Method and system for measuring variation in fuel quantity during a multi-pulse fuel injection event

Abstract

The invention provides a method for analyzing and optimizing fluid injection into an internal combustion engine via a common rail system. Once various injection parameters are determined for a given injection system, these data can be used to model the effects of sequential injection events for the system. A processor can then be used to run the model and adjust the sequential fuel injection events to optimize engine performance and fuel usage.

Description

Technical Field

[0001] This disclosure generally relates to fuel injectors, and in particular to high-pressure fuel injectors for internal combustion engines. Background Technology

[0002] Fuel injectors are typically used to control the flow of fuel into each cylinder of an internal combustion engine. A fuel injector is usually designed to move a valve to open a port, injecting a specific amount of fuel into the corresponding cylinder, and then move the valve to close the port, stopping the fuel injection. Some fuel injection systems are configured to inject fuel into the cylinders in multiple injections within a single engine cycle, rather than once per cycle; this is known as multi-pulse fuel injection. Typically, multi-pulse fuel injection consists of two pulses (e.g., a “pilot” pulse followed by a “main” pulse) or three pulses (e.g., a pilot pulse followed by a main pulse followed by a “follow” pulse) spaced at set intervals; however, many other combinations of two, three, or more pulses are common.

[0003] The fundamental problem with multi-pulse events is that pulses following other pulses are affected by those preceding them. For optimal fuel economy (based on brake fuel consumption rate, BSFC), emissions (based on NOx emissions), and noise and vibration (or noise, vibration, and harshness, NVH) reasons, pilot-master operation is typically positioned with very small intervals (the time interval between pulses). The effects of fuel-fuel interaction are significant with small intervals. Due to fuel-fuel interaction, subsequent pulses (the master pulse or another pilot pulse) will deliver more or less fuel than an equivalent single-pulse event, depending on the pulse interval and accumulator pressure, pilot injection quantity, and master injection quality. Adding more pulses complicates the effects further. In some cases, tight pilot-master intervals can cause armature "bounce" in the fuel injection system due to multiple injections.

[0004] While this pulse interaction can be accounted for to some extent in combustion chart calibration of command injection quantity, rail pressure, and pulse interval, this approach is far from ideal. This type of calibration is typically performed using nominal (or small-sample) injector hardware. Existing methods have consistently relied on open-loop fuel interaction compensation, which is affected by variations in fuel injector performance due to normal production variations and aging-related drift. This variability negatively impacts the engine's expected performance in terms of torque output, emissions, NVH, and fuel economy for a given fuel command.

[0005] Therefore, further contributions to this technical field are still required. The various aspects of the invention disclosed herein provide better and more effective control over these events. Summary of the Invention

[0006] Various embodiments of this disclosure relate to methods and systems for optimizing the injection of fluid into an internal combustion engine via a common rail system. The methods include: receiving from a sensor a quantity of fuel-fuel interaction between a pilot pulse and a main pulse during a multi-pulse fuel injection event by a processing unit; determining, based on the quantity of fuel-fuel interaction, an adjustment to be made to the pilot pulse or the main pulse using a fuel-fuel interaction model relating to the multi-pulse fuel injection event; and performing the determined adjustment on the pilot pulse or the main pulse by the processing unit.

[0007] The method may further include increasing the interval between the pilot pulse and the main pulse by the processing unit to allow sensors to measure the amount of fuel-fuel interaction between the pilot pulse and the main pulse. The determined adjustment may include a change in the amount of fuel delivered during the main pulse. The adjustment may be determined using a fuel-fuel interaction model that takes one or more of the following as inputs: initial pressure, commanded pulse interval, amount of fuel delivered by the pilot pulse, or amount of fuel delivered by the main pulse.

[0008] The method may further include adapting a fuel-fuel interaction model based on operating conditions and fuel-fuel interaction, wherein the operating conditions include one or more of the following: initial pressure, pulse interval of the command, fuel-fuel amount of the pilot pulse, or fuel-fuel amount of the main pulse. The method may also include temporarily disabling the pump connected to the common rail system while the amount of fuel-fuel interaction is being measured. The fuel-fuel interaction model may include a lookup table. The amount of fuel-fuel interaction can be filtered using a Kalman filter to generate predicted fuel-fuel interaction values.

[0009] The method may further include a processing unit comparing a predicted fuel-fuel interaction value with a target main pulse fuel quantity and determining an adjusted on-time fuel injection. When the target main pulse fuel quantity is greater than the predicted fuel-fuel interaction, an adjusted fuel quantity can be calculated by calculating the difference between the target main pulse fuel quantity and the predicted fuel-fuel interaction, and this adjusted fuel quantity is used to determine the adjusted on-time fuel injection. Furthermore, when the target main pulse fuel quantity is not greater than the predicted fuel interaction, an adjusted fuel quantity can be calculated based on the target main pulse fuel quantity and the predicted fuel interaction, and this adjusted fuel quantity is used to determine the adjusted on-time fuel injection. The adjusted on-time provides the adjusted fuel quantity for delivery during the main pulse.

[0010] An engine fuel system as disclosed herein may include a guide rail; a plurality of fuel injectors fluidly coupled to the guide rail, the fuel injectors being configured to inject fuel therefrom; and a control system including at least one sensor and a processing unit operatively coupled to the plurality of fuel injectors, the at least one sensor being configured to measure the amount of fuel-fuel interaction between a pilot pulse and a main pulse during a multi-pulse fuel injection event. The processing unit may be configured to: determine, based on the measured amount of fuel-fuel interaction, using a fuel-fuel interaction model relating to the multi-pulse fuel injection event, to determine an adjustment to be made to the pilot pulse or the main pulse; and to make the determined adjustment to the pilot pulse or the main pulse.

[0011] The processing unit can increase the interval between the pilot pulse and the main pulse to allow the sensor to measure the amount of fuel interaction between the pilot pulse and the main pulse. The determined adjustment may include changes in the amount of fuel delivered during the main pulse. The adjustment can be determined using a fuel interaction model that takes one or more of the following as inputs: initial pressure, commanded pulse interval, pilot pulse fuel quantity, or main pulse fuel quantity. The processing unit can also be configured to adapt the fuel interaction model based on the operating conditions and fuel interaction of multiple injectors, including one or more of the following: initial pressure, commanded pulse interval, pilot pulse fuel quantity, or main pulse fuel quantity. The processing unit can also be configured to temporarily deactivate multiple injectors connected to the guide rail when measuring the amount of fuel interaction.

[0012] While several embodiments have been disclosed, other embodiments of this disclosure will become apparent to those skilled in the art from the following detailed description of the illustrative embodiments shown and described herein. Therefore, the drawings and detailed description should be considered illustrative and non-limiting in nature. Attached Figure Description

[0013] These embodiments will be more readily understood in light of the following description and in conjunction with the accompanying drawings, wherein like reference numerals denote like elements. These depicted embodiments should be construed as illustrative of this disclosure and not as limiting in any way.

[0014] Figure 1 It is a graph showing the measurement of total guide rail pressure drop caused by multi-pulse events under specified normal operating intervals.

[0015] Figure 2 This is a graph showing the measurement of the total guide rail pressure drop caused by multi-pulse events under a forced large interval.

[0016] Figure 3This is a flowchart illustrating an implementation of a software algorithm executed by a control unit to control the timing and amount of multi-pulse fuel injection.

[0017] Figure 4A It is a plot of the intervals (ms) of the collected data versus the Q interaction (mg).

[0018] Figure 4B It is a plot of the interval (ms) versus Q interaction (mg), with the collected data minus the data collected at very low intervals.

[0019] Figure 4C It is superimposed on Figure 4B The segmented 1-D lookup table least squares estimate on the plot.

[0020] Figure 5A This is a graph of the gain leader relative to the leader expressed in mg. The actual and extrapolated y-intercepts determine the values ​​of x(1) and x(2).

[0021] picture x B is the gain. 主量 A graph relative to the principal quantity expressed in mg. The values ​​of x(3), x(4), and x(5) are determined by the actual and extrapolated y-intercepts; the extrapolated x-intercepts are used. Figure 6A The original experimental data for the interval versus the Q interaction are shown. Figure 6B The graph shown is generated using coefficients estimated using a least squares lookup table. Figure 6C A graph showing the lookup value relative to the interval time is displayed, and Figure 6D The residuals calculated for the fit for each sample are shown.

[0022] Figures 7A to 7D The residual plot is shown. Figure 7A Showing the residuals relative to Q p ; Figure 7B Showing the residuals relative to Q m ; Figure 7C The residuals are shown relative to the hydraulic interval; and Figure 7D The histogram of the residuals from the least squares fit is shown.

[0023] Figure 8 This is a box-and-whisker plot for coefficients c1, c2, c3, c4, c5, c6, and c7. The mean, standard deviation, minimum, and maximum values ​​of each plotted coefficient are shown in tabular form below the plot.

[0024] Figure 9This is an I-MR chart for coefficients c1, c2, c3, c4, c5, c6, and c7. The N value, mean, overall standard deviation relative to each coefficient, and standard deviation within each coefficient are shown in tabular form below the I-MR chart of the coefficients.

[0025] Figure 10 It is a flowchart of the measured delivery of fuel via multi-pulse injection into the internal combustion engine.

[0026] Figure 11 This is a plot of the refueling error (y-axis) for each sample determined after adjustments for multi-pulse events based on simulation, relative to the x-axis determined for each sample at a fuel rail hydrostatic pressure of 500 bar.

[0027] Figure 12 This is a plot of the refueling error (y-axis) for each sample determined after adjustments for multi-pulse events based on simulation, relative to the x-axis determined for each sample at a fuel rail hydrostatic pressure of 1500 bar.

[0028] Figure 13 This is a flowchart illustrating a method according to an embodiment disclosed herein.

[0029] Throughout the accompanying drawings, corresponding reference numerals indicate the corresponding parts. Although the drawings illustrate embodiments of the invention, they are not necessarily drawn to scale, and some features may be enlarged to better illustrate and explain the invention.

[0030] While this disclosure is open to various modifications and alternatives, specific embodiments are illustrated by way of example in the accompanying drawings and described in detail below. However, it is not intended to limit this disclosure to the specific embodiments described. Rather, this disclosure is intended to cover all modifications, equivalents, and alternatives that fall within the scope of this disclosure as defined by the appended claims. Detailed Implementation

[0031] In the following detailed description, reference is made to the accompanying drawings, which form part of this detailed description, and specific embodiments for practicing the present disclosure are illustrated in the drawings by way of illustration. These embodiments are described in sufficient detail to enable those skilled in the art to practice the present disclosure, and it should be understood that other embodiments may be used and structural changes may be made without departing from the scope of the present disclosure. Therefore, the following detailed description should not be considered limiting, and the scope of the present disclosure is defined by the appended claims and their equivalents.

[0032] Throughout this specification, references to "an embodiment," "implementation," or similar language mean that a specific feature, structure, or characteristic described in connection with an embodiment is included in at least one embodiment of this disclosure. The phrases "in one embodiment," "in an embodiment," and similar language throughout this specification may, but do not necessarily, refer to the same embodiment. Similarly, the term "implementation" is used to refer to an implementation having a specific feature, structure, or characteristic described in connection with one or more embodiments of this disclosure; however, unless explicitly indicated otherwise, an implementation may be associated with one or more embodiments. Furthermore, the features, structures, or characteristics of the subject matter described herein may be combined in one or more embodiments in any suitable manner.

[0033] The embodiments and examples in this disclosure provide methods and systems for measuring, adapting, and compensating for quantity changes (fuel interactions) occurring in subsequent pulses of a multi-pulse fuel injection event for injectors with variable characteristics. The embodiments and examples can be implemented in an engine fuel system including a rail (also referred to as a “common rail”), a plurality of fuel injectors fluidly coupled to the rail, and a control system coupled to the fuel injectors. The control system may include sensors and a processing unit that receives measurements acquired by the sensors to perform calculations and determinations as further explained herein. The sensors can be any suitable sensors capable of measuring quantity changes, such as fuel interactions between pulses. The processing unit can be any suitable processor, such as a central processing unit, system-on-a-chip, or integrated circuit in any suitable computing device. The processing unit adapts to and compensates for the quantity changes.

[0034] This compensation, in terms of timing and / or interval adjustment, can be created by understanding the injection characteristics of each individual injector, fuel-fuel interaction measurements, rail pressure and temperature, and the command's start time and pulse intervals. Systems based on the multi-pulse compensation algorithm disclosed herein individually determine and compensate for fuel-fuel interaction errors for each injector for multi-pulse operation. The algorithm is capable of adapting to manufacturing variations and aging-related changes. Therefore, the algorithm increases fuel economy benefits and improves emissions and NVH by achieving tighter fuel-fueling and more precise timing for each pulse during multi-pulse operation.

[0035] Figure 1 and Figure 2 A measurement strategy for measuring fuel-fuel interaction during pilot + main operation is illustrated. Figure 1 and Figure 2In the diagram, when the pump is turned on or activated, the guide rail pressure 101 is in normal operation and maintained at a specific level. When the pump is turned off or deactivated, the guide rail pressure 101 decreases due to measurement by the pilot + main operation 102. Figure 1 During operation, the pilot-master interval 103 remains the same as during normal operation. The total pressure drop consists of the pressure drops caused by the pilot quantity, master quantity, and interaction quantity. The pressure drop is proportional to the amount of fuel added, given the geometry of the sonic and high-pressure common rail system.

[0036] Total fuel measurements can be written as the sum of individual contributions, as shown below:

[0037] Q 总1 =Q 先导 +Q 主 +Q 相互作用 (Equation 1)

[0038] For systems employing closed-loop fuel control (CLFC) based on single-pulse measurements, the leader quantity (Q) present in the presence of a subsequent pulse can be calculated or measured using methods known in the art. 先导 In some examples, the sensor is used to measure the leader (Q). 先导 Total (Q) 总1 The unknown quantity (Q) is also measured, for example, using a sensor. Therefore, the unknown quantity is the principal quantity (Q). 主 ) and interaction quantity (Q) 相互作用 By measuring Q 主 Q can be calculated using equation (1). 相互作用 Now for reference. Figure 2 In order to measure the principal quantity (Q) more accurately 主 In the pilot + main operation 102, a forced ratio is applied between the pilot and main pulses. Figure 1 The large interval 200 between the pilot and master pulses, 103, makes the fuel interaction approximately zero. The pulse interval 103 between the pilot and master is only for measuring the master quantity (Q). 主 The purpose was to change to a larger interval of 200, such as Figure 2 As shown. Figure 2 The described total fuel consumption measurement (Q) 总2 ) was written as:

[0039] Q 总2 =Q 先导 +Q 主 (Equation 2)

[0040] In equation (2), the total amount (Q) 总2 There is no contribution from fuel interactions, i.e., Q 相互作用=0, because the leader and master are placed further apart, and there is no detectable pulse-to-pulse interaction. Therefore, it is possible to measure the total fuel amount (Q) based on equation (2). 总2 Subtract the leading quantity (Q) from the input. 先导 To calculate the principal variable (Q) 主 ).

[0041] Once the principal quantity (Q) has been measured 主 Then, equation (1) is used to calculate the total amount (Q). 总1 Subtract the leading quantity (Q) from the input. 先导 ) and principal quantity (Q) 主 To calculate the refueling interaction (Q) under close spacing. 相互作用） ,as follows:

[0042] Q 相互作用 =Q 总1 -Q 先导 -Q 主 (Equation 3)

[0043] Experience with fuel-fuel interactions suggests that subsequent pulses (main pulses or pilot pulses) deliver more or less fuel compared to an equivalent single-pulse event. Test data and / or injector simulation combined with system identification techniques are used to create fuel-fuel interaction models involving multi-pulse injection events. The model inputs may include operating conditions such as one or more of the following: initial pressure, commanded pulse interval, commanded pilot quantity (fuel delivery amount of the pilot pulse), or main quantity (fuel delivery amount of the main pulse). Model parameters may include injector characteristics such as: hydraulic injection duration, injection start delay, and injection end delay. Model outputs may include the actual fuel delivery amount and the actual timing of the second pulse. If desired, other injection parameters, such as injection start, injection end, duration, or the centroid of the injection pulse, may also be expressed as outputs using formulas.

[0044] Figure 3 A flowchart illustrating an implementation of a software algorithm executed by a control unit to control the timing and amount of multi-pulse fuel injection is shown. In the topmost box 301, a measurement strategy for the leader, main, and multi-pulse interactions is determined, see equation (3) above. In the middle box 302, a fuel interaction model is created such that the model is configured to adapt to manufacturing variations and aging-related variations, for example, such that the adapted leader-main interaction is lower than the default leader-main interaction. In the bottom box 303, the timing of the pulses is changed, for example by shortening the duration of the main pulse (e.g., by adjusting the timing of the pulses). Figure 3 (as shown in the diagram) and / or shifting the timing of the pulse (e.g., earlier or later) to compensate for fuel interaction errors in the fuel interaction model.

[0045] The following describes some examples of experiments and simulations that can be performed based on this disclosure.

[0046] In one example, bench testing was conducted. The effect of the pilot pulse on the quality of the main fuel quantity injected during a multi-command fuel injection event in a single-cylinder event was measured. Variables considered to affect this parameter included: the amount of the pilot pulse, the interval between pulses within the commanded fuel injection, the guide rail pressure, and the characteristics of the individual fuel injector.

[0047] Multiple test plans were performed using six (6) near-nominal injectors. Specific variable changes are as follows:

[0048] 1. Lead dose: 1 mg to 5 mg (2 mg)

[0049] 2. Main dosage: 4mg to 130mg (4mg to 130mg)

[0050] 3. Hydraulic interval: 0.05ms to 1ms (0.05ms to 0.7ms)

[0051] 4. Guide rail pressure: 500 bar to 2100 bar (500 bar and 1500 bar)

[0052] Data was collected from 840 test points in each of the three runs, resulting in a dataset containing 2520 data points for each injector. The values ​​in parentheses above were used to obtain the 2520 data points shown in the figure.

[0053] Then, the bench test data is analyzed. Now, refer to... Figure 4A and Figure 4B The change in Q interaction, expressed in milligrams (mg), relative to hydraulic interval time, expressed in milliseconds (ms), was measured. Figure 4A The original data obtained is shown. Figure 4B It shows Figure 4A The image shows the original data after editing to remove data points collected at very low intervals. Figure 4B The data shown will be further analyzed, as explained below.

[0054] Now for reference Figure 4C It shows the method for using based on Figure 4B The data shown represents the selected points in the base lookup table. The values ​​of the lookup table were created by performing a 1-D least squares fit with a resolution of 15 points (shown as connected by continuous white lines), with intervals of 0.05 to 0.7 milliseconds for each test plan. This lookup table can be called the "base lookup table" because this base lookup is calculated to estimate the coefficients and the final lookup, taking into account the amount Q of the leader pulse. p The number of main pulses Q mThe data used in this fitting are related to the spacing between them and the influence of guide rail pressure. Figure 4B The data presented is the same.

[0055] A model was then developed to predict the interaction between multiple jet events using the following equation (Equation 4):

[0056]

[0057] In equation (4), V 增益 Indicates vertical scaling, H 偏移 This represents any horizontal offset in the data. In the equations, S represents the hydraulic interval in milliseconds (ms), and Q represents the interaction in milligrams (mg). Q relative to S is based on a lookup table of 10 to 20 calibrable breakpoints. p and S p Based on Q i Q i+1 S i and S i+1 It is determined by measurement or calculation.

[0058] Then, based on equation (4), the following equation (equation 5) is calculated:

[0059]

[0060] Q 相互作用 It is the amount of fuel interaction, gain. 先导量 The gain is due to the leader. 主量 The gain is due to the principal quantity, where P is pressure, and the table... k-1 The sum of values ​​in table k is obtained from a lookup table, with intervals between them. k-1 The interval k is the interval between the pilot pulse and the main pulse, the interval Msmt is the interval between the pilot injection and the main injection event for which measurement is performed, and C P It is the guide rail pressure coefficient, and These are offset coefficients. In equation (4), each variable other than pressure P and interval Msmt is called a coefficient and can be determined offline or estimated online, as explained below.

[0061] Coefficients 1 and 2 are attributed to Q p The gain (i.e., the gain of the leader); coefficients 3, 4, and 5 are attributed to Q. mThe gain (i.e., the main quantity); factor 6 is attributed to the pressure gain; and factor 7 is the offset for level adjustment. For example, the values ​​of factors 3, 5, and 7 are calibration values ​​determined offline for appropriate injector data (such as data obtained from the U.S. Department of Energy). The values ​​of factors 1, 2, 4, and 6 are estimated using pressure drop measurements (e.g., measured using a flow meter). Examples of such flow meters used may include flow meters manufactured by AIC Systems AG in Basel, Switzerland.

[0062] Gain 先导量 Gain 主量 and C P For online estimation; and the table k-1 ,surface k ,interval k-1 , interval k and These are calibration values ​​to be determined offline. Based on publicly available information, it is understood that different estimation and / or calibration methods can be used to derive appropriate values, such as obtaining data from the U.S. Department of Energy and measuring pressure drop measurements, such as those using a flow meter. In some examples, p-value tests are used to analyze data, where coefficients leading to greater variability have higher p-values. To create robust models, coefficients with higher p-values ​​can be selected to generate models simulating the interactions of jet events. In addition to p-values, single-value moving range (I-MR) tests can also be performed, where the results can show the level of variation for each given variable.

[0063] To determine the gain in equation (5) 先导量 The value of Qp can be used to execute the following algorithm, where Qp = the precursor:

[0064] (1) For Qp < Qp_cal:

[0065] (2) Others: Gp = x(2)

[0066] In the algorithm described above, Qp_cal is defined as a calibrable Qp threshold. Figure 5A An algorithm is shown that graphically depicts how the gain of the leader is affected by the leader. Dashed lines indicate higher pressure.

[0067] To determine the gain in equation (5) 先导量 The value of Qm can be used to execute the following algorithm, where Qm = principal variable:

[0068] (1) For Qm <Qmid:

[0069] (2) For Qmid <Qm<Qmh:

[0070]

[0071] (3) For Qm>Qmh:

[0072] Figure 5B An algorithm is shown graphically illustrating how the gain of the principal quantity is affected by the leader quantity. Dashed lines indicate higher pressure. In the algorithm described above, the values ​​of x(1) to x(5) are coefficients, where x(1), x(2), and x(4) are estimated online, while x(3) and x(5) are estimated offline.

[0073] Now for reference Figures 6A to 6D , Figure 6A The experimental data Q interaction is plotted as a function of the interval time (ms). Figure 6B The data estimated using coefficients estimated using a least-squares lookup table of values, determined using the method disclosed above, are plotted as Q-interactions relative to time intervals (ms). Figure 6C Only the estimated lookup table values ​​above are shown, plotted as Q-interactions relative to the interval time (ms). Figure 6D The fit residuals for each collected sample are shown. Statistical analysis of the residuals for the principal variables, "leader," "principal," and "hydraulic interval," indicates no obvious unmodeled trend.

[0074] Now for reference Figures 7A to 7D This shows the residual value relative to Qp( Figure 7A ), Qm( Figure 7B ) and hydraulic interval ( Figure 7C The plot of the residuals and the histogram of the least squares fit (LSF) of the residuals. Figure 7D The σ value for LSF fitting is 2.089 mg / stk.

[0075] Referring to Table 1, p-value tests were used to analyze the data. Coefficients that lead to greater variability have higher p-values. To create robust models, only coefficients with high p-values ​​were used to generate models simulating the interactions of jet events. The p-values ​​of the coefficients are indicated in Table 2.

[0076] Group N average value 95% CI Standard deviation 95% CI c1 6 11.171 (9.4294，12.913) 1.6597 (1.0360，4.0706) c2 6 5.0514 (3.5101，6.5926) 1.4687 (0.9168,3.6021) c3 6 -0.26118 (-6E-01，0.1125) 0.35604 (0.2222,0.8732) c4 6 0.85430 (0.0521，1.6565) 0.76437 (0.4771，1.8747) c5 6 -0.28219 (-6E-01，0.0667) 0.33244 (0.2075，0.8153) c6 6 14.176 (12.533，15.819) 1.5658 (0.9774,3.8402) c7 6 4.531E-04 (-8E-03，0.0094) 0.0085208 (0.0053，0.0209)

[0077] Table 1: P-value test of coefficients

[0078]

[0079] Table 1 (continued): P-value test of coefficients

[0080] coefficient# p-value 1 0.583 2 0.661 3 0.493 4 0.090 5 0.045 6 0.256 7 0.629

[0081] Table 2: P-values ​​of the coefficients, taken from Table 1

[0082] Now for reference Figure 8 The box-and-whisker diagram illustrates the length of the box, and the whisker length corresponds to the amount of change of a given coefficient. Now refer to... Figure 9 A one-valued moving range (I-MR) test was performed, where the I-MR chart shows the level of variation for each given variable. Table 2 (p-values) Figure 8 (Box and whisker diagram) and Figure 9 The test results mentioned in (I-MR) are compiled so that the weighted results of these tests are summarized in Table 3.

[0083]

[0084]

[0085] Table 3: Compilation results of the three tests (p-value, box-and-whisker plot, and I-MR) performed on the coefficients.

[0086] Of the seven (7) coefficients analyzed, four (4) of the seven (specifically, coefficients 1, 2, 4, and 6 in the example shown) were considered sufficiently high to effectively account for virtually all the variability of the data and generate a robust model, and were therefore selected for conditioning. Thus, the remaining three (3) coefficients (coefficients 3, 5, and 7 in the example shown) were treated as constants during the modeling process. The process noise covariance (in the form of a 4x4 matrix Q) was created by selecting a dataset collected for a single cylinder. This database was used to estimate the four coefficients chosen for the selected cylinder. In this example, the process was repeated for all six (6) cylinders that generated six distinct sets of data. The covariance of the four coefficients and the six repetitions was calculated.

[0087] The gain leader (lead), gain principal (principal), and pressure-related coefficients were selected for tuning. See Table 4, where coefficients 1 and 2 are selected for the gain correlation due to the leader, coefficient 4 is selected for the gain correlation due to the principal, and coefficient 6 is selected for the gain correlation due to the pressure.

[0088]

[0089] Table 4: Coefficients and their explanations

[0090] The noise covariance matrix of the coefficients (e.g., matrix Q-4x4) is selected for adaptation by the following process: (1) estimating the dataset for a single cylinder for the selected four coefficients, (2) analyzing the dataset for each cylinder for a six-cylinder engine (a total of six datasets), and (3) calculating the covariance between the four coefficients of the six datasets.

[0091] Now for reference Figure 10The diagram illustrates a process 1000 for adjusting fuel injection into an internal combustion engine based on four coefficients identified as sufficient to model multi-pulse events. The total amount of fuel injected in each multi-pulse injection event 1002 is the amount of fuel Q in the target main pulse. Mo 1004 and the fuel quantity Q in the pilot injection measured in situ 先导 The sum of 1006. The output of the process is a tuned multi-pulse injection event optimized for fuel timing and quantity in the pilot and main injection events. To further refine the correlation of the coefficients and the predictive completeness of the model, the input is processed by a Kalman filter 1008. The Kalman filter 1008 filters the input interaction values ​​using a linear quadratic estimate or joint probability distribution of the interaction values ​​measured over multiple time frames, and then outputs the predicted fuel-fuel interaction Q. Int The value of 1010.

[0092] The key decision point in the model is predicting the fuel-fuel interaction Q. Int 1010 and the target main pulse Q Mo Comparison of the relative quantities of 1014 and 1012. If Q Mo 1014 is greater than Q Int 1010, then from Q Mo Subtract Q from the value of 1014 Int The value of 1010 (as shown in box 1016) is used to generate the adjusted quantity Q. 调适 1018. Then, Q 调适 1018 is processed by the fuel injection start-up time conversion algorithm (FON) 1020 to generate an adapted start-up time. 调适 1022, where "opening time" is defined as the actual injection time or the time interval during which the fuel injector remains open. If Q Mo Not greater than Q Int The adjustment amount Q is then determined using the following equation (shown in box 1024). 调整 :

[0093]

[0094] Then process Q through FON 1020 调整 Start time with output adjusted start time value 调适 1022. Opening Time 调适 The value of 1022 is converted to produce an output used to adjust the on-time of the parameters of the multi-pulse jet event 1002. 调整 1026. Then, obtain the total fuel consumption measurement value Q. 总 1028 and use it as the input for the next loop of process 1000.

[0095] The model's ability to reduce fuel loss caused by the interaction between the pilot fuel injection pulse and the main fuel injection pulse was evaluated. The adjusted on-time fuel injection quantity was compared with the adjusted fuel injection quantity (adjusted fuel injection quantity - (total fuel injection quantity - predicted interaction)) determined at a fuel rail hydrostatic pressure of 500 bar. Reference is now made. Figure 11 The plot shows the refueling error for each sample, determined after adjusting for the multi-pulse event based on simulation (y-axis) relative to each sample (x-axis). The error in the original interaction between the leader pulse and the main pulse (green line, 1101) is significantly greater than the error in the compensated residual interaction (blue line, 1102). For reference, Figure 11 This includes the line indicating the idealized interaction, i.e., the x-axis, where the fueling error for each sample is zero (black line, 1103). Measurements of the adjusted mean residual interaction between pulses are also shown on the same plot (red line, 1104).

[0096] The accuracy of the simulation was further tested by comparing the adjusted fuel injection rate at the start time with the adjusted fuel injection rate (adjusted fuel injection rate - (total fuel injection rate - predicted interaction)) determined at a fuel rail hydrostatic pressure of 1500 bar. (See now for reference.) Figure 12 The plot shows the fueling error for each sample, determined after adjusting for the multi-pulse events based on simulation (y-axis) relative to each sample (x-axis). The error in the original interaction between the leader and the main event (green line, 1201) is significantly larger than the error in the compensated residual interaction (blue line, 1202). For reference, Figure 12 This includes the line indicating the idealized interaction, i.e., the x-axis, where the fueling error for each sample is zero (black line, 1203). Measurements of the adjusted mean residual interaction between pulses are also shown on the same plot (red line, 1204).

[0097] right Figure 11 and Figure 12 Analysis of the data presented in the paper shows that adjusting the fuel delivery parameters based on the model of this invention results in an average reduction of 76% in the interaction between pulses during multi-pulse refueling events.

[0098] Figure 13 The following are illustrated according to some implementation schemes. Figure 3The algorithm is described in the following steps. In step 1301, the algorithm, or more specifically, the processing unit of the fuel injection system (such as a central processing unit, system-on-a-chip, or any other suitable computing device) operating according to the algorithm, measures the amount of fuel-fuel interaction between the pilot operation and the main operation during a multi-pulse fuel injection event. That is, the algorithm measures the amount of interaction between the pilot operation and the main operation and records the time interval between the pilot operation and the main operation. Then, in step 1302, the algorithm determines the amount of adjustment needed in the next pilot operation and the main operation in the multi-pulse fuel injection event to compensate for the fuel-fuel interaction. This determination is made, for example, by inputting measurements such as the injection characteristics of each individual injector, the fuel-fuel interaction, rail pressure and temperature, and the command activation time and the interval between operations.

[0099] In step 1303, the processing unit executes the adjustments determined by the algorithm output. For example, the adjustments may include increasing the interval between the pilot and main operations by a value determined by the algorithm. In some examples, the adjustments may also include changing the actual amount of fuel delivered during each operation. In some examples, the algorithm incorporates a lookup table that determines how much fuel-fuel interaction occurs for the indicated interval between the pilot and main operations / pulses. The lookup table can be modified or adapted based on the injector's injection characteristics and / or operating conditions. The algorithm also uses a fuel-fuel interaction model involving multi-pulse injection events, where one or more of the initial pressure, the pulse interval of the command, the pilot amount of the command, or the main amount can be input. After step 1303, the algorithm returns to step 1301 to remeasure the amount of fuel-fuel interaction to observe whether the previously determined adjustments effectively reduce fuel-fuel interaction.

[0100] The subject matter of this disclosure may be embodied in other specific forms without departing from the scope of this disclosure. The described embodiments should be considered illustrative in all respects and not limiting. Those skilled in the art will recognize that other implementations consistent with the disclosed embodiments are possible. The detailed description above and the examples described therein are presented for illustrative and descriptive purposes only and not for limitation. For example, the described operations can be performed in any suitable manner. These methods can be performed in any suitable order while still providing the described operations and results. Therefore, embodiments of this disclosure are contemplated to cover any and all modifications, variations, or equivalents falling within the scope of the basic principles disclosed above and claimed herein. Furthermore, while the above description describes hardware in the form of processor execution code, hardware in the form of state machines, or dedicated logic capable of producing the same effect, other structures are also contemplated.

Claims

1. A method for optimizing injection of fluid into an engine via a common rail system, comprising: determining, by a processing unit, a fueling interaction amount indicative of an amount of fueling interaction between a pilot pulse and a main pulse of a multi-pulse fuel injection event measured by a sensor; The processing unit determines, based on the amount of the fuel-fuel interaction, using a fuel-fuel interaction model relating to the multi-pulse fuel injection event to determine the adjustment to be made to the pilot pulse or the main pulse, the determination comprising: (a) subtracting the fuel-fuel interaction from the target main injection pulse amount to determine a first adjustment fuel amount as the adjustment in response to the target main injection pulse amount exceeding the fuel-fuel interaction amount; and (b) using the following equation to determine the fuel-fuel interaction amount Q in response to the target main injection pulse amount not exceeding the fuel-fuel interaction amount. Int Subtract the target main injection pulse quantity Q Mo And add an adjustment item to determine a second adjusted fuel quantity Q that is different from the first adjusted fuel quantity. 调整 As for the aforementioned adjustments: and adjusting, by the processing unit, the pilot pulse or the main pulse based on the determination.

2. The method of claim 1, further comprising increasing, by the processing unit, a separation between the pilot pulse and the main pulse to allow the sensor to measure the amount of fueling interaction between the pilot pulse and the main pulse.

3. The method of claim 1, wherein the adjustment based on the determination comprises a change in an amount of fuel delivered during the main pulse.

4. The method of claim 1, wherein the adjustment is determined using a fueling interaction model that involves one or more of the following as inputs: an initial pressure, a commanded pulse separation, a fueling amount of the pilot pulse, or a fueling amount of the main pulse.

5. The method of claim 1, further comprising adapting the fueling interaction model based on a working condition and the fueling interaction, the working condition comprising one or more of the following: an initial pressure, a commanded pulse separation, a fueling amount of the pilot pulse, or a fueling amount of the main pulse.

6. The method of claim 1, further comprising temporarily disabling a pump coupled to the common rail system while the amount of fueling interaction is being measured.

7. The method of claim 1, wherein the fueling interaction model comprises a lookup table.

8. The method of claim 1, wherein the quantity of fueling interaction is filtered through a Kalman filter to produce a predicted fueling interaction value, the method further comprising: comparing, by the processing unit, the predicted fueling interaction value to a target main pulse fuel amount and determining an adjusted open duration fuel injection.

9. The method of claim 8, wherein when the target main pulse fuel amount is greater than the predicted fueling interaction, an adapted fuel amount is calculated by calculating a difference between the target main pulse fuel amount and the predicted fueling interaction, the adapted fuel amount used to determine the adjusted open duration fuel injection.

10. The method of claim 8, wherein when the target main pulse fuel amount is not greater than the predicted fueling interaction, an adjustment fuel amount is calculated based on the target main pulse fuel amount and the predicted fueling interaction, the adjustment fuel amount used to determine the adjusted open duration fuel injection.

11. An engine fuel system, comprising: a rail; a plurality of fuel injectors fluidly coupled to the rail, a fuel injector configured to inject fuel therefrom; a control system comprising at least one sensor and a processing unit operably coupled to the plurality of fuel injectors, the at least one sensor configured to measure an amount of fueling interaction between a pilot pulse and a main pulse during a multi-pulse fuel injection event, the processing unit configured to: determining a quantity of fueling interaction indicative of an amount of fueling interaction between a pilot pulse and a main pulse of the multi-pulse fuel injection event; Based on the measured amount of fuel-fuel interaction, a fuel-fuel interaction model involving the multi-pulse fuel injection event is used to determine the adjustment to be made to the pilot pulse or the main pulse, wherein, in order to determine the adjustment, the processing unit performs one of the following: (a) in response to the target main injection pulse amount exceeding the fuel-fuel interaction amount, subtracting the fuel-fuel interaction amount from the target main injection pulse amount to determine a first adjustment fuel amount as the adjustment; and (b) in response to the target main injection pulse amount not exceeding the fuel-fuel interaction amount, using the following equation to calculate the fuel-fuel interaction amount Q. Int Subtract the target main injection pulse quantity Q Mo And add an adjustment item to determine a second adjusted fuel quantity Q that is different from the first adjusted fuel quantity. 调整 As for the aforementioned adjustments: and adjusting the pilot pulse or the main pulse based on the determined adjustment.

12. The engine fuel system of claim 11, wherein the processing unit increases a separation between the pilot pulse and the main pulse to allow the sensor to measure the amount of fueling interaction between the pilot pulse and the main pulse.

13. The engine fuel system of claim 11, wherein the determined adjustment comprises a change in an amount of fuel delivered during the main pulse, and the adjustment is determined using a fueling interaction model that involves one or more of the following as inputs: initial pressure, commanded pulse separation, pilot pulse fuel amount, or main pulse fuel amount.

14. The engine fuel system of claim 11, the processing unit is further configured to adapt the fueling interaction model based on a working condition of the plurality of injectors and the fueling interaction, the working condition comprising one or more of the following: initial pressure, commanded pulse separation, fueling amount of the pilot pulse, or fueling amount of the main pulse.

15. The engine fuel system of claim 11, the processing unit is further configured to temporarily deactivate the plurality of injectors coupled with the rail while measuring the amount of fueling interaction.

Images