- Email: [email protected]
Evaporation system modeling of the utility boiler aiming at real-time estimation of the heat flux into water walls
13th IFAC Symposium on Large Scale Complex Systems: Theory and Applications July 7-10, 2013. Shanghai, China
Evaporation system modeling of the utility boiler aiming at real-time estimation of the heat flux into water walls Tong Yua, Jingqi Yuanb* * a
Corresponding author. Phone & Fax: +862134204055
Department of Automation, Shanghai Jiao Tong University, 800 Dongchuan Road, 200240 Shanghai, PR China. e-mail: [email protected]
b
Department of Automation, Shanghai Jiao Tong University, 800 Dongchuan Road, 200240 Shanghai, PR China. e-mail: [email protected]
Abstract A lu mped-parameter model for natural circu lation dru m-boilers is proposed to calculate the heat flu x transferred into water wall tubes. The complicated dynamics of the entire evaporation system is described, including dru m, downcomers and riser co mponents. Based on the mass and energy balances, the model is characterized by only a few physical parame ters. The heat flu x into water wall tubes can be determined without any combustion conditions of the furnace. The model validation is carried out by using the industrial data of a coal fired power p lant. The dynamic p roperties of the heat flu x may be captured over a wide range of operating, wh ich may potentially be applied for real time estimation o f heat generation in the furnace and lower heating value (LHV) of the coal. Key words: lumped-parameter model, mass and energy balances, dynamics, heat flux into water walls, power plants
1.
namic and static characteristic of the boiler, which are also able to capture the changing of the heat flu x into water wall tubes precisely. However, for many power plants in Ch ina, because of low-quality coal, the slagging of the boiler is very frequent, which causes the instability of the heat transfer coefficient of the furnace. The lack of sensors makes it difficu lt to obtain temperature d istribution on both sides of the furnace and risers. The uncertain and drastic undulation of LHV o f coal [6, 7] may have a negative influence on calculation results of those data driven models . Due to these problems, the real time and fidelity of the heat flu x into water wall tubes calculated by models may not be guaranteed, which becomes the bottleneck of the application of those models in the actual production process of power p lants in Ch ina. Other researchers, fo r examp le J. Pan et al [8], used the emp irical parameters associated with operation conditions to make fitting between the heat flu x transferred into water wall tubes and the temperature o f water wall surface. But such fitting formulas are only suitable to static characteristic of the evaporation system of boiler. In this paper, a model of evaporation system is presented, which is mainly based on mass and energy balances. The moderately co mp lex lu mped-parameter model is developed to capture the key dynamic properties over a wide operating range and achieve online estimat ion of the heat flu x into water wall tubes without the heat transfer coefficient or the temperature distribution of surfaces which can
Introduction
The heat flu x into water wall tubes of the utility boiler is a key variab le in the boiler operation monitoring, wh ich is also necessary in many advanced boiler control strategies [1, 2]. In previous investigations, one way to calculate this variable is to establish first principal models based on the equation of heat conduction of two phases, which requires experimentally measured data such as temperature distribution and heat transfer coefficient o f internal surfaces of the furnace or water wall tubes. H. E. Emara-Shabaik et al [3] required temperature distribution and heat transfer coefficient of the internal furnace surface to establish the model of heat conduction of two phases . In the investigation of J. Taler et al [4], the heat flu x of mixture of water and steam in risers measured by additional sensors installed in risers was used to calculate heat transfer coefficient of the water wall tubes and the temperature of mixture of water and steam within water wall tubes . Another way to achieve the calcu lation of the heat flu x into water wall tubes is to establish data driven models based on ANN and SVM. H. Rusinowski et al [5] established the data driven model by employing ANN with some experimentally measured data such as lower heating value (LHV) of coal, the temperature of fluegas and heat transfer coefficient of the furnace surface. All of the methods mentioned above are effective in describing both d y-
978-3-902823-39-7/2013 © IFAC
581
10.3182/20130708-3-CN-2036.00108
IFAC LSS 2013 July 7-10, 2013. Shanghai, China
hardly be obtained fro m industrial p roduction data. The model is based on physical principles and needs less parameters determined by construction data. 2.
taken away fro m the dru m and flo wing into the superheater, Ddrain is the mass flow rate of d rain ing of boiler, Vw is the total volume of saturated water in evaporation system, ρw is the density of saturated water, Vs is the total volume of saturated steam in evaporation system, ρs is the density of saturated steam. Because the total volume of evaporation system is constant, the following equation may be obtained:
Modeling
A schematic view of a boiler evaporation system is shown in Fig. 1. The unsaturated feedwater is supplied fro m economizer to dru m at mass flow rate of Dfw. Saturated water in dru m flow into downcomers at mass flo w rate of Ddc. At the end of downcomers , the saturated water flows into water wall tubes, where radiat ion energy fro m boiler co mbustion Qww is supplied and makes saturated water within tubes shift to the mixture o f saturated water and steam. The mixtu re of water and steam is forced by gravity to rise along the risers and flows back to the dru m (at mass flow rate of Drs +Drw), where saturated water and steam is separated. Saturated steam is taken away fro m drum to the superheater and the turbine at mass flow rate of Dds by dru m pressure p d . Saturated water is back to d ru m and joins the next circulat ion. Meanwhile saturated water is taken away fro m dru m at mass flow rate of Ddrain by the draining of the boiler.
Combining (1) and (2):
D fw Dds Ddrain (Vw d w Vs d s ) ( w s ) dVw dt
h
fw
Ddrain
(3)
dt
(4)
Similarly, equation based on energy balance of evaporation system may be obtained:
Drw,Drs
pd
dt
d w d s dVw D fw Dds Ddrain (Vw dt Vs dt ) dt w s
Dds Dfw
(2)
dVs dVw dt dt
Ddc
D fw
hs Dds hw Ddrain Q
ww
d (Vw w hw Vs s hs cm M mTm ) dt
(5)
Where Qww is the heat flu x transferred fro m the boiler into working med iu m in water wall tubes; cm , Mm and Tm represent the specific heat, mass and temperature of metallic part of evaporation system, respectively. h fw represents the specific enthalpy of feed water, wh ich may be calculat ed by using pressure and temperature of feedwater according to Industrial Formu lation 1997(IF 97)Equations for Region 1[9]wh ich was published by International Association for the Properties of Water and Steam in 1997. h w and h s represent specific enthalpy of saturated water and steam, respectively. Combining (2) and (5):
Qww
Fig. 1. Schematic view of the boiler evaporation system
h fw D fw hs Dds hw Ddrain Qww
Because unsaturated water fro m feedwater is heated to saturated water instantly by saturated steam in dru m as long as it flows into the dru m, the total evaporation system is considered full o f saturated steam and saturated water. Total mass balance of the evaporation system is exp ressed as the following equation:
[Vw ( w dhw hw d w ) dt dt dh d s s Vs ( s hs ) cm M m dTm )] dt dt dt dV w (hw w hs s ) dt
D fw Dds Ddrain d (Vww Vs s )
Defining specific enthalpy of subcooling of feedwater:
dt
(1)
Where Dfw is mass flow rate of feedwater fro m economizer, Dds is the mass flo w rate of saturated steam that is
hsub hw h fw
582
(6)
(7)
IFAC LSS 2013 July 7-10, 2013. Shanghai, China
and specific condensation enthalpy:
kVrw 0.3
hc hs hw
(8)
w
d w dhw hc s dt w (Vr kVrw Vdw ) dt w s
(9)
d s dhs hc w dt s (Vr kVrs Vds ) dt w s
hc s h h h ) D fw c w Dds c s Ddrain w s sub w s w s
Where Vdc is total volu me of downco mers; Vr is total volume of risers; Vdw and Vds are volume of water and steam in dru m, respectively; k Vrw and k Vrs represent the proportional coefficients of water and steam volume in risers, respectively. According to Industrial Formulat ion 1997(IF97), physical property parameters of saturated water and steam in evaporation system(density and specific enthalpy e.g.) are polynomial functions of dru m pressure p d (Equations for Region 4[9]), so all derivatives of physical property parameters of saturated water and steam with respect to time may be expressed as derivatives of those with respect to drum pressure p d then equation(10) can be obtained:
3.
Results and conclusion
The calculation result of the heat flu x into water wall tubes (Qww) and the comparison between Qww and the real-time load of power unit are shown in Fig.2. Real Load of power unit(MW)
Qww dpd Ip ( hc s hsub ) D fw w s dt h c w Dds hc s Ddrain w s w s
(10)
320 300 280 260 240 220 200 180 160 140 0
dpd
dpd
w dhw dpd
hc s d w dpd (Vr kVrw Vdw ) w s
s dhs dpd
hc w d s dpd (Vr kVrs Vds ) w s
(11)
340 320
Heat Flux Qww(MW)
dpd
10000 20000 30000 40000 50000 60000 70000 80000
Time(s) (a)
Where the inertia coefficient of drum pressure Ip is:
Ip Vdc w dhw Vdc hw d w cm M m dTm
(13)
Where Pload is the real-time load of power unit; mid load is the middle value of the setting load of power unit which is obtained from boiler operation manual. The other two parameters of inertia coefficient of dru m pressure Ip in equation (11) are volu me of saturated water and steam in dru m, wh ich are represented by Vdw and Vds , respectively. These two parameters may be calculated via geometrical structure and water level of dru m Hdrum . Because the geometrical structure of the dru m is certain, the fidelity of parameters Vdw and Vds relies on the measurement accuracy of the water level of drum Hdrum . According to equation (10), Qww, the heat flu x into water wall tubes may be calculated. The static term of the heat flu x Qww depends on status of feedwater, draining of boiler and steam taken fro m dru m, the dynamic term of Qww is completely driven by the pressure of drum p d .
dh d dT Q Vdc w w Vdc h w cm M m m dt dt dt
(
(12)
kVrs 1 kVrw
Put (4), (7), (8) into (6):
ww
Pload midload 7 midload
300 280 260 240 220 200 180 160 0
10000 20000 30000 40000 50000 60000 70000 80000
Time(s)
In order to calculate Ip by using equation (11), several parameters are needed to be identified, t wo of which are k Vrw and k Vrs . These two parameters may be obtained by the following two empirical equations, respectively [10]:
(b) Fig.2. The comparison between (a) the real load of power unit and
583
IFAC LSS 2013 July 7-10, 2013. Shanghai, China
(b) the heat flux Qww
Fig.2 shows that the heat flu x into water wall tubes (Qww) is changing fro m 180 MW at mediu m load (150MW) to 315 MW at high load (300MW), which has an appro ximate proportional relationship with real load o f the power un it . However, when the load undergoes sudden changes , the heat flu x (Qww) will have drastic fluctuations due to the fast changing of drum pressure. According to designed operation parameters fro m the boiler operation manual, the percentage of the heat flu x into water wall tubes (Qww) is 42% of total heat generation in the furnace, which means the total heat generation in the furnace is about 428 MW at med iu m load (150MW) and 750 MW at high load (300MW). In the investigation of B. Zwaan [11], the gross efficiency of subcritical units is 35% at med iu m load and 40% at high load. The same efficiency calcu lated by estimat ion of total heat generation in the furnace and corre sponding load of the power unit is also 35% (150MW/428MW) at med iu m load and 40% (300MW/750MW) at high load, which proves the fidelity of the model. The calculation period of Qww is the same as the period of raw data of the measurement collected fro m DCS servers of power plants, which is 5 seconds in this application. Such period may basically meet the requirement of advanced control strategies with period of minutes and guarantee the real time performance of them. Some researchers, for example, J.S. Chandok et al [12] have already developed the effective approaches to achieve real t ime estimation of heat of furnace exit gas. These approaches can be combined with the method of real-time estimation of the heat flu x developed in this paper to achieve real-time estimation of gross power fro m the boiler combustion. Furthermore the LHV of coal could be calculated by using estimation of gross power mentioned above and coal mass feeding rate which can be obtained from DCS of fire power plants. Acknowledgements This work was supported by the Special Funding for Research and Manufacture of Major Equip ment by Shanghai Municipal Economic and Informat ization Co mmission (Grand No. ZB–ZB YZ-01112634) and the Natural Science Foundation of China (Grand No. 61025016). References [1] K. J. Astrom, and R. D. Bell (2000). Drum-boiler dynamics. Automatica, 36, 363-378. [2] W. Langsona, I. Chryssochoosb, S.V. Rakovićb, and D. Q. M ayne (2004). Robust model predictive control using tubes. Automatica, 40, 125–133. [3] H. E. Emara-Shabaik, M . A. Habib, and I. Al-Zaharna (2009). Prediction of risers’ tubes temperature in water tube boilers. Applied Mathematical Modeling, 33, 1323-1336. [4] J. Taler, P. Duda, and B. Weglowski (2009). Identification of local heat flux to membrane water-walls in steam boilers. Fuel, 88, 305-311. [5] H. Rusinowski, and W. Stanek (2010). Hybrid model of steam boiler. Energy, 35, 1107-1113.
584
[6] J. Huang, G. Cheng, and X. Chi (2012). Control strategy for main steam pressure of combustion system of pulverized coal boiler. International Conference on Measurement, Information and Control, 2, 805-808. [7]H. Liu, H. Tan, X. Xiong, L. Yao, Y. Niu, Y. Liu, and T. Xu (2013). Prediction of calorific value of coal using real power plant data. Cleaner Combustion and Sustainable World, 3, 705-711. [8] J. Pan, D. Yang, and H. Yu (2009). M athematical modeling and thermal -hydraulic analysis of vertical water wall in an ultra-supercritical boiler. Applied Thermal Engineering, 29, 2500-2507. [9] Equations of IAPWS-IF97 A summary by Bernhard Spang, Hamburg, Germany, at The Chemical Engineers’ Resource Page URL: http://www.cheresources.com/staff.shtml [10] P. K. Chawdhry, and B.W. Hogg (1989). Identification of boiler models. Control Theory and Applications, 136, 261-271. [11]B. Zwaan (2005). Will coal depart or will it continue to dominate global power production during the 21st century . Climate Policy, 5, 445-453 [12] J.S. Chandok, I.N. Kar and Suneet Tuli(2008). Estimation of furnace exit gas temperature (FEGT) using optimized radial basis and back-propagation neural networks. Energy Conversion and M anagement, 49
Evaporation system modeling of the utility boiler aiming at real-time estimation of the heat flux into water walls Tong Yua, Jingqi Yuanb* * a
Corresponding author. Phone & Fax: +862134204055
Department of Automation, Shanghai Jiao Tong University, 800 Dongchuan Road, 200240 Shanghai, PR China. e-mail: [email protected]
b
Department of Automation, Shanghai Jiao Tong University, 800 Dongchuan Road, 200240 Shanghai, PR China. e-mail: [email protected]
Abstract A lu mped-parameter model for natural circu lation dru m-boilers is proposed to calculate the heat flu x transferred into water wall tubes. The complicated dynamics of the entire evaporation system is described, including dru m, downcomers and riser co mponents. Based on the mass and energy balances, the model is characterized by only a few physical parame ters. The heat flu x into water wall tubes can be determined without any combustion conditions of the furnace. The model validation is carried out by using the industrial data of a coal fired power p lant. The dynamic p roperties of the heat flu x may be captured over a wide range of operating, wh ich may potentially be applied for real time estimation o f heat generation in the furnace and lower heating value (LHV) of the coal. Key words: lumped-parameter model, mass and energy balances, dynamics, heat flux into water walls, power plants
1.
namic and static characteristic of the boiler, which are also able to capture the changing of the heat flu x into water wall tubes precisely. However, for many power plants in Ch ina, because of low-quality coal, the slagging of the boiler is very frequent, which causes the instability of the heat transfer coefficient of the furnace. The lack of sensors makes it difficu lt to obtain temperature d istribution on both sides of the furnace and risers. The uncertain and drastic undulation of LHV o f coal [6, 7] may have a negative influence on calculation results of those data driven models . Due to these problems, the real time and fidelity of the heat flu x into water wall tubes calculated by models may not be guaranteed, which becomes the bottleneck of the application of those models in the actual production process of power p lants in Ch ina. Other researchers, fo r examp le J. Pan et al [8], used the emp irical parameters associated with operation conditions to make fitting between the heat flu x transferred into water wall tubes and the temperature o f water wall surface. But such fitting formulas are only suitable to static characteristic of the evaporation system of boiler. In this paper, a model of evaporation system is presented, which is mainly based on mass and energy balances. The moderately co mp lex lu mped-parameter model is developed to capture the key dynamic properties over a wide operating range and achieve online estimat ion of the heat flu x into water wall tubes without the heat transfer coefficient or the temperature distribution of surfaces which can
Introduction
The heat flu x into water wall tubes of the utility boiler is a key variab le in the boiler operation monitoring, wh ich is also necessary in many advanced boiler control strategies [1, 2]. In previous investigations, one way to calculate this variable is to establish first principal models based on the equation of heat conduction of two phases, which requires experimentally measured data such as temperature distribution and heat transfer coefficient o f internal surfaces of the furnace or water wall tubes. H. E. Emara-Shabaik et al [3] required temperature distribution and heat transfer coefficient of the internal furnace surface to establish the model of heat conduction of two phases . In the investigation of J. Taler et al [4], the heat flu x of mixture of water and steam in risers measured by additional sensors installed in risers was used to calculate heat transfer coefficient of the water wall tubes and the temperature of mixture of water and steam within water wall tubes . Another way to achieve the calcu lation of the heat flu x into water wall tubes is to establish data driven models based on ANN and SVM. H. Rusinowski et al [5] established the data driven model by employing ANN with some experimentally measured data such as lower heating value (LHV) of coal, the temperature of fluegas and heat transfer coefficient of the furnace surface. All of the methods mentioned above are effective in describing both d y-
978-3-902823-39-7/2013 © IFAC
581
10.3182/20130708-3-CN-2036.00108
IFAC LSS 2013 July 7-10, 2013. Shanghai, China
hardly be obtained fro m industrial p roduction data. The model is based on physical principles and needs less parameters determined by construction data. 2.
taken away fro m the dru m and flo wing into the superheater, Ddrain is the mass flow rate of d rain ing of boiler, Vw is the total volume of saturated water in evaporation system, ρw is the density of saturated water, Vs is the total volume of saturated steam in evaporation system, ρs is the density of saturated steam. Because the total volume of evaporation system is constant, the following equation may be obtained:
Modeling
A schematic view of a boiler evaporation system is shown in Fig. 1. The unsaturated feedwater is supplied fro m economizer to dru m at mass flow rate of Dfw. Saturated water in dru m flow into downcomers at mass flo w rate of Ddc. At the end of downcomers , the saturated water flows into water wall tubes, where radiat ion energy fro m boiler co mbustion Qww is supplied and makes saturated water within tubes shift to the mixture o f saturated water and steam. The mixtu re of water and steam is forced by gravity to rise along the risers and flows back to the dru m (at mass flow rate of Drs +Drw), where saturated water and steam is separated. Saturated steam is taken away fro m drum to the superheater and the turbine at mass flow rate of Dds by dru m pressure p d . Saturated water is back to d ru m and joins the next circulat ion. Meanwhile saturated water is taken away fro m dru m at mass flow rate of Ddrain by the draining of the boiler.
Combining (1) and (2):
D fw Dds Ddrain (Vw d w Vs d s ) ( w s ) dVw dt
h
fw
Ddrain
(3)
dt
(4)
Similarly, equation based on energy balance of evaporation system may be obtained:
Drw,Drs
pd
dt
d w d s dVw D fw Dds Ddrain (Vw dt Vs dt ) dt w s
Dds Dfw
(2)
dVs dVw dt dt
Ddc
D fw
hs Dds hw Ddrain Q
ww
d (Vw w hw Vs s hs cm M mTm ) dt
(5)
Where Qww is the heat flu x transferred fro m the boiler into working med iu m in water wall tubes; cm , Mm and Tm represent the specific heat, mass and temperature of metallic part of evaporation system, respectively. h fw represents the specific enthalpy of feed water, wh ich may be calculat ed by using pressure and temperature of feedwater according to Industrial Formu lation 1997(IF 97)Equations for Region 1[9]wh ich was published by International Association for the Properties of Water and Steam in 1997. h w and h s represent specific enthalpy of saturated water and steam, respectively. Combining (2) and (5):
Qww
Fig. 1. Schematic view of the boiler evaporation system
h fw D fw hs Dds hw Ddrain Qww
Because unsaturated water fro m feedwater is heated to saturated water instantly by saturated steam in dru m as long as it flows into the dru m, the total evaporation system is considered full o f saturated steam and saturated water. Total mass balance of the evaporation system is exp ressed as the following equation:
[Vw ( w dhw hw d w ) dt dt dh d s s Vs ( s hs ) cm M m dTm )] dt dt dt dV w (hw w hs s ) dt
D fw Dds Ddrain d (Vww Vs s )
Defining specific enthalpy of subcooling of feedwater:
dt
(1)
Where Dfw is mass flow rate of feedwater fro m economizer, Dds is the mass flo w rate of saturated steam that is
hsub hw h fw
582
(6)
(7)
IFAC LSS 2013 July 7-10, 2013. Shanghai, China
and specific condensation enthalpy:
kVrw 0.3
hc hs hw
(8)
w
d w dhw hc s dt w (Vr kVrw Vdw ) dt w s
(9)
d s dhs hc w dt s (Vr kVrs Vds ) dt w s
hc s h h h ) D fw c w Dds c s Ddrain w s sub w s w s
Where Vdc is total volu me of downco mers; Vr is total volume of risers; Vdw and Vds are volume of water and steam in dru m, respectively; k Vrw and k Vrs represent the proportional coefficients of water and steam volume in risers, respectively. According to Industrial Formulat ion 1997(IF97), physical property parameters of saturated water and steam in evaporation system(density and specific enthalpy e.g.) are polynomial functions of dru m pressure p d (Equations for Region 4[9]), so all derivatives of physical property parameters of saturated water and steam with respect to time may be expressed as derivatives of those with respect to drum pressure p d then equation(10) can be obtained:
3.
Results and conclusion
The calculation result of the heat flu x into water wall tubes (Qww) and the comparison between Qww and the real-time load of power unit are shown in Fig.2. Real Load of power unit(MW)
Qww dpd Ip ( hc s hsub ) D fw w s dt h c w Dds hc s Ddrain w s w s
(10)
320 300 280 260 240 220 200 180 160 140 0
dpd
dpd
w dhw dpd
hc s d w dpd (Vr kVrw Vdw ) w s
s dhs dpd
hc w d s dpd (Vr kVrs Vds ) w s
(11)
340 320
Heat Flux Qww(MW)
dpd
10000 20000 30000 40000 50000 60000 70000 80000
Time(s) (a)
Where the inertia coefficient of drum pressure Ip is:
Ip Vdc w dhw Vdc hw d w cm M m dTm
(13)
Where Pload is the real-time load of power unit; mid load is the middle value of the setting load of power unit which is obtained from boiler operation manual. The other two parameters of inertia coefficient of dru m pressure Ip in equation (11) are volu me of saturated water and steam in dru m, wh ich are represented by Vdw and Vds , respectively. These two parameters may be calculated via geometrical structure and water level of dru m Hdrum . Because the geometrical structure of the dru m is certain, the fidelity of parameters Vdw and Vds relies on the measurement accuracy of the water level of drum Hdrum . According to equation (10), Qww, the heat flu x into water wall tubes may be calculated. The static term of the heat flu x Qww depends on status of feedwater, draining of boiler and steam taken fro m dru m, the dynamic term of Qww is completely driven by the pressure of drum p d .
dh d dT Q Vdc w w Vdc h w cm M m m dt dt dt
(
(12)
kVrs 1 kVrw
Put (4), (7), (8) into (6):
ww
Pload midload 7 midload
300 280 260 240 220 200 180 160 0
10000 20000 30000 40000 50000 60000 70000 80000
Time(s)
In order to calculate Ip by using equation (11), several parameters are needed to be identified, t wo of which are k Vrw and k Vrs . These two parameters may be obtained by the following two empirical equations, respectively [10]:
(b) Fig.2. The comparison between (a) the real load of power unit and
583
IFAC LSS 2013 July 7-10, 2013. Shanghai, China
(b) the heat flux Qww
Fig.2 shows that the heat flu x into water wall tubes (Qww) is changing fro m 180 MW at mediu m load (150MW) to 315 MW at high load (300MW), which has an appro ximate proportional relationship with real load o f the power un it . However, when the load undergoes sudden changes , the heat flu x (Qww) will have drastic fluctuations due to the fast changing of drum pressure. According to designed operation parameters fro m the boiler operation manual, the percentage of the heat flu x into water wall tubes (Qww) is 42% of total heat generation in the furnace, which means the total heat generation in the furnace is about 428 MW at med iu m load (150MW) and 750 MW at high load (300MW). In the investigation of B. Zwaan [11], the gross efficiency of subcritical units is 35% at med iu m load and 40% at high load. The same efficiency calcu lated by estimat ion of total heat generation in the furnace and corre sponding load of the power unit is also 35% (150MW/428MW) at med iu m load and 40% (300MW/750MW) at high load, which proves the fidelity of the model. The calculation period of Qww is the same as the period of raw data of the measurement collected fro m DCS servers of power plants, which is 5 seconds in this application. Such period may basically meet the requirement of advanced control strategies with period of minutes and guarantee the real time performance of them. Some researchers, for example, J.S. Chandok et al [12] have already developed the effective approaches to achieve real t ime estimation of heat of furnace exit gas. These approaches can be combined with the method of real-time estimation of the heat flu x developed in this paper to achieve real-time estimation of gross power fro m the boiler combustion. Furthermore the LHV of coal could be calculated by using estimation of gross power mentioned above and coal mass feeding rate which can be obtained from DCS of fire power plants. Acknowledgements This work was supported by the Special Funding for Research and Manufacture of Major Equip ment by Shanghai Municipal Economic and Informat ization Co mmission (Grand No. ZB–ZB YZ-01112634) and the Natural Science Foundation of China (Grand No. 61025016). References [1] K. J. Astrom, and R. D. Bell (2000). Drum-boiler dynamics. Automatica, 36, 363-378. [2] W. Langsona, I. Chryssochoosb, S.V. Rakovićb, and D. Q. M ayne (2004). Robust model predictive control using tubes. Automatica, 40, 125–133. [3] H. E. Emara-Shabaik, M . A. Habib, and I. Al-Zaharna (2009). Prediction of risers’ tubes temperature in water tube boilers. Applied Mathematical Modeling, 33, 1323-1336. [4] J. Taler, P. Duda, and B. Weglowski (2009). Identification of local heat flux to membrane water-walls in steam boilers. Fuel, 88, 305-311. [5] H. Rusinowski, and W. Stanek (2010). Hybrid model of steam boiler. Energy, 35, 1107-1113.
584
[6] J. Huang, G. Cheng, and X. Chi (2012). Control strategy for main steam pressure of combustion system of pulverized coal boiler. International Conference on Measurement, Information and Control, 2, 805-808. [7]H. Liu, H. Tan, X. Xiong, L. Yao, Y. Niu, Y. Liu, and T. Xu (2013). Prediction of calorific value of coal using real power plant data. Cleaner Combustion and Sustainable World, 3, 705-711. [8] J. Pan, D. Yang, and H. Yu (2009). M athematical modeling and thermal -hydraulic analysis of vertical water wall in an ultra-supercritical boiler. Applied Thermal Engineering, 29, 2500-2507. [9] Equations of IAPWS-IF97 A summary by Bernhard Spang, Hamburg, Germany, at The Chemical Engineers’ Resource Page URL: http://www.cheresources.com/staff.shtml [10] P. K. Chawdhry, and B.W. Hogg (1989). Identification of boiler models. Control Theory and Applications, 136, 261-271. [11]B. Zwaan (2005). Will coal depart or will it continue to dominate global power production during the 21st century . Climate Policy, 5, 445-453 [12] J.S. Chandok, I.N. Kar and Suneet Tuli(2008). Estimation of furnace exit gas temperature (FEGT) using optimized radial basis and back-propagation neural networks. Energy Conversion and M anagement, 49