Gas turbine intake air hybrid cooling systems and their rational designing

The general trend to improve the fuel efficiency of gas turbines (GT) at increased ambient temperatures is turbine intake air cooling (TIAC) by exhaust heat recovery chillers The high efficiency absorption lithium-bromide chillers (ACh) of a simple cycle are the most widely used, but they are not able to cool intake air lower than 15°C because of a chilled water temperature of about 7°C. A two-stage hybrid absorption-ejector chillers (AECh) were developed with ejector chiller as a low temperature stage to provide deep air cooling to 10°C and lower. A novel trend in TIAC by two-stage air cooling in chillers of hybrid type has been proposed to provide about 50% higher annual fuel saving in temperate climatic conditions as compared with ACh cooling. The advanced methodology to design and rational distribute the cooling capacity of TIAC systems that provides a closed to maximum annual fuel reduction without oversizing was developed.


Introduction
The efficiency of gas turbines (GT) decreases with arising the ambient air temperature at their inlet [1,2]. The general trend to improve the fuel efficiency of GT at increased ambient air temperatures is turbine intake air cooling (TIAC) by exhaust heat recovery chillers [3]. A reduction of the chiller sizes with maximum annual fuel saving is possible due to rational design cooling capacity excluding oversizing [4,5] and its rational distribution with small deviation of current loads from a design value [6,7]. In order to realize this the overall band of current cooling loads is to be divided into two ranges: the first unstable load range, following the fluctuations of current loads (thermal "turbulences"), and the second range of comparatively stable loads ("laminarized" thermal load range). The cooling capacity of the chillers is to be designed to cover the thermal "turbulences" by the absorption lithiumbromide chiller (ACh) chillers with a high coefficient of performance (COP) not effected by load fluctuations considerably [8]. The further air subcooling takes place within the comparatively stable "laminarized" thermal load range) and can be covered by ejector chiller (ECh) as the most simple in design and cheapest but considerably effected by load changes [9]. The application of such hybrid absorption-ejector chiller (AECh) enables to cover actual loading in two-stage air cooler with boost high temperature water stage and low temperature refrigerant stage [10].
The purpose of the study is to develop the advanced hybrid TIAC systems and the improved methodology of * Corresponding author: nirad50@gmail.com their designing with rational distribution of the overall design cooling capacity between unstable ("turbulent") thermal load range for ambient air precooling in the boost high temperature stage of the air cooler (AC) by ACh and a stable ("laminarized") load range for further air subcooling to the target temperature in the low temperature stage by ECh that provides practically twice reduction of a design boost thermal load and about 50% higher annual fuel saving as compared with ACh gained due to applications of TIAC systems with hybrid AECh.

Literature review
In a number of investigations the combustion engine intake air cooling (EIAC) including TIAC is considered as subtechnologies for combined cooling, heating and power (CCHP) [11], or trigeneration [12,13]. A lot of researches are focused to improve the performance of air cooling systems by intensification of heat transfer in evaporators and condensers [14].
The technical innovations in waste heat recovery [15,16] including transport application [17,18] might be successfully applied in TIAC: two-stage intake air cooling [10], deep exhaust heat utilization [19,20]. The heat potential for converting in refrigeration can be increased due to low-temperature condensation [21,22].
Advanced methods as ANSIS [23] and statistical methods for processing monitoring data can be used for optimizing the cooling loads according site climatic conditions [24,25] and along ship voyages [26]. The sinusoidal curve was proposed for daily thermal load fluctuations [27] to match current cooling demands.
Practically all the typical design methods [28,29] issue from the assumption of a design cooling capacity to cover maximum cooling needs over the full range of yearly operating conditions, that inevitable leads to considerable cooling system oversizing and requires to define a correct design cooling load excluding overestimation.

Research methodology
A reduction of the chillers design cooling capacity is possible by determining its rational value to provide closed to maximum annual fuel saving as the first step of the methodology for designing the TIAC system and further distribution of the available cooling capacity in response to the current demands as the second step.
The annual fuel saving Be of the GT due to inlet air cooling is assumed as a criterion to determine a rational design cooling capacity Q0 of the TIAC system. With this the current fuel reduction Be have been summarized over the year: where: Δta = tamb -ta2 -current intake air temperature drop, K or °C; tamb and ta2 -ambient air and air temperature at the air cooler outlet, K or °C; Ne -turbine power output, kW; τ -time interval, h; bet -specific fuel reduction for 1K (1°С) air temperature drop, assumed 0.7 g/(kWh·K) for UGT10000 [30]. It is preferably to analyze the fuel reduction in dependence on specific cooling capacity q0 as the overall its value Q0 referred to air mass flow rate Ga = 1 kg/s: where: ξ -specific heat ratio; сma -moist air specific heat, kJ/(kg·K. According to the method developed the fluctuations of the current effect in GT fuel reduction Be are considered by the rate of their annual increment ∑Be as relative annual fuel saving increment ΣBe /Q0 referred the cooling capacity needed. A such methodological approach makes it possible to increase the accuracy of the results due to excluding the approximation of the current changeable values of Be . This is a principally novelty versus a generally accepted approach to cover the maximum current demands to reach the maximum annual value ∑Be that leads to oversizing. There are two methods developed: the first -by using the annual fuel reduction ∑Bf dependence on the design cooling capacity of the chiller to choose its rational value Q0.rat , that provides closed to maximum annual fuel reduction ∑Be , and the second -according to the maximum rate of annual fuel reduction ∑Be increment ∑Be /Q0 to choose optimum design cooling capacity Q0.opt , that provides minimum sizes of the chiller and TIAC system ( Fig.1,a).
The rational value of design cooling capacity Q0.rat , providing a closed to maximum annual fuel reduction ∑Be is associated with the second maximum rate of annual fuel reduction ∑Be increment within its range beyond the first maximum rate: Q0 >Q0.opt and ∑Be >∑Be.opt accordingly. With this a relative parameter ∑(Be -Be.opt )/Q0 is used as indicator to choose a rational value Q0.rat (Fig.1,b).
The optimum Q0.opt and rational Q0.rat cooling capacities for ta2 = 10, 15 and 20 °С were calculated for temperate climatic conditions of Voznesensk, Nikolaev region, southern Ukraine, 2017 year (Fig. 1). Fig.1. Annual fuel reduction ∑Be and its relative values ∑Be /Q0 referred to design cooling capacity Q0 over the whole range of ∑Be (a) and values ∑(Be-Be.opt )/Q0•beyond the optimum values of ∑Be.opt and Q0.opt (b) for cooling ambient air to ta2 = 10, 15 and 20 °C.
A maximum rate of annual fuel reduction ∑Be increment ∑Be /Q0•for ta2 = 10 °C takes place at the optimum design cooling capacity Q0.opt of about 900 kW ( Fig.1, a). A maximum rate of annual fuel reduction increment ∑(Be -Be.opt )/Q0 within the range beyond the value ∑Bf.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 ∑Be.rat = 250 t that is very closed 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, revealed at the lowered current thermal loads on the air cooler (AC) at the inlet of GT, to cover the peaked current thermal loads.
Because of great uncertainty of unstable boost ("turbulent") load range magnitude its design value q0.b should be determined by a remaining principle as a difference between the overall design cooling capacity q0.10 for the whole process of cooling the ambient air to the target temperature ta2 = 10 °C and its basic stable ("laminar") load range q0.10-15 for subcooling air from a threshold air temperature of about 15 °C after ACh to ta2 = 10 °C: q0.b = q0.10 -q0.10-15 , where q0.10-15 = q0.10 -q0.15 .

Results
The further development of the methodology of TIAC system designing involves distribution of the overall design cooling capacity between unstable ("turbulent") thermal load range for ambient air precooling in the boost high temperature stage of the air cooler (AC) by ACh and a stable ("laminarized") load range for further air subcooling to the target temperature in the low temperature stage by ECh (Fig. 2). The values of specific cooling capacities q0.15 needed for cooling ambient air to ta2 = 15 °С, rational cooling capacities q0.10rat , q0.15rat and q0.20rat for cooling ambient air to ta2 = 10, 15 and 20 °С accordingly, the basic cooling capacity as difference q0.10-15 =q0.10 -q0.15 , needed for cooling air from ta2 = 15 °С to ta2 = 10 °С, available residual boost cooling capacities q0.b10-15 and q0.b10-20 are calculated for climatic conditions in Voznesensk, Nikolaev region in July 2017 (Fig. 2).
As Fig. 2 shows, with cooling the ambient air to ta2 = 15 °C the fluctuations of the current thermal loads q0.15 are gradual, that points to significant amount of an excessive cooling capacity in the temperate daily hours. At the same time, when air is cooled from ta2 = 15 °C to ta2 = 10 °C, the fluctuations in the thermal load q0.10-15 = q0.10 -q0.15 are comparatively small. Thus, the temperature of cooled air ta2 = 15 °C can be assumed as the threshold temperature for shearing the overall design thermal load on the TIAC system q0.10rat into a comparatively 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 basic stable part q0.10-15 = q0.10 -q0.15 of a design cooling capacity q0.10rat = 35 kJ/kg (Fig. 1). Accordingly, the remaining part of q0.10rat is used for precooling the ambient air to the threshold temperature ta2 = 15 °C and determined as boost cooling capacity q0.b10-15rat = 35 -q0.10-15 (Fig. 2,b). The unstable q0.15 thermal load range can be covered by ACh as well as the stable q0.10-15 thermal load range -by ECh (Fig.3 Fig. 3. A hybrid two-stage TIAC system in AECh: ACHT and ACLT -high and low temperature stages of air cooler; Exp. Valve -expansion valve. As Fig. 2, a shows, the available boost cooling capacity q0.b10-15rat generally covers current thermal loads q0.15 for precooling ambient air to ta2 = 15 °C. Furthermore, even less available boost cooling capacity q0.b10-20rat also covers the current loads q0.15 except quite short periods of daylight hours (Fig.2, b).
As Fig. 4, a testifies, the available boost cooling capacity q0.b10-20 designed for cooling air to ta2 = 20 °С in general case is enough to cover the current cooling demands q0.15 for deeper cooling air to ta2 = 15 °С. The current deficit of design cooling capacity q0.20def =q0.15 -q0.b10-20 can be covered through using the daily accumulated excessive refrigeration energy. This statement is also approved by the continuously arising curve of the summarized excess of available design refrigeration energy over its current deficit Σ(q0.20 τ )exc = Σ(q0.b10-20 -q0.15 )τ .
As Fig. 4, b shows, rational designing of TIAC systems provides decrease of installed cooling capacities of the chillers and TIAC systems in the whole by the values of Δq0. 10,15,20 , i.e. by 15 to 20 % compared with their maximum magnitudes q0.10,15,20max , calculated according conventional practice of designing. The rational distribution of the installed cooling capacity of ACh enables to reduce a design boost load by the value Δq0.15-20 = q0.15rat -q0.20rat (Fig. 4, b), i.e. practically twice as compared with q0.15rat .
In temperate climatic conditions the application of rationally designed hybride two-stage TIAC systems with combined AECh enables to provide about 50% higher annual fuel saving ∑B10rat at q0.10rat as compared with ∑B15rat at q0.10rat for ACh (Fig. 4,b) and can be considered as a novel prosperous trend in TIAC.

Conclusions
A novel trend in TIAC by two-stage air cooling in combined AECh is proposed to provide about 50% higher annual fuel saving in temperate climatic conditions. An advanced methodology is developed to determine a rational design cooling capacities of TIAC systems that provides closed to maximum annual fuel saving and decrease of installed cooling capacity by 15 to 20 % as compared with conventional TIAC designing practice.
A novel approach to designing the TIAC systems through rational distribution of the overall design cooling capacity between unstable ("turbulent") thermal load range for ambient air precooling in the boost high temperature stage of the air cooler by ACh and a stable ("laminarized") load range for further air subcooling to the target temperature in the low temperature stage by ECh. Such two-range distribution of the overall cooling capacity provides the favorable thermal loading conditions for operation of ECh at practically stable loads and realization of the advantages of ACh and ECh in combined AECh (high COP and deep air cooling).
Because of great uncertainty of unstable boost ("turbulent") load range magnitude its design value is determined by a remaining principle as a difference between the overall design cooling capacity for cooling the ambient air to the target temperature and its basic stable ("laminar") load range for subcooling air from a threshold temperature of the ambient air precooled in ACh. Issuing from daily fluctuations of excces and deficit of cooling capacity a new hypothesis to divide a design thermal load of unstable boost range in two load parts with daily accumulated excess of available design cooling capacity to cover daily deficit and to reduce a design boost load practically twice as result.