JPH01166112A - Position control device - Google Patents

Abstract

PURPOSE:To make earlier the judgement time of whether an on-off valve should be driven by operating an expected position where hydraulic actuator stops and comparing this expected position with a target position. CONSTITUTION:An actual position (y) of a hydraulic actuator 1 is detected by an actual position detecting means 3 and by the comparison of the detected actual position (y) with a target position R, an action mode U of whether an on-off valve 2 is on-operated or off-operated is selected by an action mode selecting means 4. When a control mode is continued, an expected position operating means 6 operates an expected stop position y' of the hydraulic actuator 1 based on the actual position y(t0) at hat time and a parameter about the dynamic characteristic of a control system. Whether the operated expected stop position y' coincides with the target position R is decided by a deciding means 7, when they coincide, an action mode switching means 8 switches the action mode so as to off-operate the on-off valve 2.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a position control device, and particularly to a position control device that performs feedback control of the spool position of a pilot valve or the piston position of a hydraulic cylinder.

(Conventional technology) Recent advances in mechatronics have been remarkable, with servo valves used to control the position of hydraulic equipment and PWM on/off valves.
(Pulse Width Modulation) method has been proposed (for example, see pages 1135 to 1139 of "Hydraulic Technology Handbook" published by Nikkan Kogyo Shimbun, January 1977).

(Problems to be Solved by the Invention) However, such a device has a drawback in that it is difficult to obtain satisfactory results in terms of both responsiveness, accuracy, and cost.

For example, those using servo valves can provide highly accurate position control, but are expensive due to their complicated structure and are susceptible to oil contamination. On the other hand, a simple on/off valve has a simple structure, is inexpensive, and is resistant to oil contamination, but it is difficult to improve control accuracy because it depends on the characteristics of the on/off valve itself and the hydraulic system.

In order to satisfy both aspects, the first step is to use an on-off valve from the viewpoint of cost, and the second step is to develop a control method that will provide high-speed response or high accuracy.

This invention was made by focusing on the problems of the conventional example. By calculating the predicted position where the hydraulic actuator will stop and comparing this predicted position with the target position, it is possible to control the on-off valve. It is an object of the present invention to provide a device that can accelerate the timing of determining whether or not to drive.

(Means for Solving the Problems) As shown in FIG. 1, this invention is provided in an oil passage to a hydraulic pressure unit 1 (e.g., a pilot valve spool or a piston of a hydraulic cylinder), and is configured to operate in response to a drive signal. The actual position of the on-off valve 2 that opens and closes and the hydraulic actuator 1 (
y), a means (for example, a potentiometer) 3 for detecting the detected actual position (y) and a predetermined target position (R).
means 4 for selecting an operation mode (U) for turning on or turning off the on-off valve 2 by comparison with the control mode; and means 5 for outputting the drive signal when the control mode is selected.
In the position control device, when the control mode is continued, the expected stop position (wa) of the hydraulic actuator 1 is determined based on the actual position (y(to)) at that time and parameters related to the dynamic characteristics of the control system. means 6 for calculating the predicted stop position (9), and the calculated predicted stop position (9) is set to the target position (
means 7 for determining whether or not R) match, and if they match, turning off the on/off valve 2; Means 8 is provided for switching the operation mode so as to perform the operation.

(Function) The expected stop position 9 is uniquely determined depending on the characteristics of the on-off valve 2 and the hydraulic system, and if 9 matches R, the stop position of the actuator 1 can be determined without actually checking.
It appears that y converges towards R.

In such a case, in the present application, the operation mode is switched so that the on-off valve 2 is turned off.
The actuator 1 will not be driven unnecessarily even if it appears to have gone too far.

This means that before the actuator 1 actually settles, it is determined whether the on-off valve 2 is to be turned on or turned off, and good responsiveness is provided even though this is feedback control.

(Embodiment) FIG. 2 is a block diagram of a control system according to an embodiment of the present invention, which constitutes a known feedback control system. The object to be directly controlled is an on-off type pilot valve 22, and this valve 22 is controlled by a hydraulic power source (for example, 301 cgf/cm"
) The hydraulic oil from 21 enters the orifice (for example, φ0.8
mm) 23 and 24, and a normally closed on/off valve 26 and 27.
It is opened and closed by the drive signal from.

28 is the actual value of piston (spool) displacement (y(t))
The signal from this potentiometer 28 is the target value of spool displacement (R(t
)), and its deviation d(t) (=R-y) is input to the controller (microcomputer) 31 via the signal amplifier 29. Note that 30 is a sample/hold circuit.

A microcomputer (for example, a single-chip one) 31 determines the switching timing of the on-off valves 26 and 27 in consideration of the characteristics of the pilot valve 22 and the hydraulic system. Here, the optical control is on-off type positioning control. The content of this control will be explained in detail below, but first an overview based on control theory will be given, and then specific control will be given.

(i) Basic operation of pilot valve There are three types of operation modes of the pilot valve 22 as follows. However, U is an input signal for mode selection.

■When U=-1, valve 26:ON Valve 27:0FF (The spool moves to the left in the diagram) ■When U=0, valve 26:OFF Valve 27:0FF (The spool stops) ■When U=1, the valve 26: OFF Valve 27: ON (The spool moves to the right in the diagram) In addition, in the following, turning on either valve 26 or 27 with U=±1 is "off operation", and turning on either valve 26 or 27 with U=0 is "off operation". 27
Turning off both is called an "off operation."

(ii) The spool displacement (y) when observing the operating mode selection conditions is the target value of the spool displacement (
The judgment is made based on whether or not it falls within a tolerance (±ε) set around R). However, the direction in which the spool moves to the right in the figure is defined as the positive direction of y.

■If R-ε≦y≦R+ε, U=O■R+y
In the case of U=-1■R-ε〉y
U=1 Now, since this kind of feedback control reflects the result of an operation in the next operation, it cannot be said that the response during transients is good, and low-cost microcomputers cannot match the calculation speed. Since there is a strike limit, an important issue is how to solve the response delay.

Therefore, in this application, we newly introduce a predicted value of the position where the spool will stop ("i (t o 1 N + 2)), and even if it is determined that U = ±1 in the conventional method, 9 will be ( R±ε), there is no need to turn off the on-off valve, and U is changed from 1 or -1 to 0. In other words, U=1 and R-ε< 9 (to +N+2) In the case of U=1→0, (2) If U=−1 and R+ε>9 (to +N+2), U=−1→0.

However, in other cases such as (i) U=1 and R-ε≧9 (to+N+2) and (i)U=-1 and R+ε≦V (to +N+2), it goes without saying that the selected U is continued.

FIG. 3 shows a summary of mode selection conditions (conditions ``■'' to ``■'' above). In addition, unless necessary, “i (
to +N+2) is also written as 9, which is simplified.

(iii) The predictive control system for spool displacement is considered as a model (discrete value model) in which pressure change and spool displacement are connected in series, as shown in the figure below. Note that the latter spool displacement model is an example of a zero-order holder tenth-order delay.

The parameters (a and bi) shown in the above figure are based on the actual response results, that is, the time when the on-off valve is turned off is to
The following equation (1
) and (2).

a=y(to 10N+2)/y(to +N+1)
−(1) bi=iy(to +i+1) y(t
o +i+2)1-aiyDo +i) y(t
o +i+1)) - (2) However, if the time during which the operation mode is 1 is set as T, then the condition for formula (2) to be satisfied is that T>NXTo, and if this condition is satisfied, formula (2 ) shall be calculated. N is the value (order of the model) obtained by dividing the time until the hydraulic system settles after turning off the on-off valve by the sampling time To (sea), and the settling time is (N+1)To. In addition, in order to obtain a series of discrete value data, the settling time ((N+1)T
The ON operation shall not be performed during step o). Fourth
The figure shows an example of the response of y. From the figure, U(to+i) is 0 when i≧1 and 1 when i≦0.

From equations (1) and (2), a and bi are functions of y, that is, y
This value is determined based on the response (dynamic characteristics of the control system). Conversely, if we use a and bi, we should be able to reproduce the response of y, and in this case, the final value of y(
to +N+2) is thus understood as the predicted value (9).

Therefore, using the initial value pato), a and bi, 9
is calculated using the following formula. However, in this case, T is To
Select an integer multiple of , with the condition that T<NXTo 6 ■If U=1, V (to +N+2)=aN + 2bato)...
・(3) ■When U=-1 9 (to+N+2)=a""2y(to)(i
v) Compensation for constant disturbance A constant disturbance is understood as a steady deviation. For example, consider the force (Fo) due to a neutral return spring as a constant disturbance.

Letting V F (to +N+2) be the predicted value of the spool displacement that takes into account the influence of this constant disturbance, this is V F (to + N+2) = v (L (10 N + 2) +
KpFoΣaN + I-i, , (5) 1→. However, K is a constant.

On the other hand, the estimated disturbance value (ttS2 term) is the difference between Wa and y after the settling time, so KF F, ΣaN + 1-+ "y(to +N+2) -v(to 1N+2)
- Calculated in (6).

Here, if the target value of spool displacement in this case is set as R9, then R・”RKpFo Σ, + l−i
, , (7) 1 Hatake 0 , and R without considering the steady deviation is R of the open equation (7)
By changing to o, the steady-state deviation will be eliminated.

The figure below shows a discrete value model of the control system when a constant external hole is considered.

(v) Parameter (a
) is assumed to be 1.0. At this time, the other parameter (bi) is the amount (C) determined by the on-time (T) of the on-off valve.
)/(ET E-ET U(UT U)-1UT E)
1...(8) is given by the following equation (9).

b= C(Y (to + 1) Y (to))
...(9) However, equations (8) and (9) are matrix equations, and the values of each matrix are as follows.

...(10a) ...(iob) ...(10c) As a result, V F (to +N+2) is expressed by the above formula (3) ~
It will be calculated from (6), but the value of Σbi is required at this time. If we put here, equation (9) becomes X(Y(to +1) Y(to ))
−(12).

Here, C
If we calculate ゛ in advance (assuming N=9), ■T=1
In the case of ■T=2, similarly, ■The case of T=N+1.

As a result, Σbi is obtained from equation (12), and equation (3), (
4) to 9(to 1N+2), but if you change the parameters every time you estimate the parameters, the parameter values will change suddenly, which may affect the estimation of the disturbance, etc. Therefore, it is preferable (1°) Also, since the actual parameter fluctuation is considered to be sufficiently slow relative to the on/off cycle of the on/off valve,
The estimated amount of movement of y due to one off (Δyo F F (
T)) as...(14a) (where a=1), ay('''') OF F (T) = (λΔ31('
)OF F (T)1Δyo F F (T)l/ (
λ+1)...(15) Cyclically Δyl ” 'L F F (T
) and substitute this value into equations (3) and (4) to obtain 9(to 1 N+2).

In Equation (15), a person uses a weighting coefficient (for example, 15
), and by using this person, sudden changes in parameters during parameter estimation can be prevented.

In addition, the result obtained this time (Δyo F F (T)) and the value stored in the table this time (ay(')o F
F (T)) is used to calculate the next movement predicted value (ay('
``')o F F (T)) will be found sequentially, so the initial value may be zero, and therefore there is no need to initialize the parameters.

(vi) Prevention of accuracy loss due to quantization When the full stroke of spool displacement is quantized, A/
In relation to the resolution of the D converter, detection accuracy may deteriorate. For example, for a full stroke of 16111111, for 8 bits, 16/28 = 0°0625mm
is the minimum unit of quantization, and this has poor accuracy.

Therefore, the deviation d(t) between the actual value and the target value is d(t)=
y(t) −R(t) ...(1
6).

Now, when the discrete value model considering disturbance is expressed by a difference equation, the following equation (17) is obtained.

y(t+1)−ay(t)×ΣbiU(t−i)
Ten Fo F Ki-.

...(17) Here, y(t) = R(t) obtained from equation (16)
+ dD) into equation (17), d(t+1) = ad(t) + ΣbiU (t - i) 10 Fo F Ki4 + aR(t) -R(t+ 1) ・=
(18).

In equation (18), assuming that R is a constant value regardless of time (1), the fourth and fifth terms (aR(t)-R
(t+1)) is a constant value. This means that constancy6
This means that it can be treated the same as the term due to L (the third term, where Fo F has a constant value). In particular, when a=1.0, aR(t) R(t+ 1)=0, so d(t)
There is no effect when using .

Therefore, when estimating the parameters (settling process mode), the target value R is held and equation (18) is made equivalent to equation (17). As a result, positioning control when d(t) is used instead of y(t) is the same as control to set d(t) to zero. In other words, it is not necessary to quantize y(t), but by quantizing d(t), the minimum unit of quantization can be made smaller and the detection accuracy can be increased, thereby reducing the amount of energy required.

Note that if a 12-bit A/D converter can be used, the detection accuracy will be sufficient, so both the y(t) and R(t) signals may be directly processed. However, in this case, A/D converters for two channels are required. Current single-chip microcomputers have 12 pins) A
Since there is no built-in /D converter type, it has a multi-chip configuration.

Having provided a theoretical overview above, next we will show how these functions are configured on the software provided to the microcomputer 31 for an example of the deviation control method, and use this figure to explain how this example is constructed. Explain the action.

The figure opens at a certain time, for example, with a sampling period of 15ee.
), the deviation d(=y−
R) is read, and it is determined whether the operating mode is the settling process mode or the control mode based on d and according to the conditions shown in FIG. 3 (steps 42 and 43).

Here, the "settling process mode" refers to a case where the value of U changes from 1 or -1 to O, whereas the "1 control mode" refers to a case other than the settling process mode. These 2
Since the control operations differ depending on the two modes, the cases will be explained below.

(I) Control mode (A) Check the previous control state (operation mode) and U=
If it is O, it is determined from the value of d whether the current U is ±1 or 0, and U=±1 or 0 is outputted corresponding to the determination result (steps 44 to 46).

(b) If the previous operation mode is U=±1, tq is determined using 9 to determine whether it is appropriate to turn off 'PJ. In other words, the predicted stop value 9 of the spool displacement when the on-off valve is turned off at the current time (to) is calculated by substituting y(y(to)) and data of a and bi into equation (3) or (4). (step 47), and this value is used to determine whether there is a need to change the operating mode.

(a) Whether U=1 and R-9 or U=-1 and R+
When ε>9 In these two cases, the spool displacement is considered to converge toward R without actually checking the stop position of the spool displacement, so there is no need to continuously turn on the on-off valve. The value of U is changed from 1 or -1 to 0 (see Figure 3).

Therefore, in this case, the ON operation is switched to the OFF operation (step 48.53).

This means that the position of the spool displacement is controlled before the spool displacement actually settles.
Although it is feedback control, it has good responsiveness and is suitable for high-speed control. Also, in this case, if the on-off valve continues to be turned on, the spool may go too far and cause a deviation, but this has been avoided.

In other words, focusing on the fact that the position at which the spool displacement will stop can be uniquely predicted depending on the characteristics of the on-off valve and the hydraulic system, the predicted value can be used to determine whether the on-off valve should be turned on or turned off. The decision was made to make up for the response delay of feedback control by advancing the timing of determining whether to do so.

Note that when U=0 is selected, constant disturbance estimation and parameter estimation preparations are made in the subsequent settling process mode, but in that case, the necessary data (9+d(to)
and T) are stored in the memory, and the target value R is held and aged (steps 50 to 52).

(b) U = 1R-ε≧9 or U=-1R
When 1ε≦9 In these cases, U=±1 is maintained so that the on-off valve is continuously turned on (step 48).

(ff) Settling process mode (U=±1→0) In this mode, constant disturbance compensation and parameter estimation are performed based on the data of d obtained in the settling process.

(b) Estimation of constant disturbance From equation (6), the constant disturbance is y(to +N+ 2 ) and 9
(step 58).

Now, let d be d obtained by substituting ψ into equation (16), then d(to +N+2)"? (to +N+2) R
o −(16^) is obtained. However, RO is time to
This is the value of R held at .

Also, d(to 1N+2)=y(to +N+2) Ro
-(1613), so by subtracting both equations, d(to + N+2) d(to 10N+2)=
y(to +N+2) 9 (to 10N+2)...
・(16C) Therefore, by finding the value of the left side of equation (16C), the difference between the right side, that is, y(to +N+2) and 9 can be found.

In this case, y(to +N+2)-9(=KF FOΣ
The fact that aN 11・0 1-; I-;) is replaced with R again (Step 5
9) Steady-state deviations caused by changes in the neutral return spring over time are eliminated.

(b) Parameter estimation As explained in (v), this requires a considerable amount of calculation time, so in step 70- shown in FIG. 5, the parameter estimation start instruction flag is only set (step 60). ,
Actual calculations are executed when the parameter estimation start instruction flag is set in another 70- shown in FIG. 6 (steps 62 to 64). That is, step 5
4, the total (N+3> cl(t) data (dDo
) -d(to + N+ 2)) is stored in memory, so by sequentially substituting these data into equation (16), the same number of data of y(t) (y(to) -y
(to +N+2)) is obtained, and by substituting these data into equations (1) and (9), a and bi are calculated. Then, the calculated results of a and bi are updated as new parameter data.

Hydraulic systems are greatly affected by environmental conditions, so always
are not necessarily the same. For example, even if the on/off valve is turned off for the same amount of time, the spool may not move as expected due to its high viscosity at low temperatures, and may even move too much at high temperatures.

Therefore, it is impossible to predict the spool stop position without considering the difference in oil temperature.

On the other hand, in optical measurement, a and bi, which determine the beam, are not fixed values, but are treated as variable values depending on the state of the on/off valve and hydraulic system at that time, so they are the optimal values for the control state at that time. is set, and a highly accurate position control system is completed. Moreover, although high accuracy is achieved, it is achieved with an inexpensive on-off valve, unlike an expensive servo valve, and a balance with cost is also achieved.

Note that in step 61, the system is initialized (RAM, I
10), initial settings of the A/D converter, timer and interrupt functions, and initialization of system variables) are performed.

(Effects of the Invention) As explained above, in these four inventions, in a device that controls the position of a hydraulic actuator by turning the on-off valve on or off, when the control mode is continued, the While calculating the expected stop position of the actuator based on the position of the actuator and the parameters related to the dynamic characteristics of the control system, it is determined whether the expected stop position matches the target position of the 7 actuators, and if they match, Since the operation mode is switched so that the on-off valve is turned off, high-speed position control can be performed even if an inexpensive and highly reliable on-off valve is used.

[Brief explanation of the drawing]

Fig. 1 is a diagram corresponding to the claims of this invention, Fig. 2 is a block diagram of a control system of an embodiment of this invention, Fig. 3 is a diagram showing the selection conditions of the operation mode of this embodiment, and Fig. 4 is a diagram of this embodiment. A waveform chart showing an example of the response of this embodiment, FIG. tj & 5, and FIG. 6 are a flowchart showing the operation contents of this embodiment. 1... Hydraulic actuator, 2... On/off valve, 3
...Actual position detection means, 4. Operation mode selection means, 5. Drive signal output means, 6. Expected position calculation means, 7. Determination means, 8. Operation mode switching means. ,
22... Pilot valve, 25... Schilling, 26.
27...On-off valve, 28...Potentiometer,
29... Signal amplifier, 30... Sample/hold circuit, 31... Microcomputer, 32... Output amplifier. Patent applicant Michio Nakano

Claims (1)

[Claims] An on-off valve that is installed in the oil passage to the hydraulic actuator and opens and closes in response to a drive signal, a means for detecting the actual position of the hydraulic actuator, and a means for turning on the on-off valve by comparing the detected actual position with a predetermined target position. In a position control device comprising means for selecting an operating mode of operation or off operation, and means for outputting the drive signal when the control mode is selected, when the control mode is continued. means for calculating an expected stop position of the hydraulic actuator based on the actual position of the hydraulic actuator and parameters related to dynamic characteristics of the control system; and means for determining whether the calculated expected stop position matches the target position; and means for switching the operation mode so as to turn off the on-off valve when the on-off valve is turned off.