HIGH-SPEED AC VOLTAGE STABILIZATION SYSTEM

V. P. OBRUSNIK, A. V. KOVZEV, T. A. SHADRIN

(Presented at the scientific and technical seminar of the Scientific Research Institute of Automated Electric Drives)

Contactless electromagnetic and magnetic–semiconductor AC voltage regulators-stabilizers are indispensable elements of most modern electronic and electrical-engineering devices. They must meet stringent requirements for the quality of the output supply voltage, reliability, and speed of response. The most widely used actuating elements are thyristor and magnetic regulators. The latter, despite their simplicity, have considerable inertia, dimensions, and weight. The use of thyristor regulators reduces the inertia of actuating elements to a full half-period [1] and reduces their dimensions in comparison with magnetic ones; however, the circuit of the phase-pulse device or pulse-width modulator (PWM) that provides thyristor control becomes considerably more complicated, protection in high-power systems is made difficult, and difficulties arise when such a regulator operates into a transformer.

At present, there is a good prospect [6] for the use of magnetic–semiconductor regulators (MSR), consisting of parallel-connected saturable reactors and a semiconductor switching element (a thyristor or transistor). MSR are analogues of regulators based on antiparallel controlled rectifiers, while possessing greater simplicity and reliability in operation, with somewhat greater weight and dimensions. They are fundamentally inertialess and, like all systems with discrete control, have only an average statistical half-period delay. This makes it possible to use MSR widely for creating high-speed AC voltage regulators-stabilizers.

Inertialess stabilization of AC voltage encounters great difficulties even when fundamentally inertialess regulators are used. In known closed-loop voltage stabilization systems [1, 2], to obtain a feedback signal proportional to the output voltage, the latter is rectified, smoothed by active-capacitive or inductive-capacitive filters, and then fed for comparison with a reference voltage. The integrating element introduces inertia into the control loop, which does not make it possible to realize the fundamental inertialessness of the actuating element. Reducing the time constant of the filter in the feedback circuit leads to an increase in the ripple level at the input of the actuating-element control circuit, which results in malfunctions, reduced accuracy, and reduced reliability of the stabilizer.

A substantial drawback of closed-loop systems is also their unsatisfactory operation in the presence of disturbances from the load side or from the supply-voltage side. Disturbances from the supply-voltage side are usually caused by voltage fluctuations, which give rise to a forced oscillatory process in the system. Elimination of these shortcomings by installing additional filters in the feedback circuit must be regarded as limited in its possibilities, since the dynamic properties of the system deteriorate and the size and weight of the stabilizer increase.

To increase the speed of response of voltage regulators or stabilizers, the integrating element is sometimes removed from the feedback (FB) circuit. For example, in [4] the FB circuit is implemented with a rectifier without an integrator, and the rectifier voltage is compared directly with the reference voltage. This provides control and maintenance of the amplitude of the output voltage, with a sharp deterioration of accuracy indices for rms and average values in nonsinusoidal waveforms.

The most promising are considered to be voltage regulators and stabilizers [3] in which the feedback signal is the instantaneous value of the error signal. Below a method of voltage stabilization is described in which inertia-free FB is used with respect to the volt-second area of the controlled voltage [5]. Its essence is explained using the example of an AC-voltage magnetic transistor regulator (MTR) in Fig. 1, whose operating principle is described in [6].

Fig. 1.

Inertia-free stabilization of the load voltage is achieved here by including in the feedback circuit a storage device (SD) (shown by a dashed line), the voltage at whose input and output is set at a new level once per period. To fulfill this condition, a thyristor switching key T9 is connected in parallel with the input of the SD. The output voltage of the storage device \(U_{\text{out SD}}\) is introduced in series and in opposition to the output voltage \(U_z\) of the trigger, implemented on transistors T4 and T5. T4 is started by the shaper on transistor T3 synchronously with the mains, and transistor T5 synchronously with the PWM.

The difference of the pulse voltages of the reference and FB, also having the form of pulses, is applied to the input of the PWM. The input of the emitter follower

the amplifier T2 is shunted by diode D1, connected in the conducting polarity for the feedback voltage and in the nonconducting polarity for the reference voltage.

The duration of the open state of T1 is determined by the duration of the pulses generated by the PWM. The latter is synchronized with the mains by means of winding \(W_3\) of transformer TR-1 and issues pulses of a definite duration at the beginning of each period. The duration of the pulse at the output of the PWM in each of the periods depends on the magnitude of the error signal, equal to the difference between the reference and feedback voltages during the formation interval of each pulse, and also on the phase at which this signal is applied. The feedback voltage is formed by the reference unit. In this case the charge of capacitor C5 passes through matching transformer TR-2, resistor R6, and diode D3 of that half-wave of the output voltage which corresponds to the open state of transistor T1. The charge of capacitor C5 depends on the ratio of the mains period and the charge time constant. When monitoring the output voltage, it is recommended to choose the charge time constant of capacitor C5, equal to \(\tau = R6 \cdot C5\), from the relations:

\[ \tau_m \leq 0.01 \cdot \frac{1}{f_c} \quad \text{— for monitoring the amplitude value,} \]

\[ \tau \simeq 0.16 \cdot \frac{1}{f_c} \quad \text{— for monitoring the rms value,} \]

\[ \tau_{\mathrm{av}} \simeq 0.195 \frac{1}{f_c} \quad \text{— for monitoring the mean value,} \]

\[ \text{where } f_c \text{ is the mains frequency } f_c = \frac{\omega}{2\pi}. \]

For this purpose resistor R6 is made adjustable, since during adjustment of the stabilizer the time constant can be corrected for the specified conditions.

The charge level of capacitor C5 is maintained until transistor T1 is blocked, when a voltage appears stepwise on the saturating choke Dn, triggering thyristor T9 through winding \(W_y\) and differentiating capacitor C4. Through the circuit of switch T9, capacitor C5 is discharged practically instantaneously to zero. In the following period capacitor C5 is charged to a new level determined by the value of the input voltage \(U_n\) in the monitored period, thereby accomplishing inertia-free monitoring of the output voltage.

The simultaneous dependence of the duration of the PWM pulse on the amplitude and phase of the error signal is used to compensate fluctuations of the mains voltage. This is provided by the T3 shaping circuit and the KS-1 stabilitron, the duration of whose conducting state changes with fluctuations of the mains-voltage amplitude. The duration of the pulses of the shaping circuit and the phase of the trigger’s switching are changed accordingly. Thus, during the interval in which the pulse is formed by the PWM, an almost ideally smoothed voltage is applied to the input, and at all other times it is equal to zero. This sharply increases the noise immunity of the control circuit and the reliability of operation of the stabilizer. The latter becomes inertia-free and, like all systems with discrete control, has an average statistical half-period delay.

According to the described principle, fast-acting single-phase and three-phase AC voltage stabilizers can also be constructed using antiparallel controlled valves both with natural and with forced commutation. In this case, however, one more feedback channel must be used.

References

1. B. N. Ivanchuk et al. Thyristor and magnetic voltage stabilizers, Moscow–Leningrad, “Energiya,” 1968.

2. A. I. Bammass; S. V. Shapiro. Stabilizers with premagnetized transformers. Moscow–Leningrad, “Energiya,” 1965.

3. Bretz Karoly. Circuit of an alternating-voltage stabilizer. Hungarian People’s Republic patent, No. 154298, class 21 d²; H02 j, published 15.06.68.

4. Wanlass Cravens. Semiconductor voltage stabilizer. U.S. patent, No. 3461376, class 323—22; G 05f, published 14.02.66.

5. B. A. Lapin. Method for stabilizing the output voltage of a valve converter with pulse-width modulation. Author’s certificate No. 265255, official bulletin, 1970, No. 10.

6. V. P. Obrusnik; A. V. Koobzev. Theory of processes in discretely controlled ferromagnetic devices not premagnetized by direct current. Izvestiya VUZov, “Electromechanics,” 1972, No. 5.