New Development of Three-phase Uninterruptible Power Supply

introduction

For a long time to come, my country's power grid will continue to be short of power, with large voltage fluctuations and serious interference. The rapid development of various industries and fields has put forward higher and higher requirements for power supply quality, especially the contradiction between the requirements of important systems, important departments and important power-consuming equipment with strong real-time performance and the actual situation of my country's power grid is becoming increasingly acute. Therefore, the uninterruptible power supply (UPS) as a green power supply with stable voltage, frequency and purity has become the focus of more and more people's attention. In order to continuously improve the performance of UPS, researchers have done a lot of research on UPS systems and proposed many circuit topologies and control strategies.

1 UPS circuit topology

The reliable operation of UPS is inseparable from the coordinated work of various modules. The following is a brief analysis of the circuit topology of the main functional modules of UPS.

1.1 Rectification and power factor correction circuit

The rectifier circuit constitutes a DC power supply device in the application. It is the interface circuit between the public power grid and the power electronic device. Its performance will affect the operation of the public power grid and the quality of power consumption. High-performance UPS requires a higher input power factor and minimizes the harmonic components of the input current.

Traditional single-phase UPS mostly adopts analog methods, and three-phase UPS mostly adopts phase-controlled rectifier circuits and voltage-type single-tube rectifier circuits.

1.1.1 Traditional three-phase phase-controlled rectifier circuit and voltage-type single-tube rectifier circuit

The phase-controlled rectifier circuit uses a semi-controlled power device as a switch, which has the following problems:

1) The existence of grid-side harmonic current will reduce the grid-side power factor of the equipment and increase reactive power;

2) Phase-controlled rectification commutation mode causes grid voltage distortion during the commutation period, which not only affects the performance of the circuit itself, but also interferes with the grid and has adverse effects on other equipment in the grid at the same grounding point;

3) The phase-controlled rectification link is a time-delay link and cannot achieve rapid regulation of the output voltage.

The voltage-type single-tube rectifier circuit is the abbreviation of a three-phase uncontrolled rectifier bridge plus a Boost circuit. Its disadvantages are: the current peak is large, which not only hinders the improvement of system power, but also increases conduction loss and switching loss; in order to maintain the improvement of the power factor on the grid side, the Boost circuit must have a certain boost ratio, which will cause the DC output voltage to be too high for the three-phase circuit.

1.1.2 Current-type three-phase bridge rectifier circuit

The current type three-phase bridge rectifier circuit is shown in Figure 1. Its advantage is that the feedback control is simple. There is no need to add current feedback to the control circuit. The input current can be sinusoided by adjusting the duty cycle of each switch tube. The voltage on the DC side is low. The disadvantage is that the input current sinusoidality is not very good. A parallel capacitor must be added on the input side to achieve phase shifting. This circuit is now becoming one of the hot topics of research. This circuit is suitable for high-power rectifier circuits and occasions where the power factor requirement is not high.

Click to see the original image

1.1.3 Voltage-type three-phase bridge rectifier circuit

The voltage-type three-phase bridge rectifier circuit is shown in Figure 2. Its characteristics are that it uses high-frequency PWM rectification technology. The device is in a high-frequency switching state. Since the on and off states of the device can be controlled, the current waveform of the rectifier is controllable. The advantage of this circuit is that it can obtain an input current in phase with the input voltage, that is, the input power factor is 1, and the harmonic content of the input current can be close to zero; energy can flow in both directions. Under normal circumstances, energy flows from the AC side to the DC side. When the DC output voltage is higher than a given value, energy flows from the DC side to the AC side, with a higher conversion efficiency. The disadvantage is that it belongs to a Boost type rectifier circuit, and the DC side voltage requirement is relatively high. This circuit is also a hot topic of research in recent years.

Click to see the original image

1.2 Battery pack and charging and discharging circuit

The battery pack is the energy storage unit of the UPS. When the mains power is normal, it absorbs energy from the mains and stores it in the form of chemical energy. Once the mains power is interrupted, it converts the stored chemical energy into electrical energy to supply power to the inverter and maintain the continuity of the load power supply. In small and medium power UPS systems, the voltage of the battery pack is usually relatively low. Therefore, a charging and discharging circuit that allows energy to flow in both directions is usually used [4]. In order to improve efficiency and simplify the circuit in high-power systems, the battery pack is usually directly connected to the DC bus.

1.3 Inverter circuit

The inverter is the core of UPS, which converts DC power into the AC power with stable voltage and frequency required by users. The following still takes three-phase inverter as the object to analyze the research hotspots of inverter in recent years.

1.3.1 Three-phase half-bridge inverter circuit

The three-phase half-bridge circuit is the most commonly used in three-phase inverter circuits. This circuit is characterized by the use of fully controlled devices to form the inverter, and has the advantages of high power density, good performance, small size and light weight. This circuit is convenient for using new control strategies to improve the quality of the inverter. However, it is difficult to achieve 100% independent load.

1.3.2 H-bridge inverter

For ultra-large capacity inverters, due to the substantial increase in power levels, new requirements are put forward for the structure of the inverter, and the H-bridge arm inverter is one of the options. The output transformer of this inverter adopts a multi-winding connection method, the primary side of the output transformer adopts 3 independent windings, and the inverter output adopts 3 independent H bridges. This is convenient to control, but the cost is relatively high.

1.3.3 Three-phase four-bridge-arm conversion technology

Since the three-leg inverter itself has inherent defects in the three-phase circuit, people began to seek new circuit structures, and thus the three-phase four-leg inverter appeared, as shown in Figure 3. This circuit structure outputs a three-phase four-wire system, the three-phase voltage can be independently controlled, and the control method is flexible, but the algorithm of this topology is relatively complex, and the PWM vector rotates in three-dimensional space. Digital control methods must be used to achieve the generation of spatial PWM waveforms. This circuit has become one of the research hotspots in recent years.

Click to see the original image

1.4 Three-phase UPS circuit

1.4.1 Traditional three-phase UPS circuit structure

The traditional three-phase UPS structure uses thyristor rectification for input and inverter for output. The battery is directly connected to the DC bus, and the rectifier also serves as a charger. The output is isolated by a transformer, which can achieve complete isolation of input and output, ensuring that disturbances in the power grid will not interfere with the load. When the city power is cut off, the battery outputs stable AC power through the inverter; when the inverter fails, the output voltage is bypassed to ensure the reliability of power supply. The main disadvantage of this structure is that it is relatively large in size and weight.

1.4.2 High-frequency chain-type three-phase UPS

In order to reduce the cost, size and weight of UPS, a high-frequency chain three-phase UPS has emerged, as shown in Figure 4. This circuit eliminates the need for a large power frequency transformer, and uses high-frequency rectification for input, which can obtain a higher input power factor and lower input harmonic current. Its disadvantages are that there is no transformer isolation between input and output, and disturbances in the power grid may cause disturbances to the output of the UPS; the output three-phase voltage forms a neutral line by the midpoint of the battery and capacitor, so the positive and negative DC voltage amplitudes must be kept equal during control, otherwise the output neutral line will have a large DC component, which is detrimental to the load and the transformer in the load; the input adopts a three-phase four-wire system, and current flows through the neutral line, which may cause neutral line potential deviation and interfere with the load; the input and output are not isolated, and the circulating current problem in parallel is difficult to solve.

Click to see the original image

1.4.3 New Line-Interactive UPS

Since both UPSs above have to go through two full-power conversions, the efficiency of the system is low. From the perspective of improving system efficiency, a series-parallel compensation type large-capacity structure has emerged, which is a new online interactive structure, as shown in Figure 5. This topology also has no transformer isolation between input and output, so it has the disadvantages of high-frequency chain UPS. The output frequency of this UPS must be kept consistent with the power grid, and the ability to suppress power grid disturbances is not strong, so the power supply quality is worse than that of traditional three-phase UPS. Its characteristic is that the energy from input to output is not converted at full power, and it is also composed of two high-frequency converters, but converter 1 can only bear 20% of the power at most. In terms of cost, this structure is cheaper. In terms of control method, converter 1 is a voltage compensator used to compensate for the distortion of the grid voltage; converter 2 is a current compensator used to compensate for the harmonic current of the load, and it serves as a full-power voltage inverter to supply power to the load when the city power is off.

Click to see the original image

1.4.4 High-frequency link UPS with input and output isolation

Since the input and output of the traditional power frequency UPS are equipped with isolation transformers, the output has good isolation characteristics, and the high-frequency chain UPS has good input characteristics, this high-frequency chain UPS with input and output isolation has appeared as shown in Figure 6. Due to the disadvantages of high-frequency rectification, an autotransformer must be connected on the input side to reduce the voltage, which increases the weight and cost of the whole machine; in addition, since the input uses a high-frequency converter, the efficiency of the whole machine is lower than that of the high-frequency chain and traditional UPS. However, since the input power factor is 1 and there is no harmonic current, the total power consumed is lower than that of the traditional three-phase UPS.

Click to see the original image

1.4.5 UPS with parallel input and output

In this circuit, the input end is composed of multiple rectifiers connected in parallel to supply power to the DC bus. At the same time, the DC bus provides DC voltage to multiple inverters. The output ends of multiple inverters are directly connected to supply power to the load at the same time. This method can increase the capacity of the UPS, increase the reliability of the system, reduce costs, and enhance maintainability. However, the more parallel modules there are, the more difficult it is to solve the current sharing problem between the modules.

2 Uninterruptible Power Supply Control Technology

With the rapid development of control theory and various microcontrollers with rich functions and excellent performance, a variety of discrete control methods have emerged. According to the number of control feedback loops, it can be divided into single-loop, dual-loop and multi-loop control. Increasing the number of feedback loops as much as possible under the conditions allowed by the hardware can improve the control effect. From the control principle, it includes digital PID control, state feedback control, deadbeat control, repetitive control, sliding mode variable structure control, fuzzy control, neural network control, space vector control and other methods.

Digital PID control has good adaptability and strong robustness; the algorithm is simple and clear, and it is easy to implement with a single-chip microcomputer or DSP. However, there are two limitations: on the one hand, the sampling quantization error of the system reduces the control accuracy of the algorithm; on the other hand, the sampling and calculation delays make the controlled system a system with pure time lag, resulting in a reduction in the stable domain of the PID controller and increasing the difficulty of design.

Predictive control can achieve very small output current distortion and strong anti-noise ability, but this algorithm requires accurate load model and circuit parameters, so it has poor robustness, and the delay caused by numerical calculation is also a problem in practical applications. Hysteresis control has fast response speed and high stability, but the switching frequency of hysteresis control is not fixed, which reduces the reliability of circuit operation and deteriorates the spectrum of output voltage, which is not conducive to system performance.

The basic idea of ​​deadbeat control is to calculate the PWM pulse width of the next switching cycle based on the inverter state equation and the output feedback signal. Therefore, theoretically, the output voltage can be very close to the reference voltage in phase and amplitude.

The output voltage error caused by this can be corrected within one switching cycle. However, deadbeat control is a control method based on an accurate mathematical model of the controlled object and has poor robustness.

Sliding control is a nonlinear control, which is characterized by discontinuity. This control can be used for both linear and nonlinear systems. This control method has strong robustness. The disadvantage is that it is difficult to obtain a satisfactory sliding surface.

Repetitive control is a control method based on the internal model principle. The purpose of using repetitive control in the inverter is to eliminate the periodic distortion of the output voltage waveform caused by the rectifier bridge load. The repetitive controller can eliminate the steady-state error caused by periodic interference. However, due to the control characteristic of repetitive control delaying one power frequency cycle, the dynamic characteristics of the UPS inverter using repetitive control alone are extremely poor.

Fuzzy control belongs to the category of intelligent control. The design of fuzzy controller does not require an accurate mathematical model of the controlled object, so it has strong robustness and adaptability. Fuzzy control is similar to traditional PD control, so this control has a fast response speed, but its static characteristics are not satisfactory. Neural network control is a control method that simulates the intelligent activities of the human brain's central nervous system. Neural networks have the advantages of nonlinear mapping ability, parallel computing ability and strong robustness, and have been widely used in the field of control, especially in the field of nonlinear systems. At present, certain results have been achieved in the design of neural network structure and learning algorithm. However, due to the limitations of the hardware system, neural network control cannot currently realize online control of the inverter output voltage waveform. Most applications use offline learning to obtain optimized control laws, and then use the obtained laws to achieve online control.

Harmonic injection PWM technology can basically achieve 100% utilization of DC bus voltage. This method is very effective for open-loop voltage control systems, but it is difficult to apply in closed-loop control systems because the initial phase of harmonic injection must be consistent with the fundamental wave and the initial phase of the voltage fundamental wave cannot be accurately located in voltage instantaneous value control.

Space vector PWM has the advantages of small current distortion, high DC bus voltage utilization and easy digital implementation, so it has been widely used in recent years. This control method also requires an accurate model of the circuit.

The above control schemes have their advantages, but also have their disadvantages. The control scheme that uses different control methods to form a composite control has been widely used in practice and has achieved good results.

3 Problems in the design and application of uninterruptible power supply

APC, a US UPS manufacturer, has summarized and summarized the five problems that UPS power supply systems are currently facing and must solve in the future:

1) Life cycle cost issues;

2) The adaptability and scalability of the uninterruptible power supply system;

3) Improve the availability of uninterruptible power supply;

4) The manageability of the uninterruptible power supply to the power supply system;

5) Serviceability issues.

4 The latest development trend of uninterruptible power supply

The development trend of uninterruptible power supply is to make UPS multi-machine parallel redundant, and use redundant parallel technology to improve the capacity and reliability of UPS; use more functional hardware equipment to achieve full digital control, so that various advanced and complex control algorithms can be used to continuously improve the performance of UPS, that is, develop towards digitalization and high frequency; further intelligence and networking of UPS will make computer networks become uninterruptible networks.

4.1 UPS multi-machine parallel technology to achieve redundancy

UPS parallel technology can bring the following benefits:

1) The capacity of the power system can be flexibly expanded;

2) A parallel redundant system can be formed to improve the reliability of operation:

3) Extremely high system maintainability. When a single power supply fails, it can be easily replaced and repaired by hot-plugging.

The parallel connection technology can form a redundant power supply system with fault tolerance. According to the information currently available, there are mainly the following redundant configuration schemes:

1) Centralized parallel control;

2) Master-slave parallel control;

3) Distributed parallel control;

4) Ring chain parallel control;

5) Wireless parallel control.

Among these parallel connection methods, centralized control is the worst from the perspective of reliability, while wireless control is the best, and has also become a research hotspot in recent years.

4.2 UPS numbers

High frequency

The original UPS used analog control methods with many limitations. With the continuous improvement of the computing speed of digital processors, various advanced digital control methods have been realized, making the design of UPS very flexible, shortening the design cycle, and greatly improving the performance. The high frequency of UPS effectively reduces the size and weight of the device, eliminates the audio noise of the transformer and inductor, and improves the dynamic response capability of the output voltage. Digital control methods have become a research hotspot in the field of AC power supply today. An inevitable development trend is that various methods penetrate each other and combine with each other to form a composite control solution. Digital composite control is a development direction of UPS control.

4.3 Intelligentization and networking of UPS

In order to adapt to the development of computer networks, UPS has begun to be equipped with RS232 interface, RS485 interface, USB interface, SNMP card and MODEM combination, becoming a part of the computer network, with the following excellent intelligent and networked features.

1) Real-time monitoring function It performs real-time high-speed sampling of various analog parameters of the UPS and the switch quantity representing the working status to achieve digital monitoring.

2) Self-diagnosis and self-protection function UPS analyzes and compares various analog parameters and working status data collected in real time and the data of key hardware equipment in the system with normal values ​​to determine whether there are hidden dangers of UPS failure. If there is a fault, an alarm will be given on the display screen of the control panel in a friendly graphical interface and text prompt according to the corresponding fault information level, or an alarm will be given on site and in the control room in the form of indicator lights, alarm beeps, or automatic phone calls, and corresponding protection actions will be taken.

3) Human-machine dialogue control mode Large UPS can provide users with a monitor LCD screen to display the workflow and parameter information in graphics and text. It can provide a visual menu for users to operate. It can also guide users to handle faults in a predetermined way by providing help and continuous prompts, effectively preventing misoperation.

4) Remote control function In the networking era, UPS should not only provide protection for the hardware devices directly powered by it, but also comprehensively protect the running programs and data in the entire network as well as the data transmission path, making it an uninterrupted network. This means that UPS should be equipped with corresponding power monitoring software and SNMP (Simple Network Management Protocol) manager to enable remote management capabilities. Users can perform remote monitoring and data network communication operations between UPS and network platform, making UPS an important part of the network system. In this way, the network administrator monitors multiple UPS through the network management software, and the managed UPS can be in the same LAN or in different LANs, and can even be managed through the Internet and included in the network management system to manage UPS.

As the future network becomes more extensive and globalized, it will inevitably lead to more complex networks, with various forms of network systems connected together. As part of the network system, UPS is required to be able to monitor on various network platforms. Moreover, with the ultra-high-speed development of the Internet, Intranet and e-commerce, users will have higher and higher requirements for network availability, which will extend UPS from protecting key network equipment to protecting the entire network path.