Gas Turbine Intake Air Hybrid Cooling Systems and a New Approach to Their Rational Designing "2279

Gas turbine intake air cooling (TIAC) by exhaust gas heat recovery chillers is a general trend to improve turbine fuel efficiency at increased ambient temperatures. The high efficiency absorption lithium–bromide chillers of a simple cycle are the most widely used, but they are unable to cool inlet air lower than 15 ◦C. A two-stage hybrid absorption–ejector chillers were developed with absorption chiller as a high temperature stage and ejector chiller as a low temperature stage to subcool air from 15 ◦C to 10 ◦C and lower. A novel trend in TIAC by two-stage air cooling in hybrid chillers has been substantiated to provide about 50% higher annual fuel saving in temperate climate as compared with absorption cooling. A new approach to reduce practically twice design cooling capacity of absorption chiller due to its rational distribution with accumulating excessive refrigeration energy at decreased thermal loads to cover the picked demands and advanced design methodology based on it was proposed. The method behind this is issued from comparing a behavior of the characteristic curves of refrigeration energy required for TIAC with its available values according to various design cooling capacities to cover daily fluctuation of thermal loads at reduced by 15 to 20% design cooling capacity and practically maximum annual fuel reduction.


Introduction
A thermodynamic efficiency of all the combustion engines falls with increasing inlet air temperature: electric power drops and specific fuel consumption grows [1,2]. The combustion gas turbines (GT) are especially sensible to ambient air temperature at the inlet [3,4]. Therefore, it is quite reasonable to convert the heat of exhaust gas to refrigeration for intake air cooling [5,6]. The general trend in enhancing the fuel efficiency of GT at increased ambient air temperatures is turbine intake air cooling (TIAC) by exhaust heat recovery chillers [7,8]. The exhaust heat is mostly converted to refrigeration by absorption lithium-bromide chillers [9,10] or aqua-ammonia chillers [11,12]. Absorption lithiumbromide chillers (ACh) are the most efficient heat transformers [13,14]. The coefficients of performance (COP) of a simple cycle ACh are about 0.7 to 0.8 [15,16], but they are unable to cool inlet air lower than 15 • C, as the temperature of the chilled water is about 15 • C [17,18]. On using a refrigerant as a coolant, it is possible to cool air to 10 • C and lower [19,20]. The ejector chillers can be applied as the simplest in design and cheapest chillers [21,22]. However, their efficiency is not so high as ACh. The COP of ECh is about 0.3 [23,24], and

General Assumptions and Hypothesis
The following assumptions for the analysis of the operation efficiency of the TIAC systems and their rational design are accepted and approved.
So far, as proposed absorption-ejector TIAC systems are the advanced versions of typical basic absorption TIAC systems, the economic comparison with the last might be conducted taking into account only the cost of extra heat exchangers of ECh (refrigerant evaporator air cooler, refrigerant condenser, and ejector) with unchanged maintenance cost, personal, etc. Thus, the considered method of design focuses on providing just initial basic data as rational technical characteristics for further complicated detailed economic analysis.
The hypotheses accepted to prove novel approaches to the principals of proposed innovative TIAC system operation in account of the actual climatic conditions are the following.
In temperate climatic conditions, when cooling load on ACh falls, the effect in reduced specific fuel consumption of GT might be still considerably enlarged due to deeper cooling ambient air at the intake of GT to the temperatures 10 • C and lower, compared with its traditional cooling to 15 • C in the Ach of a simple cycle. According to this hypothesis, a deep GT intake air cooling is considered as a perspective trend of enhancing turbine efficiency [60].
In temperate climatic conditions, the fluctuations of current thermal loads on TIAC systems caused by variable actual ambient air parameters take place within the range of thermal load for cooling ambient air beyond 15 • C, i.e., within the range of the thermal load on ACh. The further subcooling air down to 10 • C and lower is conducted within a comparatively stable range of loading and can be covered by less-efficient ECh (COP is 0.2 to 0.3).
The absorption-ejector TIAC system provides two-stage cooling air consequently to 15 • C by chilled water from ACh and further subcooling to 10 • C and lower by boiling refrigerant from ECh [60].
The further reserves to enhance the efficiency of TIAC systems are within the range of fluctuated current thermal loads and can be realized by utilizing the excessive cooling capacity accumulated at lowered thermal load on ACh to cover picked loads. Such ambient air precooling would enable a nearly twofold reduction of the design cooling capacity of ACh.
The approach used in the present advanced designing method consists of considering the changeable excess (deficit) of the available design cooling capacity over current needs by the character of the curve of excessive refrigeration energy summarized over a certain period (monthly or yearly). With this, a continuously arising or stabilized character of the curve of the summarized available refrigeration energy excess (according to its design value) over its current needs indicates that daily accumulated excess of refrigeration energy covers daily needs and the thermal accumulator capacity is enough to cover daily thermal load fluctuation. A current drop in the summarized excess of available refrigeration energy indicates the need to increase a thermal accumulator's capacity or design the cooling capacity accordingly.
The annual fuel reduction ∑B is used as a primary criterion.
The fluctuations of the current changeable effect gained due to cooling (fuel reduction, power output increase, etc.) at varying thermal loading of TIAC systems [58] are considered through summarizing the current values of the effect and is evaluated by different rate of its annual value increment in response to the design of TIAC system cooling capacities [60]. Such an approach makes it possible to exclude the errors of 15 to 20%, caused by the approximation of the fluctuated current thermal loads in TIAC systems in typical designing  [59] and to avoid design cooling capacity overestimation of about 20% and simplify the calculation considerably.

The Computation Algorithm
The proposed methodology for the TIAC system design includes a determination of the chiller cooling capacity rational value to provide close to the maximum annual fuel saving as the first step and further distribution of the available cooling capacity in response to the current thermal demands as the second step.
The gas turbine annual fuel saving ΣB e due to the intake air cooling is assumed as a primary criterion to define a TIAC system rational design cooling capacity Q 0 .
The current fuel reduction B e is summarized over the year: where: ∆t a = t amb − t a2 -current turbine intake air temperature drop, K or • C; t amb and t a2 -ambient air and cooled air temperatures, K or • C; P e -turbine power output, kW; τ-time interval, h; b et -specific fuel reduction for 1 K (1 • C) air temperature drop, assumed as 0.7 g/(kWh·K) for UGT10000 [61,62]. The real input data of the on-site actual ambient air temperature t amb and relative humidity ϕ amb are used by applying the program "meteomanz" (http://www.meteomanz. com, accessed on 5 December 2021).
It is reasonable to analyze the fuel reduction's dependence on the specific cooling capacity q 0 as its absolute total value Q 0 refers to the unit air mass flow rate G a = 1 kg/s: where: ξ-specific heat ratio of the total latent and sensible heat extracted from the air being cooled to sensible heat extracted; c ma -moist air specific heat, kJ/kg·K. The variations in the current turbine fuel reduction B e are considered by the rate of their annual increment ΣB e , calculated as relative annual fuel saving increment ΣB e /Q 0 , referring to the cooling capacity required. Such a methodological approach enables increasing the accuracy of the results due to excluding the approximation of the current changeable values of B e and corresponding thermal loads with deviations about 10 to 15% [59]. This is a principally methodological novelty versus a generally accepted approach to cover the maximum current cooling demands to reach the maximum annual fuel saving ΣB e that leads to inevitable chillers design cooling capacity overestimation.
There are two methods developed: the first-by using the annual fuel saving ΣB e dependence on the chillers design cooling capacity to determine its rational value Q 0.rat , that provides closed to maximum annual fuel saving ΣB e , and the second-according to the maximum rate of annual fuel saving increment ΣB e /Q 0 to determine the optimum design cooling capacity Q 0.opt , that provides minimum sizes of the chillers and TIAC system ( Figure 1a).
The rational value of design cooling capacity Q 0.rat that provides a close to maximum annual fuel saving ΣB e is associated with the second maximum rate of the annual fuel saving increment ΣB e /Q 0 within its range beyond the first maximum rate: Q 0 > Q 0.opt and ΣB e > ΣB e . opt , respectively. With this, a relative parameter Σ(B e − B e.opt )/Q 0 is used as indicator to determine a rational value Q 0.rat (Figure 1b).
The optimum Q 0.opt and rational Q 0.rat values of the cooling capacities for t a2 = 10, 15, and 20 • C were calculated for temperate climatic conditions of Voznesensk, Mykolayiv region (southern Ukraine), 2017 year ( Figure 1). A maximum rate of annual fuel reduction ΣBe increment ΣBe/Q0•for ta2 = 10 °C takes place at the optimum design cooling capacity Q0.opt of about 900 kW (Figure 1a). A maximum rate of annual fuel reduction increment Σ(Be − Be.opt)/Q0 within the range beyond the value ΣBe.opt = 190 t corresponding to Q0.opt = 900 kW takes place at the rational design cooling capacity Q0.rat = 1450 kW and provides annual fuel reduction ΣBe.rat = 250 t that is very close to its maximum value 260 t but at a reduced design cooling capacity Q0.rat = 1450 kW less than Q0.max = 1700 kW by 15%.
The rational distribution of a design cooling capacity in response to the current thermal loads, as the second step of the methodology, requires comparing the available cooling capacity of the chillers with current cooling loads to determine the excessive available cooling capacity to cover the peaked current thermal loads.
Because of the great uncertainty of magnitude of the unstable boost ("turbulent") load range, its design value q0.b should be determined by a remaining principle as the difference between the overall design cooling capacity q0.10rat for cooling the ambient air to the target temperature ta2 = 10 °C and its basic stable ("laminar") load range q0.10-15 for subcooling the air from a threshold air temperature of about 15 °C after ACh to ta2 = 10 °C: q0.b10-15rat = q0.10rat − q0.10-15, where q0.10-15 = q0.10 − q0. 15.
The flow chart of the calculation procedure is presented in Appendix A.

Results
The further development of the methodology of the TIAC system designing consists of the distribution of the overall design cooling capacity between the unstable ("turbulent") thermal load range for ambient air precooling as a boost range covered by ACh and a stable ("laminarized") load range as a basic one for further air subcooling to the target temperature by ECh ( Figure 2).
The values of the current specific cooling capacities q0.15 required for cooling ambient air to 15 °C; rational design cooling capacities q0.10rat, q0.15rat, and q0.20rat for cooling ambient air to 10, 15, and 20 °C, respectively; the basic current cooling capacity q0.10-15 required for cooling air from 15 °C to 10 °C; and the available current residual boost cooling capacities q0.b10-15 and q0.b10-20 are calculated for climatic conditions in Voznesensk, Mykolayiv region (southern Ukraine) in July 2017 ( Figure 2).     A maximum rate of annual fuel reduction ΣB e increment ΣB e /Q 0 for t a2 = 10 • C takes place at the optimum design cooling capacity Q 0.opt of about 900 kW (Figure 1a). A maximum rate of annual fuel reduction increment Σ(B e − B e . opt )/Q 0 within the range beyond the value ΣB e . opt = 190 t corresponding to Q 0.opt = 900 kW takes place at the rational design cooling capacity Q 0.rat = 1450 kW and provides annual fuel reduction ΣB e . rat = 250 t that is very close to its maximum value 260 t but at a reduced design cooling capacity Q 0.rat = 1450 kW less than Q 0.max = 1700 kW by 15%.
The rational distribution of a design cooling capacity in response to the current thermal loads, as the second step of the methodology, requires comparing the available cooling capacity of the chillers with current cooling loads to determine the excessive available cooling capacity to cover the peaked current thermal loads.
Because of the great uncertainty of magnitude of the unstable boost ("turbulent") load range, its design value q 0.b should be determined by a remaining principle as the difference between the overall design cooling capacity q 0.10rat for cooling the ambient air to the target temperature t a2 = 10 • C and its basic stable ("laminar") load range q 0.10-15 for subcooling the air from a threshold air temperature of about 15 • C after ACh to t a2 = 10 • C: q 0.b10-15rat = q 0.10rat − q 0.10-15 , where q 0.10-15 = q 0.10 − q 0. 15 .
The flow chart of the calculation procedure is presented in Appendix A.

Results
The further development of the methodology of the TIAC system designing consists of the distribution of the overall design cooling capacity between the unstable ("turbulent") thermal load range for ambient air precooling as a boost range covered by ACh and a stable ("laminarized") load range as a basic one for further air subcooling to the target temperature by ECh ( Figure 2).
The values of the current specific cooling capacities q 0.15 required for cooling ambient air to 15 • C; rational design cooling capacities q 0.10rat , q 0.15rat , and q 0.20rat for cooling ambient air to 10, 15, and 20 • C, respectively; the basic current cooling capacity q 0.10-15 required for cooling air from 15 • C to 10 • C; and the available current residual boost cooling capacities q 0.b10-15 and q 0.b10-20 are calculated for climatic conditions in Voznesensk, Mykolayiv region (southern Ukraine) in July 2017 ( Figure 2). The rational design values of cooling capacities q0.10rat, q0.15rat, and q0.20rat for cooling air to 10, 15, and 20 °C are determined by developed method (Figure 1).
As Figure 2 shows, with cooling ambient air to 15 °C the fluctuations of the current thermal loads q0.15 are gradual, that points to inevitable considerable excess of design cooling capacity in the temperate daily hours. At the same time, when air is subcooled from 15 °C to 10 °C, the fluctuations in the thermal load q0.10-15 are comparably small. Thus, the temperature of cooled air ta2 = 15 °C can be reasonably assumed as the threshold value for shearing the overall design's thermal load on the TIAC system q0.10rat into a comparably stable ("laminarized") load range q0.10-15 and the boost unstable ("turbulent") range of ambient air precooling. So, the stable load value q0.10-15 is chosen as a basic stable part q0.10-15 of the design cooling capacity q0.10rat = 35 kJ/kg ( Figure 2a). Accordingly, the remaining part of q0.10rat is used for precooling the ambient air to the threshold temperature ta2 = 15 °C and determined as the boost cooling capacity q0.b10-15rat ( Figure 2b). The unstable q0.15 thermal load range can be covered by the ACh, as well as the stable q0.10-15 thermal load range by ECh ( Figure 3).  Figure 2. Current values of specific cooling capacities q 0.15 needed for cooling ambient air to t a2 = 15 • C; rational design cooling capacities q 0.10rat , q 0.15rat , and q 0.20rat for t a2 = 10, 15, and 20 • C, respectively; current cooling capacities q 0.10-15 for cooling air from t a2 = 15 to 10 • C and current residual boost part q 0.b10-15 available for cooling air t a2 = 15 • C (a) and boost part q 0.b10-20 available for cooling air to t a2 = 20 • C (b).
The rational design values of cooling capacities q 0.10rat , q 0.15rat , and q 0.20rat for cooling air to 10, 15, and 20 • C are determined by developed method (Figure 1).
As Figure 2 shows, with cooling ambient air to 15 • C the fluctuations of the current thermal loads q 0.15 are gradual, that points to inevitable considerable excess of design cooling capacity in the temperate daily hours. At the same time, when air is subcooled from 15 • C to 10 • C, the fluctuations in the thermal load q 0.10-15 are comparably small. Thus, the temperature of cooled air t a2 = 15 • C can be reasonably assumed as the threshold value for shearing the overall design's thermal load on the TIAC system q 0.10rat into a comparably stable ("laminarized") load range q 0.10-15 and the boost unstable ("turbulent") range of ambient air precooling. So, the stable load value q 0.10-15 is chosen as a basic stable part q 0.10-15 of the design cooling capacity q 0.10rat = 35 kJ/kg ( Figure 2a). Accordingly, the remaining part of q 0.10rat is used for precooling the ambient air to the threshold temperature t a2 = 15 • C and determined as the boost cooling capacity q 0.b10-15rat (Figure 2b). The unstable q 0.15 thermal load range can be covered by the ACh, as well as the stable q 0.10-15 thermal load range by ECh ( Figure 3).
According to the proposed circuit of the hybrid two-stage TIAC system, the ACh provides ambient air cooling to 15 • C in a high-temperature air cooler AC HT by chilled water with temperature of about 7 • C, and ECh provides further cooling air to 10 • C and lower in a low-temperature air cooler AC LT by refrigerant boiling at the temperatures of about 5 • C to 2 • C. The thermal load of the ECh is stabilized due to ambient air cooling to threshold temperature of about 15 • C. A steam generated in the exhaust gas boiler is used for feeding to generators of ACh and ECh. The heat of the condensation is extracted from the ACh and ECh condensers through cooling towers. According to the proposed circuit of the hybrid two-stage TIAC system, the ACh provides ambient air cooling to 15 °C in a high-temperature air cooler ACHT by chilled water with temperature of about 7 °C, and ECh provides further cooling air to 10 °C and lower in a low-temperature air cooler ACLT by refrigerant boiling at the temperatures of about 5 °C to 2 °C. The thermal load of the ECh is stabilized due to ambient air cooling to threshold temperature of about 15 °C. A steam generated in the exhaust gas boiler is used for feeding to generators of ACh and ECh. The heat of the condensation is extracted from the ACh and ECh condensers through cooling towers.
As Figure 2a shows, the available boost cooling capacity q0.b10-15rat generally covers current thermal loads q0.15 for precooling the ambient air to ta2 = 15 °C. Furthermore, even less available boost cooling capacity q0.b10-20rat also covers the current loads q0.15 except for quite short periods of daylight hours (Figure 2b).
It is quite evident that the further reserves to enhance the efficiency of TIAC systems are within the unstable boost load range of q0.b10-15rat for precooling ambient air to the temperature of 15 °C in ACh and can be realized by utilizing the excessive cooling capacity accumulated at a lowered thermal load on ACh to cover picked loads, which enables reducing the design cooling capacity of ACh.
Therefore, the next step in further enhancing the efficiency of the TIAC systems and development of their design methodology is issued from the behavior of daily available boost the thermal load q0.b10-20 and the current thermal load q0.15 fluctuations (Figure 4). As Figure 2a shows, the available boost cooling capacity q 0.b10-15rat generally covers current thermal loads q 0.15 for precooling the ambient air to t a2 = 15 • C. Furthermore, even less available boost cooling capacity q 0.b10-20rat also covers the current loads q 0.15 except for quite short periods of daylight hours (Figure 2b).
It is quite evident that the further reserves to enhance the efficiency of TIAC systems are within the unstable boost load range of q 0.b10-15rat for precooling ambient air to the temperature of 15 • C in ACh and can be realized by utilizing the excessive cooling capacity accumulated at a lowered thermal load on ACh to cover picked loads, which enables reducing the design cooling capacity of ACh.
Therefore, the next step in further enhancing the efficiency of the TIAC systems and development of their design methodology is issued from the behavior of daily available boost the thermal load q 0.b10-20 and the current thermal load q 0.15 fluctuations ( Figure 4).  As Figure 4 shows, the available boost cooling capacities q0.b10-20rat for cooling ambient air to ta2 = 20 °C generally cover the current thermal loads q0.15 for precooling ambient air to ta2 = 15 °C. It is approved by comparing the excesses of available design boost cooling capacities q0.b10-20rat − q0.15 over required q0.15 for ta2 = 15 °C and deficit of design boost cooling capacity q0.15 − q0.b10-20rat ( Figure 5). As Figure 4 shows, the available boost cooling capacities q 0.b10-20rat for cooling ambient air to t a2 = 20 • C generally cover the current thermal loads q 0.15 for precooling ambient air to t a2 = 15 • C. It is approved by comparing the excesses of available design boost cooling capacities q 0.b10-20rat − q 0.15 over required q 0.15 for t a2 = 15 • C and deficit of design boost cooling capacity q 0.15 − q 0.b10-20rat ( Figure 5). As Figure 4 shows, the available boost cooling capacities q0.b10-20rat for cooling ambient air to ta2 = 20 °C generally cover the current thermal loads q0.15 for precooling ambient air to ta2 = 15 °C. It is approved by comparing the excesses of available design boost cooling capacities q0.b10-20rat − q0.15 over required q0.15 for ta2 = 15 °C and deficit of design boost cooling capacity q0.15 − q0.b10-20rat ( Figure 5). Because of the large fluctuations in the available design boost cooling capacity q 0.b10-20rat for cooling ambient air to 20 • C, the corresponding excess available boost cooling capacity q 0.b10-20rat − q 0.15 over required q 0.15 for cooling air 15 • C, and the deficit of the boost cooling capacity q 0.15 − q 0.b10-20rat , it is problematic to estimate the daily accumulated excess of the design cooling capacity over its required value q 0.15 for cooling air to 15 • C during the month ( Figure 5).
Proceeding from this, a new approach to share the thermal loads within the unstable boost range q 0.b10-15rat in two parts, q 0.b10-20rat and q 0.15-20 , and analyze daily accumulated excess q 0.b10-20rat − q 0.15 of available boost cooling capacity q 0.20rat over cooling needs q 0.15 for cooling air to 15 • C is proposed. The accumulated excess of available boost cooling capacity q 0.b10-20rat − q 0.15 can be used to cover daily deficit q 0.15 − q 0.b10-20rat at peaked loads. On comparing current values of excess q 0.b10-20rat − q 0.15 and deficit q 0.15 − q 0.b10-20rat as resulting refrigeration energy accumulated Σ(q 0.b10-20rat − q 0.15 )τ, the hypothesis to use a reduced value q 0.20rat designed for cooling air to 15 • C to cover the current cooling needs q 0.15 has been approved ( Figures 5 and 6). This results in reduction of design boost thermal load range of q 0.b10-15rat and q 0.15rat , accordingly.
There following correlations are applied: Current excesses of available rational design boost cooling capacity q 0.b10-20rat − q 0.15 ; Current deficit of design boost cooling capacity q 0.15 − q 0.b10-20rat ; Summarized excess of available accumulated boost refrigeration energy Σ(q 0.b10-20rat − q 0.15 )τ. As Figure 5 testifies, the available boost cooling capacity q 0.b10-20rat designed for precooling ambient air to t a2 = 20 • C is generally enough to cover the current cooling needs q 0.15 for cooling air to t a2 = 15 • C. The current deficit of design boost cooling capacity q 0.15 − q 0.b10-20rat can be covered by using the daily accumulated excessive refrigeration energy. This statement is approved by the continuously arising curve of the summarized excess of available accumulated boost refrigeration energy over its current needs Σ(q 0.b10-20rat − q 0.15 )τ ( Figure 6). As Figure 5 testifies, the available boost cooling capacity q0.b10-20rat designed for precooling ambient air to ta2 = 20 °C is generally enough to cover the current cooling needs q0.15 for cooling air to ta2 = 15 °C. The current deficit of design boost cooling capacity q0.15 − q0.b10-20rat can be covered by using the daily accumulated excessive refrigeration energy. This statement is approved by the continuously arising curve of the summarized excess of available accumulated boost refrigeration energy over its current needs Σ(q0.b10-20rat − q0.15)τ ( Figure 6). A practically continuously arising character of the curve of the summarized excesses of the available boost refrigeration energy over its current needs Σ(q0.b10-20rat − q0.15)τ indicates the addition reserve to reduce a design boost cooling capacity for instance through using the optimum design boost cooling capacity q0.b10-20opt instead of its rational value q0.b10-20rat ( Figure 7).
As it is seen, the values of the available optimum design boost cooling capacity q0.b10-20opt for ambient air precooling to ta2 = 20 °C are less than the current values of the thermal load q0.15 needed for cooling ambient air to ta2 = 15 °C in the beginning of the month (from 9 to 13 July) and after 18 July. This leads to dropping the curve of the summarized excessive values over the current needs Σ(q0.b10-20opt − q0.15)τ as compared to the curve Σ(q0.b10-20rat − q0.15)τ of the summarized excessive available accumulated refrigeration energy according to rational design (Figure 7). A sharp falling character of the summarized excessive accumulated refrigeration energy curve Σ(q0.b10-20opt − q0.15)τ indicates the lack of daily accumulated refrigeration energy to cover current increased thermal loads q0.15. A practically continuously arising character of the curve of the summarized excesses of the available boost refrigeration energy over its current needs Σ(q 0.b10-20rat − q 0.15 )τ indicates the addition reserve to reduce a design boost cooling capacity for instance through using the optimum design boost cooling capacity q 0.b10-20opt instead of its rational value q 0.b10-20rat (Figure 7). The reason for this is a considerable reduced value of optimum design cooling capacity q0.10opt = 24 kJ/kg preliminary accepted as compared with rational value q0.10rat = 35 kJ/kg. Therefore, the next step of analysis focuses the choice of reduced design cooling capacity of TIAC system to minimize the duration of falling part of the summarized excessive available accumulated refrigeration energy curve Σ(q0.b10-20opt − q0.15)τ.
As it is seen, the values of the available optimum design boost cooling capacity q 0.b10-20opt for ambient air precooling to t a2 = 20 • C are less than the current values of the thermal load q 0.15 needed for cooling ambient air to t a2 = 15 • C in the beginning of the month (from 9 to 13 July) and after 18 July. This leads to dropping the curve of the summarized excessive values over the current needs Σ(q 0.b10-20opt − q 0.15 )τ as compared to the curve Σ(q 0.b10-20rat − q 0.15 )τ of the summarized excessive available accumulated refrigeration energy according to rational design (Figure 7). A sharp falling character of the summarized excessive accumulated refrigeration energy curve Σ(q 0.b10-20opt − q 0.15 )τ indicates the lack of daily accumulated refrigeration energy to cover current increased thermal loads q 0.15 .
The reason for this is a considerable reduced value of optimum design cooling capacity q 0.10opt = 24 kJ/kg preliminary accepted as compared with rational value q 0.10rat = 35 kJ/kg. Therefore, the next step of analysis focuses the choice of reduced design cooling capacity of TIAC system to minimize the duration of falling part of the summarized excessive available accumulated refrigeration energy curve Σ(q 0.b10-20opt − q 0.15 )τ.
As Figure 8 shows, the current values of the excess of the available boost cooling capacity q 0.b10-20m , based on the middle value q 0.10m = (q 0.10rat + q 0.10opt )/2, in the most days in July are higher than the current thermal loads q 0.15 . However, from 11 to 13 and 21 to 28 July, the values q 0.b10-20m are lower than q 0.15 and corresponding current excess of the available boost cooling capacity q 0.b10-20m − q 0.15 is less than the value of its deficit q 0.15 − q 0.b10-20m (Figure 9). Due to increased middle boost cooling capacity q 0.10m as compared with its optimum value q 0.10opt , a drop in the curve of the summarized excess of the available accumulated refrigeration energy Σ(q 0.b10-20m − q 0.15 )τ for the middle value q 0.10m is not so sharp as compared with the value of Σ(q 0.b10-20opt − q 0.15 )τ for the rational value q 0.10rat (Figure 9).
The analysis of the summarized values of the excessive available accumulated refrigeration energy Σ(q 0.b10-20 − q 0.15 )τ for different design cooling capacities q 0.10 makes it possible to determine their reasonably reduced design value to minimize the lack of daily accumulated refrigeration energy to cover current increased thermal loads q 0.15 ( Figure 10).
The proposed method of the thermal load analysis based on the character of the summarized excessive available accumulated makes it possible to reduce a design cooling capacity due to its rational distribution and simultaneously to cover the peaked current thermal loads by the excessive refrigeration energy accumulated daily.
The calculation results of reducing the installed cooling capacities of the chillers due to their rational distribution in response to actual current thermal loads according to proposed designing methodology are presented in Figure 11. As it was mentioned, the annual fuel reduction ΣB e is assumed as a primary criterion. marized excessive available accumulated makes it possible to reduce a design cooling capacity due to its rational distribution and simultaneously to cover the peaked current thermal loads by the excessive refrigeration energy accumulated daily.
The calculation results of reducing the installed cooling capacities of the chillers due to their rational distribution in response to actual current thermal loads according to proposed designing methodology are presented in Figure 11. As it was mentioned, the annual fuel reduction ΣBe is assumed as a primary criterion. Reduction of design cooling capacity of ACh boost stage Δq0.15-20 = q0.15rat − q0.20rat due to its rational distribution.
As Figure 11 shows, the rational designing of the TIAC systems provides a reduction of the installed cooling capacities of the chillers by the values of Δq0.10,15,20, i.e., by 15 to 20% compared with their maximum magnitudes q0.10,15,20max, calculated issuing from the conventional practice of designing. In addition, the rational distribution of the installed cooling capacity of ACh makes it possible to reduce a design boost thermal load by the value Δq0.15-20 = q0.15rat − q0.20rat, i.e., practically twice as high compared with the initial value q0.15rat.
As well as the proposed absorption-ejector TIAC systems are the advanced versions of typical basic absorption TIAC system, the economic comparison might be conducted taking into account only the cost of extra heat exchangers of ECh (refrigerant evaporatorair cooler, refrigerant condenser, and ejector) with unchanged maintenance cost, personal costs, etc.
As a basic version, a TIAC system for the gas turbine of the power output of 10 MW and air mass flow rate Ga = 40 kg/s is accepted for easy application of the calculation results for turbines with other characteristics. According to Figure 11, the rational specific cooling capacities q0.10rat = 35 kW/(kg/s) for the TIAC system with AECh and q0.15rat = 24 kW/(kg/s) for conventional ACh or their total values Q0.10rat = 1400 kW and Q0.15rat = 960 kW are assumed. The annual specific fuel reduction Σb10rat = 47 kg/kW and Σb15rat = 26 kg/kW or   The following correlations are used: Reduction of the design cooling capacity of AECh ∆q 0.10 = q 0.10max − q 0.10rat and of ACh ∆q 0.15 = q 0.15max − q 0.15rat for cooling to t a2 = 15 • C and ∆q 0.20 = q 0.20max − q 0.20rat for t a2 = 15 • C due to the rational design cooling capacities compared with traditional maximum values; Reduction of design cooling capacity of ACh boost stage ∆q 0.15-20 = q 0.15rat − q 0.20rat due to its rational distribution.
As Figure 11 shows, the rational designing of the TIAC systems provides a reduction of the installed cooling capacities of the chillers by the values of ∆q 0.10,15,20 , i.e., by 15 to 20% compared with their maximum magnitudes q 0.10,15,20max , calculated issuing from the conventional practice of designing. In addition, the rational distribution of the installed cooling capacity of ACh makes it possible to reduce a design boost thermal load by the value ∆q 0.15-20 = q 0.15rat − q 0.20rat , i.e., practically twice as high compared with the initial value q 0.15rat .
As well as the proposed absorption-ejector TIAC systems are the advanced versions of typical basic absorption TIAC system, the economic comparison might be conducted taking into account only the cost of extra heat exchangers of ECh (refrigerant evaporator-air cooler, refrigerant condenser, and ejector) with unchanged maintenance cost, personal costs, etc.
As a basic version, a TIAC system for the gas turbine of the power output of 10 MW and air mass flow rate G a = 40 kg/s is accepted for easy application of the calculation results for turbines with other characteristics. According to Figure 11, the rational specific cooling capacities q 0.10rat = 35 kW/(kg/s) for the TIAC system with AECh and q 0.15rat = 24 kW/(kg/s) for conventional ACh or their total values Q 0.10rat = 1400 kW and Q 0.15rat = 960 kW are assumed. The annual specific fuel reduction Σb 10rat = 47 kg/kW and Σb 15rat = 26 kg/kW or corresponding gas turbine annual total fuel reduction ΣB 10rat = 470 t and ΣB 15rat = 260 t are gained due to the intake of air cooling to 10 • C and 15 • C, accordingly. With this, a deeper air subcooling from 15 • C to 10 • C in ECh of the total cooling capacity Q 0.10-15rat = Q 0.10rat − Q 0.15rat , i.e., 440 kW, provides additional annual fuel reduction ΣB 10-15rat = ΣB 10rat − ΣB 15rat about 210 t as compared with conventional TIAC system with ACh.
Thus, the following additional heat exchanges of ECh are needed: refrigerant evaporator-air cooler with cooling capacity of Q 0.10-15rat = 440 kW and refrigerant condenser with thermal load of Q condenser = Q 0. 10-15rat (1 + COP), i.e., about 1760 kW, where COP of ECh can be accepted as 0.3. Taking into account the cost of additional heat exchangers of about USD 150,000 according to [63], increased by 10% for the mounting of ejector and heat exchangers, the cost of the additional equipment is about USD 180,000. At the same time, the cost of gas fuel annually saved is about USD 300,000. Proceeding from the price of gas of USD 1000 per 1000 m 3 , the payback period is about a year.
Because of the fluctuations in the price of gas fuel and the cost of heat exchangers from various manufacturers the economic analysis has to be conducted for a quite definite case. The considered method of designing focuses to provide the rational technical characteristics as initial basic data for further complicated detailed economic analysis.

Conclusions
A novel trend in TIAC by two-stage air cooling in combined AECh with ACh as a high-temperature stage for ambient air cooling to cover the range of fluctuated thermal load and ECh as a low-temperature stage for further air subcooling at practically stable thermal load is proposed to provide about 50% higher annual fuel saving in temperate climatic conditions.
The reserve of further enhancing the efficiency of the TIAC system by the rational distribution of design cooling capacity within the range of fluctuated thermal load is revealed. It can be realized by using the excess of ACh design cooling capacity accumulated at lowered current thermal loads to cover the peaked loads that enables an almost twofold reduction in the installed cooling capacity of ACh.
An advanced methodology is developed to determine a rational design cooling capacity of TIAC systems that provide close to the maximum annual fuel savings at decreased installed cooling capacity by 15 to 20% as compared with conventional TIAC designing practice.
Issuing from the daily fluctuations of the excess and deficits of cooling capacities, a new methodological approach is proposed to define a design cooling capacity of ACh reduced practically twofold due to its rational distribution through accumulating excessive refrigeration energy at lowered current thermal loads to cover their picked values by boost air precooling.  ξ specific heat ratio of the total heat (latent and sensible) rejected from air to its sensible heat τ time interval h ϕ amb ambient air relative humidity % ∆b e specific fuel consumption decrease g/kWh ∆t air temperature decrease K, • C ∑B 10,15,20 annual total fuel reduction due to cooling engine intake air to 10, 15, 20 • C t ∑b 10,15,20 annual specific fuel reduction (per 1 kW engine power output) due g, kg, t to cooling engine intake air to 10, 15, 20 • C  Subscripts  a  air  amb  ambient  b  boost  def  deficit  exc  excess  max  maximum  m  mean  opt  optimum  rat  rational