JPH0813404B2 - Control method of molten metal level in twin-belt continuous casting machine - Google Patents

Description

Description: FIELD OF THE INVENTION The present invention relates to continuous casting, and more particularly, to maintaining a molten metal level in a mold of a twin-belt continuous casting machine to a constant level. Regarding adjustment of injection flow rate.

[Conventional technology]

Molten metal is supplied from the tundish to the continuous casting mold through a nozzle, and molten metal from a ladle is supplied to the tundish. Although the amount of molten metal in the tundish fluctuates, in order to safely and stably obtain high-quality products, the molten metal is supplied to the mold at a predetermined flow rate corresponding to the slab drawing speed and It is necessary to maintain the level (melt level) at a predetermined value. Conventionally, in order to adjust the molten metal surface level, the nozzle is used as a sliding nozzle to adjust the throttle opening of the nozzle and adjust the flow rate,
Alternatively, a linear motor is installed in the nozzle, etc., and electromagnetic force is applied to the molten metal in the nozzle, and the frequency of the AC voltage applied to the linear motor is adjusted to adjust the moving speed of the electromagnetic force. To adjust the flow rate (for example, Japanese Utility Model Laid-Open No. 44-17619 and Japanese Patent Laid-Open No. 60-99458).
Issue).

According to the twin-belt type continuous casting machine, since a slab with a small thickness can be directly manufactured, many of the rolling steps are omitted,
Great effects such as energy saving and equipment cost reduction can be obtained.

[Problems to be Solved by the Invention]

In the conventional melt level control in continuous casting, the melt level is detected, and the deviation between the melt level and the target value is multiplied by the feedback calculation coefficient to obtain the injection flow rate correction value.
Feedback control is performed to adjust the injection flow rate into the mold by adjusting the injection flow rate adjustment means such as a sliding nozzle and a linear motor. Even if this is applied to a twin-belt continuous casting machine, fluctuations in the molten metal level will occur. Is large and the casting quality is low, and operating problems are likely to occur.

Because the twin-belt continuous casting machine has a drawing speed of 5 to 10 times faster than the conventional continuous casting machine, and the cross-sectional area ratio of the nozzle to the mold is 1/3 to 1/10, the injection flow rate is This is because the influence coefficient of the fluctuation of the value on the fluctuation of the molten metal level is 10 times or more.

The mold of the twin-belt continuous casting machine is composed of a pair of opposed endless casting belts and a pair of opposed moving short sides sandwiched between them, and has a small internal volume of the mold and a thickness of the drawn slab. Since it is thin, the casting speed is relatively high. If the casting speed is high, a relatively large change in the molten metal level is likely to occur, and if the molten metal level fluctuates, the cooling conditions etc. will change.
Not only does the quality of the cast slabs vary, but troubles are likely to occur in the casting operation. Therefore, in order to maintain the quality of the cast slab continuously and stably and high, it is important to have a highly responsive and stable injection flow rate control, that is, a molten metal level control, to the molten metal level fluctuation.

It is an object of the present invention to provide a molten metal level control method which is highly responsive to fluctuations in the molten metal level in a twin-belt type continuous casting machine and stably maintains the molten metal level at a constant value. And

[Means for solving the problem]

The level of the molten metal in the mold is determined by the casting speed (slab drawing speed) and the flow rate of the molten metal injected into the mold. Therefore, the injection flow rate depends on the level of molten metal in the tundish, the level of the molten metal in the mold (this is kept substantially constant), and the setting of the injection flow rate adjusting means (sliding opening in the case of a sliding nozzle, in the case of a linear motor). Its energizing power and / or frequency). When the casting speed is changed, the setting of the injection flow rate adjusting means is changed in response to the change, so that the casting speed (constant) does not affect the fluctuation of the molten metal level during the stable period except the changing transition period. However, the level of molten metal in the tundish always fluctuates, and the level of the molten metal also fluctuates during the transitional period of changing the casting speed, and the casting speed that should be constant may also fluctuate. It acts as a disturbance on the level control of the molten metal in the mold.

Therefore, in the present invention, the amount of molten metal in the tundish and the withdrawal rate of the slab are measured to obtain the change rate of the withdrawal rate, and the feedback calculation coefficient is changed based on the measured amount of the molten metal in the tundish, The sum of the adjustment amount of the injection flow rate corresponding to the obtained change rate of the drawing speed and the value obtained by multiplying the molten metal level deviation value by the changed feedback calculation coefficient is obtained, and the injection flow rate adjusting means is based on the obtained sum. Adjust the injection flow rate with.

[Action]

There are various methods for improving the response characteristics of the feedback control system, but the results of the study by the inventors of the present application are as follows.
It was concluded that it is best to take advantage of the above-mentioned unique properties of the target process. That is, a control method is adopted in which a feedback control system is combined with a feedforward control system according to the amount of molten metal in the tundish and the withdrawal speed of the slab.

Specifically, as shown in the control block diagram shown in FIG. 6 or FIG. 7, the fluctuation of the injection flow rate into the mold due to the fluctuation of the molten metal level in the tundish is adjusted through a feedback calculation coefficient or by a coefficient. Is compensated as quantity,
In addition, during the transition period of the drawing speed change and during the drawing speed change, adjustment for canceling the molten metal level fluctuation corresponding to the change of the drawing speed is performed predictively (priority).

Therefore, the response of the molten metal level control in the mold is high, and the molten metal level is stably maintained at a constant value.

Other objects and features of the present invention will become apparent from the following description of embodiments with reference to the drawings.

FIG. 1 shows an apparatus configuration for carrying out the present invention in one aspect,
FIG. 2 shows an outline of the appearance of the continuous casting machine shown in FIG. The molten metal 2 of the tundish 1 is poured into the mold through a flat nozzle 3 having a narrow cross section in the X direction and a wide width in the Y direction perpendicular to the paper surface and having a rectangular cross section. In this embodiment, the mold is a twin-belt mold, and two mold belts 4 having a width wider than the width of the long side of the flat nozzle 3 (Y direction) and facing each other with the nozzle 3 in between, and the flat nozzle 3 are provided. It is composed of two short side molds 13 having a thickness wider than the width of the short side (X direction) and facing each other with the nozzle 3 in between.

The short side mold 13 is shown in detail in Japanese Patent Application Nos. 62-328080 and 62-328082.

The casting belt 4 is stretched around a driving roller 5. The drive roller 5 is rotationally driven at a predetermined speed by a DC motor 7 via a speed reducer 6. A finger speed generator (tachogenerator) 8 is connected to the motor 7 and generates an AC voltage having a frequency proportional to the rotation speed of the motor 7. This AC voltage is converted by the pulse processing circuit 11 into a pulse signal having a frequency proportional to the frequency and having a constant pulse height and pulse width. The F / V converter 12 generates a voltage (speed voltage) of a level proportional to the frequency and gives it to the motor driver 9. The F / V converter 12 also converts the voltage (casting speed voltage) into a voltage indicating the required speed of the molten steel in the nozzle 3 (molten steel injection speed: injection flow rate corresponding to the casting speed) (casting speed voltage × mold opening area). / Nozzle opening area), and the analog voltage v ₀ indicating this required speed v ₀
A / D conversion input port of ADr.

The motor driver 9 determines the rotation speed of the motor 7 based on the target speed (voltage) given by the motor controller 10, the feedback speed (voltage) given by the F / V converter 12, and the armature current (torque) of the motor 7. Adjust the armature current so that This allows the motor 7
Rotates at the target speed specified by the motor controller 10. That is, the belt 4 moves at the target speed.

A pair of linear motors 3A and 3B are installed to face each other with the long side (Y direction) of the flat nozzle 3 interposed therebetween. The relationship between these linear motors and the flat nozzle 3 is shown in FIG.

In this embodiment, the linear motors 3A and 3B each have a shape in which a stator of an induction motor having a three-phase star-shaped connection is developed in a plane, and each phase is formed in a slot between magnetic poles facing a rotor (molten steel in the nozzle 3). The coil is stored. By applying a three-phase alternating current having a predetermined phase relationship to each phase coil, an electromagnetic transfer force (deceleration force) from the bottom to the top in the Z direction is generated in the molten steel in the nozzle 3, and an alternating current applied to the two-phase electric coil is generated. By changing the voltage, an electromagnetic transfer force (acceleration force) from the top to the bottom in the Z direction is generated in the molten steel in the nozzle 3.

Each phase output line of the three-phase AC power supply circuit 24 has a thyristor inverter 23 that performs bidirectional conduction control for each line.
Also, the phase coils of the linear motors 3A and 3B are connected via the phase order switching circuit 22. The thyristor inverter 23 is
Application of the AC voltage of each phase to the linear motors 3A, 3B is conducted when a conduction trigger pulse is received from the thyristor driver 25 at each of the positive half wave and the negative half wave of the AC voltage, and the zero crossing of the AC voltage is performed. It becomes non-conducting at the point.

A power factor improving capacitor 21 is connected to a connection line between each phase coil of the linear motors 3A and 3B and each phase line of the three-phase AC. In this embodiment, in order to reduce the eddy current loss of the molten steel in the nozzle 3, the frequency of the three-phase alternating current is 100 to 500 Hz.
Since the range is preferable, it is set to 120 Hz. That is,
The three-phase AC power supply circuit 24 is connected to each of the three-phase output lines.
It outputs an AC voltage of 120Hz, which is 120 ° out of phase with each other. The power capacity of the linear motors 3A + 3B is 2,800 KVA at 120 Hz, and the capacitor 21 is set to 2,800 KVA corresponding to this. Conventionally, since there is no power factor improving capacitor, the required capacity of the inverter 23 is 2,800 KVA, but the capacity of the inverter 23 is 1,200 KVA due to the connection of the capacitor 21.
And, it is much smaller, which greatly reduces the cost of power supply equipment.

FIG. 5 shows the relationship between the energizing currents of the linear motors 3A and 3B and the deceleration rate. As shown in FIG. 1, this is when the phase sequence switching circuit 22 is set to "deceleration", and the molten steel injection speed becomes lower as the linear motor current is increased (the mold level is Lowers). Relay driver
When 27 is energized to drive the relay contact of the phase-order switching circuit 22 downward, "acceleration" is set, and the molten steel pouring speed increases as the linear motor current increases (the mold level rises). In FIG. 5, the horizontal axis represents the value of the current supplied to the linear motors 3A and 3B, and the vertical axis represents the molten steel velocity Voff in the nozzle 3 when the linear motor is not used.
Shows the ratio Vr of the molten steel speed Von when driven by a linear motor.

Referring again to FIG. 1, below the linear motor 3A, the level of the molten metal (distance of the molten metal from the video camera 28) Ld
A video camera 28 for detecting is taken, which picks up an image of a portion of the short-side mold 13 in contact with the molten metal surface and supplies a video signal to a signal processing circuit 29. Signal processing circuit
In 29, the color image data processing cuts out the tangent line between the molten metal surface and the short-side mold (on the screen where the inner surface of the short-side mold is imaged, high-temperature color) is cut out, and its position is either the upper or lower direction of the screen. The distance Ld is calculated by determining whether or not the position is present, and data indicating this is given to a microprocessor (hereinafter referred to as CPU) 30.

The load sensor 31 generates an analog voltage corresponding to the weight of the tundish 1 and supplies it to the signal processing circuit 32.
The signal processing circuit 32 processes the signal and supplies the CPU 30 with data indicating the molten metal amount W in the tundish.

The CPU 30 is further provided with a start / end signal and a target level L ₀ from a host computer or operation panel (not shown).
The target value of the distance of the molten metal surface from the video camera 28 is given, and the frequency divider 31 gives a pulse obtained by dividing the pulse indicating the speed (output pulse of the pulse processing circuit 11).

CPU30 is the standard value W of the molten metal amount W in the tundish.
_The deviation dW from ₀ is calculated to calculate the PI control calculation coefficients Kp and Ki corresponding to dW, and the casting speed change rate dv ₀ is calculated based on v ₀ which is proportional to the casting speed, and the corresponding adjustment amount is calculated. Calculate Kd · dv ₀ (Kd is a coefficient). Moreover, the deviation dL of the detection level Ld given by the signal processing circuit 29 with respect to the target level L ₀ is calculated, and the PI control adjustment amount is calculated based on the deviation dL. The adjustment amount Kd ・ dv ₀ is added to the control adjustment amount to calculate the injection speed vi of the molten steel into the mold for making the deviation dL zero, and this speed vi
To calculate the linear motor energization current value, the conduction angle of the thyristor inverter 23 (phase angle to turn on)
The voltage data Vf indicating the conduction angle is supplied to the D / A converter 26. The D / A converter 26 converts the data Vf into an analog voltage Vf and gives it to the thyristor driver 25. The thyristor driver 25 generates a voltage that gradually increases in proportion to the increase of the AC voltage phase with respect to the zero cross point for each of the three phases, compares it with the analog voltage Vf, and when the former reaches the latter, A trigger pulse is generated and applied to the gate of the thyristor of the inverter 23. Upon receipt of this trigger pulse, the thyristor becomes conductive and becomes non-conductive at the next zero cross point.

The control operation of the CPU 30 is shown in FIGS. 4a and 4b. First, refer to FIG. 4a. CPU30 when the power is turned on (Step 1: In the following, the word step is omitted in parentheses)
Sets the input / output port to the signal level in the standby state, clears the internal registers, counters, timers, etc., gives a "ready" signal to the host computer or operation panel, and then sends control data (calculation constants, timing It waits for the data such as constants that determine the control parameters) and the start signal to be sent. When the control data is sent, it is read and written in the specified register (internal RAM) (2,3).

When the start signal arrives, the interrupt INT is enabled (4), the timer T ₀ (a program timer that takes the time period of the time T ₀ ) is started, and the timer T ₀ is timed out (5,
6).

By enabling the interrupt INT, the CPU 30 executes the interrupt processing shown in FIG. 4b every time the frequency divider 31 generates one pulse. Explaining this, when the frequency divider 31 generates one pulse, the timer T ₀ is started (restarted) (1
0), the tundish weight W given by the signal processing circuit 32 is read, the deviation dW = W−W ₀ (W ₀ is a standard value) is calculated and written in the register Acw (29). Next, write the contents of register Acv ₂ (previous v ₀ read value) to register Acv ₁ (30),
The voltage v ₀ of the A / D conversion input port ADr is digitally converted, read, and written to the register Acv ₂ (31). And the rate of change dv ₀ =
The contents of the register ACV ₂ - Register ACV ₁ of the content, to calculate the writing to the register Acvd (32).

Next, the CPU 30 reads the level detection level Ld and the level target level L ₀ (11, 12). Then, the deviation dL is calculated, and this is written in the register Acd (13, 14). Next, PI (proportional integral) control of the proportional coefficient Kp = calculates the Kp ₀ + ap · dW written into register Acp (33). Note that Kp ₀ is a proportional coefficient when the standard weight is W ₀ , and is a constant in this case. αp is a constant that determines the proportional coefficient correction amount corresponding to the weight deviation dW, and dW is the content of the register Acw (tundish weight deviation). Next, the PI control integration coefficient Ki = Ki ₀ + αi · dW is calculated and written to the register Aci (34). Ki ₀ is an integration coefficient when the standard weight is W ₀ , and is a constant in this case. αi is a constant that determines the integration coefficient correction amount corresponding to the weight deviation dW.

Next, the CPU 30 sets the injection flow rate correction amount K corresponding to the casting speed change rate.
Calculate d · dv ₀ and write to register Acvc (35).

The CPU 30 then multiplies the deviation dL by the above-mentioned proportional constant Kp (contents of the register Acp) and writes it in the register Ac ₃ (15).
Next, the data of the accumulation registers R _{1 to} R _{n and} the data of R _{n -1} are
Write to Rn, write data from Rn _-2 to Rn _-1 , discard the oldest data (of Rn), and transfer the remaining data to registers R _{2 to} R _n (16 to 18) , Free register R ₁
The value obtained by multiplying the deviation dL by the integration constant Ki (contents of the register Aci) is written in (19). Then, the sum total of the data in the registers R _{1 to} R _n (the integrated amount of the correction value) is obtained and written in the register Ac ₄ (20). Then, the required value of molten steel in the nozzle 3 (injection flow rate) vi = v ₀ + Kp · dL + ΣKi · dL + Kd · dv ₀ is calculated by adding the adjustment value corresponding to the casting speed change rate to the output value of PI control (21). Next, the ratio Vr of the required speed vi to the target speed v ₀ in the nozzle 3 (proportional to the casting target speed) is calculated and written in the register Ac ₅ (22). From the data table previously written in the internal memory, Read linear motor current data Ii corresponding to Vr and write to register Ac ₆ (2
3). Next, the conduction phase angle data Vf that causes the current Ii is
It is read from the data table written in advance in the internal memory and written in the register Ac ₇ (24). And register Ac ₃ ,
It is determined whether the sum of the data of Ac ₄ and Acvc (correction amount for the target injection flow rate v ₀ ) is positive or negative (25), that is, whether the linear motor should be accelerated or decelerated, and positive (acceleration) In the case of, H is output to the relay driver 27 (2
7). As a result, the relay contact of the phase sequence switching circuit 22 is driven downward, and the linear motors 3A and 3B are connected to the inverter 23 for acceleration (driving downward in the Z direction). Negative (deceleration)
In the case of, L is output to the relay driver 27 (26).
As a result, the relay contact piece of the phase sequence switching circuit 22 comes to the position shown in FIG. 1, and the linear motors 3A and 3B are connected to the inverter 23 in deceleration (upward drive in the Z direction). Then CPU30
Updates and outputs the data Vf of the register Ac ₇ to the D / A converter 26 (28).

As described above, the driving directions and the driving forces of the linear motors 3A and 3B are corrected according to the detected value Ld, the molten steel amount in the tundish, and the casting speed change rate.

The interrupt process described above is executed each time the frequency divider 31 generates a pulse, and the integrated value of the deviation value obtained in each of the past n interrupt processes is written in the register Ac ₄ . It is embedded.

Time limit T ₀ of the timer T ₀ is, the continuous casting machine shown in Figure 2,
This is a time slightly longer than the period Tm of the pulse generated by the frequency divider 31 at the designed minimum speed. Therefore, DC motor 7, tachogenerator 8, pulse processing circuit
When 11 and the frequency divider 31 is normal, since the frequency divider 31 before the timer T ₀ is time-over occurs a pulse, no timer T ₀ is time over. Therefore, in the steady state, the interrupt process shown in FIG. 4b is repeatedly executed.

If for some reason the frequency divider 31 does not generate a pulse once during T ₀ , the interrupt process (Fig. 4b) will not be executed and the timer
When T ₀ times out, the CPU 30 proceeds from step 6 of FIG. 4a to step 30 to give an alarm signal to the host computer or operation panel (30A). Then, the timer T ₀ is started (restarted) (31A), input reading (A), PI control output value calculation (B), phase angle calculation (C), drive direction calculation (D) and output ( Execute E) and wait for the timer T ₀ to expire. The contents of these processes (A to E) are described in Section 4b.
This is the same as the processing content of steps 11 to 28 shown in the figure. Therefore, for example, when the frequency divider 31 does not generate any pulse at all, the above-described processing is executed in the T ₀ cycle.

The pulse generated by the frequency divider 31 determines the PI control sampling period in order to shorten the sampling period in inverse proportion to it when the casting speed is high. When the frequency divider 31 does not generate any pulses, the sampling cycle is T _0, which is a relatively long constant value.

When the end signal arrives from the host computer or the operation panel (7), the CPU 30 returns to the initialization (2). That is, the standby state (linear motor stop) is entered.

In the above-mentioned embodiment, a linear motor is used as the injection flow rate adjusting means, but a conventionally known sliding nozzle or a mechanically equivalent throttle opening for adjusting the nozzle opening is used, or any other automatic flow rate adjusting method is used. It may be changed to an adjustable means. Further, even when the linear motor is used as in the above embodiment, the injection flow rate may be adjusted by adjusting the current value or the frequency instead of adjusting the injection flow rate by adjusting the voltage value. On the other hand, the casting speed is measured by detecting the moving speed of the mold belt, but since the short side mold also moves at the same speed, the moving speed of the short side mold may be detected to measure the casting speed. .

〔The invention's effect〕

In any case, in the present invention, the injection flow rate of the molten metal into the mold is adjusted in accordance with the melt level deviation value dL, the molten steel amount W in the tundish and the casting speed change rate dv ₀ .
Fluctuations in the injection flow rate into the mold due to fluctuations in the level of molten metal in the tundish are compensated, and the fluctuations in the molten metal level corresponding to the changes in the casting speed are canceled out during the transitional period of the change of the drawing casting speed and the fluctuation of the casting speed Adjustments are carried out predictively. Therefore, the response of the molten metal level control in the mold is high, and the molten metal level is stably maintained at a constant value.

[Brief description of drawings]

FIG. 1 is a block diagram showing a device configuration for implementing the present invention in one aspect. FIG. 2 is a perspective view showing the appearance of the continuous casting machine shown in FIG. FIG. 3 is a perspective view showing the appearance of the linear motors 3A and 3B shown in FIG. 4a and 4b are flow charts showing the control operation of the microprocessor 30 shown in FIG. FIG. 5 is a graph showing the relationship between the current values of the linear motors 3A and 3B and the speed of the molten steel in the nozzle 3. 6 and 7 are control block diagrams specifically showing the control method of the present invention. 1: Tundish (Tundish) 2: Molten steel, 3: Flat nozzle (nozzle) 3A, 3B: Linear motor (injection flow rate adjusting means) 3As, 3Bs: Stator core, 3Ac, 3Bc: Electric coil 4: Casting belt (Endless casting Belt: 5: Drive roller, 6: Reducer 7: DC motor, 8: Tachogenerator 9: Motor driver, 10: Motor controller 11: Pulse processing circuit, 12: F / V converter 13: Short side mold (moving short side) 15: Cooling pad, 16: Small diameter split roll 17: Slab, 21: Capacitor for power factor improvement of linear motor 22: Phase sequence switching circuit, 23: Thyristor inverter 24: Three-phase AC power supply circuit, 25: Thyristor driver 26 : D / A converter, 27: Relay driver 28: Video camera, 29: Signal processing circuit 30: Microprocessor, 31: Load sensor 32: Signal processing circuit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Michiaki Kikunaga No. 1 Nishinosu, Oita-shi, Oita Pref. Nippon Steel Co., Ltd. Inside Oita Works (72) Hidetoshi Yuyama No. 1 Nishinosu, Oita-shi, Oita Pref. Steel Works Co., Ltd. Oita Steel Works (56) Reference JP 64-53747 (JP, A) JP 63-192545 (JP, A) JP 62-168652 (JP, A) JP 59 -30460 (JP, A)

Claims (4)

[Claims] 1. A twin belt for injecting molten metal from a tundish through a nozzle into a mold constituted by a pair of opposed endless casting belts and a pair of opposed short sided molds sandwiched therebetween. In the continuous casting machine, the injection flow rate adjusting means for the molten metal is arranged in the nozzle, the level of the molten metal in the mold is detected, and the deviation value of the detected molten metal level with respect to the molten metal level target value is multiplied by the feedback calculation coefficient. In the molten metal level control method for controlling the molten metal level in the mold to the molten metal level target value by adjusting the molten metal flow rate based on the measured value, the molten metal amount in the tundish and the withdrawal of the slab The speed is measured, the change rate of the drawing speed is obtained, and the feedback calculation coefficient is changed based on the measured molten metal amount in the tundish, and the change of the drawing speed A sum of an adjustment amount of the injection flow rate corresponding to the rate and a value obtained by multiplying the deviation value by the changed feedback calculation coefficient, and adjusting the injection flow rate by the injection flow rate adjusting means based on the sum. A method for controlling the molten metal level of a twin belt type continuous casting machine. 2. A twin belt for injecting molten metal from a tundish through a nozzle into a mold constituted by a pair of opposed endless casting belts and a pair of opposed short sided molds sandwiched therebetween. In the continuous casting machine, the injection flow rate adjusting means for the molten metal is arranged in the nozzle, the level of the molten metal in the mold is detected, and the deviation value of the detected molten metal level with respect to the molten metal level target value is multiplied by the feedback calculation coefficient. In the molten metal level control method for controlling the molten metal level in the mold to the molten metal level target value by adjusting the molten metal flow rate based on the measured value, the molten metal amount in the tundish and the withdrawal of the slab The speed is measured, the rate of change of the drawing rate is determined, the amount of adjustment of the injection flow rate corresponding to the amount of molten metal in the tundish, and the amount of adjustment of the injection amount corresponding to the rate of change of the drawing rate, A value obtained by multiplying the molten metal level deviation value by a feedback calculation coefficient and a sum of the values, and adjusting the injection flow rate by the injection flow rate adjusting means based on this sum, the twin-belt type continuous casting machine Level control method. 3. An injection flow rate adjusting means is arranged so as to sandwich the nozzle and generates an electromagnetic transfer force in an extending direction of the nozzle, and a voltage, current or frequency of a level specified for the linear motor. It is composed of an electric energizing circuit for energizing. The molten metal level control method for a twin belt type continuous casting machine according to claim (1) or (2). 4. The twin-belt continuous casting according to claim (1) or (2), wherein the drawing speed is measured by detecting the moving speed of the casting belt or the moving short side. Level control method of machine.