A Nitrogen Oxides Emission Prediction Model for Gas Turbines Based on Interpretable Multilayer Perceptron Neural Networks

2020 ◽  
Author(s):  
Dawen Huang ◽  
Shanhua Tang ◽  
Dengji Zhou
Keyword(s):  
Neural Networks ◽  
Nitrogen Oxides ◽  
Prediction Model ◽  
Working Conditions ◽  
Gas Turbines ◽  
Middle Layer ◽  
Operating Conditions ◽  
Nox Emission ◽  
Outlet Temperature ◽  
Output Layer

Abstract Gas turbines, an important energy conversion equipment, produce Nitrogen Oxides (NOx) emissions, endangering human health and forming air pollution. With the increasingly stringent NOx emission standards, it is more significant to ascertain NOx emission characteristics to reduce pollutant emissions. Establishing an emission prediction model is an effective way for real-time and accurate monitoring of the NOx discharge amount. Based on the multi-layer perceptron neural networks, an interpretable emission prediction model with a monitorable middle layer is designed to monitor NOx emission by taking the ambient parameters and boundary parameters as the network inputs. The outlet temperature of the compressor is selected as the monitorable measuring parameters of the middle layer. The emission prediction model is trained by historical operation data under different working conditions. According to the errors between the predicted values and measured values of the middle layer and output layer, the weights of the emission prediction model are optimized by the back-propagation algorithm, and the optimal NOx emission prediction model is established for gas turbines under the various working conditions. Furthermore, the mechanism of predicting NOx emission value is explained based on known parameter influence laws between the input layer, middle layer and output layer, which helps to reveal the main measurement parameters affecting NOx emission value, adjust the model parameters and obtain more accurate prediction results. Compared with the traditional emission monitoring methods, the emission prediction model has higher accuracy and faster calculation efficiency and can obtain believable NOx emission prediction results for various operating conditions of gas turbines.

scholarly journals A Comparative Study on Fault Detection Methods for Gas Turbine Combustion Systems

Energies ◽  
2021 ◽  
Vol 14 (2) ◽  
pp. 389
Author(s):  
Jinfu Liu ◽  
Zhenhua Long ◽  
Mingliang Bai ◽  
Linhai Zhu ◽  
Daren Yu
Keyword(s):  
Fault Detection ◽  
Gas Turbine ◽  
Gas Turbines ◽  
Operating Conditions ◽  
Detection Methods ◽  
Distribution Model ◽  
Outlet Temperature ◽  
Combustion System ◽  
Comparison Results ◽  
Gas Turbine Combustion

As one of the core components of gas turbines, the combustion system operates in a high-temperature and high-pressure adverse environment, which makes it extremely prone to faults and catastrophic accidents. Therefore, it is necessary to monitor the combustion system to detect in a timely way whether its performance has deteriorated, to improve the safety and economy of gas turbine operation. However, the combustor outlet temperature is so high that conventional sensors cannot work in such a harsh environment for a long time. In practical application, temperature thermocouples distributed at the turbine outlet are used to monitor the exhaust gas temperature (EGT) to indirectly monitor the performance of the combustion system, but, the EGT is not only affected by faults but also influenced by many interference factors, such as ambient conditions, operating conditions, rotation and mixing of uneven hot gas, performance degradation of compressor, etc., which will reduce the sensitivity and reliability of fault detection. For this reason, many scholars have devoted themselves to the research of combustion system fault detection and proposed many excellent methods. However, few studies have compared these methods. This paper will introduce the main methods of combustion system fault detection and select current mainstream methods for analysis. And a circumferential temperature distribution model of gas turbine is established to simulate the EGT profile when a fault is coupled with interference factors, then use the simulation data to compare the detection results of selected methods. Besides, the comparison results are verified by the actual operation data of a gas turbine. Finally, through comparative research and mechanism analysis, the study points out a more suitable method for gas turbine combustion system fault detection and proposes possible development directions.

An Estimation of 3-Way Catalyst Performance Using Artificial Neural Networks During Idle Speed

2004 ◽  
Author(s):  
P. N. Botsaris ◽  
D. Bechrakis ◽  
P. D. Sparis
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Steady State ◽  
Expert Knowledge ◽  
Operating Conditions ◽  
Experimental Results ◽  
Outlet Temperature ◽  
Practical Applications ◽  
Idle Speed ◽  
Artificial Neural

The intelligent control as fuzzy or artificial is based on either expert knowledge or experimental data and therefore it possesses intrinsic qualities like robustness and ease implementation. Lately, many researchers present studies aim to show that this kind of control can be used in practical applications such as the idle speed control problem in automotive industry. In this study, an estimation of an automobile three-way catalyst performance with artificial neural networks is presented. It may be an alternative approach for an on board diagnostic system (OBD) to predict the catalyst performance. This method was tested using data sets from two kind of catalysts, a brand new and an old one on a laboratory bench at idle speed. The catalyst operation during the “steady state” phase (the phase that the catalyst has reached its operating conditions and works normally) is examined. Further experiments are needed for different catalyst typed before the methods is proposed generally. It consists of 855 elements of catalyst inlet-outlet temperature difference (DT), hydrocarbons (HC), and carbon monoxide (CO) and carbon dioxide (CO2) emissions. The simulation: detects the values of HC, CO, CO2 using the DT as an input to our network forms a neural network. Results showed serious indications that artificial neural networks (or fuzzy logic control laws) could estimate the catalyst performance adequately depending their training process, if certain information about the catalyst system and the inputs and output of such system are known. In this study the “steady state” period experimental results are presented. In this paper the “steady state” period experimental results are presented.

ANN - SQP Approach For NOx Emission Reduction In Coal Fired Boilers

2012 ◽  
Vol 13 (3) ◽  
Author(s):  
Ilamathi Balamurugan ◽  
Selladurai V. Gounder ◽  
Balamurugan Kulendran
Keyword(s):  
Flue Gas ◽  
Field Experiments ◽  
Fitness Function ◽  
Load Condition ◽  
Operating Conditions ◽  
Nox Emission ◽  
Optimum Level ◽  
Outlet Temperature ◽  
Boiler Load ◽  
Feed Forward Back Propagation

Abstract In this research paper, predictive modelling of NOx emission of a 210 MW capacity pulverized coal-fired boiler and combustion parameter optimization to reduce NOx emission in flue gas is proposed. The effects of oxygen concentration in flue gas, coal properties, coal flow, boiler load, air distribution scheme, flue gas outlet temperature and nozzle tilt were studied. The data collected from parametric field experiments were used to build a feed-forward back-propagation artificial neural net (ANN). The coal combustion parameters were used as inputs and NOx emission as outputs of the model. The ANN model was developed for full load condition and its predicted values were verified with the actual values. The algebraic equation containing weights and biases of the trained net was used as fitness function in sequential quadratic programming (SQP) to find the optimum level of input operating conditions for low NOx emission. The results proved that the proposed approach could be used for generating feasible operating conditions.

Model Order Reduction and the Heat Transfer Modelling of a Heat Exchanger by Using Artificial Neural Approach

2011 ◽  
Vol 6 (1) ◽  
Author(s):  
Yahya Chetouani
Keyword(s):  
Neural Network ◽  
Experimental Data ◽  
Neural Networks ◽  
Heat Exchanger ◽  
Black Box ◽  
Operating Conditions ◽  
Outlet Temperature ◽  
Input Output ◽  
Reliable Model ◽  
Artificial Neural

The main aim of this paper is to establish a reliable model of a process behavior under the normal operating conditions. The use of this model should reflect the true behavior of the process in the whole way and thus distinguish a normal mode from the abnormal modes. In order to obtain a reliable model for the process dynamics, the black-box identification by means of a NARMAX model has been chosen in this paper. It is based on the neural networks approach. The main advantage of the proposed approach consists in the natural ability of neural networks in modeling non-linear dynamics in a fast and simple way and in the possibility to address the process to be modeled as an input-output black-box, with little or no mathematical information on the system. This paper will show the choice and the performance of the neural network in the training and the test phases. A study is related to the number of inputs, and of hidden neurons used and their influence on the behavior of the neural predictor. Three statistical criterions, Aikeke’s information criterion (AIC), Rissanen’s Minimum Description Length (MDL), and Bayesian information criteria (BIC), are used for the validation of the experimental data. In order to illustrate the ideas proposed concerning the dynamics modelling, a heat exchanger is used. The outlet temperature is modeled according to the inlet temperature. The model is implemented by training a Multilayer Perceptron artificial neural network with input-output experimental data. Satisfactory agreement between identified and experimental data is found and results show that the model successfully predicts the evolution of the outlet temperature of the process.

Experimental and Numerical Study on Emission Characteristics of the Double Annular Swirler Under Different Pilot Fuel Ratios

2018 ◽  
Author(s):  
Chao Zong ◽  
Yaya Lyu ◽  
Desan Guo ◽  
Chengqin Li ◽  
Tong Zhu
Keyword(s):  
Nitrogen Oxides ◽  
Combustion Chamber ◽  
Gas Turbine ◽  
Gas Turbines ◽  
Numerical Study ◽  
Nox Emission ◽  
Pollutant Emission ◽  
Emission Characteristics ◽  
Micro Gas Turbine ◽  
Pilot Fuel

Micro gas turbine is one of the ideal prime movers for small-distributed energy systems. It can effectively reduce the emission of greenhouse gases and nitrogen oxides. Moreover, the use of micro gas turbines will contribute to burning fossil fuels in a much cleaner way. The staged combustion technology is the favorite way for low pollution combustion chamber such like. Therefore, the influence of the proportion of pilot fuel in the combustion chamber on pollutant emission deserves further study. The object of this research is the Double annular swirler (Das), which was applied to a 100 kW micro gas turbine combustion chamber. The combustion performance and emission characteristics under different Pilot Fuel Ratios (PFR) were obtained in prototype experimental system. Under the experimental conditions, Computational fluid dynamics (CFD) method was applied to research the reacting flow field and the formation of NOx in the combustion chamber and then analyze the influences of PFRs on combustion process. Experimental results show that the NOx emission of Das decreased at first and then increased with the augment of PFR. When PFR was near to 11%, the per unit NOx emission concentration reached its minimum. The numerical simulation agreed well with the experimental data. Further analysis of the simulation results indicate that there is a strong correlation between Φlocal and NOx concentration. When it is lower than a certain value, the number of nitrogen oxides will be significantly reduced. The value has a lot to do with the inlet air temperature and the pressure of the combustion chamber under the design condition, and it needs to be confirmed by calculating the adiabatic temperature. Simultaneously, we also find that although the percentage of total air flowing into the combustor remains unchanged, the increase of PFR would reduce the airflow ratio in inner swirler. This implies that for some particular combustion chambers, special attention should be paid to the changes in air allocation caused by PFR.

Adaptive NOx Control on Refinery Gas in Industrial Gas Turbines

2012 ◽  
Author(s):  
Philipp Geipel ◽  
Erik Bjerklund ◽  
Magnus Persson ◽  
Arturo Manrique Carrera
Keyword(s):  
Gas Turbines ◽  
Response Times ◽  
Water Injection ◽  
Operating Conditions ◽  
Nox Emission ◽  
Current Case ◽  
Parametric Curve ◽  
Emission Limits ◽  
Nox Control ◽  
Refinery Gas

The current work explores the operation of an industrial gas turbine (25 MWe) unit working under variable NOx emission limits (from 75 to 120 mg/Nm3) and variable fuel composition. The latter is typical for flared streams in oil process industry (refineries/well operation/petrochemicals). In the current case the nitrous oxides (NOx) emission limits are dependent on the mixing ratio of natural gas (NG, 89%vol. methane) and refinery gas (RG, 40% vol. H2, 56%vol. CH4, C4+) that is used as fuel for the engine. This mixture is process dependent and varies in matter of minutes from 100% NG to 100% RG. Siemens has faced this challenge by using a SGT-600 unit (25MWe) operating in non-premixed combustion mode with a novel water injection strategy, the adaptive water injection for the control of NOx. The adaptive water injection for NOx control has been developed after evaluating different NOx control strategies. It consists of a closed loop, in which NOx emission measurements are used as the main control parameter, linked to a fixed parametric curve based on the compressor discharge pressure. Fast response times and fail–safe strategies have also been developed and tested under operating conditions of the GT. The water injection method could cope with the variation of load, fuel supply mixture and NOx limit without any flame instability problems. This is done while maintaining the optimal amount of water needed to achieve the emissions target. The robustness of the system has been satisfactory over the tested period and seems a viable solution.

Model Based Diagnostics of AE-T100 Micro Gas Turbine

2016 ◽  
Author(s):  
Mariam Mahmood ◽  
Alessio Martini ◽  
Alberto Traverso ◽  
Enrico Bianchi
Keyword(s):  
Gas Turbines ◽  
Clean Energy ◽  
Operating Conditions ◽  
Outlet Temperature ◽  
Ambient Temperatures ◽  
Micro Gas Turbine ◽  
Fuel Flexibility ◽  
Stable Operating ◽  
The Stability ◽  
Steady State Simulation

The growing environmental impacts and dwindling supply of conventional fuels have led to the development of more efficient and clean energy systems. Micro gas turbines (mGT) have emerged as energy conversion technology, which offer promising features like high fuel flexibility, low emissions level, and efficient cogeneration of heat and power (CHP). Numerical simulation is a vital tool to predict the off-design performance of mGT cycles, and it also helps in cycle optimization. Starting from a model available at Ansaldo Energia, for steady state simulation of mGT T100 cycles based on user requirements, within the cooperation between University of Genova (Unige) and Ansaldo Energia, a new more comprehensive simulation tool has been developed through the incorporation of additional components, features, and involving a more detailed mathematical approach. The most important upgrades involved a number of different air path flows and the power electronics, which takes into account the power consumption from auxiliary components as well as the generator and inverter efficiencies. Once the model has been verified against the existing tools, it was used in real operating conditions at the Ansaldo Energia test rig. The mGT performance has been assessed for different power levels, starting from 100 kW (nominal power) to 60 kW and then back to 100 kW, with 10 kW steps. The two tests at 100 kW operating conditions have been carried out with two different ambient temperatures: 20°C and 25°C, respectively. Data have been acquired under stable operating conditions, considering the recuperator cold outlet temperature as the stability indicator. Finally the new model AE-T100 has been used also for diagnosis of the whole mGT cycle. The model has been successfully applied to a special mGT equipped with an on-purpose damaged recuperator, identifying the causes of performance degradation.

A Matlab Neural Network Application for the Study of Working Conditions

2013 ◽  
Vol 837 ◽  
pp. 310-315 ◽  
Author(s):  
Daniel Constantin Anghel ◽  
Alexandru Ene ◽  
Nadia Belu
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Working Conditions ◽  
Experimental Results ◽  
Output Layer ◽  
Linear Relationships ◽  
The Neural Network ◽  
Manufacture Industry ◽  
The Neural Networks ◽  
Network Application

The paper presents a method based on the neural networks to study of working conditions, for the workstations from the manufacture industry. The neural networks were chosen because they excel in gathering difficult non-linear relationships between the inputs and outputs of a system. The neural network was simulated with Matlab. In this paper, we considered as relevant for the study of working conditions, 6 input parameters: temperature, humidity, noise, luminosity, load and frequency. The neural network designed for the study presented in this paper has 6 input neurons and 3 neurons in the output layer. Some experimental results obtained through simulations, are presented in the final part of the paper.

An Efficient Prediction Method for the Azimuthal Migration of Combustion Inhomogeneity in Multi-Stage Cooled Turbines

2021 ◽  
Author(s):  
Qingfu He ◽  
Zhongran Chi ◽  
Shusheng Zang
Keyword(s):  
Steady State ◽  
Working Conditions ◽  
Gas Turbines ◽  
Three Dimensional ◽  
Prediction Method ◽  
Outlet Temperature ◽  
Multi Stage ◽  
Unsteady Simulation ◽  
Efficient Prediction ◽  
Turbine Exhaust

Abstract The outlet temperature of combustor is commonly monitored by thermocouples at the turbine exhaust. In order to establish the corresponding relationship between the temperature measured by each thermocouple and the working state of each burner, the azimuthal migration of the combustion hot/cold streaks in the multi-stage turbines needs to be quantified. Experiments to measure this migration have high cost and considerable error. It is also difficult to quantify the migration under multiple working conditions. Three-dimensional full-annulus unsteady simulation can obtain this migration. But the unsteady simulation of a single working condition could take several weeks, which is too expensive for engineering usage. A method named Steady-state Computation of Azimuthal Migration (SCAM) is proposed in this paper. By establishing and solving the transport equation of the migration angle, the azimuthal migration of hot/cold streaks can be predicted by steady-state numerical simulation using the mixing plane at rotor-stator interface. The migration computed by this method is compared with the full-annulus unsteady simulation results in multiple working conditions. The results of SCAM method show good agreement with full-annulus simulations, while costing only 0.01% of the CPU hours. It is also found that the error of SCAM is mainly caused by the fixed boundary value at coolant source terms. The optimal spanwise location of the thermocouples at turbine exhaust is discussed based on the results. The method proposed could be applied to the fault diagnosis and precise repair of the combustors of gas turbines.

Sign in / Sign up

Export Citation Format

Share Document