Share this

Optimized design of high-power frequency converter

2026-04-06 05:42:56 · · #1

Abstract: This paper proposes a unique optimization scheme to address practical engineering problems encountered in high-power frequency converters. The main optimizations include the application of a multilayer busbar, the addition of an anti-magnetic bias circuit, and improvements to the centralized overcurrent protection and absorption circuits. Practical testing has proven that these improvements significantly enhance the reliability of the high-power frequency converter and greatly improve the overall performance. Keywords: Multilayer busbar, frequency converter, overcurrent protection, anti-magnetic bias circuit 0 Introduction With China's accession to the WTO, factories are importing more and more high-power electrical equipment, and the power supply capacity of production lines is increasing, leading to a growing demand for high-power frequency converters and a very promising market prospect. Currently, high-power frequency converters on the market exhibit various problems during use, such as grid pollution and overall reliability issues, causing significant inconvenience to users. To meet market demands, there is an urgent need to develop high-reliability and environmentally friendly high-power frequency converters. Therefore, our company focused its efforts and fully complied with the requirements of the national military standard. In view of the characteristics of large current in the main circuit power devices of high power frequency converters, large power dissipation of the circuit, serious heat dissipation problems and large pollution to the power grid, we mainly optimized the design of high power frequency converters from the aspects of rectifier circuit, absorption circuit, overcurrent protection and DC bus, and developed a high-reliability 700KVA three-input three-output high power frequency converter. 1 Optimization scheme 1.1 Improvement of rectifier circuit [1] For three-phase input frequency converters, three-phase bridge uncontrolled rectification is generally adopted, and capacitor filtering is adopted on the DC side. The phase of the fundamental component of the input current of this circuit is roughly the same as the phase of the power supply voltage, so the fundamental power factor is close to 1. However, the harmonic component of its input current is very large, which causes serious pollution to the power grid and also makes the total power factor very low. Harmonics cause additional harmonic losses in the components of the public power grid, reduce the efficiency of power generation, transmission and power consumption equipment, and a large number of harmonics flowing through the neutral line will cause the line to overheat or even cause a fire. Harmonics can also interfere with nearby systems, and in severe cases, prevent the systems from functioning properly.

As shown in Figure 1, a 12-pulse rectifier circuit is used. A 30° phase shift forms a series double-junction circuit. By utilizing different connections of the transformer's secondary windings, the phases of the two sets of three-phase AC power supplies are staggered by 30°, causing the rectified output voltage to pulsate 12 times in each AC power supply cycle; hence, it is a 12-pulse rectifier circuit. The transformer's secondary windings are connected in star and delta configurations to form two sets of voltages with a 30° phase difference and equal magnitude, which are then connected to an interconnected rectifier bridge.

That is, the harmonic order of the input current is 12K±1, and its amplitude is inversely proportional to the order and decreases. This greatly reduces the 5th and 7th harmonic currents on the grid side. 1.2 Design of DC bus [2] At present, the power buses of small and medium power frequency converters in China are mainly of the following types: 1) Printed circuit board bus is mainly used for small power frequency converters. The disadvantage is that the current passing through is small. 2) Round copper wire is the most commonly used power bus, which is suitable for medium power frequency converters. The disadvantage is that the parasitic inductance is large. 3) Narrow copper strip with a width of 2 or 3 cm and a thickness of 2 or 3 mm is suitable for medium power frequency converters. The disadvantage is that the parasitic inductance is large. As power increases, the aforementioned power buses become unsuitable, leading to several problems. During the switching process of high-power inverter power supplies, high voltage spikes are generated due to the parasitic inductance on the DC bus between the DC energy storage capacitor and the IGBT device, as well as the inductance of the IGBT module itself. These voltage spikes can cause overheating of the devices and sometimes even lead to IGBT malfunction, exceeding the device's rated safe operating area and causing damage. Therefore, it is essential to limit the voltage spikes generated during switching to within acceptable limits. There are generally two methods to reduce voltage spikes: one is to increase the gate drive resistor to reduce di/dt, but selecting a suitable gate drive resistor is difficult. If the drive resistor is too large, it reduces dv/dt, prolongs the turn-on and turn-off times, and increases switching losses; the other is to reduce the distributed inductance of the DC circuit power bus. Since the above-mentioned power buses all have different drawbacks, a multilayer power bus is adopted. Multilayer power buses are based on electromagnetic field theory. The interconnects are made with a flat cross-section; the thinner and wider the cross-section, the smaller the parasitic inductance. Opposite currents flowing through adjacent conductors cancel each other out, further reducing parasitic inductance. A multilayer power bus consists of thin, wide copper busbars stacked together, with each layer isolated by a high-insulation-strength material. The distance between the busbar poles is relatively uniform to reduce mutual inductance. Each copper busbar is reliably insulated from other layers at the required terminal locations, ensuring that terminals with different potentials are exposed on the same plane, facilitating the connection of all components in the main circuit. Multilayer power buses are used to combine IGBTs, rectifier modules, heat sinks, and capacitors. Connections between the multilayer power bus and components are achieved using different terminals and connectors, resulting in very low contact resistance between the contact surfaces and the busbar. This significantly reduces parasitic inductance, minimizing Ldi/dt overvoltage stress and ensuring optimal device operation. Based on the electromagnetic theory, the following assumptions are made: 1) The length of the stacked power bus is much greater than its width, and the width is much greater than its thickness and the distance between the two positive and negative plates. Therefore, the magnetic induction intensity of the stacked power bus is the same in the length direction. 2) The stacked power bus is a non-ferrous material, and the current flowing through it is I, which is uniformly distributed. Therefore, the current density in the width direction is d = I / b. The inductance of the stacked power bus is: (1) Where L is the inductance, including the inner inductance and the outer inductance of the stacked power bus. l is the length of the stacked power bus. b is the width. w is the thickness. a is the distance between the two positive and negative plates. μr is the permeability of the plate. When (b/2a) approaches ∞, substituting into equation (1) yields (2). If the distance between the two positive and negative plates is close to zero and w is much smaller than b, the inductance L is zero. Therefore, the smaller the distance between the two positive and negative plates, the better. The multilayer power bus uses four layers of plates, from bottom to top: insulating plate 1, copper negative plate, insulating plate 2, and copper positive plate. Figure 2 shows the external dimensions of the four-layer plate, where 1 is the copper negative plate, 2 is the copper positive plate, and 3 is the insulating plates 1 and 2.

Besides reducing parasitic inductance, using a multilayer power bus can also reduce electromagnetic interference in space. According to electromagnetic field theory, when a current flows through a conductor, a magnetic field is generated at any point in space. Let the magnetic field at a point in space be measured by the permeability B of the magnetic field. The permeability in space can be approximated by the permeability μ in vacuum, then B changes proportionally with the current. When two currents of the same magnitude but opposite directions are passed through the multilayer power bus, the B generated at a point in space is the vector sum of the B generated by the two currents in opposite directions. According to Biot-Shahr's law, the magnitude of the combined B at any point in space is: B = where I: current in the plates; l1: distance from the center point of the positive plate to a point in space; l2: distance from the center point of the negative plate to a point in space; θ: angle between the lines connecting the point in space to the center points of the two plates. From the above formula, it can be seen that when the current on the plates in the multilayer power bus is constant, to minimize the electromagnetic interference at a point in space, the distance between the positive and negative plates should be minimized, i.e., the center distance between the positive and negative plates should be minimized. 1.3 Adding a DC bias circuit. With the increasing power of frequency converters, the bias of the main transformer must be considered. The consequences of bias are severe; at best, it increases power consumption and temperature rise in the transformer and power semiconductor modules, and increases mechanical noise in the transformer. At worst, it can damage power devices and prevent the frequency converter from operating normally. Therefore, to improve the reliability of high-power frequency converters, an anti-bias circuit must be added. To solve the DC bias problem in SPWM full-bridge inverters, firstly, power switching transistors with consistent saturation voltage drop and storage time characteristics are selected for the SPWM full-bridge inverter to reduce pulse width distortion and drive delay in the control circuit. Secondly, an air gap is added to the transformer core to increase the core's magnetic reluctance and improve the transformer's resistance to DC bias. Finally, an anti-bias circuit is used. Since the excitation current in the output transformer typically accounts for only 2% of the primary current, the detection of the DC component of the primary current must first filter out the fundamental and high-frequency components in the excitation current before amplifying the remaining DC component for control. The DC component of the excitation current can be extracted by first detecting the primary current of the transformer using a Hall current sensor, then passing it through an active filter, and finally sending it to the PID controller. The block diagram of its dual closed-loop control principle is shown in Figure 3. In practice, negative feedback is introduced into the inverter's output current to limit the DC component in the main circuit, thus preventing the transformer from generating bias. This method of anti-bias circuit adjustment achieves automatic adjustment of DC bias, effectively preventing its generation at various operating points. The advantages of this scheme are that it shares a detection device with overcurrent protection, saving costs; when DC bias occurs, the transformer excitation current increases rapidly exponentially, making it more sensitive than voltage correction methods.

Figure 3. Block diagram of dual closed-loop control principle

1.4 Improvement of the Absorption Circuit Generally, for overvoltage between collector and emitter in a frequency converter, the circuit shown in Figure 4 is used to install a buffer circuit to suppress overvoltage between the collector and emitter. For low power, the combination of Figures 4A and 4B is used; for medium and high power, the combination of Figures 4A and 4C is used. However, as the power of the frequency converter increases further, this combination becomes unsuitable. A brief analysis follows.

Figure 4 Buffer absorption circuit

Figure 5 shows a typical turn-off voltage waveform using a buffer snubber circuit.

Figure 5 Typical turn-off voltage waveform

In Figure 5, the spike ΔV1 in the Vce initial voltage is caused by the parasitic inductance LS of the snubber circuit and the forward recovery of the snubber diode. Its main component depends on the parasitic inductance LS: di/dt is the di/dt at the moment of turn-off or diode recovery. The worst-case di/dt is close to 0.02 A/ns × IC. If the limit of ΔV1 is determined, the maximum allowable inductance of the snubber circuit can be estimated from di/dt. Assuming the IGBT's peak operating current ICM is 800 A, and ΔV1 is limited to 150 V, the worst-case di/dt is approximately:

Calculations show that high-power IGBT circuits require snubber circuits with extremely low inductance. As the IGBT current increases, the distance between the collector (C) and emitter (E) increases, requiring more and more snubber capacitors and resistors. This also makes the PCB board for the snubber circuit very large, and it's difficult to keep the parasitic inductance of the snubber circuit below a few tens of nH. Limiting ΔV1 to 150V becomes challenging, and it may even exceed 500V, causing overvoltage and damaging the IGBT. In summary, for IGBTs exceeding 600A, using a snubber circuit may be less effective than not using one at all. A poorly designed snubber circuit may increase voltage spikes. Therefore, we removed the snubber circuit board and only connected a non-inductive capacitor in parallel between the collector (C) and emitter (E), along with a multilayer power bus. Practice has proven this method to be safe and reliable. 1.5 Improvement of overcurrent protection [3] Figure 6 is a schematic diagram of the new IGBT overcurrent protection method. The new scheme is similar to the previous IGBT overcurrent protection method in that it still adopts the method of combining distributed overcurrent protection and centralized overcurrent protection. The differences are: First, the new scheme uses Hall sensors instead of magnetic current transformers; Second, Hall sensors are used to detect the primary current of the transformer instead of the output current; Third, a fast optocoupler is connected in series in the distributed overcurrent detection channel, and the distributed overcurrent protection channel is used to respond to the requirements of the centralized overcurrent signal. The overcurrent protection circuit inside the drive module is used to implement soft turn-off of the IGBT instead of hard turn-off. The second pin of comparator A is connected to the voltage value converted from the primary current of the transformer. O1 is a fast optocoupler HCPL4504, whose output transistor is connected in series with a fast recovery diode. When the primary current of the transformer does not exceed the set threshold, there is no concentrated overcurrent signal, and the input diode of the optocoupler is in the conducting state. The optocoupler inserted in series in the distributed overcurrent detection channel will not affect the distributed overcurrent protection function. When the primary current of the transformer exceeds the set threshold, a concentrated overcurrent signal is generated. At this time, the input diode of the optocoupler is quickly turned off, and the output transistor of the fast optocoupler is quickly turned off. Pin 1 of M57962 will be floating. If the IGBT is still in the conducting state at this time, the overcurrent protection circuit inside the drive module will activate, implementing soft turn-off protection for the IGBT. In this way, whether it is distributed overcurrent protection or concentrated overcurrent protection, soft turn-off protection can be implemented for the IGBT to prevent damage to the IGBT caused by excessive turn-off voltage.

Figure 6. New IGBT overcurrent protection

2. Actual product image

Figure 7 Phase-shifting transformer Figure 8 Appearance view

3 Conclusion Utilizing the above improvements and strictly adhering to national military standards, we developed three 700KVA high-power frequency converters with three inputs and three outputs. They have been operating safely and without faults for more than six months and have received unanimous praise from users. References [1] Huang Dahua, Bie Lisheng. Selection of structural type of 12-pulse rectifier transformer. Proceedings of the National Symposium on Application of Electronic Components, 2001, p. 1 [2] Zhang Yunzhou, Xu Guoqing, Shen Xianglin. Application of new laminated power bus in EV motor driver. Power Electronics Technology, February 2005, Vol. 39, No. 1, p. 53 [3] Lu Jialin, Su Yanmin, Bai Xiaoqing, Shi Tao. IGBT driving and protection in 30KVA inverter power supply. Power Technology Application, No. 8, 2000, p. 24 [4] Smith, KM Jr; Smedley, KM Engineering design of lossless passive soft switching methods for PWM converters. II. With nonminimum voltage stress circuit cells, Power Electronics, IEEE Transactions on, 2002 (17) 6: 864~873 [5] MITSUBISHI ELECTRIC Mitsubishi Electric IGBT and Smart Power Module Application Manual

Read next

CATDOLL 128CM Hedi

Height: 128cm Weight: 19kg Shoulder Width: 30cm Bust/Waist/Hip: 57/52/63cm Oral Depth: 3-5cm Vaginal Depth: 3-15cm Anal...

Articles 2026-02-22
CATDOLL 132CM Luisa Silicone Doll

CATDOLL 132CM Luisa Silicone Doll

Articles
2026-02-22
CATDOLL Coco 95CM TPE

CATDOLL Coco 95CM TPE

Articles
2026-02-22
CATDOLL Maruko 88CM TPE Doll

CATDOLL Maruko 88CM TPE Doll

Articles
2026-02-22