JPH05297904A - Method and device for selecting optimum control system - Google Patents

Abstract

PURPOSE:To select the optimum control system at all times corresponding to a plant state and display the best suitableness between the current plant state and each control system by deciding and displaying the aptitudes of respective control systems in states of a controlled system at respective points of time according to the relation between the input state of the controlled system and the aptitudes of the respective control systems. CONSTITUTION:A water feed controller 100 has plural control systems composed of PI controllers 101, fuzzy controllers 102, neurocontrollers 103. A control system switching device 104 previously stores the relation between the state of the controlled system (feed water control valve 66) and the aptitudes of the respective control systems that the water feed controller 100 possesses in the state. The switching device 104 judges the aptitudes of the water feed control valve 66 at respective points of time according to the stored relation. The switching device 104 displays the aptitudes of the respective control systems or the state of the valve 66 according to the state of the valves 66 at this point of respective times inputting to the switching device 104, the respective controllers 101-103 and the normal/abnormal state of input information to them, and performs switching to the best control system.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for selecting and displaying an optimum control method in a control apparatus having a plurality of control methods.

[0002]

2. Description of the Related Art As a publicly known example close to the present invention, there is an applicable controller (Japanese Patent Laid-Open No. 230402/1990), which is an adaptive controller for changing a calculation formula or a calculation parameter according to the characteristics of a controlled object. The technique of selecting an adaptive controller is based on the relationship between the manipulated variable and the controlled variable, which is the input information of the controlled object. Another known example is a process control device (Japanese Patent Laid-Open No. 2-8903), which uses fuzzy control as a backup of PI control and fuzzy control when PID control becomes unstable during use. It is to switch to. Furthermore, a control method (Japanese Patent Laid-Open No. 1-250103)
No.) is for switching the PID and fuzzy when the stability of the control system becomes poor. It mainly shows the method for switching to the fuzzy control during the tuning of the PID controller. is there.

[0003]

With regard to the applicable controller, a plurality of control methods including PID control, neuro control, etc. are not switched, only the applicable controller is selected, and the control switching is performed by the manipulated variable. And only from the control amount. Therefore, no consideration is given to switching the control based on various detector information of the plant. The process control device uses fuzzy control as a backup of the PID and switches when the PID becomes unstable. However, the process control device switches the control method appropriate for the plant state and the controlled object becomes unstable. The control range is generally wider for PI control than for fuzzy, and this process control unit has PI control.
If you use fuzzy to back up
There is a problem that it is not possible to compensate for being able to control more stably than I. In addition, the control system is switched after the control system becomes unstable, so it is necessary to switch depending on the plant condition and switch the control system before the control system becomes unstable. I'm not here. Further, in these conventional techniques, since the input signal to each control system is common, there is little difference in control performance according to the state of the controlled object between control systems, and common control is performed due to abnormal input signals. There is a problem in that the control functions of each control method are inferior to each other because there is a high possibility that the performance will deteriorate. Further, these conventional techniques do not consider the function of displaying the relationship between the plant state and the control range and the function of guiding the optimum control method. In view of the above-mentioned circumstances, an object of the present invention is to always select the optimum control method corresponding to the plant state and display the present plant state and the optimality of each control method.

[0004]

Means for Solving the Problems In order to determine the suitability of each control method corresponding to the state of the controlled object, provide the information to the operator, and select the optimum control method, the state of the controlled object is set in advance. The suitability relationship of each control method that the control device has in that state is stored, and the suitability of each control method in the state of the control target at each time is determined based on the stored relationship. This determination result is displayed to the operator, the optimum control method is selected, and the selected control method is switched.

[0005]

The determination of the control method is based on the relationship between the state of the control object and the aptitude of each control method, which is input in advance, and the aptitude and the optimum value of each control method in the plant state at each time point input from the detector. Find the control method. The suitability display of each control method displays the suitability and optimum control method of each control method obtained by the above determination. The switching is performed by switching the control method according to the optimum control method determined by the above determination and the operator's request.

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a system configuration diagram showing a case where the present invention is applied to a water supply control device of a boiling water nuclear power plant. The water supply control device 100 controls the water supply flow rate (flow rate flowing through the pipe 68) for keeping the water level of the nuclear reactor 1 within a predetermined range, and the control of the water supply flow rate is constant as shown in FIG. The turbine-driven water supply pump (not shown) using only part of the steam flowing through the pipe 69 and only when controlling the opening of the water supply control valve 66 provided on the outlet side of the motor-driven water supply pump 8 that rotates with the torque No.) is installed in parallel with a motor-driven water supply pump (not shown) under different operating conditions, and its rotation speed is controlled (actually, the flow rate of steam to be supplied is controlled by opening / closing a valve). There are two types, and in the case of control by a motor driven water supply pump, there are two types: controlling a small valve used only at start / stop and controlling a large valve mainly used during rated operation. , Control according to these controlled objects In addition to the conventional technology for changing the parameters, it is also possible to use the method of changing the control method according to the plant condition shown in the present invention. There is a difference, but since it is basically performed by changing the valve opening, in order to make the explanation easier to understand, as shown in FIG. 1, one motor-driven water supply pump 8 and one water supply flow rate control valve 66 are provided. The case of controlling the water supply flow rate only will be described. A predetermined amount of water is contained in the nuclear reactor 1, which becomes steam due to the heat generated from the core 2, and the turbine 11 and the generator 13 are rotated through the main steam pipe 69 to generate electricity. The steam that has worked in the turbine 11 returns to water and is stored in the condenser 12. The water stored in the condenser is boosted by the condensate pump 70 and the feed water pump 8, and then adjusted by the feed water control valve 66 to a flow rate such that the water level in the reactor 1 becomes a predetermined value. It will be returned to the furnace 1. As described above, the purpose of the water supply control system is to keep the reactor water level within a predetermined range and to control it stably, even during transient times and abnormal times.

In the conventional control method, as main detector information for determining the reactor water level, the reactor water level signal 58 from the water level gauge 51, the main steam flow rate signal 59 from the main steam flow meter 52 and the feed water flow meter 53. P based on the water supply flow rate signal 60 from
Only I control was performed. This method has already been applied to many plants, and the theory of response and stability of the control system has been established and is clear, and since it is not delicate, it always performs reliable control. It applies to all conditions. In other words, since the input signal is relatively small and the control is basically based only on the reactor water level, the abnormalities of other detectors and the plant became abnormal, and the process amount could not be detected or showed an abnormal value. However, the control can be continued by cutting off the detector information if necessary.
Specifically, the reactor water level signal 58 and the main steam flow rate signal 5
9. It will not be affected even if detector signals other than the feed water flow rate signal 60 become abnormal, and the main steam flow rate signal 5
9. Even if one or both of the feedwater flow rate signals 60 becomes abnormal, by cutting off these two signals, it is possible to continue the control only with the reactor water level signal. Therefore, according to the PI control, although it is not necessarily optimal for all the states of the plant, reliable control is performed, and even if the plant is in an abnormal state that cannot be considered during normal operation, the reactor water level can be changed. It has a feature that it keeps operating in a direction to keep it within a predetermined range.

However, plant parameters that affect the water level of the nuclear reactor 1 are not limited to the input signals to the PI controller described above. For example, other detectable information is installed in the reactor core 2. Reactor output signal 6 calculated by the arithmetic unit 54 based on the signal from the neutron detector 6
1, a reactor pressure signal 62 from a pressure gauge 55 attached to the main steam pipe 69, a recirculation flow rate signal 64 from a flow meter 57 of the recirculation line 3 attached to control the reactor output, There is a furnace purification system flow rate signal 63 from the flow meter 56 of the furnace purification system piping 5 attached to bypass the cooling water from the circulation line 3 to the water supply system piping 68 and purify this cooling water. In addition, 4 shows a recirculation pump. In PI control, since only the water level and feed water-main steam flow rate deviation are observed, when a disturbance such as the above parameters fluctuates, the fluctuation is detected and an operation is performed to keep the water level stable in advance. It is relatively difficult to do so, and the effects are likely to be directly expressed in the form of fluctuations in the reactor water level. As a method of preventing this, various control methods such as optimal control and AI control (fuzzy, neuro, etc.) can be used in combination. In the present embodiment, the case where fuzzy control and neuro control are used in combination is shown. In the case of PI control as well, as in the above-described conventional technique, a simulated 3
Since one input signal is handled, it is possible to increase the number of input signals and convert it to a water level equivalent signal.
Basically, the relationship between the water level and other parameters is complicated, and it is difficult to convert easily, or even if a conversion function based on a detailed physical model is set, it will not be as large as a nuclear plant. Since it is difficult to create a detailed physical model itself in a large scale plant, it is difficult or ineffective to perform control based on a lot of plant information by PI control. Also, regarding fuzzy and neuro, control rules are created based on the knowledge and experience of human beings such as operators (fuzzy), or data (input signal) from the detector installed in the actual plant and the optimum value at that time. Since the neural network is constructed (neuro control) from the relationship of the control valve 66 opening (output) signal, there is a problem that it is difficult to perform stable control only within a range where these data are available. Therefore, in the present embodiment, the most suitable control method is selected according to the plant state. There are various conceivable ways of combining PI, fuzzy, and neuro control, the number of input signals of each control method, and the processing flow of each input signal. Here, PI control is the same as that of the conventional technique, fuzzy control. Regarding, regarding the reactor water level signal 58, the main steam flow rate signal 59, the feed water flow rate signal 60, the reactor output signal 61, the reactor pressure signal 62.
In regard to the neuro control, the case where the recirculation flow rate signal 64 and the furnace cleaning system flow rate signal 63 are further added to the input signal of the fuzzy control is shown as an input signal. Also,
For each combination of control methods, P that realizes PI control
An I control device 101, a fuzzy control device 102 that realizes fuzzy control, and a neuro control device 103 that realizes neuro control are provided in parallel in the water supply control device 100, and the output of each control device 101 to 102 is controlled by the control method switching device 104. It shall be switched according to the plant condition and operator's request.

FIG. 2 shows the relationship between the main plant parameters and the control range of each control method. In the example shown in FIG. 2, the reactor water level signal 5 is used as the plant parameter.
Controlling which control method the plant state at that time is based on a total of five plant parameters including a main steam flow rate signal 59, a feed water flow rate signal 60, a reactor output signal 61, a reactor pressure signal 62, and a recirculation flow rate signal 64. It indicates whether or not it falls within the range. The plant state shown in FIG. 2, the control devices 101 to 103 and the input signals 58 to 62 to them,
According to the normal / abnormal conditions of 64 and the operator's manual request, the control system switching device 104 causes the control devices 101 to 101 to operate.
A determination is made which output signal of 104 is used to control the feedwater control valve 66. First, regarding the neuro control, in this case, the neuro control can be selected only when all the input signals 58 to 62 and 64 are included in the area 23 including the areas 25 to 27 in FIG. In particular, when all the input signals 58 to 62, 64 are in the area 26, the neuro control can most stably control the water level of the reactor 1, and therefore the operator is informed of this fact. It will be. Further, even if all the input signals are in the area 23, if one or more signals are out of the area 26, PI or fuzzy control, which is a control method other than the neuro control, is more suitable. Specifically, if the fuzzy control stays within the areas 26 and 27, and if the fuzzy control extends to the area 25, the PI control becomes the optimum control. In this case as well, information to that effect is provided to the plant operator. Present. Next, regarding fuzzy control, in this case, area 2
The area 22 which is a combination of 5 to 29 becomes the fuzzy control range,
The fuzzy control can be selected only when all the input signals 58 to 62, 64 are included therein. In particular, when the signals 58 to 62 and 64 remain in the regions 26 to 28, the fuzzy control is the most suitable control method, and therefore the operator is notified of that fact. However, in this case, as described above, the signals 58 to 62, 64 are all in the area 2
If it is within 6, the neuro control is the most suitable control, and the operator is notified of that fact. Even if the signals 58 to 62, 64 are in the area 22, if at least one signal is in the area 25 or 29, the PI control is more suitable. Tell the operator to that effect. Finally, P
Regarding the I control, in this case, the PI control can be selected in all the regions 21 (or the regions 24 to 30). In this case, any one of all the signals 58 to 62, 64 is selected.
When one or more signals are in the areas 24, 25, 29 and 30, the PI control is the most suitable control method, and this is notified to the operator. In other cases, the most suitable control method in each state is as described above.

Next, selection of a control method when an input signal to the control devices 101 to 103 or each control device becomes abnormal will be described below. If the neuro control device or the signals 63 and 64 become abnormal during the selection of the neuro control system, the fuzzy control system will be automatically selected by the control system switching device 104. One or more signals of 62 and 64 are shown in FIG.
, If any one or more of the signals 58 to 62 is abnormal, or if the fuzzy control device 102 itself is abnormal, then P
The I control method will be automatically selected. In this case, if either or both of the signals 59 and 60 become abnormal, the PI control is performed by a single element using only the signal 58. When the signal 58 becomes abnormal, the manual control mode is automatically selected. In addition, while the fuzzy control mode is selected, the fuzzy controller 102, signals 59 to 6
If 2 is abnormal, the control method switching device 104 automatically selects PI control. However, signal 58-
62 and 64 are all in the area 26 in FIG.
Neurocontrol is automatically selected only if all signals 58-64 are normal and neurocontrol device 103 is normal. When the signals 59 and 60 are abnormal, the single element PI control is performed, and when the signal 58 is also abnormal, the manual switching is performed as in the case where the neuro control is selected. Finally, in the case of PI control, only when either one or both of the input signals 59 and 60 is abnormal, single-element PI control is performed only by the water level signal 58, and when the signal 58 is abnormal, manual operation is performed. It will automatically switch to control. Next, when the PI controller 101 is abnormal, all the signals 58 to 62, 64 are included in the area 26, and only when the neuro controller and the signals 58 to 64 are normal, switching to the neuro control is performed. Is done automatically. Also, all signals 58-62, 64
, Even if it is in the area 26,
When one or more of the signals 63 and 64 is abnormal, the fuzzy control is automatically switched. In this case, of course, the fuzzy controller 102 and the signal 5
8 to 62 should be normal. Next, the PI control device 101 is abnormal, and the signals 58 to 62 and 64 are in the area 26.
If it is within 28, it automatically switches to fuzzy control. In this case as well, the fuzzy controller and signal 5
It is clear that 8 to 62 need to be normal, and if all of the signals 58 to 62, 64 are in region 26 and the neurocontroller and signals 58 to 64 are normal, then Switching to the neuro control is as described above. When the PI control device becomes abnormal in a state other than the above, the control is automatically switched to the manual control.

The plant state (control device 10 shown above
1-103 and signals 58-64 include normal / abnormal states)
The control method switching device 104 in the water supply control device 100 of FIG. 1 selects the control method suitable for the above and the control method, and FIG. 3 shows a schematic configuration diagram of only those related parts. In FIG. 3, the central control panel 200 is a core part of a man-machine interface that connects an operator and many control devices other than the water supply control device 100 to operate the plant. In FIG. 3, only the minimum required signal connection is shown to explain the present embodiment, but in order to perform actual operation, the water supply control device 10
Even if it is only between 0 and 0, a larger number of signal connections are required, but in FIG. 3, those not directly related to the description of this embodiment are omitted. From the central control panel 200 to the water supply control device 100
The signals output to are the reactor water level setting signal 201, the control method request signal 202, and the valve opening request signal 203 used when manually controlling the feedwater control valve 66. The control system switching device 104 is a device that plays a central role in selecting a control system, and has a relationship between the plant state shown in FIG. 2 and a control range of each control system (including an optimal control system), and each of the previously described control systems. When the control device and the input signal become abnormal, which control device is used as a backup in each plant state (including the control device and signal abnormality), and the relation of the control device that cannot be selected in that state is stored in advance. Based on this stored data, the suitability of the control method (apparatus) according to the plant condition and the request from the operator is determined, and the optimum control method (apparatus) is selected. In order to make this selection, the main signals taken in by the control method switching device 104 are the signals 58 to 6 from the detectors provided in the plant.
4, the abnormal signals 204 to 206 from itself and the control method request signal 202 from the operator due to self-diagnosis performed in each of the control devices 101 to 102. All of these signals are combined via the input / output processing unit 105. In the control system switching device 104, the input signals 58 to 64, 20 are input on the basis of the relationship between the plant state (including the control device and the normal / abnormal state of the input signal) and the control range of each control system as described above.
2, 204 to 206, the suitability of the control system is determined, the optimum control system is selected, and the selected control system is sent to the control system switching unit 209 in the water supply control device 100 as the control system switching signal 207. The input signals 58 to 64
The abnormality diagnosis is performed in the input / output signal processing device 105, and the result is sent to the control system switching device 104 as a signal 210. The switching itself is performed in the control system switching unit 209, in which the output signals 301 to 303 from each control device and the valve opening request signal 203 manually by the operator are input, and the control system switching signal 207 is input. Switch according to. Control method switching unit 20
The signal selected in 9 will be sent to the valve actuator 67 as the control signal 65. The relationship between the plant state obtained by the control system switching device 104 and the control range of each control system is sent to the central control panel as a signal 208.
Here, it is provided to the operator as driving operation information.

Although various methods of providing operation information to the operator can be considered, this embodiment will explain a case where the drawings shown in FIGS. 4 and 2 are displayed on the CRT or the control panel. In the case of display on the CRT screen, a screen simulating operation buttons and display lamps mounted on the hardware board will be created. Since this is created by software processing, there are few restrictions on creation. Therefore, here, a case where the control panel is composed of hardware devices is shown as an example. There are various conceivable cases in which the diagram shown in FIG. 2 is provided on the central control board, but in this embodiment, small LEDs are arranged corresponding to each signal in order to display the value of each of the six input signals. Other lines and letters are written on the board with paint etc. to provide information. From this panel, it is possible to check the relationship between the control device and the plant state that does not include the abnormality of the input signal and each control range. The panel specifically embedded in the board is as shown in FIG. LED
Are filled in 5 units per scale of each detection signal, and the corresponding LED lights up from the value of each detector. L
Depending on the area where the ED is lit, the plant state at that time (not including the abnormalities of the control device and the signals 58 to 62, 64)
It is possible to know which control method is most suitable for each, and whether or not each control method can be selected, and it is also possible to confirm the temporal change of the plant state.

Next, the guidance function relating to the selection method of each control method and the control method other than the above will be described with reference to FIG.
Will be explained. The function shown in FIG. 4 can be realized by a method of embedding a hardware device on the CRT screen or the panel, but in the present embodiment, the hardware device is embedded on the board for the same reason as in the case of FIG. The case will be described.
In FIG. 4, a total of 2 is displayed on the plant status display panel 400.
0 lamps are lined up, and the control method switching device 10
In accordance with the signal 208 from No. 4, the lamps 401 to 404 indicating the optimum control method, the lamps 405 to 408 indicating whether each control method is selectable or not, and the reason why the control method is not selectable are shown at each time point. Illustrated by lamps 409-420. Specifically, taking the case of neuro control as an example, the neuro control device 103 and its input signal 5
8 to 64 are normal, and signals 58 to 62,
If all 64 are in area 26, then lamp 4
04 and 408 are turned on, and the area 25 and
If it has spread to 27, only the lamp 408 is turned on. Next, when the neuro control cannot be selected, the lamps 404 and 408 remain off.
The reason is that the areas of the signals 58 to 62 and 64 are 25 to 27.
If it spreads to other parts, the lamp 412 is turned on, the neuro control device 103 is abnormal, and the signal 20
When 4 is output, the lamp 416 is turned on, and when at least one of the signals 58 to 64 is abnormal, the lamp 420 is turned on. Lamps 401 to 40 for manual, P / I and fuzzy control
3,405-407,409-411,413-41
With respect to 5, 417 to 419, the lighting is performed in the same manner as in the above neuro control. Here, in FIG.
2 shows a signal processing flow for controlling lighting of 2 to 404 and 406 to 408.

Further, the operator selects the control method based on the information on the panel 400 and the panel shown in FIG. 6, which is a push button 450 to a selection button and a display lamp. 454. When each button is pressed, the request signal 20 for the corresponding control method is sent.
2 is output, and the part of the control method actually selected is turned on. Therefore, for example, P / I
While the control is selected, only the push button 452 is lit. If the lamp 408 is lit in this state, the push button 45 is lit.
By pressing 4, the push button 452 is turned off (the push button 454 is turned on), but no matter how much the push button 454 is pressed in the state where the lamp 408 is not turned on, the push button 452 that is turned on does not disappear. Further, the push button 453 does not light up either. When the manual control button 451 is pressed and manual control is selected, the water supply flow rate control valve 66 is directly controlled by the open push button 455 and the close push button 456 of the water feed control valve 66. In this case, manual control is performed while confirming the actual valve opening by the indicator 457 and the valve opening request signal 65 by the indicator 458. The automatic push button 450 can be selected only when any one or more of the optimum lamps 402 to 404 are lit, and when this push button is pressed,
When the push button 450 is lit, the control method corresponding to the lit lamp among 402 to 404 is automatically selected. The lamp 405 does not go out unless the part related to the manual control of the water supply control device 100 and related input signals fails, and the lamp 401 only goes out when the lamps 406 to 408 are off. It will be lit. While the push button 451 is lit, this is automatically excluded only when the lamp 405 is turned off. In this case, the push button 451 is turned off and the push button 450 is turned on, and the automatic mode is entered. .. Regarding the signal processing method performed in the PI control device 101, the fuzzy control device 102, and the neuro control device 103, no special function or signal processing is particularly necessary. Since the fuzzy control and the neuro control can be used, the description thereof will be omitted in this embodiment. As described above, according to the present embodiment, the plant operator is provided with the relationship between each plant state and the control range of each control method, the optimum control method at that time, and the applicability of each control method. Moreover, it is possible to easily select the optimum control method according to the state of the plant.

[0015]

As described above, according to the present invention,
It is possible to make the most of the characteristics of each control method and always select the optimum control method according to the plant condition. In addition, control method switching is not performed because the controlled object becomes unstable, so the controlled object can continue stable control at all times and the relationship between the plant state and each control method. From the display, even if the operator is changing the state of the plant, whether the control method selected at that time is changing the plant state in the optimal direction or Whether it is in the opposite direction can be confirmed promptly, and if necessary, backup can be performed by manually switching the control method or by manual control.

[Brief description of drawings]

FIG. 1 is a system configuration diagram when the present invention is applied to a water supply control system of a boiling water nuclear power plant.

FIG. 2 is a diagram showing a relationship between a plant state and suitability of each control method.

FIG. 3 is a system configuration diagram of a water supply control device.

FIG. 4 is a panel view of a control panel displaying a control method selection button, a plant state at each time point, an optimum control method, and whether or not each control method can be selected.

FIG. 5: Lamps 402 to 40 displaying the optimum control method
4 and lamps 406-408 showing selectable control schemes
FIG. 6 is a signal processing flow chart for turning on.

[Fig. 6] Plant status and suitability of each control method (control range)
Figure of a panel displaying relationships between.

[Explanation of symbols]

 1 Reactor 100 Water Supply Control Device 101 PI Control Device 102 Fuzzy Control Device 103 Neuro Control Device 104 Control Method Switching Device 66 Water Supply Control Valve

Claims (6)

[Claims] 1. A control device having a plurality of control methods stores in advance the relationship between the state of a controlled object and the suitability of each control method possessed by the control device in that state, and based on this stored relationship, A method for selecting an optimum control method, which comprises determining the suitability of each control method in a state of a controlled object at a time point and displaying the result of the determination. 2. A control device having a plurality of control methods stores in advance the relationship between the state of a controlled object and the suitability of each control method possessed by the control device in that state, and each is stored based on this stored relationship. An optimum control method selection method characterized by determining the suitability of each control method in a state of a controlled object at a time point, displaying the determination result, and selecting an optimum control method. 3. In a control device having a plurality of control methods, each time point of the control object is determined based on information about the relationship between the control range of each control method and the state of the control object which is predetermined and the state of the control device. To control the state at
A method for selecting an optimum control method, which is characterized by displaying how suitable each control method is. An optimum control method selection method characterized by determining the suitability of each control method in the state of the controlled object at each time point based on the stored relationship, displaying the determination result, and selecting the optimum control method. 4. The optimum control method selection method according to claim 1, wherein a PID control method is used as a backup of the fuzzy control method. 5. A control device having a plurality of control methods stores in advance the relationship between the state of a controlled object and the suitability of each control method possessed by the control device in that state, and each is stored based on this stored relationship. A control method switching device that determines the suitability of each control method in the state of the controlled object at the time point is provided, and the state of the controlled object at each time point input to this control method switching device, each control device, and the correctness of the input information to these An optimal control method selecting device characterized by outputting and displaying the optimality or suitability of each control method or the state of a controlled object according to an abnormal state, and selecting and switching the optimal control method. 6. The optimum control method selection device according to claim 5, wherein the control device having a plurality of control methods comprises a PI control device, a fuzzy control device, and a neuro control device.