ScienceDirect. Highly flexible hot gas generation system for turbocharger testing

Transcription

1 Available online at ScienceDirect Energy Procedia 45 ( 2014 ) th Conference of the Italian hermal Machines Engineering Association, AI2013 Highly flexible hot gas generation system for turbocharger testing Rodolfo Bontemo a, Massimo Cardone a, Marcello Manna a *, Giovanni Vorraro a a Università degli Studi di Naoli Federico II Diartimento di Ingegneria Industriale, via Claudio 21, 8015, Naoli, Italy Abstract he aer resents the design and rototying of a gas generator system for turbocharger exerimental testing caable of delivering a wide range of flow rates with adequate thermodynamic characteristics. he system feeds the turbocharger test section with an hot gas stream of rescribed mean and time varying ressure and temerature in order to fully san the oerating domain of the device with controlled accuracy. Comared with the more conventional gas combustor system, it allows for a safer rig oeration ensured by the structural robustness of a four stroke diesel engine and an easier turbine inlet flow control. he latter is achieved by means of an external suercharging station and a modern ICE electronic control unit so that mass flow rate, ressure and temerature values can be set indeendently. he steady comressor and turbine erformance mas can be obtained oerating the rig according to a conventional rocedure, i.e. collecting a set of flow rate ressure ratio data oints ( g, π) for given hot gas roerties. Alternatively using more advanced oerating modes unsteady testing is ossible to reroduce the comlexities characterizing the driving cycles required by the latest Euroean regulations he Authors. Published by by Elsevier Elsevier Ltd. Ltd. Oen access under CC BY-NC-ND license. Selection and eer-review under under resonsibility of AI of AI NAZIONALE. Keywords: turbocharging, hot gas generation system, rig 1. Introduction Latest trends emerging in the internal combustion engine industry have witnessed the renaissance of turbocharging as the technology caable to bring significant changes to several areas of automotive engineering. * Marcello Manna. el.: ; fax: address: marcello.manna@unina.it he Authors. Published by Elsevier Ltd. Oen access under CC BY-NC-ND license. Selection and eer-review under resonsibility of AI NAZIONALE doi: /j.egyro

2 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) Pros of the alication of turbocharger to ICE are well known and, in the downsizing concet, involve a reduction of the brake secific fuel consumtion (BSFC) and of the ollutant emissions. In order to achieve these benefits, an accurate matching is requested, so that reliable oerating mas of both comressor and turbine are very imortant. Unfortunately, turbocharger manufacturers usually offer few data of oor quality to their customers, disregarding, for instance, the otential occurrence of comressor instabilities (rotating stall and surge) or the effects of ulsating flow conditions at turbine inlet. Very often information about the turbine oerating conditions are rovided with a single continuous line with no details about the effects of the turbocharger rotational seed and even efficiency data are tyically not detailed. Sometimes turbine data have been measured in continuous and cold flow rigs, whose flow conditions differ very much from the real ones. Commercial test rigs such as the Austrian AVL or the German FEV [1,2] are both based on the use of a combustion chamber, and with their large flow rate and temerature san offer the ossibility to obtain accurate erformance data, tyically in steady flow conditions. Different solutions have been roosed by other research centers [9,10] where unsteadiness is tyical introduced inserting on the turbine inlet manifold a rotating valve. he Sanish rig of the CM-Motores of the University of Valencia is instead based on the use of an internal combustion engine as hot gas generator, and unsteadiness at turbine inlet is accounted for inserting on the manifold a valve train system of a recirocating ICE [3,4]. Similarly to the reviously discussed rig, the resent facility is also based on the use of an internal combustion engine as hot gas generator, while differences mainly concern the cold gas generator system feeding the ICE. he rig is caable of roducing accurate steady and unsteady data according to a fully automated electronically controlled rocedure. Nomenclature AC alternate current Greek BMEP brake mean effective ressure π ressure ratio BSL best straight line θ dimensionless temerature / ref CGG cold gas generator DC direct current Subscrits DCU data acquisition and rocess control unit a air FS full scale f fuel HGG hot gas generator g gas ICE internal combustion engine in inlet a air mass flow rate out outlet f fuel mass flow rate ref reference g gas mass flow rate a air ressure g gas ressure PID roortional-integrative-derivative controller PLCU art load control unit a air temerature g gas temerature C turbocharger I turbine inlet temerature S test section 2. he rig design targets As already mentioned in the introduction the latest engine develoments heavily relies on the use of turbochargers in nearly all ICE alications both in the roulsion and industrial context. For this reason all manufacturers roduce family of turbochargers, characterized by different sizes and erformance mas. In order to

3 1118 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) maximize the flexibility of a test rig, all of these eculiarities have to be taken into account at the design stage. herefore a reliminary in-deth study of the rig oerating enveloe has to be carefully erformed before carrying out an executive design of the rig. As reviously described, the easiest way to test turbochargers is based on the use of a commercial rig like the Austrian AVL [1] or the German FEV [2]. he core device of these systems is a combustion chamber where hot gas to be sent to the turbocharger are roduced with the hel of a diffusion flame, and ressurized by a searate comressor. he design of the latter tye of rig begins with the definition of the range of the most imortant thermo-fluid dynamic characteristics of the hot gas fed to the test article, that is, the mass flow rate g, the temerature g and the ressure g. Obviously, the maximum thermal ower and fuel mass flow rate f mainly deend on the choice of the ( g, g ) air, and they have to comly with the thermal and mechanical resistance of the materials. Generally the gas flow rate is rovided by a seed controlled comressor, while the hot gas temerature is controlled by different air/fuel settings. While simle to build and rather diffused in the scientific community, this tye of rig has some severe limitations. First of all, the thermo-fluid dynamic characteristics of the hot gas entering the test section are difficult to be selected indeendently. hen, at every oerating condition a minimum gas temerature must be guaranteed to allow for a regular combustion rocess. Moreover, the gas ressure and flow rate are steady in nature and therefore they are far from the tyical oerating conditions of a turbocharger. Finally, from a safety oint of view a combustion chamber requires the adotion of secial and costly measures. For these reasons a different methodology was followed during the early design stages of the resent rig and several objectives were looked for, namely: (. a a ma,, ). m f Gas Generator ( m. g, g, g ) in ( m. g, g, g ) out urbocharger Fig. 1 Concetual hot gas generator layout indeendent choice of the mass flow rate, temerature and ressure of the hot gas, temoral modulation of the above characteristics, comlete control and ease of oeration of the rig in a fully automated fashion. On account of the above requirements it was quickly concluded that the best solution was rovided by an Internal Combustion Engines oerated through a dedicated control strategy. While gasoline engines rovide higher exhaust gas temerature comared to diesel engines, the latter has been referred on the basis of safety consideration as well as lower oerating and maintenance costs. he ga in the turbine inlet temerature (I) can be filled with the adotion of an after-burner to be inserted on the exhaust manifold and taking advantage of the lean exhaust gas mixture. Such a device will enhance the gas generator flexibility offering the ossibility to adjust the I according to the test article requirements, a feature that would be lost adoting a gasoline engine as gas generator. herefore the gas generator device is based on a modern light-duty turbocharged, direct injection diesel engine which has been derived of its turbocharger to reserve the exhaust gas energy content and has been externally suercharged with an ad-hoc designed comression station, ensuring the attainment of the original erformance of the ICE, in terms of brake mean effective ressure (BMEP). Once more to secure the largest ossible rig flexibility, the comression station has been conceived as a truly indeendent (from the ICE) system with an autonomous mechanical driving unit.

4 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) he rig layout he exerimental facility is located in the laboratories of the DII Diartimento di Ingegneria Industriale of the Università di Naoli Federico II. he lab is newly built and comlies with the most actual standards for safety, ollution control and energy saving. he turbocharger test rig can be considered consisting of four main sections, namely a cold gas generator (CGG), a hot gas generator (HGG), a test section (S) and a data acquisition and rocess control unit (DCU) (see Fig. 2). Atlas Coco GA 55 VSD Air ank EPP4 2" Pressure Control Valve Byassed Flow Hot gas generator Diesel Engine 4 cylinder - 2.5L Hot Gas Byass Valve Stack Screw Comressor Control Room Brake Controller Cold gas generator Digital I/O High Seed DAQ Heat Exchanger Electric Brake Borghi&Saveri m Comressor Backressure Control Valve rm Lubrication oil ank m est Section urbine Backressure Control Valve Fig. 2 he test rig general layout 3.1. he cold gas generator As mentioned in 2, the CGG should rovide the ICE with the requested air flow rate at the aroriate boost ressure and temerature levels. hose data were collected through a detailed exerimental camaign carried out on the turbocharged engine, and served as design target for the CGG concetion. he CGG control rocedure will be dealt with in 4. Atlas Coco GA 55 VSD Air ank PID Controlled Pressure Valve Byassed flow o hot gas generator Screw Comressor Heat Exchanger Fig. 3 he cold gas generator layout

5 1120 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) he core of the cold gas generator device consists of two medium size screw comressors, viz. a two stage, intercooled Atlas-Coco GA55VSD and a one stage all-urose secial screw comressor. he Atlas-Coco is of the oil injected tye and disoses of a ost rocessing unit searating the lubricating oil of the comressed air, and of a 2000 liters reservoir used as caacitive unit. At full ower (55 kw) the two intercooled stages can deliver u to 3500 kg/h at a maximum gauge ressure of 12 bar. he station is oerated with a closed loo control system setting the aroriate rm value of the variable seed AC driving motor as to match the target flow rate. Downstream the reservoir a PID controlled ressure regulator is located. his device sets the boost ressure level requested by the HGG and it is able to cover a wide range of oerating conditions with an high frequency resonse. Inut data to the CGG control system come from the HGG control system. he all-urose secial screw comressor is driven by a 37 kw AC electric motor and oerated at variable seed through a frequency converter. It can deliver u to 600 kg/h at maximum ressure of 2.5 bar. Similarly to the Atlas- Coco, the comressor is oerated with a closed loo control strategy in order to match the target ressure and flow rate as requested by the HGG control system. When the hot gas generator is disabled, the test section can be directly fed by the CGG still offering a remarkable enthaly dro to the turbine. In these circumstances, due to the low temeratures reached during exansion, a dehumidifier should be used to revent condensation he hot gas generator As mentioned in 3 the hot gas generation is granted by a direct injection light duty diesel engine which has been derived of its turbocharger to maximize the turbine inlet temerature and ressure. he ICE ower is delivered to a Borghi&Saveri FE 350SA water cooled eddy current brake which has been secifically designed for reliable transient and steady long time oeration. he brake whose maximum seed, torque and ower are rm, 1500 Nm and 295 kw resectively, is equied with a digital controller and ermits accurate, smooth and re-defined braking at constant or variable seed. When oerated at constant torque (res. seed) the accuracy is better than 0.5% FS (±3 rm). he brake cooling water is treated by a 240 kw cooling tower oerating in closed loo, benefitting of a 10 m 3 storage tank. Control Room Brake Controller Digital I/O Personal Comuter Electric Brake Borghi&Saveri From Cold Gas Generator Diesel Engine 4 cylinder - 2.5L o est Section Fig. 4 he hot gas generator layout 3.3. he test section he test section is hysically searated from the HGG to avoid heat transfer interference between the turbocharger and the HGG, and it can accommodate turbochargers of different sizes, the largest dimension being dictated by the HGG characteristics. he C comressor and turbine are inserted in two searate circuits each of which is equied with the necessary instrumentation. he turbine inlet iing connects the HGG exhaust manifold to the turbine inlet flange through a double branch thermally insulated circuit. A 150 liters cylindrical reservoir is installed on one of the two branches in order to dum out the ressure oscillation roduced by the HGG, if desired. A set of automated remotely controlled valves determines the quality and amount of flow the turbine is fed with. More recisely, a by-ass and a backressure valve define the flow rate and the exansion ratio seen by the turbine for each rescribed value of the ICE BMEP. he turbine outlet iing connects the turbine outlet flange to the stack, a 300 mm large, 20 m long insulated stainless steel circular duct equied with an extraction centrifugal fan so that the whole exhaust ie is slightly below atmosheric ressure. On the comressor side, the suction and delivery

6 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) ies are connected to the test cell and to the stack resectively, in an oen loo scheme. A closed loo configuration, tyically adoted to enlarge the comressor oerating enveloe, can easily be imlemented. he turbocharger is lubricated and cooled by means of a closed loo lubrication system. he circuit is sulied with a 10W40 synthetic oil from the discharge side of a volumetric um and it is then cooled before being returned to a 30 liters reservoir. he volumetric flow rate is controlled through a byass system on the basis of the ressure signal collected by means of a Druck PX-600 transducer located on the um delivery flange, and sent to the DCU. Performance of both comressor and turbine are assessed through direct real time measurements in terms of the relevant thermo-fluid dynamic roerties: inlet-outlet temeratures and ressures, mass flow rates and C rotational seed. emeratures on the cold and hot manifolds are measured with either tye K thermocoules or P-100 thermo-resistances with range and accuracy of C (tye K), C (P-100), ±2.8 C (tye K), better than 0.25% FS (P-100), resectively. Mean ressures in a 0-6 kpa range are measured with silicon deformable membrane transducer with an accuracy better than 0.25% FS BSL. wo high-frequency resonse Kulite WC-312M ressure transducers, with a maximum frequency resonse of 72kHz are used for transient analysis. Mean flow rates on the hot and cold sides are measured with two Rosemount 3095MFA ressure based meters with an oerative range of 900kg/h (scalable u to 5000 kg/h with different sensor housing) and accuracy better than 0.8% FS BSL. ime deendent flow rates on the cold side are measured with a Bosch HFM-5 hot film meter with an accuracy better than 3% FS and a frequency resonse of 30Hz. he turbocharger rotational seed is acquired with an eddy current sensor flush mounted on the comressor casing near the imeller inlet shroud, detecting the blade assing frequency corresonding to a maximum rotational seed of rm. All transducers signals are routed into the control room via cables electrically and magnetically insulated to limit the external noise thus maximizing the signal to noise ratio. he hardware and software devices, as well as the brake controller, are managed with a high seed data acquisition system collecting all transducer signal digitally. he rig control arameters are handled with the hel of several dedicated I/O modules such as the National Instruments PCI-6133 Data Acquisition and the cfp-1808 Programmable Automation Controller units, also accounting for safety issues. More recisely, the PCI-6133 and the cfp-1808 deal with the high frequency resonse and the steady state signals, resectively. Finally, all I/O hardware is controlled via a Virtual Instrument home-develoed code running under LabVIEW standard, that ermits to acquire, control and save all the oerative arameters of the rig in a nearly automated manner. 4. he rig control strategy One of the key oint to address when oerating the HGG is related to the suercharging strategy to adot in order to reserve and/or maximize the enthaly dro the turbine can take advantage of. he CGG relying on the two comression systems (the general urose screw comressor and the Atlas Coco GA55VSD) described in 3.1 is used to suercharge the ICE following an ad-hoc designed control strategy run by the DCU. When the general urose screw comressor serves as CGG, the DCU generates a DC signal which is sent to the frequency converter, setting the desired seed of the electrical motor whose maximum ower is 37kW (see Fig. 5). By so doing different oerating conditions of the comressor and therefore different boost ressure levels can be defined. In other words changing the AC motor rm it is always ossible to match exactly the engine suercharging requirements through a single control arameter. Different oerating modes are instead ossible with the Atlas Coco GA55VSD comressor. his requires a more in-deth descrition of the system. he Ga55VSD screw comressor is a two-stage intercooled machine Inverter Electric Motor Screw Comressor Double Pie Heat Exchanger ICE Fig. 5 he general Screw comressor flow chart

7 1122 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) owered by a 55kW variable frequency AC motor, whose delivery flange is connected to a 2000 liters reservoir. he comressed air system control strategy is mainly aimed at defining the system dynamics, i.e. the changes in demand over time in the most efficient way. Here the rimary concern is related to the caability of the system to meet the demand requirements romtly, without running into oerating inefficiencies and/or malfunctioning. herefore, art load oerations are achieved through a high erformance ressure regulator system which is fully controlled electronically by the DCU. he system is based on the Parker EPP4 architecture which is equied with a rogrammable PID. he PID arameters are maniulated with a rorietary software and a USB connection in order to adjust the dynamic resonse of the system. aken collectively the comression station can deliver any mass flow between 0 and 3500 kg/h at any ressure between 1 and 12 bars. A counter flow double-ie heat exchanger is laced downstream the junction of the two comression station and it may be used to cool the comressed air using as cold sink the water flow rate rocessed by the cooling tower. Atlas Coco GA55VSD Inverter Electric Motor Screw Comressor Reservoir Part Load Control Unit Doub le Pie Heat Exchanger ICE Fig. 6 Atlas Coco GA55VSD screw comressor layout 5. Oerating the rig In this section we resent a few relevant results concerning both the CGG and the HGG to demonstrate the otential of the roosed system. Before removing the turbocharger from the ICE, the latter has been characterized carrying out a series of dedicated extensive tests aimed at qualifying the erformance of the turbocharging system, and more recisely of the comressor. he objective is to define the boost ressure level ensured by the turbocharger for each value of the mass flow rate swallowed by the engine. his is best understood through Fig. 7 where the oerative enveloe of the original comressor matched to the ICE is resented. It is worth noting that every (,π) air lying outside the shaded area of Fig. 7 cannot be obtained with the original setu simly because they do not resect the turbocharger-ice matching conditions. Fig. 8 resents the external suercharging conditions offered by the general urose single stage screw comressor relacing the turbocharger comressor on the engine intake side. It can be seen that thanks to the variable seed oeration allowed by the frequency converter, the original oerating ma of the engine, defined by the shaded area, is readily sanned. However, since the (,π) airs on the oerating ma result from the comressor-ice matching, and the controlling element to oerate the comressor at the art-loads is its seed, it is difficult to set a re-determined value of both and π without resorting to a PID controller. his limitation is overcome by the Atlas Coco comressor, which, thanks to PLCU, induces the oerating ma resented in a Fig. 9. he advantages offered by the PLCU system are clearly enlightened by the Figure where new (,π) airs lying outside the shared area are resent. Fig. 7 he original comressor enveloe matched to the ICE Fig. 8 he general urose screw comressor-ice matching

8 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) Fig. 9-he Atlas Coco extended oerating enveloe Fig. 10- he hot gas oerating lane at 30% ICE load Fig. 11- Comressor erformance ma: ressure ratio Fig. 12- Comressor erformance ma: olytroic efficiency On the hot side the state sace is three-dimensional, i.e. the hot gas characteristics are described in terms of (π g, g,θ g ) trilets, the turbine inlet temerature being a arameter deending on the ICE load. Fig. 10 reorts the oerating lane on the hot side obtained with a 30% load; higher ressure ratios for identical flow rates can be obtained increasing the engine load (results not shown). Figures 11 and 12 show a tyical steady state result obtained with the automated data acquisition system, reorting ressure ratio and olytroic efficiency vs. flow rate of the comressor of a medium size turbocharger for automotive alications. he dashed line in Fig. 11 denotes the occurrence of surge as detected by the high frequency ressure transducers and defined once the ressure signals begin to exhibit areciable unsteadiness qualified in terms of its rms values. Fig. 13- he HGG transient maneuver

9 1124 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) Fig. 14- Comressor ressure and rm time histories during maneuver: 45% throttling Fig. 15- Comressor ressure and rm time histories during maneuver: 22% throttling Fig. 16-Closed u view of Fig. 15 Fig. 17-Comressor oerating line during maneuver: red line, 0.1s samling, 22% throttling; blu line, 25s samling, 45% throttling; dashed line enveloe 25s maneuver o demonstrate the otential of the rig in terms of its unsteady caabilities, we resent in Figures the effects of a simle test maneuver carried out on the HGG on the erformance of the comressor. More recisely the ICE rotational seed and the brake torque are simultaneously changed in time as illustrated in Fig. 13 by the black (rm) and blue (throttle) curves. More comlex maneuvers are of course ossible. he resonse of the system in terms of ICE rm as measured by the eddy current encoder, showing a erfect corresondence with the imosed signal, is also reorted on the same figure (red curve). he effects of the maneuver on the comressor erformance are documented in figures 14 and 15 resenting the temoral traces of the suction and delivery ressure signals together with the turbocharger rm for 45% and 22% backressure valve throttling, resectively (100% corresonds to full oening). While for 45% throttling the ressure ratio exhibit a smooth behavior in time, in the 22% throttling case the situation is different and the occurrence of dee surge flow conditions are clearly visible. It is a eriodic henomenon with a neat dominant frequency of about 19 Hz which is best visualized in Fig. 16. In terms of global erformance the effects of the two maneuvers are also reorted in Fig. 17 in the -π lane. While for the 45% throttling the trajectory of the (,π) air is ractically a single-valued curve, in the 22% throttling case, the occurrence of dee surge flow conditions determines hysteresis cycles, whose enveloe is described by the dashed line of fig. 10. herefore quasi-steady aroach would yield reasonable results in the first case only. 6. Conclusions he aer has resented the concetion, design and rototying of a hot gas generation system for turbocharger exerimental testing. hanks to its flexibility, the HGG has been shown caable of delivering a wide range of flow

10 Rodolfo Bontemo et al. / Energy Procedia 45 ( 2014 ) rates with adequate ressure and temerature values allowing for an easy and comlete sanning of the oerating domain of the tested equiment with controlled accuracy. Steady exerimental results of the comressor erformance ma have confirmed the quality of the control system of the air comression station serving as CGG. Unsteady results referring to a time varying maneuver obtained simultaneously changing the quality and quantity of hot gas sent to the tested article demonstrated the otential of the rig reorting accurate data of the comressor characteristic under dee surge flow conditions. Acknowledgements he authors are indebted with Cav. Gianaolo Borghi of Borghi&Saveri for suorting this work with first class equiment. References [1] htts:// [2] htt:// [3] Lujian JM, Bermudez V, Serrano JR, Cervellò C. est bench for turbochargers grous characterization. SAE Paer [4] Galindo J, Serrano JR, Guardiola C, Cervello C. Surge limit definition in a secific test bench for the characterization of automotive turbocharger. Elsevier Exerimental hermal and Fluid Science 30 (2006) [5] Zinner K. Suercharging of internal combustion engines: fundamentals, calculations, examles. Sringer- Verlag; 1978 [6] Watson N, Janota MS. urbocharging the internal combustion engine. Macmillan; 1982 [7] Naundorf D, Bolz H, Mandel M. Design and imlementation of a new generation of turbo charger test benches using hot gas technology. (2001) SAE Paer [8] Stemler E, Lawless P. he design and oeration of a turbocharger test facility designed for transient simulation. (1997) SAE Paer , [9] Rajoo S, Romagnoli A, Martinez-Botas RF Unsteady erformance analysis of a twin-entry variable geometry turbocharger turbine. Energy 38 (2012) [10] Caobianco M, Marelli S. Exerimental analysis of unsteady flow erformance in an automotive turbocharger turbine fitted with a waste-gate valve. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering (2011) 225: 1087