Analysis of Circuits Supplying Thomson Coil Actuator Operating in Vacuum Contact Units of DC and AC Ultra-Fast Circuit Breakers

Abstract

The use of vacuum-hybrid DC circuit breaking methods allows the short-circuit current to be switched off in a shorter time, resulting in a reduction in the arc burning time. This requires the use of a drive, such as the Thomson Coil Actuator TCA, capable of providing a short response time for opening the vacuum interrupter VI, regardless of its rated current. The IDD is powered by a pre-charged capacitor, which, together with the drive coil, forms an LC oscillating circuit that, when switched on by a thyristor, generates a current pulse of several kA with a frequency above 1 kHz. The paper investigates the effect of modifying the basic IDD power supply circuit by adding semiconductor diodes to shape the current pulse and improve its performance. The authors also focused on exploring the impact of the connection quality and their length and the associated loss in drive force while proving that a circuit with a reverse diode on the IDD coil is most beneficial and that the effect of the circuit on the front of the current pulse can significantly slow down the drive.

1. Introduction

The hybrid switch combines the advantages of a mechanical and a solid-state switch. This technology can find application in railway traction [1,2,3], for which the authors of the articles emphasize the need to use a high-speed drive to open the contacts. Switches of this type are used in marine solutions [4,5,6], and as protection for medium-voltage equipment [7,8,9]. The development of renewable energy technologies is resulting in increased interest in high voltage direct current HVDC transmission lines and their protection [10]. In studies [11,12,13], the authors indicate that the lack of a suitable hybrid switch hinders the development of this type of line. Chang Peng et al. [14] described a vacuum contact unit VCU rated for 15 kV and 630 A, which uses an inductive–dynamic drive IDD, also known as Thomson coil actuator TCA. This type of drive is also described in publications [15,16], where it cooperates with an SF₆ interrupter.

The use of vacuum interrupters in hybrid switches in DC circuits isolates the arc from the environment and reduces or eliminates the arc occurrence time. From an operational point of view, hybrid switches eliminate vacuum interrupter erosion and significantly reduce contact erosion, making the device almost maintenance-free. In addition, the cooperation of the vacuum interrupters with the IDD makes it possible to limit short-circuit currents to values not achievable by switches with a magnetic blow-out, which require a significant protection zone above the extinguishing chamber.

This paper focuses on a crucial component of every vacuum hybrid switch, called VCU, which was mentioned before. A theoretical analysis of the interaction of the IDD—the VCU main drive—with the three types of power supply circuit configurations was carried out in order to determine its effect on the performance of the drive in both the response time and displacement of the movable components of the VCU. These parameters will decide the performance of the whole switch, regardless of its topology and current-breaking method. Laboratory verification of the theoretically determined parameters on the two selected VCUs was carried out by measuring these times and displacements of the VCU using a high-speed camera. The investigated VCU rated for 2 kA and 4 kA were used in the Energy Extraction System at CERN [17] and are the first device of its type known to the authors, which were tested, and its mechanical durability is suitable for commercial applications and cooperation with available market vacuum interrupters VIs.

2. The IDD Drive

The IDD is a high-power device that, with great acceleration and while under the influence of a current pulse with a maximum value of several kA and a frequency of (1.2 ÷ 2) kHz, allows the opening of the circuit breaker contacts. In combination with the vacuum interrupter VI, it forms the vacuum contact unit VCU, shown in Figure 1a.

The actuator consists of a coil (9) and a disk (10) attached to a support (11), which is mechanically connected to the drive rod (7) and then to the movable contact (3) of the VI through an insulator (4). The appropriate contact pressure is provided by a set of several springs (12).

The masses of the moving components and the contact pressure value of the VI increase as the switch-rated current increases, making it necessary to adjust the repulsion force F_N developed to ensure the assumed dynamics of the drive.

The VCUs are used in hybrid DC switches, enabling a fast-breaking of short-circuit currents or switching between different circuits. One of the VCUs, shown in Figure 1b, is the main switching component of an energy extraction system EES that operates at CERN, protecting the superconducting magnet systems with a rated current of 2 kA. It is capable of achieving a response time of less than 300 $\mathsf{\mu }$s and an initial velocity of the VI contacts of about 2 m/s. Achieving the required performance of the IDD is possible by selecting the appropriate design parameters of the coil and disk, as well as by changing the configuration of the power supply circuit and adjusting the voltage to which the capacitor is charged.

3. The IDD Power Supply Circuit

The simplest VCU power supply circuit is shown in Figure 2a. It consists of a capacitor C charged to a voltage u_c, which, together with a drive coil C_N and a disk d characterized by the inductance L, forms a circuit with oscillatory properties under certain conditions. When thyristor T is switched, a single pulse of current i_CN(t) appears in the circuit, which is described by Equation (1):

$$\begin{array}{c}\hfill i_{CN}\left(t\right)={\displaystyle \frac{u_c}{{\omega }_{0}L}}{e}^{-\alpha t}sin{\omega }_{0}t\end{array}$$

(1)

where u_c—capacitor voltage, damped natural frequency ${\omega }_0=\sqrt{{\omega }_n^2-{\alpha }^2}$, undamped natural frequency ${\omega }_n={\displaystyle \frac{1}{\sqrt{LC}}}$, damping attenuation $\alpha ={\displaystyle \frac{R}{2L}}$.

From here, the equation for the maximum value of current I_max can be obtained by Formula (2):

$$\begin{array}{c}\hfill I_{max}={\displaystyle \frac{u_c}{\omega L}}{e}^{-{\displaystyle \frac{a}{w}}arctan({\displaystyle \frac{\omega }{\alpha }})}sin[arctan({\displaystyle \frac{\omega }{\alpha }})]\end{array}$$

(2)

The frequency f of the circuit is expressed by Formula (3):

$$\begin{array}{c}\hfill f={\displaystyle \frac{1}{2\pi }}\sqrt{{\displaystyle \frac{1}{(L_C+L_W)C}}-{\left({\displaystyle \frac{R_C+R_W}{2(L_C+L_W)}}\right)}^2}\end{array}$$

(3)

where L_C—IDD inductance, L_W—circuit inductance, C—value of capacitor, R_C—resistance of the C_N, R_W—circuit resistance.

The example of the waveforms of coil current i_CN and voltage on the capacitor u_C, along with the marked response time of the VCU, is shown in the Figure 3.

The current waveform parameters, shown in Figure 3, are entered into theAnsys Maxwell 2019, and defined as excitation, where the maximum value of the force F_N developed by the IDD can be determined. To perform this, it was necessary to create the 2D IDD model using the axis-symmetric properties of the drive, which is shown in Figure 4. The Eddy Current module was used to calculate the F_N.

This module uses an automatically generated adaptive mesh, where entered data allow for the calculation of the magnetic vector potential A and the electric potential $\varphi$ based on Equations (4) and (5):

$$\begin{array}{c}\hfill \nabla \times {\displaystyle \frac{1}{\mu }}(\nabla \times A)=(\sigma +j\omega \epsilon )(-j\omega A-\nabla \varphi )\end{array}$$

(4)

$$\begin{array}{c}\hfill {\int }_{\Omega }(\sigma +j\omega \epsilon )(-j\omega A-\nabla \varphi )d\Omega =I_t\end{array}$$

(5)

where A—vector magnetic potential, $\varphi$—scalar electric potential, $\mu$—magnetic permeability, $\omega$—pulsation, $\sigma$—specific conductivity, $\epsilon$—electrical permeability, I_t—total current flowing in conductors, $\Omega$—conductor cross-section.

The electromagnetic force F_N (6) acting on the disk has two components F_AC and F_DC, that can be calculated from Equations (7) and (8):

$$\begin{array}{c}\hfill F_N=F_{AC}+F_{DC}\end{array}$$

(6)

$$\begin{array}{c}\hfill F_{AC}={\displaystyle \frac{1}{2}}\int Re|\overline{J}\times \overline{B}|dV\end{array}$$

(7)

$$\begin{array}{c}\hfill F_{DC}={\displaystyle \frac{1}{2}}\int Re|\overline{J}\times \overline{B^{\ast }}|dV\end{array}$$

(8)

where J—current density in a conductor, B—magnetic induction.

4. Effect of CNT Circuit Impedance on the Force Generated by IDD

The circuit supplying the drive (Figure 2a) is a high-voltage system, usually of considerable size. However, the wires and connections between components become crucial to achieving the assumed amplitude of the current. They can significantly affect the inductance of the circuit, causing an increase in the wave impedance, and in the result the decrease in the value of the coil current and the force developed by the drive, as well as a decrease in its efficiency.

In order to demonstrate this effect, calculations in Matlab Simulink were carried out for three selected drives, which were equipped with coils with the number of turns equaling 8, 18, and 28. Additionally, three lengths of connections between components were assumed: 1, 2, or 3 m. The obtained results were compared with the calculations of an ideal circuit, in which the lengths of the wires connecting the various components of the circuit will be ignored. The excitation was a current with an amplitude equal to I_max = 5 kA and a frequency of f = 1200 Hz. The obtained graphical results, showing the waveforms of currents i_CN in the coils and voltage u_C on the capacitor, are shown in Figure 5, Figure 6 and Figure 7, while the maximum values of currents together with calculated F_N are summarized in

Table 1. Effect of change in connection length in the CNT configuration of the IDD equipped with a coil C_N, with the number of turns equal to 8, 18, and 28, on the repulsion force F_N.

	l	Imax	f	FN
Coil Turns	[m]	[kA]	[Hz]	[kN]
8	0	5	1200	6.56
	1	4	1034	4.07
	2	3.69	977	3.42
	3	3.42	925	2.91
18	0	5	1200	28.4
	1	4.66	1145	24.4
	2	4.54	1126	23.2
	3	4.41	1106	21.9
28	0	5	1200	60.9
	1	4.82	1170	56.6
	2	4.76	1162	54.9
	3	4.69	1151	53.3

.

The greatest influence of the wires connecting the circuit on the maximum value of the current occurs in circuits with drive coils of several turns. In the case of a drive with an 8 turn coil and 3 m long wires, the current amplitude is equal to 3425 A and the frequency to 925 Hz, which is 68% and 77%, respectively, compared to the ideal case. As a result, the force generated by the drive decreased more than twice, from a value of 6560 N to a value of 2910 N. Even for the most favorable modeled circuit, with wires equal to 1 m in length, the efficiency of the drive decreases significantly as the drive force developed drops by about 40%.

As the number of turns of the drive coil increases, the differences in maximum values of current and frequency compared to an ideal circuit decrease. For a drive coil with 18 turns connected to the circuit with wires of 1m or 3m lengths, the calculated amplitudes of currents and their frequencies are, respectively, 4665 A and 1145 Hz and 4414 A and 1106 Hz. For these conditions, the calculated driving forces with respect to the value of 28,400 N decrease to 86% and 77%, or 24,439 N and 21,881 N.

Therefore, analyzing the effect of the length of the wires connecting the capacitor C, inductor C_N, and thyristor T is important, as these circuits are relatively extensive and bring significant inductance into the calculations in relation to the inductance of the C_N. In practice, drive coils with the number of turns in the range of (13 ÷ 23) are most commonly made.

As can be seen from Figure 3 the circuit supplying the IDD is a low-efficiency circuit, since the excitation presence is very short, and the moment of contact separation of the VI signaling the beginning of movement occurs after the current disappears. In the case of the short impulsive forces, which cause the movement of significant masses, the law of force momentum (9) can be applied:

$$\begin{array}{c}\hfill \overrightarrow{F}\ast \Delta t=m\ast \overrightarrow{\Delta v}\end{array}$$

(9)

This equation shows that a force $\overrightarrow{F}$ acting on a mass m at time $\Delta$t causes a change in the momentum of that mass. An increase in the time $\Delta$t of the force will cause an increase in the velocity $\overrightarrow{\Delta v}$, and, thus, an increase in displacement. Confirmation of this relationship is presented later in the article. It is possible to increase the time $\Delta$t of the current flowing through the circuit by attaching a reverse diode to the drive coil (see Figure 2b) or in parallel to the thyristor (see Figure 2c). With unchanged power circuit conditions, an improvement in the drive performance can be achieved. To illustrate the above relationships, calculations and measurements of current and voltage waveforms were made for several selected circuits.

5. Selected Configurations of Power Supply Circuits for the IDD

5.1. CNT + D1 Configuration

The modified circuit from Figure 2a by attaching a reverse diode D1 in parallel to the drive coil, called CNT + D1, is shown in Figure 2b. The current waveforms in the C_N and diode D1 are shown in Figure 8a while the voltage waveforms on the capacitor C are shown in Figure 8b. The diode D1 does not affect the amplitude of the circuit current but causes its deformation in the descending part from the moment when the EMF of self-induction in the coil exceeds the value—opposite to it—of the voltage on the internal resistance of the drive coil. Commutation of the drive coil current to the diode then takes place, the circuit current decreases to zero with great steepness, while the coil current decreases according to an exponential function with a time constant depending on the inductance L and resistance R of the coil as illustrated in Figure 8a. As a result, the duration of current flow through the drive coil increases, that is, the duration of occurrence of the decreasing drive force proportional to the square of the current flowing through the coil increases. Thus, there will be an increase in the force F_N, resulting in an increase in the displacement of the moving components of the VCU.

In the CNT + D1 circuit, simulations were carried out to determine the effect of the length of the wires connecting the diode D1 to the coil C_N. The influence on the current acquired by the diode and the voltage on the capacitor after overcharging were investigated. A drive coil made of 18 turns was selected, and a wire connection between the coil and the capacitor of 1 m was considered. For this stage of consideration, three connection lengths between the coil and the diode were assumed: 0.75 m, 0.5 m, 0.25 m, and 0 m—the ideal case. The calculation results are shown in Figure 8.

As can be seen, the length of the connection wires in the coil–diode circuit slightly delays the commutation of current to the diode and causes the drive capacitor to overcharge to a value of 132 V, slightly higher than the voltage equal to 40 V in an ideal circuit. Comparing the initial energy stored in the drive capacitor, equal to about 280 J, with the energy after overcharging, equal to about 4 J, it can be seen that virtually all the energy stored in the drive capacitor C dissipates in the circuit, so the effect of the size of the coil–diode drive circuit is practically negligible.

The effect of the length of the coil–diode connection on the waveforms in the power supply circuits of drives with coils, where the number of turns is 8, 18, and 28, was also checked, as shown in Figure 9. In this case, the length of the coil-capacitor connection was assumed to be 1 m, and the coil–diode circuit was assumed to be 0.5 m.

From the presented waveforms, it can be seen that in the case of drive coils with a small number of turns, there is a strong influence, on the rate of commutation, of the wires connecting the diode to the drive coil.

5.2. CNT + D2 Configuration

A second way to increase the duration of the excitation is to attach diode D2 counterclockwise to thyristor T, as shown in Figure 2c. The second half-wave of the capacitor current will flow through the diode, and the full period of the sinusoidal current will flow through the drive coil C_N. Since the force produced by the drive is a quadratic function of the current, the time of its action on the moving masses will increase significantly. However, thanks to the compact design of the thyristor–diode circuit, the problem associated with the wires connecting these components disappears.

6. Experiment

Two selected VCUs for rated currents of 2 kA and 4 kA, differing significantly in the masses of moving components and VI contacts force, shown in Figure 1, were tested. IDDs with identical parameters were installed in them. VCU performance was tested in the circuit shown in Figure 10. The main circuit was connected in one of three ways: CNT, CNT + D1, or CNT + D2. Voltage waveforms on the capacitor and the VI contacts were recorded, as well as the current flowing in the drive coil, making it possible to determine the response time of the VCU. The displacement of movable contact, in the first milliseconds of movement, was checked by the high-speed camera Photron APX.

6.1. VCU for 2 kA

The results of the measurements for all the considered configurations of the power supply circuits, and for the VCU at 2 kA, are shown in Figure 11.

Table 2. Average values of the 2 kA VCU response times.

	Circuit Configuration
	CNT	CNT + D1	CNT + D2
Uc [V]	[$\mathsf{\mu }$s]	[$\mathsf{\mu }$s]	[$\mathsf{\mu }$s]
400	439	440	452
600	300	323	305
800	285	281	277
1000	269	270	269

shows the average values of the response times, which are calculated on the basis of ten successive openings of the VCU, for each of the four values of capacitor voltage u_c, which are regulated within the limits (400 ÷ 1000) V.

The presented oscillograms show the results of measurements with a voltage on the supply capacitor equal to about 1000 V and a current in the drive coil equal to about 4 kA. In all circuit configurations, the measured opening time remains at a constant value equal to about 270 $\mathsf{\mu }$s. Also at lower values of voltage u_c on the capacitor, the opening time increases, but it does not depend on the circuit configuration, and small differences should be ignored.

The opening time of the 2 kA VCU is influenced by the initial value of the coil current, specifically its first half-wave. This current, in interaction with the current in the disk, generated by the magnetic field of the coil, causes the occurrence of the maximum value of the force developed by the drive. The moving elements of the VCU begin to move at a speed v resulting from the maximum value of acceleration a produced by the force developed by the drive. However, the value of the acceleration of the masses after the start of movement will decrease, as the driving force disappears and, in addition, the resisting force increase. It should be assumed that movable VCU masses will also significantly affect the rate of acceleration decrease. To clarify the above relationships, the movement of the masses was measured with a high-speed digital camera at u_C = 1000 V, and the results are shown in Figure 12.

The smallest displacements were obtained for the basic CNT circuit, where the movable components of the VCU moved 1 mm after a time equal to 4 ms. Attaching diode D2 to thyristor T, increased the displacement to 1.5 mm after a time equal to about 4 ms. The CNT + D1 configuration turned out to be the most effective, because the obtained displacement reached 2 mm. This proves the validity of relation (9) linking the increase in coil excitation duration with the increase in displacement, which is guaranteed by more complex types of circuits.

However, the circuit CNT + D2 causes the capacitor C to be overcharged twice with a near-sinusoidal current, the first amplitude of which is about 4 kA and the second about 2 kA. Therefore, it is possible to re-charge the capacitor faster, because the voltage on the capacitor after the drive triggers in this circuit stops at 300 V.

In addition to the advantages of the individual circuits supplying the drive coil, derived from the recorded waveforms, the displacement curves of the moving contact, measured with a high-speed camera, become crucial in their evaluation. In the case of VCU cooperation with hybrid methods of arc quenching, it is required to provide as large a displacement as possible in a short time.

6.2. VCU for 4 kA

To determine the effect of the increase in the VCU moving element masses and VI contact pressure on drive performance, the 4 kA VCU was tested for each circuit. The capacitor was charged to a voltage in the range of (1200 ÷ 1600) V.

Table 3. Average values of the 4 kA VCU response times.

	Circuit Configuration
	CNT	CNT + D1	CNT + D2
Uc [V]	[$\mathsf{\mu }$s]	[$\mathsf{\mu }$s]	[$\mathsf{\mu }$s]
1200	952	777	888
1400	771	692	749
1600	679	677	670

shows the average values of the response time for ten measurements for each type of circuit. It can be observed that at higher contact pressures and moving element weights, the circuit configuration significantly affects the opening time, especially the CNT + D1. However, as the voltage on the capacitor increases, this difference disappears.

From Table 3 it can be seen that in order to achieve the required performance of the IDD, in circuits with large moving masses of the VCU, it is necessary to change the design of the drive, voltage regulation on the capacitor may be insufficient.

7. Conclusions

This article presents the VCU as a component of a switch operating in an EES system, the purpose of which is to protect superconducting circuits. To achieve this, a hybrid switch is used, which allows the current to be evacuated from the main circuit as quickly as possible and dissipate in the parallel resistor. This is possible thanks to the short response time of the IDD, where the advantages of vacuum technology in switching off currents were also introduced. The resulting comparison of the three different circuits used to power the IDD shows the features of each of them. The following specific conclusions are drawn from the study:

• In order to shorten the VCU’s response time, the voltage of capacitor C must be increased, as shown in Table 2 and Table 3. The increase in this voltage will cause an increase in the amplitude of the current i_CN, according to Equation (1). At the same time, a difference was obtained for the response times of the 4 kA VCU for different supply circuit configurations, but this did not exceed 10 %. For a 2kA VCU, no major impact was observed.
• In accordance with Equation (9), increasing the drive coil current flow time by adding a D1 or D2 diode will increase the displacement of the moving VCU components, with the CNT + D1 configuration achieving the best result. The addition of diode D1 to the CNT circuit allows the magnetic energy stored in drive coil C_N to be used more efficiently. In the CNT circuit, this energy is not fully utilized in the IDD but only re-charges capacitor C to a voltage slightly lower than the initial voltage. From the start of the conduction of diode D1, the current in the drive coil decreases aperiodically with a time constant L/R, which depends on the parameters R, L of the drive coil. In addition, the CNT + D1 circuit allows capacitor C to discharge to almost zero and reduces the charging time of the drive capacitor, allowing it to reach readiness for the next operation sooner.
• A drive coil with at least a dozen turns is less sensitive to additional losses due to the higher resistance of the supply circuit connections.
• Efforts should be made to shorten the connections between the components of the IDD power circuit, especially between the drive coil C_N, capacitor C, and thyristor T, as it shapes the first part of the i_CN.
• The above simulation of the influence of the power circuit configuration on the dynamics of the drive operation can provide helpful material for designers of VCU of hybrid DC switches. The exact impact of individual parameters will depend on the specification of the VCU.