For systems which operate with only a single three-way catalytic converter, the pollutants must be in a state of chemical balance in or- der that the conversion level for all three pollutant constituents is as high as possible. This necessitates a stoichiometric A/F-mix- ture composition with λ = 1.0, which means that the “window” in which the A/F ratio must be located is very narrow. The only so- lution is to apply closed-loop control to the adjustment of the A/F mixture ratio. Open- loop control of fuel metering is not accurate enough.
Direct-injection gasoline engines are run with A/F mixtures which deviate from stoi- chiometric. Closed-loop control can also be used on these systems.
Design and construction
A Lambda oxygen sensor (Fig. 1, Pos. 3a) is located upstream of the pre-cat (4). The sen-
sor signal USais inputted to the engine ECU
(7). In order to do so, either a two-step Lambda sensor (two-step control) or a broad-band Lambda sensor (continuous-ac- tion Lambda control) must be used. A fur- ther Lamda oxygen sensor (3b) can be situ- ated downstream of the main catalytic con- verter (5). This is always a two-step sensor, and it delivers the sensor signal USb. This
form of control is known as two-sensor con- trol.
Operating concept
Using the Lambda control loop, deviations from a specific A/F-ratio can be detected and corrected. The control principle is based on the measurement of the residual oxygen in the exhaust gas. This is a measure for the composition of the A/F mixture supplied to the engine (2).
Two-step control
The sensor voltage USagenerated by the two-
step Lambda oxygen sensor upstream of the pre-cat (4) is high in the rich range (λ < 1) and low in the lean range (λ > 1). Since the sensor voltage jumps abruptly at λ = 1, the two-step Lambda oxygen ensor can only dif- ferentiate between rich and lean A/F mixtures. 82 Catalytic emissions control Lambda control loop
Figure 1 1 Air-mass meter 2 Engine 3a Lambda oxygen
sensor upstream of the pre-cat (two-step Lambda sensor, or broad-band Lambda sensor)
3b Two-step Lambda sensor downstream of the main catalytic converter (only if required; on gasoline direct injection with integral NOxsensor) 4 Pre-cat (three-way
catalytic converter) 5 Main cat (On mani- fold injection: three- way converter; on gasoline direct injec- tion: NOxaccumula- tor-type converter) 6 Injectors 7 Engine ECU 8 Input signals US Sensor voltage UV Injector-triggering voltage
VE Injected fuel quantity
Air Exhaust gas
Fuel VE 3a 3b UV USa 1 2 4 5 6 7 8 USb
Functional diagram of the Lambda closed-loop control
1
æ
UM
K1
The sensor output signal is converted to a binary signal in the engine ECU and used as the input signal for the Lambda closed-loop control as implemented using software. The Lambda control has a direct influence on the A/F mixture formation and sets the correct A/F ratio by adapting the injected fuel quan- tity. The manipulated variable comprises a step change and a ramp, and its control di- rection changes with each jump of the sen- sor voltage. In other words, a jump of the manipulated variable causes the A/F mixture to change. This change is first of all very abrupt, and then it follows a ramp. With a high sensor voltage (“rich” A/F mixture), the manipulated variable adjusts in the “lean” direction, and for a low sensor voltage (“lean” A/F mixture) in the “rich” direction. This so-called two-step control enables A/F mixture to be closed-loop controlled to val- ues around λ = 1.
Shaping the manipulated variable’s char- acteristic curve asymmetrically compensates for the Lambda sensor’s typical false signal caused by variations in A/F mixture forma- tion (rich/lean shift).
Continuous-action Lambda control The broad-band Lambda sensor outputs a continuous voltage signal USa. This means
that not only the Lambda area (rich or lean) can be measured, but also the deviation from λ = 1 so that the Lambda control can react more quickly to an A/F mixture devia- tion. This leads to better control behaviour with highly improved dynamic response. The broad-band Lambda oxygen sensor can measure A/F mixtures which deviate from
λ= 1. This means that (in contrast to the two-step control), such A/F mixtures can also be controlled. The control range covers
λ= 0.7...3.0 so that continuous Lambda control is suitable for the “rich” and “lean” operation of engines with gasoline direct in- jection.
Two-sensor control
When it is situated upstream of the pre-cat, the Lambda oxygen sensor (3a) is heavily stressed by high temperatures and untreated exhaust gas, and this leads to limitations in accuracy. On the other hand, locating the sensor downstream of the main catalytic converter (3b) means that these influences are considerably reduced.
The only problem here though is that a sin- gle downstream sensor would be far too “sluggish” due to the exhaust gases taking so long to reach it. The principle of two-sensor control relies upon the upstream sensor controlling the “lean” and “rich” shift, while the downstream sensor is part of a “slow” corrective closed control loop responsible for additive changes.
Lambda closed-loop control of gasoline di- rect injection
The NOxaccumulator-type catalytic con-
verter has two different functions. During lean-burn operation, NOxaccumulation and
CO oxidation must take place. In addition, at λ = 1, a stable three-way function is needed which provides for a minimum level of oxygen-accumulation.
The Lambda sensor upstream of the cat- alytic converter monitors the stoichiometric composition of the A/F mixture.
Together with the integrated NOxsensor, the
two-step Lambda sensor downstream of the NOxaccumulator converter not only takes
part in the two-sensor control but also mon- itors the behaviour of the combination O2
and NOxaccumulator (detection of the end
of the NOxrelease phase).