METHOD FOR CARRYING OUT A FORECAST FOR AN AIR FLOW MANAGEMENT SYSTEM FOR AN INTERNAL COMBUSTION ENGINE

Abstract

Method for performing a forecast for an airflow management system for an internal combustion engine, comprising: Setting a throttle position signal in response to a detected change in airflow; and Monitoring an airflow variation compensation value that corresponds to a setting level of the control signal; characterized by Generating a metric for the degree of coking of the throttle body (116) based on at least a plurality of residual error values ​​and the airflow variation compensation value; and Perform at least one reaction action based on the metric of the degree of coking of the throttle body (116) that exceeds a predetermined threshold.

Description

TECHNICAL FIELD

[0001] The present invention relates to a method for performing a prognosis for an air flow management system for an internal combustion engine according to the preamble of claim 1, as is essentially known from DE 103 00 592 A1.

[0002] With regard to the further state of the art, reference is made to the documents DE 10 2007 041 224 A1 and DE 10 2008 014 062 A1. INTRODUCTION

[0003] Certain control systems monitor sensor data associated with a corresponding vehicle system and can diagnose faults in these sensors. Such technology is reactive and may be limited to the current state without assessing the severity of the fault or predicting sensor degradation. This may result in the control system being unable to predict the future condition of the sensors and / or their remaining useful life.

[0004] The invention is based on the object of providing a method for carrying out a forecast for an air flow management system for an internal combustion engine, which can react sensitively to measurement errors. SUMMARY

[0005] This object is achieved by a method having the features of claim 1.

[0006] An airflow management system for an internal combustion engine includes an intake section for receiving ambient air and a mass flow sensor (MAF) configured to sense the mass flow of air directed through the intake section. The airflow management system also includes a throttle body including a valve for selectively restricting airflow from the intake section and a throttle position sensor (TPS) that senses an opening value of the throttle body. The airflow management system further includes an intake manifold in fluid communication with the throttle body and configured to direct airflow to each of a plurality of combustion cylinders, and a manifold air pressure (MAP) sensor configured to sense air pressure at the intake manifold.The airflow management system further includes a controller programmed to monitor signals from each MAF sensor, TPS, and MAP sensor and generate at least one residual error value based on a difference between a model-based value and a corresponding monitored signal. The controller is also programmed to generate an airflow variation compensation (AVC) value in response to a deviation between an actual throttle valve open area and a target throttle valve open area and to generate a throttle body coking metric value based on the at least one residual error value and the AVC value. The controller is further programmed to cause at least one responsive action in response to the throttle body coking level exceeding a predetermined threshold.

[0007] A prognostic system is provided for an airflow management system for an engine, including a mass air flow (MAF) sensor configured to sense the mass flow of air flowing through an intake portion, a throttle position (TPS) sensor configured to sense an opening degree of a throttle body downstream of the intake portion, and a manifold air pressure (MAP) sensor configured to sense the air pressure at an intake manifold downstream of the throttle body. The prognostic system includes a controller programmed to receive signals from each of a group of sensors, including at least the MAF sensor, TPS, and the MAP sensor.The controller is also programmed to adjust a throttle position control signal in response to a detected airflow variation and generate an airflow variation compensation (AVC) value corresponding to a degree of adjustment of the valve position control signal. The controller is further programmed to store at least one mathematical model for estimating throttle body contamination in memory based on the signals received from the array of sensors and the AVC value. BRIEF DESCRIPTION OF THE DRAWINGS Fig. Figure 1 is a schematic diagram of an airflow management system for an internal combustion engine. Fig. 2A to Fig. Figure 2C illustrates a mathematical model for each of a variety of sensor values. Fig. Figure 3 is a diagram of an algorithm for calculating a variety of residual error values. Fig. Figure 4 is a fault table associated with a variety of sensor faults of an airflow management system. Fig. 5A to Fig. 5D are plots of sensor residual error values ​​for a MAF sensor failure. Fig. 6A to Fig. 6D are plots of sensor residual error values ​​for a MAP sensor error. Fig. 7A to Fig. 7D are plots of sensor residual error values ​​for a TPS error. Fig. 8A Fig. 8B are flowcharts of a forecasting algorithm for an air management system. Fig. Figure 9 is a schematic diagram of an algorithm for calculating the coking degree metric of a throttle body. Fig. 10 is a representation of the metric of a throttle body coking level over multiple driving cycles of a vehicle. Fig. Figure 11 is a flowchart of a throttle body coking prediction algorithm. DETAILED DESCRIPTION

[0008] With reference to Fig. 1, the air management system 100 for an engine provides an airflow that supports combustion during engine operation. The engine includes a block with a number of combustion cylinders 102, within which a reciprocating cylinder operates according to the timing of the combustion cycle. The engine may be a gasoline engine, diesel engine, or any other type of engine requiring a metered airflow to regulate combustion conditions. Air is supplied to each of the cylinders 102 via an intake manifold 104, which includes a plurality of skids to direct air to each cylinder. An air duct 106 is arranged to supply fresh air from the ambient air to the engine as needed. The air is drawn in at an inlet 108 along a direction 110. An air filter 112 may be provided to reduce particulates and other contaminants in the engine airflow.

[0009] The air intake system 100 includes a variety of sensors that collect data about various airflow conditions to optimize combustion. A mass air flow (MAF) sensor 114 is located near the air intake section and outputs a signal indicative of the mass flow rate of air flowing through the air passage 106. The mass air flow information is used to deliver the correct mass of air to the engine, corresponding to the amount of fuel intended for combustion, according to the engine's power demand. The density of the intake air can vary as it expands and contracts with temperature and pressure. Thus, the MAF 114 may also include an intake air temperature (IAT) sensor, which provides thermal data as input to adapt to such temperature changes. The output signal of the MAF sensor enables precise control of the engine's air-fuel ratio.

[0010] The air intake system 100 also includes a throttle body 116 for regulating the amount of air directed from the intake 108 to the intake manifold 104. The throttle body includes a variably positionable air control valve for varying the size of the airflow restriction and thus measuring the amount of air directed to the intake manifold. In some examples, the valve of the throttle body 116 is a throttle plate having a rotating disc 118 that changes angular position between a first position substantially aligned with the airflow along the passage (e.g., direction 110) and a second position substantially perpendicular to the airflow along the passage. A throttle position sensor (TPS) 120 outputs a signal indicative of the angular position of the disc 118.For example, although a throttle valve is provided, other devices capable of regulating variable airflow according to engine requirements may also be suitable according to the present invention. Throttle body 116 may further include a throttle intake air pressure (TIAP) sensor for outputting signals indicative of the air pressure into the throttle valve.

[0011] Depending on the amount and conditions of air directed to the intake manifold, the air directed to the cylinders 102 to support combustion may vary in pressure. Additionally, devices that force the air, such as turbocharger compressors, may further affect the pressure of the air flowing through the air management system 100. A manifold absolute pressure (MAP) sensor 122 monitors the air pressure in the intake area of ​​the manifold. According to some examples, the MAF sensor provides an open-loop signal to a controller 124 to predict airflow information, while the MAP 122 provides closed-loop feedback to make minor corrections to the predicted air mass by adjusting the throttle position.Failures in the air intake system sensors can cause rough engine idle, engine hesitation, and suboptimal fuel consumption, resulting in a malfunction indication and / or reduced engine performance. Some common failure modes include contamination or corrosion-related degradation due to aging and sensor signal drift. According to aspects of the present invention, systematic diagnostic and prognostic methods are provided for detecting and predicting degradation of air management system sensor performance.

[0012] A controller 124 is provided to monitor and control the operation of the air management system 100. The controller 124 includes one or more digital computers, each including a microprocessor or central processing unit (CPU), read-only memory (ROM), random access memory (RAM), electrically programmable read-only memory (EPROM), a high-speed clock, analog-to-digital (A / D) and digital-to-analog (D / A) circuitry, input / output (I / O) circuitry and devices, and corresponding signal conditioning and buffering circuitry. The controller 124 also stores a number of algorithms or computer-executable instructions required to issue commands for performing actions in accordance with the present invention.

[0013] The controller communicates with a number of sensors to receive data indicative of the performance of various vehicle components. The controller 124 receives signals from each of the MAF 114, the TPS 120, and the MAP sensor 122. Other sensors, such as air temperature sensors and additional pressure sensors at various locations along the air management system 100, also output signals to the controller 124. Although each of the sensors is referred to in the singular, any number of sensors may be arranged at various locations to provide signals representative of the aforementioned data and other data.

[0014] Various elements of the controller 124 may also be located outside the vehicle, such as at a central processing location. In particular, some components and / or functions of the controller 124 may be located / performed within the vehicle, and other components and / or functions of the controller 124 may be located remotely from the vehicle, with data transferred therebetween as necessary. In these cases, the controller 124 may also be capable of wireless communication using an internal transceiver. The transceiver may be configured to exchange signals with a number of external components or systems. The controller 124 is further programmed to exchange information using a wireless communication network 126.Thus, the controller 124 may be in wireless communication with an external server 128 that performs at least some of the processing described in the present invention. In further examples, the controller 124 periodically uploads measurement data to the server 128, which stores aggregated data, performs data analysis, and generates forecast messages. The use of an external server may help reduce the need for data processing and storage in the vehicle. The server may store one or more model-based calculation algorithms, discussed in more detail below, to facilitate the air management system's forecasting.

[0015] The controller 124 may also store one or more algorithms representing mathematical models of various physical aspects of the operation of the air management system 100. These mathematical models of the operation of the air management system 100 may be used to predict system performance. Model-based assessments of system health may be performed using foundational mathematical models. That is, input signals received by the controller may be recognized to illustrate certain signature system behaviors associated with known faults, such as component degradation or impending failure. As mentioned above, known faults associated with the air management system include, but are not limited to, contamination, corrosion, and sensor signal drift.

[0016] In some examples, parity equations are used to refine the monitoring and control of the air management system. Model-based estimates of certain operating values ​​are generated during vehicle operation with given fixed parameters. The difference between the measured outputs and the model-based estimate outputs should be close to zero under ideal conditions. In the event of a fault, one or more process behaviors differ from the model-based behavior because the models are structured to mimic fault-free cases. The deviations can be determined, for example, using transfer functions or using state-space formulations. A certain residual deviation can be chosen so that the deviation values ​​are only affected by certain types of faults to be detected.The deviations may vary continuously, at least due to fluctuations in the raw input data and modeling errors. To overcome the fluctuations and errors, characteristics of deviations are derived to eliminate the influence of noise and reduce the overall data burden. Depending on the difficulty of detecting a particular error, the associated deviation may be calculated at a unique sampling rate and / or have unique sensitivity to other deviation types associated with different error types. In some examples, thresholds against which the deviations are compared may be adaptive thresholds. That is, a threshold may be automatically set based on the character of the input data (e.g., rate of change of the input data, direction of the trend of the input data, the shape of the change function of the input data).In general, the arrangement of deviations is chosen to sensitize the deviations to disturbances and at the same time to make them robust against disturbing effects.

[0017] The controller may store algorithms in memory that include mathematical models for a number of different system attributes. With reference to Fig. 2, a throttle model 200 describes the flow through the throttle body and is used to estimate the mass air flow through the throttle body as a function of ambient air pressure, estimated MAP, throttle position, and intake air temperature. The throttle model is quasi-steady and uses a first-order lag filter to model dynamic airflow effects through the throttle body. The throttle model utilizes an effective flow area of ​​the throttle body as a function of the signals output from the TPS, IAT, and TIAP sensors.

[0018] The following equation (1) represents an exemplary function for estimating the mass air flow through the throttle. MA^Ft=maflag∗TIAPt∗ψ∗AeffR∗(IAT+273.15)+(1−maflag)∗MA^Ft−1 maflag represents the first-order lag filter and can be varied in a range from 0 to 1 to control a weighting between the current parameters that determine the estimation of the mass air flow MÂF t and influence the previous estimate of the mass air flow. MÂF t-1 . TIAPt represents the inlet absolute pressure of the throttle valve, R represents an ideal gas constant for air (e.g. about 287 m 2 / (s 2 *K). A effrepresents the effective open area of ​​the throttle body based on the signal from the TPS. ψ is a compressible flow function for air flowing through the air management system. ψ may depend at least in part on the pressure ratio across the throttle, Pr, where Pr is based on the throttle inlet pressure relative to the pressure at the manifold inlet pressure. According to some examples, ψ = 0.685 for Pr < 0.5283. For 0.5283 ≤ Pr ≤ 1.0, ψ is defined by the following equation (2). ψ=7∗(Pr1,428−Pr1,714)

[0019] Based on Fig. 2, a first intake manifold model 220 describes the intake manifold and is used to estimate MAP as a function of the mass flows into the manifold (from throttle and exhaust gas recirculation (EGR)) and the mass flows out of the manifold caused by engine pumping. The intake manifold model is also quasi-steady and accounts for the dynamics of the manifold by integrating the effect of small incremental flow changes over time. The flow into the manifold from the throttle is estimated using the throttle model. The engine flow model uses a model to determine volumetric efficiency and relies on the intake manifold model to properly account for the influence of altitude, camshaft timing, and cylinder deactivation on volumetric efficiency.

[0020] The following equation (3) represents a first exemplary function for determining MAP MA^P1t=MA^P1t−1+[Δt∗Tcharge∗R∗(MA^Ft−1+EG^Rt−EF^Rt)Volintake] represents Δt stands for a loop execution time (e.g., t ≤ 0.1 sec). Vol intake stands for the volume of the intake manifold in cm 3 (e.g. as determined during vehicle calibration). T charge stands for the charge temperature and is included to account for the air density in the intake manifold and the influence of the EGR flow on the temperature of the gas in the intake manifold. EGR t stands for the EGR flow and EF̂R t for the engine flow rate.

[0021] With further reference to Fig. 2, a second intake manifold model 240 is similar to the first intake manifold model 220 described above, but the MAF sensor data is used directly in the model instead of the throttle model for the throttle air intake. The following equation (4) represents a second exemplary function for determining MAP. MA^P2t=MA^P2t−1+[Δt∗Tcharge∗R∗(MAFt−1+EG^Rt−EF^Rt)Volintake]

[0022] As mentioned above, a previous value of the MAF sensor MAF t-1 used as input in equation (4), compared with a calculated value MÂF t-1 in the first example of the MAP function of equation (3).

[0023] The residual values ​​for the modeled air intake system values ​​are calculated based on differences between sensor-independent estimates and the actual measured values. The modeled estimates of MAF, MAP1, and MAP2 from the models described above are compared with the actual measured values ​​of one or more sensors. The comparison generates three residual error values, a first residual error value, MAF rt according to the MAF sensor signal from the throttle model is calculated according to the following equation (5). MAFrt=MAF−MA^Ft

[0024] A second residual error value, MAP1 rt , and a third residual error value, MAP2rt, corresponding to the MAP sensor signals of the first and second intake manifold models, respectively, are calculated according to the following equations (6) and (7). MAP1rt=MAP−MA^P1t MAP2rt=MAP−MA^P2t

[0025] A fourth residual error value TPS rt , corresponding to the TPS signal is generated by multiplying the first residual error of the MAF signal by the second residual error of the MAP signal according to the following equation (8). TPSrt=MAFrt∗MAP1rt

[0026] Based on the behavior of the residual error values, a condition assessment of the air management system can be developed, as well as the identification of specific types of sensor errors. The values ​​are tracked over time and can relate to an error in the signals emitted by one or more sensors. With reference to Fig. 3, a system 300 is provided for generating and storing residual values ​​of the air management system when stable operating conditions are detected (e.g., during each drive cycle). The residual values ​​resulting from the sensor signal error are recorded and stored in memory. The controller may be programmed to monitor at least one operating parameter that may indicate the presence of a stable operating condition. In some examples, stability may be characterized by a variance of less than a threshold range over a predetermined period of time.

[0027] With further reference to Fig. 3, the operating range 302 may correspond to a specific set of stable conditions represented by parameter 1 up to parameter m. For example, the operating range 302 may represent driving at a constant speed. In particular, parameter 1 may correspond to a throttle position and parameter 2 to an engine speed. In some further examples, idle and / or coasting conditions may have sufficient stability to assess the sensor error. The controller may be programmed to monitor other parameters up to parameter m, where the set of parameters together signify the presence of constant driving within the operating range 302. According to the example of Fig. 3 Each of the individually monitored parameters must be stable for the operating state to be considered stable. In some alternative examples, a subset of fewer than all monitored parameters may be appropriate for the entire operating range to be considered stable. The algorithm can be configured to monitor a specific number of stable parameters simultaneously. In further alternatives, the algorithm can be arranged to monitor a predefined list of specific parameters.

[0028] In response to the detection of a stable operating condition, the controller is programmed, as described above, to calculate and store a variety of residues. In the example of Fig. 3, each MAF residual, MAP1 residual, MAP2 residual, and TPS residual is calculated and stored in a measurement log 304. In addition to the raw data values ​​for other operating parameters, the residual values ​​can also be stored in a memory.

[0029] The controller can be further programmed to monitor various stable operating ranges according to other operating conditions. With reference to Fig. 3, the controller is programmed to monitor parameter 1 through parameter k for the parameter set that characterizes the stable conditions for operating region 306. Parameters 1 through k may be the same as or different from parameters 1 through m of operating region 302. In one example, operating region 306 corresponds to a second steady-state driving condition during the same drive cycle. In further examples, different stable operating conditions may correspond to different vehicle maneuvers or functions (e.g., idling in park mode, idling while driving, driving the vehicle, etc.). A particular set of monitored parameters for particular operating regions may include a particular set of parameters compared to other operating regions. During a particular drive cycle, the controller may detect any number of operating regions based on the presence of stable parameter sets.

[0030] As already explained, residual values ​​and other relevant parameters are stored via a memory on the controller. This data can also be transferred to an external server configured to store and process the data for maintenance status, vehicle inventory data aggregation, and other monitoring functions.

[0031] According to aspects of the present invention, the air management algorithm is designed to track the progression of residual errors across the operating range. Over time and multiple drive cycles, both the source and direction of the error may point to one or more specific sensors. As the residual error magnitudes approach predefined thresholds for various error severities, the progressive trends provide an early indication of a potential error condition and intensity. The controller may store data for each of the tracked residual error values ​​according to plot 310. The horizontal axis 312 represents drive cycles, and the vertical axis 314 represents a specific error. Curve 316 represents an exemplary residual value plotted over time. In the example of plot 310, the residual value increases in a positive direction as the sensor ages and the error increases.As explained in more detail below, a given sensor may also exhibit residual displacements in a negative direction depending on the sensor type and failure mode.

[0032] The air management algorithm may include storing any number of thresholds indicating a state of health relative to the plotted residual curve 316. Values ​​less than a first threshold 318 may be indicative of nominal operating conditions at which the sensor can be considered fully intact and the controller may not respond. In response to the residual value increasing to a range between the first threshold 318 and a second threshold 320, the sensor may be assumed to have a low severity fault. In this case, the controller may issue a first prognostic message to provide an indication of the degraded sensor condition. The first prognostic message may include a severity indicator and may continue to be issued while the sensor operates in the state range between the threshold 318 and the threshold 320.In at least one example, the prediction message is transmitted to a non-vehicle diagnostic server external to the vehicle.

[0033] The residual value further increases the occurrence of certain combustion faults, as discussed above, and may result in an increase in frequency and / or severity. If the residual value 316 increases above the second threshold 320, the controller may issue a warning message indicating an impending failure of the sensor. The impending fault message may persist while the sensor operates in a range of states between the second threshold 320 and a third threshold 322. The impending fault message may act as a different severity indicator and include increased urgency relative to the first prognostic message. Additionally, the impending failure message may have a different audience than the first prognostic message.

[0034] If the residual value increases above the third threshold 322, the controller may determine that the sensor has failed based on the degree of increase in the residual error. In this case, the sensor's performance may be degraded to such an extent that a notification for urgent vehicle maintenance may be issued to prevent powertrain shutdown due to malfunctions in the air management system. A multi-level prognostic notification system, as described herein, may provide different information about the sensor's condition at different stages of its service life. Each of the different prognostic notifications may serve as a severity indicator depending on the operating conditions. The prognostic system may also provide a vehicle owner with information to proactively schedule vehicle service before the vehicle actually fails.

[0035] With reference to Fig. 4, Table 400 is a visual representation of a "reasoning" algorithm that analyzes combinations of symptom characteristics to isolate specific air management sensor failures. The fault table 400 associates a predetermined fault type with each of several combinations of fault signature components. Column 402 represents a set of sensors to be monitored (e.g., MAF sensor, MAP sensor, TPS sensor, etc.). Column 404 identifies a series of sensor failure mode types related to shifts in residual error. Each failure mode type may or may not include complete sensor failure, with each fault signature based on symptoms of degradation in the performance of specific sensors. Columns 406, 408, 410, and 412 represent residual error value trends for each of the MAF residual values, MAP1 residual values, MAP2 residual values, and TPS residual values.Lines 414, 416, 418, 420, 422, and 424 correspond to unique fault signatures that indicate specific sensor failure modes. Each fault signature contains a unique combination of residual fault trends, capable of indicating component degradation before a reduction in performance is perceived by the driver. Although six different fault types are provided as an example, any number of faults can be specified and assigned to a specific fault signature.

[0036] As discussed above, data for each of the residual error trends can be continuously collected during a drive cycle. Depending on the behavior of the individual residual error values, certain trend combinations relate to fault signatures that indicate specific sensor shifts. The unique signatures allow the controller to proactively isolate specific sensor faults before the fault is perceived by the driver. Residual error trends designated with "(+)" refer to a positive directional residual error trend. Similarly, residual error trends designated with "(-)" refer to a negative directional residual error trend.

[0037] In a first example, as shown in line 414, a positive direction of the MAF residual error simultaneously with a negative direction of the MAP2 residual error indicates a positive output shift of the MAF sensor.

[0038] In a second example, as shown in line 416, a negative direction of the MAF residual error simultaneously with a positive direction of the MAP2 residual error indicates a negative output offset of the MAF sensor.

[0039] In a third example, as shown in line 418, a positive direction of the MAP1 residual error simultaneously with a positive direction of the MAP2 residual error indicates a positive output shift of the MAP sensor.

[0040] In a fourth example, as shown in line 420, a negative direction of the MAP1 residual error simultaneously with a negative direction of the MAP2 residual error indicates a negative output shift of the MAP sensor.

[0041] In a fifth example, as shown in line 422, a negative directional trend of the MAF residual error, along with a negative directional trend of the MAP1 residual error, indicates a positive output shift of the TPS sensor. And, as mentioned in a previous example, the TPS residual is based on a multiplication of the MAF residual and the MAP1 residual. Thus, the TPS residual also shows a positive directional trend, consistent with the positive shift of the TPS sensor values.

[0042] In a sixth example, as shown in line 424, a positive directional slope of the MAF residual error concurrent with a positive directional slope of the MAP1 residual error indicates a negative output shift of the TPS sensor. Since the TPS residual is based on a multiplication of the MAF residual and the MAP1 residual, the TPS residual also exhibits a positive directional slope, which coincides with the negative shift of the TPS sensor values.

[0043] With reference to the Fig. 5A to Fig. 5D, a series of charts are provided to graphically display the residual error trends associated with a variety of failure modes. Plot 500 provides MAF residual error along vertical axis 502. MAF sensor drift is displayed along horizontal axis 504, with no-error setpoints near the center, positive shifts toward the right of the plot, and negative shifts toward the left of the plot. Curve 506 represents MAF residual error behavior in the presence of MAF sensor drift. Plot 520 provides MAP1 residual error along vertical axis 508 and MAF sensor drift along horizontal axis 504. Curve 510 represents MAP1 residual error behavior in the presence of MAF sensor drift. Plot 540 provides MAP2 residual error along vertical axis 512 and MAF sensor drift along horizontal axis 504.Curve 514 represents the MAP2 residual error behavior in the presence of MAF sensor drift. Plot 560 provides TPS residual error along vertical axis 516 and MAF sensor drift along horizontal axis 504. Curve 518 represents the TPS residual error behavior in the presence of MAF sensor drift.

[0044] A relative comparison of plots 500, 520, 540, and 560 reveals that the trend behavior of certain residual error values ​​is consistent for certain MAF sensor shifts. For example, in the presence of a positive directional change of the MAF sensor, the MAF residual 506 is shifted in an increasing direction and the MAP2 residual 514 is shifted in a decreasing direction. At the same time, both the MAP1 residual 510 and the TPS residual 518 remain within the rated operating ranges. These conditions generally correspond to the failure mode of row 414 of the aforementioned fault table 400.

[0045] Conversely, in the presence of a negative direction MAF sensor shift, the MAF residuals 506 are shifted in a decreasing direction and the MAP2 residuals 514 are shifted in an increasing direction. At the same time, both the MAP1 residual 510 and the TPS residual 518 remain within the rated operating ranges. These conditions generally correspond to the failure mode of line 416 of the aforementioned fault table 400.

[0046] According to some examples, a history of a particular residual error value is detected by the air management algorithm when the residual error value exceeds a predetermined severity level. Conversely, if the residual error values ​​of a particular monitored value remain within a predetermined range of normal operating limits, the monitored value is assumed to be in a normal maintenance state and no forecast message is generated. With further reference to the Fig. 5A through 5D, multiple progressive severity levels are provided for all illustrations. A first set of thresholds 522 indicates when a residual error value exits a nominal operating range into a low severity state. In response, the algorithm may issue a corresponding low severity prognosis message indicating the condition of the sensor and reminding the user and / or service technician of an upcoming maintenance schedule. A second set of thresholds 524 is used to determine when the failure mode condition deteriorates from a low severity state to a medium severity state. In response, the algorithm may issue a corresponding medium severity prognosis message indicating the deterioration condition and illuminate a fault indicator or other persistent warning message.If a residual error value exceeds a third set of thresholds 526, the algorithm includes determining that the failure mode is in a high-severity state, including an imminent or present sensor failure. In response, the algorithm may include engaging a limp-home or other reduced engine operating modes to protect components from damage caused by air management system fault conditions. The first set of thresholds 522, the second set of thresholds 524, and the third set of thresholds 526 may be set to a different failure magnitude for each of the various monitored residual error values.While the exemplary ranges are bounded by a pair of thresholds symmetrically spaced around a nominal value of zero, it should be noted that each of the different thresholds may be asymmetrically spaced around a non-zero value.

[0047] With reference to the Fig. 6A to Fig. 6D, a series of charts are provided to graphically display the residual error trends associated with a variety of failure modes. Plot 600 provides MAF residual error along vertical axis 602. MAP sensor drift is displayed along horizontal axis 604, with no-error setpoints near the center, positive shifts toward the right side of the plot, and negative shifts toward the left side of the plot. Curve 606 represents MAF residual error behavior in the presence of MAP sensor drift. Plot 620 provides MAP1 residual error along vertical axis 608 and MAP sensor drift along horizontal axis 604. Curve 610 represents MAP1 residual error behavior in the presence of MAP sensor drift. Plot 640 provides MAP2 residual error along vertical axis 612 and MAP sensor drift along horizontal axis 604.Curve 614 represents the MAP2 residual error behavior in the presence of a MAP sensor shift. Plot 660 provides TPS residual error along the vertical axis 616 and MAP sensor drift along the horizontal axis 604. Curve 618 represents the TPS residual error behavior in the presence of a MAP sensor shift.

[0048] A relative comparison of plots 600, 620, 640, and 660 reveals that the trend behavior of certain residual error values ​​is consistent for certain MAP sensor shifts. For example, in the presence of a positive directional change of the MAP sensor, the MAP1 residual error 610 is shifted in an increasing direction and the MAP2 residual error 614 is shifted in a decreasing direction. At the same time, both the MAF residual error 616 and the TPS residual error 618 remain within the nominal operating ranges. These conditions generally correspond to the failure mode of row 418 of the aforementioned fault table 400.

[0049] Conversely, in the presence of a negative directional MAP sensor offset, the MAP1 residual error 610 tends to decrease, and the MAP2 residual error 614 also tends to decrease. At the same time, both the MAF residual 606 and the TPS residual error 618 remain within their rated operating ranges. These conditions generally correspond to the failure mode of row 420 of the aforementioned fault table 400.

[0050] With reference to the Fig. 7A to Fig. 7D, a series of charts are provided to graphically display the residual error trends associated with a variety of failure modes. Plot 700 provides MAF residual error along the vertical axis 702. TPS sensor drift is displayed along the horizontal axis 704, with no-error setpoints near the center, positive shifts toward the right side of the plot, and negative shifts toward the left side of the plot. Curve 706 represents MAF residual error behavior in the presence of TPS sensor drift. Plot 720 provides MAP1 residual error along the vertical axis 708 and TPS sensor drift along the horizontal axis 704. Curve 710 represents MAP1 residual error behavior in the presence of TPS sensor drift. Plot 740 provides MAP2 residual error along the vertical axis 712 and TPS sensor drift along the horizontal axis 704.Curve 714 represents the MAP2 residual error behavior in the presence of a TPS sensor shift. Plot 760 provides TPS residual error along the vertical axis 716 and TPS sensor drift along the horizontal axis 704. Curve 718 represents the TPS residual error behavior in the presence of a TPS sensor shift.

[0051] Similar to the previous examples, a relative comparison of plots 700, 720, 740, and 760 reveals that the trend behavior of certain residual error values ​​is consistent in the presence of TPS sensor shifts. For example, in the presence of a positive TPS sensor shift, the MAF residual 706 tends to decrease, and the MAP1 residual 710 also tends to decrease. Since the TPS residual 718 is based on a multiplication of the MAF residual 706 by the MAP1 residual 710, the TPS residual 718 exhibits a positive trend due to a positive directional shift of the TPS sensor values. At the same time, the MAP2 residual 714 remains within the rated operating ranges. These conditions generally correspond to the failure mode of row 422 of the aforementioned fault table 400.

[0052] Conversely, with a negative directional shift in the TPS sensor values, the MAF residual 706 is trending upwards, and the MAP1 residual 710 is also trending upwards. And since the TPS residual 718 is based on multiplying the MAF residual 706 by the MAP1 residual 710, the TPS residual 718 also exhibits a positive trend due to a negative directional shift in the TPS sensor values. At the same time, the MAP2 residual 714 remains within the rated operating ranges. These conditions generally correspond to the failure mode of line 424 of the aforementioned fault table 400.

[0053] With reference to Fig. 8, method 800 illustrates an algorithm for detecting shifts in air management system sensor data and performing a sensor health prediction based on residual trends. In step 802, the algorithm includes collecting residual error data for a plurality of air management system sensors in response to detecting steady-state operating conditions. As discussed above, this may occur multiple times over a single drive cycle and / or across multiple different drive cycles. The sensors to be monitored may include at least one MAF sensor, one MAP sensor, and one TPS. Information from additional sensors that output air health data may also be monitored and correlated with sensor shifts and / or changing health behavior.

[0054] At step 804, the algorithm includes detecting whether at least two sensor residual error values ​​exhibit trends that deviate from the nominal operating ranges. If at step 804, fewer than two sensor residual values ​​exhibit such deviations, the algorithm includes returning to step 802 and continuing to monitor the sensor residual values.

[0055] If at step 804, at least two residual error values ​​exhibit trends that deviate from the nominal operating ranges, the algorithm includes assessing which residual error exhibits trends and generating predictive determinations based on at least the residual error types, directionality, and severity. If at step 806, both the MAP1 and MAP2 residual error values ​​exhibit a deviation outside a predetermined nominal operating range, the algorithm includes determining at step 808 that the MAP sensor may be experiencing a reading drift.

[0056] At step 810, the algorithm includes assessing the magnitude of the residual error trends. Specifically, the algorithm includes determining whether the MAP1 residual error or the MAP2 residual error is approaching a threshold. As discussed above, a number of different thresholds may be arranged depending on the severity of the residual error values ​​and the proximity of impending failure of one or more sensors. Depending on the notification scheme, step 812 may include generating and sending a severity message indicating MAP sensor degradation. In some examples, this severity is sent for all residual errors outside of the nominal operating ranges. In further examples, lower deviations are recorded and stored to monitor operating status, and sensor degradation messages are only sent in response to residual errors that exceed higher thresholds (e.g.,medium or high severity thresholds).

[0057] If at step 806 both the MAP1 and MAP2 residual values ​​do not deviate from the nominal operating ranges, the algorithm includes assessing whether both the MAF residuals and the MAP residual values ​​deviate from the nominal operating conditions at step 814. If both values ​​exhibit sufficient deviation, the algorithm includes determining at step 816 that the MAF sensor may be experiencing a reading drift.

[0058] At step 818, the algorithm includes assessing the magnitude of the residual error trends. Specifically, the algorithm includes determining whether the MAF residual error or the MAP2 residual error is approaching a threshold. As discussed above, a number of different thresholds may be arranged depending on the severity of the residual error values ​​and the proximity of impending failure of one or more sensors. Depending on the notification scheme, step 820 may include generating and sending a severity notification indicating MAF sensor degradation.

[0059] If at step 814 both the MAF and MAP2 residual values ​​do not deviate from the nominal operating ranges, the algorithm includes assessing at step 822 whether both the MAP1 residual values ​​and the TPS residual values ​​deviate from the nominal operating conditions. If both values ​​exhibit sufficient deviation, the algorithm includes determining at step 824 that the TPS sensor may be experiencing a reading drift. As described above, the relationship between the TPS residual, the MAF residual, and the MAP1 residual is chosen such that various combinations may be suitable for predicting the TPS sensor. In particular, either a MAF-TPS residual error combination or a MAP1-TPS residual error combination may be sufficient to assess the health of the TPS sensor.

[0060] At step 826, the algorithm includes assessing the magnitude of the residual error trends. Specifically, the algorithm includes determining whether the MAP1 residual error or the TPS residual error is approaching a threshold. As discussed above, a number of different thresholds may be arranged depending on the severity of the residual error values ​​and the proximity of impending failure of one or more sensors. Depending on the notification scheme, step 828 may include generating and sending a severity notification indicating TPS sensor degradation.

[0061] According to further examples, combinations of residual failure trends can be used to predict other components of the air management system. For example, throttle valve degradation and failure prediction can be determined based on the behavior of the MAP residual, MAF residual, and / or TPS residual. By analyzing the residual values, throttle deposit problems can be detected before they affect engine performance. "Coking" can be caused by burnt oil depositing on surfaces and can lead to flow-restricted passages. Mixtures of soot and oil as part of the combustion process. If these deposits are significant enough, they can effectively restrict airflow, causing the engine to lag and stall. The engine control module is initially programmed with a desired airflow versus the throttle position table based on a new throttle body.As deposits build up in the throttle body, reduced airflow can cause problems if the engine control module fails to correctly 'learn' new airflow characteristics relative to throttle position values. It may be desirable for the control module to relearn the amount of airflow through a restricted throttle body and, over time, adapt to coking by accurately increasing the amount of bypass airflow.

[0062] As explained above, airflow across the throttle body can be measured and estimated using various sensors. Based on the accumulated residual errors, the algorithm compensates for variations in the MAF sensor, the TPS sensor, the MAP sensor, and throttle body deposits, such as coking. Symptoms associated with throttle body coking may include, in particular, reduced engine idle speed and / or engine stalling. In this case, there are no direct diagnostic codes corresponding to throttle body contamination to trigger engine calibration changes to adjust and "relearn" idle conditions.

[0063] The air management system controller may be programmed to generate throttle body airflow variation (AVC) compensation to alter a relationship between throttle position and open flow area. This compensation may be provided to account for accumulated residual errors. According to some examples, the AVC value is generated in response to a deviation between an actual open area and a desired throttle open area. Compensation may be achieved by adjusting a throttle position signal, where the AVC value may be expressed as a percentage correction. Consequently, the AVC adjustment compensates for airflow restrictions in the throttle body by changing an effective airflow area to achieve a desired opening.The AVC percentage is calculated based on the cumulative MAF residual errors, which can also be a ratio between the detected error and the estimated error. To minimize MAF residual errors, the relationship between throttle position and effective flow area can be adjusted so that the actual throttle area is larger or smaller than the desired throttle area to achieve the desired airflow. The AVC value is then calculated by dividing the amount of compensation provided by the amount of total allowable compensation.

[0064] The required adjustment level and the corresponding AVC value can provide clues to throttle body deposits, such as coking. According to some examples, an air management algorithm involves tracking a history of airflow variation compensation and the sensor's residual errors over time (depending on the operating range). As the residual error magnitudes approach predefined thresholds (at various error severities), the progressive trends provide an early warning of a potential throttle body coking condition.

[0065] With reference to Fig. 9, a graph 900 shows an algorithm for calculating a throttle body coking severity score. Each of the TPS residual error, MAF residual error, and MAP residual error values ​​is converted to a normalized value.

[0066] According to at least one example, each residual is normalized to a percentage error representation relative to the data output that represents a maximum error with respect to each individual sensor according to equations (9), (10), and (11) below. TPS%=TPSrtTPS Remaining maximum limit MAF%=MAFrtMAF Remaining maximum limit MAP%=MAPrtMAF Remaining maximum limit

[0067] With further reference to Fig. 9, the TPS residual error is normalized at 902, the MAF residual error is normalized at 904, and the MAP residual error is normalized at 906. Each of the normalized error values ​​can be used in conjunction with the AVC value to calculate a metric of the throttle body coking degree at 908. Equation (12) below is an exemplary equation for calculating the coking degree. Degree of coking=β∗log(w1∗AVC%)∗log(w2∗min{TPS%,100})∗log(max[w3∗min{MAF%,100},w4∗min{MAP%,100}])

[0068] In equation (11) above, β represents a scaling factor that can be used to normalize the overall coking level value within a desired range. In some examples, β is based on the vehicle's mileage and / or the AVC value. Multiple weighting factors w1, w2, w3, and w4 can be applied to the normalized residual error values, depending on the desired sensitivity of the parameter for accurate coking detection. In some examples, the overall coking level ranges from 0 to 100. Due to the coking of the throttle body over time, the residual error values ​​increase with the age of the vehicle.

[0069] With reference to Fig. 10, plot 1000 is an example of the temporal progression of a coking degree. The vertical axis 1002 quantifies the value of the coking degree of a throttle body. The horizontal axis 1004 denotes the vehicle's driving cycles, with the number of driving cycles being zero on the left side of the plot and the number of driving cycles increasing from the left to the right section of the plot. Curve 1006 denotes the value of the coking degree, which, according to the exemplary plot, increases as a function of the number of driving cycles. A certain degree of coking of the throttle body is permissible without affecting the operation of the air management system. However, tracking the history of the residual error values ​​via the throttle coking metric can help predict the presence of throttle body coking.

[0070] A first operating range 1008 may include a range of coking level values ​​1006 where air management system operation remains within an acceptable range and no action is taken in response to throttle body coking. A second operating range 1010 denotes a low coking level, which may reflect the onset of contaminant accumulation on the throttle body. An exemplary response to coking level values ​​within the second operating range 1010 may include uploading stored residual and coking level measurements to an external server for further predictive analysis and / or comparison with data corresponding to other vehicles, such as within a shared fleet.

[0071] A third operating range 1012 denotes a moderate coking severity, which may trigger an incrementally more urgent response. For example, an algorithm may be configured to issue an initial warning indicating the need to clean the throttle body when the vehicle is serviced as part of a maintenance schedule or repair of other vehicle components. Specifically, the coking level may be integrated with oil life monitoring and / or other maintenance schedules and routine maintenance checks. A passive service flag may be issued for inspection and maintenance of the throttle body depending on the coking level value. In this way, a service technician may receive an instruction for preventative cleaning of the throttle body.

[0072] A fourth operating range 1014 indicates a high coking severity, which may trigger a further, more incremental response. For example, a warning indicator may be activated in the vehicle or a message may be sent to a user device indicating an active throttle body maintenance requirement. A maintenance status may also be output to a user to provide an indication of the remaining throttle body life related to the coking condition.

[0073] A fifth operating range 1016 denotes a critical emergency coking condition that may trigger a high-level response. For example, an imminent failure service notification may be communicated to a user or vehicle service provider. Furthermore, a control algorithm may involve entering a mode with reduced engine functionality (e.g., a low-power mode) to prevent or mitigate damage to the engine or air management system.

[0074] With reference to Fig.11, method 1100 illustrates an algorithm for performing throttle body coking condition prediction. At step 1102, the algorithm includes collecting AVC values ​​and residual error values, as described above, under various vehicle operating conditions. If the AVC is less than a threshold Th1 at step 1104, the algorithm remains in normal operation and continues to monitor AVC values ​​and residual error values ​​of the air management system.

[0075] If the AVC value is greater than the threshold Th1 at step 1104, the algorithm includes assessing whether the residual error values ​​deviate from the nominal ranges at step 1106. As explained above, for example, for a plurality of residual error values, corresponding thresholds are stored and characterized as limits for nominal operating ranges. If the residual error values ​​are within the nominal ranges at step 1106, this may indicate that the increase in the AVC value is due to a causal factor besides throttle body coking. Therefore, the algorithm includes returning to step 1102 and continuing to monitor AVC values ​​and residual error values ​​related to the air management system.

[0076] If, at step 1106, a number of residual error values ​​are outside the nominal ranges, the algorithm involves entering a throttle body prediction subroutine to assess the conservation status relative to the coking status. Step 1108 involves normalizing the error values ​​as explained above. At step 1110, the algorithm involves calculating a metric value for throttle body coking as a function of the AVC value and the residual error values.

[0077] Depending on the coking severity value relative to one or more thresholds, the algorithm involves creating a response action. As explained above, a plurality of severity thresholds may be stored, corresponding to a number of different response actions. If no threshold is met at step 1112, the algorithm involves returning to step 1108 and monitoring the normalized residual error values, and then recalculating the coking severity.

[0078] If the coking level metric value approaches a predefined threshold at step 1112, the algorithm includes mapping the current operating conditions to the corresponding threshold and / or operating range and determining the corresponding response action. In some examples, the response action corresponds to the nearest determined threshold. In further examples, the response is based on the coking level metric having a value within an operating range between two thresholds. After determining the appropriate response action at step 1114, the algorithm issues a command to execute one or more response actions at step 1116.

[0079] The processes, methods, or algorithms disclosed herein may be provided and / or implemented by a processing device, controller, or computer, which may include any existing programmable electronic control unit or a dedicated electronic control unit. Likewise, the processes, methods, or algorithms may be stored as data or executable instructions by a controller or computer in a variety of ways, including, without limitation, permanent storage on non-writable storage media, such as ROM, and as alterable information on writable storage media, such as floppy disks, magnetic tapes, CDs, RAM, and other magnetic and optical media. The processes, methods, or algorithms may also be implemented in a software-executable object.Alternatively, the processes, methods, or algorithms may be embodied in whole or in part with suitable hardware components, such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), state machines, controllers, or other hardware components or devices, or a combination of hardware, software, and firmware components.

Claims

[1] A method for performing a forecast for an airflow management system for an internal combustion engine, comprising: Setting a throttle position signal in response to a detected airflow change; and Monitoring an airflow variation compensation value corresponding to an adjustment degree of the control signal; characterized by Generating a metric for the degree of coking of the throttle body (116) based on at least a plurality of residual error values ​​and the airflow variation compensation value; and Performing at least one responsive action based on the metric of the degree of coking of the throttle body (116) exceeding a predetermined threshold. [2] The method of claim 1, wherein the at least one responsive action includes generating a severity indicator based on the coking metric of the throttle body (116) exceeding a predetermined severity threshold. [3] The method of claim 2, further comprising storing a plurality of progressive severity thresholds, wherein each of the thresholds corresponds to a unique set of response actions. [4] The method of claim 1, wherein the at least one response action includes providing an error signal associated with at least one of a residual mass flow rate error, a residual throttle position error, and a residual manifold air pressure error. [5] The method of claim 4, wherein the error signal includes a prognostic message indicative of a maintenance status of at least one of a mass air flow sensor (114), a throttle position sensor (120), and a manifold absolute pressure sensor (122). [6] The method of claim 4, wherein the airflow variation compensation value, the metric of the degree of coking of the throttle body (116) and / or the error signal are transmitted to an external server (128). [7] The method of claim 1, wherein generating the at least one residual error value includes a residual error of the throttle position sensor (120), a residual error of the mass air flow sensor (114), and a residual error of the manifold air pressure.

Images