JPH01254891A - Plasma control system - Google Patents

Abstract

PURPOSE:To enable sure control of plasma only by several shots of discharges while the position and shape the plasma are stably maintained by performing an arithmetic operation on the detected signal of the electric current of a poloidal coil and a set current pattern by using a fixed function and automatically modifying the next discharging current pattern. CONSTITUTION:A plasma position which is an object to be controlled (y) is found based on an instruction reference (x) indicating the given shape of plasma. The electric current of the electric power supplied from the plasma to a poloidal coil 26 whose power source is a thyristor power source 25 is detected by means of a DC converter 27. The current patter to the coil 26 is set by means of a preprocessing waveform setting device 22 in advance. A plasma transmission function generator 23 performs an arithmetic operation on the current pattern and a detected output by using a fixed function. In accordance with the result of this arithmetic operation, the discharging preprocessing pattern of the device at the next time is automatically modified.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a plasma control system that controls the position and shape of plasma in a nuclear fusion device and maintains the plasma stably for a predetermined period of time.

[Conventional technology]

FIG. 8 is a block connection diagram of a conventional plasma control system. In the figure, 1 is a buffer memory that holds a pre-pro waveform signal created using a pre-pro waveform signal creation program prepared in advance in a small computer 9. The pattern generator 2 is a feed bank calculator 2a that corrects the pre-pro waveform signal held in the pattern generator 1 with a feedback signal from the plasma control device 8 as a current reference signal for the poloidal coil 5 corresponding to the plasma current 1p. is the current of the poloidal coil 5 which is the output of the DC current detector 7 and the feed bank calculator 2
3 is a current control circuit that controls the gate pulse signal to the thyrisk rectifier 4 by the output of the comparator 2a, 4 is a thyrisk rectifier that obtains the current to be applied to the poloidal coil 5, 6 is a plasma, and 8 is a The plasma control device 9 outputs a feedback signal based on the signal generated by the plasma, and the plasma control device 9 controls the execution of a pre-pro waveform signal creation program, etc., and also controls the graph ink display device 10, the keyboard device 11, and the flexible disk device 12. It is a small computer that does In addition,'
This graphic display device 10 is a small computer a.
The keyboard device II displays the pre-pro waveforms, numerical values, etc. created by the user using the pre-pro waveform signal creation program prepared in 9, and the keyboard device II displays the numerical values, symbols, etc. necessary for the user to create the pre-pro waveform signal. The flexible disk device 12 is used to input information, and is used to store programs used in a small computer such as a pre-pro waveform signal creation program, or to accumulate numerical values, constants, etc. of the pre-pro waveform signal created.

Next, the operation will be explained. All the functions provided in the small computer Ia9 are provided as programs, and they can be selected and executed from the menu. First, input the initial values necessary to create the pre-pro waveform. The input initial values are placed in a predetermined initial value storage area of the main memory of the small computer 9, and can be output to the printer device 13 if necessary, so that they can be corrected and corrected if necessary. Quoting the initial value of jjlj ON, the pre-pro waveform (
Create No. 3.

This device is a fusion experiment bag; it is a control system that uses a pre-pro waveform signal and a feedback signal together in σ, etc.
A small calculation stage is adopted as a pre-pro waveform setting device, and a pre-pro waveform is created in an interactive manner with the user using a pre-pro waveform creation program prepared in advance on the small computer 9, and this is used as a pattern connected to the small computer 9. It is held at a predetermined memory location in the buffer memory in the generator 1, and outputs a pre-pro waveform signal in response to a timing signal.

[Problem to be solved by the invention]

Since the conventional plasma control system is configured as described above, it is possible to reduce steady-state deviation and disturbance due to external disturbances.
In order to achieve this while maintaining stability, the creation of patterns that make subtle changes must be done entirely manually, and the next pattern must be determined by the operator's judgment after the experimental data has been determined. However, there were other problems, such as the work being too time-consuming and making it difficult to set patterns with high precision.

This invention was made to solve the above problems, and when an experiment is performed using a certain pattern, the pattern for the next experiment is automatically corrected according to a predetermined function set in advance using the experimental data. plasma control n
The purpose is to obtain a system.

(Means for Solving the Problems) The plasma control system according to the present invention supplies electric power to a poloidal coil by detecting a current using a current transformer or the like, and uses the detected output and a preset current pattern for the poloidal coil. A calculation is performed on the output using a fixed function, and based on the result of this calculation, the next pre-pro pattern of discharge by the pre-pro waveform setting device is automatically corrected.

[For production]

The plasma control TSL system of the present invention uses a small computer in the system to compare and calculate the set value of the poloidal coil current set by the pre-pro waveform setting device and the actual value of the poloidal coil TI current being measured, and calculates the result of this calculation. The current supplied to the poloidal coil is automatically controlled using a cyrisk or the like in order to correct the error.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. 1st
The figure is a plasma control block diagram, 21 is a power supply section, 2
2 is a pre-pro waveform setting device, 23 is a plasma transfer function generator, 24 is a controller, 25 is a Cyrisk power supply, 2
6 is a poloidal coil, 27 is a DC current converter (hereinafter referred to as
DCCT).

In this circuit, an instruction base lx indicating the shape of the plasma (for example, a circular cross section or a circular cross section) is given, and the purpose of this circuit is to 1) determine the plasma position to be controlled, that is, the control target y. Since it is difficult to learn and control the plasma shape itself, we measure the plasma shape when a certain current pattern is applied to a poloidal coil, and then perform learning control based on the plasma position correction to determine the next discharge. Poloidal coil? Create an itm pattern.

FIG. 2 is a block connection diagram of the plasma control device.

A tokamak-type fusion device consists of a toroidal coil that confines plasma and a poloidal coil that controls the plasma shape and position.

The S coil 35d among the poloidal coils is the main poloidal coil, and generates a magnetic field that suppresses the force that causes the large circumference of the donut-shaped plasma to expand.

Vup: l i JLz32d, Vtow -1 il 3
The 3d H-S coil is used as an auxiliary to the S coil 35d to control the fast movement of plasma. The Q coil 34d produces a magnetic diverter with practical power, and the OH coil 36d is an ohmic heating coil, which produces plasma current mainly by electromagnetic induction.

As mentioned above, the plasma position and shape are controlled by each voloidal current, but the necessary current for each coil depends on various factors such as pressure, and it is appropriate to find the optimal value by experimenting and learning the configuration of the present invention. It is. 31 is the plasma control panel, and is the control panel? It is installed in room II. 32 is V
Uρ power board A, 33 is V LQII power board, 34 is Q power board, 354; ts? J source board, 36 is OH? li [board,
Reference numeral 37 has Hti[1"t?, and is installed at each site. 38 is a data logger board, which receives the current signal from each power supply board from 32 to 37, performs data logging, and converts it into an optical signal. It is converted and sent to the plasma control panel 31 in the control room.The data logger panel 38 is located at the site and is installed near each control panel 32 to 37.The plasma control panel 31 includes a controller 31a. , General Purpose Interface Bus Interface (GENER)
AL PURPO3E INTERFACE B
us INTERFACE (hereinafter referred to as GPIBI/F) 316 and an optical CPIB link 31C are housed therein. The VUP power supply panel includes a coil voltage feedback resistor 32a, an insulation amplifier 32b for insulating and extracting the voltage signal from the main circuit, and a current reference analog signal sent as an optical signal from the plasma control panel 31. a controller 32C that controls the Vup coil 32d, a thyristor 32e that controls the current of the Vup coil 32d, a DC current detector (hereinafter referred to as DCCT) that measures the coil current, and an isolation amplifier 32g that sends the coil current signal to the data logger. is stored.

Regarding the power supplies 33 to 37, the same suffix indicates the same thing.

38a is a data logger, and the DCCT coming from each power supply
Receives a current signal and performs analog-to-digital conversion (hereinafter referred to as A/
D conversion) and serial data to optical CP I
After converting it into an optical signal through the B IJ link 38h, it is sent to the plasma control panel 3I. The thick line in the figure indicates that the signal is an optical signal.

Next, the plasma control operation will be explained.

The waveform signal set in the pre-pro waveform setting device 22 is sent to the controller 31a, and after performing calculations with the plasma measurement signal to create a current reference signal, it is converted into an analog optical signal and sent to each electric tX board 32 to 38. At the same time, a data acquisition start number 13 is sent to the data logger 38a of the data logger board 38, and the data logger 38a starts measurement. On the other hand, each controller 32c to 37c of the TUB boards 32 to 37 receives the current reference (No. 3), calculates a coil voltage feedback signal and a coil current feedback signal, and uses the calculated output as a sirisk gate signal. In addition, the DCCT current signal is transmitted to the isolation amplifier 3.
2. 37g to the data logger 38a, and after data logging, the plasma control panel 3
1 is input to the pre-pro waveform setting device. In the pre-pro waveform setting device, the following learning control is performed to create the next setting signal.

FIG. 3 is a configuration diagram of the learning control software system.

41 is a batch disk operating system that operates in single jog mode, and 42 is a command analyzer that manages initial value setting programs, waveform setting programs, etc., and starts these programs according to instructions from the operator. The program 643 is an initial value setting program T: Yes, each coil 32d to 37d
The first of 21 il data, electric power Xj, voltage limit value, etc.! This is used to set, display, and change Ul value parameters. 44 is a 1tL setting program for learning control, which is configured as one module of the above-mentioned initial jUl value setting program 43, and displays and changes setting values for data/4J sampling time, learning control channel allocation, and data logger manual allocation. 45 is a waveform setting program, which has the function of manually creating and modifying the current waveform flowing through the poloidal coil in the main memory of the 1γ machine. 46 is a learning control program, which is started by the waveform setting program 45 and is used to start and run a control program.
Control the data logger 38a via 1C, read the poloidal coil current waveform, perform learning control calculations according to operator instructions, and transfer the waveform data calculated using learning control to a DigiCrew analog converter (hereinafter referred to as a D/A converter). This is a program that has the function of transferring data to the puffa memory of

Here, the initial 1 full value setting program, t3, starts the initial value setting program 43 from the command ant riser 42, displays the initial ill value setting menu on the screen, and, in accordance with the operator's instruction, displays the initial value setting menu of the learning control 31! Jump to IJI 1a setting routine. In the learning control initial value setting routine, current set values are displayed and changed for the following three items.

(1) Selection of data sampling time (2)
Channel number for learning control and measurement channel assignment of data logger 38a 1 phase,
That is, 1.6 (i7 msec and twice that, 3.33
There are two options: 3m5ec, and the current setting value can be displayed and changed.

To assign channel numbers and data logger measurement channels for 0 learning control, first display the current setting values on the screen and change them using the operator device.

To select the measurement interrupt input address of the O data logger 38a, display it on the screen and press i! shall be translated.

After all input is completed, the screen is displayed, and the input value is entered manually from the keyboard to determine whether or not to replace the current setting value with the input value. The value is stored in the value setting table, and the data sampling period, etc. are initialized for the data logger.

If the character "N" is input, do nothing and move on to the next process.If any other character is input, sound Hell,
shall be re-entered. After the above processing is completed, the process returns to the initial value setting menu according to an operator instruction.

After the waveform setting program 45 is activated by an operator instruction from the command analyzer, a waveform setting menu is displayed on the screen and a learning control program is started.

FIG. 4 shows the entire flowchart of the learning side JLL program.

The learning side 2111 program is composed of the following three modules.

(1) Waveform data transfer to control [1-ra 31a (
2) Control of Cheetaroker 38a (:3)
Learning: 1,101 calculations First, the waveform setting program 45 performs learning control.

When the program 46 is started, first, the controller 31a
A routine for transferring waveform data to is started (step (51)), and the waveform data of the channel for which learning control is to be performed is transferred to the buffer memory of the D/A converter in the controller 31a. Next, check whether there is an error in the transferred data using the step knob (52)), and if there is no error, start the control routine of the data logger 38a (step 53)), and transfer the current waveform data to the data logger 38a. The data is then transferred to the main memory area of the 1γ machine. Next, it is determined whether or not there is an error in the transferred data (step (54)). If there is no error, a learning control calculation routine is started (step (55)), and learning control calculation is performed according to the operator's instructions. If learning control is to be performed as is by the operator's command input (step (56)), the program returns to the beginning of the learning control program 46 (step (57)) and repeats the same process. If it is necessary to make corrections to the learning-controlled waveform, return to the waveform setting program 45, use the waveform correction function of this program 45 to make corrections according to the operator's instructions, and return to the learning I11 control program again. . Then, when the learning control is finished, the process returns to the command analyzer 42.

FIG. 5 shows a flowchart of a data transfer routine to the buffer memory in the controller 31a.

First, to perform a transfer to buffer memory data,
Displays the buffer memory being written on the display (step (62)), and then outputs the waveform data in the common block to the general-purpose digital human output mechanism based on the learning control allocation channel set in the learning control initialization routine. The waveform data is transferred to the buffer memory of the D/A converter in the controller 31a via the (step (i3)), and then the waveform data in the buffer memory is read and compared with the transferred data to ensure accurate transfer. Check if the
step (64)).

While data is being transferred, the contents of the check are displayed on the screen so that the operator can understand the current status.

If the process ends normally, the error flag is set to zero and the process ends (step (65)).

If a defect occurs in the transferred data, the waveform data is transferred to the buffer memory again (step 66), and the buffer memory is read out and compared.

If a defect occurs again, the details of the defect are output to the printer (step (67)), the error flag is set to 1, and the process is terminated (step (68)).

When transferring data, the waveform data in the computer is data in engineering units in single-precision real number format, and the data in the buffer memory of the D/A converter is 12-bit data set to ±IOV full scale. The coefficients are stored in single precision real number format for each channel. Waveform data is stored in single-precision real number format for each channel from the beginning.
FIG. 6 shows a flowchart of the data logger control routine, which first resets the error flag and overflow flag to zero (step (71)), and then resets the error flag and overflow flag to zero (step (71)). Control the data logger connected to the computer, write a program in the data logger to start measurement when an interrupt signal is input to the data logger interrupt address, and start it (Step (72)) Screen Displays that the data locator is in standby mode7
Third step (73)) and wait until measurement is completed (step (74)). When the measurement is completed, display "Data measurement complete" and "Data transfer in progress" (step (75)), repeat step (76)) for 6 channels based on the data loss channel allocation table,
38a is read into the main memory area (step (77)). Next, it is determined whether there is a CPI+3 error (step (78)). If there is no CPI+3 error, the code conversion described later is performed. In addition, the measurement interval of the data logger is data sampling time of 1 + 667 + m.
5ec 1 m5ec, 3.333m5ec 2
m5ec, and the difference is corrected by time axis interpolation 84 using linear interpolation (step (84)).

Data from the data logger is sent in ASCII code, but this is converted into an integer with a length of 1+nd (step (79)). At this time, CPIB ink face CP
IB I/F data transfer error.

If the sent ASCII code is not normally converted into a numerical value, a code conversion error is determined (step (80)), and the old 1JJ11 data is read from the data logger again (step (81)). If an error occurs again, output the data logger channel where the error occurred and the status of the data logger to the printer, set "3" to the error flag, and end the process (step (82)
)). Also, the status of the data logger is read (step (83)).

If no error occurs during measurement with the data logger, time axis interpolation is performed to check the status of the data logger (steps (85) and (86)), and if an error occurs, Set "2" to the error flag and end the process (step (87)), and display the status of the error floor and channel (step (88)).
). At this time, the scale overflow flag in the status is checked (step (89)), and the overflow flag in the learning control program is checked (step (90)).

FIG. 7 shows a flowchart of learning control calculation.

First, display the coordinate axes on the screen (step (+00)
), channel fish, time axis, control method, etc. manually using the keyboard (101)), learning 1
．． The channel designation and time axis for performing the first reading are determined to be either time expression in ms units or Itji number expression color expression. Next, perform the previous settings and display the measured waveform (
Step (102)). For example, the setting waveform of the previous shot is shown by a dashed line, and the measured waveform is shown by a broken line. Next, input the learning control section and coefficient α (step (103)
)). Specifically, α0 in equation (1). Input the value of T,,T,'. Next, learning control/total 1γ is performed according to the second part of f, 4-r, +α, (d, -f,), and the next set waveform data is obtained (m is the channel number and 1<m<6). That is, 2fl calculations are performed (step (104)).

Next, the next setting waveform is displayed (step (105)).
) studies? After the l ifi'l f7ff operation is completed, the result is added as a solid line. Next, input whether to perform learning control calculations for other channels. If learning control is to be performed for two or more channels, repeat the above operation (step 7).
(106)).

To f, =r), +αa Cd 5(Ll
-f m'('Ll)...2(11 (T, ≦t≦T,') However, fa't'w' Target value of previous discharge at m-th channel, time t f+ll'l'Ll: Target value d+*ll for the next discharge at time t on the m-th channel: Data of the previous discharge corresponding to the m-th channel at time t α-: Correction coefficient To for the m-th channel: Learning side in calculation start time 'r, '
:Learning control calculation end time In this embodiment, a DC type current transformer is used as the current detector, but the same purpose can be achieved with a 1m shunt or an AC type current transformer. Further, the SIRISK power supply used to control the coil current may be a transistor power supply, and the calculation formula may be modified.

[Effects of the Invention] As shown in L below, according to the present invention, the set value and the measured value of the poloidal coil current are calculated by -7 according to a certain function, and the poloidal coil current for the next discharge is calculated according to the calculation result. Since the configuration is configured to automatically correct the position and shape of the plasma, it is possible to reliably control the position and shape of the plasma with just a few shots of discharge while maintaining stability.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a plasma control system according to an embodiment of the present invention, FIG. 2 is a circuit diagram of a plasma control i1 [1 system, and FIG. 3 is a system configuration diagram of learning control software. Figure 4 is a flowchart of the entire learning control program, Figure 5 is a flowchart of the buffer memory data transfer routine, Figure 6 is a flowchart of the data logger control routine, Figure 7 is a flowchart of learning control, and Figure 8 is the conventional flowchart. It is a block connection diagram of a plasma control system. 2■ is the power supply unit, 22 is the pre-pro waveform setting bag, 23 is the plasma transfer function generator, 25 is the Cyrisk power supply, 26
is a poloidal coil, 27 haho. Idal coil current detector.

Claims (1)

[Claims] (1) In a plasma control system that sets a current pattern of the poloidal coil, superimposes this current signal on a feedback signal of the position and shape of the plasma, and controls the plasma based on this superimposed signal, detects the current of the poloidal coil, A plasma control system characterized in that a calculation is performed using a certain function on this detection signal and the current pattern set above, and the current pattern for the next discharge is automatically corrected based on the calculation result.