GB2552516A

GB2552516A - Method of controlling a fuel injector

Info

Publication number: GB2552516A
Application number: GB1612992.6A
Authority: GB
Inventors: Baur Peter
Original assignee: Delphi Automotive Systems Luxembourg SA
Current assignee: BorgWarner Luxembourg Automotive Systems SA
Priority date: 2016-07-27
Filing date: 2016-07-27
Publication date: 2018-01-31
Anticipated expiration: 2036-07-27
Also published as: GB201612992D0; GB2552516B

Abstract

Disclosed is a method of controlling a fuel injector. A fuel injector includes a valve having an electrically operated actuator, such as a solenoid actuator or piezo-electric actuator, which is used to move a valve needle. The needle is moved away from a valve seat on application of an activation signal so as to open the valve to dispense fuel. The activation signal has a profile which includes at least one actuation pulse which moves and/or holds the needle in a valve open position and at least one subsequent braking pulse adapted to slow the needle on subsequent closing of the valve. The method comprises the steps of detecting whether a valve re-opening event occurs during an operational cycle of the valve and changing an inter-pulse delay between the end of the activation pulse and the start of the braking pulse for subsequent operational cycles if a re-opening event occurs. A re-opening event is determined by integrating the current or voltage signal across the actuator over a time window and determining if re-opening has occurred if the integrated signal is at or below a threshold value. The method prevents unwanted re-opening of a valve during valve closure which arise from pressure fluctuations in the fuel supplied to the injector whilst providing controlled deceleration of the needle during closure of the valve to prevent noise and minimise wear of the valve.

Description

(54) Title of the Invention: Method of controlling a fuel injector Abstract Title: Method of controlling a fuel injector (57) Disclosed is a method of controlling a fuel injector. A fuel injector includes a valve having an electrically operated actuator, such as a solenoid actuator or piezo-electric actuator, which is used to move a valve needle. The needle is moved away from a valve seat on application of an activation signal so as to open the valve to dispense fuel. The activation signal has a profile which includes at least one actuation pulse which moves and/or holds the needle in a valve open position and at least one subsequent braking pulse adapted to slow the needle on subsequent closing of the valve. The method comprises the steps of detecting whether a valve re-opening event occurs during an operational cycle of the valve and changing an inter-pulse delay between the end of the activation pulse and the start of the braking pulse for subsequent operational cycles if a re-opening event occurs. A re-opening event is determined by integrating the current or voltage signal across the actuator over a time window and determining if re-opening has occurred if the integrated signal is at or below a threshold value. The method prevents unwanted reopening of a valve during valve closure which arise from pressure fluctuations in the fuel supplied to the injector whilst providing controlled deceleration of the needle during closure of the valve to prevent noise and minimise wear of the valve.

GENERAL DESCRIPTION OF CLOSING SPEED CONTROL STRATEGIES

CLOSING SPEED SENSING WINDOW ^{6 a}

-------:R PULSE DELAY

3-BRAKING PULSE

5-REOPENING DETECTION WINDOW

500 ' 1000 1500 2000 ° 2500 3000 3500 4000 τΐΜΕ(μ₃) FIG. 1A

/9

GENERAL DESCRIPTION OF CLOSING SPEED CONTROL STRATEGIES

(Λ) ΞΟνίΊΟΛ ΗΟ133ΓΝΙ S (stO S/J_N3IAI3DV1dSIC] 3HNId

500 '1000 1500 2000 ^u 2500 3000 3500 4000

TIME Oxs) FIG. 1 A

2/9

CNG064 PINTLE DISPLACEMENT TRACE EVLUTION OVER LEARNING PHASE (Λ) ΞΟνίΊΟΛ

500 1000 1500 2000 2500 3000 3500 4000 4500 5000

TIME (μ,ε) FIG. 1 B

o	o	o	o	o	o	o	o	o
o	LO	o	LO	o	LO	o	LO
	co	co	CM	CM	τ—	τ—

(sri) _LN3IA30V1dSIO 3LLNId

3/9

OOOCO^|-C\JOOO<O^|-C\JO amvAivoisxHd

diTIVAIVOISAHd

21.5 22 22.5 23 23.5 24 24.5 25 25.5 2( ^{TIME (ms)} FIG. 1 C

4/9 (N o T^- o

LO t

o o

j ~ LO

IMPACT OF UNKNOWN RAIL PRESSURE CHANGE WITH DRIVE SCHEME V16

TIME (με) FIG

o	o	o	o	o	o	o	o	o
o	LO	o	LO	o	LO	o	LO
	co	co	C\J	C\J	t—	i—

(ωτί) INRIABOVIdSia 311Nld

5/9

EVOLUTION OF INTERPULSE DEL7\YVS PRESSURE

6/9

PRESSURE SIGNAL DELAY

TIME (με) FIG (ava) aanssaaa

7/9

PINTLE DISPLACEMENT VS. TIME

^{1 1} γ TJ- C\J CjJ (lut!) ad (s/\) 30V110A (SA) ΞΟΫΠΟΛ adivaodiNi

-2000 -1500 -1000 -500 0 γι 500 1000

TIME AFTER FINAL FT (με) FIG 5c

8/9

9/9

PQ

Method of Controlling a Fuel Injector

Field of the Invention

This disclosure relates to methods of controlling actuation of fuel injectors. It has particular but not exclusive application to method of reducing fuel injector valve reopening,

Background

Solenoid or piezo electric actuated fuel injectors typically are controlled by pulses sent to the actuator of a fuel injector which act to open a fuel injector valve and allow fuel to be dispensed. Such actuators act to displace (via the armature or piezo stack of the actuator) a pintle and needle arrangement, of the valve to move the a needle away from a valve seat. In such a state the valve is open and when the pulse falls there is no power to the actuator and the valve is forced to a closed position e.g. via a spring.

Pulse profiles may vary and may comprise a series of pulses to operate the solenoid. There may be an initial activation pulse, provided in order to start to move the needle away from the valve seat, thereafter the pulse and thus power to the actuator is reduced. After a short while this may be followed by a hold phase where a reduced level of power is applied to keep the valve in the open position.

These pulses may be regarded as fueling pulses. Thereafter the pulse and this voltage is reduced to close the valve. This may be followed by one or more braking pulses which act to slow the movement of pintle and needle when closing.

Various valve closing speed control strategies are known which concentrate on limiting axial stresses on moving parts. In such methodologies, stresses can be controlled (i.e. limited) by detecting armature closing speed, and consequential control of inter-pulse delay (timing between fueling/activation and braking pulses) to fine tune resistive force.

A problem with such strategies is that it has been determined by the inventors that there is a problem of valve reopening, during pintle closure, which is undesirable, and this is in particular caused by fluctuations in the pressure supplied to the fuel injector. Although pressure measurements of the fuel supply (e.g. from a common rail) to the injector can help avoid this problem, there is often a delay between the registered fuel pressure e.g. a the ECU and the actual pressure over sudden pressure change. An object is to overcome such problems. So to reiterate, because there is an inherent delay between the actual pressure (eg. in the common rail) and that registered by an engine ECU at any instantaneous time point, this can lead to valve reopening, i.e. reopening of the valve before it should do (prematurely).

So to recap, due to the high stroke of CNG DI injectors in particular, there is a high risk of excessive injector seating velocities that would lead to reduced injector lifetime and increased noise. The application of a second injector pulse during closing can reduce this seating velocity. In case this second injector pulse is activated too early during injector closing, there is a risk of an unwanted injector reopening event. By detecting this event the second pulse (braking pulse) timing can be corrected. A known technique provides a method of injector reopening detection which is realized by calculating the derivative of the fdtered injector voltage after the second injector pulse (braking pulse) If the voltage derivative exceeds a calibratable threshold within a given detection window, the injector is considered to reopen. A problem with this technique is that the filtering and derivation process consumes a lot of processor resources and is sensitive to noise / glitches in the measurement signal.

It is an object of the invention to overcome these problems.

Statement of the Invention

In one aspect is provided a method of controlling a fuel injector, said fuel injector including an electrically operated actuator adapted to actuate a needle controlled valve, adapted such that consequent to providing said actuator with an activation signal, said needle is moved away from a valve seat so as to open the valve to dispense fuel, wherein said activation signal has a profile which includes at least one actuation pulse, adapted to move and/or hold said needle to a valve open position and at least one subsequent braking pulse adapted to slow said needle/pintle on subsequent closing of said valve as the needle returns to the valve closed position, comprising:

a) detecting whether a valve re-opening event in an operational cycle of the valve occurs,

b) if so, varying the inter-pulse delay between the end of said activation pulse and the start of the braking pulse for subsequent operational cycles;

wherein in step b) re-opening event is determined by integrating the current or voltage signal over a time window, and determining if re-opening has occurred if said integrated signal is at or below a threshold value.

Said re-opening event may be detected in a time window starting from the end of the last breaking pulse.

Said inter-pulse delay may be increased if it is determined that re-opening occurs.

Said inter-pulse delay may be increased by a pre-determined increment.

If it is determined there is no re-opening of the valve in step a) the further following steps may be implemented:

i) determining the closing time/duration of the valve by detecting a closing event; and ii) increasing the inter-pulse delay if the closing time duration is above an upper target value; and/or iii) reducing the inter-pulse delay if the closing time is below a lower target value.

In step i) the closing event may be detected in a time window starting form the end of a first braking pulse.

The closing event time window may finish at the end of the last braking pulse.

The problems of high processor resource need for the reopening detection and the sensitivity to noise in the measurement signal are solved by implementing an integration of the injector voltage in the given detection window.

Brief Description of Drawings

The invention will now be described with reference by way of example and with reference to the following figures of which:

Figure la shows a plot of the characteristics of the voltage across the solenoid actuator against time during one injection cycle (i.e. consequent to application of a corresponding pulse profile), as well as the pintle/needle displacement 2 and injector fueling command 3 received from an engine controller.

Figure lb shows a plot of the characteristics of the voltage (applied) across the solenoid actuator against time during an injection cycle (i.e. consequent to a corresponding pulse profile), versus the pintle/needle displacement for two different pulse profiles.

Figure lc is similar to figure la and shows typical plots of solenoid (actuation) voltage and corresponding current against time.

Figure 2 shows the problems of varying supply pressure of fuel and the effect this has on pintle displacement for a set control strategy;

Figure 3 shows how a chart of the inter-pulse delay is provided dependent on the fuel supply pressure;

Figure 4 shows the delay between the actual fuel pressure supply and that which is the ECU control is processing;

Figures 5a, b and c illustrates one method of detecting a re-opening event;

Figure 6 shows a block diagram represents a simple method which includes a block to calculate the sum of the measurement elements sample by sample;

Figure 7a shows a flow chart of methodology according to one example; and

Figure 7b shows a flowchart of further refined methodology according to one aspect.

Detailed Description of the Invention

Figures la shows a plot of the characteristics of the voltage 1 across the solenoid actuator against time during an injection cycle (i.e. consequent to a corresponding pulse profile), versus the pintle/needle displacement 2 and injector fueling 3 command form engine controller.

The plot shows the injector command pulse 3 that is sent e.g. from an engine ECU or injector controller. This control pulse effectively is determined by e.g. the ECU as a result of the fuel demand, and in a simple example comprises a single pulse of variable length; the variable which is set is thus length of the pulse which determines essentially how long a fuel injector valve is to be opened, which determines the corresponding amount of fuel to be dispensed.

Plot f shows the actual voltage across the actuator eg. solenoid of the fuel injector as a result of the command pulse. The applied voltage across the injector is based on the command/control pulse 3, and may comprise a series of phases which will be described hereinafter. It is to be noted that for a particular command pulse different voltage profiles may be applied to the actuator (e g. solenoid, according to design strategy). It is to be noted that the voltage trace measured across a solenoid actuator will be somewhat influenced by the movement of the valve/solenoid actuator by virtue of induced current/voltage which by will have an influence on the voltage.

As can be seen on figure la, once injector command is received, in a first phase 4 a relatively high initial (activation) pulse voltage 4 is applied to the actuator in order to actuate it, which causes the pintle/needle to start moving and the needle to move away from the valve seat. The pintle starts to move rapidly. After a short while as the pintle starts to move, the drive pulse (voltage) is reduced and a short time after this when the pintle reaches maximum displacement, the drive pulse decreased to a relatively low level called the “hold” phase 5. This is maintained for a set time and then the pulse is reduced to zero once the fueling command is inactive, and the armature/pintle is forced to move back in the opposite direction to close the valve. The period between the start of the first phase 4 and the end of the second phase 5 may be regarded as the fueling phase as shown in the figure by arrow 7. Aspects of the invention are not limited to such profiles and the term “fueling phase” may be regarded at the time period from the start of a first pulse to activate the solenoid actuator to the time after the last activation pulse when the current is reduced to zero, less or substantially close to zero.

A braking pulse 6 may be then applied during the closing phase to control and sense closing of the pintle. The braking pulse may be composed of two elements. A first braking phase/component or pulse 6a where voltage is provided to the coil to generate magnetic force to counter act pintle movement over its closing, and a second braking phase 6b. The duration of the braking pulse duration may be regarded as the time between the start of the first braking pulse component and the end of the last braking pulse component as shown by the arrow 6.

To achieve lowest closing speed, the braking pulses provide a counterforce, carefully controlled to provide a balanced condition. Any sudden change in closing force, like change of operating pressure, would result in undesired injector control status.

There is an inter-pulse delay that is defined as time between the end of the last fueling pulse and the start of the first braking pulse and is shown by arrow 8.

In aspects of the invention, which will be described hereinafter, sensing may be performed and such sensing can take place in a re-opening detection window shown by arrow 9.

Optionally a closing speed sensing window is shown by arrow 10 where within this window, closing speed is sensed via closing time detection. Closing time is defined as the time between the start of the breaking pulse and the time the valve is closed, when the pintle displacement is zero in the figure.

In a simple example valve reopening is detected and the inter-pulse delay varied according to the detection of this.

Figures lb shows a plot of the characteristics of the voltage (applied) across the solenoid actuator against time during an injection cycle (i.e. consequent to a corresponding pulse profde), versus the pintle/needle displacement for two different pulse profdes A and B having different inter-pulse delay durations. Reference numeral 10 denote the voltage plot for profde A and reference numeral 11 shows the plot for case B.

For both cases the initial activation pulse and hold pulse are the same. In the first plot 10 of voltage there is a longer inter-pulse delay as shown by arrow C than the inter-pulse delay shown by arrow D in the second plot 11. The respective pintle movement is shown for the two pulse profiles 10 and 11. By controlling inter-pulse delay, pintle closing speed can be reduced.

This can be monitored via closing time detection. Horizontal arrows E and F represent closing time for the two variant of inter-pulse delay. Plots 12 and 13 show the pintle displacement for cases A and B respectively. There can be a problem of the valve re-opening if the inter-pulse delay is not set correctly.

Figure lc is similar to figure la and shows typical plots of solenoid (actuation) voltage (bottom plot) and corresponding current (top plot) against time during an injection cycle. The plot shows the duration and nature of the fueling pulse, braking pulse, inter-pulse delay and detection time windows which may be used in aspects of the invention.

It is important that the inter-pulse delay is carefully controlled to provide reduced closing speed but without the problem of valve reopening due to a longer inter-pulse delay. As will be explained hereinafter, supply pressure variation of fuel to the injector will have a large effect on the valve opening and closing (i.e. pintle movement) characteristics. Fine tuning of resistive force can be achieved via modulation of inter-pulse delay and controlled by sensing of closing time as observed in figure lb.

Figure 2 shows plots of pintle displacement for the same pulse profile but with varying fuel supply pressure to the fuel injector /valve. The problems of varying supply pressure of fuel and the effect this has on pintle displacement for a set control strategy are clearly seen. If there is a sudden reduced pressure in the fuel supply to injectors, this causes an increase of the closing force acting on the valve which consequentially leads to an increase in the valve closing speed. This increase stresses and noise.

Where there is an increase in the pressure in the fuel supply to the injector, this reduces the closing force and causes the pintle to reopen which not only increases stress and noise but also results in significant fuel increase. Any sudden change in closing force, like change of operating pressure, would result in undesired injector control status. There is also a danger that if the braking pulses are too high and /or too long, or the inter-pulse interval between braking pulses is too small there is a danger of the valve re-opening.

Various closing speed algorithms are known to prevent reopening events by setting pre-leamed value the inter-pulse delay as function of pressure. Figure 3 shows how a chart of the inter-pulse delay is provided dependent on the fuel supply pressure; this can be stored in an controller as a map and used in such control methods.

However, over fast transient pressure fluctuation, the pressure data refresh rate can be much slower than pressure ramp up; in other words there is a delay between the actual fuel pressure supply 15 and that which is the ECU control is processing 16 as shown in figure 4. At highest engine speed, this results in undesired reopening events. While refresh rate can be addressed via improvement of hardware pressure acquisition system, it will impact cost and potentially be intrusive to existing EMS system.

Detailed Description of the Invention.

In a simple embodiment, the voltage across the solenoid is monitored to detect if re-opening occurs. The inter-pulse delay may be varied and controlled accordingly.

Optionally the inter-pulse delay may be varied (i.e. controlled) as a function of the detection of any valve re-opening as well as other parameters such as pressure. Thus such methodology may be used with known methods of closing speed control that will detect possibility of reopening failure and apply instantaneous corrective action prior fueling failure.

Detection of reopening event is be performed through analysis of the voltage signal across actuator terminals. This is preferably performed over a time window starting from after the braking pulse. The inventors have determined that if voltage signal is integrated over a time window; this yields information enabling the determination of whether re-opening has occurred.

In a simple aspect of the invention, the voltage or current signal across the solenoid actuator is monitored preferably in a predetermined time window, and the signal integrated. The integrated signal is then analysed to determine if re-opening was likely or has occurred.

Reopening Detection

Figure 5 shows how the re-opening of the valve in a fuel injector can be detected . Figure 5a shows several plots for the pintle displacement against time. The plots above line Z are for where valve reopening occurs and those below the line are those acceptable characteristics where re-opening does not occur. FT stands for “Fast Transition” - it is the recirculation phase where upper and lower bridge transistors are open and the induced coil current is pushed in the boost capacitor. During this phase, the coil voltage is inverted (going negative).

Figure 5c shows the voltage across the solenoid terminals versus time. The vertical line Y1 at time 248 micro seconds indicates the start of the detection window in one examples. The vertical line Y2 at 1050 micro seconds indicates the end of detection window in one example.

Figure 5b shows the integral of the voltage trace against time and thus represents the surface (area) between the voltage curve (traces) and the OV-axis, for a number of different events. Those where reopening occurs are shown by reference “R” and those where re-opening does not occur are shown by reference numeral “S”. These correspond to the similarly referenced traces in figures 5a and 5c

As can be seen if the integral over the sampling period is above a threshold level, reopening occurs, if less than the threshold then re-opening does not occur. Thus in one example the voltage trace is integrated over the sampling window (predefined time period) and then compared to a threshold. If more than this threshold then re-opening is deemed to have occur, conversely if the value is less than the threshold then re-opening is deemed not to have occurred..

Since a negative voltage is integrated, a lower integral result means a larger surface. If the integral is lower than a threshold, the cycle is classified as reopening event. The first plot in Figure la shows the corresponding pintle displacement traces. A pintle displacement of ~ 360 micro meter means the injector is fully open, 0 micro meter means fully closed. The soft landing pulse will reduce the impact speed of the pintle. If the soft landing pulse is positioned too early in the closing phase, it will generate more opening force and the injector will not move directly to the closed position right away but will first open again and then be pushed by the spring to the closed position with higher speed what is undesired. Plot 5b shows the integral of the voltage curve (between 0 and the negative decay voltage within the detection window) versus the corresponding pintle closing time (pintle displacement < 10 micro meter). The horizontal line ZZ in the middle plot indicates the threshold. All integral values below this threshold are reopening events. This represents a smooth line with very little influence of noise on the signal and therefore is more robust compared to prior art method.

Figure 6 shows a block diagram represents a simple method which includes a block to calculate the sum of the measurement elements sample by sample. No additional calculation step is required. The end result can directly be used to compare to a threshold. The integration method does not require filtering nor derivative calculation. A simple sum of the elements of the array of measured voltage value is sufficient. This sum can be compared to a threshold value. The comparison result will decide if a reopening did occur or not.

This is one general aspect re-opening is detected in a re-opening time window (by such an above described method), and if re-opening is detected, the inter-pulse delay is varied. The inter-pulse delay may also be controlled and varied on the basis of such a determination and one or more other variables such as pressure of fuel supply to the fuel injector, closing time, and such like. Thus the method can be used with other known methodologies

Detailed Example

Figure 7a shows a flow chart of methodology according to one example of how a pulse profile for a fuel injector is set. The chart effectively shows a method to control the closing of a fuel injector valve, and incorporates a valve reopening detection feature.

In initial step SI, the pulse width of the control signal (e.g. from the engine ECU ) is read. Optionally the fuel supply pressure to the fuel injector is determined

In step S2, the particular pulse profile is determined on the basis of the information from step S1. The fuelling pulse will be set and also the braking pulse is selected.

In step S3, the inter-pulse delay is set and may be determined based upon fuel pressure and/or valve characteristics. The inter-pulse delay may thus be determined as a function of fuel and pulse width. The relationship may be stored e.g. as maps/look-up charts in e.eg the ECU.

For example it may be determined by also be determined by linear interpolation between adjacent pressure values.

The inter-pulse delay may also varied according to feedback from valve re-opening detection as will be described hereinafter.

During the subsequent control of the fuel injector, the actuator is operated according to the profde set in the previous steps including applying the braking pulse in step S4.

After this, there is a step of determining if reopening of the valve is detected: at step S5.This may be performed as explained above.

If re-opening is detected, this may be recorded at step S6, and at step S7 the inter-pulse delay is increases by e.g. an incremental value Y ps (microseconds). In optional step S8, the MAPs, tables or look-up charts relating the inter-pulse delay to pressure and/or pulse width may be up-dated accordingly e.g. the offset in inter-pulse delay learning table, is updated. This th eMAps used may be updated in a self-learning method.

The value of Y (ps) is then sent to step S3 where a new inter-pulse delay is set based on this.

As mentioned additionally the system may be self- learning and hence tables and map used to calculate inter-pulse delay may be updated according to the results. This is done according to an example in step S8.

It is to be noted that any or more of the above steps can be performed within a engine ECU or within a separate injector I fuel injection system controller.

Further Example.

Figure 7b shows a further refined example. The method is similar to that described above but is refined in terms of including additional steps. These additional steps provide a method where the inter-pulse delay is refined according to closing time (response characteristics). Steps Si to S8 are identical to the method steps above.

In step S5 if no valve re-opening has been detected, the method proceeds to step S9 where the closing time/response/duration is determined. This may be performed by known techniques. In one example the closing time may be determined from the maximum time after application of a (first) braking pulse (or from application of the braking pulse) that the detected voltage is zero or substantially zero. The closing time/duration may be determined in a closing sensing window which is preferably from the end of the first braking pulse to the end of the last braking pulse, as shown in figure la.

In step S9 if the closing time CT is on target i.e. is lies within a set threshold from target CT the procedure will return to step S3. The inter-pulse delay will not be altered for the next injection cycle.

If not the process proceeds to step S10 where if the current closing time is substantially less than the target (i.e. below the minimum threshold) the inter-pulse delay is reduced incrementally e.g. by X ps (microseconds). If the closing time (response) is greater than a threshold above target then the inter5 pulse delay is increased by X ps. After step 10 the method proceeds to step S8. Again the system may be self- learning and hence tables and map used to calculate inter-pulse delay may be updated according to the results of the closing time response.

Claims

1. A method of controlling a fuel injector, said fuel injector including an electrically operated actuator adapted to actuate a needle controlled valve, adapted such that consequent to providing said actuator with an activation signal, said needle is moved away from a valve seat so as to open the valve to dispense fuel, wherein said activation signal has a profile which includes at least one actuation pulse, adapted to move and/or hold said needle to a valve open position and at least one subsequent braking pulse adapted to slow said needle/pintle on subsequent closing of said valve as the needle returns to the valve closed position, comprising:

2. A method as claimed in claim 1 wherein said re-opening event is detected in a time window starting from the end of the last breaking pulse.

3. A method as claimed in any preceding claim wherein said inter-pulse delay is increased if it is determined that re-opening occurs.

4. A method as claimed in claim 3 wherein said inter-pulse delay is increased by a predetermined increment.

5. A method as claimed in claim 1 to 4 where, if it is determined there is no re-opening of the valve in step a) the further following steps are implemented:

6. A method as claimed in claim 5 where in step i) the closing event is detected in a time window starting form the end of a first braking pulse.

7. A method as claimed in claim 6 where the closing event time window finishes at the end of the last braking pulse.

Intellectual

Property

Office

Application No: GB1612992.6 Examiner: Mr Alastair Kelly