Application Note HFR09
www.cambustion.com
Combustion instability detection during catalyst heating phase of GDI cold start
Introduction
The combustion quality of engines is always important but especially so when the catalyst has not yet
reached sufficient temperature to neutralise engine-out pollutants. Usually, the three-way catalyst
surface temperature required for significant HC reduction is above 300oC and the vehicle’s cold start
calibration is designed to reach these temperatures as quickly as possible.
However, the methods employed to increase the exhaust gas temperature can inadvertently lead to
combustion instability and therefore tailpipe emissions of high HC at a time when the catalyst is not
fully converting them.
Spark retard is a typical technique which provides a late burn and high exhaust gas temperatures.
Multiple fuel injections are also often applied during the catalyst heating phase with the main fuel
injection occurring during the intake stroke (to promote good mixing of the cylinder’s homogeneous
charge) but with secondary injection(s) occurring during the compression stroke to try to aid
combustion stability by providing a rich pocket of gas at the spark plug’s electrodes. The retarded
spark and relatively poor fuel vaporisation are factors which make achieving combustion stability
particularly difficult during this crucial part of the engine’s operation.
Fast FIDs are widely used by calibrators to identify combustion problems and to study the real time
cycle-by-cycle HC emissions during this critical phase. Their help in identifying the spark retard limits
and other factors associated with combustion quality are often used in conjunction with fast CO& CO2
measurements (for fast lambda) as well as cylinder pressure information.
Example: Vehicle and Engine
A 2008 model year Peugeot 308 (Euro 4) with 1.6 litre turbocharged GDI was used for this testing on a
robot-driven chassis dynamometer (Figure 1). The vehicle had both a close-coupled starter catalyst
and a main catalyst packaged together.
Figure 1: Vehicle on chassis dynamometer
Fast response gas analyzers (including the fast FID) were position at the pre-catalyst and post-catalyst
sampling points to provide the engine-out and tailpipe data (Figure 2). A standard bench emissions
analyzer was used to sample the emissions from the dilution tunnel.
Page 1 of 5
Figure 2: Emissions sampling points
The vehicle was driven on the New European Drive Cycle (NEDC) from 20oC ambient conditions and
the results from the cold start and catalyst heating phase were logged.
Page 2 of 5
Results
The cumulative tailpipe HC mass emissions (shown in blue in the upper plot of Figure 3) show the
light-off characteristics of the 3-way catalyst from the cold start, but it is clear that a significant
contribution to the total cycle’s tailpipe HC emissions (which total 1g after 1200s) were associated with
the catalyst heating phase (which, from other observations ends at 68s). The catalyst is largely active by
the start of the 2nd acceleration at 53s but there is evidence of some breakthrough from poor combustion
during the gear change at 57s; a spike appearing in the tailpipe.
0.5
40000
Black - Fast FID engine out
Red - Fast FID tailpipe
Blue - Cumulative Tailpipe (tunnel)
Orange - NEDC drive cycle
35000
0.45
0.4
0.35
25000
0.3
20000
0.25
0.2
15000
HC Emissions (g)
HC Emissions (ppm C)
30000
0.15
10000
0.1
5000
0.05
0
0
0
20
40
60
80
100
7000
70
6500
60
6000
50
5500
40
5000
30
4500
20
Spark Timing
4000
10
3500
0
3000
-10
2500
-20
2000
-30
1500
-40
1000
-50
500
-60
Engine Speed (RPM)
0
-70
0
20
40
60
80
100
Time (s)
Figure 3: Cold start HC measurements and spark timing
Most of the combustion instability occurred during the first gear cruise from 18-27s. The catalyst
heating strategy of this vehicle included double fuel injection events and retarded spark timing as
shown in the lower graph in Figure 3.
Page 3 of 5
Spark Timing (°BTDC)
Engine Speed (RPM)
Time (s)
The data for the cold start and first gear cruise is expanded from 0-30s in Figure 4 in an attempt to
study this more closely.
40000
0.2
Black - Engine Out Emissions
Red - Tailpipe Emissions
Orange - NEDC drive cycle
35000
0.18
0.16
HC Emissions (ppm C)
30000
0.14
25000
0.12
20000
0.1
0.08
15000
0.06
10000
0.04
5000
0.02
0
0
0
5
10
15
20
25
30
Time (s)
200000
50000
180000
45000
160000
40000
140000
35000
120000
30000
100000
25000
Red - Engine-out CO2
Black - Engine-out HC
Blue - Engine RPM
80000
20000
60000
15000
40000
10000
20000
5000
0
[HC] (ppm C1)
[CO2] (ppm)
Figure 4: First 30s of cold start and catalyst heating phase
The data shows the (largely unconverted) HC passing through the catalyst with a few very short
duration single poor combustion events in the engine-out HC between 5 and 10s. There are some high
HC emissions associated with the transient during the first acceleration at 11s but the 1st gear cruise (a
steady state condition where one might expect combustion to be stable), during which the spark timing
has been retarded to approximately 3 degrees after Top Dead Centre, is exhibiting sporadic combustion
problems which are translating to the high tailpipe HC emissions shown in Figure 3.
0
0
5
10
15
20
25
30
Time (s)
Figure 5: Fast CO2 engine-out gas data complementing the combustion instability observations
A fast response NDIR was also used to measure the engine-out CO2 simultaneously with the fast FID
and the data from this is shown in red in Figure 5. The sporadic absence of CO2 indicates that
combustion was either absent or incomplete from those cycles which are producing the high HC levels;
the peaks of engine-out HC aligning precisely with the troughs in the CO2 signal.
Page 4 of 5
On reviewing this data, the calibrator could adjust spark timing, fuel injection timings and quantities
and other parameters to eliminate the combustion instabilities.
As described previously, this data was recorded after the turbo and shows the total engine-out
emissions from all 4 engine cylinders. Further analysis of each cylinder by positioning the fast FID and
NDIR sample probes in each of the individual exhaust ports would identify the precise cycles and
cylinders with poor combustion and, via the NDIR’s fast CO& CO2 data, the combustion lambda could
be calculated.
Conclusion
The fast FID is an effective calibration tool for identifying the limits of spark retard and other engine
control parameters during the cold start and catalyst heating phase of GDI engine operation. This can
help optimise combustion and minimise tailpipe emissions.
The HC signature provides information about the quality of combustion and the precise timing of such
issues. It furthermore provides a direct measurement of the real time emissions of this legislated
pollutant.
Page 5 of 5