Method for operating a fuel injection system

Abstract

A fuel injection system (10, fig 1) in an internal combustion engine comprises a fuel injector (2, fig 1) coupled to a fuel rail (1, fig 1), and a pressure sensor (6, fig 1) that monitors fuel pressur

Description

Technical field The present invention generally relates to internal combustion engines, more specifically to an improved method for operating a fuel injection system. Background Art Fuel injectors are typically controlled by generating pulses which are sent to the actuators of the fuel injectors. The amount of fuel injected typically depends on the length of a pulse sent to the actuator. Typically, an Engine Control Unit (ECU) adjusts the pulse length as a result of the demand quantity of fuel to be injected. The demand quantity of fuel is conventionally stored in a map which relates it to the engine speed and the torque demand. Characteristics of fuel injectors may vary over time, e.g. as a result of wear. It is thus important to calibrate injection systems periodically to account for variations in the lifetime of their injectors. Techniques are known which apply learning strategies, whereby injector characteristics are freshly determined, and the injectors are consequently appropriately controlled. The amount of fuel effectively injected by a fuel injector is commonly estimated using pressure drop analysis (PDA), whereby one or more sensors monitor the pressure in a fuel rail which supplies fuel to the injectors. During an injection event, the pressure within the rail drops as fuel is released from the injector. This pressure drop can then be used to estimate the amount of fuel injected, as is described in GB 2533104 A. This estimate may then be compared with the demand quantity of fuel to identify discrepancies, recalibrate the injection system and compensate for individual injector behaviors, or trigger an alert if they exceed a certain threshold. However current techniques have limitations, especially for high-pressure systems and low fuel quantity estimation, as they fail to accurately account for static leaks. Technical problem It is an object of the present invention to provide a method for operating a fuel injection system in an internal combustion engine, which overcomes the aforementioned drawbacks and accurately accounts for static leaks when estimating injected fuel quantities. General Description of the Invention This object is achieved by a method for operating a fuel injection system as claimed in claim 1. The fuel injection system comprises at least one fuel injector coupled to a fuel rail, and pressure sensor configured to monitor pressure in the fuel rail. Injection events are operated by applying drive pulses to the fuel injector to inject fuel into a cylinder of the engine. The invention provides a method implementing an injected fuel quantity estimation routine which comprises the steps of: - acquiring raw pressure data for a pre-injection period before a given injection event, and a post-injection period after said given injection event; - determining a first slope parameter by performing a linear regression on the raw pressure data for the pre-injection period; - determining a second slope parameter by performing a linear regression on the raw pressure data for the post-injection period; - determining a mean slope parameter based on the first slope parameter and the second slope parameter; - determining compensated pressure data based on the raw pressure data and the mean slope parameter; - determining a compensated pressure level for the pre-injection period and a compensated pressure level for the post-injection period based on the compensated pressure data; - determining a pressure drop from the difference between the compensated pressure level for the pre-injection period and the compensated pressure level for the post-injection period; - determining an injected fuel quantity from the determined pressure drop; - using the injected fuel quantity to control operation of the engine. The method according to the invention has been found to be less sensitive to variations / noise in the pressure data and to more accurately account for static leaks. Indeed, the inventive methods effectively cancels out the contribution of static leaks by determining a compensated pressure data based on the mean slope parameter. At the same time, the inventive method is less sensitive to noise as it determines pressure levels before and after the injection event from this compensated pressure data. Typically, the raw pressure data for the pre-injection period and post-injection period are acquired by means of the pressure sensor. The step of determining the injected fuel quantity from the determined pressure drop is conventionally achieved using a calibrated mapping. In particular, a pressure drop analysis (PDA) mapping can be used, which relates injected fuel quantities to fuel rail pressure, fuel temperature and pressure drop. The inventive method is applicable to engines operating on liquid fuel (gasoline, diesel, biofuels, alcohols, synthetic fuels...) and on gaseous fuels (hydrogen, CNG...). Advantageously, the first slope parameter and / or the second slope parameter are determined using a least square method. In embodiments, the mean slope parameter is computed as an average of the first slope parameter and the second slope parameter. In embodiments, a linear model is defined with the mean slope parameter, and the compensated pressure data is determined by subtracting said linear model from the raw pressure data. For example, the compensated pressure data may be computed as the difference between the raw pressure data and a linear function with the mean slope parameter as its slope: P_comp = P_raw - (a3 * index), where Praw is the raw pressure data, a3 is the mean slope parameter and index represents crank angle or time. In embodiments, the compensated pressure level for the pre-injection period is determined from a median value of the compensated pressure data for the preinjection period, and / or the compensated pressure level for the post-injection period is determined from a median value of the compensated pressure data for the postinjection period In embodiments, the injected fuel quantity is used to recalibrate the fuel injection system. In particular, the injected fuel quantity may be used to determine an injector specific correction factor. In embodiments, when operating a liquid fuel injection system, the pre-injection period begins after the end of a pumping event, and the post-injection period ends before the start of a subsequent pumping event. In embodiments, when operating a gaseous fuel injection system, the pre-injection period begins after the closure of a shut-off valve connecting the fuel rail to a fuel source, and the post-injection period ends before the opening of said shut-off valve. In embodiments, the beginning of the post-injection period is offset from the end of the injection event by a predetermined delay. This delay is calibrated to allow the injector pintle / needle to come to rest on a standstill closed position. The invention further provides a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method as described above. The invention further provides a fuel injection system for an internal combustion engine, the fuel injection system comprising at least one fuel injector coupled to a fuel rail to inject fuel into a cylinder of the engine, a pressure sensor being configured to monitor pressure in the fuel rail and a processing unit configured to execute the method as described above. Brief Description of the Drawings A preferred embodiment of the invention will now be described, by way of example, with reference to the accompanying drawings in which: Fig. 1 is a schematic view of a liquid fuel delivery system; Fig. 2 is a schematic view of a gaseous fuel delivery system; Fig. 3 is a flowchart of an embodiment of the method according to the invention; Fig. 4 is a graph showing various plots illustrating pressure variations around an injection event. Description of Preferred Embodiments Figure 1 shows a schematic view of a liquid fuel delivery system 10 for a vehicle, in which the methods according to embodiments of the invention may be used. Such fuel system is well known and will only be briefly described. The fuel delivery system 10 comprises a common fuel rail 1 (or accumulator) fluidly connecting the fuel therein to a plurality of injectors 2 (e.g. solenoid actuated fuel injectors). The circuit typically includes an in-tank electrical fuel pump 3, a fuel filter 4, and a high-pressure pump 5. A high-pressure sensor 6 is located on the fuel rail 1 to measure the fuel pressure inside the fuel rail 1. A high-pressure valve 8 is provided on the fuel rail 1, and opens when the pressure exceeds a predetermined safety threshold (which may be passive or controllable). Reference sign 7 indicates a backleak circuit including a backleak pressure regulator and an injector return line. Backleak circuits 7 are typically used for indirectly controlled fuel injectors, such as Diesel injectors which include a solenoid-controlled valve that hydraulically controls the needle. Backleak circuits 7 are generally not required for injectors where the pintle is directly controlled via the solenoid, typically via an armature coupled to the pintle, such as gasoline injectors. The fuel system 10 also includes an engine control unit (ECU) 36 connected to the various operating components of the fuel delivery system, i.e. the fuel pumps, sensors and injectors. The ECU 36 typically includes at least one processor and memory and is configured, inter alia, to control fuel injection. The ECU 36 is configured to operate the present method via hardware and / or software. Figure 2 shows a schematic view of a gaseous fuel delivery system 100, in which the methods according to embodiments of the invention may also be used. The gaseous fuel delivery system 100 comprises a gaseous fuel tank 102, a pressure regulator 104, and a fuel rail 106 coupled to a plurality of gas injectors 110. A pressure sensor 111 is located on the fuel rail 106 in order to measure the gaseous fuel pressure inside the fuel rail 106. The pressure regulator 104 is serially connected between the gaseous fuel tank 102 and the fuel rail 106 by means of piping 108. The gaseous fuel tank 102 is configured to store pressurized gaseous fuel at pressures of up to 700 bars. The pressure regulator 104 comprises shut-off means, such as an arrangement of valve with at least one shut-off valve. The pressure regulator 104 further comprises pressure regulating means, such as an arrangement of valve (typically electronically controlled) configured to decreases the flow pressure downstream thereof to a nominal working pressure range, e.g. around 5 to 40 bar. The shut-off means and the pressure regulating means may be distinct elements or may be combined as a single element. The gaseous fuel delivery system 100 also includes an engine control unit 112 connected to the various operating components of the fuel delivery system, i.e. the pressure regulator 104, pressure sensor 111 and injectors 110. The ECU 112 typically includes at least one processor and memory and is configured, inter alia, to control fuel injection. The ECU 112 is configured to operate the present method via hardware and / or software. Diesel systems generally involve significantly higher fuel rail pressures compared to gasoline systems. This results in significantly more pressure waves, and increase the impact of fuel temperature variations. Furthermore, static leaks must be considered, especially for low fuel quantity estimation or when temperature variations are important. For Diesel engines, these static leaks occur mainly at the injectors 2 to the backleak circuit 7, resulting in a decrease of pressure in the fuel rail 1. Likewise gaseous fuel systems, in particular hydrogen systems, are susceptible to static leaks due to the relatively small molecular size of their fuel. For gaseous fuel engines, these static leaks occur mainly at the pressure regulator 104 towards the fuel injectors, resulting in an increase of pressure in the fuel rail 106. Figure 3 is a flowchart representing an embodiment of the inventive method. Figure 3 will be described below in conjunction with figure 4, which illustrates various plots of pressure against index, for a liquid fuel engine with e.g. the fuel delivery system shown on figure 1. The inventive method may be carried out by a processing unit of a vehicle, such as its ECU. In an initial step S1, raw pressure data in the fuel rail is acquired over an analysis window by the pressure sensor 6 and mapped against index, as represented by plot P_raw on figure 4. The index is typically a measure of the crank angle, but can alternatively be a measure of e.g. time or other physical quantity directly depending thereon. The pressure in the fuel rail may be measured by the pressure sensor 6 with a sampling period of e.g. 20 ps. The analysis window comprises a pre-injection period T_pre, an injection event T_inj, and a post-injection period T_post. The injection event may span the period during which an opening pulse is sent by the ECU to the injector. Alternatively, the injection event may span the hydraulic open time, i.e. the period during which the injector is effectively open, thus accounting for the injector’s opening delay and closing response. The pre-injection period T_pre begins at the start of the analysis window and finishes at the start of the injection event. The post-injection period T_post begins at the end of the injection event and finishes at the end of the analysis window. The beginning of the post-injection period T_post may be slightly offset from the end of the injection event by a predetermined margin / delay. On Gasoline Direct Injection (GDI) configuration, the acquisition window may start e.g. 360 engine degrees before combustion Top Dead Centre (TDCcomb) and finish 60 engine degrees after TDCcomb. On Diesel configuration, the acquisition window may start 180 engine degrees before combustion TDCcomb and finish 180 engine degrees after TDCcomb. To limit hydraulic disturbance for Diesel systems, the high-pressure pumping can be turned off during the PDA pressure acquisition. To limit hydraulic disturbance for gaseous fuel systems, the shut off valve of the pressure regulator 104 may be closed. In subsequent step S2, a first slope parameter a1 is obtained by performing a linear regression on the raw pressure data P_raw for the pre-injection period T_pre. Likewise a second slope parameter a2 is obtained by performing a linear regression on the raw pressure data P_raw for the post-injection period T_post. The linear regressions are performed using e.g. the least square method. The first and second slope parameter are represented on figure 4 by linear plots A1 and A2, respectively. In subsequent step S3, a mean slope parameter a3 is computed from the first slope parameter and the second slope parameter. Preferably, the mean slope parameter a3 is computed from the average of the first slope parameter a1 and the second slope parameter a2. Said average may be weighted. Alternatively said average may be straight, i.e. a3 = al^n2. In subsequent step S4, a compensated pressure data P_comp is computed based on the raw pressure data P_raw and the mean slope parameter a3. More specifically, the compensated pressure data P_comp may be computed as the difference between the raw pressure data P_raw and a linear function with the mean slope parameter as its slope, i.e. P_comp = P_raw - (a3 * index + b), where b is an optional offset which may depend on the initial index of the analysis window. In embodiments, the optional offset may be ignored, i.e. b=0. In subsequent step S5, a compensated pressure level for the pre-injection period T_pre and a compensated pressure level for the post-injection period T_post based on the compensated pressure data P_comp are computed. The compensated pressure levels may be determined by computing the median value of the compensated pressure data P_compfor the pre-injection period T_pre and the postinjection period T_post, respectively. The compensated pressure level for the pre-injection period T_pre is represented on figure 4 by constant plot P1, whilst the compensated pressure level for the post-injection period T_post is represented on figure 4 by constant plot P2. Constant plots PT and P2’ are represented on figure 4 and represent the average value of the raw pressure data P_rawfor the pre-injection period T_pre and the post injection period T_post, respectively. The values represented by PT and P2’ would normally be used for conventional injected fuel estimation. However, as it can be seen, PT and P2’ differ significantly from P1 and P2. More importantly, their respective difference PT-P2’ and P1-P2 also significantly differs. Thus, a pressure drop analysis performed on the raw pressure data P_raw (as in known in the prior art) would yield different results as one performed on the compensated pressure data P_comp. According to the invention, in subsequent step S6, a pressure drop is computed from the difference between the compensated pressure level for the pre-injection period T_pre and the compensated pressure level for the post-injection period T_post. In subsequent step S7, an injected fuel quantity is computed using conventional methods from the pressure drop (P1-P2). For example, a mapping relating the injected fuel quantity to the fuel rail pressure, the fuel temperature and the pressure drop may be used. This injected fuel quantity estimate can then be used by the ECU to control operation of the engine. In preferred embodiments, the injected fuel quantity estimate is used to recalibrate the fuel injection system. The fuel injection system may be calibrated by, e.g. updating an injector-specific map relating injected fuel quantity, fuel rail pressure and pulse duration. The method according to the invention has been found to be less sensitive to variations / noise in the pressure data and to more accurately account for static leaks.

Claims

1. A method for operating a fuel injection system (10) in an internal combustion engine comprising at least one fuel injector (2) coupled to a fuel rail (1), whereby injection events are operated by applying a drive pulse to the fuel injector (2) to inject fuel into a cylinder of the engine, a pressure sensor (6) being configured to monitor pressure in the fuel rail (1), the method implementing an injected fuel quantity estimation routine which comprises the steps of:- acquiring raw pressure data (P_raw) for a pre-injection period (T_pre) before a given injection event (T_inj), and a post-injection period (T_post) after said given injection event (TJnj);- determining a first slope parameter (a1) by performing a linear regression on the raw pressure data (P_raw) for the pre-injection period (T_pre);- determining a second slope parameter (a2) by performing a linear regression on the raw pressure data (P_raw) for the post-injection period (T_post);- determining a mean slope parameter (a3) based on the first slope parameter (a1) and the second slope parameter (a2);- determining compensated pressure data (P_comp) based on the raw pressure data (P_raw) and the mean slope parameter (a3);- determining a compensated pressure level for the pre-injection period (P1) and a compensated pressure level for the post-injection period (P2) based on the compensated pressure data (P_comp);- determining a pressure drop from the difference between the compensated pressure level for the pre-injection period (P1) and the compensated pressure level for the post-injection period (P2);- determining an injected fuel quantity from the determined pressure drop;- using the injected fuel quantity to control operation of the engine.

2. Method according to any of the preceding claims, whereby the first slope parameter (a1) and / or the second slope parameter (a2) are determined using a least square method.

3. Method according to any of the preceding claims, whereby the mean slope parameter (a3) is computed as an average of the first slope parameter (a1) and the second slope parameter (a2).

4. Method according to any of the preceding claims, whereby a linear model is defined with the mean slope parameter (a3), and whereby the compensated pressure data (P_comp) is determined by subtracting said linear model from the raw pressure data (P_raw).

5. Method according to any of the preceding claims, whereby the compensated pressure level for the pre-injection period (P1) is determined from a median value of the compensated pressure data (P_comp) for the pre-injection period (T_pre), and / orthe compensated pressure level for the post-injection period (P2) is determined from a median value of the compensated pressure data (P_comp) for the postinjection period (T_post).

6. Method according to any of the preceding claims, whereby the injected fuel quantity is used to recalibrate the fuel injection system.

7. Method according to any of the preceding claims when operating a liquid fuel injection system, whereby the pre-injection period (T_pre) begins after the end of a pumping event, and the post-injection period (T_post) ends before the start of a subsequent pumping event.

8. Method according to any of the preceding claims when operating a gaseous fuel injection system, whereby the pre-injection period (T_pre) begins after the closure of a shut-off valve connecting the fuel rail to a fuel source, and the postinjection period (T_post) ends before the opening of said shut-off valve.

9. Method according to any of the preceding claims, whereby the beginning of the post-injection period (T_post) is offset from the end of the injection event (T_inj) by a predetermined delay.

10. A computer program comprising instructions which, when the program is 5 executed by a computer, cause the computer to carry out the method according to any of the preceding claims.

11. A fuel injection system (10, 100) for an internal combustion engine, the fuel injection system comprising at least one fuel injector (2, 110) coupled to a fuel rail (1, 106) to inject fuel into a cylinder of the engine, a pressure sensor (6, 111) 10 being configured to monitor pressure in the fuel rail (1, 106) and a processing unit (36, 112) configured to execute the method according to any one of claims 1 to 9.

Images