Design of Waste Heat Recovery System based on ORC for a Locomotive Gas Turbine

This is an excerpt from a technical paper, presented at the Asian Congress on Gas Turbines (ACGT) and written by Abdul Nassar, Nishit Mehta, Oleksii Rudenko, Leonid Moroz, and Gaurav Giri. Follow the link at the end of the post to read the full study!


Gas turbines find applications in aerospace, marine, power generation and many other fields. Recently there has been a renewed interest in gas turbines for locomotives. (Herbst et al., 2003) Though gas turbines were first used in locomotives in 1950 – 1960’s, the rising fuel cost made them uneconomical for commercial operation and almost all of them were taken out of service. The diesel locomotives gained popularity and presently locomotives are operated by diesel engines and electric motors. The emission levels in diesel locomotives have raised concerns among the environmentalists, leading to stringent emission norms in recent years. One of the solutions to reduce emission for these locomotives is to switch to LNG fuel which requires huge investment in upgrading the engines to operate with LNG. The other alternative is Gas Turbine based locomotives and this has gained renewed interest with RZD and Sinara Group of Russia successfully operating LNG based Gas Turbine-electric locomotives. Fig. 1 shows the GT1-001 freight GTEL from Russia, introduced in 2007. It runs on liquefied natural gas and has a maximum power output of 8,300 kW (11,100 hp). Presently, this locomotive holds the Guinness record for being the largest gas turbine electric locomotive (Source: Though there have been a lot of improvements in gas turbines, the thermal efficiency is still not very high unless the exhaust heat is efficiently utilized by a bottoming cycle.

Fig. 1 Russian GT1_001 gas turbine locomotive

Converting the gas turbine into a combined cycle unit, with a bottoming steam cycle, is employed in case of several land-based and marine applications; however, such an option is not practical in a locomotive gas turbine due to the requirements of steam generators, steam turbines and other auxiliaries. The next best alternatives are to utilize either an organic Rankine cycle (ORC) or a supercritical carbon dioxide cycle (sCO2) to extract heat from the exhaust of the gas turbine and convert it into useable energy in the bottoming cycle (Rudenko et al., 2015; Moroz et al., 2015a; Moroz et al., 2015b; Nassar et al., 2014; Moroz et al., 2014). Supercritical carbon dioxide cycles, operating in a closed-loop Brayton cycle, are still in research phase. There is not much practical experience in deploying an sCO2 unit for propulsion gas turbines even though there is considerable research currently in progress. Hence, the obvious choice is to incorporate an ORC based system which is compact, modular and easy to operate. The same concept can also be implemented in any gas turbine application, be it a land-based, power generation, or marine application.

As the first step in this study, a gas turbine cycle was designed with a target power output of 9 MW to be used as Gas Turbine Electric Locomotive (GTEL). The base cycle was designed using a simple open cycle configuration which was subsequently modified to incorporate a recuperator to increase the overall thermal efficiency of the cycle. It is well known that using intercoolers between compression stages can result in reduced work for compressor thus resulting in improved cycle efficiency. At the same time, it requires additional work for extracting heat from the compressed air. (Cohen et al., 1996). In addition to the use of intercoolers and recuperators, the authors in the present work propose the use of a waste heat recovery system (WHRS) with the locomotive gas turbine to increase the power output as well as the overall efficiency of the system. The waste heat available in this cycle is at two locations, one at the exhaust and the other between the compressors (intercooler). Here, an ORC based system was proposed as the bottoming cycle which extracted heat from the compressed air and also from the exhaust after recuperation. Multivariable parametric studies were performed to optimize the pressure ratios of the compressors on either side of the intercooler. The operating pressures and temperatures of the ORC system were also determined by such studies. To keep the overall target power output to about 9 MW, the gas turbine was resized based on the additional power that could be generated in the bottoming cycle.


The conceptual thermodynamic cycle for the basic gas turbine unit based on the Brayton cycle was developed using commercial software AxCYCLE™. The software provides flexibility in building the conceptual cycle, by using appropriate available components, as well as in choosing the parameters to be assigned as fixed constraints in the boundary conditions without the need for any geometric information. The boundary conditions, used for the present design, are shown in Table 1. The cycle layout is shown in Fig. 2a. The overall efficiency of the cycle is 34.5%, which can be considered as idealistic, since the effect of cooling in turbines, as well as inlet and outlet pressure losses were not considered.

Table1: Boundary conditions and parameters assigned for the components
Table1: Boundary conditions and parameters assigned for the components

The cycle was then modified to incorporate the recuperator. In the modified cycle, the compressed air from the compressor passes through a recuperator and gets heated by the exhaust gases from the
turbine. This resulted in efficiency of the cycle increasing to 35.56% (Fig.2b). The boundary conditions and component efficiencies were all retained same as in the open cycle. A pinch type recuperator model was considered with temperature difference of 10 degrees at pinch point. Pressure losses in the recuperator were specified as 0.1 bar on the hot side and twice that value on the cold side. The value of heat recovery coefficient was estimated to be 0.9. It was found that the use of recuperator reduces the fuel flow to 0.522 kg/s compared to that of 0.538 kg/s in the base cycle, for the same power output, resulting in increase of cycle efficiency with recuperator.

After the addition of the recuperator in the base cycle, it was found that the exhaust temperature after recuperator is 4830C. Since the compressor outlet temperature itself is high, not much heat is transferred in the recuperator to the compressed air. If the compressed air temperature at the outlet of compressor is reduced, retaining the same compressor pressure ratio, the possibility of higher heat transfer in the recuperator is possible which can possibly improve the cycle efficiency further. This is precisely what the addition of an intercooler does with an added benefit of lower work input to the compressor after intercooling. Thus, two opportunities opened up to increase the power output as well as the efficiency in the cycle. The first was to modify the base cycle, by addition of an intercooler, to allow more heat transfer in the recuperator resulting in lower exhaust temperature; and the other was to utilize a part of the remaining heat from the recuperator exhaust in a waste heat recovery system. Both modifications were systematically applied in combination in this work. As the next step, the intercooler was combined into a waste heat recovery system to make full use of the waste heat from all the sources

Fig. 2: Gas turbine cycle without (a) and with (b) recuperator for 9 MW power output
Fig. 2: Gas turbine cycle without (a) and with (b) recuperator for 9 MW power output

First, to reduce the overall compression work and the temperature of the compressed air, an intercooler was added and the compression was divided in two parts with the temperature of air after intercooling fixed to 1000C. To utilize energy from the recuperator exhaust, an ORC system consisting of a turbine, an air-cooled condenser, a pump and a heat exchanger was added (Fig. 3). Pressure drops on the hot and cold sides of the heat exchanger were assumed to be 0.1 bar and 0.2 bar, respectively. The fluid used in the ORC cycle was R245fa based on previous experience of Moroz et al. (2013) in working with this fluid. The ORC system was designed for a maximum working pressure of 32 bar and the turbine inlet temperature was 2000C. Since our objective was to maintain the total power at about 9 MW, the power output from the gas turbine was iteratively reduced to about 8.21 MW. At this rating of the gas turbine, the ORC system produced about 0.839 MW to give the total net power desired. The overall system efficiency had increased to 45.6%

Gas turbine cycle with an added intercooler, recuperator
Fig. 3: Gas turbine cycle with an added intercooler, recuperator and ORC cycle

After solving the cycle of Fig. 3, it was found that considerable amount of heat was transferred through the intercooler at a fairly high temperature. This heat is wasted in the form of temperature rise in the cooling circuit. Also, some additional work is utilized to run the cooling flow to the intercooler (not simulated here) which is a penalty to the total output from the system. Here, instead of using a cooling circuit to remove heat from the compression stages, this heat was directed to the ORC system such that it could utilize heat from both the exhaust as well as the intercooler. Fig. 4 shows the thermodynamic cycle of the gas turbine unit with such a configuration for the ORC system.

the exhaust and from the intercooler
Fig. 4: Gas turbine cycle with an ORC cycle extracting heat from the exhaust and from the intercooler

As shown in Fig. 5, a parametric study was performed to obtain the ideal values of pressure ratios for the two compressors as well as the output power of the gas turbine cycle. The feature of generating solutions from multiple variables using Low Discrepancy Sequence (LDS) model which is built inside AxCYCLE™ was used to perform this study, wherein the pressure ratios for both compressors and the generator power from the gas turbine cycle were chosen as variables for the study. Net power production and cycle thermal efficiency were the chosen objective functions. Two additional parameters were also assessed during the study, viz. turbine outlet temperature in GT cycle and ORC mass flow rate. The pressure ratios were varied between 2 to 7 for both compressors and the gas turbine electrical power output was allowed to change between 7.5 to 7.9 MW. 400 combinations of variable values were chosen for calculation by a quasi-random Sobol sequence, within the specified variable ranges and the calculated efficiency are shown in the design space (Fig. 5a).

Fig. 5: Selection of compressor pressure ratios based on
Fig. 5: Selection of compressor pressure ratios based on thermal efficiency (color contours) and required net power

The calculated results show a wide range of output parameters. A refinement was done by filtering out the solutions which gave net power production lower than 9 MW and higher than 9.05 MW. An additional filter was applied to the calculated values of the turbine outlet temperature in gas turbine system. Solutions with the turbine outlet temperatures lower than 5000C and higher than 6000C were also filtered out. After the refinement, all but 9 out of the 400 solutions were filtered out (Fig. 5b). Out of the remaining 9 combinations of the variable values, the one that gave the best thermal efficiency was chosen for further study. This particular design had the first compressor pressure ratio as 5.36, the second compressor pressure ratio as 3.95 and a gas turbine electric power production of 7.89 MW. The cycle in Fig. 4 was recalculated with these values of the three variables. The calculated efficiency, in this case, was 47.94% with the compressed air temperature at the inlet of the intercooler as 229.560C.


The efficiencies and component performances at this stage of cycle design were based on assumed values and previous experiences. However, to get realistic values of the performance, preliminary design needed to be performed and this was carried out using the commercial software AxSTREAM® as it allows designing the complete turbomachinery from conceptual stage with very few inputs. The preliminary design module of AxSTREAM®, based on an inverse task solver, generates thousands of designs allowing to choose the optimal yet feasible solution by considering multiple objective parameters, including manufacturability in a short span of time.

Fig. 6: Preliminary designs of compressor modules and comparison of various designs

Here, both compressors were designed simultaneously, considering them as two different modules and taking in to account the pressure drop and temperature change in between the two modules due to the presence of the intercooler. Fig. 6 shows the comparison of different geometries created for both compressors in the preliminary design module. The best design consisted of a total of 11 stages (6 stages in the first compressor module and 5 in the second) with total-total efficiencies of 89.85% and 84.37% for the modules, respectively. The compressors generated a combined pressure ratio of 20.79 with a mass flow rate of 20.65 kg/s at an optimal speed of 14300 rpm. Various designs generated during this phase were further evaluated, and additional constraints were specified to filter the solutions and to choose the optimal flow path that met both geometric and performance criteria.

Fig. 7: Flow path of the designed compressor.
Fig. 7: Flow path of the designed compressor.

The selected design was analyzed in the meanline solver and blade twist was incorporated for the first four stages based on the diameter to length ratio (D/L < 10). Streamline calculations were performed after reassigning the pressure loss in the intercooler and maintaining the cooler outlet temperature as per the cycle parameters. After detailed design, the total-to-total efficiencies for both compressors modules were increased to 90.64% and 87.05%, respectively, at the design point. Fig. 7 shows the compressor flow path with the intercooler. The total power consumption in the process of compression is about 8.5 MW out of which the first compressor requires 4.17 MW whereas the second compressor after the intercooler consumes 4.41 MW…..

Read the full paper here

Interested in learning about the tools used in this study? Explore our AxSTREAM Platform here or request a trial here

Leave a Reply

Your email address will not be published. Required fields are marked *