menu

ORegen: GE Waste Heat Recovery

system to reduce CO2 emissions and
increase plant efficiency
BY ANDREA BURRATO / GLOBAL SERVICES, GE OIL & GAS

Abstract
The current energy market is directly influenced by different dynamics that drive future
development. Energy demand is growing and the traditional sources used for satisfying the energy
request are less able to fulfill the growing demand. The "easy oil" era is over, after recent events,
the use of nuclear energy results less attractive. Coal, though still abundant is considered highly
contaminant of the environment, because of its greenhouse emissions. Another market trend is
reducing CO 2 emissions; in some countries carbon taxes have been introduced, or will be
introduced in the near future.

All these trends bring out the need to use energy more efficiently, avoiding waste and finding ways
to increase efficiency of energy production.

One approach to increase efficiency is the waste heat recovery from the plants of energy
Download a pdf of this article conversion. This article will outline a system that recovers waste heat from gas turbine exhaust to
increase the total efficiency. The efficiency of a simple cycle gas turbine varies from 30 to 40% at
the best cases, and it may rise by at least 10 points.

The traditional Rankine Cycle with water/steam has limitations that make difficult its application in systems where the load of the turbine over time is not constant or the variation of the load is limited
or where water is scarce or not available, or where the temperature of the waste heat source is not so high.

This article introduces the Organic Rankine Cycle (ORC) developed by GE, denominated ORegen, a system to recover the waste heat from gas turbines exhaust.

ORegen is a system capable of recovering the waste heat from gas turbines exhaust, operating at variable load from 50 up to 100% and in mechanical drive application, or located in areas where
the water is absent or scarce.

Cycle description
The Organic Rankine Cycle (ORC) is a thermodynamic cycle based on the well known Rankine cycle that uses an organic fluid as working fluid. ORegen is the ORC developed
by GE Oil & Gas.

The cycle includes two closed loops, a first loop used to extract the heat from the waste heat source that uses diathermic oil, and a second loop where the waste heat is
converted into power.

The working fluid is usually a high molecular weight fluid, with a boiling temperature at atmospheric pressure lower than that of water, which allows the exploiting of low
grade waste heat sources, and a positive slope in the saturated vapor curve in the TS diagram, to allow a dry expansion.

Figure 1 Figure 2
ORC cycle T-S diagram
In the cycle the working fluid is pumped, heated, vaporized, superheated in the heat exchangers with diathermic oil, and then expands in a turbine, generating the power
that is converted into electricity by the alternator. The fluid is then condensed closing the circuit.

A regenerative heat exchanger is located after the turbine in order to increase the system efficiency and to reduce the load at the condenser.

The ORegen uses Cyclopentane (C5H10) as working fluid.

The system is capable of producing power from a few MW up to 17 MW with a single unit, and is sized to recover the power from all the GE O&G gas turbines (GT) and can
also be applied to GE Power & Water and other manufacturers GT.

The use of the intermediate oil loop allows recovery of waste heat also from multiple gas turbines that are operated in parallel, as often happen in oil and gas applications. In
this case a single diathermic oil circuit that passes through the exhaust of all units recovers the heat that is then used by the ORegen system.

The configuration selection depends on 2 main parameters, the number of the GT installed at site and the maximum turboexpander power, which is 17 MW.
If only one turbine is installed at site, there are two possible configurations: direct configuration if all the exhaust energy can be exploited in a single ORegen system, or
multi cycle if more ORegen systems are required.

If more than one gas turbine is installed at site, there are 2 possible configurations available: parallel oil when each GT is equipped with a WHRU and parallel gas when the
exhaust ducts of all the GTs are conveyed to a single WHRU.

Another advantage of the use of the intermediate hot oil loop is the plant layout flexibility that gives to installation on existing plant. Only the WHRU has to be located close to
the gas turbine, the rest of the plant can be located, where space is available: hot oil is transported from the WHRU to the heat exchangers with small pipes, 10 to 16 inches
diameter, depending on the size of the plant. This is an important feature, granting flexibility to plants that are often crowded with equipment that cannot be relocated easily.

next: Page 2/
 menu

ORegen: GE Waste Heat Recovery system to reduce CO 2 emissions and increase plant efficiency Download

/previous: Page 1
next: Page 3/

Working fluid selection

The selection of the working fluid is a system design decision with major implications for the performance and the architecture of the plant. While there are many choices
available for working fluids, there are also many constraints on the selection that relate to the thermodynamic properties of the fluid as well as consideration of safety and
potential environmental impact.

Among others, the condenser pressure gives a first criterion. It was chosen to be always above atmospheric pressure in order to seal the process fluid against any
infiltration from the surroundings. Therefore, if ambient air is used as a cooling medium, the boiling temperature at approximately 1 bar should be sufficiently higher than
the ambient temperature. This requirement is satisfied by Cyclopentane, which, at 1 bar, evaporates at about 50C.

Other desired characteristics are: a high molecular weight, because it will reduce the size of the turbine; high critical pressure, because it allows greater expansion.

After eliminating many possibilities in the pre-screening phase, the final three potential fluids selected have been: R-245fa, Thiophene and Cyclopentane. The analysis
highlighted Cyclopentane as the most suitable compromise fluid for this type of application, giving fair or high performance in all categories. Its main thermo-physical
properties (at ambient pressure) are shown in the following figure.

Fluid Property
Boiling point
14.9 84.1 49.3
(affects condenser cost)
Freezing point
-103C -38C -94C
(operability in cold environment)
Molecular weight
134 84.1 70.1
(increase lowers turbine cost)
Velocity of vapor line
Fair Good Fair
(increase lowers condenser cost)

Cyclopentanes main thermo-physical properties at atmospheric pressure are as follows:

Another important characteristic of Cyclopentane (as well as of other hydrocarbons and refrigerants) is the shape of the saturated vapour curve as viewed in a temperature-
entropy diagram. This curve, that for water has a negative slope everywhere, shows a positive slope for portions of the saturationline. This fact, called retrograde behaviour,
has major implications for Rankine cycles. Normal fluids like water, in fact, require considerable superheat in order to avoid excessive moisture at the turbine exhaust.
Retrograde fluids, instead, allow expansion from the saturated vapour line into the superheated region avoiding any moisture during the expansion process. The retrograde
behaviour also allows recuperating thermal energy from the hot vapour at the discharge of the expander, thus increasing the overall cycle efficiency.

To complete the working fluid assessment, an additional analysis evaluating the long-term behaviour of the fluid and the fluid-metal interaction was carried out. The working
fluid long-term behaviour analysis shows a residual risk due to Cyclopentane decomposition behaviour leading to a negligible long-term cycle performance loss of 2% over
10 years.

Intermediate loop fluid selection

The purpose of the intermediate fluid is to transfer heat from the WHRU installed in the GT exhaust duct to the heat exchanger system heating the working fluid. The
selection criteria for the hot oil takes into account the following two main parameters:
1. Minimum pumpability temperature (PIN)
2. Temperature operational range

The operational range is defined as the interval between the PIN and the maximum operating bulk temperature (PAX). The wider the operational range, the more
suitable the fluid is. The intermediate fluid selection was performed considering three different potential fluids:
Boiling point: 49.3C (121F)
Freezing point: -94C (-137F)
Molecular weight: 70.1

Fluid A: PIN= 0C PAX=345C

Fluid B: PIN=-40C PAX=315C
Fluid C: PIN=-35C PAX=330C

Fluid C was selected as the more suitable for this application since it does not need an oil heater and pipe heat tracing in most applications and it is able to withstand
sufficiently high temperature to optimize the cycle efficiency. The main characteristics of the selected intermediate fluid are:

PIN -35C
PAX 330C
Maximum (recommended) 360C
film temp
Auto-ignition temperature 412C
Flash point 120C

Turbomachinery development
The core of the ORegen plant is the Turboexpander, the rotating machine that converts the gas energy in mechanical power. The Turboexpander has been developed by
GE Oil & Gas Advanced Technology team, leveraging the experience of the Rotoflow design.

The ORegen recovery cycle turboexpander generator consists of two radial stage casings
directly coupled with the gearbox to form an integrally geared architecture machine. It is coupled
to a synchronous generator via a dry flexible coupling. The two stages are connected through an
inter-stage pipe.

The expander wheels are mounted on the high-speed shafts coupled by means of a Hirth
serration. The two wheels drive a bull gear rotating at the output speed of the generator (1,500 or
1,800 rpm).

The product requirements and aero design output were flown down to the system and
component levels and a robust design approach was applied to guarantee that the selected
design would meet the requirements at all operating conditions. The DFSS approach (application
of Six-Sigma principles to the design of products and their manufacturing) was used to set up the probabilistic optimization problem (design variables, constraints, and
objective function) and determine the optimal solution of the key performance requirements.

Among the main aspects that were analyzed with a statistical DFSS approach were:
IGV design
Thrust analysis
Selection of seals

The process pressure is controlled by the IGV system. This system consists of a four-bar mechanism that drives the variable nozzle and controls the nozzle throat area. A
high accuracy hydraulic cylinder, regulated by process pressure, guarantees the precise positioning of the IGV.

The design of this system took into account the variation contributed by each individual component due to manufacturing and assembly tolerances and it was proven to be
capable of controlling the nozzle angle with very low uncertainty.

The scope of thrust analysis was to assess the thrust on the gearbox at all operating conditions. The design was required to meet the specification limits of the gearbox
collar and axial thrust. The preliminary thrust assessment based on the analytical method highlighted the sensitivity of the thrust load to the boundary conditions (outer
diameter pressure, position of balance holes, back wheel seal position, etc.). To produce a robust thrust balance design, a cross-functional analysis, with inputs and
verifications from CFD, secondary flows and structural analysis was performed. Uncertainties from each variable (pressure distribution along the wheels, aerodynamic
forces, position and tolerances of seals, etc.) were combined into a Montecarlo analysis to get the overall thrust load variability over different operating conditions.

The sensitivity study led to the optimization of the diameter and position of the balance holes and a suitable position of the back wheel seal diameter to ensure that the
thrust loads on the thrust collars and low speed shaft thrust bearing would be within the recommended values (minimizing losses). The risk of thrust inversion was also
evaluated. The selected design meets the specification limits with a very good safety factor.

The gas sealing system is required to be capable of minimizing process gas losses during operation and avoiding air infiltrations during standstill conditions. Leakage during
operation (escape of process fluid) would contribute to pollution and deteriorate the plant performance by decreasing the inlet enthalpy flux. This leakage may be recovered
by condensation. Leakage at standstill conditions (air infiltrations into the circuit) would deteriorate the plant performance by increasing the condenser pressure.

Both dry gas seals (DGS) and oil seal options were investigated. A standstill ring whose purpose is to seal the machine during standstill conditions was also investigated. This
device is a metallic ring clamped against the rotor by injection of nitrogen once the machine is stopped to prevent both the escape as well as infiltration of gas. The main
configurations assessed are:
Single DGS with standstill ring
Tandem DGS with standstill ring
Double DGS without standstill ring
Oil seal with standstill ring

The double DGS with standstill ring was selected. This solution is the most effective for meeting design requirements for both escape and infiltration issues. The sealing
capability was also evaluated statistically and the probability of leakage was found to be close to zero.

The lube oil system is integrated into the base plate (the oil reservoir is contained within the beams of the base plate). The seal gas and local control system are mounted on
the same structural steel base plate. The lube oil cooler will be mounted on a structural steel frame installed on concrete.

The turboexpander base plate and generator base plate are separated. The double base plate arrangement offers several advantages; a lighter base plate beam can be
used for the generator and to facilitate coupling the machine with generators belonging to different companies. It also facilitates the transportation of the machine from the
fabrication site to the final installation site.
 menu

ORegen: GE Waste Heat Recovery system to reduce CO 2 emissions and increase plant efficiency Download

/previous: Page 2

Secondary flows and thermal assessment

An accurate design of fluid systems and a careful evaluation of the thermal condition of any fluid machine are crucial to achieving satisfactory performance and life targets.
To ensure proper and safe operation of the ORegen recovery cycle expander, special care was devoted to these aspects, and a great deal of effort was spent in this
respect to guarantee the quality of the design of the machine.

The design process took advantage of several advanced design tools: a proprietary flow network solver, YFT was used to simulate the secondary air system; CFD analysis
were performed for some of the main cavities to provide support to the YFT modelling and increase the level of detail of the analyses where needed; full 3D P-thermal
models were developed to carry out the thermal analysis both at steady state and during transients.

YTF simulations were at the core of the entire fluid systems analysis and thermal design activity. The machine architecture entails an almost complete decoupling of the flow
networks of the two expansion stages as well as of their thermal behaviour; interaction occurs only through the inter-stage pipe. As a consequence a dedicated YFT model
was developed for each module.

The interaction in the fluid dynamics was taken into account by setting appropriate boundary conditions (BCs) in the flow network and in the CFD models for each operating
condition. The definition of each set of BCs was based on accurate CFD simulations of the aerodynamic behaviour of the nozzles and wheels. The models were developed
based on specific geometric features (e.g., cavity shape and size, gaps between mating components) and internal GE design practices were intensively used to accurately
model specific critical components (such as abradable seals and brush seals).

Once tuned using CFD and field experience or test data when available YFT models became a very powerful design tool through which the design space could be quickly
explored. Transfer functions were defined to correlate appropriately selected parameters driving the response of the system in terms of a set of dependent variables
relevant to the design of critical aspects (such as leakage and thrust assessment). A transfer function providing the pressure distribution in secondary flow network as a
function of the load was developed. Based on the pressure distribution, the thrust generated by the flow out of the main flow path was calculated. This contribution was
added to the thrust produced by momentum variation in the main channel and a DFSS approach was used to statistically define the overall thrust occurring for each
particular load condition.

The thermal analysis was also carried out using a different model for each stage, the thermal interaction between the two occurring again only through the inter-stage pipe.
This approach also made it possible to keep the size of the model within acceptable levels. The interaction was modelled by creating a simple standalone model of the pipe,
and therefore, the thermal coupling was achieved by a macro-iterative process to match the temperature and the heat flux occurring at each flange. The same approach
was used to model the thermal coupling of each module with the gearbox casing.

Structural assessment
The results of the CFD and thermal analysis were useful for both the structural assessment of single components as well as for understanding the rotor-to-stator
positioning. Under operating conditions this has a vital impact on clearances (and therefore performance and thrust) at hot conditions during the steady state operations of
the machine and during transients. A simplified FEM model of the whole turboexpander was built for this assessment and was used for a preliminary assessment of the
stress of the casings early in the design phase and for the evaluation of the casing-rotor relative displacements. In addition, it allowed the assessment of the forces acting on
any stator structure connection. These interactions were taken into account by setting appropriate BCs in the finer models that were successively built and analysed for each
component.

FEM structural analysis and life assessments performed as per GE internal design practices (including ASME verification of the casings and flanges) demonstrated that the
stress levels are within the life requirements of all components.

The HP and LP wheels are 3-D (three-dimensional formed shape) with 16 airfoils and shrouded with a full cover disk. In order to keep the bulk wheel stress state at
acceptable levels, the wheel tip peripheral speed was kept under control from the very initial phase of the aerodynamic design. During the mechanical design, starting from
the aero flow path, the whole 3D wheel model was designed, including the wheel features such as labyrinth seals and balancing holes. Based on 3D CAD models, as for the
other components, FEM models were prepared and used for thermal-mechanical and modal calculations.

The major limitations for turboexpander wheels are mainly the yielding stress and High Cycle Fatigue (HCF). Those wheels, being uncooled components, are generally not
affected by high temperature gradients, therefore Low Cycle Fatigue (LCF), normally, does not constitute the major life limiting factor. At the same time, thanks to the limited
maximum temperatures reached in continuous operation, even creep is normally not a major limiting factor. Nevertheless LCF and creep assessment were also performed,
as well as oxidation, corrosion and erosion life assessments.

FEM structural analyses run at maximum continuous speed, maximum temporary overshoot speed and wheel over speed test (room temperature) conditions found stresses
at levels well below the life requirements specified in internal GE design practices.

Among the causes that can lead to cracking and failure of a turboexpander wheel, HCF induced failures are treated as a primary design issue. The correct evaluation of the
dynamic behavior of the wheels is invaluable to avoid potential resonance crossings. This is achieved starting from the knowledge of which stimuli are affecting the wheels
and their natural frequencies and modal shapes; in fact, resonances occur when the stimulus matches the frequency and modal shape of the structure. The design was
focused on reducing the number of stimuli to a minimum. Stimuli are caused by flow irregularities imposed on the wheel: the upstream IGV, the diffuser struts if present, etc.,
and their identification is usually straightforward and shows the SAFE (Singh's Advanced Frequency Evaluation) interference diagrams for the HP and LP wheels for the 17
MW machine. Both the stress and modal requirements specified by GE internal design practices are fulfilled with a good safety margin

Case study
ORegen output, when coupled to GE O&G gas turbines in direct configuration, is shown in the following table.

The input data are the gas turbines outlet flow and temperature calculated at ISO conditions.

System
GT power Exhaust flow Exhaust temp GT efficiency ORC output efficiency
GT Model (kW) (kg/sec) (C) (%) (MWe) (%)
PGT25 (*) 23,261 68.9 525 37.7 6.9 48.9
PGT25+ (*) 31,364 84.3 500 41.1 7.9 51.5
PGT25+ G4 (*) 33,973 89.0 510 41.1 8.6 51.5
 MS5001 (*) 26,830 125.2 483 28.4 11.3 40.4
MS5002B (*) 26,100 121.6 491 28.8 10.8 40.7
MS5002C (*) 28,340 124.3 517 28.8 12.4 41.4
MS5002D (*) 32,580 141.4 509 29.4 13.8 41.9
MS6001B (*) 43,530 145.0 544 33.3 15.6 45.2
LM6000PC (**) 43,397 125.6 454 41.7 9.7 51.1

GE Oil & Gas developed the ORC application to recover electrical energy from GT waste heat using, as a case study, a plant where 2 MS5002C are installed in a compression
station, one running at full load and one in stand-by condition. This study shows the capability of the system to produce electricity using the energy that would normally be
released into the atmosphere from the gas turbine.

The starting data for the case study are the environmental conditions and the gas turbine exhaust data.
GT Power 25300 kW
GT heat rate 12814 kJ/kWh
GT efficiency 28.1%
Exhaust temperature 527C
Exhaust flow 116.8 kg/sec
Site elevation 100 m a.s.l.
Ambient temperature 25C
Relative humidity 60%

The solution for this plant is to connect 1 ORegen system to the 2 gas turbines in parallel oil configuration.

Each GT is equipped with its own WHRU, but only one is running.

The power that the system is able to generate is the following:

Gross power 11,500 kW
Net Power 9,800 kW

The difference between the 2 values is the parasitic load of the plant that is due mainly to three items: hot oil pumps, Cyclopentane pumps and air condenser fans. The net
power represents the recovered power that can be effectively used. In the hypothesis that the system runs for 8,500 hours per year, the ORegen system is capable to
produce 83,000 MWh of power per year.

The total plant efficiency is increased by more than 10 points:

Total plant power 35,100 kW
(GT+ ORegen)
GT efficiency 38.9% +10.8 points

These factors make the ORegen system an ideal choice for those applications that are in geographic areas where water is either scarce or difficult to treat. Moreover
ORegen can be installed also on existing plants as a revamping, requiring only the modification of the exhaust duct to install the WHRU. The complete automation of the
ORegen system allows for complete remote operations and monitoring.

References
1. Reducing the CO2 footprint by means of overall cycle efficiency increasing F. Cinelli, A. Miliani , G. Seghi GASTECH 2009, Abu Dhabi, UAE, 2009

2. The ORegen waste heat recovery cycle: reducing the CO2 footprint by means of overall cycle efficiency improvement P. Del Turco, A. Asti, A. Scotti del Greco, A. Bacci, G.
Landi, G. Seghi ASME Turbo Expo 2011, Vancouver, Canada, ASME GT2011