicals (OH*, CN*, CH* and C2*) are captured using a spectroscopic imaging system. The features of the images are then identified based on the best M-term approximation of contourlet coefficients. The relationships between the features of radical images and NOx emissions are finally established through the use of the Radical Basis Function network. The test results obtained on a biomass-gas fired test rig show the effectiveness of the proposed technical approach for the prediction of NOx emissions.
Conference Paper
Nonlinear identification using a novel neural network paradigm, namely enggenes, is investigated. A set of MATLAB functions for the training and simulation of eng-genes based neural models are described. These functions are then used to investigate the effectiveness of the technique applied to two nonlinear dynamical systems. Experimental data from a pH neutralisation plant and simulation data from a physical model of a CSTR process are used to generate 'eng-genes' models. The results are compared with conventional neural models of these plants, showing that simple neural models with better performance and improved transparency are obtainable using the enggenes paradigm.
Article
The boiler combustion process of coal-fired power plant is a very complicated MIMO system with high nonlinearity and strong coupling. The LSSVM(Least Square Support Vector Machine) is applied to build the boiler combustion model based on the property test data and the nonlinear MPC(Model Predictive Control) is applied to optimize the control of boiler combustion process. The improved ACO(Ant Colony Optimization) is proposed to solve the nonlinear optimization problem of MPC algorithm, which extracts the target individuals dynamically and stochastically to lead the global search of ant colony while carries out the small step search nearby the optimal ant. Case study indicates its effectiveness.
Article
Viscosity represents a key indicator of product quality in polymer extrusion but has traditionally been difficult to measure in-process in real-time. An innovative, yet simple, solution to this problem is proposed by a Prediction-Feedback observer mechanism. A 'Prediction' model based on the operating conditions generates an open-loop estimate of the melt viscosity; this estimate is used as an input to a second, 'Feedback' model to predict the pressure of the system. The pressure value is compared to the actual measured melt pressure and the error used to correct the viscosity estimate. The Prediction model captures the relationship between the operating conditions and the resulting melt viscosity and as such describes the specific material behavior. The Feedback model on the other hand describes the fundamental physical relationship between viscosity and extruder pressure and is a function of the machine geometry. The resulting system yields viscosity estimates within 1% error, shows excellent disturbance rejection properties and can be directly applied to model-based control. This is of major significance to achieving higher quality and reducing waste and set-up times in the polymer extrusion industry.
Article
Increasing penetration of renewable energy sources to the power grid has prompted new ramping scenarios to dispatchable thermal power plants to balance the variability caused by intermittent renewable supplies. With many thermal power plants designed to be base-loaded, ramping of the power output results in increased emission of pollutants. This study develops a dynamic data-driven model of a coal-fired utility boiler that estimates NOx and CO emissions simultaneously. Given a production schedule of a power plant, estimation of NOx and CO emissions for 3 h into the future is performed that can be further utilized in a dynamic optimization algorithm to minimize the emissions over a horizon. It is observed that a dynamic model always has a higher prediction accuracy than a static model, when training and forecasting of the models are concerned. Application of dynamic and steady-state optimization also results in reduced emissions as compared to historical plant emissions.
Article
In this study, porous mullite ceramics were prepared by gel-casting using fused mullite particles as the raw materials, ρ-Al2O3 and polysilicon waste as the binders, and starch as the pore forming agent. In addition, the effects of pore forming agent content on the apparent porosity, compressive strength, pore size distribution, thermal analysis, and pressure drop of porous mullite ceramics were studied. The results show that the as-prepared porous mullite ceramics have high apparent porosity, compressive strength, and gas permeability. Mullite phase formed in situ by adding polysilicon waste and ρ-Al2O3, which bonded the fused mullite particles and improved the performance of porous mullite ceramics. With increasing soluble starch content, the apparent porosity of the samples increased from 49.50 to 62.67%, and the compressive strength decreased from 4.98 MPa to 1.35 MPa. Furthermore, uniform pore size distribution and high gas permeability fully meet the requirements for the application of high-temperature flue gas filtration.
Article
Ten established, data-driven dynamic algorithms are surveyed and a practical guide for understanding these methods generated. Existing Python programming packages for implementing each algorithm are acknowledged, and the model equations necessary for prediction are presented. A case study on a coal-fired power plant’s NOx emission rates is performed, directly comparing each modeling method’s performance on a mutual system. Each model is evaluated by its root mean squared error (RMSE) on out-of-sample future horizon predictions. Optimal hyperparameters are identified using either an exhaustive search or genetic algorithm. The top five model structures of each method are used to recursively predict future NOx emission rates over a 60-step time horizon. The RMSE at each future timestep is determined, and the recursive output prediction trends compared against measurements in time. The GRU neural network is identified as the best candidate for representing the system, demonstrating accurate and stable predictions across the future horizon by all considered models, while satisfactory performance was observed in several of the ARX/NARX formulations. These efforts have contributed 1) a concise resource of multiple proven dynamic machine learning methods, 2) a practical guide explaining the use of these methods, effectively lowering the “barrier-to-entry” of deploying such models in control systems, 3) a comparison study evaluating each method’s performance on a mutual system, 4) demonstration of accurate multi-timestep emissions modeling suitable for systems-level control, and 5) generalizable results demonstrating the suitability of each method for prediction over a multi-step future horizon to other complex dynamic systems.
Article
In the coal-fired power generation system, it is necessary to predict the NO x emissions of power station boilers when it comes to the step to spray ammonia to ensure that NO x emissions do not exceed national standards. Using traditional machine learning algorithms in the modeling of power station boilers will require features selection and steady-state extraction, which is not suitable for practical applications. In order to reduce the NO x prediction error rate under variable operating conditions, a multi-model fusion algorithm S3LX combined with linear regression, XGBoost, and long-short-term memory recurrent neural network is proposed to model the NO x emission prediction of power station boilers. The preprocessing data scheme suitable for power station boiler data sets is proposed and implemented in this paper, which can perform numerical processing, data cleaning and data standardization for boiler’s data and features. A 7-day historical operating data set of a unit in Guangzhou Shajiao C Power Plant was used as the training set and test set and was used to build the NO x emission prediction model after data preprocessing. Results show that compared with traditional machine learning algorithms, S3LX has good prediction ability under varying conditions with an average error of 4.28%. Compared with the average prediction error of the multi-layer perceptron 9.16%, SVM 7.37%, S3LX makes the error significantly reduced and satisfies the actual engineering demand.
Article
The cement calcination process suffers serious pollutant emission problems, especially nitrogen oxides (NOx) emission. Traditional methods mainly use process physical model to optimize the operation of a cement calcination process for NOx reduction. However, physical modeling of NOx emission in the rotary kiln is too complicated because of the difficulties with determining model parameters. To address this challenge, the present study proposes to apply data-driven modeling and model-based real-time cement calcination process optimization (RTO) for reducing NOx emission. The Gaussian mixture regression (GMR) model based on an improved just-in-time (IJIT) learning is introduced for modeling NOx in the kiln tail. Data preprocessing based on multivariate empirical mode decomposition (MEMD) is used and an improved similarity strategy taking into account time-space information is utilized for local learning relevant sample selection to address problems associated with nonlinear and time-varying characteristics. The RTO problem is solved by a particle swarm optimization (PSO) algorithm and optimized set values of decision variables are obtained. Finally, the proposed modeling and optimization approach is applied to practical cement calcination process data. The result shows that the proposed modeling strategy has better performance than traditional strategy and optimization approach performs better than previous conditions in calcination NOx emission reduction.
Article
The emission of NOx is of great concern to designers and operators of most industrial furnaces and boilers. The pulverized coal flame in utility scale boilers is also of great importance, affecting the levels and distribution of temperature and heat flux. Numerical studies of combustion and heat transfer processes in energy conversion systems can describe how the fuel chemical energy is converted into thermal energy with high efficiency and acceptable emission. Although there is much technology now available to compute complex flows in energy systems, development of submodels describing individual processes, as well as comprehensive CFD codes are increasing worldwide. A comprehensive 3D differential mathematical model and software were previously developed in-house and validated against experimental data. A practical motivation was to solve operation problems in tangentially-fired furnaces of the power plant Kostolac-B 350 MWe boiler units. The software is aimed for prediction of processes and operation situations in utility boiler pulverized coal-fired furnaces and it is adapted to be used by engineering staff dealing with the process analysis in boiler units. Characteristics of the model are Eulerian-Lagrangian approach to multiphase flow, k-ε turbulence model, particles-to-turbulence interactions modeled by PSI Cell method, diffusion model of particle dispersion, six-flux method for radiation modeling, heterogeneous reactions in kinetic-diffusion regime on the basis of experimentally obtained case-study coal kinetic parameters, within a "shrinking core" concept and with respect to the model of char oxidation, as well as homogeneous reactions controlled by chemical kinetics or turbulent mixing. In addition, submodel describing formation and destruction of thermal and fuel NOx has been developed and validated against available data obtained by monitoring of NOx emission from boiler units. The main motivation for this study was to achieve optimal position of flame with acceptable levels of NOx emission. The flame position depends on many influencing parameters. Selected predictions of pulverized coal flame geometry and position are given in the case-study furnace under different operating conditions, like fuel and air distribution. Even when both the fuel nitrogen content and the combustion temperature are not very high, the emission of NOx may still surpass environmental limits if the combustion process is not managed correctly. It is therefore essential to understand the NOx formation process so that the NOx emission can be controlled. Although post-combustion clean-up is viable, modifying combustion process often controls NOx most economically. In air staging method, e.g., the portion of combustion air is introduced downstream, through special, over-fire-air ports. In this work, the numerical study has been performed to achieve both NOx emission reduction and favorable position of flame in the case-study furnace, by investigating the impact of pulverized coal distribution over the burner tiers, without need for construction changes. The contributions of fuel and thermal NOx are reported as well. The results of the model can help in increasing combustion efficiency, lowering emission of pollution, fuel savings and corresponding economy and enviromental benefits during the facility exploitation.
Article
The accurate and rapid measurement of NOx concentration is critical to reduce the NOx emission of flue gas and to increase the efficiency of de-NOx system in coal-fired power plant. The measurement delay is one of the key problems for measurement of NOx concentration in continuous emission monitoring systems (CEMS). Furthermore, CEMS is unable to make NOx concentration measurement during the purging process of the sampling tube from. Although soft sensor is an attractive delay-free and economical measurement approach for monitoring NOx concentration, its measurement accuracy is low. A fusion measurement method is proposed for the NOx concentration measurement of flue gas based on Kalman filter and data fusion techniques. Numerical simulations are carried out with real operation historical data in a coal-fired power plant to evaluate the performance of the proposed method. Results demonstrated that the fusion measurement method based on Kalman filter and data fusion techniques is capable of improving response time and measurement accuracy of NOx concentration.
Chapter
Many thermal power generation plants rely on combustion of pulverised coal carried out in large furnaces. Design and improvement of these furnaces can be effectively assisted by using numerical modelling with Computational Fluid Dynamics (CFD) techniques to develop a detailed picture of the conditions within the furnace, and the effect of operating conditions, coal type, and furnace design on those conditions. The equations governing CFD models of pulverised coal combustion are described, with a focus on sub-models needed for devolatilisation, combustion and heat transfer. The use of the models is discussed with reference to examples of CFD modelling of brown coal fired furnaces in the Latrobe Valley in Australia and black coal fired furnaces described in the literature. Extensions to the CFD models that are required to tackle specific industrial and environmental issues are also described. These issues include control of NOx and SOx emissions and the effect of slagging and fouling on furnace and boiler operation.
Chapter
The characteristics of wind power result in unique challenges for system operators when integrating large penetrations of wind generation into power systems. This chapter discusses some of the power system operational challenges associated with large penetrations of wind generation, such as increased reserve requirements and the costs associated with increases in the variable operation of conventional generators. A number of power system optimization techniques with wind generation are discussed, namely the fuelsaver approach, deterministic optimization, rolling commitment and stochastic optimization. Also, a discussion of certain flexibility solutions which may reduce the system costs of increased wind penetration levels is provided.
Chapter
A dynamic model of a coal fired boiler is proposed to predict the formation of NOx during plant operation. Since evaluation of the model at hand should only take few minutes, computationally expensive CFD simulations are not feasible. Instead, it is proposed to represent the boiler as a network of ideal reactors. Gas phase reactions are modeled using a detailed kinetic mechanism; additional consideration is necessary for the heterogeneous reactions on char particles. In the reported preliminary case studies, radiative heat transfer is not considered.
Article
Reducing the denitration cost of coal-fired boilers is important to enhance the competitiveness of power generation companies. This study proposes a real operation data-based denitration cost optimization system that guides operators in economically adjusting the operation parameters of boilers. A data-driven least square support vector machine (LSSVM) learning method is utilized to predict the denitration cost of a coal-fired boiler. Back propagation (BP) is used here to select the input variables to simplify the model. With the pre-built BP-LSSVM-based denitration cost model, the genetic algorithm (GA) is then applied to obtain offline optimizations at the typical operating load points, which results in an Offline Optimal Expert Database (OOED). Once a load command is received, fuzzy association rule mining (FARM) is employed to extract the relationship between the operating load point and the optimal adjustable variables (AVs) in the OOED, thereby achieving the online denitration cost optimization of the power plant. For comparison, a single LSSVM method is also employed to build a denitration cost prediction model, and the GA and FARM proposed in this study are compared too. The results show that, compared with the single LSSVM method, the BP-LSSVM method not only predicts more accurately but also lowers the model complexity. In addition, considering the denitration cost, optimization time, and update time, the proposed BP-LSSVM-GA-FARM-based denitration cost optimization system is always better than traditional optimization methods.
Article
Owing to the complexity and inherent instability in polymer extrusion there is a need for process models which can be run on-line to optimize settings and control disturbances. First-principle models demand computationally intensive solution, while `black box' models lack generalization ability and physical process insight. This work examines a novel `grey box' modelling technique which incorporates both prior physical knowledge and empirical data in generating intuitive models of the process. The models can be related to the underlying physical mechanisms in the extruder and have been shown to capture unpredictable effects of the operating conditions on process instability. Furthermore, model parameters can be related to material properties available from laboratory analysis and as such, lend themselves to retuning for different materials without extensive remodelling work.
Article
Subject of this paper is the process variant optimal use of primary measures for nitrogen oxide minimi ration in coal-fired power plants with respect to load dependent operation, variable railing capacity and coal qualities.
Article
In this paper, a MLP (Multi-Layer Perceptron) model is developed for long time period prediction of NOx emission in a coal-fired power generation plant. In order to achieve this purpose, a novel training algorithm is used to improve the generalisation capacity of the model. The application results show the merits of this MLP model.
Article
Computational modelling of NOx formation in a 275 MWe utility boiler was investigated. The aim was to explore the use of simplified NOx chemistry when applied to a 3-D utility furnace, and ascertain if the correct exit NO trends and magnitudes could be obtained under a range of furnace operating conditions. The furnace was frontwall-fired with 24 burners in six groups of four. Furnace air-staging in the form of overtired air was available in the furnace. Combustion and modelling studies were carried out with two coals of different fuel-nitrogen content. A number of NOx control measures were investigated, including: burners out of service, overtired air and excess air. The computational framework included: Lagrangian particle tracking; turbulent particle dispersion; the discrete transfer thermal radiation model; the standard k-ε model for flow turbulence; and a simple NOx chemistry model. The NOx model contained a hydrocarbon re-burn mechanism, and the model also included the NOx precursors NH3 and HCN. A sensitivity analysis was included, to test the relative importance of selected NOx modelling parameters. The computational grid size necessary to carry out a 3-D NOx analysis of the combustor was also investigated. Furnace temperature, heat-transfer and exit NOx predictions were compared with site measurements. Detailed axial in-flame measurements of oxygen, gas temperature and NO were compared with predictions in the nearburner field of the furnace. Thermal NOx was shown to contribute greater than 20% of the total exit NO. Reasonable predictions could be obtained by means of simplified NOx chemistry combined with a relatively coarse computational grid. The predicted furnace exit NO concentrations differed by 0-30% from those measured.
Article
From measurements carried out on flat premixed hydrocarbon/oxygen argon (or helium) flames, into which small amounts of ammonia, or cyanogen are added, overall reaction rates of formation of NO and N//2 are determined. From similar measurements effected on nitrogen-diluted ethylene/oxygen flames, an overall rate of prompt NO formation is obtained. The discussion of these rate constants indicates that the relative importance of HCN molecules as intermediates in the fuel NO mechanism increases according to the following sequence of primary fuel nitrogen compounds: ammonia, cyanogen and molecular nitrogen; this last is found to behave like a true fuel nitrogen compound in the early flame stages.
Article
An understanding of the complex reaction mechanisms involved in the formation of nitrogen oxides from the combustion of fossil fuels provides a basis for the design and application of NOx control strategies. This experimental study is concerned with the formation of HCN and NH3 (as the dominant NOx precursors) from 130 kW turbulent-spray flames operating in standard and externally air-staged modes. Detailed nitrogenous species-concentration measurements from a series of nitrogen-doped gas-oil flames (using pyridine, pyrrole, quinoline, benzonitrile, benzylamine and phenylbenzylamine) supported an NO-formation route where fuel-nitrogen is initially converted to HCN, which subsequently decays to NO via NH3. Although variations were found in the developments in concentration and peak concentration levels of HCN and NH3 with each additive for an equivalent fuel-nitrogen concentration, the differences in the final NO emissions were small. Comparison of the experimental NO concentration profiles from combustion of gas oil with the same fuel containing 0.45% by mass nitrogen (by doping with pyridine) enabled the developments of thermal and fuel-NO to be followed separately for both combustion modes. For an air-staged flame operating at a primary zone (fuel/air) equivalence ratio φ, of 1.21, thermal-NO was reduced by 21% relative to an unstaged flame at the same overall stoichiometry of φ2 = φ1 = 0.85. A fuel-NO reduction of 33% was accompanied by significantly increased in-flame production of HCN.
Article
This paper concerns the mathematical modelling of fuel nitric oxide (NO) in the combustion of pulverized coal. A fuel-NO post-processor has been written and applied, along with a general mathematical model of pulverized-coal combustion, to predict NO emissions from two industrial-type burner configurations. The fate of fuel-bound nitrogen is simulated by a simplified reaction mechanism, in which the contributions from the volatile and char nitrogen contents are distinguished. Validation studies have been performed by means of data recently acquired in the Imperial College furnace. A reasonable degree of prediction ability is demonstrated; however, when attempts are made to trace the origins of predictive difficulties, the defects in the post-processor cannot be separated from those of the parent code for the flow and combustion.
Article
NOx emission dispatching is a method to reduce the amount of oxide of nitrogen produced by thermal generating units to meet given power demands. This requires adequate models relating NOx emissions to the active power generation of the unit. Least-squares estimation procedures work best when measurement errors are Gaussian. For the non-Gaussi an errors and Gaussian errors, the iteratively reweighted least-squares (IRWLS) procedure with optimized weight functions and their associated weight constants can be used to estimate, refine and fine tune parameters of identified models. The form of the emissions model is restricted to those with monotonically increasing derivatives due to subsequent minimization process requirements. We explore the application of some transformations to the form of model chosen. Moreover, emission models are developed using the IRWLS procedure. The results obtained using the iteratively reweighted least-square procedures are compared with the least squares.
Article
Computational fluid dynamics (CFD) modelling is now widely applied as an industrial plant development and process optimisation tool. The steady increase in computer power over recent years has enabled process engineers to model reacting multi-phase flows in a realistic geometry with good mesh resolution. As a result, the number of applications of CFD to industrial processes is also growing rapidly and increasing in sophistication. This paper reviews some of the recent applications of the CFX-4 code [CFX-4.3: Solver Manual, AEA Technology Engineering Software, 1999] to the power generation and combustion industries. The aim is to illustrate what can be done and also to identify trends and those areas where further work is needed. Examples include coal-fired low-NOx burner design, furnace optimisation, over-fire air, gas reburn, and laminar flames. It is argued that the trend is for CFD models to become more comprehensive and accessible by being coupled to other process models and embedded in automated information and process control systems.
Article
An orthogonal parameter estimation algorithm is derived which allows each parameter in a linear–in–the–parameters non–linear difference equation model to be estimated sequentially and quite independently of the other parameters in the model. The algorithm can be applied for any persistently exciting input and provides both unbiased estimates in the presence of correlated noise and an indication of which terms to include in the model. Several simulated examples are included to demonstrate the effectiveness of the algorithm.
Book
In this chapter, we present a detailed review of fundamental concepts, theory, definitions, and methods that are used in nonlinear regression analysis of pressure and rate transient data. Nonlinear parameter estimation coupled with statistical methods is simply referred to as nonlinear regression analysis. It has become a standard analysis procedure for interpreting pressure-transient data in the last two decades because unlike conventional methods of graphical methods used in pressure transient data analysis, it allows one to quantify the uncertainty in the final estimated formation parameters and model uniqueness in the presence of noisy (or inexact) data and uncertainty about the true, but unknown, reservoir/well model.The parameter estimation method based on the maximum likelihood estimation (MLE) is introduced and compared with the least-squares estimation (LSE), a most widely used and known estimation method in petroleum engineering. Although MLE is not widely used in petroleum engineering, it is, by far, the most commonly used method of parameter estimation in the statistics literature. Associating uncertainty into measurements by careful construction of the objective functions is best done within the concept of maximum likelihood, particularly when history matching multiple sets of pressure transient data sets, e.g., multi-well system and interval pressure transient tests. The objectives functions incorporating available prior information into parameter estimation within the framework of Bayesian methodology are given. Nonlinear parameter estimation problems for overdetermined and underdetermined problems are also covered. The chapter includes minimization methods and algorithms (while constraining the parameters within a feasible region) that can be used for minimizing various objective functions arising from MLE and LSE formulations with or without prior information. Computation of various statistics [e.g., 95%, confidence intervals for parameters, correlation coefficients for parameters, standard deviation of residuals, root-mean-square (RMS) errors] in assessing the uncertainty in estimated parameters and goodness of fit is provided.In this chapter, a number of example applications are presented using both real field and synthetic pressure transient tests to illustrate the use of nonlinear regression analysis based on the maximum likelihood and least-squares estimations.
Article
The application of comprehensive combustion modeling technology as applied to fossil-fuel combustion process is discussed. The review is divided into three main parts. First, a brief review of the state-of-the-art of the various components or submodels that are required in a comprehensive combustion model is presented. Second, the kinds of data required to evaluate the predictions of comprehensive combustion codes are considered. Third, representative applications of comprehensive combustion models are summarized, and three sets of model simulations are compared with experimental data.
Article
This paper considers the modelling of NOx formation during pulverized coal combustion with a view to developing low-NOx combustors and some aspects of the gasification of pulverized coal. The approach is based on the availability and ease of use of information for computational modelling of combustion chambers and particularly NOx formation. The approach taken is that a sufficiently low level of NOx emissions can be achieved by careful combustor design without the significant use of flue gas clean-up processes such as SCR or SNCR.
Article
In the past several methods have been elaborated for the identification of nonlinear dynamic systems. Most of the methods assume that the structure of the system is given a priori. Therefore they are in reality parameter estimation algorithms and structure identification is thus usually performed by repeated parameter estimation. However in nonlinear system theory several methods are known to determine the structure of a system. In this paper structure identification of block-oriented (especially cascade) models, of semi-linear dynamic models with signal-dependent parameters and of nonlinear dynamic models being linear in the parameters will be considered. Different structure selection methods are summarized based on step and impulse tests, frequency response measurements, correlation analysis, repeated reproducible tests and normal operating data.
Article
Following discussions on acidification in the eighties, the European Community came to a decision on NOx emission by power stations in the framework of Directive 88/609/EC on the limitation of emissions of certain pollutants from Large Combustion Installations. Within the Member States different strategies and technical measures have been applied to meet the obligations. It appears that the target for the year 1998 was already met in 1990 for the average emissions of LCI in the Community as a whole. As an example of recent developments the power station emissions in the Netherlands will be presented. The European Community is developing a Community strategy to combat acidification. The Electricity Supply Industry has serious reservations about the strategy. This will be discussed together with a case study for the Netherlands. It will be shown that RAINS is not an appropriate tool for quantitative allocation of emission reduction goals to the Member States of the European Community. A strategy is preferred that proceeds by successive stages, accompanied by monitoring of the environmental improvements. This approach enables us to integrate the progress achieved in environmental quality and emission reduction techniques into modelling works and cost/benefits assessment studies.
Article
Research in NOx formation and control has advanced significantly since Zeldovich first postulated the thermal NO formation mechanism in 1943. This paper reviews the history of NOx control implementation over this time period, with an emphasis on the role that research has played on NOx control technology, development and implementation. The discussions are divided by decades, to facilitate comprehension of the progress made. Key representative research during each decade is reviewed, covering topics from fundamental combustion studies to the effect of chemical additives on selective non-catalytic NOx reduction performance. NOx control regulations that drove (and continue to drive) the technology are also discussed. The primary focus is on Southern California and Federal regulations. Southern California was chosen because it has always had the strictest NOx emissions limits in the U.S.A., while other states and the federal government have followed at a somewhat slower pace. Federal regulations are reviewed since they represent the minimum control requirements nationwide. The application of NOx controls discussed ranges from the simple combustion modifications, implemented in the late 1950s and early 1960s, through the development of ultralow NOx burners and the retrofit of advanced selective catalytic reduction systems of today.
Article
Input-output representations of non-linear discrete-time systems are discussed. It is shown that the NARMAX (Non-linear AutoRegressive Moving Average with eXogenous inputs) model is a general and natural representation of non-linear systems and contains, as special cases, several existing non-linear models. The problem of approximating non-linear input-output systems is also addressed and several properties of non-linear models are highlighted using simple examples.
Conference Paper
First Page of the Article
Conference Paper
When considering NO<sub>x</sub> development in power stations, experiments show that its formation is slower than the combustion process. Therefore in this paper the modelling of NO<sub>x</sub> emissions is decoupled from the combustion process. A NO<sub>x</sub> model for Kilroot Power Station in Northern Ireland is then derived from the extended Zeldovich mechanism. The physical parameters in the model are determined from experiments. The comparison of the simulated values with the real measurements is also presented.
Correlation of NOx emission with basic physical and chemical characteristics of coal
  • J M Pohl
Pohl, J. M., et al. (1983). Correlation of NO x emission with basic physical and chemical characteristics of coal. Proceedings of the 1983 joint symposium on stationary NO x control, V-2 EPRI CS-3182.
Neural networks for NOx control
  • K F Reinschmidt
  • B Ling
Reinschmidt, K. F., & Ling, B. (1994). Neural networks for NO x control. Proceedings of the American power conference, Vol. 56, Part 1, Chicago (pp. 354–359).
Long-term NO x predictions of three models over test period 4
  • K Li
Long-term NO x predictions of three models over test period 4. K. Li et al. / Control Engineering Practice 12 (2004) 707–723
A case study of fundamental grey-box modelling
  • K Li
  • S Thompson
  • G R Duan
  • J Peng
Li, K., Thompson, S., Duan, G. R., & Peng, J. (2002). A case study of fundamental grey-box modelling. Preprints of the 15th IFAC world congress, Barcelona, July. Ljung, L. (1987). System identification: Theory for the user. Englewood Cliffs, NJ: Prentice-Hall.
A MLP prediction model for power generation NO x emission Proceedings of IFAC sixth international workshop on algorithms and architectures for real-time control
  • S Thompson
  • K Li
Thompson, S., & Li, K. (2000). A MLP prediction model for power generation NO x emission. In V. Hern! andez & G. W. Irwin (Eds.), Proceedings of IFAC sixth international workshop on algorithms and architectures for real-time control, AARTC'2000, Palma de Mal-lorca, Spain (pp. 108–114).
Technology status report: NO x control for pulverised coal-fired power plant Overall reaction rates of NO and N 2 formation from fuel nitrogen. 15th symposium (international) on combustion
  • Coal
Coal R&D programme (1997). Technology status report: NO x control for pulverised coal-fired power plant. Harwell: ETSU. De Soete, G. G. (1975). Overall reaction rates of NO and N 2 formation from fuel nitrogen. 15th symposium (international) on combustion (pp. 1093–1102). Pittsburgh: The Combustion Institute.
Progress report on the development of a generic NO x control intelligent system (GNOCIS) Coal R&D Program
  • K Holmes
  • L W Mayes
Holmes, K., & Mayes, L. W. (1994). Progress report on the development of a generic NO x control intelligent system (GNOCIS). Coal R&D Program, Project Profile 103, ETSU.
Implementing NOx control
  • Muzio