JPH10201274A - Speed controller for rotary machine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent the deterioration of the time-base resolution of the measurement of rotational speed due to rotational speed in a discrete time system for controlling a rotary machine, by updating a control input to the rotary machine at the time intervals of a sampling time. SOLUTION: A rotational-speed measuring means 202 measures rotational speed on the basis of the time intervals of a pulse signal from a pulse generating means 201, and a control means 300 controls a control input to the rotary machine 200 by a discrete time system on the basis of the rotational speed. The rotational-speed measuring means 202 measures rotational speed from the time intervals of a pulse signal immediately before a time when the control input is updated and a pulse signal immediately after the updated time, a sample-time setting means 301 of the control means 300 sets time intervals obtained by multiplying the time intervals of the pulse signal by an integer of 1 or more as a sampling time, a control-input processing means 302 updates the control input corresponding to rotational speed and the sampling time, and a control-input update means 303 updates the control input to the rotary machine 200 at the time intervals of the sampling time at that time.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improvement of a rotation speed control device of a rotary machine having means for generating a pulse signal corresponding to a rotation position and measuring a rotation speed from a time interval of the pulse signal. It is.

[0002]

2. Description of the Related Art As an apparatus for controlling the rotation speed of a rotary machine, for example, as shown in FIG. 16, a DC motor 51 and a tachometer 52 for measuring the rotation speed are combined with a controller 50 having an amplifier. Is conventionally known.

[0003] Referring to FIG.
Generates a voltage proportional to the rotation speed of the DC motor 51 in accordance with the principle of the induced electromotive force. The A / D is built in a controller 50 such as a microcomputer.
A / D conversion is performed by a converter to obtain the rotation speed of the motor.

Since the characteristics of the DC motor 51 do not depend on the rotation angle, the difference between the rotation speed of the motor and the target value is obtained regardless of the rotation angle of the motor, and this difference is multiplied by an appropriate gain. The obtained value is converted into a voltage value by a built-in D / A converter, and is given to the DC motor 51, so that the rotation speed of the motor can be controlled to a target value. For example, if the DC motor 51 has a first-order lag characteristic expressed by 1 / (S + 1) and performs the above simple feedback control with the feedback gain = 5, the gain = 1
In the step response, the control characteristic becomes as shown by the line a in FIG. 18 and a steady deviation remains.

[0005] Therefore, if PI control (proportional integral control) is performed so that the integral gain is set to 5 so that the steady-state deviation is not left and the integral gain is set to 5, the characteristic shown by the line b in FIG. It is widely known in theory.

A microcomputer widely used as a controller for controlling a motor is a discrete-time (discrete) system based on a clock signal. Therefore, the above-described feedback control must be executed as a discrete time. .

Therefore, as shown in FIG. 17, a controller 60 mainly composed of a microcomputer is provided with an integrator 61 to control a motor 51 having a first-order lag characteristic as expressed by the above equation (1). When the same control as described above is performed with the rotation speed sample time T set to 0.1 second, the proportional gain set to 5, and the integral gain set to 1, the control result as shown by the line c in FIG. It is widely known as control theory.

However, the above considerations make it possible to provide a control input to the DC motor 51 by multiplying a value at a certain discrete time by a gain as an ideal discrete time system.

That is, Y (k) is a control input, U (k) is a target value, e (k) is the rotation speed of the motor 51 obtained from the tacho generator 52, IG (k) is an internal variable, and proportional gain is P. , IG (k) = IG (k−1) + 0.1 × (U (k) −e (k)) Y (k) = P × (U ( k) −e (k)) + I × IG (k) (2) This is a case where a control law represented by the following equation is executed.

The above equation (2) satisfies causality as a discrete time system.
Since the calculation time of the microcomputer is required, such a simple control law is required, and when a more complicated control law taking into account the characteristics of the control target is performed, the required calculation time is required. ) Is difficult to realize.

Then, IG (k−1) = IG (k−2) + 0.1 × (U (k−1) −e (k−1)) Y (k) = P × (U (k−1) ) -E (k-1)) + I.times.IG (k-1) (3) The control rule is convenient for realizing the microcomputer.

However, as shown in the above equation (3), simply 1
If the control law is delayed by the sample (K-1),
As shown by a line d in FIG. 18, a control result different from that of the line c is obtained.

Further, the processing timing is as follows: IG (k−2) = IG (k−3) + 0.1 × (U (k−2) −e (k−2)) Y (k) = P × If (U (k−2) −e (k−2)) + I × IG (k−2)... (4) is delayed by two samples (k−2), FIG. As shown by the line e, the control result is much different from the line c.

In the example shown in FIG. 18, a relatively long sample time is used so as to clarify the difference between the sample times. However, shortening the sample time does not change the essence of the problem. .

Since the above-mentioned equations (2), (3) and (4) are distinctly different systems as the discrete time control system, such a result is natural.

That is, in the control of the discrete time system,
Since a completely different control law is executed depending on which information of the sample time is used, the difference of the sample time can be obtained by applying the corresponding control law depending on which information of the sample time is used. It is possible to minimize the difference in the control result due to.

That is, in contrast to the control device shown in FIG.
9, a controller 6 including a correction filter 62
If the control rule based on 0 'is used, as shown in FIG. 20, as compared with the line b without one sample delay, the line c when the information U (k-1) delayed by one sample is used, as shown in FIG. Even if the information of one sample delay is used, almost the same control result can be obtained.

Therefore, what is important in the control of the discrete time system is not to acquire the information without a sample delay, but to clarify which sample time the information is, and to make the control side according to the time of the acquired information. Is important in performing appropriate control, and if control is performed based on information one sample or more in advance, the time required for calculation can be extended, and control can be performed with a device having a simpler configuration. It can be realized.

[0019]

In the prior art shown in FIG. 16, the rotational speed of the DC motor 51 is
The tachometer 52 is used to obtain a measurement value at a desired time, but when a microcomputer is used as the controller 50, the time interval at which a pulse is generated is measured using an encoder that outputs a pulse according to the rotation angle. This is more general because the device for obtaining the rotation speed is simpler.

However, when the rotation speed is measured by the above-mentioned conventional control device and the rotation speed is controlled by the discrete time system as described above, the following three problems occur.

The first problem is that the rotational speed calculated by the microcomputer measures the time interval of the pulse from the encoder, so that the discrete time system requires a discrete time T (k). Since it is calculated not as an instantaneous value but as an average value within a pulse time interval and takes an average value within a certain time interval, there is a problem that the resolution of the measured value in the time axis direction deteriorates.

Further, since the deterioration of the resolution in the time axis direction depends on the time interval of the pulse, that is, the rotation speed of the rotating machine, there is a problem that the resolution is not constant.

Therefore, it is not possible to reduce the influence by increasing the number of pulses generated per rotation of the encoder and making the time interval between the pulses relatively small with respect to the sample time of the discrete time system. However, when the time interval between pulses is short, sufficient measurement accuracy cannot be obtained unless the resolution of the minimum time for measuring time, that is, the system clock cycle of the microcomputer is set finer at high speed rotation. The increase in the frequency of the clock makes simple control by the microcomputer difficult. For example, when the encoder generates 10 pulses per rotation, the time interval is 10 milliseconds when the rotation speed is 10 rotations / second, and when the control sampling time is 10 milliseconds, the rotation speed is measured. Is the average value of the entire sample time for controlling. Assuming that this would cause a problem in control performance, if the encoder generates 1000 pulses per rotation and obtains 0.1% measurement accuracy in the case of high-speed rotation of 1000 rotations / sec, 1
A time step of nanoseconds is required, and the use of a microcomputer using a 1 GHz system clock is required.

A second problem is that the time at which the measurement is performed is defined by the time at which the pulse is generated, which is independent of the discrete time step defined by the discrete time system.

That is, there arises a problem that it is appropriate to use the measured rotation speed as a measurement result at a time interval specified by the discrete time system.

When a pulse is input from the encoder at the timing shown in FIG. 21A, the measurement time is determined by T (k) at which the discrete time system executes control.
It is considered appropriate to approximate to the time point, but when a pulse is input at the timing shown in FIG. 21B, the measurement time is set to T (k) rather than to the time point of T (k). It is considered appropriate to approximate the time point of (k + 1).

This problem becomes a very important problem that cannot be ignored when the sample time of the discrete time system is made as long as possible to reduce the cost or when a more complicated control law is executed.

The third problem is that, as shown in FIG. 21C, when the rotating speed of the rotating machine is very low, the pulse may not arrive within the sample time set by the discrete time system. According to the idea of a discrete-time system, there is no input unless a pulse arrives, that is, it must be determined that the rotating machine has stopped,
In reality, the rotating machine is not always stopped, and may be rotating at an extremely low speed.

In order to avoid such a problem, there are known measures such as increasing the number of pulses output per rotation, setting a longer sample time in a discrete time system, or extrapolating from past pulses. For example, in the case of high-speed rotation in which a large number of pulse signals per unit time are input by using a frequency divider as in Japanese Patent Application Laid-Open No. 59-193361, the pulse signal is divided and used. On the other hand, in the case of a low-speed rotation in which a small number of pulse signals are input, there is a method in which the number of frequency divisions of the pulse signal is reduced. Some are known to change, but these are aimed at improving the measurement accuracy,
As a result, there is an effect that the time interval required for measuring the rotation speed is fixed to some extent. That is, it is conceivable that the measurement cycle and the control cycle are not synchronized even when a discrete time system having a predetermined fixed sample time is used. However, in the above-described conventional example, the time taken to take the average pointed out in the first problem becomes long in high-speed rotation, and the time-axis resolution is deteriorated even in high-speed rotation in which the time-axis resolution is relatively good. As a result, the control performance is reduced.

The present invention has been made in view of the above problems, and in a discrete time system for controlling a rotating machine, it is possible to prevent the time axis resolution of rotation speed measurement from deteriorating depending on the rotation speed. Another object of the present invention is to provide a speed control device for a rotating machine that can reliably measure and control the rotation speed even when the rotation speed is extremely low.

[0031]

According to a first aspect of the present invention, as shown in FIG. 22, a pulse generating means 201 for generating a pulse signal corresponding to the rotational position of a rotary machine 200, Time interval TPuls of pulse signal
rotation speed measuring means 20 for measuring the rotation speed on the basis of e
And control means 300 for controlling a control input to the rotary machine 200 by a discrete time system based on the rotation speed.
In the speed control device for a rotary machine, the rotation speed measuring means 202 calculates a time interval TPulse between a pulse signal generated immediately before the time when the control input is updated and a pulse signal generated immediately after the time. While measuring the rotation speed, the control means 300 includes a sample time setting means 301 for setting a time interval obtained by multiplying the time interval TPulse of the pulse signal by a predetermined integer j of 1 or more as a sample time; A control input calculating means 302 for updating a control input according to the rotation speed measured by the 202 and the sample time; a control input updating means 303 for updating a control input to the rotating machine 200 at the time interval of the sample time; Is provided.

In a second aspect based on the first aspect, the control input updating means includes a time interval TPulse of the pulse signal from a time at which the pulse signal is generated.
Is multiplied by a predetermined integer j, and a control input is updated after a lapse of a time obtained by adding a value α obtained by multiplying the time interval TPulse of the pulse signal by １ ／.

In a third aspect based on the first aspect, the control input updating means includes a time interval TPulse of the pulse signal from a time when the pulse signal is issued.
The control input is updated after a lapse of a time obtained by adding a value obtained by multiplying a time interval TPulse of the pulse signal by a predetermined constant 1 / m to a time interval obtained by multiplying the pulse signal by a predetermined integer j.

In a fourth aspect based on the third aspect, the sample time setting means sets the constant 1 / m to a small value when the rotating machine accelerates, while the rotating machine decelerates. In this case, the constant 1 / m is set to be large.

According to a fifth aspect of the present invention, there is provided a pulse generating means 201 for generating a pulse signal corresponding to the rotational position of the rotary machine 200, and a rotational speed based on a time interval TPulse of the pulse signal from the pulse generating means 201. A rotation speed measuring means 202 for measuring the rotation speed, and a control means 300 for controlling a control input to the rotating machine 200 based on the rotation speed by a discrete time system. 202 is a time interval TPuls between a pulse signal generated immediately before the time when the control input is updated and a pulse signal generated immediately after the time.
e, while measuring the rotation speed, the control means 300
Is a sample time setting means 301 for setting a time interval obtained by multiplying the time interval TPulse of the pulse signal by a predetermined integer j equal to or greater than 1 as a sample time, and a pulse interval for comparing the time interval TPulse of the pulse signal with a predetermined value. When the comparison result between the comparing means 304 and the pulse interval comparing means 304 is that the time interval TPulse of the pulse signal is less than a predetermined value, the sample time is changed to a preset value, while the time interval TPulse is not less than a predetermined value. In this case, a sample time changing means 305 for setting a value set by the sample time setting means, and a control input in accordance with the rotation speed measured by the rotation speed measuring means 202 and the sample time from the sample time changing means 305. The control input calculating means 302 for updating and the updating of the control input to the rotary machine 200 And a control input update means 303 to perform a sample time changing means 305 configured sample time intervals in.

[0036] In a sixth aspect based on the fifth aspect, the rotation speed measuring means is configured to determine whether the time interval TPulse of the pulse signal is less than a predetermined value as a result of comparison by the pulse interval comparing means. A control prohibiting means is provided for measuring the rotation speed based on a pulse signal generated after updating the control input and for inhibiting the rotation speed measurement after the rotation speed measurement until the next control input update is performed.

The seventh aspect of the present invention provides a rotating machine 200 having a plurality of rotating shafts, a plurality of pulse generating means 201 for respectively generating pulse signals corresponding to the rotational positions of these rotating shafts, A rotation speed measurement unit 202 that measures the rotation speed based on the time interval TPulse of the pulse signal from the pulse generation unit 201, and a control unit that controls a control input to the rotary machine 200 based on the rotation speed by a discrete time system. 300, wherein the rotational speed measuring means 2 is provided.
02 measures the rotation speed from the pulse signal generated immediately before the time at which the control input is updated and the time interval TPulse of the pulse signal generated immediately after, while the control means 300 calculates the rotation speed of each pulse signal. Time interval TPuls
e, a longest pulse interval selecting means 306 for selecting the largest one, and a time interval obtained by multiplying the time interval TPulse of the pulse signal selected by the longest pulse interval selecting means 306 by a predetermined integer j of 1 or more. Sample time setting means 301, which is set as: a control input calculating means 302, which updates a control input according to the rotation speed measured by the rotation speed measuring means 202 and the sample time;
Control input updating means 303 for updating the control input to the rotary machine 200 at the time interval of the sample time.

In an eighth aspect based on any one of the first, fifth to seventh aspects, the sample time setting means includes a time interval TPuls of the pulse signal.
When calculating a time interval obtained by multiplying e by a predetermined integer j, a constant changing means for changing the integer j according to the time interval TPulse of the pulse signal is included.

[0039]

According to the first aspect of the present invention, the control input for controlling the rotation speed of the rotating machine is updated at a time interval obtained by multiplying the time interval TPulse of the pulse signal by a predetermined integer j of 1 or more. At the same time, the calculation of the control input is performed by calculating the time interval TP between the pulse signal generated immediately before the time when the previous control input was updated and the pulse signal generated immediately thereafter.
Calculated using the rotation speed obtained from ulse and the time interval of the control input update, the sample time of the discrete time system is always longer than the time interval of the pulse of the rotation measurement, and the rotation measurement is performed within the sample time of the discrete time system. The third problem of the conventional example that the pulse does not arrive cannot occur, and the rotation speed is obtained by the time interval between the pulse signals immediately before and after the time when the control input was last updated, so that the discrete time The second problem shown in the above-described conventional example that the time value is unknown when viewed as a system is also solved. Further, the time interval TP of the rotation measurement pulse
When ulse is longer, the sample time of the discrete time system is also an integer multiple thereof, so that the time axis resolution with respect to the sample time of the discrete time system of the rotation speed measurement becomes constant in the entire rotation speed range. As described above, the time axis resolution does not decrease or the measurement cannot be performed during the extremely low-speed rotation.

As described above, the concept of the sample time in the discrete-time system makes it possible to synchronize the sample time of the discrete-time system with the time interval of the pulse to be measured (rotation speed measurement) to the rotation speed. Each coefficient can be changed in real time in accordance with a change in the sampling time such as the integration operation by the information before one or more samples, and the three problems shown in the conventional example can be solved to solve the problem of the rotating machine. The accuracy of speed control can be greatly improved.

According to a second aspect of the present invention, after the control input is updated last time, the time interval between the pulse signal generated in accordance with the rotation of the rotary machine and the time interval of the pulse signal is set to 0.
When a new control input is updated after a lapse of a time obtained by adding a predetermined value α to a time interval multiplied by a predetermined integer j including
The control input is calculated using the pulse signal generated immediately before the time at which the previous control input was updated, the time interval between the pulse signals generated immediately thereafter, and the sample time, and at the same time, the predetermined value α Is set to の of the time interval of the pulse signal, so that when viewed as a discrete-time system, it is possible to more appropriately solve the second problem of the above-described conventional example in which the value of the time is unknown, Further, the third problem shown in the conventional example, that is, the problem that the rotation speed measurement pulse does not come within the sample time of the discrete time system, is solved, and the control input can be accurately performed even at extremely low speed rotation of the rotating machine. Can be obtained, the maximum margin can be obtained for acceleration / deceleration changes, and the introduction of α allows the control means to secure the time required for calculation. Become.

According to a third aspect of the present invention, a value obtained by multiplying the time interval TPulse of the pulse signal by a predetermined constant 1 / m is added to a time interval obtained by multiplying the time interval TPulse of the pulse signal by a predetermined integer j. When applied to a rotating machine having different acceleration / deceleration characteristics, for example, a transmission for a vehicle, the case where the rotation measurement pulse is not located at the center of the sample time of the discrete time system is one of the third problems shown in the conventional example. In this way, the problem that the measurement pulse does not come within the sample time of the discrete-time system during extremely low-speed rotation can be easily eliminated, and accurate speed control can be performed over the entire speed range from high-speed rotation to low-speed rotation. Becomes possible.

According to a fourth aspect of the present invention, when a value obtained by multiplying a time constant TPulse of a pulse signal by a predetermined constant 1 / m is added to a time interval obtained by multiplying the time interval TPulse of the pulse signal by a predetermined integer j. When the rotating machine accelerates, the constant 1 / m is set small, while when the rotating machine decelerates, the constant 1 / m is set large. If the present invention is applied to a machine, even in a transient state in which the sampling time interval changes every moment, such as during acceleration / deceleration, the time required for the calculation by the control means can be reliably secured, and control accuracy can be secured. It is.

According to a fifth aspect of the present invention, when the time interval TPpulse of the pulse signal is less than a predetermined value, the sample time interval is changed to a preset value. in case of,
By performing the control at a predetermined fixed sample time, it is possible to secure the calculation time at the time of high-speed rotation, and to perform accurate control over the entire speed range.

According to a sixth aspect of the present invention, when the time interval TPpulse of the pulse signal is less than a predetermined value, the rotation speed is measured by the pulse signal generated after the control input is updated, and after the rotation speed measurement, the next rotation is measured. Since the rotation speed measurement is prohibited until the control input is updated, in a situation where a large number of measurement pulses are generated during high-speed rotation, the pulse measurement process is prevented from increasing the calculation load on the control means, and various calculations are averaged. By doing so, control accuracy during high-speed rotation can be ensured.

According to a seventh aspect of the present invention, when a plurality of rotation speeds are measured based on time intervals of pulse signals generated from a plurality of rotation axes to control the rotation speed, the rotation speed of the rotating machine is changed. By calculating the control input as the sample time interval by multiplying the largest value among the time intervals of the plurality of pulse signals by an integer j of 1 or more, for example,
Even when controlling the speed ratio between the input shaft and the output shaft, such as the speed ratio control of a continuously variable transmission of an automobile, accurate control can be performed over the entire speed range.

According to an eighth aspect of the present invention, when calculating a time interval obtained by multiplying the pulse signal time interval TPulse by a predetermined integer j, the integer j is converted to the pulse signal time interval TPulse.
The sampling time interval changes according to the time interval of the measurement pulse, so that it is possible to secure the calculation time of the control means at the time of high-speed rotation, and to improve the control accuracy in the entire speed range. Can be secured.

[0048]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 shows an example of controlling a motor 80 connected to a pulse encoder 81 by a controller 82 mainly composed of a microcomputer.

The controller 82 includes a CPU 83, a RAM
84, a ROM 85, a capture register 86, a free-run counter 87, a D / A converter 88, a timer 89, and a clock generator 79, and the pulse from the pulse encoder 81 is supplied to the capture registers 86 and C
It is input to the PU 83 and sent to the CPU via the D / A converter 88.
A motor 80 as a rotating machine is driven by a control input OUT from 83.

In this example, the pulse encoder 81 connected to the motor 80 outputs 10 pulses each time the motor makes one revolution.
It is assumed that the pulse is output twice.

From the pulse encoder 81 to the controller 8
The pulse input to 2 is supplied to the capture register 86 as a trigger signal.

The capture register 86 reads and holds the value of the free-run counter 87 that increments the system clock generated by the clock generator 79 in synchronization with the falling edge of the input pulse.

On the other hand, the pulse input from the pulse encoder 81 is also input to the CPU 83, and the CPU 83 is notified of the occurrence of the pulse input by the interrupt signal IRQ1.

The timer 89 has a register (not shown) readable and writable by the CPU 83, decrements a predetermined value set by this register in synchronization with the system clock, and interrupts every time the value of the register becomes 0. A signal IRQ0 is generated to notify that the value of the register has become 0.

The CPU 83 executes the processing shown in FIGS. 2 and 3 stored in the ROM 85, issues a predetermined command to the D / A converter 88, and the D / A converter 88 outputs a voltage corresponding to the command. When the motor 80 is driven by this voltage, the D / A converter 88 holds this voltage at the same time.

Next, the control operation performed by the CPU 83 will be described with reference to the flowcharts of FIGS.

First, the flowchart of FIG. 2 is executed every time the interrupt signal IRQ0 from the timer 89 is generated. In step 91, the value of the variable Timer is set to True, that is, 1 and the value TR of the register of the timer 89 is set to TR. Is set to a preset value of the constant MaxTimer.

In the next step 92, the value of the variable 0UT is sent to the D / A converter 88, and this time is set as the control input update time.

On the other hand, the flowchart of FIG. 3 is executed each time the pulse interrupt signal IRQ1 from the pulse encoder 81 is generated. In step 93, the value of the variable Now_c is substituted for the variable Last_c,
The value CR of the capture register 86 is substituted for the variable Now_c.

In step 94, when it is determined that the value of TimerFlag is True, that is, 1 is obtained,
Proceeding to the processing of step 95, the variable TPulse (pulse interrupt signal IRQ1 generation interval) and the variable Tsampl
While calculating e (timer interrupt signal IRQ0 generation interval), otherwise, the process ends.

The variables TPpulse and Tsample performed in step 95 are obtained from Last_c and Now_c in step 93 based on the following equations.

TPulse = Now_c-Last_c Tsample = (j + 1) × TPulse Here, j is a predetermined constant.

In the next step 96, step 9
The timer register value TR set in step 1 is reset as follows: TR = Tsample- (MaxTimer-TR) At the same time as setting so that the timer interrupt signal IRQ0 is generated after a lapse of Tsample from the previous timer interrupt, TimerFlag is set. Is set to False, that is, 0, and the processing from step 95 to step 98 is prohibited until the next timer interrupt occurs.

In step 97, the rotational speed Rev of the motor 80 is set to the number of generated pulses per rotation of the pulse encoder 81 = 10 and the pulse time interval TPulse.
From the following equation.

Rev = Clock / TPulse / 10 Clock is the cycle of the system clock.

Further, a deviation E is obtained from a difference TRev-Rev between the rotation speed Rev of the motor 80 and a predetermined target rotation speed TRev.
rr is determined.

Next, at step 98, an integral value IErr of the error Err is calculated as follows: IErr = IErr + Tsample × Err Further, based on the integral value IErr, 0UT as a control input is calculated as follows. Is done.

OUT = P × Err + I × IErr where P is a proportional gain and I is an integral gain.

By the processing of FIGS. 2 and 3 as described above,
The control input OUT is updated at the timing shown in FIG.

That is, in FIG. 4A, the time Ts
When the timer interrupt signal IRQ0 occurs in (k),
The control input OUT is sent to the D / A converter 88 from step 92 in FIG.
rFlag = True is set.

Therefore, the processing after step 95 in FIG. 3 is executed, and the pulse interrupt I immediately after time Ts (k) is performed.
TPulse0 is calculated in step 95 from the time Tp (k) of RQ1 and the value of Tp (k-1) immediately before, and the time Tsam is j times this value, and in the example of FIG.
ple0, and Tsam from time Ts (k).
After the elapse of ple0, the next timer interrupt time Ts (k
+1).

At the time Ts (k + 1), the next control input OUT is sent to the D / A converter 88.

In the above processing, the control input OUT at the time Ts (k + 1) outputs the pulses Tp (k) and Tp (k-1) emitted at times before and after the time Ts (k) to the rotation speed Rev of the motor 80. Ts
When the constant j for calculating the sample is relatively large, the rotation speed Rev can be treated as a measurement value at the time Ts (k), and the resolution in the time axis direction is good.

When the motor 80 is started, it goes without saying that the variables are initialized so that the above processing can be performed without any problem.

In this way, the control input OUT for controlling the rotation speed of the motor 80 is updated at the time interval obtained by multiplying the time interval TPpulse of the pulse signal by a predetermined integer j of 1 or more, and at the same time, the operation of the control input OUT is calculated. From the pulse signal Tp (k-1) generated immediately before the time Ts (K) at which the control input OUT was updated last time and the time interval TPulse between the pulse signal Tp (k) generated immediately after the time Ts (K). Since the calculation is performed using the rotation speed and the time interval of updating the control input, the sampling time of the discrete time system is always longer than the time interval of the rotation measurement pulse, and the rotation measurement pulse does not arrive within the discrete time system sample time. The third problem shown in the conventional example described above cannot occur, and the rotational speed of the motor 80 is changed to the time Ts (K Pulse signal Tp just before and just after the
(K) and Tp (k-1), which solves the second problem shown in the above-mentioned conventional example that the value at which time the value is unknown when viewed as a discrete-time system is solved. You.

Further, the time interval TPu of the rotation measurement pulse
When lse is longer, the sample time of the discrete time system is also an integer multiple thereof, so that the time axis resolution with respect to the sample time of the discrete time system of the rotation speed measurement becomes constant in the entire rotation speed range. As described above, the time axis resolution does not decrease or the measurement cannot be performed during the extremely low-speed rotation.

As described above, by the concept of the sample time in the discrete time system, the sample time of the discrete time system and the time interval of the pulse whose rotational speed is measured can be calculated as follows.
It is possible to synchronize with the rotation speed, and it is possible to change each coefficient in real time according to the change of the sampling time such as the integration operation by the information before one sample or more. By solving the problem, the accuracy of speed control of the rotating machine can be dramatically improved.

FIG. 5 shows the second embodiment. Step 96 of the first embodiment shown in FIG.
The other configuration is the same as that of the first embodiment.

In step 100 of FIG. 5, the variable α is set as 1/2 of the value of the variable Tsample indicating the interval of generation of the timer interrupt signal IRQ0 calculated in step 95. In the following step 101, the value TR of the register of the timer 89 is set. Is calculated as TR = α + j × TPulse, and similarly to step 96,
Set TimerFlag to False = 0. As a result, Tsample, which is a value obtained by adding 1 to the constant j,
The control input is updated at the time interval of.

The above processing will be described first with reference to FIG. 6A when the constant j is 0.

In the case of j = 0, if the pulse interrupt IRQ1 occurs at the time Tp (k), the variable α
Is set as 1/2 of the value of Tpulse0, and j is 0
Therefore, a timer interrupt IRQ0 occurs at the time of Ts (K + 1) after α from Tp (k), and the control input 0UT is output to the D / A converter 88.

At this time, if TPulse is constant,
Tsample is a sample time that matches TPulse.

The control input 0UT output at the time of Ts (k + 1) is calculated as the rotation speed Rev of the motor 80 at the time interval of Tpulse0, and this value is Ts (k).
Therefore, the maximum approximation accuracy can be obtained as a discrete-time system by using a control law designed to be the number of revolutions at the time of Ts (k) since the average value is centered on the time.

An example in which the constant j is set to 1 is shown in FIG.
(B) of FIG.

In this case, in the pulse interrupt IRQ1 of Tp (k + 1), the processing after step 94 is not performed, and
When the constant j is increased, Tsample0, which is the sampling interval of the discrete time system, is TPulse
Since 0 becomes shorter, the time taken for averaging becomes shorter, and the time resolution improves.

Therefore, from the time Tp (n) at which the pulse signal generated according to the rotation of the motor 80 is generated,
When updating to a new control input OUT after a lapse of a time obtained by adding a predetermined value α to a time interval obtained by multiplying the time interval TPulse of the pulse signal by a predetermined integer j including 0, the control input OUT is controlled by the previous control. The pulse signal generated at the time immediately before the time at which the input was updated, the time interval of the pulse signal generated immediately thereafter, and the time obtained by adding the sample time to the time interval of the pulse signal are calculated at the same time. Since the predetermined value α is set to の of the time interval of the pulse signal, it is possible to more appropriately solve the second problem of the above-described conventional example that it is unclear which time the value is when viewed as a discrete time system. In addition, the third problem shown in the conventional example, that is, the problem that the rotation speed measurement pulse does not come within the sample time of the discrete time system, is solved, A control input can be accurately obtained even during rotation, and a maximum margin for acceleration / deceleration change can be obtained.

Also, by introducing the above α, it becomes possible for the control means to secure the time required for the calculation, and it is possible to set the integer j to 0, and to allow the deterioration of the time axis resolution. In this case, the sampling time can be reduced as compared with the first embodiment.

In the present embodiment, the variable α is set to TPul
Although it is set to a value of ｓｅ, if the acceleration / deceleration characteristics of the rotating machine are not uniform, such as a rotating machine such as a transmission of an automobile, an arbitrary constant m is used instead of １ ／, and α By setting = TPulse × 1 / m, optimal rotation speed control can be performed according to the acceleration / deceleration characteristics.

When the acceleration / deceleration characteristics of a rotating machine are not uniform, such as in a transmission of an automobile, the constant m may be changed in accordance with the direction of acceleration / deceleration.
Assuming that m = m1 on the acceleration side and m = m2 on the deceleration side, the relationship m1> m2 is set.

Therefore, on the acceleration side, 1 / m decreases on the acceleration side, while 1 / m increases on the deceleration side, and the time interval TPpulse of the pulse interrupt IRQ1 gradually decreases. By setting the time interval from time Tp (k) to Ts (k + 1) short,
The time from time Ts (k + 1) to Tp (k + 1) can have a margin, and the next pulse arrives earlier than time Ts (k + 1). That is, the time Tp (k + 2)
Can be prevented from becoming earlier than Ts (k + 1), and on the deceleration side, time Tp (k + 1) is set to Ts (k + 1).
+1) can be prevented, and accurate speed control can be performed in the entire speed range from high-speed rotation to low-speed rotation of the motor 80.

Further, the constant 1 / m is set small when accelerating, while the constant 1 / m is set large when decelerating. If applied, even in a transient state where the sampling time interval changes every moment, such as during acceleration / deceleration, the time required for the calculation of the control means can be reliably secured,
Control accuracy can be ensured.

FIG. 7 shows a third embodiment, in which a step 102 for performing a conditional branch by TPulse between steps 95 and 100 shown in FIG. 5 of the second embodiment.
A is added, and step 102B for performing processing when the condition of step 102A is satisfied is added.
Other configurations are the same as those of the second embodiment.

At step 102A, if the variable TPpuls indicating the timer interrupt signal IRQ0 generation interval is smaller than a predetermined value, the routine proceeds to step 102B, where Tsa
mple is set to a constant value Tconst to prevent the arithmetic processing in the CPU 83 from being unable to keep up.

In the second embodiment, TPulse
Is extremely short, a timer interrupt IRQ0 may occur earlier than the control input 0UT is calculated.However, when TPpulse is smaller than a predetermined value, Tsample is set to a constant value Tconst. Thus, the operation time of the CPU 83 can be ensured.

In the third embodiment, TPulse
When T is small, Tsample is set to a constant value Tconst
In this case, Tsam is used for TPulse.
ple is sufficiently long, the first to third problems associated with the pulse measurement do not occur, and if the response speed of the motor 80 is restricted, it is meaningless to shorten the sampling time more than necessary. Therefore, the control performance does not deteriorate.

Also, the time interval TPulse of the pulse signal
Is smaller than a predetermined value, the sampling time interval is changed to a preset value Tconst. In the case of high-speed rotation where the period of the measurement pulse time interval TPlus is short, control is performed with a predetermined fixed sample time Tconst. By securing the calculation time at the time of high-speed rotation, accurate control can be performed in the entire speed range.

FIG. 8 shows the fourth embodiment. Step 100 of the second embodiment shown in FIG.
The other configuration is the same as that of the second embodiment.

In step 103, the pulse interrupt signal I
A variable TPpuls indicating the RQ1 occurrence interval is calculated in the same manner as in step 95, and the variable j and the variable Tsam are calculated.
The calculation of the ple and the variable α are performed.

J = int (γ / TPulse + β) Tsample = (j + 1) × TPulse α = TPulse / 2 where γ and β are preset values, and int is a function for rounding the operation result to an integer.

In step 103, the variable j is calculated by dividing the predetermined value γ by the value obtained by adding the predetermined value β to TPpulse. Therefore, as Tpulse becomes smaller, j becomes smaller.
Therefore, as described above, even when TPpulse is small, the calculation processing time of the CPU 83 can be reliably ensured.

FIGS. 9 and 10 show a fifth embodiment, and FIG. 9 is a diagram in which step 91 of the timer interrupt processing shown in FIG. 2 of the first embodiment is replaced with steps 104 to 106. Reference numeral 10 denotes a pulse interrupt process executed in place of FIG. 3 of the first embodiment, and the same components as those of the first embodiment are denoted by the same reference numerals, and redundant description is omitted.

FIGS. 9 and 10 solve the problem that the pulse interrupt signal IRQ1 frequently occurs particularly at high speed rotation, and the processing time of the CPU 83 is shortened.

In FIG. 9, the timer interrupt IRQ0
Has occurred, if it is determined in step 104 that TPuIse is equal to or greater than a predetermined value, the flow advances to step 105 to set the variable Count to 1 and set the variable Flag to False.
Set to se.

On the other hand, if TPpulse is equal to or less than the predetermined value, the variable Count is set to 0, the variable Flag is set to True, and pulse interruption is permitted. Step 92 is the same as in FIG.

In FIG. 10, the pulse interrupt is enabled,
If it occurs, the variable Now is set in step 107.
_c is assigned to Last_c, the value of Now_c is set to the value CR of the capture register 86, and the variable C
Increment out.

At step 108, if the value of the variable Count is 2, the routine proceeds to step 95, where the same processing as in the first embodiment is performed, and the routine proceeds to step 109.

At step 109, if the value of the variable Flag is False, the routine proceeds to step 111, at which the value of the register TR of the timer 89 is set to TR = Tsample- (MaxTimer-TR), as in step 96 of the first embodiment. ).

On the other hand, if the value of Flag is True,
Proceeding to step 110, the pulse interrupt is prohibited.

The operations in steps 97 and 98 are the same as those in the first embodiment.

In the above processing, when TPpulse is larger than the predetermined value, the same processing as in the first embodiment is performed, that is, the variable TimerFlag in FIGS. 2 and 3 of the first embodiment becomes the variable Count, and Only the conditions are changed, and the operation is the same.

On the other hand, when Tpulse is smaller than the predetermined value, the operation is as shown in FIG.

That is, at time Ts (k), step 9
2, after the control input OUT is output to the D / A converter 88, the pulse interrupt is permitted.
ulse0 is the difference between Tp (k) and Tp (k + 1).

A pulse after two pulses of Tp (k) and Tp (k + 1) are inputted, that is, Tp (k)
+2) and Tp (k + 3) are not processed because interrupts are prohibited.

Then, from time Ts (k) to Tconst
After a lapse of time, it becomes Ts (K + 1), and again at step 10
4 are performed, and Tp (k + 4), Tp (k + 5)
Are measured.

As a result, the pulse interval becomes extremely short at the time of high-speed rotation, and the calculation processing can be reduced even in a situation where many pulses are generated within the time of Tconst.

That is, steps 107 and 10
In addition to not performing the processing of step 8, the processing of saving the register of the CPU 83 associated with the pulse interrupt processing is eliminated, so that the calculation load of the microcomputer is remarkably reduced as compared with the conventional example.

Further, in this embodiment, the pulse measurement at the time of high-speed rotation is not located at the center of the time step of the discrete time system, that is, Ts (k) is intermediate between times Tp (k) and Tp (K + 1). However, during high-speed rotation, the effect is small because the time interval between pulses is short.

Further, when the CPU 83 is in a multitasking environment, in order to avoid deadlock (uncontrollable state due to competition of a plurality of processes) accompanying the interrupt process, it is sufficient to perform the process of permitting the interrupt in step 105 as well. In addition, the calculation load of the CPU 83 accompanying the interrupt processing can be further reduced, and the control at the time of high-speed rotation of the motor 80 can be performed with high accuracy.

Thus, the pulse signal time interval TPul
When se is less than a predetermined value, the rotation speed is measured by the pulse signal generated after the control input update, and after the rotation speed measurement, the rotation speed measurement is prohibited until the next control input update is performed. In a situation where a large number of measurement pulses are generated during high-speed rotation, it is necessary to prevent the pulse measurement processing from increasing the calculation load on the control means, and to perform various calculations on average to ensure control accuracy during high-speed rotation. Can be.

Further, the time interval TPulse of the pulse signal
When a time interval obtained by multiplying by a predetermined integer j is calculated, the integer j is changed according to the time interval TPulse of the pulse signal. Therefore, the sample time interval changes according to the time interval of the measurement pulse. In such a case, it is possible to ensure the calculation time of the control means.

FIGS. 12 to 15 show a sixth embodiment, in which the present invention is applied to a continuously variable transmission for automobiles.

In FIG. 12, a continuously variable transmission as a rotating machine includes a primary pulley 141, a secondary pulley 142, a drive belt 144, and a hydraulic actuator 14.
5, the driving force of the engine 140 is transmitted from the primary pulley 141 to the driving belt 14
4, and is transmitted to the secondary pulley 142 to drive the output shaft 143.

The hydraulic actuator 145 continuously changes the groove widths of the primary pulley 141 and the secondary pulley 142 so that the primary pulley 1
41 and the secondary pulley 142 are driven by the drive belt 144.
, Ie, the gear ratio, can be continuously changed.

The hydraulic actuator 145 has hydraulic control means (not shown) including an actuator responsive to a control input OUT from the controller 146, and the like.
The hydraulic actuator 145 is connected to the primary pulley 14
By displacing the first and secondary pulleys 142 in the axial direction, the drive belt 144 is moved along the inclined surfaces of both pulleys, and the gear ratio is set to an arbitrary speed ratio by changing the contact radius ratio.

In order to control the gear ratio more accurately, the primary pulley 141 has a pulse encoder 8
1a, a pulse encoder 81b is disposed on the secondary pulley 142, and these pulse encoders 81a, 81
The output pulse 1b is input to the capture registers 86a and 86b of the controller 146, and controls the speed ratio based on the ratio of the rotational speeds of both pulleys.

The controller 146 mainly composed of a microcomputer replaces the capture register 86 of the controller 82 of the first embodiment shown in FIG. 1 with two capture registers 86a and 86b corresponding to the pulse encoders 81a and 81b. The configuration is the same as that of the first embodiment except that the time interval between two types of pulses can be measured.

Next, FIGS. 13 to 15 are flowcharts showing an example of control performed by the controller 146.
FIG. 13 shows an interrupt process based on an interrupt signal IRQ from the timer 89, as in the above embodiment.

FIG. 14 shows an interrupt process in which a pulse from the pulse encoder 81b connected to the output shaft 143 is set as the interrupt signal IRQ1, and FIG. 15 shows a pulse from the pulse encoder 81a connected to the engine 140 as the interrupt signal. An interrupt process for IRQ2 will be described.

First, in FIG. 13, when a timer interrupt IRQ0 occurs, in step 151, Ti
With merFlag set to True, interrupt processing shown in FIGS.
lag_a and PulseFlag_b are set to False, and the value TR of the register of the timer 89 is set to a constant MaxTi
Set mer.

In step 92, as in the first embodiment, the control input OUT is output to an actuator (not shown) of the hydraulic actuator 145.

The flowchart of FIG. 14 is executed when an interrupt signal IRQ1 is generated from the pulse encoder 81a on the primary pulley 141 side, and is the same as steps 93 to 94 of FIG. 2 of the first embodiment. And at Step 94, TimerFla
When the value of g is True, the processing of step 151a is performed.

In the step 151a, the variable TPulse
(Pulse interrupt signal IRQ1 generation interval) TPulse
_a is calculated by the following equation.

TPulse_a = Now_c−Last_c Further, PulseFlag_a is set to True.

Next, in step 152a, PulseFla set in the flowchart of FIG.
It is determined whether g_b is True, and if it is True, the process proceeds to step 153.

In this step 152a, PulseF
By checking lag_b, timer interrupt IRQ0
Occurs, interrupts IRQ1 and IRQ2 occur, and pulse encoder 81a and pulse encoder 81b
After confirming that the pulse signals from
The processing after step 153 is performed.

In step 153, the pulse encoder 8
Time interval TP between two types of pulse signals from 1a and 81b
ulse_a and TPulse_b are compared, and step 1
In step 54, the value of j times the time interval of the pulse signal having the longer time interval is set as Tsample.

Next, at step 96, the sample time of the discrete time system is set as the sample time of step 154 or 15 described above.
Based on Tsample having a long time interval set in 5, the value TR of the timer register is reset as follows: TR = Tsample- (MaxTimer-TR) The timer interrupt signal is output after the elapse of Tsample from the previous timer interrupt. At the same time as setting so that IRQ0 is generated, TimerFlag is set to False, that is, 0, and the processing after step 94 is prohibited until the next timer interrupt occurs.

In step 156, the ratio of the time interval of the pulse signal is calculated as the speed ratio Rtio of the continuously variable transmission as follows.

Rtio = TPulse_a / TPulse_b Next, a speed ratio deviation Err is calculated from the difference between the speed ratio Rtio and the target speed ratio TRtio as follows.

Err = TRtio-Rtio The target gear ratio TRtio is a value set according to the driving state of the vehicle.

Next, at step 98, as in the first embodiment, the integrated value IErr of the error Err is calculated, and then 0UT is calculated by the control input.

On the other hand, the flowchart of FIG. 15 executed when the pulse interrupt IRQ2 occurs also corresponds to FIG.
4 is the same as that of the flowchart of FIG.
In 51b, PulseFlag_b is set to True, and in step 152b, PulseFlag_a is checked, and the rest is configured in the same manner as the float in FIG.

By executing the processing of FIGS. 13 to 15 in accordance with each of the interrupt signals IRQ1 and IRQ2, the time signals of the pulse signals IRQ1 and IRQ2 from the two rotating shafts (the primary pulley 141 and the secondary pulley 142) are output. The generation interval Tsamp of the timer interrupt signal IRQ0 for updating the control input OUT from the pulse signal having the longer interval
le and set Tsample to the pulse interval TP
Since ulse_a or TPulse_b is set by multiplying it by a constant j of 1 or more, even when controlling the speed ratio of two rotating shafts, it is possible to accurately control the number of rotations from high speed rotation to extremely low speed rotation in the same manner as in the above embodiment. , And control accuracy of the speed ratio of the continuously variable transmission can be improved.

In this way, as shown in the conventional example, in order to synchronize the sample time of the discrete time system with the time interval of the pulse measured in the rotation measurement, the sample time of the discrete time system is changed every moment. , While preventing the time resolution of rotation measurement from deteriorating depending on the rotation speed for a discrete time system,
The timing of the rotation measurement for the time step of the discrete time system is undefined, and the rotation measurement pulse does not arrive within the sample time of the discrete time system at extremely low speed rotation, and the rotation measurement is not performed correctly. Can be avoided, and the speed control of the rotating machine can always be accurately performed from high-speed rotation to extremely low-speed rotation. It is possible to dramatically improve.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a motor and a controller according to an embodiment of the present invention.

FIG. 2 is a flowchart of a timer interrupt process, illustrating an example of control performed by the controller.

FIG. 3 is a flowchart of a pulse interruption process.

FIG. 4 is a timing chart showing an operation, showing a relationship between a pulse generation interval and a control input;

FIG. 5 is a flowchart of a pulse interruption process according to the second embodiment.

FIG. 6 is a timing chart showing the operation, showing the relationship between the pulse generation interval and the control input.

FIG. 7 is a flowchart illustrating a pulse interruption process according to the third embodiment.

FIG. 8 is a flowchart of a pulse interruption process according to the fourth embodiment.

FIG. 9 is a flowchart of a timer interrupt process according to the fifth embodiment.

FIG. 10 is a flowchart of a pulse interruption process.

FIG. 11 is a timing chart showing the operation,
The relationship between the pulse generation interval and the control input is shown.

FIG. 12 shows a sixth embodiment, and is a block diagram of a continuously variable transmission and a controller.

FIG. 13 is a flowchart of a timer interrupt process.

FIG. 14 is a flowchart of a pulse interruption process on the primary pulley side.

FIG. 15 is a flowchart of a pulse interruption process on the secondary pulley side.

FIG. 16 is a block diagram of a motor and a controller, showing a conventional example.

FIG. 17 is a block diagram of a motor and a controller, showing another conventional example.

FIG. 18 is a graph showing control characteristics of a conventional example, showing the relationship between time and control input.

FIG. 19 is a block diagram of a motor and a controller, showing still another conventional example.

FIG. 20 is a graph showing control characteristics of a conventional example, showing the relationship between time and control input.

FIG. 21 is an explanatory diagram showing control timings of a conventional example, showing the relationship between pulse generation intervals and control inputs.

FIG. 22 is a diagram corresponding to claims corresponding to the first to eighth inventions.

[Explanation of symbols]

 Reference Signs List 80 motor 81 pulse encoder 82 microcomputer 83 CPU 84 RAM 85 ROM 86 capture register 87 free run counter 89 timer 140 input shaft 141 primary pulley 142 secondary pulley 143 output shaft 144 drive belt 145 hydraulic actuator 200 rotating machine 201 pulse generating means 202 rotation Speed measuring means 300 Control means 301 Sample time setting means 302 Control input calculating means 303 Control input updating means 304 Pulse interval comparing means 305 Sample time changing means 306 Longest pulse interval selecting means

Claims (8)

[Claims] 1. A pulse generating means for generating a pulse signal corresponding to a rotational position of a rotary machine, and a time interval TPu between pulse signals from the pulse generating means.
a rotation speed measurement unit that measures a rotation speed based on the rotation speed; and a control unit that controls a control input to the rotation machine based on the rotation speed by a discrete time system. While the speed measurement means measures the rotation speed from the pulse signal generated immediately before the time when the control input is updated and the time interval TPulse of the pulse signal generated immediately after, the control means calculates the rotation speed of the pulse signal. Sample time setting means for setting a time interval obtained by multiplying the time interval TPulse by a predetermined integer j equal to or greater than 1 as a sample time; and updating a control input according to the rotation speed measured by the rotation speed measurement means and the sample time. Control input calculating means, control input updating means for updating the control input to the rotating machine at the time interval of the sample time, Speed control device for a rotating machine, characterized in that it includes. 2. The method according to claim 1, wherein the control input updating means includes a time interval TP of the pulse signal from a time at which the pulse signal is issued.
2. The control input is updated after a lapse of time obtained by adding a value α obtained by multiplying a time interval TPulse of the pulse signal by １ ／ to a time interval obtained by multiplying ulse by a predetermined integer j. 3. Speed control device for rotating machinery. 3. The method according to claim 1, wherein the control input updating means includes a time interval TP of the pulse signal from a time when the pulse signal is issued.
2. The control input is updated after a lapse of a time obtained by adding a value obtained by multiplying a time interval TPulse of the pulse signal by a predetermined constant 1 / m to a time interval obtained by multiplying the pulse number by a predetermined integer j. A speed control device for a rotary machine according to claim 1. 4. The sampling time setting means sets the constant 1 / m small when the rotating machine is accelerating, and sets the constant 1 / m when the rotating machine is decelerated.
The speed control device for a rotary machine according to claim 3, wherein is set to be large. 5. A pulse generating means for generating a pulse signal corresponding to a rotational position of a rotary machine, and a time interval TPu between pulse signals from the pulse generating means.
a rotation speed measurement unit that measures a rotation speed based on the rotation speed; and a control unit that controls a control input to the rotation machine based on the rotation speed by a discrete time system. While the speed measurement means measures the rotation speed from the pulse signal generated immediately before the time when the control input is updated and the time interval TPulse of the pulse signal generated immediately after, the control means calculates the rotation speed of the pulse signal. Sample time setting means for setting a time interval obtained by multiplying the time interval TPulse by a predetermined integer j equal to or greater than 1 as a sample time; pulse interval comparing means for comparing the time interval TPulse of the pulse signal with a predetermined value; If the comparison result of the interval comparison means is that the time interval TPpulse of the pulse signal is less than a predetermined value, While changing the preset value, the time interval T
When Pulse is equal to or more than a predetermined value, a sample time changing unit that sets a value set by the sample time setting unit, and a rotation speed measured by the rotation speed measuring unit and a sample time from the sample time changing unit. Control input calculating means for updating the control input; and control input updating means for updating the control input to the rotating machine at a time interval of the sample time set by the sample time changing means. Speed control device for rotating machinery. 6. If the time interval TPpulse of the pulse signal is less than a predetermined value in the comparison result of the pulse interval comparing means, the rotation speed measuring means determines the rotation speed by the pulse signal generated after updating the control input. 6. The speed control device for a rotary machine according to claim 5, further comprising a control prohibition unit that measures the rotation speed and prohibits the measurement of the rotation speed until the next control input update is performed after the measurement of the rotation speed. 7. A rotary machine having a plurality of rotating shafts, a plurality of pulse generating means for respectively generating pulse signals corresponding to the rotational positions of the rotating shafts, and a pulse signal from the plurality of pulse generating means. In a speed control device for a rotary machine, comprising: a rotation speed measurement unit that measures a rotation speed based on a time interval TPulse; and a control unit that controls a control input to the rotary machine based on the rotation speed by a discrete time system. The rotation speed measurement unit measures the rotation speed from a pulse signal generated immediately before the time at which the control input is updated and a time interval TPulse of the pulse signal generated immediately afterward. Time interval TPulse of each pulse signal
A longest pulse interval selecting means for selecting the largest one among the above, and a time interval obtained by multiplying a time interval TPulse of the pulse signal selected by the longest pulse interval selecting means by a predetermined integer j of 1 or more as a sample time. Sample time setting means, control input calculating means for updating a control input according to the rotation speed measured by the rotation speed measuring means and the sample time, and updating the control input to the rotating machine at a time interval of the sample time. A speed control device for a rotary machine, comprising: a control input updating unit configured to perform the control input by the control unit. 8. The sample time setting means changes the integer j according to the pulse signal time interval TPulse when calculating a sample time interval obtained by multiplying the pulse signal time interval TPulse by a predetermined integer j. The rotation speed control device for a rotary machine according to any one of claims 1, 5 to 7, further comprising a constant changing unit that changes the rotation speed.