C OVER STORY EMIS SIONS REDUCTION
CO 2 SAVING POTENTIAL OF MORE ACCURATE EXHAUST TEMPERATURE SENSORS In modern diesel engines, regeneration of the particulate filter is controlled by temperature sensors. Watlow, in cooperation with FEV, has used the example of a commercial vehicle diesel engine to study the influence of temperature sensor precision on the vehicle’s fuel consumption and therefore its CO2 emissions in various test cycles.
10
AUTHORS
DR. STEFAN SCHMIDT is Key Account Manager Europe for Diesel Emission Systems at the Watlow GmbH in Kronau (Germany).
DR. JÖRN BULLERT is Technical Manager Europe for Diesel Emission Systems at the Watlow GmbH in Kronau (Germany).
DR. MARKUS SCHÖNEN is Project Manager Commercial, Industrial and Large Engines at the FEV GmbH in Aachen (Germany).
VERENA KOLL is Research Assistant at the Institute for Combustion Engines at RWTH Aachen University (Germany).
FUEL CONSUMPTION STRONGER IN THE FOCUS
One of the major problems of our time is the anthropological climate change. More and more countries try to put a stop on this development by more stringent emission legislation. Thereby especially CO2 emissions, i.e. fuel consumption, gets in the focus. Right at this point it creates a trade-off situation, because the aftertreatment systems need a certain operation temperature which requires fuel combustion. To solve this trade-off, manufacturers need to find new solutions and solution approaches. Usually the focus is put on improved combustion, better thermal management, catalyst coatings, applied sensors and software models. The temperature sensor influence was hardly investigated so far. Watlow, among others a manufacturer of the accurate, active temperature sensor called ExactSense (ES), investigated the influence of temperature sensors with higher accuracy on the aftertreatment system. Besides the influence on fuel consumption, also the single components were investigated. For this purpose a numerical software model was chosen, in which an engine for Stage IV emission limit with its aftertreatment system could be investigated under different offhighway operation profiles. SIMULATION SOF TWARE, ENGINE / AFTERTREATMENT SYSTEM AND INVESTIGATED OPERATION CYCLES
The calculations were performed with the 1D simulation Software called PROcal-EAS, ❶. The software was developed by FEV, who also performed the calcula-
tions. The software allows the simulation of an engine with its aftertreatment system under different load profiles. FEV feeds their investigation results from real aftertreatment systems as well as the engine maps developed in house into the model as basis. In a first step the software calculates the engine raw emissions, which, in a second step, are fed into the EATS (engine aftertreatment system) model. This part performs the calculations for the aftertreatment components and provides results for temperature profiles, emission values and fuel consumption values for all states of operation. The investigations have been done on a state-of-the-art commercial vehicle engine usable for Stage IV, Tier 4 final and Euro VI legislation. The aftertreatment system contains a DOC (diesel oxidation catalyst) a particulate filter (DPF) and a SCR catalyst (selective catalytic reduction). Results shown here focus only on the DOC/DPF part of the aftertreatment system. Consequently this requires the usage of the temperature sensors up- and downstream of the DPF. Heating up of the exhaust system was done with engine management only. Every one of the temperature sensors showed its own characteristic which was fed into the temperature model. This facilitates usage of the individual temperature signals for the calculation. Despite the NRTC (non-road transient cycle) a second operation cycle was investigated. This load profile represents a typical agricultural machine with increased idle speed, which in this case is about 15 to 17 % of the maximum engine speed. Three further engine speeds about 40, 70 and almost 100 % were simulated. At lower engine speeds only
❶ Simulation tool FEV PROcal-EAS 06I2014
Volume 75
11
C OVER STORY EMIS SIONS REDUCTION
Therefore here only the results for the negative accuracy limits of the sensors are shown. Fuel saving potentials shown here represent the minimum savings.
Trigger by soot load DPF load case 1 DPF load case 1
Regeneration start threshold
Trigger by max. time
DPF RESULTS
Soot load
❹ shows the typical soot loading curve Balance point
Regeneration success threshold Rgn. phase
Rgn. phase
Load phase Load phase Time
❷ Scheme of soot loading investigations
INVESTIGATION METHOD DOC/DPF
The engine part of the simulation software calculates the amount of soot on the DPF with the input from the two temperature sensors. Herein the soot loading is considered as well as the soot reduction by passive- or active regeneration [1]. An active regeneration is initiated as soon as the soot load reaches a threshold value. In that case the temperature is raised by engine management (post injection) until soot starts to burn off. If the soot loading reaches the lower threshold value the active regeneration is stopped and the soot loading starts again, ❷. Depending on the system layout it could reach a stable soot loading (balance point) where the amount of new soot accumulated and the effect of passive regeneration balance out. For system safety reasons even if a balance point is established an active regeneration event is forced after every 50 h of operation. The values of regeneration start threshold and regeneration success threshold were set to 75 and 10 %, respectively. TEMPERATURE SENSOR ACCURACIES
Temperature sensor influence was investigated by means of four sensors with different accuracies, ❸. A resis-
12
tance temperature sensor (RTD), the one used to configure the system originally (“standard”), a Watlow active temperature sensor (“ES sensor”) utilising the self-developed non-standard thermocouple and an optimised version of the last sensor which is technically realisable today (“optimised”). Finally a sensor with zero tolerance was simulated to find out the maximum potentials. This sensor (“ideal”) shows the real temperature. The performance and accuracy characteristics of smart thermocouples were published in 2008 and 2011 [2, 3]. In the simulation only the minimum and maximum tolerance limits of the sensors were investigated. The sensors with positive drift underestimated the soot load and overestimated the soot burned off during the active regeneration. This requires a recalibration of the model.
25 20 Temperature deviation [°C]
low loads were required. For higher engine speeds mainly loads over 50 % of the maximum load were applied.
under active regeneration as a function of the sensor accuracies. As described earlier, besides the “ideal” sensor, only the negative accuracy limit sensors are shown. Right after the regeneration starts, the soot load decreases linearly and then converges to the lower threshold which is finally crossed. As expected the “ideal” sensor is the first, followed by the “optimised” sensor, the “ES sensor” and the “standard sensor”. The different times under active regeneration conditions could be directly converted into fuel consumption. If we look at the case where an active regeneration is started after 50 h of operation, the following results are obtained (the saving potentials are all related to the fuel consumption of the “standard sensor” over the whole operation time (50 h) plus the regeneration): The saving potential under NRTC conditions vary from 0.27 (ideal sensor) over 0.21 (optimised sensor) down to 0.17 % (ES sensor). For the customer cycle saving potentials in the range of 0.11 (ideal sensor) to 0.06 % were found. The “optimised” sensor showed 0.07 % fuel saving potential. The lower saving potential in the customer cycle could be explained as follows. The average temperature in this cycle is higher than in the NRTC. Higher temperature means more efficient pas-
15
Standard (+)
10
ES sensor (+)
5
Optimised (+) Ideal
0 -5
0
200
400
600
800
1000
Optimised (-) ES sensor (-)
-10
Standard (-)
-15 -20 -25 Exhaust gas temperature [°C]
❸ Uncertainty of different exhaust gas temperature sensors depending on the absolute temperature
sive regeneration and therefore the balance point is at 40 % soot load. Less fuel and regeneration time is needed to clean a filter loaded with only 40 % which results in less potential saving by the improved sensors. A similar tendency, but with higher values, can be observed if only the active regeneration event is considered, ❺. The “ES sensor”, operated under NRTC conditions, shows a potential saving of 21.8 % of fuel compared to the “standard sensor”. “Optimised-“, respectively “ideal sensor”, show values of 26.3 and 34.45 %.
In the customer cycles savings in the range from 17 to 29.6 % were calculated for the “ES-“ and the “ideal sensor”, respectively. Percentage of fuel could be a quite abstract number. If this is converted this into kg of fuel, ❻, a potential saving (“ES sensor” in NRTC conditions) of 4.1 kg fuel can be found. Maximum saving in an NRTC with an ”ideal sensor” is 6.4 kg fuel. It needs to be mentioned that the savings are reached only by sensor exchange. No software adaption has been done. Savings of 2.2, 2.6 and 3.9 kg fuel were calculated for the
20 18
Standard ES sensor Optimised Ideal
70 Normalised soot load [%]
Normalised soot load [%]
80
60 50
16 Regeneration success
14 12 10 8 6 4 2
40
0 2000
2200
2400
30
2600 2800 Time [s]
3000
3200
3400
“ES-“, the “optimised-” and the “ideal sensor” under customer cycle operation conditions. SUMMARY AND OUTLOOK
The investigation presented above show that temperature sensors could have a noticeable influence on fuel consumption. Besides the fuel saving potential other system components participate on this positive effect as well. For example oil dilution is less and DOC lifetime can be extended. Details on this results are shown in [4]. However, the results shown above are only valid for the system and its layout investigated here. Depending on the engine, aftertreatment layout and operation conditions the results might vary. Investigations on the SCR system described in the setup section have been performed as well, but are discussed in [4]. Potential next steps could be to verify the results obtained here on a real system or to check the impact if the high accurate sensors are used from the beginning of the development.
20 10 0
Regeneration success 0
500
1000
1500
2000
2500
3000
3500
Time [s]
❹ Schematic decrease of soot loading during an active regeneration for the different sensors
REFERENCES [1] Schönen, M.: Echtzeitsimulation katalytischer Dieselpartikelfiltersysteme unter Berücksichtigung der Partikelcharakteristik. Dissertation, RWTH Aachen, 2010 [2] Culbertson, D. P.; Harvey, D. D.; Kovacevich, S.: The Development of Active Thermocouples for Diesel Exhaust Temperature Measurement. SAE paper 2008-01-2492 [3] Culbertson, D. P. et al.: Advances in Exhaust Temperature Sensing and their Applicability for Diesel Emission Diagnostics. SAE On-Board Diagnostic Symposium, 2011 [4] Schmidt, S.; Bullert, J.; Schönen, M. et al.: Contribution of High Accuracy Temperature Sensors Towards Fuel Economy and Robust Calibration. SAE paper 2014-01-1548
THANKS The authors would like to thank Eric Brueckner, Maurice Smeets and Roland Boehner, FEV GmbH, Aachen (Germany), for the support on the simulation work. They further acknowledge Julian Tan, Pete Hermann, Robert Aasand
❺ Fuel saving potential during regeneration event compared to “standard sensor”
Magdi Khair, all Watlow, Richmond (USA), and Christian Winkler, Watlow GmbH, Kronau (Ger-
ES SENSOR
OPTIMISED
IDEAL
❻ Fuel saving
Customer cycle
2.2 kg
2.6 kg
3.9 kg
NRTC
4.1 kg
4.9 kg
6.4 kg
potential of the sensors during an active regeneration
SENSOR TYPE
06I2014
Volume 75
many), for their financial and technical support on the sensor technology and development.
13
