How To Improve Power Quality

There are many devices and measures to improve power quality. New devices with high-power power electronic devices as core units can be used to effectively suppress or offset various short-term and transient disturbances in the power system, while conventional measures are well suited for steady-state voltage adjustment. Power quality control devices can be divided into the following three categories according to their functions: reactive compensation devices, filters, and unified power quality conditioners (UPQCs) that focus on solving transient power quality problems. In order to make the power quality control device give full play to its design function, it is crucial to adopt accurate and efficient analysis and control methods. First, it is necessary to obtain timely and accurate information about the "source", such as three-phase voltage, three-phase current, neutral current, and neutral-to-ground voltage, and then analyze these source information in real time and quickly to obtain the required control information. The control device uses appropriate control methods to produce corresponding actions based on these control information, and finally obtains the ideal compensation effect.
1. Extraction of disturbance signals
For power quality problems such as voltage fluctuations and flickers, harmonics, and three-phase imbalance, which change relatively slowly and last for a long time, the symmetrical component method and harmonic analysis method are the most commonly used time domain analysis methods. They are characterized by simple mathematical expressions and clear physical concepts. However, the time domain analysis method has a large amount of calculation and takes a long time, and cannot achieve real-time and online control. Therefore, the transformation method must be used to quickly and accurately obtain the required control signal. As the most classic signal processing method, Fourier transform plays an important role in power quality detection. At present, the discrete Fourier transform (DFT) and fast Fourier transform (FFT) of various algorithms have become the basis of spectrum analysis and harmonic analysis.
For power quality disturbances such as voltage drop, voltage rise, instantaneous pulse, and instantaneous voltage interruption, because of its short duration and great randomness in the occurrence time, Fourier transform can no longer meet the requirements, so new signal analysis methods must be used, such as windowed Fourier transform, short-time Fourier transform, and wavelet transform. In addition, combining traditional analysis methods with emerging intelligent methods is also a trend in analyzing power quality problems.
The detection and analysis of harmonic current is another important aspect of power quality analysis. Existing harmonic current detection methods include detection methods based on Fryze power definition, analog bandpass filter detection methods, FFT detection methods based on frequency domain analysis, synchronous determination methods, adaptive detection methods, instantaneous detection methods of distorted current based on instantaneous reactive power theory, etc. In addition, there are time-varying harmonic detection methods based on wavelet transform, harmonic current detection methods based on phase discrimination principle, and harmonic detection methods based on artificial neural networks. Among them, the harmonic current detection method based on the instantaneous reactive power theory proposed by H. Akagi et al. in 1984 has strong real-time performance and has been widely used in active filtering. However, this method ignores the influence of zero-sequence components. When the voltage is distorted, the harmonic current obtained is different from the actual value. The dq0 transformation based on the generalized instantaneous reactive power theory can detect the harmonic current more accurately and in real time.
2. Control strategy
Once the information about the power quality problem is detected and analyzed, an effective control method must be used to eliminate or suppress this information. The control method used is closely related to the type of power quality problem and the control device.
Some traditional devices used for steady-state voltage adjustment, such as shunt capacitors, shunt reactors, transformer taps, etc., are mechanical. They react slowly to power quality problems, have imprecise control, and have limited adjustment capabilities. In the past, manual control methods were generally used. Now, some devices use automatic switching methods. Their control strategies include very simple open-loop control and modern control strategies such as fuzzy control and intelligent control.
There are more control methods for power quality control devices based on power electronics technology and connected to the power system through converters, such as SVG (static var generator), APF (active power filter), DVR (dynamic voltage restorer), DSTATCOM (i.e. parallel DVR), UPQC, etc. The PWM control technology for converters is currently the most commonly used control method. By adjusting the conduction angle ∆ and the modulation pulse width Ｈ, the active or reactive exchange between the energy storage device and the power grid can be controlled in four quadrants, and the harmonics on the AC side can be effectively suppressed. According to the extracted power quality disturbance signal, the trigger signal of the final converter is determined. At present, the control methods that are widely studied and applied are as follows:
a. PID control: This is the most commonly used method in the power system. It has perfect theory, strong robustness, good stability, high steady-state accuracy, and is easy to implement in engineering. Classical PID control uses typical control modules such as proportional, integral, and differential, plus several correction networks, which can improve the dynamic and steady-state performance of the system. However, PID control also has shortcomings such as overshoot response, poor ability to perturb system parameters and load disturbance resistance. Therefore, variable parameter PID control and control methods such as combining PID with variable structure control have emerged.
b. Hysteresis comparison control: At present, the most widely used control method for tracking harmonic currents is hysteresis comparison control. The principle of hysteresis comparison control is to compare the controlled quantity with its given value within a given range to determine the switching timing of the switching element of the power converter. Hysteresis comparison control has the advantages of fast response speed, high control accuracy, easy implementation and no need to understand load characteristics; the main disadvantage is that the switching frequency is not fixed, there is serious phase interference when used in three-phase three-wire system, and the controlled quantity often cannot be effectively controlled when the load is switched. Combining with vector control and other methods can effectively overcome the above disadvantages.
c. Space vector control: The principle of space vector control is to obtain the DC quantity (dq) based on the two-phase rotating coordinate system by Park transformation from the measured AC quantity (abc) based on the three-phase stationary coordinate system, realize decoupling control, and have good steady-state performance and transient performance. Conventional vector control methods require complex sine and inverse tangent function operations, which are generally processed by DSP; in order to shorten the real-time operation time and reduce the requirements for hardware, some simplified algorithms can be used.
d. Deadbeat control: K.P.Gokhale et al. first proposed the inverter deadbeat control method in 1987. Its main idea is to deduce the switch control quantity of the next cycle based on the system state equation and the current state information, and finally achieve the purpose of making the output quantity track the input quantity. The use of deadbeat control can eliminate steady-state errors and end the transition process in the shortest time; however, it also has disadvantages such as poor robustness, large transient response overshoot, strong real-time calculation and high hardware requirements. The use of deadbeat control with disturbance state observer or optimal predictive control technology can greatly improve the performance of deadbeat control.
e. Feedback linearization: The direct feedback linearization (DFL) method converts the original system into a linear system by accurately compensating for the nonlinear factors of the system, which can be controlled by linear control theory.
f. Nonlinear robust control: Considering that SMES (superconducting energy storage device) will be affected by various uncertainties during actual operation, interference can be introduced into the deterministic model of SMES to obtain a nonlinear second-order robust model. For this nonlinear model, the feedback linearization method can be applied to make it globally linearized, and then the control law of all linear systems can be used for control; or the robust control theory can be directly used to design the controller. The most typical representative of the robust control theory based on the optimization of certain performance indicators is the H∞ control theory pioneered by Canadian scholar G. Zames in 1981. This theory has now developed to a relatively mature level and has become a powerful tool for analyzing and designing uncertain systems.
g. Adaptive control: The actual SMES system will inevitably be affected by load disturbances and changes in other environmental factors during operation. It is obviously difficult to achieve satisfactory results by using a conventional controller to adapt to various changes with a set of unchanged controller parameters. The adaptive control method can identify the system model online, and then adjust the controller parameters in time according to the system model and control indicators to achieve high-precision control.
h. Fuzzy logic control: When designing a controller using the "frequency domain method" of classical control theory and the "time domain method" of modern control theory, the precise mathematical model of the controlled object must be known. Although adaptive control and self-correcting control have greatly reduced the requirements for modeling accuracy, they require the use of a large amount of prior data and require online identification of the model. The algorithm is complex and the amount of calculation is large, which limits its scope of application. As an intelligent control method, fuzzy control does not require an accurate mathematical model for the system. By describing the system characteristics fuzzily, the cost of obtaining the dynamic and static characteristics of the system can be greatly reduced. Fuzzy control has strong robustness and is insensitive to external interference, process parameter changes and nonlinear factors. However, fuzzy control has steady-state errors and is prone to small-scale oscillations near the operating point. Other control methods can be combined with fuzzy control, such as variable structure control and artificial neural networks, to improve the performance of fuzzy control.
i. Artificial neural network (ANN): Artificial neural networks have adaptive and self-organizing capabilities, and can learn the nonlinear relationship between input and output based on them without the need for a mathematical model of the system; the fault tolerance and adaptability of ANN can cope with many uncertain factors in the operation of complex systems and improve the anti-interference ability of the system; the inherent parallel structure and parallel processing capabilities of ANN enable it to quickly process large amounts of data in the system.
In short, the surplus and shortage of reactive power is an important factor affecting the deviation of the power supply voltage. Traditional power quality testing methods have limitations. Haiyida Energy Technology has developed the EPDS™ intelligent power distribution system, i.e., the power quality monitoring and improvement system. It has established an online power quality monitoring network covering the entire network, as well as a unified and open monitoring and management platform, which dynamically monitors the power quality level of the power grid, and then transforms the interfering loads that seriously affect the power quality of the power grid, effectively improving the power quality management level. It also uses modern measurement and control technology, data processing and communication technology to achieve management and control of all power distribution and power system facilities at the user end, including power supply lines to terminal power equipment, at an economically reasonable cost, greatly improving the operation and management efficiency of power distribution and power systems and facilities, and reducing operating costs.