An effective method to suppress magnetic bias in single-phase and three-phase sine wave inverters-EEWORLD

Collect

Abstract: The causes of DC bias in SPWM inverter power supply are analyzed, the anti-bias methods currently used are compared, and a new method is proposed to suppress DC bias by using the midpoint voltage of each bridge arm as feedback. It can be applied to the anti-bias design of single-phase and three-phase inverter power supply.

Keywords: sine wave pulse width modulation inverter; bias magnetization; suppression

1 Introduction

In SPWM switching converters, the bias of the main transformer can be said to be a common problem. It is just that the degree of performance is different in various applications. The consequences of bias are very serious. At the least, it will increase the power consumption of the transformer and power semiconductor module, increase the temperature rise, and increase the mechanical noise of the transformer (when the switching frequency or modulation frequency is within the audible range). In severe cases, it will damage the power device and make the inverter unable to work normally. Therefore, anti-bias is one of the key issues of switching inverter power supply.

Based on the comparative analysis of the different magnetization processes of PWM and SPWM transformer cores, this paper proposes a new method for SPWM inverter power supply to suppress transformer bias magnetism, that is, to suppress DC bias magnetism by using the midpoint voltage of each bridge arm of the inverter bridge as feedback. It has been successfully applied in 400Hz single-phase and three-phase series variable frequency power supplies, verifying the practicability and reliability of this method.

2 Magnetization process of transformer core and comparison of bias suppression methods

The electromagnetic process of the main transformer core of the switching inverter power supply satisfies the electromagnetic induction law just like the ordinary transformer. For the convenience of analysis, the winding resistance, leakage inductance, transformer distributed capacitance, etc. can be considered to be zero. In this way, the voltage u1 applied to the primary winding of the transformer _is balanced with the winding induced potential. Therefore,

u ₁ = N ₁ = N ₁ Sk _T （1）

Where: B is the magnetic induction intensity of the iron core;

S is the cross-sectional area of the core;

N ₁ is the number of turns of the primary winding;

k _T is the effective coefficient of the core area;

φ is the main magnetic flux of the transformer.

From formula (1), we can get the magnetic induction intensity

B ( t ) = u ₁ d t + B _r (2)

Where: Br is the magnetic induction intensity _in the iron core at t = 0.

_{For the convenience of analysis, formula (2) is written in incremental form and considering that in PWM and SPWM} inverters, u1 is a pulse with constant amplitude, so the magnetic induction increment becomes

Δ B ( t ) = （3）

Therefore, the magnetic induction increment Δ B ( t ) becomes a linear function of time. For the full-bridge PWM inverter circuit, under normal circumstances, the square wave "volt-second" area of the transformer in the forward and reverse directions is equal, and the magnetic induction intensity of the iron core is proportional to the square wave pulse width. The change is shown in Figure 1 (a), and the magnetization curve is symmetrical to the origin. However, the width of each pulse in the SPWM inverter circuit is different and changes with the change of the carrier ratio. The size of Δ B ( t ) is proportional to the SPWM pulse width. Its voltage waveform and the waveform of the magnetic induction intensity in the iron core are shown in Figure 1 (b). At this time, the magnetization curve is symmetrical to the origin in one fundamental wave cycle.

(a) PWM type transformer core magnetic induction intensity

(b) Magnetic induction intensity of SPWM transformer core

Figure 1 Transformer primary voltage and magnetic induction intensity

When the primary side of the transformer contains a DC component, the square wave "volt-second" area in the positive and negative directions of the PWM type conversion circuit is no longer equal, and the magnetic flux will gradually increase in a certain direction, eventually causing the transformer core magnetic induction intensity to exceed the saturation magnetic induction intensity and saturate, and the magnetization curve will no longer be symmetrical to the origin. In the SPWM type conversion circuit, when a DC component is contained, a constant magnetic flux will be generated in the transformer core. As a result, the transformer magnetic flux will no longer be a sine wave with the same positive and negative directions in the fundamental wave period, and its range will change from ± Δφ _1m in normal times to - Δφ _1m + φ _d ~ + Δφ _1m + φ _d . The transformer magnetic induction intensity variation range will also change from - B _1m ~ + B _1m in normal times to B _d - B _1m ~ B _d + B _1m . The magnetization curve will no longer be symmetrical to the origin, resulting in damage to the semiconductor switch tube ^[5] .

Many scientific and technological workers have proposed some methods to reduce magnetic bias based on their own engineering practice and have achieved good results ^[1]~[5] . Some of these methods are only applicable to PWM DC converters ^[1]~[3] . The magnetic bias can be eliminated by correcting the pulse width of each switching cycle, and there is no impact on the output waveform. However, the pulse width of each switching cycle of the SPWM sine wave inverter is not the same. Using this method will cause a serious deviation from the SPWM mode, resulting in modulation distortion, and ultimately causing the output waveform to be distorted. For the SPWM sine wave inverter, reference [4] uses electronic switches to simulate the upper and lower switches of the same bridge arm. This method does not take into account the discreteness of the power switch tube; reference [5] connects a sampling transformer in series with the primary of the main transformer, requiring that the working characteristics of the transformer and the main transformer are completely consistent, and accurately reflects the working state of the main transformer. Due to the discreteness of many factors such as materials and devices, there may be certain difficulties in practice. At the same time, some of the above methods also require the use of high-priced components such as current sensors and sample holders, which will have an adverse effect on cost-sensitive power conversion devices. Moreover, for three-phase inverters, the control circuit will become extremely complicated. Therefore, for sinusoidal wave inverters, it is very meaningful to seek a simple, effective anti-bias magnetic method that is applicable to both single-phase and three-phase inverters.

3 A new method to suppress magnetic bias in sine wave inverter

In general, the asymmetry of positive and negative pulses of the inverter bridge SPWM wave is the fundamental cause of magnetic bias. The specific reasons for the asymmetry of positive and negative pulses of the SPWM wave are:

1) Differences in switching speed of power semiconductor modules (IGBTs) (device discreteness or inconsistency);

2) Differences in on-state voltage drop of power semiconductor devices (IGBT) (same as above);

3) The difference in transmission delays of various signals.

In addition, if the circuit design is improper, the process will also produce bias. In summary, no matter what method the SPWM control signal is generated by [sine wave vs. triangle wave, single chip microcomputer, or application-specific integrated circuit (ASIC)], bias always exists, but the degree is different. In order to make the transformer work in an ideal (or reasonable) state, the better way is to make the positive and negative directions of the transformer magnetization curve symmetrical to the origin O. For the variable frequency power supply in the SPWM working mode, if the positive and negative areas of the output sine wave represented by the SPWM pulse wave output at the midpoint of the inverter bridge (before filtering) are equal, it means that the output waveform does not contain a DC component, that is, the DC component is zero, and the transformer has no bias; conversely, if the positive and negative areas are not equal, the DC component is not zero. The existence of DC voltage on the primary side of the transformer is the fundamental reason for the unequal volt-second areas in the positive and reverse directions, which causes bias. How to detect the DC voltage and correct it through an appropriate circuit is the key to suppressing bias.

3.1 Principle of Bias Elimination

The main circuit of the sinusoidal full-bridge inverter circuit is shown in Figure 2. The cross-phase corresponding switches (S1, S4) and (S2, S3) in the two bridge arms _form two _switch groups _respectively . _The driving signal of the inverter bridge switch tube is the SPWM driving pulse obtained by comparing the sine wave with the triangle wave. Therefore, the voltage u AO between _the first bridge arm of the inverter bridge and the 0 point can be expressed as the sum of the DC component U _ad and the fundamental component and a series of harmonic components; similarly, the voltage u _B0 between the other bridge arm of the inverter bridge and the 0 point can be expressed as the sum of the DC component U _bd and the fundamental component and a series of harmonic components. It can be obtained that the DC voltage U _AB = U _ad - U _bd on the primary side of the transformer , when U _ad = U _bd , U _AB = 0. This coincidence is difficult to occur. Even if the difference is controlled to be zero, it is difficult to satisfy the difference of other phases to be zero for a three-phase inverter. If the output voltage u _A0 , u _B0 , u _C0 of each bridge arm is corrected respectively so that their respective DC components are all zero, even if U _ad , U _bd , U _cd in each phase is zero, then the DC component in the primary winding of the output transformer is eliminated. This is the unified method for eliminating bias magnetism in single-phase and three-phase sinusoidal wave inverters.

Figure 2 Schematic diagram of full-bridge inverter main circuit

3.2 Control block diagram

FIG3 is a control block diagram of an anti-bias magnetic circuit of a bridge arm in an SPWM sine wave inverter. I is an inverter bridge, 0 is a DC midpoint, and _Rp is used to adjust the midpoint. II is a low-pass filter used to detect the DC component U _xd ( x = 1, 2, 3) in the high-voltage SPWM pulse sequence. _{U xd is sent to the PI regulator for error amplification, and its output signal uc is used as a control signal of the pulse width modulator (PWM) to make its output pulse width track the change of U xd. The average value u 0} obtained _thereby is used to correct _the reference _sine of the control circuit to make it symmetrical to the horizontal axis. If the output sine wave of the inverter bridge arm is biased upward, it is adjusted and corrected to make it biased downward. The U _xd of the midpoint of each bridge arm is made zero, thereby achieving the purpose of eliminating the bias magnetic. The whole process is realized in a closed-loop dynamic manner.

Figure 3 Control block diagram of the bias suppression circuit

3.3 Circuit Implementation

The hardware implementation of bias suppression is shown in Figure 4. Low-pass filter II is implemented by RC passive low-pass filter, where the time constant τ = RC . Because it is filtering the DC midpoint 0, a CBB capacitor with a withstand voltage greater than _400V should be selected. The core circuits III and IV can be formed into a push-pull structure using a double-ended output PWM control IC (such as TL494, SG3535, SG3524, etc.). The pulse transformer T is used to isolate and obtain the required voltage gain. u0 is adjustable. _The different u0 will _result in different potentials at point a, and the voltage sent to the adder will also be different. If there is no bias, u0 = ₀ , and u0 is always dynamically closed-loop regulated near zero.

Figure 4 Bias correction circuit

The application method of the three-phase circuit is exactly the same as that of the single-phase full-bridge, except that three integrated chips and three small pulse transformers (Φ26, 2K magnetic tank as magnetic core) are used. In practical applications, the three-way control circuit is made on a small printed circuit board (but isolated from each other). If it is a single-phase sine wave inverter power supply, only two circuits are installed, which has good versatility.

Figure 5 compares the output voltage waveforms of the power supply with and without the anti-bias magnetic circuit in a 400 Hz sine wave inverter. The waveforms are sampled by Tektronix's TDS201 oscilloscope. As can be seen from Figure 5 (a), the transformer has a bias magnetic field, which causes the waveform to be distorted. Figure 5 (b) shows the voltage output waveform of the same main circuit after the anti-bias magnetic circuit is installed. The waveform is a sine wave with THD < 3%.

(a) Output voltage of power supply without anti-bias circuit

(b) Output voltage of power supply with anti-bias magnetic circuit

Figure 5 Comparison of power supply output voltage by anti-bias magnetic circuit

Vertical axis: 50V/division, horizontal axis: 500μs/division

4 Conclusion

By detecting the DC voltage component at the midpoint of each arm of the inverter bridge, an automatically adjustable DC adjustment amount is added to the sinusoidal signal accordingly, thereby effectively solving the DC bias problem of the inverter power supply. This method can be used not only for single-phase full-bridge inverter power supply, but also for three-phase inverter power supply, avoiding the influence of three-phase power supply coupling on the bias. Regardless of the cause of the bias (control circuit or main circuit), it can be effectively corrected. It is a more practical new method to eliminate the bias. After the successful design of this method in the early 1990s, it has been successively used in KZD, TAC, ATO, ATT and other series of products, single-phase and three-phase 400Hz sinusoidal inverter power supplies with power levels ranging from 2kVA to 100kVA. The temperature rise of the transformer is reduced by about 20℃, the total harmonic content is reduced by 3 to 4 percentage points (THD<3%), and the electrical characteristics of the device are improved. Long-term use shows that this method has many advantages such as simple circuit, good effect, low cost, high reliability, etc., and can produce good economic benefits.

Reference address：An effective method to suppress magnetic bias in single-phase and three-phase sine wave inverters

Previous article：Design of a Unipolar Frequency-doubling Voltage-type SPWM Soft-Switching DC/AC Inverter
Next article：DIY production: detailed illustration of the inverter circuit production process

Popular Resources
Popular amplifiers