JPS6347544A - Torque detecting device for power train - Google Patents

Abstract

PURPOSE:To enable a stable condition free from periodic noises to be maintained, by determining values of the correction factor applied to the production of noise elements, depending on a variation in the operating condition of an engine. CONSTITUTION:The operating condition of the engine which drives a power train is detected by an engine operating condition detecting means 6, and according to the output thereof, a variation in the operating condition is detected by an operating condition variation detecting means 7. A correction factor is determined by a correction factor determining means 8, based on the variation in the operating condition of the engine. A greater variation in the operating condition of the engine requires a value of the correction factor alpha to be changed to the extent in which no essential effect may be exerted on the production of noise elements, if the periodic noises from a detected torque signal d(t) are found difficult to be removed, thus prohibitting the noise elements to be corrected or reducing the degree of such correction. Therefore, the stable condition free from noises can be maintained even under varied operating conditions of the engine.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention provides a torque detection means for detecting the torque of a rotating shaft of a power train, and periodic noise contained in a detected torque signal inputted from the torque detection means. The present invention relates to a torque detection device for a power train, which includes noise removing means that sequentially corrects and generates noise components and subtracts them from the detected torque signal to obtain a target corrected torque signal.

[Conventional technology]

A torque detection device that detects the torque of a rotating shaft of a power train, for example, the output shaft of an automatic transmission, removes periodic noise from this output shaft torque, and obtains a corrected torque signal used for power train speed change control, etc. There is something like the one shown in FIG. 13 (see Japanese Patent Application No. 60-271668).

This device, which is currently being proposed by the applicant, uses a previously known synchronous application means to convert the torque signal input from the detection sensor into a true signal or DC level. This method effectively removes only the periodic noise that occurs. This synchronous application means 1 basically uses a detected torque signal d (t) obtained by waveform-shaping the output of a torque detection sensor that detects the torque of the rotating shaft;
A reference pulse signal x(t) that has a phase correlation with the periodic noise contained in this detected torque signal d(t) is input, and in synchronization with this reference pulse signal x(t), each phase of the periodic noise is The noise component y(t) of ). In other words, the noise component y (t) is expressed as y (t)=
y-1(t) + α · e (t) α: Correction coefficient, created by sequential correction according to O〈α〈l.

A modified torque signal e (t) is obtained according to e (t) = d (t) - y (t). However, simply creating a noise component according to y (t) = y-1 (t) + α · e (t) will result in the noise component y (t) if the true signal value to be obtained is a DC component. As a result, the DC component that becomes the true signal value is removed from the corrected torque signal e(t) obtained by e(t) = d (t) - y (t). That will happen.

Therefore, in the above application, the present applicant proposed that the detected torque signal d
The noise component y (
t) from the corrected torque signal e(t) obtained at that time.
A subtraction means 3 is provided for subtracting the average level dm of the detected torque signal d(t) extracted by synchronous adaptation means l. It is assumed that the signal contributes to the creation of the noise component y (t) in .

This will be explained in detail below based on FIG. 14.

An automatic transmission lO that transmits power to the power train transmits engine power to an input shaft 12 via a torque converter 11, and further transmits the input of this input shaft 12 to a transmission 1 having a fluid friction element.
3 to the output shaft 14 at a predetermined rotation ratio and rotation direction to rotate the propeller shaft 15.

To detect the output shaft torque of such an automatic transmission 10, it is common to use a magnetostrictive torque sensor 16 that alternately magnetizes the output shaft 14 and detects changes in the magnetic field accompanying torque fluctuations. The output signal of the magnetostrictive torque sensor 16 is basically an AC signal corresponding to alternating magnetization whose amplitude is changed according to the magnitude of the output shaft torque, as illustrated in No. 812(a). However, due to eccentricity of the output shaft 14, material unevenness, etc., this output signal further periodically fluctuates in synchronization with the rotation of the output shaft 14. To actually detect the output shaft torque □, this output signal is subjected to aftereffect rectification and high-frequency components are removed using a low-pass filter (see Fig. 15).
The waveform shown in b) is processed, and the actual output shaft torque is detected based on this detected torque signal d (t). In the detected torque signal d (t) shown in FIG. 15(b), its DC component corresponds to the actual torque. On the other hand, in order to remove periodic fluctuations included in this detected torque signal d(t), that is, periodic noise, the seventh
As shown in the figure, an electromagnetic induction type rotation sensor 13 is disposed in the vicinity of a predetermined number of parking gears 17 provided on the output shaft 14, and a reference pulse having a phase correlation with periodic noise is sent from the rotation sensor 13. Obtain the signal x (t). Then, the synchronous adaptation means l inside the signal processing circuit 20 equipped with a noise removal device receives the reference pulse No. (1) is input, and the desired corrected torque signal e (t) is obtained while removing periodic noise as follows. That is, when the detected torque signal d(t) obtained from the magnetostrictive torque sensor 16 and the signal waveform of the reference pulse signal x(t) obtained from the one-rotation sensor 18 are as shown in FIG.
In order to obtain a corrected torque signal e (t) obtained by removing periodic noise from the detected torque signal d (t) at a particular phase t, the noise component y- generated at the same phase in the period immediately before the relevant period is 1(t) and the corrected torque signal e-1(t) obtained in the previous cycle that is fed back.
The following calculation is performed using the deviation signal [e-1(t)-dml obtained by subtracting the average level dm of the detected torque signal d(t) from .

Y (t)=y-1(t)+α[e-1(t)-dm]
α: correction coefficient, 0 × α × l and according to e (t) = d (t) y (t), a modified torque signal e(t) is obtained. If such correction processing is performed sequentially for each phase from the first cycle, a certain phase (
The periodic noise y(t) for 1) will converge to the value obtained by subtracting the average level signal dm from the corrected torque signal e(t) as the correction is repeated, and the resulting corrected torque The signal e (t) is obtained by removing the periodic noise y (t) from the detected torque signal d (t).

Then, the corrected torque signal e (t) obtained in this way
is used for speed change control of a power train as follows (see Japanese Patent Application No. 193950/1982). In FIG. 17, reference numeral 16 denotes a torque sensor for detecting the output shaft torque of the automatic transmission lO as described above, and 10
Reference numeral 1 denotes the aforementioned noise removing means for removing periodic noise from the detected torque signal d(t) inputted from the torque sensor 16. This noise removing means 101 outputs a corrected torque signal e(t).

Reference numeral 102 denotes a variation range detection means for determining the variation range Δe of the output shaft torque (see FIG. 18) during the shift operation of the automatic transmission 10 based on this corrected torque signal e (t), and 103 indicates the output shaft torque at the start of the shift. e
This is a shift operation evaluation means that determines the quality of the shift operation based on the torque fluctuation rate K based on the output shaft torque e6 at the end of the shift and the previously determined torque fluctuation width Δe. Specifically, the smaller the variation rate K is, the better the shifting operation is, and the larger the variation rate K is, the worse the gear shifting operation is, with a larger shift shock.

Reference numeral 104 denotes a fluid pressure adjusting means for correcting the setting of the supply pattern of the fluid pressure supplied to the fluid formal friction element 105 for shifting in response to such a shift operation evaluation. Therefore, with this arrangement, the corrected torque signal e (t) obtained from the noise removing means 101 is used to control the fluid friction element 105 in an appropriate pressure increase pattern that does not cause a shift shock based on the torque value. It can pressurize and reduce shift shock in the powertrain. If the evaluation result of the shift operation evaluation means 103 under the same conditions is favorable, the fluid pressure adjustment means 104 thereafter supplies pressure to the fluid pressure type friction element with the same fluid pressure pattern.

(Problem to be solved by the invention) By the way, the torque applied to the rotating shaft of the power train is
It fluctuates with changes in the operating state of the engine of the vehicle that includes the power train. Here, changes in the operating state of the engine include, for example, changes in the throttle opening of the engine, changes in the intake air amount of the engine, changes in the intake pipe pressure of the engine, changes in the rotational speed of the engine, and the like.

FIG. 19 shows an example of the relationship between changes in the engine throttle opening and fluctuations in the output shaft torque, taking the engine throttle opening as an example.

In the figure, (a) is a change in engine throttle opening, (b) is a torque signal indicating output shaft torque, (C) is a detected torque signal including periodic noise, and (d) is a detected torque signal d ( The average level dm extracted by the average level extracting means 2 is superimposed on t).

As is clear from this figure, when the output shaft torque fluctuates due to changes in the engine throttle opening, the detected torque signal d (t) extracted by the average level extraction means 2
The difference between the average level dm of dm and the actual DC level of the detected torque signal d(t) at the time of its extraction becomes large.

This is because the average level dm depends on the level state of the detected torque signal d (t) before its extraction, and in particular, the detected torque signal d (t) as shown in FIG. 5(C).
This is because the level change before the extraction point is large. If the difference between the average level dm extracted in this way and the actual DC level of the detected torque signal d (t) is large, this is the difference between the target corrected torque signal e (t) and the average level dm.

The noise component y (t) generated based on the difference signal [e (t) - d (t) ] deviates from the actual value. Therefore, the engine operating state 1, for example, the engine throttle opening state If d(t) fluctuates greatly, the noise component y(t) created to remove periodic noise will be disturbed, and the detected torque signal d(t
) will fail to remove periodic noise from As a result, periodic noise remains in the corrected torque signal e (t), and there is a problem that it is not possible to reliably control the power train's shift to reduce shift shock based on the corrected torque signal e (t). will occur.

Therefore, one object of the present invention is to prevent the generation of a modified torque signal e (t) on which the noise component y(t) is superimposed, and to ensure the speed change control of the power train.

(Means of the Present Invention) As shown in the first factor, the power train shift control bag according to the present invention includes a torque detecting means 4 for detecting the rotating shaft torque of the power train, and a detected torque input from the torque detecting means. The noise component y at each phase is calculated using a predetermined correction coefficient α in synchronization with the reference pulse signal x (t), which has a phase correlation with the periodic noise contained in the signal d (t).
(t) while sequentially modifying the detected torque signal, and subtracts this noise component from the detected torque signal to obtain a target corrected torque signal e(t). In order to achieve the above purpose, the noise removing means 5 extracts the average level dm of the detected torque signal d(t) and calculates the value from the corrected torque signal e(t) obtained at the time. An engine condition detecting means 6 detects the operating condition of the engine that drives the power train, and a predetermined intermediate time period is detected based on the output of the engine condition detecting means 6. a state change amount detection means 7 for detecting a change machine in the operating state; and a correction coefficient determining means 8 for determining an lfi correction distortion correction coefficient based on the engine operating state change needle detected by the state change amount detection means 7. Be prepared.

(Function) According to the present invention, since the correction coefficient can be determined based on the amount of change in the operating state of the engine, when the operating state of the engine changes significantly, in other words, the noise generated for the purpose of removing periodic noise. If the component y (t) is disturbed and periodic noise cannot be successfully removed from the detected torque signal d(t), the value of the correction coefficient α can be set to substantially affect the creation of the noise component y (t). Since it is possible to prohibit or reduce the degree of modification of the noise component y(t) by changing the noise component y(t) to "fl", a stable noise removal state can be maintained even if the operating state of the engine changes.

(Example of the invention) Embodiments of the present invention will be described below with reference to the accompanying drawings.

Fig. 2 shows a powertrain torque detection device a according to the present invention.
It is a block diagram showing an example.

Reference numeral 8 denotes a rotation sensor for detecting the rotation of the output shaft that is the object of torque detection, and the magnetostrictive torque sensor 16 and rotation sensor 18 are installed in the same manner as shown in FIG.

Reference numeral 40 denotes a throttle opening sensor that detects the throttle opening as an idle state of the engine.
A known potentio-type throttle opening sensor as shown in No.-1116 is used. The throttle opening sensor 40 outputs a detection signal having a relationship as shown in FIG. 3, for example, with respect to the throttle opening.

Returning to FIG. 2, 20 is a signal processing circuit that creates a corrected torque signal based on each detection signal from the magnetostrictive torque sensor 16, the rotation sensor 18, and the throttle opening sensor 40, and as a whole, the It has functions.

To explain the specific configuration of the signal processing circuit 20, the detection signal from the magnetostrictive torque sensor 16 (FIG. 8(a)
), an aftereffect rectifier circuit 22 that performs aftereffect rectification of the signal from this AC amplification circuit 21, and a low-pass filter 23 that removes high frequency components of the signal from the full-wave rectifier circuit 22. On the other hand, it has a waveform shaping circuit 24 that shapes the waveform of the detection signal from the rotation sensor 18 and converts it into a rectangular pulse signal. here,
The signal output from the above low-pass filter 23 is: 58
The waveform becomes as shown in Figure (b), and the rectangular pulse signal from the waveform shaping circuit 24 becomes a rectangular output with the same number N as the number of teeth of the parking gear every time the output shaft of the transmission rotates once. Further, 25 is a main control circuit in the signal processing circuit 20, and this main control circuit 20 uses the signal from the low-pass filter 23 as the detected torque signal d(t), and also refers to the signal from the waveform shaping circuit 24. Input as a pulse signal x (t). In addition, the detection signal output by the throttle opening sensor 40 is referred to as the throttle opening signal R(t).
Enter as . Specifically, this main control circuit 20 includes:
It has a CPU 30, ROM 31, and RAM 32, and
It has an interface circuit 33 composed of an A/D conversion circuit, a D/A conversion circuit, and the like.

26 is an addition/subtraction circuit that subtracts the noise component y(t), which is the output of the main control circuit 25, from the detected torque signal d(t), which is the output of the low-pass filter 23, and the output of this addition/subtraction circuit 26 is used to perform the desired correction. The torque signal e (t) is output from the signal processing circuit 20. Note that the state change amount detection means and correction coefficient determination means according to the present invention are realized as functions of the main control circuit 25 in this embodiment. Further, an example of the waveform of each of the above signals is shown in FIG.

Next, the above-mentioned mounting operation will be explained according to the flow shown in FIG. 5.

The processing flow is divided into two interrupt operations.

The flow shown in FIG. 5(a) calculates the amount of change in the throttle opening based on the throttle opening signal R(t) which is the output of the throttle opening sensor 40, and calculates the noise component based on the amount of change. It shows the process of determining the correction coefficient α used to create y. This process is performed at regular intervals, for example, every 4 m5ec.

Looking at the specific process, first, the throttle opening signal R value sampled at the time of a time interrupt is A/D converted and stored in the common register R (0-1). Next, the throttle opening signal R value sampled at the time of a time interrupt is stored in the common register R (0-1). is calculated and stored in register r (0
-2) The calculation in this process is obtained from the relationship between the throttle opening signal R and the actual throttle opening r as shown in FIG. Furthermore, the subsequent step (0-
The previous throttle opening r obtained by the process in 7). Calculate the throttle opening change amount H by calculating Lo and the throttle opening obtained this time "H" l r 0LII r l (0-3)
, and this throttle opening change amount H is stored in advance in ROM3.
Compare with the value of the predetermined level stored in 1 (0-4
).

This predetermined value is a comparison value for determining whether the value of the output shaft torque due to a change in the throttle opening has an adverse effect on the production of noise components, and is determined experimentally. Here, if the throttle opening change QH is smaller than a predetermined value, it means that the amount of change in the throttle opening r is not large enough to adversely affect the creation of the noise component, and therefore it is used to create the noise component. Set the correction coefficient α to αI (0<α, cl). In this case, the correction coefficient α is set in the range (0<α1×l) so that the noise component y can be substantially newly created.
It is determined to be a relatively large value within the range.

On the other hand, when the throttle opening change H exceeds a predetermined value, the correction coefficient
to a value α2 (0≦α2<αt < 1) within a small value range that has no substantial effect on the creation of noise components, for example, “0” (0-6), or when the correction coefficient α2 is “0”. In this case, the creation of noise components will be prohibited.

Next, in order to calculate the next throttle opening change amount H,
The current throttle opening degree is stored in the register r.LD (0-7). Through the above processing, the correction coefficient α used to create the noise component is determined.

Next, the creation of noise components that is performed in synchronization with the rotation of the rotation sensor 18 will be explained based on the flow shown in FIG. 5(b).

First, the reference pulse signal passed through the waveform shaping circuit 24 and the detected torque signal passed through the low-pass filter 23 are
They are as shown in Figures x (t) and d (t). Then, the main control circuit 25 (CPU 30) receives the reference pulse signal x.
(t) is used as an interrupt signal, and the reference pulse signal x
Each time (t) rises, a series of interrupt processing shown in FIG. 5(a) is performed. Here, the reference pulse signal X (
1) rises every time the output shaft of the transmission rotates by 1/N, and each rotational position (phase) at which the signal rises corresponds to the subscript k in FIG. 5(a).

Regarding the specific processing, the torque signal value d sampled at the time of interruption of the reference pulse signal is A/D converted and stored in the common register d (1-1).
2, the noise component yk calculated and stored at the same position (phase) last time is read out to the common register y, and (1-
2), the noise component y stored in this common register y is D
/A conversion and output (1-3).

When the noise component y is output from the main control circuit 25 in this way, this noise component y is subtracted from the detected torque signal d(t) at that point in the addition/subtraction circuit 26, and the corrected torque signal e is generated by the signal processing concerned. It is output from the circuit 20.

Furthermore, in the main control circuit 25, in order to obtain an output similar to the output (corrected torque signal e) from the addition/subtraction circuit 26, the detected torque signal d value stored in each of the common registers d and V is used. and the noise component y, find the difference e = dy (1-4), and then, the obtained e, the noise component y stored in the common register y, and the noise component y, obtained by the calculation described later. The average level d m and the correction coefficient α determined in steps (0-4) to (0-6) above
Based on.

A new noise component is calculated according to y + α fist (e − d m )0〈αcl, and the calculated value is newly stored in the common register y, and (1-5),
The noise component yk in the RAM 32 is rewritten to the noise component stored in the register y (1-5).

When the noise component yk in the RAM 32 is rewritten with the noise component newly calculated at the rotational position (phase) as described above, the average level dm to be used in the calculation of step (i-s) described above is extracted. Transition.

The register Wa stores the sum of past rotations (one cycle) of the detected torque signal d values sampled every 17N rotations of the output shaft (every 3607N degrees of phase) as described above. When the detected torque signal d value is sampled at the relevant position (phase) (1-1), the detected torque signal d value previously sampled at the same position (phase) and stored in the RAM 32 is stored in the register Wa. (2-1), add the newly sampled detected torque signal sand d value to the subtracted value, and register W.
Store in a.Thus, the register Wa maintains the state in which the sum of the detected torque signal d values sampled in the past cycle is stored. After that, the detected torque signal dk value in the RAM 32 is rewritten with the newly sampled detected torque signal d value (2-3), and the wholesale value stored in the register Wa is divided by the number of samplings N in the corresponding period and is averaged. Calculate the level dm (2-4). That is, this average level dm is the average value of the past cycle of the detected torque signal d(t).

From then on, every time the reference pulse signal x(t) rises, the main control circuit 25 samples the detected torque signal d value, outputs the noise component y, creates a new noise component y, and controls the average level dm. Processing such as extraction is repeated. During this process, each time the series of processes described above is completed, a register indicating the rotational position (phase) is incremented, and each time the value reaches N, the register k is reset to "O" (3-1 ), (3-2), (3-3).

As described above, the average level dm of the past cycle of the detected torque signal d(t) is calculated, and the value obtained by subtracting this average level dm from the corrected torque signal e value (e - d m ) and the previously calculated noise component y , and a correction coefficient α determined according to the braking condition of the vehicle.

A new noise component y is calculated according to y+α-(e-dm).

When the throttle opening r does not vary greatly, the correction coefficient becomes α1, and the noise component y(t) output from the main control circuit 25 gradually approaches the true noise component as shown in FIG. , accordingly, the output of the addition/subtraction circuit 26, ie.

The corrected torque signal e (t) converges to the true signal (DC level corresponding to the torque) included in the signal d(t).

On the other hand, when the throttle opening is greatly fluctuating, the value of the correction coefficient α is set to α2 = 0, so the calculation of the new noise component y is essentially y ← y, and the correction of the noise component y is will no longer be carried out.

As mentioned above, FIG. 4 shows the throttle opening signal R(t) which is the output of the throttle opening sensor 40, the correction coefficient α determined according to this throttle opening signal R(t), and the detected torque signal d. (t), its average level dm, noise component y, and the detected torque signal e(t) which is the output of this device.

In the same figure, while the engine throttle opening ``is fluctuating greatly, the correction coefficient α becomes ``0'';
” during the period from time t1 to time t2, modification of the noise component y is virtually prohibited. In this process, the main control circuit 2
The noise component (1) output from 5 becomes the noise component that is stored in the RAM 32 and is not rewritten (1-2) (1-3).

Here, time 1 in FIG. If the previously created noise component (the noise component already stored in the RAM 32) has been sufficiently focused into the true noise component through the process shown in Figure 9, the result obtained from the arithmetic circuit 2-6 is The corrected torque signal (output signal of the signal processing circuit 20) is the fourth
As shown in Figure e(t), it follows the torque fluctuation.

In addition, in the above process, step (2-1) or 2-4
) The average level dm obtained by the process deviates from the DC level at a point where the DC level of the detected torque signal fluctuates greatly, as shown in dm shown in FIG. 4, for example.

In this way, according to this embodiment, the correction coefficient α is set to "0" so as to prohibit the correction of the noise component y while the throttle opening degree r fluctuates greatly. It is possible to maintain a stable state of noise removal.

It should be noted that even if the value of the correction coefficient α is determined to be a small value that does not substantially affect the creation of the noise component so that the degree of correction of the noise component y is extremely small while the throttle opening degree r fluctuates greatly. Almost the same effect can be obtained. In addition, in this embodiment, the correction coefficient α is set to the minimum value only when the throttle opening is changing, but considering the delay in response of the output shaft torque to changes in the throttle opening, the throttle opening is changed. The correction coefficient α may be kept at the minimum value for a predetermined period of time after the correction is completed.

FIG. 6 shows a flow for determining the correction coefficient α based on the intake air amount of the engine. In the case of this second embodiment, a sensor for detecting the intake air amount of the engine is required in place of the throttle opening sensor 40. The specific process flow is basically the same as that of the above embodiment, and includes A/D converting the intake air amount signal Q value sampled from the sensor at the time of the 1-hour interrupt and storing it in the common register Q (2-1). , From this Q value, the intake air amount q
is calculated and stored in register q (2-2).The calculation in this case is determined based on the relationship between the intake air amount signal Q and the actual intake air amount as shown in FIG. Then, from the previous intake air amount qOLD obtained in the previous process and the intake air amount QOLD calculated this time, H2←IQOLD-Ql
Find the amount of change H2 in the intake air amount by calculating (2-3),
The method for determining the correction coefficient α based on the amount of change H2 is the same as in the previous embodiment.

FIG. 8 shows the flow when determining the correction coefficient α based on the intake pipe pressure of the engine. In the case of this third embodiment, a sensor is required to detect the intake pipe pressure of the engine. The processing flow in this case is basically the same as that of the above embodiment, and the intake pipe pressure signal P value sampled from the intake pipe pressure sensor at the 1-hour interrupt is converted to A/D.
In this case, it is exchanged and stored in the common register P (3-1), and the intake pipe pressure p is calculated from this P value and stored in the register P (3-2).

The pressure p is determined from the relationship between the intake pipe pressure signal p and the actual intake pipe pressure p as shown in FIG. and.

From the intake pipe pressure poc, D obtained in the previous process and the intake pipe pressure P obtained this time, calculate the amount of change H3 in the intake pipe pressure by calculating H3←l POLO-PI (3-3).
The method for determining the correction coefficient α based on H3 is the same as in the first embodiment.

FIG. 1O and No. 11 show a processing flow when determining the correction coefficient α based on the amount of change in the rotational speed of the engine. In the case of this fourth embodiment, an optical crank angle sensor built into the distributor, for example, is used to detect the rotational speed of the engine.

First, the process shown in the flowchart of FIG. 10 is for calculating the engine rotation speed, and is performed in synchronization with the rise of the crank angle signal w(t) output from the crank angle sensor. The crank angle signal w (t) rises every time the crankshaft rotates by 2 degrees, as shown in FIG. 12(b).

Looking at the specific processing, first, the crank angle signal w(
CPU sampled and stored at the same time as t) rises
The value FRC of the built-in free run counter 30 is stored in register C (4-1).

This free run counter is a 1, for example, 16-bit counter that increments by 1 every psec.

Next, from the free run counter value COLD obtained in the previous processing obtained in the process of step (4-4) described later and the free run counter value C obtained this time, the crank is calculated by calculating D E LTA" l C - C0LD I. Find the time interval DELTA between the rises of the crank angle signal w (t) that rises every 2 degrees on the axis (4-2
), here, the free run counter is incremented every 1 ILsec, so the determined time interval is DELTAxlO
−' seconds.

moreover.

The engine rotational speed w [RPM] is determined by the calculation (4-3). The time interval between 2 degrees of the crankshaft is DELT
Since AxlO-' seconds, one rotation of the crankshaft (360°
) It takes DELTAxlO-'x180 seconds to rotate.

In other words, it rotates l/DELTAX 10-'X180 per second. Since it is the number of revolutions per minute, it is further multiplied by 60 to obtain the rotation speed w [RPM] from the above formula.

Next, the current free run counter value C is stored in the register COLD for the next time interval DELTA calculation (4-4).

Through the above processing, the rotational speed W of the engine used in the processing shown in the flowchart of FIG. 11 is determined.

The flow shown in FIG.
In the process performed for each ec, a change amount H4 in the rotational speed W is determined based on the engine rotational speed W, and a correction coefficient α used to correct the noise component y is determined based on this change amount H4.

Looking at the specific process, first, the steps (
The previous engine rotational speed W obtained in the process of 5-7).

The engine rotational speed change H4 is calculated from LD and the rotational speed W obtained at that time by calculating H4←l W ot, o W l (5-3
).

This engine speed change H4 is compared with a predetermined value previously stored in the ROM 31 (5-4). This h is a value obtained experimentally based on the relationship between the rank angle and the engine rotation speed as shown in Fig. 12 (a), and the change in the torque value due to the change in the engine rotation speed has an adverse effect on the creation of the noise component y. This is the comparison value for whether or not to give.

Here, when the engine rotation speed change H4 is less than a predetermined value, the change in engine rotation speed is small, so the correction coefficient α is set to α1 (0<α1 × 1) as described above (5-5). .

On the other hand, when the engine rotational speed change H4 exceeds the predetermined 4tih, the change in the engine rotational speed W is large, so as mentioned above, the correction coefficient α is set to α2 smaller than α1, for example, “0” (5-6) .

Next, for the next engine speed change H4 calculation,
The engine rotational speed W obtained at that time is stored in the register W. Store it in the LD (5-7).

With the above processing, the processing for determining the correction coefficient α used to create the noise component is completed.

As described above, since the device according to the present invention determines the correction coefficient used to create noise components from various engine operating conditions, the correction coefficient can be changed depending on the amount of change in the engine operating condition. , speedy and stable noise removal can be achieved.

(Effect of the invention) According to the present invention, as has been explained further.

Since the value of the correction coefficient used to create the noise component is determined according to the amount of change in the operating state of the engine,
Even if the amount of change in the engine operating state is large, that is, the noise component created to remove periodic noise is disturbed, and as a result, periodic noise cannot be removed successfully from the detected torque signal, the correction coefficient The value of the noise component creates a substantial shadow! By changing the value to i1 within a small value range without any periodic noise, it is possible to maintain a stable state in which periodic noise is removed.

Therefore, even if the rotating shaft torque of the power train varies greatly while the vehicle is running, the power train can be reliably controlled based on a stable corrected torque signal from which periodic noise has been removed.

[Brief explanation of drawings]

Fig. 1 is a claim correspondence diagram showing the configuration of the present invention, Fig. 2 is a block diagram showing an example of a device that realizes the power train torque detection device according to the present invention, and Fig. 3 is a diagram showing the throttle opening r.
FIG. 4 is a diagram showing the relationship between each signal waveform and the correction coefficient when the operating state of the engine changes, and FIG. 5 is the processing flow of the device shown in FIG. 2. FIG. 6 is a flowchart according to the second embodiment of the present invention, FIG. 7 is a flowchart showing the relationship between the intake air amount detection signal and the actual intake air amount, and FIG. 8 is a flowchart according to the second embodiment of the present invention. A flowchart according to the third embodiment, FIG. 9 is a graph showing the intake pipe pressure detection signal and the actual intake pipe pressure, and FIGS. 10 and 11 are flowcharts showing the fourth embodiment of the present invention. Figure 12 is a graph showing the relationship between engine speed and crank angle, Figure 13 is a block diagram showing a conventional noise removal device, and Figure 14 is an example of a device that detects the torque of the output shaft in an automatic transmission. Figure 15 is a diagram showing the waveforms of output signals etc. from the magnetostrictive torque sensor, Figure 16 is a diagram showing the waveforms of each signal used in the periodic noise removal device, and Figure 17 is a diagram showing the waveforms of the output signals from the magnetostrictive torque sensor. FIG. 18 is a block diagram showing the control equipment; FIG. 18 is a diagram showing an example of the output shaft torque waveform of the automatic transmission during gear shifting; FIG. 19 is a diagram showing an example of the waveform of each signal when the operating state of the engine changes. . l...Synchronous adaptation means 2...Average level extraction means 3...Subtraction means 4-)
- Luk detection hand 5... Noise removal means 6...
Engine state detection means 7...State change amount detection means 8...Modification coefficient determination means Fig. 1 Fig. 3 Throttle darkness 71r Fig. 5 Fig. 6 (tfHz+-ru) oa on! It (kg/h1
Fig. tO Fig. 13 Fig. 15 Fig. 16
−−−−−−−−−−−−−−−−−−−Fig. 17

Claims (1)

[Scope of Claims] Torque detecting means for detecting the rotating shaft torque of the power train; Noise removing means that sequentially corrects and generates a noise component in each phase of periodic noise using a predetermined correction coefficient, and subtracts this noise component from the detected torque signal to obtain a target corrected torque signal. In the powertrain torque detection device, the noise removal means extracts an average level of the detected torque signal and subtracts that value from the corrected torque signal obtained at the relevant point in time to eliminate deviations that contribute to the creation of noise components. an engine state detecting means for detecting the operating state of the engine driving the power train; and a state change amount detecting means for detecting the amount of change in the operating state of the engine based on the output of the engine state detecting means. A torque detection device for a power train, comprising correction coefficient determining means for determining the correction coefficient based on the amount of change in the operating state of the engine detected by the amount of change in state detection means.