Effects of a Dual-Loop Exhaust Gas Recirculation System and Variable Nozzle Turbine Control on the Operating Parameters of an Automotive Diesel Engine

Abstract: Reduction of NOX emissions and fuel consumption are the main topics in engine development, forcing the adoption of complex techniques and components, whose interactions have to be clearly understood for proper and reliable operations and management of the whole system. The investigation presented in this paper aimed at the development of integrated control strategies of turbocharging, high pressure (HP) and low pressure (LP) exhaust gas recirculation (EGR) systems for better NOX emissions and fuel consumption, while analyzing their reciprocal influence and the resulting variations of engine quantities. The study was based on an extended experimental program in three part load engine operating conditions. In the paper a comparison of the behavior of the main engine sub-systems (intake and exhaust circuits, turbocharger turbine and compressor, HP and LP EGR loops) in a wide range of operating modes is presented and discussed, considering open and closed loop approaches for variable nozzle turbine (VNT) control, and showing how these affect engine performance and emissions. The potential of significant decrease in NOX emissions through the integration of HP and LP EGR was confirmed, while a proper VNT management allowed for improved fuel consumption level, if an open loop control scheme is followed. At higher engine speed and load, further actions have to be applied to compensate for observed soot emissions increase.


Introduction
Among the different available technologies to limit NO X emissions in internal combustion engines (ICEs), exhaust gas recirculation (EGR) has a long and consolidated history, with investigations started about fifty years ago, both on gasoline [1] and diesel engines [2].Since then, a number of evolution steps were applied to EGR systems, through the fitting of electronic control and exhaust gas cooling, in order to enhance their management while achieving higher effectiveness.In gasoline engines, EGR can also reduce knock occurrence at high load, thus limiting mixture enrichment and improving fuel consumption.In recent years, further developments were considered, starting from the adoption of low pressure (LP) circuit [3,4], as an alternative or coupled to the conventional high pressure (HP) loop [5][6][7].Other aspects are related to the integration of EGR circuits with the main engine sub-assemblies, such as fuel injection equipment [8,9] and turbocharging systems [10], either in a single-stage configuration fitted with variable nozzle turbine (VNT) [11] or according to two-stage concepts [12].Moreover, other available cleaning options have to be matched to EGR, such as Miller cycle [13] or advanced combustion processes [14,15], extending its use to the whole range of ICEs, such as in marine and stationary applications.Finally, the growing interest in alternative fuels, is leading to further issues to be deepened [16][17][18].
Current engines are therefore very complex systems, whose performance, fuel consumption, emissions are strongly influenced by each component's behavior and where the number of control variables is continuously increasing.Investigations dealing with the development of suitable management strategies to optimize engine outcomes are therefore mandatory in order to balance and achieve different goals, taking into account the growing worldwide legislation pressure on thermal and chemical emissions.
As broadly discussed in a previous paper [19], most of the available experimental results presented in the literature are focused on large displacement engines [6,20,21].On the other hand, investigations on dual loop EGR systems were developed mainly through numerical simulations [22][23][24], limiting test bench activities to model validation.Consequently, little experimental information on the application of hybrid EGR systems in current downsized automotive diesel engines is available.Moreover, as the integration of VNT and EGR systems is approached as in [11], the requirements and problems of control system components and algorithms have increased, or the proper turbocharger matching issues, as in [20].Therefore, the study presented in this paper aimed at enlarging the knowledge on the reciprocal interactions of turbocharging, HP and LP EGR systems when different control strategies are applied to achieve lower NO X emissions and fuel consumption.The first goal required the optimization of HP and LP contribution to the overall EGR rate, while the second target demanded the exploitation of the potential offered by the VNT management, thus making experimental planning and results analysis more complex.
Tests were made on a Euro 5 downsized diesel engine (FCA, Turin, Italy), selecting three operating conditions belonging to the engine working area related to the Type Approval cycle.A wide number of engine operating parameters were measured, together with fuel consumption, NO X and soot emissions.
The influence of dual-loop EGR and turbocharging systems control on intake, exhaust and turbocharger quantities is analyzed and discussed, while showing the prevailing effects of VNT open loop control scheme on fuel consumption, allowing to increase VNT opening, and proving that a larger contribution from the LP loop is requested to reduce NO X .

Engine Test Facility
The experimental campaign was made on an automotive Euro 5 DI diesel engine, fitted on a steady state eddy current dynamometer test bench.Table 1 lists engine main technical characteristics, while Table 2 presents measured parameters, together with the relevant instrumentation.For each instrument, range and accuracy are also shown.Pressure and temperature were acquired in selected sections of the engine intake and exhaust, turbocharging and EGR systems.Moreover, a proper control of coolant, lubricant and charge air temperatures was applied through thermostatic circuits or managing water flow rate to obtain a suitable intercooler efficiency level.
Other measurements were related to the displacement S of the nozzle ring push rod, for which a linear potentiometer was used.Comparing S with its maximum and minimum value (S MAX and S MIN ), the VNT opening degree A VNT was evaluated through the equation: ranging from 0% to 100%, when varying the flow area from minimum to maximum.For low frequency measuring process in steady-state conditions, an automatic data acquisition system was used, including a multichannel scanner, an IEEE bus, a GPIB interface and a personal computer.In-house developed codes in Labview ® and Excel ® environment allowed to collect and process data, calculating statistical parameters of measured quantities (average, standard deviation, variation coefficient) and the main engine operating parameters.In-cylinder pressure diagrams were acquired through a dedicated system, sampling different signals at high frequency rates.More details on the measurement equipment used in this study are presented in [19].

Engine Management System
The engine management system was based on an open electronic control unit (ECU), fitted with an EPROM emulator module (Magneti Marelli, Bologna, Italy), and an ETAS ® MAC2F interface connecting the ETK module (ETAS, Stuttgart, Germany) to a dedicated PC.The INCA (ETAS, Stuttgart, Germany) software was used to show engine operating parameters, to select maps stored in the ECU and to change control variables in real time according to the experimental program.
In particular, HP EGR system was managed through a closed loop scheme, comparing the relative air-fuel ratio (AFR) set point with the actual value, given by the measured air mass flow rate and the calculated fuel mass flow rate.The duty-cycle (DC) of the electric HP EGR valve (DC EGR ) was then modulated according to the resulting difference, thus adapting its opening degree.
Turbocharger turbine was controlled through open or closed loop strategies, selected by the ECU according to the engine operating condition.The first was applied by setting proper duty-cycle values of an electro-pneumatic valve (DC VNT ), thus obtaining fixed values of A VNT .As the resulting pressure level is below the atmospheric pressure, the maximum duty-cycle (DC VNT = 100%) corresponded to the fully closed position (A VNT = 0).The comparison of an intake pressure set point with the measured level allowed to apply the closed loop control scheme, resulting in proper changes of DC VNT to reduce the calculated difference.

High and Low Pressure Exhaust Gas Recirculation Circuits
Figure 1 shows engine scheme, displaying the standard HP EGR system and the prototype LP EGR circuit, added for the investigations on the integrated control of EGR and turbocharging systems according to a widely adopted solution [20,21].A selection of engine operating parameters is also presented in the figure, referring to those discussed in Section 3. A diesel oxidation catalyst and particulate filter were not fitted on the engine during the tests, to evaluate the influence of the hybrid EGR system on raw soot emissions, even if compressor rotor and blades were more stressed by particles.
Energies 2017, 10, 47 4 of 18 the fully closed position (AVNT = 0).The comparison of an intake pressure set point with the measured level allowed to apply the closed loop control scheme, resulting in proper changes of DCVNT to reduce the calculated difference.

High and Low Pressure Exhaust Gas Recirculation Circuits
Figure 1 shows engine scheme, displaying the standard HP EGR system and the prototype LP EGR circuit, added for the investigations on the integrated control of EGR and turbocharging systems according to a widely adopted solution [20,21].A selection of engine operating parameters is also presented in the figure, referring to those discussed in Section 3. A diesel oxidation catalyst and particulate filter were not fitted on the engine during the tests, to evaluate the influence of the hybrid EGR system on raw soot emissions, even if compressor rotor and blades were more stressed by particles.On the other hand, engine management and tests development were simplified, due to the absence of the required periodic filter regeneration phases.In order to reproduce suitable pressure levels along the exhaust circuit and in the LP EGR loop, a throttle valve (TV) was fitted at its terminal (Figure 1), whose settings will be analyzed when discussing Table 3.The LP EGR valve was manually controlled through a vacuum signal, while the EGR mass flow rate in the long route loop depended also on the opening position of the above mentioned TV.
EGR rate (fEGR) is expressed as the ratio between recirculated gas and total engine mass flow rate (Equation ( 2)): Following a common approach [21,25], fEGR was evaluated through measurements of ambient, intake and exhaust carbon dioxide concentrations, allowing to apply Equation (3): Estimated levels of MEGR corresponded to the total mass of recirculated gases when both circuits were used or to the mass flowing in the activated circuit when only one was operated.In the first case, the contribution from each loop was calculated through a simplified energy balance between the intake air and the recirculated gases from the two circuits, considering measured temperature On the other hand, engine management and tests development were simplified, due to the absence of the required periodic filter regeneration phases.In order to reproduce suitable pressure levels along the exhaust circuit and in the LP EGR loop, a throttle valve (TV) was fitted at its terminal (Figure 1), whose settings will be analyzed when discussing Table 3.The LP EGR valve was manually controlled through a vacuum signal, while the EGR mass flow rate in the long route loop depended also on the opening position of the above mentioned TV.
EGR rate (f EGR ) is expressed as the ratio between recirculated gas and total engine mass flow rate (Equation (2)): Following a common approach [21,25], f EGR was evaluated through measurements of ambient, intake and exhaust carbon dioxide concentrations, allowing to apply Equation (3): Estimated levels of M EGR corresponded to the total mass of recirculated gases when both circuits were used or to the mass flowing in the activated circuit when only one was operated.In the first case, the contribution from each loop was calculated through a simplified energy balance between the intake air and the recirculated gases from the two circuits, considering measured temperature levels for each flow, neglecting heat losses and variations of constant pressure specific heat with temperature.
It is well known that one of the consequences of EGR technique application is the reduction of oxygen concentration at the engine intake (X O 2 i ), representing the dilution effect [1,2,18].Therefore, the maximum EGR rate, which can be applied without unacceptable penalties in soot emission and in engine performance, depends on AFR corresponding to the requested engine load, being higher at low brake mean effective pressure (bmep) values.At the same time, oxygen content in the exhaust/recirculated gases (X O 2 e ) depends on the same parameter.Useful relationship to link intake and exhaust oxygen concentrations, EGR level and engine load were demonstrated in [25], where a charge dilution index (CDI) was also introduced according to Equation (4): to normalize the effect of load on EGR and to allow the comparison of EGR rate in different engine operating conditions [25].
In this investigation, intake and exhaust oxygen concentrations and CDI values were estimated according to the measured levels of air and EGR mass flow rates and to the calculation of in-cylinder relative AFR following the relationship presented in [25,26].

Testing Procedure
Selection of engine operating conditions was referred to low and medium levels of bmep and rotational speed n, as they are prevailing when dealing with type approval and real world operations in the automotive field.Moreover, different interactions between EGR and turbocharging systems were outlined in these conditions, allowing to analyze a wide range of relationship between the engine operating quantities.The first column of Table 3 presents the identification number of each selected working point, together with the relevant values of n and bmep.Considering the mass and gearbox ratios of cars belonging to the B and C-segment equipped with the tested engine, conditions can be referred to the new European driving cycle (NEDC).In particular, points No. 1 and 2 represent the average operations related to speed levels between 30 and 90 km/h, while point No. 3 is related to the extra-urban acceleration from 100 to 120 km/h.Tests aimed at the assessment of the potential in NO X reduction resulting from the application of a LP EGR system, used either simultaneously with the standard short route circuit or as a stand-alone option.To compensate the negative influence of long route loop on fuel consumption, VNT control was added in order to reduce engine pressure gradient and pumping losses.
To achieve these objectives, four different control variables were considered in all the working points.In operating conditions No. 1 and 2, each set of test was defined according to a fixed value of relative AFR and to the position of the throttle valve at the end of the exhaust circuit.For every experimental set, the HP EGR circuit was used in the first operating mode.Then, LP EGR loop was activated, gradually increasing the relevant valve opening degree.In these modes, A VNT was maintained at its standard value by fixing the corresponding duty-cycle in the ECU map.At the maximum LP EGR valve opening level, the VNT control was added, increasing A VNT .
In operating condition No. 3, turbine was controlled by the ECU according to the closed loop scheme based on the intake pressure set point.Therefore, experimental sets were identified through fixed levels of three control variables (i.e., relative AFR, exhaust TV position and intake pressure).The first mode was again related to recirculation from the short route loop only, while a growing contribution from the LP circuit was then added by managing the relevant valve.
Table 3 summarizes the schedule of the experimental campaign, showing selected values of relative AFR, valve settings for exhaust throttle and, for condition No. 3, intake pressure.In the first case, standard levels stored in the ECU represented the reference condition when exhaust gas were recirculated only through the HP circuit.As regards the different settings of the exhaust throttle valve, position 1 corresponded to the pressure level at the turbine exit when a regenerated diesel particulate filter (DPF) was considered.More closed settings were actuated in positions 2 and 3, leading to an increase in recirculated gases from long route loop, while position 0 corresponded to a more opened setting.

Results and Discussion
Results are presented and discussed in this section considering selected parameters, which outline the most significant interactions between EGR and turbocharging systems.Graphs are built as a function of LP proportion, which is the ratio between LP and total mass flow rates of recirculated gases.Based on the definition of each test set, the analysis of every curves has to be made starting from the left hand side (M LPEGR /M EGR = 0, HP EGR loop activated).Moving towards the right side, contribution from the LP loop grows, until the HP EGR circuit is deactivated (M LPEGR /M EGR = 1).To obtain this operating mode, suitable A VNT control was applied in points No. 1 and 2 [19], while just increasing LP contribution in point No. 3 forced the ECU to close the HP EGR valve.In this testing condition, it was also possible to obtain different modes characterized by the null level of HP proportion, by changing A VNT value until a threshold limit of exhaust smoke (i.e., 3 FSN) was reached.
Each graph shows results for selected sets of the three operating conditions, in order to highlight differences in observed trends.For points No. 1 and 2, continuous lines are associated to standard levels of VNT opening degree, while dotted lines are referred to A VNT control with increased values.In each figure, reference values are shown, corresponding to observed levels when applying the standard HP EGR setting fixed by the engine manufacturer.
Discussion is developed considering the behavior of engine quantities referred to the intake circuit (Section 3.1), EGR circuits (Section 3.2), turbocharger (Section 3.3) and exhaust circuit (Section 3.4).Finally, trade-offs between fuel consumption, NO X and soot emissions are analyzed (Section 3.5).

Intake Circuit
Levels of compressor mass flow rate M c are shown in Figure 2.This quantity is given by the sum of air and LP EGR rate mass flow rates.Taking into account that the first parameter (not shown) is constant in operating conditions No. 1 and 2 for the standard level of A VNT , while presented slight reduction in point No. 3, trends show the increasing contribution from the LP circuit when opening the relevant valve.Variations are apparent when changing relative AFR (in particular for operating modes with HP EGR only).When higher values of A VNT were applied in points No. 1 and 2, reductions of both contributions were observed.Due to the different VNT control in 2500 × 8 condition, a rising trend is observed in the whole range of LP proportion.Figure 3 presents the behavior of intake mass flow rate Mi, given by the sum of air and recirculated gases from both the EGR loops.This parameter can also be evaluated as the sum of compressor and HP EGR mass flow rates.Changes between points No. 1 and 2 on one hand and point No. 3 on the other are shown, always due to the alternative VNT control schemes.In the first case, the increase proves that LP EGR overcomes the reduction in HP EGR for constant AVNT values, while in the second, observed levels are almost constant, as variations in contribution from the two EGR circuits are balanced.Figure 3 presents the behavior of intake mass flow rate M i , given by the sum of air and recirculated gases from both the EGR loops.This parameter can also be evaluated as the sum of compressor and HP EGR mass flow rates.Changes between points No. 1 and 2 on one hand and point No. 3 on the other are shown, always due to the alternative VNT control schemes.In the first case, the increase proves that LP EGR overcomes the reduction in HP EGR for constant A VNT values, while in the second, observed levels are almost constant, as variations in contribution from the two EGR circuits are balanced.Figure 3 presents the behavior of intake mass flow rate Mi, given by the sum of air and recirculated gases from both the EGR loops.This parameter can also be evaluated as the sum of compressor and HP EGR mass flow rates.Changes between points No. 1 and 2 on one hand and point No. 3 on the other are shown, always due to the alternative VNT control schemes.In the first case, the increase proves that LP EGR overcomes the reduction in HP EGR for constant AVNT values, while in the second, observed levels are almost constant, as variations in contribution from the two EGR circuits are balanced.Opening the VNT leads to a reduction in intake mass flow rates, but with different effects to LP proportion, which may decrease (one set of point No. 1), increase (second set of point No. 1, all sets of point No. 2) or remain at level = 1.0 (point No. 3).Measured values of intake pressure p i are displayed in Figure 4.The constant trend in 2500 × 8 condition is obviously due to the VNT closed loop control scheme.In points No. 1 and 2, the activation of LP loop and the growth of its contribution leads to an increase in this quantity, outlining one of the effects of the different interaction between this circuit and the turbocharging system, which is more remarkable at intermediate level of n and bmep.When controlling VNT opening degree, a significant reduction in p i is obviously observed, with final levels going below the reference values for the adopted A VNT settings.
Energies 2017, 10, 47 8 of 18 bmep.When controlling VNT opening degree, a significant reduction in pi is obviously observed, with final levels going below the reference values for the adopted AVNT settings.When opening AVNT, the strong decrease in intake pressure overcame the influence of T1, leading to reduced T2 levels.In point No. 3, as boost pressure is kept constant, only the temperature increase at the compressor inlet influenced T2, resulting in a lower growth.
Following the engine intake circuit (Figure 1), after the compressor two main components affects charge flow, modifying mass rate, pressure and temperature, i.e., the intercooler and the HP EGR valve.Values of intake temperature are depicted in Figure 6, showing the beneficial influence of LP EGR [7].When its contribution increased (in particular up to a level around 70%), a significant reduction in ti was observed, due to the cooling effect of intercooler on this share of recirculated gas.bmep.When controlling VNT opening degree, a significant reduction in pi is obviously observed, with final levels going below the reference values for the adopted AVNT settings.When opening AVNT, the strong decrease in intake pressure overcame the influence of T1, leading to reduced T2 levels.In point No. 3, as boost pressure is kept constant, only the temperature increase at the compressor inlet influenced T2, resulting in a lower growth.
Following the engine intake circuit (Figure 1), after the compressor two main components affects charge flow, modifying mass rate, pressure and temperature, i.e., the intercooler and the HP EGR valve.Values of intake temperature are depicted in Figure 6, showing the beneficial influence of LP EGR [7].When its contribution increased (in particular up to a level around 70%), a significant reduction in ti was observed, due to the cooling effect of intercooler on this share of recirculated gas.When opening A VNT , the strong decrease in intake pressure overcame the influence of T 1 , leading to reduced T 2 levels.In point No. 3, as boost pressure is kept constant, only the temperature increase at the compressor inlet influenced T 2 , resulting in a lower growth.
Following the engine intake circuit (Figure 1), after the compressor two main components affects charge flow, modifying mass rate, pressure and temperature, i.e., the intercooler and the HP EGR valve.Values of intake temperature are depicted in Figure 6, showing the beneficial influence of LP EGR [7].When its contribution increased (in particular up to a level around 70%), a significant reduction in t i was observed, due to the cooling effect of intercooler on this share of recirculated gas.

Exhaust Gas Recirculation Circuits
Analysis of the EGR circuits' behavior is a fundamental aspect of the investigation, as they influence engine and turbocharger performance, aside from determining emission levels.Considering the overall operation of the two loops, trends of total EGR mass flow rate are presented in Figure 7. Increasing levels are apparent when adding LP contribution, with similar slope in the different operating conditions, with a few exception (especially for the 2000 × 5 point when applying a high HP EGR rate).When opening VNT, observed values show significant differences according to the experimental points, as MEGR remained quite constant in 1500 × 2, while reducing in 2000 × 5, due to the prevailing effect of HP EGR decrease.For 2500 × 8, the whole curve corresponded to an increase in AVNT, leading MLPEGR and MEGR to rise, while MHPEGR reduced up to its complete exclusion.Non-dimensional parameters are usually adopted to evaluate EGR levels, comparing it to intake charge through the EGR rate.As previously stated (Section 2.3), a Charge Dilution Index was recently introduced in [25], to normalize the effect of load on EGR while comparing EGR rate in different

Exhaust Gas Recirculation Circuits
Analysis of the EGR circuits' behavior is a fundamental aspect of the investigation, as they influence engine and turbocharger performance, aside from determining emission levels.Considering the overall operation of the two loops, trends of total EGR mass flow rate are presented in Figure 7. Increasing levels are apparent when adding LP contribution, with similar slope in the different operating conditions, with a few exception (especially for the 2000 × 5 point when applying a high HP EGR rate).When opening VNT, observed values show significant differences according to the experimental points, as M EGR remained quite constant in 1500 × 2, while reducing in 2000 × 5, due to the prevailing effect of HP EGR decrease.For 2500 × 8, the whole curve corresponded to an increase in A VNT , leading M LPEGR and M EGR to rise, while M HPEGR reduced up to its complete exclusion.

Exhaust Gas Recirculation Circuits
Analysis of the EGR circuits' behavior is a fundamental aspect of the investigation, as they influence engine and turbocharger performance, aside from determining emission levels.Considering the overall operation of the two loops, trends of total EGR mass flow rate are presented in Figure 7. Increasing levels are apparent when adding LP contribution, with similar slope in the different operating conditions, with a few exception (especially for the 2000 × 5 point when applying a high HP EGR rate).When opening VNT, observed values show significant differences according to the experimental points, as MEGR remained quite constant in 1500 × 2, while reducing in 2000 × 5, due to the prevailing effect of HP EGR decrease.For 2500 × 8, the whole curve corresponded to an increase in AVNT, leading MLPEGR and MEGR to rise, while MHPEGR reduced up to its complete exclusion.Non-dimensional parameters are usually adopted to evaluate EGR levels, comparing it to intake charge through the EGR rate.As previously stated (Section 2.3), a Charge Dilution Index was recently introduced in [25], to normalize the effect of load on EGR while comparing EGR rate in different Non-dimensional parameters are usually adopted to evaluate EGR levels, comparing it to intake charge through the EGR rate.As previously stated (Section 2.3), a Charge Dilution Index was recently introduced in [25], to normalize the effect of load on EGR while comparing EGR rate in different engine operating conditions.The relationship between EGR rate and CDI is presented in Figure 8, identifying observed/calculated values according to the experimental points.
Energies 2017, 10, 47 10 of 18 engine operating conditions.The relationship between EGR rate and CDI is presented in Figure 8, identifying observed/calculated values according to the experimental points.Notwithstanding CDI was obtained through the estimation of intake and exhaust oxygen concentrations, a good correlation with EGR rate is shown, which is in line with trend presented in [25].Therefore, this parameter will be further investigated, taking into account its potential to develop general correlation with combustion parameters derived by in-cylinder pressure diagrams processing (pressure derivative and rate of heat release) for modeling purposes.
On the other hand, an extended range for the two parameters is apparent for 1500 × 2, progressively reduced when considering condition No. 2 (2000 × 5) and No. 3 (2500 × 8), as it could be anticipated by the corresponding levels of engine load and relative AFR.Reference values are always at the lower hand side of the intervals, confirming the significant increase in EGR rate allowed by the prototype LP circuit.Taking into account this outcome, it can be concluded that the prototype LP loop present a satisfactory sizing, balancing the request of recirculated flow rate in the lowmedium range of engine speed and load and reproducing proper levels of pressure losses in the circuit.

Turbocharger
Turbocharging systems are key elements to define engine performance and operations, while their control is a fundamental aspect to achieve satisfactory levels of fuel consumption and emissions, while obtaining acceptable engine transient response.
Being strictly related, the first two parameters discussed in this section are engine pressure gradient, whose behavior is represented in Figure 9, and VNT opening degree, shown in Figure 10.
Engine pressure gradient has probably the strongest relations with all the involved sub-systems and components, interacting with HP EGR mass flow rate, engine pumping losses and fuel consumption, compressor and turbine operating points and nozzle turbine control strategies.
At constant AVNT level and when only HP EGR loop is activated (LP proportion = 0), its value depended on HP EGR valve opening, that is, on relative AFR level.Introducing and increasing LP EGR contribution, observed trends depended on VNT control scheme.In open loop (i.e., in points No. 1 and 2), a significant increase of pressure gradient was apparent when rising LP proportion, until the VNT opening compensate for this effect (dotted lines in Figures 9 and 10), leading its values below the reference ones, with positive influence on fuel consumption.
In closed loop (condition No. 3), ECU opened nozzle turbine (Figure 10) when LP EGR circuit was activated, in order to keep the intake pressure at set point level, compensating for its contribution to compressor mass flow rate and modifying compressor working point.Consequently, AVNT shows an increasing trend (Figure 10), leading to slight decrease of exhaust pressure (i.e., turbine inlet pressure).Notwithstanding CDI was obtained through the estimation of intake and exhaust oxygen concentrations, a good correlation with EGR rate is shown, which is in line with trend presented in [25].Therefore, this parameter will be further investigated, taking into account its potential to develop general correlation with combustion parameters derived by in-cylinder pressure diagrams processing (pressure derivative and rate of heat release) for modeling purposes.
On the other hand, an extended range for the two parameters is apparent for 1500 × 2, progressively reduced when considering condition No. 2 (2000 × 5) and No. 3 (2500 × 8), as it could be anticipated by the corresponding levels of engine load and relative AFR.Reference values are always at the lower hand side of the intervals, confirming the significant increase in EGR rate allowed by the prototype LP circuit.Taking into account this outcome, it can be concluded that the prototype LP loop present a satisfactory sizing, balancing the request of recirculated flow rate in the low-medium range of engine speed and load and reproducing proper levels of pressure losses in the circuit.

Turbocharger
Turbocharging systems are key elements to define engine performance and operations, while their control is a fundamental aspect to achieve satisfactory levels of fuel consumption and emissions, while obtaining acceptable engine transient response.
Being strictly related, the first two parameters discussed in this section are engine pressure gradient, whose behavior is represented in Figure 9, and VNT opening degree, shown in Figure 10.
Engine pressure gradient has probably the strongest relations with all the involved sub-systems and components, interacting with HP EGR mass flow rate, engine pumping losses and fuel consumption, compressor and turbine operating points and nozzle turbine control strategies.
At constant A VNT level and when only HP EGR loop is activated (LP proportion = 0), its value depended on HP EGR valve opening, that is, on relative AFR level.Introducing and increasing LP EGR contribution, observed trends depended on VNT control scheme.In open loop (i.e., in points No. 1 and 2), a significant increase of pressure gradient was apparent when rising LP proportion, until the VNT opening compensate for this effect (dotted lines in Figures 9 and 10), leading its values below the reference ones, with positive influence on fuel consumption.
In closed loop (condition No. 3), ECU opened nozzle turbine (Figure 10) when LP EGR circuit was activated, in order to keep the intake pressure at set point level, compensating for its contribution to compressor mass flow rate and modifying compressor working point.Consequently, A VNT shows an increasing trend (Figure 10), leading to slight decrease of exhaust pressure (i.e., turbine inlet pressure).Measured levels of turbocharger rotational speed nTC are presented in Figure 11.Trends confirm the influence of LP EGR contribution and AVNT control in operating points No. 1 and 2. It is interesting to note that levels attained at higher VNT settings are anyway higher than the reference ones, which should help in transient operations starting from these part load conditions.In point No. 3, increase of turbine mass flow rate (which differs from compressor one, Figure 2, only for fuel mass flow rate) overcame VNT opening at reduced values of HP proportion, resulting in a slight rise in turbocharger speed.Measured levels of turbocharger rotational speed nTC are presented in Figure 11.Trends confirm the influence of LP EGR contribution and AVNT control in operating points No. 1 and 2. It is interesting to note that levels attained at higher VNT settings are anyway higher than the reference ones, which should help in transient operations starting from these part load conditions.In point No. 3, increase of turbine mass flow rate (which differs from compressor one, Figure 2, only for fuel mass flow rate) overcame VNT opening at reduced values of HP proportion, resulting in a slight rise in turbocharger speed.Measured levels of turbocharger rotational speed n TC are presented in Figure 11.Trends confirm the influence of LP EGR contribution and A VNT control in operating points No. 1 and 2. It is interesting to note that levels attained at higher VNT settings are anyway higher than the reference ones, which should help in transient operations starting from these part load conditions.In point No. 3, increase of turbine mass flow rate (which differs from compressor one, Figure 2, only for fuel mass flow rate) overcame VNT opening at reduced values of HP proportion, resulting in a slight rise in turbocharger speed.

Exhaust Circuit
Trends of turbine inlet pressure p3 are shown in Figure 12.In points No. 1 and 2, again LP EGR effect and AVNT control play a major role in determining the observed behavior.In point No. 3, VNT control, jointly with the almost constant trends in Mi, led to a slight reduction in p3.Temperature levels at turbine inlet and exit (T3 and T4) show minor variations in the different operating modes.As an example, values at the turbine exit are presented in Figure 13, showing slight reductions in points No. 2 and 3 when increasing LP contribution and a different influence of VNT opening in conditions 1 and 2. Taking into account the position of the DPF, it can be concluded that DPF regeneration would not be significantly affected by the integrated control of EGR and turbocharging systems.

Exhaust Circuit
Trends of turbine inlet pressure p 3 are shown in Figure 12.In points No. 1 and 2, again LP EGR effect and A VNT control play a major role in determining the observed behavior.In point No. 3, VNT control, jointly with the almost constant trends in M i , led to a slight reduction in p 3 .

Exhaust Circuit
Trends of turbine inlet pressure p3 are shown in Figure 12.In points No. 1 and 2, again LP EGR effect and AVNT control play a major role in determining the observed behavior.In point No. 3, VNT control, jointly with the almost constant trends in Mi, led to a slight reduction in p3.Temperature levels at turbine inlet and exit (T3 and T4) show minor variations in the different operating modes.As an example, values at the turbine exit are presented in Figure 13, showing slight reductions in points No. 2 and 3 when increasing LP contribution and a different influence of VNT opening in conditions 1 and 2. Taking into account the position of the DPF, it can be concluded that DPF regeneration would not be significantly affected by the integrated control of EGR and turbocharging systems.Temperature levels at turbine inlet and exit (T 3 and T 4 ) show minor variations in the different operating modes.As an example, values at the turbine exit are presented in Figure 13, showing slight reductions in points No. 2 and 3 when increasing LP contribution and a different influence of VNT opening in conditions 1 and 2. Taking into account the position of the DPF, it can be concluded that DPF regeneration would not be significantly affected by the integrated control of EGR and turbocharging systems.

Fuel Consumption and Emissions Trade-Off
In the previous sections, an extended analysis of the influence of control strategies for both the EGR circuits and the turbocharger on engine operating parameters was developed, in order to outline the potential of LP EGR circuit fitted to a state-of-the-art diesel engine while showing the strong interactions with controlled systems.
As energy consumption and pollution issues are the main driving forces of engine development, it is mandatory to conclude the results discussion summarizing the main achievements in these fields, reporting the main important trade-offs between brake specific fuel consumption (bsfc), NOX emission (bsNOX) and soot emission (bsS).
Figure 14 shown a selection of these curves for fuel consumption and nitrogen oxides emission, confirming the different behavior of some engine quantities in the tested operating conditions.Considering the reference values shown in Figure 14 (Euro 5 settings), significant reduction were achieved for bsNOX, ranging between −51% (point No. 2) and −64% (point No. 3), that is, around the requested improvement from Euro 5 to Euro 6 limits.

Fuel Consumption and Emissions Trade-Off
In the previous sections, an extended analysis of the influence of control strategies for both the EGR circuits and the turbocharger on engine operating parameters was developed, in order to outline the potential of LP EGR circuit fitted to a state-of-the-art diesel engine while showing the strong interactions with controlled systems.
As energy consumption and pollution issues are the main driving forces of engine development, it is mandatory to conclude the results discussion summarizing the main achievements in these fields, reporting the main important trade-offs between brake specific fuel consumption (bsfc), NO X emission (bsNO X ) and soot emission (bsS).
Figure 14 shown a selection of these curves for fuel consumption and nitrogen oxides emission, confirming the different behavior of some engine quantities in the tested operating conditions.

Fuel Consumption and Emissions Trade-Off
In the previous sections, an extended analysis of the influence of control strategies for both the EGR circuits and the turbocharger on engine operating parameters was developed, in order to outline the potential of LP EGR circuit fitted to a state-of-the-art diesel engine while showing the strong interactions with controlled systems.
As energy consumption and pollution issues are the main driving forces of engine development, it is mandatory to conclude the results discussion summarizing the main achievements in these fields, reporting the main important trade-offs between brake specific fuel consumption (bsfc), NOX emission (bsNOX) and soot emission (bsS).
Figure 14 shown a selection of these curves for fuel consumption and nitrogen oxides emission, confirming the different behavior of some engine quantities in the tested operating conditions.Considering the reference values shown in Figure 14 (Euro 5 settings), significant reduction were achieved for bsNOX, ranging between −51% (point No. 2) and −64% (point No. 3), that is, around the requested improvement from Euro 5 to Euro 6 limits.Considering the reference values shown in Figure 14 (Euro 5 settings), significant reduction were achieved for bsNO X , ranging between −51% (point No. 2) and −64% (point No. 3), that is, around the requested improvement from Euro 5 to Euro 6 limits.Table 4 lists the corresponding values for controlled parameters: it is interesting to note the prevailing effect of LP contribution to NO X decrease in the three tested points to achieve the mentioned variations.It has also to be outlined that proposed strategies are referred to steady state operating conditions, as the available test bench doesn't allow to perform dynamic tests.In real world operations, transient corrections will probably be requested [27,28], to take into account different time delays of high and LP EGR circuits, mainly related to their length.Referring to fuel consumption, a significant decrease can be observed in the corresponding modes of operating conditions 1 (−4.6%) and 2 (even if at a lower extent, −2.1%), obtained with A VNT levels around 17% (Table 4), i.e., higher than the reference ones (Figure 10).In the third condition, no variations were attained due to the different VNT control.
The comparison of bsfc trends with those presented in Figure 9 for engine pressure gradient confirms the strict relationship between these parameters, outlining the prevailing effect of pumping losses over every variations on combustion development induced by EGR activation and control.
Soot and NO X emissions trade-off is presented in Figure 15.In this case, penalties in soot are probably acceptable in point No. 1 (+23.3%,referring to the mode with the best NO X reduction) and No. 2 (+48.4%),while in condition No. 3 bsS is more than doubled, requiring further actions to limit this increase, such as a change in fuel injection pressure.The different order of magnitude of measured bsS variations are related to relative AFR levels (Table 3).
Energies 2017, 10, 47 14 of 18 controlled parameters: it is interesting to note the prevailing effect of LP contribution to NOX decrease in the three tested points to achieve the mentioned variations.It has also to be outlined that proposed strategies are referred to steady state operating conditions, as the available test bench doesn't allow to perform dynamic tests.In real world operations, transient corrections will probably be requested [27,28], to take into account different time delays of high and LP EGR circuits, mainly related to their length.Referring to fuel consumption, a significant decrease can be observed in the corresponding modes of operating conditions 1 (−4.6%) and 2 (even if at a lower extent, −2.1%), obtained with AVNT levels around 17% (Table 4), i.e., higher than the reference ones (Figure 10).In the third condition, no variations were attained due to the different VNT control.
The comparison of bsfc trends with those presented in Figure 9 for engine pressure gradient confirms the strict relationship between these parameters, outlining the prevailing effect of pumping losses over every variations on combustion development induced by EGR activation and control.
Soot and NOX emissions trade-off is presented in Figure 15.In this case, penalties in soot are probably acceptable in point No. 1 (+23.3%,referring to the mode with the best NOX reduction) and No. 2 (+48.4%),while in condition No. 3 bsS is more than doubled, requiring further actions to limit this increase, such as a change in fuel injection pressure.The different order of magnitude of measured bsS variations are related to relative AFR levels (Table 3).From a more general point of view, observed trends confirm that it is possible to move trade-off curves towards lower values, while modifying their trend through VNT control.In operating condition No. 3, a more conventional behavior is apparent, while control variables (in particular, engine intake pressure) have to be varied in a more significant way in order to modify the position of trade-off curves.From a more general point of view, observed trends confirm that it is possible to move trade-off curves towards lower values, while modifying their trend through VNT control.In operating condition No. 3, a more conventional behavior is apparent, while control variables (in particular, engine intake pressure) have to be varied in a more significant way in order to modify the position of trade-off curves.
Trends of NO X and soot emissions can be related to measurement and processing of in-cylinder pressure diagrams, referring their values to maximum pressure and heat release, respectively.
Figure 16 shows bsNO X values as a function of maximum in-cylinder pressure.In points No. 1 and 2, observed complex behavior is related to EGR fraction and VNT opening degree, influencing intake pressure and mass flow rate and the development of combustion process.In point No. 3, a more expected trend is apparent, with NO X reduction corresponding to lower levels of p MAX for each set.
Trends of NOX and soot emissions can be related to measurement and processing of in-cylinder pressure diagrams, referring their values to maximum pressure and heat release, respectively.
Figure 16 shows bsNOX values as a function of maximum in-cylinder pressure.In points No. 1 and 2, observed complex behavior is related to EGR fraction and VNT opening degree, influencing intake pressure and mass flow rate and the development of combustion process.In point No. 3, a more expected trend is apparent, with NOX reduction corresponding to lower levels of pMAX for each set.Figure 17 presents bsS values as a function of heat released in premixed mode of combustion referred to the main injection, calculated from heat release curves as defined in [29].It is well known [30], that a lower soot emission is associated to a more intense premixed combustion.Therefore, an inverse relationship is expected between these parameters, as shown in [29] and is apparent in Figure 17 for one test set for each working point.Moreover, the same correlation was verified for all the operating modes of point No. 3, allowing to assess the linear link for a wider range of soot emissions and premixed heat release values, as proved by the high level of determination coefficient.Figure 17 presents bsS values as a function of heat released in premixed mode of combustion referred to the main injection, calculated from heat release curves as defined in [29].It is well known [30], that a lower soot emission is associated to a more intense premixed combustion.Therefore, an inverse relationship is expected between these parameters, as shown in [29] and is apparent in Figure 17 for one test set for each working point.Moreover, the same correlation was verified for all the operating modes of point No. 3, allowing to assess the linear link for a wider range of soot emissions and premixed heat release values, as proved by the high level of determination coefficient.
Trends of NOX and soot emissions can be related to measurement and processing of in-cylinder pressure diagrams, referring their values to maximum pressure and heat release, respectively.
Figure 16 shows bsNOX values as a function of maximum in-cylinder pressure.In points No. 1 and 2, observed complex behavior is related to EGR fraction and VNT opening degree, influencing intake pressure and mass flow rate and the development of combustion process.In point No. 3, a more expected trend is apparent, with NOX reduction corresponding to lower levels of pMAX for each set.Figure 17 presents bsS values as a function of heat released in premixed mode of combustion referred to the main injection, calculated from heat release curves as defined in [29].It is well known [30], that a lower soot emission is associated to a more intense premixed combustion.Therefore, an inverse relationship is expected between these parameters, as shown in [29] and is apparent in Figure 17 for one test set for each working point.Moreover, the same correlation was verified for all the operating modes of point No. 3, allowing to assess the linear link for a wider range of soot emissions and premixed heat release values, as proved by the high level of determination coefficient.

Conclusions
The experimental study on the development of integrated control strategies of turbocharging and high and LP EGR systems in an automotive diesel engine allowed us to study in depth different aspects referring to the reciprocal interactions between the different subsystems and to the influence on engine behavior.Among the achieved outcomes, some major points can be summarized:

•
Results are strongly influenced by the applied VNT management scheme: among the selected operating conditions, open loop control is applied at low levels of speed and bmep (i.e., in 1500 × 2 and 2000 × 5 test point), while boost pressure is closed loop controlled when increasing these parameters (i.e., in 2500 × 8 condition).

•
When activating and increasing LP EGR contribution, intake and exhaust parameters (mass flow rate, temperature and pressure levels) are significantly modified.In particular, the increase of engine pressure gradient and compressor/intake mass flow rate leads engine and turbocharger to work in conditions similar to those without EGR.

•
With VNT open loop control, fuel consumption is strongly influenced by engine pressure gradient, representing an index of pumping losses.Applying proper settings of VNT position, significant benefit can be achieved for this quantity (−4.6% and −2.1% in operating condition No. 1 and 2, respectively).

•
Referring to the standard engine behavior (levels measured in Euro 5 configuration), reduction between 51% and 64% for NO X emission was obtained, through the increase of EGR rate and a prevailing (points No. 1 and 2) or complete (point No. 3) contribution from LP loop.As a consequence, soot emission was increased, but while a reasonable rise was observed in points No. 1 and 2 (with significant benefits obtained through VNT opening), excessive penalties affected point No. 3, therefore requiring the adoption of further measures to compensate for this negative aspect.

•
EGR technique can be enhanced through the integration of HP and LP circuits.As the second option has different interactions with turbocharger compressor and turbine, a wide potential is offered by the integrated control of these systems, not only to improve fuel consumption and NO X emissions, but also to help engine in transient conditions.

Energies 2017, 10 , 47 7 of 18 Figure 2 .
Figure 2. Compressor mass flow rate as a function of LP proportion in different operating modes.

Figure 3 .Figure 2 .
Figure 3. Engine intake mass flow rate as a function of LP proportion in different operating modes.Opening the VNT leads to a reduction in intake mass flow rates, but with different effects to LP proportion, which may decrease (one set of point No. 1), increase (second set of point No. 1, all sets of point No. 2) or remain at level = 1.0 (point No. 3).Measured values of intake pressure pi are displayed in Figure 4.The constant trend in 2500 × 8 condition is obviously due to the VNT closed loop control scheme.In points No. 1 and 2, the activation of LP loop and the growth of its contribution leads to an increase in this quantity, outlining one of the effects of the different interaction between this circuit and the turbocharging system, which is more remarkable at intermediate level of n and

Energies 2017, 10 , 47 7 of 18 Figure 2 .
Figure 2. Compressor mass flow rate as a function of LP proportion in different operating modes.

Figure 3 .Figure 3 .
Figure 3. Engine intake mass flow rate as a function of LP proportion in different operating modes.Opening the VNT leads to a reduction in intake mass flow rates, but with different effects to LP proportion, which may decrease (one set of point No. 1), increase (second set of point No. 1, all sets of point No. 2) or remain at level = 1.0 (point No. 3).Measured values of intake pressure pi are displayed in Figure 4.The constant trend in 2500 × 8 condition is obviously due to the VNT closed loop control scheme.In points No. 1 and 2, the activation of LP loop and the growth of its contribution leads to an increase in this quantity, outlining one of the effects of the different interaction between this circuit and the turbocharging system, which is more remarkable at intermediate level of n and

Figure 4 .
Figure 4. Engine intake pressure as a function of LP proportion in different operating modes.Values of compressor exit temperature T2 are presented in Figure5.Major variations are observed for experimental points No. 1 and 2, as compressor exit temperature is influenced by the rise in compressor inlet temperature T1 (as a consequence of LP EGR activation and ranging between 5 and 20 °C) and pi.

Figure 5 .
Figure 5. Compressor exit temperature as a function of LP proportion in different operating modes.

Figure 4 .
Figure 4. Engine intake pressure as a function of LP proportion in different operating modes.

Figure 4 .
Figure 4. Engine intake pressure as a function of LP proportion in different operating modes.Values of compressor exit temperature T2 are presented in Figure5.Major variations are observed for experimental points No. 1 and 2, as compressor exit temperature is influenced by the rise in compressor inlet temperature T1 (as a consequence of LP EGR activation and ranging between 5 and 20 °C) and pi.

Figure 5 .
Figure 5. Compressor exit temperature as a function of LP proportion in different operating modes.

Figure 5 .
Figure 5. Compressor exit temperature as a function of LP proportion in different operating modes.

Energies 2017, 10 , 47 9 of 18 Figure 6 .
Figure 6.Engine intake temperature as a function of LP proportion in different operating modes.

Figure 7 .
Figure 7.Total EGR mass flow rate as a function of LP proportion in different operating modes.

Figure 6 .
Figure 6.Engine intake temperature as a function of LP proportion in different operating modes.

Energies 2017, 10 , 47 9 of 18 Figure 6 .
Figure 6.Engine intake temperature as a function of LP proportion in different operating modes.

Figure 7 .
Figure 7.Total EGR mass flow rate as a function of LP proportion in different operating modes.

Figure 7 .
Figure 7.Total EGR mass flow rate as a function of LP proportion in different operating modes.

Figure 8 .
Figure 8. Relationship between EGR rate and charge dilution index in tested operating modes.

Figure 8 .
Figure 8. Relationship between EGR rate and charge dilution index in tested operating modes.

Energies 2017, 10 , 47 11 of 18 Figure 9 .
Figure 9. Engine pressure gradient as a function of LP proportion in different operating modes.

Figure 10 .
Figure 10.VNT opening degree as a function of LP proportion in different operating modes.

Figure 9 . 18 Figure 9 .
Figure 9. Engine pressure gradient as a function of LP proportion in different operating modes.

Figure 10 .
Figure 10.VNT opening degree as a function of LP proportion in different operating modes.

Figure 10 .
Figure 10.VNT opening degree as a function of LP proportion in different operating modes.

Figure 11 .
Figure 11.Turbocharger rotational speed as a function of LP proportion in different operating modes.

Figure 12 .
Figure 12.Turbine inlet pressure as a function of LP proportion in different operating modes.

Figure 11 .
Figure 11.Turbocharger rotational speed as a function of LP proportion in different operating modes.

Figure 11 .
Figure 11.Turbocharger rotational speed as a function of LP proportion in different operating modes.

Figure 12 .
Figure 12.Turbine inlet pressure as a function of LP proportion in different operating modes.

Figure 12 .
Figure 12.Turbine inlet pressure as a function of LP proportion in different operating modes.

Figure 13 .
Figure 13.Turbine exit temperature as a function of LP proportion in different operating modes.

Figure 14 .
Figure 14.Trade-off between brake specific NOX emissions and fuel consumption in different operating modes.

Figure 13 .
Figure 13.Turbine exit temperature as a function of LP proportion in different operating modes.

Energies 2017, 10 , 47 13 of 18 Figure 13 .
Figure 13.Turbine exit temperature as a function of LP proportion in different operating modes.

Figure 14 .
Figure 14.Trade-off between brake specific NOX emissions and fuel consumption in different operating modes.

Figure 14 .
Figure 14.Trade-off between brake specific NO X emissions and fuel consumption in different operating modes.

Figure 15 .
Figure 15.Trade-off between brake specific NOX and soot emissions in different operating modes.

Figure 15 .
Figure 15.Trade-off between brake specific NO X and soot emissions in different operating modes.

Figure 16 .
Figure 16.Relationship between maximum in-cylinder pressure and brake specific NOX emissions in different operating modes.

Figure 17 .
Figure 17.Relationship between heat released in premixed mode and brake specific soot emissions in different operating modes.

Figure 16 .
Figure 16.Relationship between maximum in-cylinder pressure and brake specific NO X emissions in different operating modes.

Figure 16 .
Figure 16.Relationship between maximum in-cylinder pressure and brake specific NOX emissions in different operating modes.

Figure 17 .
Figure 17.Relationship between heat released in premixed mode and brake specific soot emissions in different operating modes.

Figure 17 .
Figure 17.Relationship between heat released in premixed mode and brake specific soot emissions in different operating modes.

Table 1 .
Characteristics of tested engine.EGR: exhaust gas recirculation; and HP: high pressure.

Table 2 .
Measured parameters and instrumentation.FSN: filter smoke number.
Table 4 lists the corresponding values for Table 4 lists the corresponding values for

Table 4 .
Optimal control strategies for NO X and fuel consumption reduction in tested steady state operating conditions.

Table 4 .
Optimal control strategies for NOX and fuel consumption reduction in tested steady state operating conditions.