The pulverised coal and its mixture with biomass are one of the most popular fuels in industrial energy. To ensure, on one hand, minimal greenhouse gas emission and, on the other hand, maximum energy production efficiency, it is necessary to monitor the combustion process of these fuels. One way to do this is to monitor the flame intensity. This is an optical, non-invasive solution, and information on the status of the process is obtained with minimal delay. The article proposes a method for identifying undesired combustion states for which the excess air coefficient is greater or smaller than the value ensuring total combustion. Three deep recurrent neural network architectures for classifying the flame intensity time series were explored. The best results were obtained using the convolutional long short-term memory model, which offered the accuracy of 86.5%–99.8%, depending on the current thermal power. The prediction time of a single data sequence was about 0.6 ms. High accuracy and low time consumption of the proposed method create an opportunity for its use in industrial combustion systems of pulverised coal and its mixture with biomass.
HiiNCTanCKChongAZS, et al. The measurement of pulverised fuel flows by high frequency acoustic emission techniques. In: Proceedings of the 7th European conference on industrial furnaces and boilers, Porto, Portugal, 18–21 April 2006. Rio Tinto: CENERTEC.
2.
WardJTanC-KChongAZS, et al. The application of artificial intelligence techniques in combustion systems. In: KalogirouS (ed.) Artificial intelligence in energy and renewable energy systems. Nicosia, Cyprus: Nova Science Publishers, 2006; pp.363–394.
3.
XuLTanCLiX, et al. On-line identification of fuel type using joint probability density arbiter and support vector machine techniques. In: Proceedings of the 2010 IEEE instrumentation & measurement technology conference, Austin, TX, 3–6 May 2010. New York: IEEE.
4.
RossTJ.Fuzzy logic with engineering applications. Chichester: John Wiley & Sons, 2009.
5.
BruschiRGiacomazziEGiuliettiE, et al. An optical technique for the identification and tracking of combustion instabilities. In: Proceedings of the 7th mediterranean combustion symposium, Sardinia, 11–15 September 2011. Sardinia: Colloquia.
6.
SmolarzAKotyraAWojcikW, et al. Advanced diagnostics of industrial pulverised coal burner using optical methods and artificial intelligence. Exp Therm Fluid Sci2012; 43: 82–89.
7.
SmolarzAWojcikWGromaszekK, et al. Artificial intelligence methods in diagnostics of coal-biomass blends co-combustion in pulverised coal burners. In: PawlowskaMPawlowskiL (eds) Environmental engineering V. Boca Raton: CRC Press, Taylor & Francis Group, 2017, pp.311–317.
8.
XuLTanCLiX, et al. Fuel-type identification using joint probability density arbiter and soft-computing techniques. IEEE T Instrum Meas2012; 61(2): 286–296.
9.
WójcikW.Application of fibre-optic flame monitoring systems to diagnostics of combustion process in power boilers. Bull Pol Ac Sci-tech Sci2008; 56(2): 177–195.
10.
Ronquillo-LomeliGRomeroCEYaoZ, et al. On-line flame signal time series analysis for oil-fired burner optimization. Fuel2015; 158: 416–423.
11.
AbdurakipovSSGobyzovOATokarevMP, et al. Combustion regime monitoring by flame imaging and machine learning. Optoelectro Instrum Data Process2018; 54(5): 513–519.
12.
LyuYChenJSongZ.Image-based process monitoring using deep belief networks. IFAC PapersOnLine2018; 51(18): 115–120.
13.
WangZSongCChenT.Deep learning based monitoring of furnace combustion state and measurement of heat release rate. Energy2017; 131: 106–112.
14.
GangopadhyayTLocurtoABoorP, et al. Characterizing combustion instability using deep convolutional neural network. In: Proceedings of the ASME 2018 dynamic systems and control conference, Atlanta, GA, 30 September–3 October 2018. New York: ASME.
15.
QiuTLiuMZhouG, et al. An unsupervised classification method for flame image of pulverised coal combustion based on convolutional auto-encoder and hidden Markov model. Energies2019; 12(13): 2585–2602.
16.
LiuYFanYChenJ.Flame images for oxygen content prediction of combustion systems using DBN. Energ Fuel2017; 31: 8776–8783.
17.
TóthPGaramiACsordásB.Image-based deep neural network prediction of the heat output of a step-grate biomass boiler. Appl Energ2017; 200: 155–169.
18.
LiNLuGLiX, et al. Prediction of NOx emissions from a biomass fired combustion process based on flame radical imaging and deep learning techniques. Combust Sci Technol2016; 188(2): 233–246.
19.
AkintayoALoreKGSarkarS, et al. Prognostics of combustion instabilities from hi-speed flame video using a deep convolutional selective autoencoder. Int J Prognost Health Manag2016; 7(023): 1–14.
20.
SarkarSLoreKGSarkarS, et al. Early detection of combustion instability from hi-speed flame images via deep learning and symbolic time series analysis. In: Proceedings of the annual conference of the prognostics and Health Management Society. Vol 6, San Diego, CA, 18–22October2015, pp. 353–362. San Diego, CA: PHM.
21.
SarkarSLoreKGSarkarS.Early detection of combustion instability by neural-symbolic analysis on hi-speed video. In: Proceedings of the 2015th international conference on cognitive computation: integrating neural and Symbolic Approaches. Vol 1583, Montreal, QC, 1January2015, pp. 93–101. Aachen, DEU: CEUR-WS.Org.
22.
SmolarzAKotyraAWojcikW, et al. Advanced diagnostics of industrial pulverized coal burner using optical methods and artificial intelligence. Exper Thermal Fluid Sci2012; 43: 82–89.
23.
RomeroCLiXKeyvanS, et al. Spectrometer-based combustion monitoring for flame stoichiometry and temperature control. Appl Therm Eng2005; 25(5–6): 659–676.
24.
KeyvanSRossowRAVelezM, et al. Combustion control experimentations at a pilot scale glass furnace. In: Proceedings of the 105th annual meeting of the American Ceramic Society, Nashville, TN, 27–30 April 2003. Nashville, TN: The American Ceramic Society.
25.
MichelJBChetelatOWeberN, et al. Flame signature as a lowcost flame control method. In: Proceedings of the 5th international conference on technologies and combustion for a clean environment, Lisbon, Portugal, 12–15 July 1999. Lisbon, Portugal: Nitric Oxide Reduction by Non–hydrocarbon Fuels – Implications for Reburning with Biomass Fuels.
26.
MartinP.Actual state of technology of flame monitoring and validation. In: Proceedings of the international conference of exhibition & training program part 2, Chicago, IL, 19–24 September 1993. Instrument Society of America.
27.
ChongHSTanCKWilcoxSJ, et al. Development of an intelligent flame monitoring system for gas-fired steel reheating furnaces. In: 5th European Thermal-sciences Conference, Eindhoven, The Nederlands, 18–22May2008. Eindhoven: Eindhofen University of Technology.
28.
ArthkampJAltemarkDKruczekHJ.Burner control using flame signals. In: Proceedings of the preprints of the 1995 international gas research conferences, Cannes, 6–9 November 1995. Cannes: Government Institutes, Incorporated
29.
MiyasatoMMMcDonnellVGSamuelsenGS.Active optimization of the performance of a gas-turbine combustor. Combust Sci Technol2005; 177(9): 1725–1745.
30.
ScottDAKingGBLaurendeauNM.Chemiluminescence-based feedback control of equivalence ratio for a continuous combustor. J Propul Power2002; 18(2): 376–382.
31.
RobyRJHamerAJJohnssonEL, et al. Improved method for flame detection in combustion turbines. J Engng Gas Turb Power1995; 117: 332–340.
32.
AbdurakipovSButakovE.Combustion anomalies detection for a thermal furnace based on recurrent neural networks. J Phys Conf Series2018; 1105: 012043.
33.
MalhotraPVigLShroffG, et al. Long short term memory networks for anomaly detection in time series. In: Proceedings of the ESANN 2015, European symposium on artificial neural networks, computational intelligence and machine learning, Bruges, 22–24 April 2015. Bruges: ESANN.
34.
ChauhanSVigL.Anomaly detection in ECG time signals via deep long short-term memory networks. In: Proceedings of the 2015 IEEE international conference on data science and advanced analytics (DSAA), Paris, 19–21 October 2015. New York: IEEE.
35.
NanduriASherryL.Anomaly detection in aircraft data using recurrent neural networks (RNN). In: Proceedings of the 2016 integrated communications navigation and surveillance (ICNS), Herndon, VA, 19–21 April 2016. New York: IEEE.
36.
GuptaSRayAMukhopadhyayA.Anomaly detection in thermal pulse combustors using symbolic time series analysis. Proc IMechEng Part I-J System Control Engineering2006; 220(5): 339–351.
LuCY.A delay-range-dependent approach to global robust stability for discrete-time uncertain recurrent neural networks with interval time-varying delay. Proc IMechEng Part I-J System Control Engineering2007; 221(8): 1123–1132.
SainathTNVinyalsOSeniorA, et al. Convolutional, long short-term memory, fully connected deep neural networks. In: Proceedings of the 2015 IEEE international conference on acoustics, speech and signal processing (ICASSP), Brisbane, QLD, Australia, 6 August 2015. New York: IEEE.
46.
ShiXChenZWangH, et al. Convolutional LSTM network: a machine learning approach for precipitation nowcasting. arXiv: 1506.04214, 2015, https://arxiv.org/pdf/1506.04214v1.pdf
BaiXLuGHossainMM, et al. Multi-mode combustion process monitoring on a pulverised fuel combustion test facility based on flame imaging and random weight network techniques. Fuel2017; 202: 656–664.
50.
SujathaKPappaNSiddharth NambiU, et al. Automation of combustion monitoring in boilers using discriminant radial basis network. Int J Artif Intell Soft Comput2013; 3(3): 257–275.
