1. Introduction With the widespread application of new relay protection devices, especially microprocessor-based protection devices, higher requirements have been placed on testing technology. Microprocessor-based relay protection testing devices, involving advanced technologies such as computer-automated testing, have become indispensable specialized equipment. To this end, the Ministry of Electric Power has issued the "Technical Conditions for Microprocessor-based Relay Protection Testing Devices," providing a basis for selecting testing devices. Some provincial power testing institutes have tested imported and domestically produced relay protection testing devices and published comparative results of various testing devices. This article, from the comprehensive perspective of the principles of intelligent electronic instruments and the requirements for relay protection testing and verification, briefly describes the technical performance and testing methods of current domestic and international microprocessor-based relay protection testing devices. 2. Electronic Circuit Structure Characteristics of the Testing Device The microprocessor-based testing device uses a human-machine interface to set the input current and voltage parameters. A program-controlled signal source guides the electronic circuitry, D/A conversion, and power amplification, outputting a three-phase current and voltage with a certain power to the relay protection device. Simultaneously, it receives feedback information from the relay protection device, makes a series of responses, records them, and achieves automatic testing results. The results are then printed in text and chart form as a report. 2.1 Human-machine dialogue Domestic test devices have Chinese menus, which can be used to enter various special test programs. The program interface is user-friendly and has a help menu. Imported test devices have English interfaces. 2.2 Parameter setting Parameters can be directly input or modified through the screen. Generally, test devices must be adjusted before leaving the factory so that the output current and voltage signals are the set values. Since the test device has sudden changes in current and voltage during automatic testing (such as the bisection asymptotic method), it is generally not monitored by instruments. In subsequent tests, regardless of the device's operating conditions, the set value is assumed to be the output value. There is a big problem here. Once the load impedance is large, the output waveform is clipped or self-excited oscillation occurs or the power amplifier circuit parameters change, causing an error between the output value and the set value, which will cause the relay protection to be mis-set. Several methods for solving these problems in domestic and foreign devices are described below. (1) The Swedish OMICRON CMC156 uses A/D measurement automatic feedback correction. Its advantage is that it can improve accuracy when the power amplifier stage is not yet saturated. If the load impedance is large, the error of automatic feedback correction will increase after the power amplifier stage is saturated. However, the test device can alarm on the computer screen when the error exceeds the standard. (2) Jiangxi Huadong Electric Power Instrument Factory JJC-1H adopts a proprietary distortion discrimination measure, which is equivalent to comparing the signal source and output with the circuit. If the effective value of the output current and voltage is more than 3% different from the computer setting value, an audible alarm will be issued. If it exceeds 5%, the device will be shut down after a delay and the error phase will be indicated to avoid incorrect setting. The JJC-111 test device is equipped with open circuit and short circuit protection for current and voltage circuits. The distortion discrimination measure also strengthens the open circuit and short circuit protection of current and voltage power supply. This is not available in other test devices in China. 2.3 Program control signal source Most test devices currently transmit the instantaneous value data calculated in the microcomputer to the D/A through the data communication line to generate a signal source. Some transmit point by point, and some transmit in batches. This has a great influence on the number of points separated by the test waveform each week. The Beijing Weite MRT-02B and Guangdong Angli 3100D test devices reportedly only have 40 points per cycle for the fundamental frequency, resulting in poor amplitude-frequency characteristics. Harmonic superposition and frequency conversion tests are only qualitative, not quantitative, measurements, leading to errors between the output test values ​​and computer settings in these two projects. The Jiangxi Huadong Electric Power Instrument Factory's JJC-1H test device, with transformerless output, has 180 points per cycle for the fundamental frequency, significantly improving amplitude-frequency characteristics. When superimposing 2nd, 3rd, 5th, and 7th harmonics, the error between its harmonic output values ​​and computer settings does not exceed 3%. Considering the time required for point-by-point transmission and communication handshake, 180 points per cycle is already the limit; otherwise, additional hardware memory would be needed for batch transmission. However, point-by-point transmission also has its advantages, enabling the simulation of long-term fault signals. Foreign test devices guarantee 256 points per cycle at any frequency, producing very smooth waveforms with excellent amplitude-frequency characteristics. At the 20th harmonic, the error between the value and the computer setting is less than 3%, resulting in excellent performance during single-frequency testing. Since batch transmission requires a pre-memory, and the pre-memory capacity is limited, the length of the simulation information is restricted during fault simulation. Foreign testing equipment tends to have smaller capacities; increasing capacity requires adding a power amplifier, increasing size and weight. The application environment must be considered when selecting a system. 2.4 Digital-to-Analog Converter (D/A) D/A converters generally use 12 bits or more. In impedance element testing, the voltage signal may drop very low, reducing the effective number of bits in the D/A converter and affecting the waveform step count. Considering the amplitude margin of superimposed harmonics, a phase voltage of 80V is generally suitable. The zero-sequence voltage 3Uo should generally be designed to be 100V. 2.5 Power Amplifier Stage Currently, there are two designs. One type is output through a transformer. The advantage is that impedance matching is achieved through the transformer, which fully utilizes the capacity of the power amplifier tube. The disadvantage is that it cannot transmit non-periodic components, and the transient response and amplitude-frequency characteristics are not good. The other type is output without a transformer. The transient response is good, and the price of high-voltage, high-current power amplifier tubes is low. The use of output without a transformer is the current trend. There are many key technologies in the design of the power amplifier stage, which are related to the hardware investment of the manufacturer. For the test device, the voltage level power amplifier can generally meet the requirements, while the current level power amplifier has the highest capacity requirement. The different parameters of the power amplifier tubes selected in the power amplifier stage result in different expressions of capacity provided by the manufacturer. The following only compares the capacity of the current power amplifier stage. (1) The parameters of Jiangxi Huadong Electric Power Instrument Factory JJC-1H are: when the maximum output power is 300VA/phase and the current is 30A/phase, the undistorted voltage output to the load is 10V. When the ambient temperature is 20℃, the 30A current can be sustained for 5 minutes. This index is currently very high both domestically and internationally. When the current is 10A/phase, the undistorted voltage output to the load is 15V, which allows for long-term operation. When the current of the fault recorder is checked on-site using JJC-1H, 24 CTs can be connected in series at the same time, while under the same conditions, the Angli 3100-D test device can only connect 6 CTs in series, indicating that the former has a larger load capacity. (2) The parameters of the Swedish CMC156, OMICRON, and CMA156 after power amplifier devices were measured by Jiangsu Provincial Electric Power Research Institute: the maximum output power is 150VA/phase, and when the current is 0-15A/phase, the undistorted voltage output to the load is 10V; when the current is 20A/phase, the allowable voltage output to the load is 6.25V; when the current is 30A/phase, the allowable voltage output to the load is only 3.85V; when the current is 50A/phase, the allowable voltage output to the load is only 1.413V. When the ambient temperature is 20℃ and the voltage is 50A/1.4V, it is allowed to operate for 8 minutes. It must be noted that when the current exceeds 15A, the limitation of the power amplifier stage is due to the heat load. At 30A, the output power is only 115VA, which shows that the load capacity is also very small. (3) The output undistorted voltage of Singapore VENUS330 and FREJA300 can only reach 3.8~3.5V, and the output load capacity is very low. (4) The parameters of Beijing Weite MRT-02B and Guangzhou Angli 3100D are: output current range 0~30A/phase, long-term allowable operating current below 10A/phase. When the output current is greater than 10A/phase, the test time should not exceed 10s, indicating that the heat capacity of the power amplifier stage is insufficient. 2.6 Automatic loading test According to the requirements of the relay protection test items, the test device can output current and voltage according to the set program and automatically find the action boundary. For example, the differential relay magnetic characteristic test requires simultaneous output of power frequency current and DC current, and automatic asymptotic search for the action boundary according to the binary method. If the DC magnetizing current is selected as 20A, the AC operating current will approach 30A, therefore the allowable duration under high current should not be too short. Some devices may experience overheating shutdown during testing under the above conditions, preventing the automatic boundary finding process from completing. 2.7 Relay Protection Feedback Information There are various types of relay protection feedback information, such as dry contact, potential type, and pulse type, and there are also positive and negative polarities. Early test devices had poor intelligent discrimination, requiring the use of changing hardware switch positions to define the processing of feedback information, which was inconvenient. Currently, most domestic devices, like foreign devices, have achieved automatic intelligent discrimination, which is very convenient. This characteristic should also be considered when selecting test devices. 2.8 Test Data Processing Distance protection operating zone test uses an automatic bisection method to asymptotically find the operating boundary, obtain a series of test results, and plot these test points on the screen. Currently, many test devices cannot use intelligent fitting methods to plot the operating zone curve, nor can they obtain the measured operating impedance, thus failing to express how much error there is between the measured impedance and the setting value. Jiangxi Huadong Electric Power Instrument Factory's JJC-1H microcomputer relay protection tester developed the following software: intelligent fitting programs for impedance element operating zone curves, precision current curves, and differential protection magnetizing characteristic curves, etc. Especially when testing the effect of superimposed harmonics on the impedance element operating zone, this program can depict a non-circular operating zone curve and generate a complete test report (see Figures 1 and 2). [IMG=Influence of Harmonics on Impedance Relay Operating Zone]/uploadpic/THESIS/2007/12/2007121917403916386Y.jpg[/IMG] Figure 1 Influence of Harmonics on Impedance Relay Operating Zone [IMG=Differential Relay Ratio Braking Characteristics]/uploadpic/THESIS/2007/12/2007121917404568648G.jpg[/IMG] Figure 2 Differential Relay Ratio Braking Characteristics 3 How to test the performance of the test device in the laboratory Laboratory testing often lacks complex relay protection but utilizes numerous instruments and testing devices. 3.1 A relatively simple method for amplitude-frequency characteristic testing is observation using a dual-line oscilloscope. The test setup is placed in a superimposed harmonic state, with one phase receiving a power frequency voltage U1 and the other a high-frequency voltage Un. The computer is set to have equal amplitudes, i.e., U1 = Un. If the amplitude-frequency characteristic is good, the two waveforms at different frequencies should have the same height on the oscilloscope. Foreign devices, at a high frequency of 1000Hz, show the same amplitude for both waveforms, and the curve remains smooth, indicating a high number of points per cycle in the signal source waveform. Due to the absence of RC smoothing, the amplitude-frequency characteristic is excellent. The Jiangxi Huadong Electric Power Instrument Factory's JJC-1H test device has 180 points per cycle, and at high frequencies below 350Hz, the amplitudes of the two waveforms are the same. The Beijing Weite MRT-02B test device, with only 40 test points per cycle and employing significant RC smoothing measures, suffers from poor amplitude-frequency characteristics. At the 5th harmonic, its amplitude error compared to the computer-set value exceeds 50%. 3.2 Linearity and Accuracy Measurement: Relay protection testing primarily ensures linearity and accuracy at 50Hz, which can be verified using general ammeters and voltmeters. For current accuracy verification, the test device should be tested point-by-point from high current down, which helps prevent overheating in devices with insufficient heat capacity. Due to the limited number of bits in the D/A converter, attention must be paid to accuracy at low voltages; for current, only the range of 0.2–30A needs to be checked. The OMICRON CMC156, employing A/D feedback for automatic accuracy correction, exhibits excellent accuracy and linearity. However, its output capacity is relatively small; often at low voltages and currents, the test device experiences higher distortion due to the limited number of bits in the D/A converter. This has little impact on microprocessor-based protection. Because the number of bits in the AD sampling of microprocessor-based protection decreases under low voltage and low current, the reading accuracy also decreases, and the step-like effect of the output waveform has a significant impact on ranging. Domestic testing devices generally have good accuracy at power frequency, but their accuracy is worse under low voltage due to the reduced number of A/D bits. 3.3 Load Capacity Measurement: When measuring the load capacity of a testing device with an oscilloscope and load resistor under undistorted output waveform conditions, it is necessary to simultaneously indicate the maximum output current and the maximum undistorted voltage of the output waveform. Load capacity testing is a crucial step, requiring an adjustable load resistor value below 1Ω and a current carrying capacity of 30A. Suitable adjustable resistors are generally difficult to find; several reactance transformers connected in series can be used as a substitute. The measurement results are shown in Table 1. When carrying 30A, the undistorted output voltage of the Jiangxi JJC-1H is 10-11V, while the undistorted output voltage of the Singapore VENUS330 is 3.8V. Clearly, the former has a much larger load capacity. 3.4 Harmonic Superposition Capability: A harmonic analyzer is used to test the ability of a test device to superimpose harmonics, simultaneously assessing accuracy and amplitude-frequency characteristics. Foreign devices, due to their superior amplitude-frequency characteristics, exhibit excellent harmonic synthesis. The Jiangxi JJC-1H also demonstrates good harmonic synthesis and rapid response when a step wave is transmitted. For test devices without transformer output, an oscilloscope can be used to observe the device's ability to output square waves and step waves. 3.5 Differential Relay Magnetizing Characteristic Test: This test is a simple method to assess whether the test device can automatically perform tests and how it processes the test results. 3.6 Impedance Relay Operating Zone Test: This test is also a simple method to assess whether the test device can automatically perform tests and how it processes the test results, used to determine its ability to superimpose harmonics or automatically plot fitting curves. 3.7 Impedance Relay Precision Current Test: This test assesses the test device's ability to perform dynamic tests. If the transient response is poor, the results of two precision current measurements will differ significantly, especially under low current conditions. This requires analysis to determine whether the problem stems from transient dispersion in the protection device or the transient characteristics of the test device. At the same time, the results of the precision current fitting curve should be evaluated, especially when the data is scattered, whether the fitting curve will produce unreasonable oscillations, so as to evaluate the quality of the plotting software (Figure 3). Performance comparison of microcomputer-based test devices for relay protection at home and abroad. [IMG=Precision current curve fitting]/uploadpic/THESIS/2007/12/2007121917405133283Z.jpg[/IMG] Figure 3 Precision current curve fitting 4 Review of software menu design The richness of the software is also a selection criterion for the test device. Generally speaking, commonly used software should have the following contents: (1) The initial value and fault value of the three-phase current and voltage can be set arbitrarily; harmonics can be superimposed; the length of the state time zone of each segment can be set; the information of the relay protection feedback contact can be intelligently judged; the critical action value of the relay can be automatically found according to a certain mode. For example, if the current and angle remain unchanged, the voltage amplitude can be changed according to the dichotomy method to find the critical action value of the impedance element. It can also keep the current and voltage constant and let the angle between the voltage and current change continuously to see the angle range of the relay entering the operating area. It can be used to measure the impedance relay with the upward circular characteristic, and can also measure the operating area of ​​the directional relay. (2) The three-phase current and voltage sources can be independently frequency-converted, and can also be changed as needed. For example, when performing the differential protection magnetic characteristic test, one phase sends DC current and the other phase sends AC current. When performing the harmonic braking characteristic of the differential relay, phase A sends harmonic current and phase B sends 50Hz current. The Singapore VENUS 330 does not achieve independent frequency conversion for each phase. (3) The three-phase current and voltage can also be automatically calculated and generated according to the input line parameters R, X and the selected fault type. FREJA300 does this well. It can input R, X; or Z, Φ; or I, Z; or U, I; or U, Φ, etc., and generate fault information accordingly. (4) The harmonics of the three-phase current and voltage cannot be arbitrarily superimposed. Because once the fundamental current and voltage are set, it is equivalent to the line parameters R, X being determined. If the harmonic current is set arbitrarily, then the harmonic voltage must meet the constraints of parameters R and X. (5) The three-phase current and voltage should be able to set multiple time zones for different states, such as normal state, fault state, non-full-phase state, reclosing to fault state, and clearing state, so as to verify the operation of the integrated reclosing. (6) The three-phase current and voltage should be able to simulate complex fault transitions. For example, after 15ms (adjustable) of normal state and single-phase ground fault state, it should switch to two-point ground fault state. This setting is necessary to assess the response mode of microcomputer protection to transitional faults. Because the first 15ms is a single-phase ground fault, and the following is a two-point ground fault, the formula for calculating the fault distance is different for single-phase ground fault and two-point ground fault. The problem faced by microcomputer protection at this time is that the data in one cycle is confused data. Which formula should it use to calculate, wait, or calculate in a mixed manner? Because the new relay protection is complex, its core is not very clear. How does it perform in operation? The microcomputer-type test device for relay protection should shoulder the responsibility of this assessment. (7) The test device should be able to generate three-phase current and voltage models when the system oscillates, for testing the oscillation disconnection device; it should also be able to generate three-phase current and voltage models (including fault first, then oscillation or oscillation first, then fault, etc.) when the system oscillates and is accompanied by a fault. (8) Advanced simulation. The fault simulation data is calculated using the EMTP electromagnetic transient program, converted into the test device's data file in a certain format, and the three-phase simulated current and voltage are sent out by D/A through power amplification. For test devices without isolation transformers that transmit computer information point by point and have more than 40 points per week, it is not difficult to achieve this. Beijing Weite MRT-02 completed the fault simulation work earlier, and then other domestic devices also completed this work. The more points per week, the more detailed the simulation. It can also calculate the short-circuit current and voltage under complex faults, and directly input the calculated current and voltage amplitude and phase angle into the test device's menu program to conduct relay simulation tests. (9) Fault information reproduction. In principle, data recorded by a fault recorder should be reproducible and output through a test setup. However, this process requires data file translation (amplitude calibration). For example, if a domestic fault recorder only records 20 points per week, it needs to be converted to 180 points per week using equidistant interpolation (for the JJC-1H test setup). For test setups that transmit data point by point, interpolation can be performed simultaneously with data transmission to achieve a more refined output. Compressed files require even more careful handling.