Comparative Evaluation of Integrated Waste Heat Utilization Systems for Coal-Fired Power Plants Based on In-Depth Boiler-Turbine Integration and Organic Rankine Cycle

To maximize the system-level heat integration, three retrofit concepts of waste heat recovery via organic Rankine cycle (ORC), in-depth boiler-turbine integration, and coupling of both are proposed, analyzed and comprehensively compared in terms of thermodynamic and economic performances. For thermodynamic analysis, exergy analysis is employed with grand composite curves illustrated to identify how the systems are fundamentally and quantitatively improved, and to highlight key processes for system improvement. For economic analysis, annual revenue and investment payback period are calculated based on the estimation of capital investment of each component to identify the economic feasibility and competitiveness of each retrofit concept proposed. The results show that the in-depth boiler-turbine integration achieves a better temperature match of heat flows involved for different fluids and multi-stage air preheating, thus a significant improvement of power output (23.99 MW), which is much larger than that of the system with only ORC (6.49 MW). This is mainly due to the limitation of the ultra-low temperature (from 135 to 75 °C) heat available from the flue gas for ORC. The thermodynamic improvement is mostly contributed by the reduction of exergy destruction within the boiler subsystem, which is eventually converted to mechanical power; while the exergy destruction within the turbine system is almost not changed for the three concepts. The selection of ORC working fluids is performed to maximize the power output. Due to the low-grade heat source, the cycle with R11 offers the largest additional net power generation but is not significantly better than the other preselected working fluids. Economically, the in-depth boiler-turbine integration is the most economic completive solution with a payback period of only 0.78 year. The ORC concept is less attractive for a sole application due to a long payback time (2.26 years). However, by coupling both concepts, a net power output of 26.51 MW and a payback time of almost one year are achieved, which may promote the large-scale production and deployment of ORC with a cost reduction and competitiveness enhancement.

selected power plant is 994 MW with the pressure/temperature of main steam of 26.01 MPa/600 • C and the temperature of reheated steam of 600 • C. The power plant is designed for a bituminous coal with the element analysis of 57.37% carbon, 4.19% hydrogen, 7.57% oxygen, 1.4% nitrogen, 0.87% sulfur and 7.3% moisture. The stream data related to all regenerative heats (RHs) have been listed in Table 1 for the turbine heat rate acceptance (THA) condition, under which the inlet conditions and back pressure of the turbine are nominal with the regenerative feedwater preheaters operating normally. More detailed descriptions of 1000 MW-level coal-fired power plants can be found elsewhere in our publications, e.g., [58][59][60][61][62][63], and will not be repeated here.

The Case with In-Depth Boiler-Turbine Integration
For current regenerative systems of coal-fired power plants, a large amount of steam at various pressure levels are extracted to heat the feedwater and condensate water, which improves the average temperature of heat absorption of the Rankine cycle, thus enhancing the thermodynamic efficiency. However, the working ability of the extracted steam will not be fully realized. The proposed integrated system (modified from [27]) aims at utilizing waste heat from the boiler cold-end to partially heat the feedwater so that some of the steam extraction can be avoided to boost the power generation and system efficiency without introducing additional heat sources. Figure 1 illustrates the configuration of the integrated WHUS system with in-depth boiler-turbine integration. A bypass flue gas duct parallel to the main air-preheater duct is introduced with two-stage successively-configured gas-water heat exchangers. Approximately one-third of the total flue gas out of the economizer enters the newly introduced bypass duct and sequentially exchanges heat in the high-temperature and low-temperature gas-water heat exchangers (HGHE and LGHE). Since the heat of the flue gas entering the main duct decreases and cannot meet the heat requirement for air preheating, two additional heat exchangers (FSHE and SSHE) are introduced for air preheating at low temperatures with the low-grade heat from the low-pressure steam extraction and the flue gas. The specifications of the main heat exchanges involved are presented in Table 2. °C and the temperature of reheated steam of 600 °C. The power plant is designed for a bituminous coal with the element analysis of 57.37% carbon, 4.19% hydrogen, 7.57% oxygen, 1.4% nitrogen, 0.87% sulfur and 7.3% moisture. The stream data related to all regenerative heats (RHs) have been listed in Table 1 for the turbine heat rate acceptance (THA) condition, under which the inlet conditions and back pressure of the turbine are nominal with the regenerative feedwater preheaters operating normally. More detailed descriptions of 1000 MW-level coal-fired power plants can be found elsewhere in our publications, e.g., [58][59][60][61][62][63], and will not be repeated here.

The Case with In-Depth Boiler-Turbine Integration
For current regenerative systems of coal-fired power plants, a large amount of steam at various pressure levels are extracted to heat the feedwater and condensate water, which improves the average temperature of heat absorption of the Rankine cycle, thus enhancing the thermodynamic efficiency. However, the working ability of the extracted steam will not be fully realized. The proposed integrated system (modified from [27]) aims at utilizing waste heat from the boiler cold-end to partially heat the feedwater so that some of the steam extraction can be avoided to boost the power generation and system efficiency without introducing additional heat sources. Figure 1 illustrates the configuration of the integrated WHUS system with in-depth boilerturbine integration. A bypass flue gas duct parallel to the main air-preheater duct is introduced with two-stage successively-configured gas-water heat exchangers. Approximately one-third of the total flue gas out of the economizer enters the newly introduced bypass duct and sequentially exchanges heat in the high-temperature and low-temperature gas-water heat exchangers (HGHE and LGHE). Since the heat of the flue gas entering the main duct decreases and cannot meet the heat requirement for air preheating, two additional heat exchangers (FSHE and SSHE) are introduced for air preheating at low temperatures with the low-grade heat from the low-pressure steam extraction and the flue gas. The specifications of the main heat exchanges involved are presented in Table 2.   Figure 1. Schematic of the case with in-depth boiler-turbine interaction.

of 22
As given in Table 2, by employing two additional heat exchangers (SSHE and FSHE), the logarithmic mean temperature difference of the whole air preheating process is controlled above 35 • C. Particularly, by introducing the one-third of the flue gas cooled from 385 • C down to 135 • C for feedwater preheating, parts of the first, second, third, fifth and sixth-stage steam extractions are saved and continue for expansion, thus leading to remarkable performance enhancement. This proposed system utilizes part of the waste heat from the exhausted flue gas with its temperature reducing from around 130 • C in the original design to 100 • C, due to the limitation of acid dew point of the flue gas, t adp . There are many methods to estimate the acid dew point [64]. Considering the acceptable deviation among various methods, the acid dew point is calculated here with the widely-employed method in China [65,66]: where t wdp is the dew point temperature of water with the corresponding water vapor pressure of the flue gas ( • C), β is the coefficient related to the excess air coefficient (β = 121) [67], S ar,zs and A ar,zs are the sulfur and ash content at as-received basis (kg/kcal), and finally the α ash is the fly-ash share in the fuel gas (0.9) [67]. Due to the relatively low sulfur content of the coal (approximately 0.87%), the acid dew point is estimated as 94.7 • C, which indicates that the outlet flue gas temperature of the novel WHUS can be reduced to 100 • C without serious corrosion problem.

The Case with ORC for Waste Heat Recovery
Since the in-depth boiler-turbine integration requires significant modification on the real equipment, the decision makers of existing power plants can hardly accept such a complex proposal due to potential risks and influences on the system reliability. Therefore, a simple end-of-pipe waste heat recovery system via ORC without any influence on the existing power plants ( Figure 2) might be easier to be accepted.
The WHUS via ORC employs a two-stage evaporator to recover the waste heat from flue gas, which is deployed between the electrostatic precipitator (ESP) and flue-gas desulfurization unit (FGD), and between the air preheater and the ESP, respectively. Evaporator 1 recovers waste heat from the flue gas between 100 • C to 130 • C. Evaporator 2 employs fluoroplastic heat exchanger technology [68] to eliminate the corrosion problem, which allows further reduction of flue-gas temperature down to 75 • C without affecting the operating efficiency of the FGD.
The ORC system without regenerating process has been considered, since the recuperative preheating does not always lead to the maximization of net additional power output [39]. In the ORC-evaporator (ORCE), the high-pressure, sub-cooled liquid-phase working fluid is vaporized and is then expanded in the ORC-turbine (ORCT) for power generation. The expanded superheated vapor  Figure 2. Schematic of the case with end-of-pipe waste heat utilization via ORC.

The Case with In-Depth Boiler-Turbine Integration and ORC
The concept considering both in-depth boiler-turbine integration and ORC is mainly for more advanced, integrated plant design and aims at a further performance enhancement compared with the cases described above. The system layout is illustrated in Figure 3 with the SSHE employed for air preheating, which indicates that only one evaporator is employed for the heat absorption of the ORC, compared with Figure 2. The flue-gas temperature entering the ESP and FGD are the same as those of the ORC WHUS.

Methodology
The simulations of all concepts described in the last section are performed by Ebsilon Professional 13 [69]. Since the performance of the involved ORC is largely affected by the selected organic working fluid for given heat source, the ORC working fluid and the corresponding evaporating and condensing pressures are first selected with genetic algorithm (GA) to maximize the power generation of ORC. Then, exergetic and economic evaluation methods are described for thermo-economic comparison of the four different cases (including the reference case).

Working Fluid Selection of ORC
For the given heat-source temperature, a pre-selection of working fluids is performed based on [32]. Eight working fluids, i.e., R134A, Pentane, RC318, R236FA, R245FA, R123, R141B and R11, are to be evaluated. It should be mentioned that here we focus on the thermo-economic evaluation; therefore, other criteria, e.g., safety, technological feasibility and environmental harmony, are not considered. For each working fluid, the genetic algorithm is employed to optimize the two operating pressures of the ORC. The maximum allowed evaporating pressure is selected by the heat-source temperature T s,in and the approach temperature difference ∆T min (see Figure 4, set as 10 • C [41]). The pinch-point temperature difference (PPTD) is set as 5 • C [41]. Note that these two temperature differences are fixed, since only thermodynamic but not economic objective is considered here.

Working Fluid Selection of ORC
For the given heat-source temperature, a pre-selection of working fluids is performed based on [32]. Eight working fluids, i.e., R134A, Pentane, RC318, R236FA, R245FA, R123, R141B and R11, are to be evaluated. It should be mentioned that here we focus on the thermo-economic evaluation; therefore, other criteria, e.g., safety, technological feasibility and environmental harmony, are not considered. For each working fluid, the genetic algorithm is employed to optimize the two operating pressures of the ORC. The maximum allowed evaporating pressure is selected by the heat-source temperature , and the approach temperature difference ∆ (see Figure 4, set as 10 °C [41]). The pinch-point temperature difference (PPTD) is set as 5 °C [41]. Note that these two temperature differences are fixed, since only thermodynamic but not economic objective is considered here.  The thermodynamic model of the sub-critical ORC is classical [39]: where the is the heat transferred in the ORC evaporator, is the mass flow of the organic working fluid, is the mass flow of the heat source (flue gas).
(2) ORC Turbine where the is the power output of the ORC turbine and is the isentropic efficiency of the ORC turbine (88% [70]).
Tc-r The thermodynamic model of the sub-critical ORC is classical [39]: .
where the . Q evap is the heat transferred in the ORC evaporator, . m wf is the mass flow of the organic working fluid, . m s is the mass flow of the heat source (flue gas).
where the . W ORCT is the power output of the ORC turbine and η ORCT is the isentropic efficiency of the ORC turbine (88% [70]). .
where the . Q cond is the heat transferred in the ORC condenser and . m c is the mass flow of the cooling water. (4) ORC Pump .
where the . W ORCP is the power consumed by the ORC pump and η ORCP is the isentropic efficiency of the ORC pump (80% [70]).
The pressure drops as well as heat losses are neglected for all ORC equipment [39]. The net power output ( . W ORC ) of the ORC is given as below: where .
W ORCC is the power consumed by the recycle cooling water pump, which is calculated with the inlet temperature of the cooling water of 20 • C.
The optimization procedure for working-fluid selection and the corresponding operating variables is illustrated in Figure 5.
where the is the power consumed by the ORC pump and is the isentropic efficiency of the ORC pump (80% [70]).
The pressure drops as well as heat losses are neglected for all ORC equipment [39]. The net power output ( ) of the ORC is given as below: where is the power consumed by the recycle cooling water pump, which is calculated with the inlet temperature of the cooling water of 20 °C.
The optimization procedure for working-fluid selection and the corresponding operating variables is illustrated in Figure 5.

Exergy Analysis
Exergy analysis is employed in this paper to explore the difference in the energy utilization of the four cases. Exergy analysis can identify the sources and the magnitudes of exergy destructions and losses of energy systems [71][72][73], thus highlighting the processes with the largest exergy destruction or loss as the key components or processes for potential system improvement [1,[74][75][76]. Note that the exergy losses do not occur at a single component level but only at the overall system level, when the system is interacting with the environment via vent streams.
The exergy balance of the component can be calculated as following formulation [77]: where the symbols , , , represent the fuel exergy, product exergy and exergy destruction,

Exergy Analysis
Exergy analysis is employed in this paper to explore the difference in the energy utilization of the four cases. Exergy analysis can identify the sources and the magnitudes of exergy destructions and losses of energy systems [71][72][73], thus highlighting the processes with the largest exergy destruction or loss as the key components or processes for potential system improvement [1,[74][75][76]. Note that the exergy losses do not occur at a single component level but only at the overall system level, when the system is interacting with the environment via vent streams.
The exergy balance of the component k can be calculated as following formulation [77]: .
where the symbols E D,k represent the fuel exergy, product exergy and exergy destruction, respectively. With the clear definitions of fuel exergy, product exergy, the exergy efficiency of a productive component k can be expressed as: The exergy balance equation of the overall system can be formulated as below: .
with the total exergy efficiency of the overall system defined as The detailed calculation procedure of chemical and physical exergies about various material flows, work and heat streams can be referred in the literature [78]. The chemical exergy of coal also calculated as the product of the higher heating value (HHV) with a constant, usually 1.02 [77]. In addition, the reference environment for chemical exergy calculation is defined at 298.15 K and 1.098 bar [77].

Economic Analysis
Basic economic assumptions employed are listed as follows: (1) The on-grid power tariff is set as 0.061 USD/kWh and the annual full-load operating hours (e.g., the annual utilization hours) are assumed as 5000 h, which indicates that the average operating load factor is below 100% due to the frequent participation in peak load regulation of large-scale coal-fired power units; (2) The auxiliary cost is fixed as 15% of the bare module cost for all components [79], which indicates a common total module factor ( f TM ) of 0.15; (3) The operation and maintenance cost (C O&M ) accounts for 4% of the annualized total capital investment (TCI) [24,80]; (4) The exchange rate of RMB to U.S. dollar is set as 6.25 CNY/USD.

Estimation of Purchased Equipment Cost
The total capital investment of a component can be estimated based on its purchased equipment cost (PEC) [26,81], which is employed for the estimation of the corresponding bare module cost (C BM ) by further considering a bare module factor ( f BM ) due to the selection of material, the operating temperature and pressure, etc. Thus, for this retrofit problem, the total capital investment cost can be calculated as follows: The purchased equipment cost of a component can be estimated by a simple widely-used scaling method [81,82], if not many investment data are available for the considered component type: where the symbols S and α denote the selected sizing parameter and scaling factor, respectively, while the PEC rerf and S ref represent the reference capital investment and scaling parameter of the  [83,84]: where the co-efficients K 1 -K 3 can be obtained by the regression and are 3.853, 0.424 and 0 for heat exchangers, 3.579, 0.321 and 0.003 for the pump, and 3.514, 0.598 and 0 for turbine [83]. For heat exchangers, pump and turbine, the sizing parameters are taken as area and shaft work, respectively. The bare module factor ( f BM ) is generally a function of the material factor ( f M ), the pressure factor ( f P ), the temperature factor ( f T ) as well as the actualization factor ( f A ) as follows [82]: The actualization factor is employed to convert the capital cost in the reference year to the actualization year, which can be based on different cost index, e.g., Marshall and Swift index (MS) and Chemical Engineering Plant Cost Index (CEPCI) [83].
For the system with only in-depth boiler-turbine integration, the TCI for retrofitting the original power plant mainly consists of the capital investment of the newly-added exchangers (HEX). The related capital investment is estimated based on the simple method (Equation (16)) with the reference cost data and all necessary factors given in Table 3. The bare module cost, in this case, is a product of all factors: For the case with only the ORC for waste heat recovery, the total capital investment comes from the evaporate, condenser, pump, and turbine, since the contribution of the fluid cost to the total investment cost is negligible, as shown in Refs. [34,42,83]. The costs of the evaporator and condenser are estimated by Equation (17) with the bare module factor obtained by [85]: and the pressure factor formulated as The coefficients B 1 , B 2 , C 1 , C 2 , C 3 are 1.53, 1.27, 0, 0, and 0 for turbines and heat exchangers, 1.8, 1.51, 0.168, 0.348 and 0.484 for pumps, respectively [83]. The costs of pumps and turbines are computed via Equations (18) and (19).
For the case with both ORC and in-depth boiler-turbine integration, the total capital investment is comprised of all newly-added heat exchangers (including the gas-air HEX, water-air HEX, ORC evaporator and condenser), pumps and turbines with the equations and coefficients listed above.
With the total capital investment calculated for a specific retrofit, the annualized investment capital cost (C TCI ) can be calculated as [81,82,86]: where the symbol i refers to the interest rate per year (8%), while the n represents the system lifespan (20 years).

Economic Performance Indicators
The feasibility of the three different retrofit concepts is further evaluated by an economic benefit, which is represented by the net annual revenue (N AR): where the EAI stands for the additional income per year of the overall plant after the specific retrofit: where the ∆P net represents the net additional power output for the full-load operation of the plant after the retrofit, the τ is the equivalent full-load operation hours per year (5000 h as mentioned above), and the C e stands for the tariff of the electricity sent to the grid. In addition, the pay-back period (Pt) for the total capital investment, a vital indicator to show the economic feasibility of a technology, is calculated as follows [87]:

Results and Discussion
The three different retrofit concepts are comprehensively evaluated and discussed as follows: the selection of working fluid of the ORC, the comparison of the thermodynamic performances and the comparison of economic performances.

The Selection of ORC Working Fluid
Eight working fluids are screened in terms of the net additional power output with respect to two operating variables: the pressures of evaporation and condensation. The optimal operating conditions of each working fluid are given in Table 4 for the system retrofitted with both the in-depth boiler-turbine integration and ORC. From the perspective of the turbine output power, the R134A is the optimal fluid with 3.07 MW total power generated, followed by the R11 (2.93 MW). The T-s diagrams of both R134A-and R11-based ORC are illustrated in Figure 6 for the thermodynamic comparison of the two cycles. It is also found in Table 4 that the power-generation difference between different working fluids is not significant. Considering the pumping work, the fluid R11 achieves the largest net power generation (2.52 MW) with the corresponding evaporation pressure of 0.48 MPa and the condensation pressure of 0.13 MPa; while the fluid R134A requires higher pump work due to a much higher evaporation pressure (2.69 MPa, which leads to a net power output of 2.41 MW). Therefore, we consider the fluid R11 as the optimal. All the calculations and comparisons related to ORC are based on this fluid and its corresponding optimal working conditions. From the perspective of the turbine output power, the R134A is the optimal fluid with 3.07 MW total power generated, followed by the R11 (2.93 MW). The T-s diagrams of both R134A-and R11based ORC are illustrated in Figure 6 for the thermodynamic comparison of the two cycles. It is also found in Table 4 that the power-generation difference between different working fluids is not significant. Considering the pumping work, the fluid R11 achieves the largest net power generation (2.52 MW) with the corresponding evaporation pressure of 0.48 MPa and the condensation pressure of 0.13 MPa; while the fluid R134A requires higher pump work due to a much higher evaporation pressure (2.69 MPa, which leads to a net power output of 2.41 MW). Therefore, we consider the fluid R11 as the optimal. All the calculations and comparisons related to ORC are based on this fluid and its corresponding optimal working conditions.

Exergy Analysis
Compared with the original net power output (993.8 MW), an additional 23.99 MW, 6.49 MW, and 26.51 MW net power is produced with only in-depth boiler-turbine integration, only the ORC, and both the ORC and boiler-turbine integration. The in-depth boiler-turbine integration contributes the most to boost the power output, since the temperature level of flue gas up to 385 °C has been considered for heat integration. The same decrease in the temperature difference of heat transfer,

Exergy Analysis
Compared with the original net power output (993.8 MW), an additional 23.99 MW, 6.49 MW, and 26.51 MW net power is produced with only in-depth boiler-turbine integration, only the ORC, and both the ORC and boiler-turbine integration. The in-depth boiler-turbine integration contributes the most to boost the power output, since the temperature level of flue gas up to 385 • C has been considered for heat integration. The same decrease in the temperature difference of heat transfer, which is the driving force of the increase in power output, allows much more heat available for power generation, due to the wide temperature range of heat source. The additional work resulted from the utilization of the waste heat of flue gas from 135 • C down to 75 • C itself is limited no matter how small the heat-transfer temperature difference is reached. However, the ORC can still be employed with the in-depth boiler-turbine integration to reach a high-level waste heat recovery of the overall system.
The exergy analysis provides more insights on how each new integration concept eventually improves the system performance, with the comparison of exergy destructions within different components for different concepts in Figure 7. Due to the increased power output with the same amount of fuel consumption, the overall exergy efficiency of the system has been increased by 0.39 percentage points when only with ORC, 1.1 percentage point when only with the in-depth boiler-turbine integration, and 1.25 percentage point when with both concepts. It is also found in the in-depth turbine-boiler integration system that the reduction of the exergy destruction within the boiler subsystem reaches 27.75 MW, far larger than that within the turbine subsystem, 0.48 MW. Therefore, it is concluded that the energy-saving improvement is mainly contributed by the performance enhancement of the boiler subsystem.
pth turbine-boiler integration system that the reduction of the exergy destruction within the boile bsystem reaches 27.75 MW, far larger than that within the turbine subsystem, 0.48 MW. Therefor is concluded that the energy-saving improvement is mainly contributed by the performanc hancement of the boiler subsystem. More specifically, the decrease in exergy destruction within the boiler island is mainly due t e improvement of air preheating (AP), which reduces its exergy destruction from 40.77 MW to 21.5 W. This is mainly because of the utilization of the two-stage air preheater exchangers as well as th o-stage gas-water heater exchangers in the bypass flue gas duct (BFGD). The heat transfe perature difference of the concept with both ORC and in-depth boiler-turbine integratio creases significantly due to the better temperature match of the hot and cold fluids after dividin e air preheating process into three parts. Although with the bypass flue-gas duct, the exerg structions within the high-temperature gas-water heat exchangers and low-temperature gas-wate at exchangers increase by 2.38 and 0.86 MW, the system recovers 26.46 MW exergy from the flu s going through the bypass flue-gas duct. Therefore, the exergy destruction within the boiler islan n be saved up to 27.75 MW when with both ORC and in-depth boiler-turbine integration, and th ergy losses from flue gas (EFG) are largely reduced as well.
For the turbine island, the variations of the exergy destruction within the turbine (TURB) an ndenser (CD) are rather small (only 1.66 MW), compared to that within the regenerative system S). The beauty of the in-depth boiler-turbine integration is that the new layout promotes som dwater and condensate water acquiring heat from the flue gas via the two-stage gas-water he changers configured in the bypass flue gas duct in the boiler island, which generally results in More specifically, the decrease in exergy destruction within the boiler island is mainly due to the improvement of air preheating (AP), which reduces its exergy destruction from 40.77 MW to 21.54 MW. This is mainly because of the utilization of the two-stage air preheater exchangers as well as the two-stage gas-water heater exchangers in the bypass flue gas duct (BFGD). The heat transfer temperature difference of the concept with both ORC and in-depth boiler-turbine integration decreases significantly due to the better temperature match of the hot and cold fluids after dividing the air preheating process into three parts. Although with the bypass flue-gas duct, the exergy destructions within the high-temperature gas-water heat exchangers and low-temperature gas-water heat exchangers increase by 2.38 and 0.86 MW, the system recovers 26.46 MW exergy from the flue gas going through the bypass flue-gas duct. Therefore, the exergy destruction within the boiler island can be saved up to 27.75 MW when with both ORC and in-depth boiler-turbine integration, and the exergy losses from flue gas (EFG) are largely reduced as well.
For the turbine island, the variations of the exergy destruction within the turbine (TURB) and condenser (CD) are rather small (only 1.66 MW), compared to that within the regenerative system (RS). The beauty of the in-depth boiler-turbine integration is that the new layout promotes some feedwater and condensate water acquiring heat from the flue gas via the two-stage gas-water heat exchangers configured in the bypass flue gas duct in the boiler island, which generally results in a significant reduction of the heat needed from the steam extractions and a boost of power output due to this suppress of steam extractions. The exergy destruction within the RS is reduced remarkably by 21.87 MW. However, the additional heat exchangers require much power to overcome the flow resistance, the exergy destruction within the other parts of the turbine (OETL), including the valves and pipeline, etc., increases also dramatically due to the pure power consumption, which eventually leads to an increase in total exergy destruction within the turbine island by 0.48 MW, compared to that of the original system. For the ORC system, the exergy destructions are 13.26 MW and 8.19 WM for the systems with only the ORC and with both the ORC and the in-depth turbine-boiler interaction. The difference is mainly due to the different layouts of the ORC systems in both concepts, since there is two-stage evaporator for the former while one-stage evaporator for the latter.
More thermodynamic insights are given in Figure 8 to further illustrate how the system is improved with different integration concepts. With only one-stage of air-preheating, the mass flowrate and temperature of the air at an intermediate stage cannot be flexibly adjusted, which means the slopes of the two straight lines in Figure 8a are fixed. The upper-terminal temperature difference and the exergy destruction of the air preheater can be reduced by an increased air-preheater area; however, the increase in the final air-preheating temperature will increase the temperature of combustion, thus increasing the exergy destruction within the remaining boiler subsystem. This eventually increases the inlet temperature of the flue gas entering the air preheater. Therefore, increasing the air-preheater area is not a way of reasonably utilizing the heat at the boiler cold end. With the integration of only ORC (Figure 8b), there is no modification of the boiler subsystem at all. Only the heat of the flue gas at a temperature lower than 100 • C is utilized, which brings additional benefits without affecting the existing plant. However, the benefit is largely limited by the available heat extracted from the flue gas at a low temperature until the due temperature.
The essence of the in-depth boiler-turbine integration is given in Figure 8c. With the splitting of the hot and cold fluids, the slopes of the Q-T profiles are in fact readily adjusted, which gives a much better opportunity for establishing better matches of temperature levels of heat transfer. Particularly, the temperature difference of the bypass flue gas duct can be reduced largely but still beyond the industrial minimum temperature difference usually employed. In addition, by the splitting of flue gas and multiple stages of air preheating, a smaller temperature difference for heat transfer can be successfully achieved without modifying the system too much. Further, coupling with the ORC system for ultra-low temperature waste heat allows for an even larger benefit can be achieved based on the in-depth boiler-turbine integration.
The grand composite curves of the varied parts of the four plants ( Figure 8e) offers a much more straightforward comparison of the benefits from different concepts. With the in-depth boiler-turbine integration, the energy pocket enclosed by the flue gas is reduced significantly, which means more heat is extracted for better utilization (power generation). The pocket reduction realized by the ORC is also illustrated, although limited. 21.87 MW. However, the additional heat exchangers require much power to overcome the flow resistance, the exergy destruction within the other parts of the turbine (OETL), including the valves and pipeline, etc., increases also dramatically due to the pure power consumption, which eventually leads to an increase in total exergy destruction within the turbine island by 0.48 MW, compared to that of the original system. For the ORC system, the exergy destructions are 13.26 MW and 8.19 WM for the systems with only the ORC and with both the ORC and the in-depth turbine-boiler interaction. The difference is mainly due to the different layouts of the ORC systems in both concepts, since there is two-stage evaporator for the former while one-stage evaporator for the latter.
More thermodynamic insights are given in Figure 8 to further illustrate how the system is improved with different integration concepts. With only one-stage of air-preheating, the mass flowrate and temperature of the air at an intermediate stage cannot be flexibly adjusted, which means the slopes of the two straight lines in Figure 8a are fixed. The upper-terminal temperature difference and the exergy destruction of the air preheater can be reduced by an increased air-preheater area; however, the increase in the final air-preheating temperature will increase the temperature of combustion, thus increasing the exergy destruction within the remaining boiler subsystem. This eventually increases the inlet temperature of the flue gas entering the air preheater. Therefore, increasing the air-preheater area is not a way of reasonably utilizing the heat at the boiler cold end. With the integration of only ORC (Figure 8b), there is no modification of the boiler subsystem at all. Only the heat of the flue gas at a temperature lower than 100 °C is utilized, which brings additional benefits without affecting the existing plant. However, the benefit is largely limited by the available heat extracted from the flue gas at a low temperature until the due temperature.
The essence of the in-depth boiler-turbine integration is given in Figure 8c. With the splitting of the hot and cold fluids, the slopes of the Q-T profiles are in fact readily adjusted, which gives a much better opportunity for establishing better matches of temperature levels of heat transfer. Particularly, the temperature difference of the bypass flue gas duct can be reduced largely but still beyond the industrial minimum temperature difference usually employed. In addition, by the splitting of flue gas and multiple stages of air preheating, a smaller temperature difference for heat transfer can be successfully achieved without modifying the system too much. Further, coupling with the ORC system for ultra-low temperature waste heat allows for an even larger benefit can be achieved based on the in-depth boiler-turbine integration.
The grand composite curves of the varied parts of the four plants ( Figure 8e) offers a much more straightforward comparison of the benefits from different concepts. With the in-depth boiler-turbine integration, the energy pocket enclosed by the flue gas is reduced significantly, which means more heat is extracted for better utilization (power generation). The pocket reduction realized by the ORC is also illustrated, although limited.

Economic Analysis
The capital investment costs of the newly-added components of the three concepts are listed in Table 5. The additional capital investment cost introduced by the new components of in-depth turbine-boiler integration is 5.327 USD million, more than that of the ORC (3.951 USD million). The turbine and heat exchangers of the ORC contribute significantly to the total capital investment, leading to the fact that the potential of reducing the turbine cost and heat exchangers expense should be investigated. The integrated system with ORC and in-depth boiler-turbine integration needs the largest capital investment, 7.550 USD million, slightly less than the sum of both, due to the cost reduction of the equipment for the integrated system. The economic feasibility of the three improved systems is further investigated by various economic indicators listed in Table 6. The EAI of the in-depth turbine-boiler integration is 7.317 million USD, three times more than that of the system with only the ORC (1.979 million USD). However, too many newly-added components, together with the high capital investment, increase the system reliability and complexity, which further results in an increased operation and

Economic Analysis
The capital investment costs of the newly-added components of the three concepts are listed in Table 5. The additional capital investment cost introduced by the new components of in-depth turbine-boiler integration is 5.327 USD million, more than that of the ORC (3.951 USD million). The turbine and heat exchangers of the ORC contribute significantly to the total capital investment, leading to the fact that the potential of reducing the turbine cost and heat exchangers expense should be investigated. The integrated system with ORC and in-depth boiler-turbine integration needs the largest capital investment, 7.550 USD million, slightly less than the sum of both, due to the cost reduction of the equipment for the integrated system. The economic feasibility of the three improved systems is further investigated by various economic indicators listed in Table 6. The EAI of the in-depth turbine-boiler integration is 7.317 million USD, three times more than that of the system with only the ORC (1.979 million USD). However, too many newly-added components, together with the high capital investment, increase the system reliability and complexity, which further results in an increased operation and maintenance cost. The EAI of the concept with only the ORC is 1.979 million USD with a C TCI of 1.462 million USD. Therefore, the system with only the ORC is less suggested to be applied solely at the current technology level. The NAR of the integrated system is 4.990 million USD, indicating its excellent economic performance and better implementation potential. The investment-payback period of the system with only the ORC is almost 2.26 years to recover the total investment, while the system with the in-depth boiler-turbine integration only needs 0.78 years. The single implementation of the ORC is less economically viable since large capital investment is needed for less gain in power output. This also indicates that the ORC technology is still quite expensive, and its economic competitiveness needs to be further enhanced by the technology developers. By integrating both concepts, the integrated system has an intermediate payback time of 1.01 years, which is still acceptable and may be helpful to promote the development and to reduce the cost of the ORC technology.

Conclusions
In this paper, as a follow-up to our previous research, we proposed a novel waste heat recovery system based on a comprehensive understanding of the ORC and in-depth system integration. Then, we comparatively evaluated three integration concepts of the large-scale coal-fired power plant to better utilize the heat at an overall system level, particularly that at intermediate and low grade. The three concepts employ only ORC, only in-depth boiler-turbine integration, and both, respectively. A comprehensive and comparative evaluation of these three concepts is performed in terms of thermodynamic and economic performances to evaluate their feasibility and competitiveness. The major conclusions are as follows: For the in-depth boiler-turbine integration system, by splitting the flue gas, the heat capacities of the hot streams can be adjusted to have a better match with available cold streams, thus fulfilling much better overall heat integration. Therefore, the multi-stage air preheating is made possible and significantly reduces the heat-transfer temperature difference. The exergy destruction of the air preheating process is reduced by 19.23 MW. By introducing this part of flue-gas heat to the turbine system, the steam extractions can be largely suppressed, thus largely boosting the power output. The in-depth boiler-turbine integration demonstrated an additional power output of 23.99 MW and high economic competitiveness (with a payback time of only 0.78 years).
The selection of the working fluid with respect to available heat sources is important to maximize the ORC performance. The best ORC working fluid found out of eight for the given flue gas temperature (100 • C) is R11, which can be fully vaporized at such a low temperature and, therefore, offer better work ability. With the utilization of its latent heat, the mass flowrate of R11 in the ORC is much lower than those of the other working fluids. With R11-based ORC, the exergy loss from the exhausted flue gas is reduced from 20.69 MW to 5.90 MW, which results in an additional power output of 6.49 MW.
Compared with in-depth boiler-turbine integration, the waste heat recovery via ORC only delivers limited effect to further increase the power output; however, its high investment cost and lower payback time make its single application hardly acceptable in a near term. However, by coupling the ORC with in-depth boiler-turbine integration, the payback time can be largely reduced, indicating a better choice to promote the large-scale deployment of ORC and cost reduction. The inlet temperature of the flue gas, • C T s,out The outlet temperature of the flue gas, • C T s,r The temperature of the flue gas at the PPTD point, • C T c,out The outlet temperature of the cooling water, • C T c,in The inlet temperature of the flue gas, • C T c,r The temperature of the cooling water at the PPTD point,