See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/314174625

ORC module on a cupola furnace

Conference Paper · June 2014

CITATIONS

0READS

233

3 authors, including:

Some of the authors of this publication are also working on these related projects:

Centrifugal compressor deisgn View project

Robust Design Optimization of Organic Rankine Cycles View project

Benoit Obert

Enertime

10 PUBLICATIONS   26 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Benoit Obert on 02 March 2017.

The user has requested enhancement of the downloaded file.

PROCEEDINGS OF ECOS 2014 - THE 27TH INTERNATIONAL CONFERENCEON

EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS

JUNE15-19, 2014, TURKU, FINLAND

1

ORC module on a cupola furnace

Clément GANTIEZa, Laurent SANCHEZa and Benoit OBERTa

aENERTIME, Puteaux, France, clement.gantiez@enertime.com

Abstract: The Organic Rankine Cycle technology allows combined heat and power (CHP) or single electrical generation from different heat sources (biomass, waste heat, geothermal, solar). For medium output power (typically lower than 1 or 2 MWe) and temperature (available heat source below 300°C), it turns out to be more cost-effective than the traditional steam Rankine cycle.

The present paper focuses on the particular case of a waste heat recovery ORC module installed on a hot blast cupola furnace in a cast iron facility. The module is directly connected to the oil loop (200°C) that was previously used to cool down the smokes of the cupola. The ORC allows reducing the electrical consumption of the former cooling system and to generate up to 1 MWe which corresponds to 30% of the internal plant electricity consumption. The characteristics of the module are described.

The ORC considered in this article is a 1MWe gross system using Solkatherm SES36 as the working fluid. The expander is an axial turbine specifically designed to match the working fluid requirements. The heat exchangers are shell and tubes (evaporator), plate heat exchangers (regenerator) and finned tubes (aero-condenser) types. The electrical generator is asynchronous.

Finally, the article focuses on the possibility to slightly modify the design of an existing ORC machine in order to match the requirements of different applications instead of using a rigid design that would lead to lower effectiveness.

Keywords: ORC, Waste Heat Recovery, CHP, Turbine, Energy, Energy efficiency, Power plant

1. Introduction

Increased environmental awareness among energy intensive industries and a recent increase of energy costs open the door to innovative techniques to produce and consume electricity. Among the available technologies, Organic Rankine Cycles are nowadays one of the best ways to recover low grade heat and generate electricity with no extra fuel consumption [1]. For medium output power (lower than 2MWe) and temperature (available heat source up to 300°C), ORC turn out to be more cost-effective than the conventional steam cycle [2].

This technology uses the same principle as the conventional steam cycle: it is composed of heat exchangers (evaporator and condenser), a pump and an expansion machine (turbine) coupled to a generator. The only difference is the working fluid, an “organic fluid”1 is used in closed loop instead of water. Among other properties, organic fluids are characterized by a boiling temperature different from water (in our case, boiling temperature is lower than water’s boiling temperature), allowing the recovery of heat from sources available at lower temperature [3,4].

Depending on the working fluid selected for the application, it can be useful to use a recuperator between the hot vapour at turbine exhaust and the cold liquid at pump exhaust to improve cycle efficiency (by recovering

1 Several categories of fluids, derived from carbon chemistry, are included in the “Organic fluids” (eg. refrigerants, hydrocarbons, siloxanes, etc.). New fluids are developed every year by chemical companies for different uses such as air conditioning, etc.

2

a part of thermal power before condensation). This has been implemented in the particular case of the ORCHID module (see Fig. 1).

Fig. 1.ORC principle

Nowadays, most of the references are geothermal or biomass applications but since the WHR project on Heidelberg cementery (ORMAT, 1999) [8], ORC on WHR applications are more and more numerous. For example there are ORC project on incinerator (Turboden Mirom Belgium) and on cement industry (Turboden, Ciments du Maroc, Morocco).

2. Fonderie et Mécanique Générale Castelbriantaise (FMGC) situation With its 330 employees (€ 70 Million of annual sales), FMGC (Chateaubriant, Loire-Atlantique, France) produces counterweights and Ship keels. The FMGC is today the European leader in its market field, producing 90,000 tons of cast per year.

The foundry is a metal forming process using used iron and coal as raw materials. The techniques used depend on the dimensions, characteristics and number of parts to be casted. In a typical industrial foundry, cast iron goes through the stages of melting, pouring, cooling, shakeout, finishing and shipping of the finished part.

Cast iron can be produced in a cupola furnace, a shaft furnace of several meters high. The cupola is a counter-current reactor where in a mixture of iron and /or steel is melting.

Mainly two types of cupola are used:

• the cold blast cupola furnace wherein the combustion air is at ambient temperature • the hot blast cupola furnace where combustion air is heated by recovery from cupola exhausts or by

independent heater (mainly gas).

3

Fig. 2.View of FMGC cupola process

The heat recovery system of a hot blast cupola (such as the FMCG) is a heat exchanger with an upper and a lower part. The upper part is used to preheat the combustion air (450°C) cooling the flue gas down from high exhaust temperature (~850°C). In the lower part of the exchanger circulates a thermal heat oil loop (see characteristics in table 1) which lowers the flue gas temperature (below 180°C) to allow treatment in bag filters before exhaust through a chimney.

Cast iron production is a continuous and steady process over the year on FMGC site (24/24 and 5 days a week). Size of FMGC hot blast cupola foundry is considered by CTIF2 as representative of French foundries.

Before the ORCHID project, most of thermal heat recovered by the oil loop was not valorized in the manufacturing process (less than 20% was used for building heating) and simply dissipated in air coolers (dry cooling towers). This meant a low energy efficiency of the foundry (high losses of thermal energy), but also an important electrical consumption (fans of cooling towers).

Table 1.FMGC thermal oil loop characteristics Parameters Unit Value

Oil temperature at heat exchanger (flue gas/oil) inlet °C 110 +/- 5°C Oil temperature at heat exchanger (flue gas/oil) outlet °C 200 +/- 10°C Thermal power dissipated in air coolers kWth 7,000 Annual electrical consumption of air coolers MWh/an 100

3. ORCHID© Module The ORC module, named ORCHID©, has been designed, developed and is supplied by ENERTIME3, a French ORC manufacturer. ORCHID is a module for power generation from waste heat as heat source and able to produce up to 1MWe gross power.

2Centre Technique des Industries de la Fonderie : French Association of cast iron producers 3www.enertime.com

4

Fig. 3.Waste Heat Recovery ORCHID© module at the FMGC foundry

3.1. Fluid selection The choice of the Solkatherm (SES36, hydrofluorocarbon, refrigerant of 3rd generation) [7] as the working fluid was made to reach a cycle efficiency above 16% with the kind of heat source available (200°C), for environmental concerns (low GWP of 1,300 and no ODP4) and for safety reasons (very low toxicity and flammability). In addition to this, Solkatherm allows a limited legal workload under Declaration according to the French ICPE classification5).

Major drawbacks of this fluid are its compatibilities: some common materials used for seals, such as Viton rubber, are prohibited due to potential interaction between the material and the fluid.

3.2. Heat exchangers Three heat exchangers are implemented in ORCHID module:

• An evaporator composed of two shell and tubes heat exchangers in series: the first one pre-heats the fluid and the second one is used for boiling and superheating.

• A recuperator: a shell and plates heat exchanger for cost and compactness reasons

• A condenser: a low finned tubes Air Cooled Condenser (ACC). This equipment - instead of water-cooled condensers and cooling towers – leads to better compactness, lower price and lower electrical consumption.

3.3. Pump A circulating pump operates at a relatively high flow rate and a relatively low pressure head. Given these operating conditions, a displacement pump was not adequate. Therefore, a centrifugal pump was chosen. Another benefit of a centrifugal pump is its low maintenance requirements and costs.

3.4. Turbine The ORCHID turbine was designed [5,6] by ENERTIME with the support of Dynfluid laboratory at ENSAM6 in Paris.

4 GWP : Global Warming Potential. ODP : Ozone Depletion Potential 5ICPE : Classified Installations for Environmental Protection 6Ecole Nationale Supérieure des Arts et Métiers or Arts et Métiers ParisTech

5

This turbine is a multi-stage axial turbine with a robust mechanical design including oil-lubricated rolling bearings and mechanical seals. The turbine stages are in-between the two bearings which allows the rotor’s critical rotational speed (first bending vibration mode) to be well above the nominal rotation speed.

The ORCHID turbine has a subsonic flow and a mean degree of reaction close to 0.5. This high degree of reaction generates a high axial thrust that is compensated by a balancing drum. The turbine is directly coupled to an asynchronous generator. Main parameters of the turbine and generator are:

Table 2.ORCHID Turbine design point characteristics Property Unit Design Point

Pressure ratio - 11.5 Inlet Temperature °C 167 Maximum Electrical Power kWe 1,000 Rotational Speed RPM 3,000 Critical Speed RPM > 4,500 Isentropic Efficiency % >82% Generator Efficiency % 95.5% Generator Frequency Hz 50

4. Implementation In February 2011, ORCHID project was awarded by the 6th TOTAL-ADEME (French Environmental Agency) call for projects for Energy Efficiency in Industry with subsequent financing allowing the

implementation of a first ORCHID© on FMGC site.

The facility is subject to regulation for Classified Installations for Environmental Protection and has been registered under Declaration to French local administration.

Piping length of 150m (total) was implemented from oil exchanger/flue gas to ORC module to feed it (location of the ORC was defined to disturb as less as possible the foundry processes and activities).

Civil works consist in a concrete slab to install 2 skids and the air cooled condenser (total weight is around 60 tons). Concrete slab is also used to collect any possible leaks from ORC module (thermal oil, organic fluid, cooling water).

The module was assembled in a workshop and delivered on site in June 2012, a few weeks after the delivery of the air cooled condenser (due to the size of this equipment, it has to be assembled on site).

Mechanical and electrical connections have been made during July 2012.The electricity generated is fed back into the factory grid to limit administrative constraints, and as a competitive option in absence of a feed-in tariff for electricity from waste

heat in the country at the time.

Commissioning was carried out during September and October 2012 consisting in the definition of programmable logic control

(PLC) parameters and modification of automated start-stop sequences. A supervision panel is installed directly on the platform. A live reporting to the foundry supervision centre is made possible through ethernet network.

Main problems that were identified during this commissioning period are:

• Organic fluid leaks have been detected at the flanges of the installation: o special washers have been installed to prevent loosening flanges caused by vibrations; o a leak detection system has been installed. It is composed of flange belts and a detection

panel: this panel detects leaks of organic fluid and switches the installation off in case.

Fig. 4Assembly of ORCHID air cooled condenser

6

• ORCHID module is stopped every weekend: due to this, air infiltrations have been detected resulting in a decrease of efficiency of the installation. A vacuum system was set up consisting of a compressor and a gas cooler to evacuate parasitic air without losing organic fluid.

Fulltime electrical production started in November 2013 and from this date the module runs as soon as foundry production is sufficient to send enough oil to the module.

5. Performance

5.1. Design performance At design point, the nominal performances of the different parts of the ORC module are listed in Table 3.

The thermal oil leaves the heat exchanger of the ORC module at 140°C before being cooled down to 120°C in another exchanger for building heating. For security reasons (in case of sudden stop of the ORC), Air Coolers still operate at minimum load and cool down the oil loop to 110°C.

Table 3.Nominal parameters and Fig. 5.T-s diagram with hot source and cold source

5.2. On-Site performance

5.2.1. Start up ORCHID© module is fully automated. The start up order is given remotely from the supervision center. Synchronisation between the asynchronous machine and the grid is also fully automated. During start up, a first 30 minute phase is required to heat up the machine (exchangers, piping and turbine) and to speed up the turbine up to synchronisation speed of 3,000 RPM. It can be observed in Fig. 6 that the time from synchronisation to full load is about 60 minutes. Electrical gross power is measured directly at the generator outlet.

7

Fig.6. Start-up characteristics.

5.2.2. ORC Efficiency It is common to present the performance of an ORC by comparing its efficiency with respect to the maximal efficiency that could be reached: the Carnot efficiency.

The integrated electrical efficiency is defined by:

Given the Carnot efficiency:

The Fraction of Carnot is:

It can be seen in Fig. 5 that the FoC efficiency depends highly on the thermal capacity available at the evaporator ( ). For a ratio of 0.6 and at around 62% of thermal inlet capacity, the FoC reaches 35%,

which is good. Data at higher ratio are still not available because the measurements have been made

during summer, when the cupola furnace was not at full load and the outdoor temperature was high. But a FoC up to 45% is expected for ratio of 0.5.

Fig. 7.Fraction of Carnot efficiency.

8

5.2.3. Turbine Performance Maximal Power Point Table 4 shows the maximal power point recorded since the commissioning of the ORC power plant. The maximum electrical power produced by the turbine is lower than the design power because the waste heat available at the FMGC is lower than the nominal heat taken for the design of the plant. This is due to a decrease in production of the foundry and also to the fact that the ORC plant has been slightly oversized to fit with the aim of ENERTIME to develop a 1MWe machine. The measured isentropic efficiency is also slightly lower than the nominal efficiency due to the use of the turbine at partial load is this case.

Table 4.ORCHID Turbine measured performance Property Unit Measured value

Pressure ratio - 6.6 Inlet Temperature °C 164.2 Electrical Power kWe 730 Isentropic Efficiency % 81.4%

5.2.4. Turbine Performance at partial load Figure 8 shows the evolution of isentropic efficiency of the turbine. These values were recorded between April 2013 and September 2013, for turbine inlet temperature varying from 140°C to 172°C and outlet pressure varying from 1.6 bars to 3.1 bars. The evolution of the efficiency with respect to the turbine electrical power remains generally flat; this is a property of multi-stage axial turbines which allows high isentropic efficiency over a broad range of conditions. The measured efficiency is compared to the efficiency computed with the help of CFD simulations and losses models. The CFD analysis was carried out with ANSYS CFX 14.0 using real gas properties for modelling the SES36 fluid [7].

Fig. 8.Measured and computed isentropic efficiency as a function of electrical power

The difference between the predicted efficiency and the measured efficiency can be explained by two types of errors: errors in the CFD calculation (numerical errors, turbulence models and fluid modelling uncertainties...) and losses estimation errors (shaft seal and lubricated bearings mechanical losses, labyrinth seal leakages...).

6. Design modification to CHP conditions The ORCHID turbine was designed to be easily adapted to other applications with low adjustment costs. In the case of CHP (eg. using a biomass boiler), the main parameter change that affects the turbine is the rise in the outlet pressure. The relatively high number of stages of the ORCHID turbine allows a high flexibility to adapt to this requirement. Indeed, by just removing the two last stages of the turbine and extending the axial

9

diffuser, it is possible to fit to the new outlet conditions with little to no losses in isentropic efficiency. Using the same method, the turbine can easily be adapted to cycles with lower temperature heat sources by removing the first stage. The mechanical design and flow path remain mostly unchanged in all these cases, allowing power generation in a large variety of applications.

7. Economics details The total investment is around 2,1M€ including ORC (75%) and integration costs (25%). The net annual production being around 4GWh, the revenues are 330 000€/year. Considering operational costs of 30 000€/year, the payback time rises 6,7 years.

8. Conclusion The ORCHID© module installed on the FMGC foundry shows good efficiency at part load conditions. Even if it has not run in full load condition (1MWe), mostly due to insufficient thermal power available, the results presented are encouraging. The ORCHID© turbine has shown its good capacity to match the requirements for different applications leading to different operating conditions (CHP, WHR) while keeping good efficiency. Further investigations are still necessary to evaluate the availability of the module.

10

Nomenclature Letter symbols: FoC Fraction of Carnot, -

Thermal power, W T Temperature, K

Electrical power, W

Greek symbols η Electrical efficiency

ηcarnot Carnot efficiency

Subscripts and superscripts c Cold

el Electrical

ev Evaporator

h Hot

References [1] Elektrotechnischingenieur, n°117, December 2008

[2] A. Schuster, S. Karellas, E. Kakaras, H. Spliethoff. Energetic and economic investigation of Organic Rankine Cycle applications. Applied Thermal Engineering, 2009, Volume 29, 1809-1817.

[3] S. Quoilin, S. Declaye, B. F. Tchanche, V. Lemort. Thermo-economic optimization of waste heat recovery Organic Rankine Cycles. Applied Thermal Engineering, 2011, Volume 31, 2885-2893.

[4] S. Quoilin, M. Van Den Broek, S. Declaye, P. Dewallef, V. Lemort. Techno-economic survey of Organic Rankine Cycle (ORC) systems. Renewable and Sustainable Energy Reviews, 2013. Volume 22, 168-186.

[5] R. H. Aungier. Turbine Aerodynamics, Axial-Flow and Radial-Inflow Turbine Design and Analysis. ASME Press, 2006.

[6] L. Moroz, Y. Govorushchenko, P. Pagur, A Uniform Approach to Conceptual Design of Axial Turbine/Compressor Flow Path, Future of Gas Turbine Technology, 3rd International Conference, 2006, Brussels, Belgium.

[7] A. P. Fröba, H. Kremer, A. Leipertz, F. Flohr, C. Meurer. Thermophysical Properties of a Refrigerant Mixture of R365mfc (1,1,1,3,3-Pentafluorobutane) and Galden® HT 55 (Perfluoropolyether). International Journal of Thermophysics, April 2007, Volume 28, Issue 2, pp 449-480.

[8] L. Y. Bronicki. Organic Rankine Cycle Power Plant for waste heat recovery. http://www.ormat.com/FileServer/e008778b83b2bdbf3a8033b23928b234.pdf

View publication statsView publication stats