JPH0713640A - Fuzzy pi temperature control method - Google Patents

Abstract

PURPOSE:To reduce the hunting width of the temperature by setting the deviation coefficient and the velocity coefficient in an area where an object temperature is lower than a target temperature and the object temperature rises smaller than the ones in the other areas. CONSTITUTION:Temperature deviation is defined as an X axis, the fluctuation velocity of the temperature is defined as a Y axis and a coordinate value in an X-Y plane 21 is obtained. The coordinate value indicates one of the squares of a loop-up table and a value inside the square corresponding to the coordinate value becomes a fuzzy inference value. The fuzzy inference value obtained in such a manner is used as the output value of a velocity type and the value of output control is calculated corresponding to the output value and supplied to a heating/cooling means. The area 12 where the object temperature is lower than the target temperature and the object temperature rises is defined as a first area and inside the area, the deviation coefficient and the velocity coefficient are set to be smaller compared with the ones in the other areas 11, 13 and 14. Thus, the hunting width of the temperature can be reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fuzzy PI temperature control method for controlling dead time + temperature of first-order lag system controlled object or dead time + second-order lag system controlled object temperature using a fuzzy rule. .

[0002]

2. Description of the Related Art A system (Japanese Unexamined Patent Publication No. 4-227502) has been proposed in which the die temperature of an extrusion molding machine, which is a control target of dead time + first-order lag system, is controlled by using a fuzzy rule. However, in this case, since the resin generates heat, it is necessary to perform both heating and cooling in order to maintain the mold at the target temperature, and the method shown below is being studied for its control.

If the temperature of the mold is called the target temperature and the difference between the target temperature and the target temperature is called the temperature deviation, then the temperature deviation is on one axis and the rate of change of the target temperature is on the other axis. Then, the fuzzy rule composed of the temperature deviation and the changing speed is expanded as numerical data, and the temperature deviation-change speed plane of the expanded data table is associated with the XY plane, and the coordinates of the XY plane are associated. The mold temperature is controlled using the fuzzy inference value.

As shown in FIG. 4, the coefficient for associating the temperature deviation-change speed plane with the XY plane is set so that the coefficient during heating is smaller than the coefficient during cooling, and fluctuations in the control output value during heating. By increasing the rate, when the mold temperature becomes lower than the target temperature, the control method is to quickly heat the temperature to the target temperature.

[0005]

When the target temperature falls and becomes lower than the target temperature in the above control, the history of the target temperature (shown by the curve 91) moves from the fourth quadrant to the third quadrant in FIG. It will be. Therefore, when the above control method is used, the same changing speed (Y corresponding to the position 92)
4) and 3rd quadrant
In the quadrant, the coefficient for associating the temperature deviation-rate of change plane with the XY plane is small on the side of the third quadrant.
When the history 91 moves from the fourth quadrant to the third quadrant,
The fuzzy inference value discontinuously increases its value.

This results in the control output rapidly increasing (indicated by 93 in FIG. 5), so that the target temperature also largely overshoots after exceeding the target temperature (indicated by 94). Since the cycle of cooling it again is repeated, there arises a problem that the hunting width of the temperature of the controlled object becomes large.

The present invention was devised to solve the above problems, and an object thereof is to provide a fuzzy PI temperature control method capable of reducing the hunting width of the temperature of the controlled object.

[0008]

In order to solve the above problems, the fuzzy PI temperature control method of the present invention uses a dead time + the temperature of the controlled object of the first-order lag system, or the dead time + the temperature of the controlled object of the second-order lag system. Is the target temperature, and the difference between the target temperature of the control target and the target temperature is the temperature deviation, the temperature deviation is in one plane and the rate of change of the target temperature is in the other axis. The fuzzy rule composed of the change rate and the change rate is developed as numerical data, and the temperature deviation-change rate plane of the developed data table is associated with the XY plane converted to an arbitrary scale, and the XY plane is set. It is applied to the control method that controls the temperature of the controlled object by using the fuzzy inference value corresponding to the coordinates of the target temperature. When the area is the first area and the other areas are the second areas, and when each of the coefficients for associating the temperature deviation and the changing speed with the coordinate values on the XY plane is the deviation coefficient and the speed coefficient, The respective values of the deviation coefficient and the speed coefficient in the area 2 are set to be smaller than the respective values of the deviation coefficient and the speed coefficient in the second area.

[0009]

Since the target temperature is higher than the target temperature and the rate of change is negative, and the target temperature is lower than the target temperature and the rate of change is negative, both regions belong to the second region. The value of the deviation coefficient and the value of the changing speed that associate the changing speed plane with the XY plane are equal in both areas. Therefore, even when the target temperature drops and crosses the target temperature, the fuzzy inference value becomes a value that continuously changes, and the control output value changes smoothly.

When the target temperature changes to the first region where the temperature rises, both the deviation coefficient and the speed coefficient are set to small values, so the output changes greatly,
The target temperature quickly rises toward the target temperature.

[0011]

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

FIG. 1 is an explanatory view showing a temperature deviation-change speed plane of an embodiment of the fuzzy PI temperature control method of the present invention.

The present embodiment is a control method in which the temperature of the die of the extrusion molding machine, which is the control object of the dead time + secondary delay system, is controlled by heating or cooling.

Therefore, the horizontal axis of the temperature deviation-change rate plane 10 shows the temperature deviation obtained by subtracting the set target temperature from the target temperature which is the temperature of the controlled object, and the vertical axis shows a constant value. It shows the rate of change of the target temperature obtained by detecting the change in temperature deviation at time intervals.

The areas 11 and 14 in which the temperature deviation is positive
2 corresponds to the value 1 on the X axis of the XY plane 21 shown in FIG. 2, and the deviation coefficient −Ec of the region 13 in which the temperature deviation is negative and the rate of change is negative is the value on the X axis. It corresponds to -1. And the speed coefficient Δ of the area 11
Ec is made to correspond to the value 1 on the Y-axis, and the velocity coefficient −ΔEc of the regions 13 and 14 is made to correspond to the value −1 on the Y-axis (region 1
1, 13, 14 correspond to the second region of claim 1).

Further, in the region (first region) 12 in which the target temperature is lower than the target temperature and the temperature deviation is negative, and the target temperature is rising and the speed coefficient is positive (first region) 12, the value is smaller than the speed coefficient ΔEc. Deviation coefficient ΔEh is XY
It corresponds to the value 1 on the Y-axis of the plane 21, and the deviation coefficient is smaller than the coefficient -Ec by the coefficient -Eh.
Corresponds to the axis value -1.

That is, X, Y on the XY plane 21 shown in FIG.
22 where each value of -1 ≦ X ≦ 1 and −1 ≦ Y ≦ 1
Indicates the temperature deviation-change rate plane regions 11 to 1 shown in FIG.
It corresponds to 4.

This means that even if the changing speed is the value A in FIG. 1, when the changing speed is the changing speed in the area 11, it corresponds to the value A11 in the XY plane 21. If the change rate is 12, the value corresponds to the value A12. This also applies to the temperature deviation.

That is, even if the temperature deviation or the changing speed is the same, they are in the area (first area).
12 and other area (second area) 1
Corresponding values on the XY plane 21 are different from those belonging to 1, 13, and 14.

On the other hand, with respect to the XY plane 21, the X axis,
Each value on the Y axis is divided into 21 coordinate values at equal intervals, and a fuzzy inference value configured with the temperature deviation on the X axis and the changing speed on the Y axis is calculated in advance for each intersection of the coordinate values. The calculated values are stored in a lookup table, which is a matrix divided into 21 in each of the horizontal and vertical directions.

From the above, in the temperature control of the die (control target) of the extruder, the temperature deviation is first calculated from the detected target temperature and the set target temperature, and then the temperature deviation changes. The rate of change is calculated from Then, the coordinate value on the XY plane 21 is calculated from the calculated values.

Since this coordinate value is a value indicating one of the cells in the lookup table, the value in the cell corresponding to the coordinate value is the fuzzy inference value.

The fuzzy inference value thus obtained is treated as a velocity type output value, and the control output value is calculated in accordance with this output value in the output processing unit for controlling the temperature of the die. Then, the calculated control output is given to the heating and cooling means of the mold.

FIG. 3 is an explanatory diagram showing changes in the target temperature and changes in the control output. The operation of the actual machine will be described below with reference to FIG.

In this embodiment, the temperature of the mold is controlled by both heating and cooling. Therefore, by comparing the target temperature with the target temperature, it is determined whether or not to start the cooling control.

When the target temperature becomes higher than the target temperature, the cooling control is started, and each value of the deviation coefficient Ec, -Ec, Eh in the temperature deviation-change speed plane 10 is calculated, and the speed coefficient ΔEc,-. The temperature deviation-change speed plane 10 and the XY plane 21 are associated with each other by calculating the respective values of ΔEc and ΔEh.

Now, assuming that the target temperature of the mold starts to rise from the temperature shown as the center O (the temperature which coincides with the target temperature), the history 31 moves in the area 11, so that the temperature deviation and the changing speed are changed. Cooling is performed according to the fuzzy inference values obtained corresponding to and. Therefore, the target temperature eventually decreases, and the history 31 moves to the area 11 or the area 14 (indicated by 32).

When the history 31 is moved from the area 11 to the area 14, the history 11 and the area 14 are respectively marked with X.
-Deviation coefficient Ec and speed coefficient ΔE associated with Y plane 21
Since c and -ΔEc have the same coefficient in the area 11 and the area 14, the fuzzy inference value is obtained as a value that continuously changes.

Since the cooling is also performed in the area 14, the history 31 moves toward the area 13 (3
However, the fuzzy inference value is continuous because the correspondence between each of the regions 13 and 14 and the XY plane 21 is the same coefficient. It will be obtained as a value that changes over time. This means that the control output value of the mold temperature does not change significantly.

When the target temperature rises due to heating, the history 31 moves from the area 13 to the area 12 (indicated by 34).

At this time, the deviation coefficient of the area 12-Eh
And the speed coefficient ΔEh are respectively set smaller than the deviation coefficient −Ec and the speed coefficient ΔEc, and the corresponding coordinate values are different on the XY plane 21, so the fuzzy inference value increases discontinuously. That means
The control output value increases rapidly. Therefore, the target temperature rises faster, and eventually exceeds the target temperature.

However, the history 31 changes from the area 12 to the area 11
Regarding the changing speed when shifting to, the heating control in the area 13 is controlled by the coefficient continuous with the area 14, and the output fluctuation rate is suppressed, and the fuzzy inference value is the speed type. Since it is controlled, the changing speed (indicated by B) is slower than the changing speed in the related art.

Therefore, the temperature deviation (indicated by C in FIGS. 1 and 3) in the area 11 remains at a small value. Hereinafter, the history 31 moves to the regions 14, 13 and 12, but since the temperature deviation (C) remains small, the temperature deviation (D) at the boundary between the regions 13 and 12 is small. Also remains small.

That is, the hunting width (W) of the mold temperature (target temperature) remains at a small value, and the accuracy in temperature control is improved.

The present invention is not limited to the above embodiment, and the control object is the dead time + second-order lag system of the extrusion molding machine. The same can be applied to the control object of the time + first-order lag system, or the dead time of the units other than the extruder and the control object of the second-order lag system.

[0036]

According to the fuzzy PI temperature control method of the present invention, the fuzzy rule is expanded as numerical data in a flat data table having the temperature deviation and the rate of change of the target temperature as axes, and the temperature of the expanded data table. The deviation-change speed plane is associated with the XY plane converted to an arbitrary scale, and is applied to the temperature control method using the fuzzy inference value corresponding to the coordinates of the XY plane. Then, a coefficient that associates each of the temperature deviation and the changing speed with the XY plane is defined as a deviation coefficient and a speed coefficient, a region where the target temperature is lower than the target temperature and the target temperature rises is the first region, and the other regions are the first regions. When referred to as the second area, the first
The respective values of the deviation coefficient and the speed coefficient in the area 2 are set smaller than the respective values of the deviation coefficient and the speed coefficient in the second area. Therefore, even when the target temperature drops and crosses the target temperature, the fuzzy inference value becomes a value that continuously changes, and the control output value changes smoothly. Further, when the target temperature is in the first region where the target temperature starts to rise, both the deviation coefficient and the speed coefficient are set to small values, so that the output greatly changes and the target temperature quickly becomes the target temperature. Since it rises toward the front, it is possible to reduce the hunting width of the temperature.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a temperature deviation-change rate plane according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an XY plane of an embodiment of the present invention.

FIG. 3 is an explanatory diagram showing changes in a target temperature and a control output according to an embodiment of the present invention.

FIG. 4 is an explanatory diagram showing a temperature deviation-change rate plane in the related art.

FIG. 5 is an explanatory diagram showing changes in a target temperature and a control output according to a conventional technique.

[Explanation of symbols]

 10 Temperature Deviation-Change Speed Plane 11, 13, 14 Second Area 12 First Area 21 XY Plane Ec, -Ec, -Eh Deviation Coefficients ΔEc, -ΔEc, ΔEh Speed Coefficient

Claims (1)

[Claims] 1. A dead time + a temperature of a control target of a first-order lag system or a dead time + a temperature of a control target of a second-order lag system is set as a target temperature, and a difference between the target temperature of the control target and the target temperature is a temperature deviation. When, and the temperature deviation on one axis, in the data table of the plane with the rate of change of the target temperature the other axis, while expanding the fuzzy rule configured by the temperature deviation and the rate of change as numerical data, The temperature deviation-rate of change plane of this expanded data table is associated with the XY plane converted to an arbitrary scale, and this X-
In a control method for controlling the temperature of the controlled object using a fuzzy inference value corresponding to the coordinates on the Y plane, a region in which the target temperature is lower than the target temperature and the target temperature rises is defined as a first region. , The other area is the second area, and the respective coefficients that associate the temperature deviation and the changing speed with the XY plane are the deviation coefficient and the speed coefficient, the deviation coefficient and the speed coefficient of the first area are A fuzzy PI temperature control method, wherein respective values are set smaller than respective values of the deviation coefficient and the speed coefficient in the second region.