JPH01162901A - Controller - Google Patents

Abstract

PURPOSE:To set an optimum PID constant under the operating conditions of a wide range by storing previously the process gains to plural or continuous process values or the dead time into a memory part and correcting the PID constant based on said stored value. CONSTITUTION:An auto-tuning part 10 obtains a PID constant and sends this to a control arithmetic part 2 and an arithmetic part 11 respectively. At the same time, the part 11 obtains the process gain G1 and the dead time L1 by a process gain file 13 and a dead time file 14 respectively. Then the part 2 calculates the controlled variable based on the process value and the PID constant. When the control is carried on and the process value is changed to V2 from V1, the part 11 obtains the process gain G2 and the dead time L2 of that time point from the process value by reference to both files 13 and 14. Then the part 11 calculates a PID constant based on the process value and sends this to the part 2. Thus the control is ensured with an optimum PID constant over a wide range of operating conditions and in a simple handling way.

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to a controller that controls a heating furnace, etc., and particularly to a controller that can always perform optimal control under a wide range of operating conditions.

<Prior art> To control the temperature of a heating furnace, etc., a control variable is calculated by applying PID control to the difference between the furnace temperature, which is a process variable, and a set value, and this control variable is used to control the heat source such as a heater. Control. An example of such a controller is shown in FIG. In FIG. 3, process variables such as temperature signals are input to an input section 1 and converted into digital signals. This digital signal is input to a control calculation section 2 composed of a microprocessor or the like. Reference numeral 3 denotes a setting section, which outputs a set value, which is a target value of the process amount, to the control calculation section 2. The control calculation unit 2 calculates the deviation between these process variables and set values, and performs proportional, differential, integral calculations, etc. on this deviation to calculate the control amount. This control allffi is converted into an analog signal by the output section 4 and output. In heating furnaces and the like, it is required to control the temperature, which is a process variable, according to a certain predetermined pattern. Therefore, a pattern of changes in the set value may be set in the 5σ constant section 3 in the form of a program. The setting section 3 outputs a setting value that changes depending on the amount of time to the control calculation section 2 according to the set program. Furthermore, the set values are often manually manipulated by the operator.

A heating furnace is required to be controlled from a relatively low temperature of several hundred degrees to a high temperature of several hundred degrees with a single controller. As the temperature rises, more heat is radiated from the furnace wall of the heating furnace, so
The ratio of temperature rise to lllfl is reduced. That is, the process gain decreases. Furthermore, at high temperatures, radiation heat transfer becomes the main component, reducing wasted time. These reductions in process gain and waste time change the optimum value of the P [D constant for control calculations, so in order to perform accurate control, it is necessary to set the optimum PrD according to the process amount.
A constant must be selected. Furthermore, in the case of a resistance furnace, the resistance value of the heating element changes significantly depending on the temperature, in some cases by a factor of 10 or more. In such cases, it is especially necessary to select the most suitable PID constant.

An auto-tuning method using a limit cycle is known as one of the methods for selecting a PID constant. This auto-tuning method will be explained based on FIG. In this auto-tuning method, as shown in FIG. 4(A), control outputs of 0% and 100% are alternately output during control. Therefore, the process amount changes periodically as shown in (B).

That is, at time point (2), one hour after the time point (2) when the control output changes from 100% to 0%, the process 1Gt-maximum value is reached and then it begins to decrease. This L is wasted time. Further, the amplitude y, which is the difference between the saturation point in the increasing direction and the saturation point in the decreasing direction of the process amount, corresponds to the process gain G. Furthermore, the time T between the time point (2) and the time point (2) when the tangent to the decreasing curve of the process amount at that time point reaches the saturation point in the decreasing direction is a time constant of the process. The relationship between the amplitude a of the actual process quantity and these values is a-yL/T-GL/T. For example, according to the well-known Ziegler-Nichols optimal adjustment method, PIDIf11t! The optimal proportionality constant P1 differential constant D1 integral constant l for l1m calculation is P-83GL
/T・・・・・・・・・(1) There is a relationship of 1-0,51. The PrD constant can be obtained from the above-mentioned amplitude a1 dead time and these equations.

The above formula (1) is an example, and various formulas are actually used depending on the purpose of the process.

Further, as a method of giving the PID constant, there are a method of giving it by a program set by the setting section 3, a method of changing it depending on the size of a set value, etc.

<Problems to be Solved by the Invention> However, such a controller has the following problems. Auto-tuning using limit cycles is a method in which the control output is changed from 0% to 100% for several cycles during control, which results in disturbance to the process and disturbance of control.

Therefore, it cannot be executed at any time.

As mentioned above, the optimal PID constant changes depending on the process amount, so the optimum PID constant for all points to be controlled is
In order to perform J@, it is necessary to perform an auto-tuning operation at an arbitrary point, which is difficult.

Further, in the method of giving it in a program or switching it by a set value, it is necessary to reset the PID constant every time the program is changed, which has the disadvantage that it is cumbersome to handle.

<Object of the Invention> An object of the invention is to provide a controller that is easy to handle and capable of controlling i and lI with an optimal P'ID constant over a wide range of operating conditions.

Means for Solving the Problems In order to solve the above problems, the present invention provides at least one of data related to process gain and data related to amount of wasted time corresponding to a plurality of or continuous process amounts. is stored in the storage unit, the process gain or waste time corresponding to the process amount at each control point is calculated from the data stored in the storage unit, and P is calculated from these data.
An ID constant is calculated by a calculation section, and a control amount is calculated from this PID constant, process amount, and set value.

<Embodiment> FIG. 1 shows a block diagram of an embodiment of a controller according to the present invention. Note that the same elements as in FIG. 3 are denoted by the same reference numerals, and explanations thereof will be omitted. In FIG. 1, reference numeral 10 denotes an auto-tuning section, which executes the limit cycle auto-tuning explained in FIG. 4 to determine the PrD constant. This determined constant is input to the control calculation section 2. 11 is a calculation unit, which calculates the PID obtained by the auto-tuning unit 10.
A constant is entered. Reference numeral 12 denotes a storage unit, which includes a process gain file 13 and a waste time file 14. The process gain file 13 and the dead time file 14 each store process gains and dead times corresponding to a plurality of pro-his amounts. These data are provided to the controller by the process operator or control system designer. The process gain file 13 and dead time file 14 are read out by the calculation unit 11. FIG. 2 shows an example of changes in process gain and dead time with respect to process amount.

In this figure, (A) is the change in process gain, (B)
represents the change in wasted time. When the process M is the temperature of the furnace in a heating furnace, the process gain and dead time become smaller as the process amount (temperature) becomes higher. The process gain file 13 and waste time file 14 store changes in process gain and waste time with respect to the process amount in the form of tables. Data corresponding to process quantities not included in the table are calculated using methods such as linear approximation.

Note that it is not necessary to obtain and store the absolute values of the process gain and dead time, but it is sufficient to store the rate of change. For example, it may be stored as 100% at room temperature, 50% at 1000°C, 40% at 1200°C, etc. In addition, since only the rate of change is required, not the absolute value, there is no need to measure process gain and dead time for each device to be controlled; if the device is similar, it can be measured using a representative device. It can be applied to other devices as well. Furthermore, in a case where the amount of heat generated by a heater changes depending on the process temperature, such as in a resistance furnace, changes in process gain can be predicted relatively easily.

Next, the operation of this embodiment will be explained. At the beginning of control (process amount PV+), the auto-tuning section 10 adjusts the PI using the limit cycle auto-tuning method described above.
The O constant is determined and output to the control calculation unit 2 and calculation unit 11. Further, the calculation unit 11 calculates the process gain G1 and dead time L1 at this time from the process gain file 13 and the dead time file 14. Next, the control calculating section 2 obtains a set value from the setting section 3 and calculates a control amount based on the process amount from the input section 1 and the PID constant obtained by the autotuning section 10. When the control progresses and the process amount changes from PV + to PV2, the calculation section 11 refers to the process gain file 13 and dead time file 14 from the process amount input from the control calculation section 2, and determines the process at that time. The gain G2 and the waste time B L 2 are determined, and the PID constant for that process amount is calculated and output to the control calculation section 2. From then on, the control calculation unit 2 executes control calculations based on the input PrD constant. Assuming that the process gain and dead time in processes flPV+ and PV2 are respectively G+ -G2, L+, and L2, and the optimum PfD constants in process ff1Pv+ obtained by the autotuning section 1o are P+ and D+1 ft, the PID in process 11 PV2 is Constants P2, D2.
12 can be calculated as P2-P+ ・G2/GI D2 = D+ −L2 /L1 12=II ・L 2 /L +. The above calculation is performed at appropriate time intervals or at each control cycle to update the P4D constant. Further, although this explanation has been made for two points, P V + and P V 2, for convenience, similar calculations are actually performed over the entire range of process quantities. Furthermore, although autotuning is performed at the beginning of control, similar results can be obtained by performing autotuning at any convenient point during control.

Note that, as is clear from the above equation (1), the proportionality constant P also depends on the ratio L/T of the dead time and the time constant d. However, the value of L/T may be kept constant since it hardly changes unless the physical quantity inside the furnace changes significantly. Also, when the relationship between the process amount and L/T can be estimated, the value stored in the process gain file 13 may be corrected by the ratio of dead time to time constant L/T, GL/T. good.

Further, in this embodiment, both the process gain and the dead time are stored, but if one is approximately constant, only one may be stored and the other may be treated as a constant.

Furthermore, although this embodiment has been described with reference to the case where it is applied to a heating furnace, it is also applicable to other processes.

Further, in this embodiment, the PID constant is determined by the auto-tuning section 10 at the beginning of control, but the PID constant may be manually set by the operator without using the auto-tuning section 10. In this case, if the optimal control state can be confirmed at any one point in process m, appropriate control states can be obtained at other process values as well.

Further, in this embodiment, the control calculation section and the calculation section are configured separately, but they may be executed by one microprocessor. Further, the data stored in the storage unit may not be in a table format, but may be approximated using a polynomial or the like, the coefficients thereof may be stored, and the calculation unit 11 may calculate the data.

<Effects of the Invention> As specifically explained based on the embodiments above, in the present invention, at least one of the process gain and dead time for a plurality of or continuous process quantities is stored in advance in the storage unit, and the stored time is stored in the storage unit in advance. The P■D constant is corrected based on the value.

Therefore, the optimum PD constant can be set over a wide range of operating conditions, and since disturbances do not occur during control unlike in auto-tuning, accurate control can be achieved.

Furthermore, even if auto-tuning is not used, it is only necessary to set the PID constant at an appropriate point, making handling extremely easy.

Although this explanation has focused on controlling a heating furnace, it is also applicable to other processes. For example, near the isoacid point for pH control in neutralization reactions, the process gain is extremely large and automatic tuning is not possible. However, in the present invention, changes in process gain and dead time due to process amount are stored in advance. By performing auto-tuning at an appropriate point and automatically correcting and applying the PID constant obtained at that time, control can be performed even near the current point m.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the controller according to the present invention, FIG. 2 is a characteristic curve diagram showing changes in process gain and dead time with respect to process amount, and FIG. 3 is a diagram showing the configuration of a conventional controller. The block diagram shown in FIG. 4 is a characteristic curve diagram for explaining the operation of auto-tuning using a limit cycle. 2... Control calculating section, 3... Setting section, 10... Auto tuning section, 11... Calculating section, 12... Storage section, 13... Process gain file, 14... Waste time file.

Claims (2)

[Claims] (1) A storage unit that stores at least one of data related to process gain or data related to wasted time corresponding to multiple or continuous process quantities, and a memory unit that corresponds to process quantities from the data stored in this storage unit. Find the process gain or wasted time and use these values to calculate the PI
1. A controller comprising: a calculation section that performs a correction calculation on a D constant; and a control calculation section that calculates a control amount from a PID constant, a process amount, and a set value determined by the calculation section. (2) The controller according to claim 1, further comprising an auto-tuning section for determining an optimal PID constant under predetermined process conditions.