JP2000170890A - Automatic transmission - Google Patents

Abstract

PROBLEM TO BE SOLVED: To stably provide a good shift quality by more accurately calculating a correction value used at the time of speed-change when a feed back control is carried out. SOLUTION: This automatic transmission has a presumption part 73 for calculating a presumption value TDM of a control value; a correction value calculation part for calculating a correction value Ts based on the control value TD and the presumption value TDM, by using a speed-change model mathematically introduced from a mechanism of an automatic transmission AT and defining a relationship of revolution numbers ω1, ω5 of the rotation shaft and the control value TD. The control value calculation part calculates the control value TD based on a correction value Ts' fed-back from the correction value calculation part.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic transmission, and more particularly, to an engagement control of a friction engagement element such as a clutch or a brake that constitutes an automatic transmission.

[0002]

2. Description of the Related Art Generally, an automatic transmission for an automobile is composed of a torque converter and a mechanical transmission mechanism. The torque converter outputs the torque of the crankshaft, that is, the output torque of the engine that is the driving engine, to the input shaft of the transmission. On the other hand, the transmission mechanism further shifts the torque of the input shaft and outputs the torque to the output shaft of the transmission. Shifting in the transmission mechanism is achieved by switching characteristics of a power transmission mechanism including a sun gear, a ring gear, a pinion gear, and the like. Switching of the power transmission characteristic, that is, shifting, is performed by engaging or disengaging a friction engagement element (hereinafter, simply referred to as an "engagement element") such as a clutch or a brake connected to the gear. At this time, from the viewpoint of shift quality, the turbine speed is smoothly changed while matching the turbine speed with a preset target speed or matching the change rate of the turbine speed with the target change rate. Feedback control is generally performed.

As a conventional feedback control method, for example, Japanese Patent Application Laid-Open No. 9-269051 discloses a method of calculating a feedback hydraulic pressure by using a fuzzy control technique. An estimated value of the calculated hydraulic pressure is obtained by fuzzy combining a plurality of grade values and calculating the feedback hydraulic pressure. Then, a deviation between the estimated value and the calculated hydraulic pressure is obtained, and the calculated hydraulic pressure is corrected based on the deviation. By performing such feedback control, in the subsequent feedback control, the turbine speed or the change rate of the turbine speed can be quickly converged to the target speed or the target change rate.

However, the correction value calculated by the fuzzy control method as in the above-described conventional technique is not a mathematically or physically guaranteed value, but a value set based on experiments and experiences. Therefore, an optimum correction value (that is, a value that is mathematically and physically calculated) is not always set in all conditions, and an error from a value to be originally applied may be different depending on the condition. It is also conceivable that a large correction value is applied. If an improper correction value is applied, improper torque is generated at the engagement element, thereby deteriorating shift quality. Also,
In the related art, there is a problem that the labor required for setting the correction value under each condition is large and a problem that the amount of calculation required for calculating the correction value in the feedback control increases.

[0005]

SUMMARY OF THE INVENTION In view of the above-mentioned problems of the prior art, it is an object of the present invention to improve the shift quality by performing more accurate calculation of a correction value used in gear shifting for performing feedback control. Is to provide a new automatic transmission capable of stably obtaining the automatic transmission.

[0006]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention is directed to a method for setting a power transmission characteristic between a rotating shaft receiving driving force from a driving engine and a rotating shaft transmitting driving force to wheels. A detecting means for detecting the rotation speed of the rotating shaft, a control value calculating means for calculating a control value for controlling the release and engagement of the engaging element, and a mathematical Estimating means for calculating an estimated value of the control value by using a shift model that defines the relationship between the rotation speed of the rotating shaft and the control value, and calculates a correction value based on the control value and the estimated value. Correction value calculation means, wherein the control value calculation means provides an automatic transmission that calculates a control value based on the correction value fed back from the correction value calculation means.

Here, as the shift model, a different shift model is prepared for each type of shift, and the estimating means preferably calculates an estimated value by using a shift model corresponding to the shift to be executed. .

It is preferable that the speed change model is a speed change model in which at least one of viscosity about a rotating shaft, inertia about a rotating shaft, and turbine torque is considered.

On the other hand, the control value calculating means calculates the control value by using a second shift model derived from the mechanism of the automatic transmission and defining the relationship between the rotation speed of the rotating shaft and the control value. Is also good.

Further, the above configuration may further include a learning means for storing a correction value at the end of the shift control as a learning value. In this case, the control value calculation means calculates the control value based on the learning value. The learning means preferably stores a learning value for each type of shift. In this case, the control value calculating means calculates a control value based on a learning value corresponding to the shift to be executed, from among a plurality of learning values stored in the learning means, when a certain type of shift is executed. It is desirable.

In addition, in addition to the above configuration, a plurality of linear solenoid valves provided for each engagement element may be further provided. In this case, each linear solenoid valve gives a hydraulic pressure according to the control value to the engagement element. Thereby, the engagement element generates a torque corresponding to the hydraulic pressure.

[0012]

In such a configuration, the correction value to be fed back is calculated using a shift model. This shift model has a physical or mathematical basis because it is derived from the relationship of the mechanism of the automatic transmission, for example, the equation of motion, the relationship between torque and rotation speed. Therefore,
The correction value calculated using this is a relatively accurate value under all conditions (regardless of the rotational speed or the magnitude of the torque), so that a stable control value can be obtained regardless of the conditions.

[0013]

FIG. 1 is a diagram showing a schematic structure of a main part of an automatic transmission (AT) as an example. The driving force from the crankshaft 9 of the engine is transmitted to a turbine shaft 11 of the transmission via a torque converter 10. A turbine shaft 11 which is an input shaft of the transmission is connected to a sun gear of the rear planetary 2. On the other hand, a reduction drive shaft 12 as an output shaft of the transmission is connected to a ring gear of the front planetary 1 and a planetary carrier of the rear planetary 2. Each member of two planetary gears 1 and 2 (sun gear, planetary carrier, ring gear)
Are connected to three multi-plate clutches (reverse clutch 3, high clutch 5, low clutch 6), two multi-plate brakes (2 & 4 brake 4, low & reverse brake 7), and low one-way clutch 8, as shown. ing. These engagement elements (clutch, brake) are selectively engaged or disengaged according to the shift speed. As a result, the transmission can perform four forward speeds and one reverse speed.

FIG. 2 is a table showing the relationship between the shift position and the engagement state of the engagement element in the above automatic transmission. In this table, a circle indicates that the corresponding engaging element is engaged, and a blank indicates that it is released. Further, the mark ◎ indicates that the engagement is performed only during the corresponding driving. Except for 1st and 2nd gear, this transmission has a clutch-to-clutch
A shift is performed. What is clutch-to-clutch shifting?
This is a shift in which the front-stage engagement element is released and the rear-stage engagement element is simultaneously engaged. On the other hand, low one-way clutch 8
In the gear shift between first gear and second gear where
The shift is achieved only by the engagement control of.

FIG. 3 is a diagram generally showing a control mechanism of the automatic transmission. This control mechanism mainly includes an engine 21, a transmission mechanism 22, a hydraulic control mechanism 23, and an electronic control unit (ECU) 24. The torque generated by the engine 21 is transmitted to the reduction drive shaft 12 via the torque converter 10, the turbine shaft 11, and the transmission mechanism 22. This shaft 1
The torque of No. 2 is transmitted to the differential gear 16 via the drive pinion shaft 15 to drive the front wheels.

The throttle opening sensor S1 is connected to the engine 2
1 is a sensor for detecting the throttle opening θ.
The engine speed sensor S2 is a sensor that detects the speed of the crankshaft 9, that is, the engine speed ωE. The turbine speed sensor S3 is a sensor that detects the speed ω1 of the turbine shaft 11. The output rotation speed sensor S4 is a sensor that detects the rotation speed ω5 of the reduction drive shaft 12 (the output shaft in the present embodiment). As a sensor for detecting these rotation speeds, for example, an electromagnetic pickup may be used.

The oil pump 36 in the hydraulic control mechanism 23
Discharges the control oil sucked from the oil pan 37. The control oil adjusted to a predetermined oil pressure by the regulator valve 38 is supplied to five linear solenoid valves 31, 32, 3
3, 34, 35. Here, a linear solenoid valve is provided for each engagement element. Each linear solenoid valve controls the engagement element corresponding to the linear solenoid valve directly and linearly (a continuous control amount is generated according to the current value) in accordance with the current value from the hydraulic control circuit 51. A system in which a solenoid valve is provided for each engagement element and the engagement of each engagement element is directly controlled by a corresponding valve is generally called direct AT.

The ECU 24 includes a CPU 41, a ROM 42,
It comprises a RAM 43, an input circuit 44, and an output circuit 45. The signals output from the four sensors S1, S2, S3, S4 are input to the input circuit 44. CPU4
1 performs various calculations according to information from these sensors. The torque command value for controlling the linear solenoid valves 31 to 34 is output to the hydraulic control circuit 51 via the output circuit 45. The hydraulic control circuit 51 refers to a table or the like that defines the relationship between the torque command value (or the hydraulic command value having a linear relationship with the torque) and the control current value.
A current value for operating each linear solenoid valve is obtained from the torque instruction value and supplied to an appropriate linear solenoid valve.

FIG. 4 is a sectional view of the linear solenoid valve. The magnetic field generated by the electromagnet in response to the control current moves the spool valve, thereby connecting the supply port and the output port. In the direct AT system using a linear solenoid valve, the hydraulic pressure can be adjusted linearly according to the control current. The control current is determined according to a control value (torque command value or oil pressure command value). The valve of the linear solenoid valve opens by an amount corresponding to the control current value, and the corresponding control oil pressure is supplied to the engagement element. as a result,
Ideally, a torque having a magnitude corresponding to the torque instruction value is generated in the engagement element. However, the torque actually generated has a value different from the torque instruction value, although the degree is different, due to a difference in friction characteristics of the engagement element due to the influence of a disturbance element described later.

By using such a linear solenoid valve, it is not necessary to use an accumulator required when a duty solenoid valve is used. Therefore, in the direct AT system using the linear solenoid valve, high-precision hydraulic control that does not cause an interlock or a feeling of thrust can be performed by controlling the current supplied to the linear solenoid valve. In this embodiment, a normally-high valve that supplies the maximum control pressure when the current is 0 is used as the linear solenoid valve. FIG.
5 is a table showing a relationship between a gear position and an open / closed state of each linear solenoid valve.

Incidentally, instead of the direct AT system, it is also possible to use a conventional hydraulic circuit combining a duty solenoid valve and a switching valve. However, in the present embodiment that requires accurate control of the hydraulic pressure, it may be preferable to use the direct AT.

(Calculation of Shift Model for Each Type of Shift) One of the features of the automatic transmission according to the present embodiment is the equation of motion for each member constituting the automatic transmission, the relational expression of the torque in the planetary gear, and its relation. The point is that feedback control for correcting the hydraulic pressure of the engagement element to be controlled is performed based on the shift model of the automatic transmission derived from the relational expression of the rotational speed. Therefore, in this section, as a premise for explaining such feedback control, a shift model is derived for each type of shift by taking the mechanism of the automatic transmission shown in FIG. 1 as an example.

FIG. 6 is a skeleton diagram of the automatic transmission shown in FIG. In the symbols in the figure, I is the moment of inertia, T is the torque, D is the viscosity coefficient, and ω is the number of rotations of the rotating shaft. The subscripts indicate members of the transmission. Here, the subscript H is the high clutch 5, the subscript R is the reverse clutch 3, the subscript D is the 2 & 4 brake 4, and the subscript L.
Denotes a low clutch 6, a suffix B denotes a low & reverse brake 7, and W denotes a low one-way clutch 8, respectively. When the equation of motion of each member in the automatic transmission shown in FIG. 6 is obtained, the following relational expression including a viscous term (a term represented by Dω) and an inertia term (a term including a differential value of ω) is obtained. can get. Here, the inertia moment I1 of the turbine shaft is a value that also takes into account the inertia moment of the engine 21 and the torque converter 10.

(Equation 1)

The relational expression relating to the torque of the planetary gears (the front planetary 1 and the rear planetary 2) is as follows. Here, the suffix 1 means the front planetary side, and the suffix 2 means the rear planetary side. The subscript S indicates a sun gear, the subscript C indicates a planetary carrier, and the subscript R indicates a ring gear.

(Equation 2)

Further, a relational expression relating to the number of rotations (rotational speed) of the planetary gears 1 and 2 is as follows.

(Equation 3)

From the above relational expressions, a shift model at power-on is derived for each type of shift. First, the relationship between the turbine rotation speed ω1, the turbine torque TT, and the torque T at the engagement element is obtained from the relational expression between the torque and the rotation speed of the planetary gears 1 and 2. In deriving this relationship, a shift in the D range will be considered, and a change in the rotational speed of the reduction drive shaft during the shift will be considered to be very small.

(Equation 4)

First, in the inertia phase control of the 1-2 shift (including both the upshift and the downshift),
The shift model is as follows when TH = 0, TW = 0, and ω2 = ω4.

(Equation 5)

Next, in the inertia phase control of the 2-3 shift (including both the upshift and the downshift),
The shift model is as follows when TW = 0 and ω2 = ω4.

(Equation 6)

In the inertia phase control of the 3-4 shift (including both the upshift and the downshift),
The shift model is as follows when TW = 0 and ω1 = ω2.

(Equation 7)

(Calculation of Estimated Torque Based on Shift Model) The shift model (Equation 5 to Equation 7) derived for each shift is modified to obtain the torque of the engagement element.

First, in the inertia phase control of the 1-2 shift (including both the upshift and the downshift),
The torque actually acting on the 2 & 4 brake 4 to be controlled is expressed as the following Expression 8 by modifying Expression 5. In addition, 1-2 shift is one way clutch to clutch (oneway clutch to clutch)
Since it is a shift, only the engagement control of the 2 & 4 brake 4 is 1
-2 shifts are achieved.

(Equation 8)

The torque actually generated by the 2 & 4 brake 4 at a certain time is the turbine speed ω1, the turbine torque TT, and the output speed ω at that time.
If 5 is known, it can be estimated using Equation 8.
For this purpose, first, the turbine speed ω1 is monitored by the turbine speed sensor S3 shown in FIG.
The output rotation speed ω is determined by the output rotation speed sensor S4.
5 and further specify the turbine torque TT in an appropriate manner. Then, by substituting these parameters into Expression 8, the torque actually acting on the 2 & 4 brake 4 can be calculated as the estimated torque value TDM.

Next, in the inertia phase control at the time of shifting from the second speed to the third speed (including both power-on and power-off), the torque of the release side 2 & 4 brake 4 is changed during the inertia phase control. It is set to 0. on the other hand,
The torque actually generated by the high clutch 5 on the engagement side to be controlled is expressed by Expression 9 by modifying Expression 6. By obtaining the three parameters ω1, ω5, and TT and substituting them into Equation 9, the torque acting on the high clutch 5 during this type of shift can be calculated as the estimated torque value THM.

(Equation 9)

Conversely, in the inertia phase control at the time of shifting from the third speed to the second speed (power on), 2 &
The torque of the four brakes 4 is set to change linearly with the elapsed time of the inertia phase control. In this case, the estimated torque value THM of the high clutch 5 on the releasing side
Can be calculated from Expression 10.

(Equation 10)

Similarly, in the inertia phase control at the time of shifting from the third speed to the fourth speed (including both power-on and power-off), the torque of the low clutch 6 on the disengagement side is changed during the inertia phase control. It is set to 0. In this case, the estimated torque value TDM of the 2 & 4 brake 4 on the engagement side to be controlled is expressed by the following equation (1) obtained by modifying the equation (7).
1 can be calculated.

[Equation 11]

On the other hand, in the inertia phase control at the time of shifting from the fourth speed to the third speed (power-on), the torque of the low clutch 6 on the engagement side becomes a linear function in accordance with the elapsed time of the inertia phase control. Set to change. In this case, the estimated torque value TDM of the release side 2 & 4 brake 4
Can be calculated from Expression 12.

(Equation 12)

From the above description, those skilled in the art will be able to easily derive a shift model at the time of a downshift with the power off, and therefore, a description of the derivation of the shift model in this state will be omitted.

The shift model of Equations 8 to 12 derived in this manner is derived based on the equation of motion of each member of the automatic transmission and the like, and the turbine speed ω
1 and the viscosity term (Va · ω
1 and Wa · ω5), and furthermore, the turbine speed ω1
(The differential of ω1 multiplied by Ha). Therefore, if the turbine rotational speed ω1, the output rotational speed ω5, and the like are known, the torque currently acting on the mechanism of the automatic transmission can be estimated relatively accurately. In particular, since these shift models include a viscous term, the torque estimation value can be calculated relatively accurately irrespective of the rotational speed band of the rotating shaft (high or low of the rotational speed) at the time of executing the shift. In this embodiment, the turbine speed ω1 and the output speed ω5 are determined by the speed sensor S.
3, it is necessary to use a value directly measured using S4, but a value estimated based on other parameters may be used as the turbine torque TT.

(Hydraulic Control of Engagement Element Using Feedback Control) The calculation relating to the feedback control described in this section is executed by the ECU 24 in FIG.
FIG. 7 is a functional block diagram for executing hydraulic control of the engagement element using feedback control. The feature of the hydraulic control in the present embodiment is that a torque correction value is obtained based on a torque estimation value calculated using the above-described transmission model of the automatic transmission and a torque instruction value, and this is a control target. The point is that feedback is provided to the hydraulic control of the engaged engagement element. FIG. 7 illustrates an example of the inertia phase control at the time of upshifting from the first speed to the second speed. However, other types of shifts may be performed by using a shift model (any one of Expressions 8 to 12). Is basically the same, except that

The torque calculation unit 71 determines, for each time elapsed from the start of the shift control, the engagement side 2 &
A torque command value TD, which is a control value related to the fourth brake 4, is calculated. In order to obtain good shift quality, it is important to smoothly change the turbine speed ω1 from the start to the end of the shift, so that the control target 2 &
The torque instruction value TD for the four brakes 4 needs to be set appropriately from such a viewpoint. The hydraulic control circuit 51 shown in FIG. 3 determines a control current such that a torque corresponding to the torque instruction value TD 'is generated in the 2 & 4 brake 4 based on the input torque instruction value TD', and It is supplied to the linear solenoid valve 32. As a method of determining the control current, for example, there is a method of using a table defining the relationship between the torque instruction value (or the hydraulic pressure instruction value) and the control current. The control current described in this table is a value determined in advance by experiments so that the control current is supplied to the linear solenoid valve so that a torque corresponding to the torque instruction value acts on the engagement element.
The linear solenoid valve 32 has a hydraulic pressure P corresponding to the control current.
Generate D. Then, the brake piston moves according to the oil pressure PD and presses the pressure plate in the 2 & 4 brake 4, so that the 2 & 4 brake 4 generates torque.

The torque instruction value TD calculated by the torque calculator 71 may be calculated using a table or the like, but in the present embodiment, the torque instruction value TD is calculated using a shift model of an automatic transmission. The torque instruction value TD is set so that the actual turbine speed ω1 changes following a preset target turbine speed ωr. A method of calculating the torque instruction value TD using the shift model and the target turbine rotational speed ωr is described in Japanese Patent Application No. 10-2980 filed by the present applicant.
No. 87223, which is incorporated herein by reference. As an example, a method of calculating the torque instruction value TD in the inertia phase control at the time of upshifting from first gear to second gear will be briefly described. In the automatic transmission shown in FIG. 6, the 1-2 shift is a one-way clutch-to-
Since the shift is performed using a clutch (oneway clutch to clutch), the shift can be achieved only by controlling the engagement of the 2 & 4 brake 4.

The shift model used for calculating the torque instruction value TD (to be distinguished from the shift model used for calculating the estimated torque value TDM, which will be described later) is expressed by Expression 13. This equation replaces the turbine rotation speed ω1 of Expression 5 with the target turbine rotation speed ωr, and provides a feedback term (fifth term) relating to the deviation between the turbine rotation speed ω1 and the target turbine rotation speed ωr, and a torque correction value Ts ′ described later.
(6).

(Equation 13)

The target turbine rotational speed ωr appearing in the first, third, and fifth terms of the above equation is finally determined by the shift from the speed ratio at the start of the shift (the speed ratio before the shift). The speed change pattern is set in advance so as to smoothly change toward the speed ratio to be reached (the speed ratio after the speed change), and the output speed ω5 during the speed change is set in advance.
Is set for each elapsed time.

FIG. 11 is a block diagram for explaining the calculation of the torque instruction value TD. The turbine speed ω1 in the automatic transmission is detected by a turbine speed sensor S3 shown in FIG. The output rotation speed ω5
(The rotation speed of the reduction drive shaft 12) is detected by the output rotation speed sensor S4. A first value is obtained by multiplying the detected output rotational speed ω5 by the F / F gain Wa. Further, a target turbine speed ωr is calculated based on the output speed ω5 and the target speed ratio (speed change pattern). A second value is obtained by multiplying the value obtained by differentiating the target turbine speed ωr by the F / F gain Ha, and a third value is obtained by multiplying the target turbine speed ωr by the F / F gain Va. Ask. Further, a fourth value is obtained by multiplying the deviation between the target turbine speed ωr and the turbine speed ω1 by the F / B gain Fa. This F / B gain Fa is
For example, a reference rotation speed deviation of 10 rad / s is set, and an appropriate value is set in consideration of how much the influence of this deviation is reflected on the torque on the release side.

On the other hand, the turbine torque TT uses an estimated value calculated based on the estimated value TE of the engine torque and the estimated value of the torque ratio. FIG. 9 is a diagram for explaining a method of calculating the turbine torque TT. First, in order to calculate the engine torque TE, the throttle opening sensor S1 is used.
And the engine speed ωE is obtained from the engine speed sensor S2. The engine torque TE refers to an engine torque table describing the correspondence between the parameters and the turbine torque TT (estimated value), and uses values corresponding to these parameters.
On the other hand, in order to calculate the torque ratio, the speed ratio (ω1 / ωE) is calculated from the turbine speed ω1 and the engine speed ωE detected by the turbine speed sensor S3. Then, by referring to a table describing the correspondence between the speed ratio and the torque ratio (estimated value), a value corresponding to this parameter is used as the torque ratio. The turbine torque TT is estimated based on the torque ratio estimated from the table and the engine torque TE. The calculation of the turbine torque TT is an example, and the turbine torque TT can be calculated by various other methods. For example, the turbine torque TT may be estimated based on other parameters such as the intake pipe pressure and the intake air amount of the engine instead of the throttle opening θ as the engine parameter. It is also possible to directly measure the turbine torque TT using a kind of torque sensor. A fifth value is determined by multiplying the determined turbine torque TT by the F / F gain Ma. By adding the above five values, a basic value of the torque instruction value TD (hereinafter, referred to as a base torque instruction value Td) is obtained.

The torque correction value Ts', which is the output of the weighted average filter 74, is added to the base torque instruction value Td by feedback control, and the torque instruction value TD is calculated. As shown in FIG. 7, the learning value Ta output from the learning unit 75 is further added to the torque instruction value TD output from the torque calculation unit 71, and the final torque instruction value TD 'at a certain time is obtained. Is determined and this is the hydraulic control circuit 5
1 is input. As a result, torque is generated by the 2 & 4 brake 4 in accordance with the torque command value TD ', and the turbine speed ω1 and the output speed ω5 change. The torque correction value Ts' and the learning value Ta will be described later.

The shift model unit 73 uses a shift model for calculating the estimated torque value TDM, and uses the torque instruction value TD
Is calculated. The shift model corresponding to the type of shift currently being executed (one of Equations 8 to 11) is used, and the shift model of Equation 8 is used for 1-2 shift. FIG. 8 is a diagram for explaining the calculation of the estimated torque. The estimated torque TDM is calculated from the turbine speed ω1 detected from the turbine speed sensor S3, the output speed ω5 detected from the output speed sensor S4, and the turbine torque TT. By multiplying the detected turbine rotational speed ω1 by the gain Va, the third term (viscosity term) in Expression 8 is obtained, and the differential value of the turbine rotational speed ω1 is calculated by the gain H
The first term (inertia term) is obtained by multiplying by a. Further, the fourth term (viscous term) is obtained by multiplying the detected output rotational speed ω5 by the gain Wa. Further, the second term is obtained by multiplying the turbine torque TT by the gain Ma. Here, the turbine torque TT can be calculated by a method similar to the method of calculating the turbine torque TT in the torque calculation unit 71. By adding the above four values, an estimated torque value TDM is obtained. The estimated torque value TDM is a value determined using the turbine rotational speed ω1 and the output rotational speed ω5 directly detected by the sensors S3 and S4.

At the time CT, the deviation Ts (C) between the torque instruction value TD (CT) calculated by the torque calculation unit 71 and the estimated torque value TDM (CT) at the same time calculated by the shift model unit 73 is calculated.
Calculate T). Assuming that the shift model used in the shift model unit 73 accurately reproduces the mechanism of the automatic transmission, if the torque indicated by the torque instruction value TD is generated in the engagement element , The deviation Ts should theoretically be zero. However, some deviation Ts actually occurs. The reason why such a deviation Ts occurs is that an unmeasurable disturbance torque Tz based on a disturbance element exists in the actual automatic transmission 72. Due to the influence of the disturbance torque Tz, an error occurs between the theoretical torque value to be generated by the torque instruction value TD and the actually generated torque value. As a disturbance element that generates the disturbance torque Tz, for example, there is a difference in friction characteristics between the engagement elements. The coefficient of friction between the drive plate and the driven plate forming the engagement element cannot be completely the same in all the engagement elements,
Actually, the engagement elements are slightly different. As the degree of the difference increases, the influence of the disturbance element increases, so that the actually generated torque has a value that cannot be ignored with respect to the torque instruction value TD. Further, the secular change of the friction characteristic can be cited as a disturbance factor. As the wear time of the friction surface progresses with the elapse of the use time of the engagement element, the torque generated by the engagement element changes with time while the same oil pressure is supplied. on the other hand,
Disturbance elements also exist on the linear solenoid valve side, and typical examples thereof include variations in hydraulic characteristics. Even if the same control current is supplied, the opening degree of each valve is not completely the same, varies between individuals, and changes over time, so that the generated hydraulic pressure slightly differs. Therefore, the torque generated by such a disturbance element is regarded as a disturbance torque Tz, and its value is set to an amount corresponding to an error between the torque instruction value and the actually generated torque. If the error of the generated torque caused by the influence of the disturbance torque Tz becomes too large to be ignored, the intended torque cannot be appropriately generated, and the shift quality is deteriorated.

The turbine rotational speed ω1 and the output rotational speed ω5 detected by the sensors S3 and S4 naturally change depending on the torque instruction value TD '. Affected.
Therefore, as in the present embodiment, an accurate shift model that reproduces the characteristics of the automatic transmission is prepared, and the turbine speed ω
1. It is effective to obtain the estimated torque value TDM by inputting the output rotational speed ω5 and the turbine torque TT. Since the estimated torque value TDM calculated in this manner is regarded as the torque actually acting on the automatic transmission 72, the deviation Ts between the value TDM and the torque command value TD is determined, so that the disturbance torque Tz is calculated. It is possible to grasp the degree of the impact. That is, it can be determined that the larger the deviation Ts, the larger the disturbance torque Tz. Therefore, if the deviation Ts corresponding to the influence of the disturbance torque Tz is fed back to the torque setting unit 71, a torque instruction value TD that cancels out the influence of the disturbance torque Tz can be calculated.

However, in this embodiment, the deviation Ts is not directly fed back to the torque setting
The correction value Ts' obtained by averaging s through the filter 74 is fed back. The reason why such an averaging process is performed is to secure the stability of the system by feeding back a stable value even when the deviation Ts changes greatly temporarily. As an averaging method, for example, a method such as a weighted average can be used. When a weighted average is used, the correction value Ts ′ (CT) that is fed back is a weight coefficient a that is appropriate for the deviation Ts (CT).
It is calculated by adding a value obtained by multiplying (0 <a <1) and a value obtained by multiplying the correction value Ts' (CT-1) at the immediately preceding time by the weight coefficient (1-a). The correction value Ts' is used to calculate the torque
The next time (CT +
A new torque instruction value TD (CT + 1) in 1) is calculated.

Further, the control system shown in FIG.
have. The learning unit 75 has a data area for storing the calculated learning value Ta, and stores the learning value Ta for each type of shift. The learning unit 75 is desirably a device that does not lose stored data even when the power supply to the system is stopped by turning off the ignition key. For example, a nonvolatile semiconductor memory such as an EEPROM can be used. The learning value Ta is a correction value Ts' at the end of the shift control (for example, at the end of the inertia phase control or at the end of the torque phase control), and the content is read out and updated every time the same type of shift is executed. Is done. When a certain shift control is executed, a learning value Ta corresponding to the shift is read out, and this learning value Ta is added to the output of the torque instruction value TD which is the output of the torque calculation unit 71. Torque command value TD for each time i during the shift control period
The learning value Ta is added to (i). As a result, the torque command value TD is corrected, and the hydraulic control circuit 51 sets the solenoid valve 32 based on the corrected torque command value TD '.
Is determined.

As a method for updating the learning value, for example, any of the following methods can be applied. (1) The correction value Ts' at the end of the control is simply added to the previous learning value Ta. (2) A value Ts ″ is obtained by multiplying the correction value Ts ′ at the end of the control by an appropriate coefficient, and the value Ts ″ is added to the previous learning value Ta. (3) A value Ts ″ is obtained by performing a dead zone process as shown in FIG. 10 on the correction value Ts ′ at the end of the control, and the value Ts ″ is added to the previous learning value Ta. (4) The value T is obtained by performing an averaging process (simple averaging, weighted averaging, etc.) of the correction value Ts' calculated at each time in the control.
s '' is obtained, and its value Ts '' is added to the previous learning value Ta.

FIG. 12 is a timing chart showing the transition of the correction value Ts' and the learning value Ta when the same type of shift is repeatedly executed. It is assumed that, for example, +1.0 (constant value) is generated as the disturbance torque Tz in a certain type of shift, and a case where the learning value Ta is set to 0 is considered. Considering such static disturbance torque,
If the learning value Ta can be made to converge to −1 by repeating the shift, the disturbance torque Tz can be canceled. In an ideal state where there is no disturbance torque Tz, the torque instruction value TD of the 2 & 4 brake 4
Is a constant value as shown by a dashed line parallel to the time axis in the timing chart of FIG.

First, immediately after the start of the shift control of the n-th shift, since the learning value Ta is 0, the disturbance torque Tz is not compensated at all. Therefore, the disturbance torque Tz
And the deviation Ts between the torque instruction value TD output from the torque setting unit 71 and the estimated torque value TDM is large. As a result, a correction value Ts' which is a weighted average value of the deviation Ts
Is also large. As time elapses from the start of the shift control, the correction value Ts' gradually decreases (T
s'<0), the torque instruction value TD gradually decreases in accordance with this change. If the correction value Ts '(last1) at the end of the shift control is, for example, -0.5, the correction value Ts' (last1) becomes the previous learning value Ta.
(= 0) and -0.5 is stored as a new learning value Ta (when the above-described learning value updating method (1) is used).

After that, when the (n + 1) -th shift of the same type is executed again, the read learning value Ta (−0.5) is always changed to the torque instruction value TD ′ output from the torque calculator 71 (ie, the shift is performed). (At all elapsed times in the control). As a result, the torque instruction value TD at the start of the shift control starts from -0.5. In this shift control, the correction value Ts' gradually decreases from 0 in accordance with an amount that is insufficient for the compensation of the disturbance torque by the learning value Ta. Accordingly, the torque instruction value TD 'also gradually decreases from 0, so that the learning value Ta
Is decreased from -0.5. If the correction value Ts '(last2) at the end of the (n + 1) -th shift control is, for example, -0.25, the correction value Ts' (last2) becomes the previous learning value Ta (-0.5).
, And −0.75 is set as a new learning value Ta.
Will be remembered inside.

Further, when the (n + 2) th shift of the same type is executed again, the learning value Ta (−0.7
5) is always added to the torque instruction value TD 'from the torque calculator 71. As a result, the torque instruction value TD 'starts from 0, but the torque instruction value TD starts from -0.75. In this shift control, the correction value Ts' is the learning value Ta
However, the compensation for the disturbance torque due to is insufficient, and the compensation torque gradually decreases from zero. Accordingly, the torque command value TD also becomes-
It gradually decreases from 0.75. Assuming that the correction value Ts '(last3) at the end of the shift control is, for example, -0.125, the correction value Ts' (last3) becomes the previous learning value Ta (-
0.75), and -0.875 is stored as a new learning value Ta.

Thereafter, each time the same type of shift control is executed, the learning value Ta stored in the learning section 75 is added to the torque instruction value Ts ', and the final correction value Ts' of the control is added.
The learning value Ta is updated based on. As a result, the learning value Ta is the amount (-1.0) that cancels the disturbance torque Tz (+1.0).
Converges to As the degree to which the learning value Ta cancels the disturbance torque Tz increases, the turbine speed ω1 approaches an ideal change.

As described above, in the present embodiment, the torque calculator 7
1 and the transmission model unit 73
The correction amount is obtained based on the estimated torque value calculated in the step (1), and the torque instruction value is corrected based on the correction amount. Since the correction amount corresponds to the disturbance torque in the automatic transmission 72, if the hydraulic control of the engagement element is performed using the corrected torque instruction value, the torque corresponding to the torque instruction value is controlled. Can be generated by the engaging element. As a result, the shift quality can be improved as compared with the conventional feedback control. In particular, the shift model used to calculate the torque estimate used in the present embodiment is derived from the mechanism of the automatic transmission and accurately reflects the relationship between the parameters related to the automatic transmission. In addition, parameters directly affected by disturbance torque,
That is, the estimated torque value is calculated using the turbine speed and the output speed. Therefore, it is possible to further effectively reduce the influence of the disturbance torque.

Further, by using the shift model as in the present embodiment, the torque estimation value can be calculated only with the sensor conventionally mounted on the automatic transmission without providing a new sensor or the like. This system can be implemented at low cost. Another advantage is that accurate feedback control can be performed with a relatively small amount of calculation.

Further, since the correction amount calculated in the previous shift is stored as a learning value without being reset, and this learning value is used in subsequent shifts, a stable and good shift quality is obtained. Can be obtained. Further, since the learning value is stored and updated for each type of shift, it is possible to prevent the shift quality from being varied for each type of shift.

(Example of Application of Shift Model in 2-3 Shift) An actual example of application of the feedback control using the above-described shift model will be described with an example of an upshift from the second speed to the third speed. The shift includes four steps such as a shift start control, a torque phase control, an inertia phase control, and a shift end control.
There are two control modes. Each control mode in the 2-3 shift is switched by a flag F23, a flag F23T, and a flag F23I. The flag F23 is a flag indicating execution of the 2-3 shift control, and is set to 1 when the 2-3 shift is permitted. FIG. 13 shows each control mode and flags F23T and F23.
6 is a table showing a relationship with I.

FIG. 14 is a timing chart in the upshift. In the figure, when the 2-3 shift command is issued, the flag F23 changes from 0 to 1. afterwards,
After a shift start control (1), a torque phase control (2), an inertia phase control (3), and a shift end control (4), 2-3
When the shift is completed, the flag F23 changes from 1 to 0. FIG. 15 is a flowchart showing a part of a control procedure in the upshift of the 2-3 shift. FIG.
5 is a flowchart showing a control procedure of torque phase control, following the flowchart of FIG. FIG. 17 is a flowchart showing a control procedure of the inertia phase control, following the flowchart of FIG. 15 or FIG. Also, FIG.
FIG. 18 is a flowchart showing a control procedure of a shift end control following the flowchart of FIG. 17.

First, the flag F23 remains at 0 until the 2-3 shift command is issued, and the flag F23T and the flag F23I
Is also initially set to zero. Therefore, step 1
The process proceeds to return according to the judgment at 01. Since this flowchart is repeatedly executed, step 102 and subsequent steps are not executed until the 2-3 shift command is issued. When the 2-3 shift command is issued, the flag F23 changes to 1 and upshifting is started. At this time, the flag F23T
Since the flag F23I and the flag F23I are both 0 (ie, the state of the shift start control (1) in FIG. 9), the process proceeds to step 10 through steps 102, 103, and 104.
Proceed to 5.

In step 105, the degree of engine blow-up, that is, the degree of slippage of the disengagement-side engagement element (2 & 4 brake 4) is determined. When slippage occurs, the turbine rotation speed ω1 becomes larger than the slippage determination rotation speed ωth ((output rotation speed ω5) * (second speed gear ratio i2) + (slip reference amount Δω)). Since no slippage occurs at the beginning of the shift, ω1 = i2 · ω5. In this case, the routine proceeds to step 106, where the control hydraulic pressure on the release side is reduced. This hydraulic pressure is calculated from the product of an appropriate linear function q23, which decreases substantially linearly, and the turbine torque TT.

Steps 101 to 10 described above
When the procedure of step 6 is repeated and the hydraulic pressure on the release side is gradually decreased linearly, the engagement element on the release side eventually slips. Due to this slip, the turbine rotation speed ω1 becomes higher than the slip determination rotation speed ωth. In this case, the process proceeds to step 107, where the flag F23T is changed from 0 to 1, and the clock CT is set to 0 (step 108). As a result, the control mode changes from the shift start control (1) to the torque phase control (2).

In step 109 in FIG. 16, it is determined whether or not the turbine speed ω1 has been pulled. At the beginning of the torque phase control (2), the turbine rotational speed ω1 is larger than the value obtained by multiplying the output rotational speed ω5 by the second speed ratio i2, and the turbine rotational speed ω1 is not drawn because the vehicle is slipping. Therefore, the process proceeds to step 110, where the torque phase control is executed.

In the torque phase control (Step 110), the hydraulic pressure on the engaging side is increased and the hydraulic pressure on the releasing side is decreased so as to maintain the slippage generated in the engaging element on the releasing side. By repeatedly executing this, both the hydraulic pressure on the engagement side and the hydraulic pressure on the release side change while the turbine speed ω1 follows the target turbine speed ωr. When such release-side and engagement-side hydraulic pressure control proceeds,
Retraction in the fast direction occurs. As a result, the turbine speed ω1 is smaller than the value obtained by multiplying the output speed ω5 by the second speed ratio i2. In this case, step 109
To step 117. Then, the flag F23T is set to 1
From 0 to 0, reset the counter value CT (step 118), change the flag F23I from 0 to 1 (step 119), and proceed to step 120 in FIG.

Since the flag F23T is 0 and the flag F23I is 1, the inertia phase control (3) is executed. At the start of the inertia phase control, the output rotational speed
The turbine speed ω1 is larger than the value obtained by multiplying the speed ratio i3. Therefore, the process proceeds from step 120 to step 131, and the inertia phase control is executed.

First, with the start of the inertia phase control, 1 is added to the counter value CT, and then the release-side torque TD
Is set to 0 (step 13
2), the hydraulic pressure PD on the release side is set to be 0 (step 133). Next, the procedure from step 134 to step 138 is executed to calculate the engagement side hydraulic pressure PH.

In step 134, a pattern value R23 (CT) corresponding to the counter value CT is specified with reference to the target pattern f2 based on the current counter value CT (see the table below as an example).

[Table 1] CT 1 2 3 4 5 6 7 8 9 10 R23 1.99 1.95 1.90 1.83 1.75 1.67 0.61 1.55 1.51 1.50 ωr 3980 3900 3790 3650 3500 3350 3210 3100 3020 3000

Taking the change pattern of the target rotational speed described in Table 1 as an example, 1.99 is obtained as the pattern value R23 (1) corresponding to the counter value CT = 1. For example, 2
Assuming that the speed ratio i2 = 2.00, the speed ratio i3 after the third speed shift i3 = 1.50, and the output speed ω5 is 2000 rpm (constant value), the target turbine speed ωr at CT = 1 is 3980 rpm.
(Step 135).

The engagement-side torque command value TH is calculated so that the turbine speed ω1 follows the target turbine speed ωr (step 136). The engagement-side torque instruction value TH is calculated based on the following shift model. Here, the previous correction value Ts' (CT-1) in the equation of this shift model
Is added. Note that the correction value Ts'
Is initially set to 0, so step 131
At the first execution of the subsequent procedure, the correction value Ts' (0)
Is 0.

[Equation 14]

The target turbine rotational speed ωr, turbine torque TT (estimated value), turbine rotational speed ω1 (actual measured value of turbine rotational speed sensor S3), and output rotational speed ω5 (actual measured value of output rotational speed sensor S4) are calculated. By substituting these as input parameters into the shift model,
The torque instruction value TH at that time is obtained. Then, the torque instruction value TH 'is obtained by adding the learning value Ta to the torque instruction value TH (step 137).

An engagement-side hydraulic pressure instruction value PH is obtained from the obtained torque instruction value TH '(step 138). In general,
The hydraulic pressure P of the engagement element has a linear function relationship with the torque T, and is determined by P = k · | T | + s (where k and s are constants and | T | is the absolute value of the torque). Therefore, a unique parameter (kH in the case of the high clutch 5,
If (sH) is set, the hydraulic pressure instruction value PH can be obtained from the torque instruction value TH '. The control current of the linear solenoid valve is determined by referring to the conversion table and the like with the hydraulic pressure PH determined in this way as an instruction value.

A series of procedures from step 139 to step 141 relate to feedback control.
First, in step 139, the transmission model unit 73
Based on Equation 9, the estimated torque THM is calculated. And
A deviation Ts between the calculated estimated torque THM and the calculated torque TH is calculated (step 139), and a weighted average thereof is used to calculate a correction value Ts' (CT) (step 140).
Then, proceed to return.

The procedure from step 131 to step 141 is repeatedly executed until it is determined in step 120 that the turbine speed ω1 is synchronized with the speed ratio of the third speed. With the progress from the start of the inertia phase control (as the counter value CT is incremented), the target turbine rotational speed ωr is, as shown in Table 1,
Decreases smoothly (without discontinuous changes). As a result, the turbine rotation speed ω1 also follows the target turbine rotation speed ωr, and thus smoothly changes over time. Further, for the reason already described with reference to FIG. 12, the correction value Ts'(<0) is set to a value that cancels the disturbance torque Tz.
Is gradually approaching.

The inertia phase control proceeds, and when the turbine speed ω1 is eventually synchronized with the third speed gear ratio (step 12).
0), the process proceeds to a step 121, wherein the learning value Ta is updated. This is because the correction value Ts' (l
ast) based on the above-described learning value updating method. Then, in step 122, the flag F23
T and the flag F23I are both set to 1 and the counter value C
After T is reset (step 123), the process proceeds to step 124 in FIG.

Since both the flag F23T and the flag F23I are 0, the shift end control (4) is executed. First, in step 124, the hydraulic pressure on the engagement side is increased, and it is determined whether or not this hydraulic pressure has reached a maximum value (step 1).
25). If the hydraulic pressure has not reached the maximum value, the process proceeds to return and waits for the next execution of this flowchart. In the next execution, the flag F23T and the flag F23I are both 1
Therefore, the process proceeds from step 104 to step 124.
This procedure is repeated until the hydraulic pressure on the engagement side reaches the maximum value. When the hydraulic pressure reaches the maximum value, the process proceeds from step 125 to step 126. as a result,
While changing the flag F23T and the flag F23I from 1 to 0, the flag F23 is also changed from 1 to 0 (step 127).
Proceed to return. Since the flag F23 has become 0, the 2-3 shift is completed according to the above procedure.

In the present embodiment, a shift model of an automatic transmission having a structure as shown in FIG. 3 is derived and described. However, the automatic transmission shown in FIG. 3 is an example, and it goes without saying that the present invention can be applied to various automatic transmissions having other structures.

Although the present embodiment has mainly been described for the case where the present invention is applied to the inertia phase control, the present invention can naturally be applied widely to control for performing hydraulic control including torque phase control. It can be used for various hydraulic controls of the engagement element within the scope of the gist.

[0081]

As described above, according to the present invention, the control value of the engagement element is estimated by substituting the detected turbine speed and output speed into the shift model of the automatic transmission. . Then, the correction amount is calculated based on the calculated control value and the estimated control value. This correction amount acts to cancel the disturbance torque in the automatic transmission, that is, the error in the torque generated due to the variation in the characteristics of the engagement element and the solenoid valve and the aging thereof. Therefore, if the control value is calculated by feeding back the correction value, it is possible to calculate the control value that cancels out the influence of the disturbance torque. Therefore, it is possible to effectively suppress deterioration of the shift quality due to the influence of the disturbance torque.

[Brief description of the drawings]

FIG. 1 is a diagram showing a schematic structure of a main part of an automatic transmission according to an embodiment.

FIG. 2 is a table showing a relationship between a shift position and an engagement state of an engagement element in the automatic transmission of FIG. 1;

FIG. 3 is a diagram generally showing a control mechanism of the automatic transmission.

FIG. 4 is a sectional view of a linear solenoid valve.

FIG. 5 is a table showing a relationship between a shift speed and an open / close state of each linear solenoid valve.

FIG. 6 is a skeleton diagram of the automatic transmission shown in FIG. 1;

FIG. 7 is a functional block diagram for executing hydraulic control of an engagement element using feedback control.

FIG. 8 is a view for explaining calculation of an estimated torque value using a shift model.

FIG. 9 is a diagram for explaining calculation of a turbine torque TT.

FIG. 10 is a view for explaining dead zone processing relating to a correction value;

FIG. 11 is a diagram for explaining calculation of a torque instruction value;

FIG. 12 is a timing chart showing changes in a correction value and a learning value when the same type of shift is repeatedly executed.

FIG. 13 is a table showing a relationship between each control mode and a flag.

FIG. 14 is a timing chart in the inertia phase control of the second to third speed upshifts;

FIG. 15 is a flowchart showing a part of a control procedure in an upshift of a 2-3 shift;

FIG. 16 is a flowchart showing a control procedure of torque phase control, following the flowchart of FIG. 15;

FIG. 17 is a flowchart showing a control procedure of the inertia phase control, following the flowchart of FIG. 15 or FIG.

18 is a flowchart showing a control procedure of a shift end control following the flowchart of FIG. 17;

[Explanation of symbols]

1 front planetary, 2 rear planetary, 3 reverse clutch, 4 2 & 4 brake, 5 high clutch, 6 low clutch, 7 low & reverse brake, 8 low one way clutch, 9 crankshaft, 10
Torque converter, 11 turbine shaft,
12 reduction drive shaft, 15 drive pinion shaft, 16 differential gear 21
Engine, 22 transmission mechanism, 23
Hydraulic control mechanism, 24 ECU, 31, 3
2,33,34,35 Linear solenoid valve, 36
Oil pump, 37 Oil pan, 38
Regulator valve, 41 CPU, 42 R
OM, 43 RAM, 44 input circuit, 45 output circuit, 51 hydraulic control circuit, 71 torque setting section, 72 automatic transmission, 73 shift model section, 74 weighted average filter, 75 learning section, S1 throttle opening sensor, S2 engine Speed sensor, S3
Turbine speed sensor, S4 output speed sensor

Claims (10)

[Claims] 1. A plurality of engagement elements for setting a power transmission characteristic between a rotating shaft receiving a driving force from a driving engine and a rotating shaft transmitting a driving force to wheels, and a detecting device detecting a rotation speed of the rotating shaft. Means, control value calculation means for calculating a control value for controlling the release and engagement of the engagement element, and a relationship between the number of rotations of the rotating shaft and the control value, which is derived from a mechanism of the automatic transmission. An estimating unit that calculates an estimated value of the control value by using a prescribed shift model; and a correction value calculating unit that calculates a correction value based on the control value and the estimated value. Means for calculating the control value based on the correction value fed back from the correction value calculation means. 2. The shift model, wherein the shift model different for each shift type is prepared, and the estimating means calculates the estimated value by using the shift model corresponding to a shift to be executed. The automatic transmission according to claim 1, wherein: 3. The automatic transmission according to claim 1, wherein the speed change model is a speed change model that considers viscosity of the rotation shaft. 4. The automatic transmission according to claim 3, wherein the speed change model is a speed change model in which inertia with respect to the rotating shaft is considered. 5. The automatic transmission according to claim 4, wherein the shift model is a shift model taking turbine torque into consideration. 6. The control value calculating means uses a second shift model derived from a mechanism of the automatic transmission and defining a relationship between a rotational speed of the rotating shaft and the control value. The automatic transmission according to claim 1, wherein a value is calculated. 7. A control device according to claim 1, further comprising: a learning unit configured to store the correction value at the end of the shift control as a learning value, wherein the control value calculating unit calculates the control value based on the learning value. The automatic transmission according to claim 1, wherein: 8. The automatic transmission according to claim 7, wherein said learning means stores said learning value for each type of shift. 9. The control value calculating means calculates the control value based on a learning value stored in the learning means and corresponding to the shift, when a certain kind of shift is executed. Item 10. The automatic transmission according to Item 8. 10. A linear solenoid valve further comprising a plurality of linear solenoid valves provided in association with each of said engagement elements, wherein each of said linear solenoid valves supplies a hydraulic pressure corresponding to said control value to said engagement element. The automatic transmission according to any one of claims 1, 2, 6, and 7, wherein: