Variable nozzle control of a variable displacement turbocharger

Abstract

When controlling boost pressure by adjusting the opening of the vanes that regulate the exhaust gas flow velocity to the variable displacement turbocharger, the optimal vane position is determined as needed, taking into account individual differences and changes in conditions, and this is continuously maintained to enable appropriate control. [Solution] When the required boost pressure on the turbocharger 13 increases from the current level to a high-power demand situation, the boost pressure at the vane 15 opening is measured using the vane 15 opening as a reference opening. Then, when the high-power demand situation returns, the boost pressure is measured again with the vane 15 opening set to a different opening than the reference opening. This process is repeated to obtain multiple boost pressure values ​​corresponding to multiple vane 15 openings. The opening that yields the highest boost pressure among the multiple vane openings is learned as the maximum boost opening of the vane 15, and the vane 15 is controlled during driving based on this maximum boost opening.

Description

Technical Field

[0001] This invention relates to the control of a variable nozzle of an automobile having a variable capacity turbocharger.

Background Art

[0002] As one of the superchargers that send compressed air to increase the output of an engine, a turbocharger that rotates a turbine using the energy of exhaust gas is used. The effect of the turbocharger depends on the inflow rate (flow velocity) of the air supplied to the turbine, so a mechanism for adjusting this as needed is provided. <00,00010>

[0003] As such a mechanism, a wastegate valve (W / G valve) may be used. In controlling this W / G valve, for example, after power is turned on, a constant drive current is passed through the valve control motor, and the positions where the valve is fully closed and fully open are learned by a position sensor. These are set as fully closed: 0% opening and fully open: 100% opening, and the opening is adjusted as appropriate between them. However, it is known that the situation once learned about the position of the valve does not continue as it is, and the optimal situation changes as appropriate. For this reason, it is proposed in Patent Document 1 to perform new learning and correction as appropriate.

[0004] On the other hand, in a variable capacity turbocharger, the output of the turbocharger is adjusted by varying the opening (angle) of the vanes provided around the turbine blades. In performing the opening control of these vanes, control for learning the opening of the vanes is known (Patent Document 2).

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Patent Document 2

Summary of the Invention

[0006] In a variable displacement turbocharger, as shown in Figures 2B-C, closing the vane 15 increases the velocity of the exhaust gas G flowing to the turbine blades, thereby increasing the turbocharger output. Opening the vane decreases the velocity of the exhaust gas flowing to the turbine blades, thus decreasing the turbocharger output. However, fully closing the vanes does not necessarily result in the fastest exhaust gas flow velocity to the turbine blades. As shown in Figure 2A, fully closing the vanes 15 narrows the gap too much, which actually reduces the exhaust gas G flow velocity to the turbine blades. In other words, unlike Patent Document 1, which controls the degree of supercharging by controlling the W / G valve, the relationship between the vane opening and the exhaust gas flow velocity to the turbine blades is such that the flow velocity increases as the vanes are closed up to a predetermined opening, but when the vane opening falls below the predetermined opening, the flow velocity decreases as the vanes are closed. Therefore, it is necessary to learn the vane opening that maximizes the flow velocity and thus the turbocharger output. Furthermore, due to aging changes such as the accumulation of deposits on the vane surface, the vane position that results in the fastest flow velocity changes. However, conventional methods, including Patent Document 1, do not disclose a method for learning the vane opening that maximizes the turbocharger output.

[0007] Therefore, the objective of this invention is to learn the vane position in a variable displacement turbocharger that maximizes the velocity of the exhaust gas flowing into the turbine and maximizes the output of the turbocharger. [Means for solving the problem]

[0008] This invention, as a first solution, A control device for a vehicle equipped with an engine having a variable displacement turbocharger that adjusts the flow rate of exhaust gas introduced into the turbine by changing the vane opening, When the required boost pressure for the turbocharger increases to a high-power demand, the boost pressure at the vane opening is measured using the vane opening as the reference opening. Then, when the high-power demand occurs again, the boost pressure is measured again with the vane opening set to a different opening than the reference opening. This process is repeated to obtain multiple boost pressure values ​​corresponding to multiple vane openings. Among the multiple vane openings, the opening that results in the highest supercharging is learned as the maximum supercharging opening of the vane. The above problems were solved by a control device that controls the vanes during operation based on this maximum supercharging opening.

[0009] Furthermore, in the first solution, this invention A second solution can be adopted in which, in the aforementioned group of situations, the learning is performed when the engine temperature is within a predetermined range, and the learning is not performed when the engine temperature is outside the predetermined range.

[0010] Furthermore, in the first or second solution, In measuring the degree of supercharging by changing the vane opening to a different degree from the reference opening, the vane opening is changed from the reference opening to either open or closed. If the degree of supercharging increases compared to the reference opening when the vane is changed to one side, then in the next high-output demand situation, the vane opening will be changed to one side and the measurement will be performed. If the degree of supercharging decreases compared to the reference opening when the vane is changed to one side, a third solution can be adopted in the next high-power demand situation, in which case the vane opening is changed to the other side and the measurement is performed.

[0011] Furthermore, in the first to third solutions of this invention, Once the predetermined conditions for the previous learning are met, the measurement and learning will be performed on a new group of supercharging levels with different vane openings. A fourth solution can be adopted, in which the reference opening of the vane in the new group is set to the position of the maximum supercharging opening in the previous group.

[0012] Furthermore, in the fourth solution, this invention If the degree of supercharging at the reference opening in the new group is equivalent to the degree of supercharging at the maximum supercharging opening in the previous group, a fifth solution can be adopted, which involves setting the reference opening in the new group as the maximum supercharging opening and ending the new group.

[0013] Furthermore, in the first to fifth solutions of this invention, If any of the aforementioned high-power requirements in one group do not meet the predetermined conditions, A sixth solution can be adopted, in which the supercharging level measured at that time is not used, and the supercharging level is measured at the same opening in the next high-output demand situation.

[0014] Furthermore, in the first to sixth solutions of this invention, In addition to the method of measuring the degree of supercharging based on the high power demand conditions and learning the maximum supercharging opening, the degree of supercharging is also measured by other methods. A seventh solution can be referenced, which evaluates the likelihood of learning obtained under the high-power demand conditions based on the measurement results from the other methods described above. [Effects of the Invention]

[0015] According to this invention, the vane opening that maximizes the supercharging level is determined from the actual supercharging level, so the vane position that reliably maximizes the supercharging level can be learned. Furthermore, since the vane position is learned based on the supercharging level in high-power demanding situations such as when accelerating from a standstill, the learning accuracy can be improved. [Brief explanation of the drawing]

[0016] [Figure 1] This functional block diagram shows an example of an embodiment of the control device according to this invention and a hybrid vehicle. [Figure 2A] Conceptual diagram of the situation where the vanes are closed too much and the degree of supercharging has decreased [Figure 2B] Conceptual diagram of the situation where the opening of the vanes is maximum and the exhaust gas is accelerating [Figure 2C] Conceptual diagram of the situation where the vanes are opened too much and the degree of supercharging has decreased [Figure 3] Flow example diagram of measurement and learning in the first learning mode [Figure 4] Graph showing the transition examples of the vanes, turbine rotation, and vehicle speed during measurement in the first learning mode [Figure 5] Graph showing the transition examples of the vanes, ignition timing, and actual torque during measurement in the second learning mode [Figure 6] Graph showing the transition examples of the vanes, turbine rotation, and intake air volume during measurement in the third learning mode [Figure 7] Flow example diagram of evaluating and learning the reliability of measurement by using different learning modes in flow pattern 1 [Figure 8] Flow example diagram of evaluating and learning the reliability of measurement by using different learning modes in flow pattern 2

Mode for Carrying Out the Invention

[0019] The control device 11 includes an arithmetic unit, temporary memory used for calculations, and a non-temporary computer-readable storage medium for storing programs and data. The control device 11 also has an interface for exchanging data and signals with each component of the vehicle 10.

[0020] The engine 12 of the vehicle 10 has a turbocharger 13 that supercharges the intake air. The turbocharger 13 is equipped with a turbine 14 that rotates in response to the flow of exhaust gas and a VG (Variable Geometry) vane 15 (hereinafter referred to as "vane") that adjusts the flow path for introducing exhaust gas to the turbine 14. If the vane 15 is opened too wide, the flow of exhaust gas G is not sufficiently restricted as shown in Figure 2C, and the degree of supercharging (turbine rotation speed) decreases. On the other hand, if the vane 15 is completely closed, exhaust gas G cannot flow in sufficiently as shown in Figure 2A, and the degree of supercharging decreases. Therefore, the opening of the vane 15 that maximizes the degree of supercharging is greater than the opening when it is completely closed. The control device 11 learns this opening of the vane 15 that maximizes the degree of supercharging (hereinafter referred to as "maximum supercharging opening") by measuring it, and controls the vane 15 based on this maximum supercharging opening to enable optimal supercharging.

[0021] When measuring the degree of supercharging, it is sufficient to measure values ​​that allow us to understand the status of the turbocharger 13, and the sensors used for measurement are not particularly limited. For example, by measuring the pressure in the intake manifold that supplies air to the combustion chamber of the engine 12 (intake manifold pressure), it can be determined that the degree of supercharging is higher when the intake manifold pressure rises more rapidly. Alternatively, by measuring the rotational speed of the turbine 14, it can be determined that the degree of supercharging is higher when the rate of increase in this rotational speed during the measurement period is high. In addition, multiple values ​​may be measured, and the degree of supercharging may be determined comprehensively from multiple pieces of information.

[0022] The opening degree of the vane 15 is controlled by the control device 11. A sensor may be provided to determine the opening degree of the vane 15. Alternatively, the opening degree of the vane 15 may be determined based on the amount of movement of the actuator that changes the opening degree of the vane 15.

[0023] In addition, it is desirable to provide sensors in various locations to measure the rotational speed of the turbine 14, the intake volume of the engine 12, and so on. The values ​​from these sensors are transmitted to the control device 11 and can be used as information to measure the degree of supercharging.

[0024] There are several possible timings for learning the vane 15 opening degree that results in the maximum supercharging opening. The first to third learning modes will be described in order.

[0025] <First Learning Mode> In the first learning mode, learning occurs when the vehicle 10 starts moving from a standstill (zero-to-start acceleration). However, in the case of a hybrid vehicle with a drive motor 16, learning occurs when the required supercharging level increases significantly due to a rapid increase in the output of the engine 12 from a steady state. These situations in which the supercharging level increases are collectively called high-output demand situations. Here, an increase in output from a steady state means that the output becomes greater than that of steady-state operation, and this is likely to occur, for example, during acceleration or uphill driving.

[0026] In a single high-power demand scenario, the degree of supercharging at one throttle opening is measured. Then, in the next high-power demand scenario (for example, zero-to-start acceleration after coming to a stop), the degree of supercharging at a different throttle opening is measured. Multiple supercharging levels are measured during multiple starts, and these are grouped together. Within this group, the degree of supercharging at each throttle opening is compared, and the throttle opening with the highest degree of supercharging is learned as the maximum supercharging throttle opening.

[0027] This example procedure is explained with the flowchart in Figure 3. The control device 11 first reads the information and history learned in previous series (S101). However, there may be no history to refer to immediately after factory shipment, immediately after factory inspection, or immediately after resetting the control device 11, which is an ECU. In that case, step S101 is omitted.

[0028] Next (S102), a check is performed to determine whether there is a learning request. If there is a learning request, learning begins (S102 → Yes → S110). If there is no learning request, measurement and learning in the first learning mode are terminated there (S102 → No). Whether or not there is a learning request can be determined by whether or not the above predetermined transition conditions have been met.

[0029] To begin learning (S110), the learning count is set to its initial value of "1". The control device 11 first checks the state of the vehicle 10 (S111). Specifically, it checks the vehicle speed and shift, and confirms whether the vehicle is in a stationary state, which is the stage before acceleration in the first learning mode (S111→S112). If the vehicle is moving, the flow returns (S112→No). If the vehicle is in a stationary state, that is, in a state of waiting for acceleration from a standstill (S112→Yes), the process proceeds to measurement and learning.

[0030] First, in the initial measurement of the boost pressure in that group, where the learning count is "1" (S113 → Count: 1), the opening of the vane 15 is set to a predetermined value based on the history (hereinafter referred to as the "reference opening") (S114). If there is no history, the reference opening is set to the maximum boost opening position, which is the default setting in the design of the vehicle 10. However, if the ignition does not remain ON and the engine 12 stops (S117 → IG OFF), the learning count is cleared (reset to the starting value of 1, S118), and the process returns to start (S102). This is because stopping and restarting the engine 12 can significantly change the environment of the turbocharger 13, and it is desirable to compare values ​​measured under conditions as close as possible to the environment, so it is better to restart the learning process.

[0031] If engine 12 starts without stopping (S117 → IG ON continues), check whether the start meets predetermined conditions (S121). These conditions are set to be those in which acceleration and accelerator pedal input are considered sufficient to measure the degree of supercharging. For example, the amount of accelerator pedal input is detected by the accelerator position sensor (APS), and the determination is made based on whether this input is greater than or equal to a predetermined value. If the conditions are not met (S121 → No), it is considered that the acceleration is insufficient and not enough to measure the degree of supercharging, so the learning count is left as is, and the system waits for the next high-power request situation (S111).

[0032] If the conditions for starting are met (S121 → Yes), the degree of supercharging is actually measured (S122). The measurement is performed from the start until the specified starting acceleration learning time has elapsed (S123 → No). This starting acceleration learning time should be between 2 and 5 seconds. If it is too short, it will not be possible to measure the degree of supercharging sufficiently. On the other hand, if it is too long, it will interfere with normal driving if the position of vane 15 is kept fixed. However, if the intake manifold pressure (intake manifold pressure) reaches a threshold that indicates it has risen sufficiently, even for a short time, it will be judged that it is being sufficiently supercharged, so the measurement is performed while checking the intake manifold pressure (SS124 → No). The measurement is completed when the specified time has elapsed (S123 → Yes) or when the intake manifold pressure is above the threshold (S124 → Yes).

[0033] At this stage, if the learning count is "1" (S125 → Yes), a comparison is made with the degree of supercharging at the maximum supercharging opening in the previous series, based on the history read in S101 (S126). For the first time, the maximum supercharging opening in the previous series is used as the reference opening (S114), and if the degree of supercharging at this time is within a range that can be considered to match the previous series (S126 → Yes), it can be determined that the environment of engine 12 has not changed enough to change the degree of supercharging, and the learning process ends there (S131). If it is determined that the degree of supercharging does not match the maximum supercharging opening of the previous series (S126 → No), the learning count is increased by 1 for the next measurement (S127 → No → S128), and the system returns to checking the vehicle status (S111). If the learning count is "2" or higher (S125 → No), it is checked whether the learning count has reached the specified number of times (S127). In the example in the figure, the specified number of times is 3, but it is not limited to this. If there are still remaining attempts (S127 → No), the learning count is increased by 1 (S128) and the process returns to checking the vehicle's status (S111). In the example shown in the diagram, for the measurement where the learning count becomes "2", the vane 15 is positioned +0.1 mm from the reference opening (S115) and the measurement is performed (S122). Furthermore, for the measurement where the learning count becomes "3", the vane 15 is positioned -0.1 mm (S116) and the measurement is performed (S122). However, this is only if the measurement result at learning count "2" is lower than the measurement result at learning count "1". If the measurement result at learning count "2" is better, the position is changed to wait at +0.2 mm instead of -0.1 mm (S116 modification).

[0034] In other words, if the boost pressure increases when the vane 15 is changed from the standard opening to either open or closed, it is advisable to change the opening further to the other side for the next measurement. As long as the boost pressure increases when the vane 15 is closed, check whether the boost pressure increases even when the vane 15 is closed further to find the maximum boost opening. On the other hand, if the boost pressure decreases when the opening is changed from the standard opening to either open or closed during the next high-power demand situation, it is advisable to change the opening to the other side from the standard opening for the next measurement. For example, if the boost pressure decreases when the vane 15 is slightly closed, it is likely that the vane 15 is closed too much, and opening it further is likely to improve the boost pressure.

[0035] If the specified number of measurements is greater than 3, measurements should be taken at positions with different opening angles in increments of 0.1 mm. The number of measurements in a group can be set as appropriate. However, if the number of measurements is too large, the situation will gradually change, making it difficult to maintain a single group, so it is preferable to keep the number of measurements at 10 or less, and more preferably 5 or less.

[0036] If the learning count reaches the specified number of times (S127 → Yes), the degree of supercharging in each instance is compared (S129), and the opening with the highest degree of supercharging is learned as the maximum supercharging opening (S130). This completes the learning for one group (S131), and the history is recorded, and the learning request flag is cleared (S140). After the learning for one group is completed, if predetermined conditions are met, measurement and learning are performed for a new group. The predetermined conditions can be, for example, continuous operation for a predetermined time, or driving a predetermined distance. In other words, it is good to set conditions that make it highly likely that the condition of the turbocharger 13 will change (due to aging, etc.) and the maximum supercharging opening will change.

[0037] Figure 4 shows a graph example summarizing the changes in vehicle speed, vane 15 opening degree, and turbine 14 rotational speed for measuring the degree of supercharging for one group in an example of the execution of the first learning mode. In the example shown in Figure 4, the turbine rotation increases most rapidly when the vane opening degree is 0.9 mm. Since the turbine rotation increases in accordance with the degree of supercharging, the highest degree of supercharging occurs at 0.9 mm.

[0038] The control device 11 should store the vane 15 opening and boost level when the engine temperature is within a predetermined range during multiple measurements of the boost level in a group, but should not store the vane 15 opening and boost level when the engine temperature is outside the predetermined range, and should measure the boost level again with the same vane 15 opening when high power is required next time. Since engine temperature affects the boost level, if the temperature is too different when measuring a particular opening compared to when measuring other openings, it becomes impossible to compare the boost levels under the same conditions. The engine temperature can be determined by using the coolant temperature of the engine 12, but is not particularly limited.

[0039] <Second Learning Mode> In the second learning mode, learning takes place when the vehicle 10 is operating at a steady speed in a state where engine 12 supercharging is not required. For example, this could be during series operation, where the engine 12 generates electricity and the vehicle continues to run on that electricity. In the case of a vehicle that runs solely on engine 12, this would be when it is operating at low speed or low load.

[0040] Figure 5 shows an example graph summarizing the changes in actual torque, vane 15 opening, and ignition timing (BTDC) in an example of the execution of the second learning mode. When the vehicle 10 is operating steadily in an operating state where supercharging of the engine 12 is not required, the vane 15 is fully open (an opening that causes the turbine 14 to barely rotate). However, when the system switches to the second learning mode, the vane 15 is gradually closed from the fully open state. The degree of supercharging at this time is measured, and the vane 15 opening just before the degree of supercharging decreases is learned as the maximum supercharging opening.

[0041] In the second learning mode, closing the vane 15 increases supercharging, which can raise the actual torque of the engine output and potentially exceed the target torque. Therefore, in conjunction with the operation of closing the vane 15, the ignition timing of the engine 12 is gradually retarded to control the generated torque to be approximately constant.

[0042] The maximum boost pressure opening can be determined by observing this ignition retardation amount. The opening of vane 15 that requires the most ignition retardation is the maximum boost pressure opening. Once the maximum boost pressure opening has been learned, the second learning mode is terminated, and the opening of vane 15 and the ignition timing are returned to their original positions.

[0043] However, ignition retardation increases exhaust temperature. Therefore, if the exhaust temperature exceeds a predetermined value, learning in the second learning mode is stopped, and the vane 15 opening and ignition timing are returned to their original settings.

[0044] Furthermore, if the vehicle 10's requested output increases during the second learning mode, the degree of supercharging must be controlled, making learning in the second learning mode difficult, and therefore the second learning mode is terminated.

[0045] Once learning is performed in the second learning mode, and after predetermined conditions are met, learning is performed again in the same manner in a new second learning mode. If the maximum supercharging opening has already been learned once in the second learning mode, instead of gradually closing the vane 15, it may be closed all at once to the maximum supercharging opening learned in the previous second learning mode, that is, the target opening of the vane 15 may be set to the maximum supercharging opening the moment the second learning mode is activated. Measurement of the supercharging degree is started at the maximum supercharging opening, and if the supercharging degree can be considered to match the value of the maximum supercharging opening in the previous second learning mode, the vane 15 is not moved any further, and that opening is learned as the maximum supercharging opening. Alternatively, measurement of the supercharging degree is started at the previously learned maximum supercharging opening, and if the supercharging degree does not match the value of the maximum supercharging opening in the previous second learning mode, the supercharging degree is measured while gradually closing the vane 15 from that opening. At this time, when closing the vane 15 from the maximum boost opening learned previously and the boost level decreases, the boost level is measured while gradually opening the vane 15. For example, if the boost level is measured by the ignition retard amount, the boost level can be considered to match the previous measurement if this ignition timing can be considered to match the previous measurement. Since the ignition retard amount is less susceptible to external influences, measuring the boost level using the ignition retard amount provides stable measurement accuracy.

[0046] Furthermore, if the environmental conditions such as temperature, humidity, and atmospheric pressure have changed significantly since the previous learning in the second learning mode, the degree of supercharging in the second learning mode may change. Therefore, the second learning mode may be performed only when the environmental conditions around the vehicle 10 are within a predetermined range (values ​​such as temperature, humidity, and atmospheric pressure are within a predetermined range).

[0047] <Third Learning Mode> The third learning mode is a learning control under special conditions of the second learning mode. Therefore, the third learning mode may be performed instead of the second learning mode. The third learning mode is performed while idling with the shift lever in a range other than the driving range (P range, N range). The control of the vane 15 during learning and the learning method for the maximum boost opening are the same as in the second learning mode, but the method of measuring the boost level is different. Since ignition retardation during idling may cause the combustion of the engine 12 to become unstable, in the idling learning mode, ignition retardation is not performed even when the vane 15 is gradually closed, and the boost level is measured from the rotational speed of the turbine 14. When the vane 15 is gradually closed, the opening of the vane 15 just before the rotational speed of the turbine 14 becomes maximum, that is, just before the rotational speed begins to decrease, is learned as the maximum boost opening. However, if the vehicle switches to the driving range during execution, the learning of the third learning mode is stopped because it is necessary to prioritize control for driving.

[0048] Figure 6 shows an example graph summarizing the changes in the intake volume of the engine 12, the opening degree of the vanes 15, and the rotational speed of the turbine 14 in an example of the execution of the third learning mode.

[0049] <Combining modes> The first, second, and third learning modes each operate under different measurement conditions (operating conditions). Therefore, it is possible to combine learning in each mode. For example, after learning in one mode, learning in another mode can be performed, and the results can be compared to verify the accuracy of the learning in the first mode.

[0050] <Flow Pattern 1> Figure 7 illustrates an example of a workflow in which learning in the first learning mode is performed in conjunction with learning in other modes. This example workflow is particularly suitable for vehicles 10 that run solely on engine 12, or for PHEV vehicles, especially when driving in mountainous areas where engine output is likely to fluctuate significantly.

[0051] First, the engine 12 is ignited in the parked vehicle 10 (S201). The control device 11 refers to the history recorded in the memory device (S202) and determines whether the time elapsed since the previous learning meets the conditions (S203). Here, it is expressed as time, but the conditions are not particularly limited and may be the number of days elapsed since the previous learning, or any other condition that can be judged as increasing the likelihood that the environment of the engine 12 has changed, such as the distance traveled, the time traveled, the amount of fuel added, or the number of engine ignitions since the previous learning. If the conditions requiring learning have not elapsed (S203 → No), it is determined that detailed learning is not yet necessary (S207), and the decision is postponed until the next engine ignition (excluding ignition after idling stop) (S208).

[0052] If the progress since the previous learning meets the conditions (S203 → Yes), a determination is made as to whether learning in the third learning mode is possible (S204), and measurement and learning are performed in the third learning mode (S205). The results from this third learning mode are used as a pre-test to determine whether or not a difference has occurred from the previous learning. If the change in the results from the third learning mode from the value in the previous learning is less than a predetermined value (S206 → Yes), it is determined that there has been no significant change in the environment of engine 12 and there is no need to relearn. In that case, the decision for learning is postponed until the next engine ignition (S207, S208).

[0053] If the results in the idling learning mode differ significantly from the results in the previous learning (S206 → No), it is determined that detailed learning is necessary, and the first learning mode is performed as the main learning (S211). This is performed using the procedure illustrated in Figure 3. The control device 11 also checks whether there have been any environmental changes to the engine 12 during one group in this first learning mode, or whether any changes have remained within a predetermined range (S212). Environmental changes to the engine 12 include, for example, changes in intake air temperature, exhaust air temperature, and atmospheric pressure around the vehicle 10. If the environmental changes remain within a predetermined range (S212 → Yes), there is no particular need to re-evaluate the learning results for that group (S213). However, if the environmental changes do not remain within a predetermined range, that is, if the environmental conditions around the vehicle 10 have changed significantly during each high-power demand situation (S212 → No), the learning content is evaluated because the impact of those environmental changes on the results may become significant (S214). To verify the learned content, learning is performed in the second learning mode (S215). By measuring the degree of supercharging while the displacement is stable, it is possible to verify whether the learning in the first learning mode is "incorrect." However, since elements related to acceleration response cannot be learned in the second learning mode, the first learning mode is used as the main learning mode, and the results of the second learning mode are used for verification. Specifically, it is verified whether the results of the second learning mode (S215) are more similar in trend to the previous learning result (learned value in S218 during the previous flow) or to the current learning result (S211) (S216). If the results of the current second learning mode are similar to the previous learning result, it is determined that the learning result in the current first learning mode is a failure (insufficient accuracy), and learning in the first learning mode is repeated (S216 → No → S211). If the results from the second learning mode are close to the results from the current learning mode, it is determined that the results from the current first learning mode have sufficient accuracy, and the results from the current first learning mode are adopted to learn the maximum supercharger opening (S216 → Yes → S217). After that, the timing of the next learning is updated (S218).

[0054] <Flow Pattern 2> Next, a flowchart example of another embodiment in which learning in the second learning mode is performed in conjunction with learning in other modes will be explained with reference to Figure 8. This flowchart example is suitable for use in a hybrid vehicle that generates electricity while the engine 12 is running in steady operation and is driven by the traction motor 16. Even in such a hybrid vehicle, the degree of supercharging to the engine 12 may be increased when high output is required instantaneously.

[0055] The overall structure of Flow Pattern 2 is similar to Flow Pattern 1 in Figure 7, and the numbered steps common to both Figure 7 and Flow Pattern 2 are basically the same, so we will omit their explanation.

[0056] For the main learning, the second learning mode is performed instead of the first learning mode (S211a). In hybrid vehicles, steady-state power generation is common, and supercharging is less frequent each time a high-power demand situation occurs, as shown in flow pattern 1 in Figure 7. Therefore, the second learning mode is more suitable as it provides a better opportunity to perform the main learning. However, even during steady-state operation, the environment of the engine 12 may change. For this reason, if the change in the environment of the engine 12 does not remain within a predetermined range (S212→No), the learning results are evaluated (S214). Here, learning is performed in the first learning mode for evaluation purposes (S215a). However, in the case of hybrid vehicles, during zero-start acceleration, the drive motor 16 is usually used instead of the output of the engine 12. Therefore, the conditions in step S112 in Figure 3 are changed and replaced with conditions that are in line with a high-power demand situation, such as "Has the required output of the engine 12 increased by a predetermined amount or more?". Note that since this is a measurement for evaluation purposes, the number of specified learning counts (S113, S127) can be fewer than when the first learning mode is used as the main learning, for example, it may be as few as once. If the previous learning result is closer to the result of the first learning mode, the learning in the second learning mode is repeated (S216 → No → S211a). If the current learning result is closer to the result of the first learning mode, the learning result from the current second learning mode is adopted and its maximum supercharge opening is learned (S216 → Yes → S217), and the timing of the next learning is updated (S218). [Explanation of Symbols]

[0057] 10 vehicles 11 Control device 12 Engines 13 Turbocharger 14 Turbine 15 Bane 16 motors 17 Batteries 18 Generators

Claims

1. A control device for a vehicle equipped with an engine having a variable displacement turbocharger that adjusts the flow rate of exhaust gas introduced to the turbine by changing the vane opening, When the required boost pressure for the turbocharger increases to a high-power demand, the boost pressure at the vane opening is measured using the vane opening as the reference opening. Then, when the high-power demand occurs again, the boost pressure is measured again with the vane opening set to a different opening than the reference opening. This process is repeated to obtain multiple boost pressure values ​​corresponding to multiple vane openings. Among the multiple vane openings, the opening that results in the highest supercharging is learned as the maximum supercharging opening of the vane. A control device that controls the vanes during operation based on this maximum supercharge opening.

2. The control device according to claim 1, wherein, in the group of conditions described above, the learning is performed when the engine temperature is within a predetermined range, and the learning is not performed when the engine temperature is outside the predetermined range.

3. In measuring the degree of supercharging by changing the vane opening to an opening different from the reference opening, the vane opening is changed from the reference opening to either open or closed. If the degree of supercharging increases compared to the reference opening when the vane is changed to one side, then in the next high-output demand situation, the vane opening will be changed to one side and the measurement will be performed. The control device according to claim 1, wherein if the degree of supercharging decreases compared to the reference opening when the vane is changed to one side, the vane opening is changed to the other side and the measurement is performed in the next high-output demand situation.

4. Once the predetermined conditions for the previous learning are met, the measurement and learning will be performed on a new group of supercharging levels with different vane openings. The control device according to claim 1, wherein the reference opening of the vane in the new group is the position of the maximum supercharging opening in the previous group.

5. The control device according to claim 4, wherein if the degree of supercharging at a reference opening in the new group is equivalent to the degree of supercharging at the maximum supercharging opening in the previous group, the reference opening in the new group is set to the maximum supercharging opening, and the new group is terminated.

6. If any of the aforementioned high-power requirements in one group do not meet the predetermined conditions, The control device according to claim 1, wherein the supercharging level measured at that time is not adopted, and the supercharging level is measured at the same opening in the next high-output demand situation.

7. In addition to the method of measuring the degree of supercharging based on the high power demand conditions and learning the maximum supercharging opening, the degree of supercharging is also measured by other methods. A control device according to any one of claims 1 to 6, which evaluates the likelihood of learning obtained by the high-power demand situation based on the measurement results of the other method described above.

Images