1. Introduction Generally speaking, an ideal power system should possess electrical energy attributes of a single frequency, a single waveform, and several voltage levels. The highest efficiency mode for electrical energy transmission occurs when voltage and current have the same waveform, frequency, and phase. This is also the optimal form of electrical energy that power product production, transmission, and conversion strive to ensure. Although the phenomenon of sinusoidal waveform distortion existed in previous power systems, its harm was not significant due to its relatively small power. However, with the rapid development of power electronics technology, the widespread application of various power electronic devices has also brought serious pollution to the power system; and modern industrial and commercial electrical equipment has also placed higher demands on power quality. Therefore, how to correctly detect harmonics, thereby suppressing grid harmonics and improving power quality has become one of the hot research topics in electrical engineering today. 2. Detection of Power System Harmonics The problem of power system harmonic detection is one of the urgent problems that electrical engineering needs to solve. Existing harmonic detection methods can be categorized by principle into analog filter methods, detection methods based on traditional power definition, detection methods based on instantaneous reactive power theory, detection methods based on Fourier transform, detection methods based on neural networks, detection methods based on adaptive cancellation principle, and detection methods based on wavelet analysis. 2.1 Commonly Used Harmonic Detection Methods 2.1.1 Analog Filter Method There are two types of analog filters: one uses a filter to remove the fundamental current component to obtain the harmonic current component; the other uses a bandpass filter to obtain the fundamental component, which is then subtracted from the detected current to obtain the harmonic current component. Analog filters are simple in principle and circuit structure, low in cost, and can filter out some inherent frequency harmonics. However, they have disadvantages: large errors, poor real-time performance, especially noticeable when the power grid frequency changes; and they are highly sensitive to circuit component parameters, with detection performance significantly deteriorating when parameters change. 2.1.2 Reactive Power-Based Methods In 1984, Japanese scholars Yasufumi Akagi et al. proposed a theory based on instantaneous reactive power, and on this basis, proposed two methods for detecting harmonic currents: the PG method and the IP-IQ method. These are currently the most widely used methods for detecting harmonic currents in active power filters (APFs), and have greatly promoted the research and development of harmonic and reactive power compensation devices. Both methods can accurately measure the harmonic values ​​of symmetrical three-phase three-wire circuits. In recent years, many scholars at home and abroad have researched and developed the theory of instantaneous reactive power, and proposed a generalized instantaneous reactive power theory. Based on this, harmonic detection methods based on the generalized instantaneous reactive power theory have been proposed and have been initially applied in engineering practice. Like the instantaneous reactive power theory, the generalized instantaneous reactive power theory is relatively convenient for real-time detection of total harmonics, but it does not meet the requirements for detecting individual harmonics. 2.1.3 Methods Based on Fourier Transform In 1822, the French mathematician J. Fourier first proposed and proved the principle of expanding periodic functions into sine series, thus laying the theoretical foundation for Fourier series (FP) and Fourier transform (FT). These two were later collectively referred to as Fourier analysis (FA). After the American mathematicians Cooly and Tukey proposed the Fast Fourier Transform (FFT) in 1965, FA truly moved from theory to practice, becoming a widely used mathematical tool. FFT is currently the most widely used harmonic detection method, with countless related research papers. Currently, most steady-state harmonic detection in power systems uses the Fast Fourier Transform and its improved algorithms. 2.1.4 Methods Based on Intelligent Control In recent years, with the continuous maturation of intelligent control, intelligent control methods have made significant progress in harmonic detection and analysis. Intelligent control methods, such as neural network methods, genetic algorithms, and fuzzy control algorithms, have achieved fruitful theoretical results in improving computing power, approximating arbitrary continuous functions, learning theory, and stability analysis of dynamic networks. They have also been applied in many fields, such as pattern recognition and image processing, control and optimization, prediction and management, and communication. Among them, the application of neural networks to harmonic detection in power systems is still in its initial stage. It mainly has three applications: (1) harmonic source identification; (2) power system harmonic prediction; and (3) harmonic detection. Applying neural networks to harmonic detection mainly involves network construction, sample determination, and algorithm selection. 2.2 The principle and application of Fourier transform method As mentioned above, there are many types of harmonic detection methods currently in use. In comparison, with the continuous improvement of computer hardware, software technology, and software functions and capabilities, a large amount of computation is no longer a major problem for Fourier transform method. Therefore, we may be able to abandon complex algorithms and use computer hardware and software combined systems to realize ideas and practices that were previously considered unsuitable. For example, with the development of virtual instrument technology, we can use virtual instrument software to write programs to complete the required functions. At the same time, using the functions provided by the software package, we can more easily implement functions and calculation steps that were previously considered very complex. Harmonic measurement based on Fourier transform is currently the most widely used method. It is based on the fundamental principle of transitioning from discrete Fourier transform to fast Fourier transform. This method calculates the harmonic order, amplitude, and phase coefficient of the current based on the current or voltage value acquired over one cycle. The harmonic components to be canceled are then passed through a Fourier transform to obtain the required error signal. Finally, an inverse Fourier transform is performed on this error to obtain the compensation signal. 2.2.1 Fourier Series Representation of Harmonies The core of the Fourier transform method is to use Fourier series to analyze and process the signal in order to obtain the time-frequency response diagram of the signal. Generally, any periodic waveform can be expanded into a Fourier series, as shown in the following formula: For the measured current and voltage waveforms, the distorted waveform can be expanded into a Fourier series, which can be expressed by the following formula: Where Ih represents the peak current of the h-th harmonic; Vh represents the peak voltage of the h-th harmonic; φh represents the phase of the h-th harmonic current; θh represents the phase of the h-th harmonic voltage; and ω0 represents the fundamental angular frequency. 2.2.2 Voltage and Current Distortion Factors The voltage distortion factor VDF, also known as the total harmonic distortion rate (THDV), is defined as: Similarly, the current distortion factor (CDF), also known as the current harmonic distortion rate (THDI), is defined as: 3 Virtual Instrument Technology A virtual instrument consists of a computer, instrument auxiliary circuits, and corresponding software. It is a product of the combination of computer technology and instrument technology. In instrument design, a dedicated virtual instrument development platform is usually used to build a graphical virtual instrument panel to complete the functions of instrument control, data acquisition, data analysis, and display. In virtual instruments, the instrument hardware only plays the role of signal conditioning, input and output, while the software is the core of the entire instrument. 3.1 Comparison with traditional instruments Virtual instruments are a brand-new instrument design concept formed by the application of computer technology in the field of instrumentation. Compared with traditional instruments, they show many advantages: (1) "test integration" and virtual instrument library; (2) users have higher participation; (3) the achievements of computer technology can be fully utilized. The above advantages of virtual instruments can greatly reduce the difficulty and cost of establishing a test laboratory; it can also speed up the system development speed; because virtual instruments have rich external interfaces. Excellent computer software achievements (such as the data processing function of MATLAB) can be fully utilized to build a powerful test system. Moreover, expert design ideas can be integrated into the design, so that even inexperienced instrument designers can complete the design of complex test systems in a short time. 3.2 Common virtual instrument systems The hardware components of a typical virtual instrument test system include the tested component, sensor component, signal conditioning and signal acquisition component (such as external or internal data acquisition card, image acquisition card and camera and conventional instruments used for auxiliary measurement and capable of computer communication, etc.) and general computer. The system software is typically written in a dedicated virtual instrument development language (such as LabVIEW). Figure 1 shows a block diagram of a general virtual instrument system implementation scheme. 4. Building a Power System Harmonic Analysis System Using Virtual Instruments LabVIEW is an excellent virtual instrument development platform. Based on a graphical language, its programming interface is intuitive and provides various control and display elements such as knobs and waveforms to create the front panel of a virtual instrument. It uses icons and connections to write programs, rather than the text statements of traditional programming languages, which is significant for developing measurement and testing systems. LabVIEW provides many functions for classic signal analysis and processing in the form of icons, such as digital filtering, window functions, correlation analysis, and spectrum analysis; it also provides various mathematical tools such as differentiation, integration, Fourier transform, and Laplace transform. Furthermore, it provides signal processing toolkits such as joint time-frequency analysis and wavelet transform. Technicians can easily build high-performance measurement instruments simply by accessing this platform and calling the control icons. 4.1 System Hardware Composition The overall design block diagram of the system is shown in Figure 2. The signal conversion circuit in the figure converts the measured voltage and current into an electrical signal from -5 V to +5 V. The signal conversion circuit also includes anti-interference and filtering circuits. The signal conditioning module uses Advantech's ADAM5017 module, and the acquisition card uses the NI6023E board from National Instruments (NI). This board has two input modes: 16-channel single-precision and 8-channel analog differential input. Its main features include: 16/8-channel analog input; effective resolution of 12 bits; channels: six differential and two single-ended; input types: mV, V, mA; input range: ±500mV, ±50mV, ±5V, ±10V, (4～20)mA, ±20mA; isolation voltage: 3000 VDC; fault and overvoltage protection: withstands ±11V voltage; sampling rate: 200kHz (total); input impedance: voltage input impedance 100GΩ, current input impedance 125Ω; accuracy: ±0.1% or better; power supply requirements: (+1～+20)VDC, etc. 4.2 System Software Structure The harmonic testing system software is a unified whole, but it can be roughly divided into three parts: data acquisition module, graphical interface display program, and harmonic calculation program. The general structure diagram is shown in Figure 3. According to the system design purpose and combined with the above program module structure diagram, the system can complete the following functions: (1) It can complete the acquisition, display and record of voltage and current waveform data; (2) It can simulate voltage and current waveform data and adjust the waveform on the interface; (3) It can calculate harmonics and generate harmonic frequency-amplitude curves; (4) It can calculate the THD value of voltage and current. According to the above requirements, a series of virtual instrument program flowcharts were written, among which the total harmonic distortion rate program module is shown in Figure 4. 4.3 Analysis Results Through the construction of the above hardware and software system, the acquisition of current and voltage signals, the monitoring and analysis of harmonic distortion phenomena, especially the use of virtual instrument software components and functions to complete the system, it was found that using the functions in LabVIEW software to analyze harmonic problems has many advantages. As shown in Figure 5, the graphical interface clearly displays the proportion of each harmonic component and the impact of each harmonic on the total harmonic distortion (THD). The THD can also be directly read. 5. Conclusion This paper uses virtual instrument software to build a power system harmonic detection and analysis system. The resulting program and graphs show that the expected functions have been largely achieved. This system also performs some work on calculating the THD, and different harmonic distortion rates can be obtained by adjusting the frequency and amplitude of the analog signal. Although these functions have been implemented, as analyzed above, to obtain more accurate harmonic calculation values, an improved windowed Fourier transform algorithm can be used. Window functions are also a strength of LabVIEW. In future work, we plan to further compare the usage of window functions based on the existing model and apply them to the calculation of the THD, hoping to obtain more satisfactory results.