0 Introduction
Digital simulation of power systems has many advantages, such as not being limited by the scale and structural complexity of the prototype system, ensuring the safety of the system being studied and tested, and having good economy and convenience. It is attracting more and more attention from power science and technology workers and plays an important role in power system planning and design, equipment research and development, and operator training [1]. As we all know, power system protection devices, especially high-voltage/ultra-high-voltage power grid protection devices, need to have sufficient reliability and be able to adapt to various working conditions of the power system. They should have sufficient sensitivity under any fault type to quickly and accurately cut off the fault, so as to ensure the stability of the power grid and the safety of the equipment. However, for such demanding devices, they only operate for a very short time in the field and are in a non-operating state for a long time. Therefore, there is very little actual fault experience available for reference, and even less opportunity for actual power grid testing. As a result, high-voltage power grid protection device products have always been a high-tech and high-threshold industry. In its development and research, actual power system experience and dynamic model testing have an important position. Compared with traditional dynamic model testing, the all-digital RTDS dynamic model system has many advantages, such as convenient model construction, multiple system models, and repeatable reproduction of similar fault conditions. It provides an extremely advantageous means for the development and research of protection devices, and thus ensures that the devices developed from it have more reliable performance.
This paper introduces the newly developed ultra-high voltage line protection device DF3621 based on our company's RTDS dynamic model system testing. It details the testing conditions under several special operating scenarios in RTDS testing, points out several research difficulties in current ultra-high voltage microprocessor-based line protection, and provides our solutions and test results for reference by peers. Further research is needed on specific issues, and the developed device requires further field trial operation evaluation.
1 RTDS Dynamic Model Testing System
1.1 RTDS Experimental Model [2]
RTDS is a digital dynamic simulation system developed by the Maniloba DC Research Centre (HVDC) in Canada, specifically designed for real-time power system research. The power system component models and simulation algorithms in this system are based on the industry-recognized and widely used EMTP and EMTDC, and its simulation results are consistent with the actual conditions of real-world systems. This system has been adopted in many countries and regions worldwide, and several organizations in my country have also introduced RTDS systems of varying sizes.
The basic component of RTDS is a rack, which is connected to each other via a bus. The number of racks determines the scale of the simulating system. By inputting real-time simulated electrical and digital inputs from RTDS to the protection device under test, and then introducing the output signal of the device into the digital input board of the simulation system, a closed-loop real-time simulation test of the protection device can be achieved.
The key to using RTDS for relay protection product testing lies in establishing a test model system. Referring to the test model system of the China Electric Power Research Institute and considering the development positioning of this device—applied to single-circuit lines at voltage levels of 220kV and above—the following line models were constructed. To more closely approximate actual field conditions and ensure the persuasiveness of test results, all line models adopted distributed parameter models, with a voltage level of 500kV.
(1) Single-power source unloaded long-line model (I)
The infinite power supply is designed to power a 400km single-circuit overhead line, with a short-circuit capacity of 3000MVA. It is primarily used for testing the transient overrun performance of distance protection systems.
(2) Dual-power-supply, dual-circuit long-line model (II)
The simulation system model is shown in Figure 1. The protection devices P1 and P2 under test are installed on the N side and L side of the NL1 line, respectively. The line voltage signal required for protection is provided by a 500kV/0.1kV capacitive voltage transformer, and the line current signal required for protection is provided by a 1250A/1A current transformer.
System N is a regional equivalent system with a short-circuit capacity of 3000MVA. Plant L is equipped with one generator M1 with an equivalent capacity of 2100MW and a transformer B1 with a capacity of 2500MVA. The maximum load capacity is 1000MW, of which motor load accounts for 65% and resistive load accounts for 35%. Each transmission line is equipped with a shunt reactor with a capacity of 150Mvar at both ends. Under normal conditions, the power flow P=1000MW, Q=480Mvar. The main parameters of the transmission lines are: z1=z2=0.0193+j0.2793Ω/km, z0=0.1788+j0.8412Ω/km, c1=0.013μF/km, c0=0.0092μF/km.
(3) Short-line ring network model (III)
The short-line ring network system model is shown in Figure 2. The main line parameters are the same as those in Model II. The protection devices under test are installed on the L and N sides of the NL line, respectively.
Plant M, Plant N, and System L are interconnected via a 500kV short-distance transmission line. Plant M is equipped with one generator M1 with an equivalent capacity of 1050MW, and Plant N is equipped with one generator M2 with an equivalent capacity of 1050MW. Plant N is also connected to a load transformer FB with a capacity of 1200MVA, which can handle a maximum load capacity of 1000MW, of which motor loads account for approximately 65% and resistive loads account for approximately 35%. System L is a regional equivalent system with two operating modes, large and small, corresponding to short-circuit capacities of 3000MVA and 20000MVA, respectively.
Special attention should be paid to the construction of certain component models due to the inherent characteristics of ultra-high voltage (UHV) power grid lines, striving to ensure that the design closely matches reality. This realism directly impacts the performance indicators of protection devices. The use of capacitive voltage transformers (CVTs) results in significant transient characteristics in the secondary signals acquired after a fault, posing a crucial challenge for distance protection in transient overrun tests and outgoing fault tests. Furthermore, due to the large attenuation time constant of UHV power grids, the non-periodic components decay slowly under certain fault conditions, leading to saturation of current transformers (CTs). This is also a critical evaluation criterion for protection devices. Therefore, the characteristics of high-voltage power grids place higher demands on protection devices. For different channel connection methods, we utilized the "programmable logic function" of the SEL protection device for simulation, achieving good experimental results.
1.2 Test Content and Results
To ensure that the developed protection device has a high starting point and meets the various requirements of high-voltage/ultra-high-voltage line protection with excellent performance, we have taken the content of SD286-88 "Technical Conditions for Dynamic Model Testing of Line Relay Protection Products" as the basic requirement. In addition, considering that this standard was formulated more than 10 years ago and some of its contents are no longer suitable for the current situation, we have combined the actual needs of the current system with reference to the enterprise standards of other domestic protection manufacturers to formulate our test content and performance requirements.
Our developed DF3621 line protection device is positioned for high-voltage/ultra-high-voltage power systems. Its basic configuration includes longitudinal distance protection as the primary protection, three-stage distance protection, staged zero-sequence protection, and inverse-time zero-sequence protection as backup, along with a complete set of auxiliary functions. The device employs an advanced and reliable software platform, with a 32-bit CPU+DSP hardware platform, providing a reliable foundation for overall protection performance. Its protection principle is comprehensive and advanced. Building upon the advantages of similar domestic protection products, it incorporates mature research results such as adaptive and pattern recognition technologies, enabling the device to meet basic protection requirements while also providing satisfactory results for faults under special operating conditions.
This RTDS test is a method used in the development phase. Therefore, while fully including all testing items from the China Electric Power Research Institute and some relevant standards from the power grid bureau, we added some test contents: complex faults, ultra-high resistance (greater than 300Ω) grounding faults, grounding faults through transition resistors during oscillations, and severe CT saturation conditions. The basic test results of DF3621 are shown in Table 1.
Other tests, such as PT disconnection, CT disconnection and saturation, manual closing of unloaded lines, and faulty lines, were also conducted. Thousands of tests have shown that the DF3621 high-voltage line protection device meets all requirements, exhibits stable and reliable performance, and provides satisfactory fault response under special operating conditions.
2 Several challenges in high-voltage line protection devices [3-6]
Microcomputer-based protection products in my country have been in practical use for over a decade. While they have integrated all previous protection functions, the actual performance improvement in terms of protection functionality has been limited. In other words, the advantages of microcomputers have not been fully utilized in terms of protection principles. This is especially true given the rapid development of software and hardware technologies, where mature advanced algorithms and intelligent analysis methods can be readily incorporated into protection systems. Of course, reliability is paramount in the development of protection devices, and we have been conducting experimental trials to address the shortcomings of current protection devices, adhering to this principle. We also hope to draw the attention of our peers to these issues.
2.1 State Recognition and Adaptation
Strictly speaking, state recognition and adaptation involve many aspects. Due to space limitations and the focus of this article, only a few key aspects will be discussed. Currently, in domestic protection devices, after the activation element operates, the subsequent fault handling process unfolds based on the activation time. For example, after a sudden change in activation, the fast phase, steady-state phase, oscillation blocking phase, tripping phase, and incomplete phase phase are executed sequentially. Other aspects, such as PTDX and acceleration, are included within this process. This approach is unreasonable. The protection handling scheme we have developed is designed for primary and secondary states outside the protection device itself. Except for simple, single-function faults where the program processing and system state are consistent, in most cases, inconsistencies arise, thus affecting the overall protection performance. Several examples are listed below:
(1) The starting element is set to be very sensitive to ensure that it can operate under various possible faults. Therefore, the operation of the starting element is often not the actual time of the fault. In order to balance the rapid clearing of the fault at the near end and the transient overshoot at the end, the distance protection generally adopts the method of opening the protection range based on the start time. This strategy will inevitably result in the transient overshoot test operation time not being too fast. In addition, for faults that occur after a period of time after start due to small disturbances, transient overshoot may occur.
(2) Currently, microprocessor-based distance protection generally adopts short-time opening, and enters the oscillation blocking processing module after 150ms (based on the start time), opening the protection by adding conditions. It should be said that this scheme focuses on reliability considerations and has positive practical significance for the stability of my country's power grid. However, the oscillation blocking program is run only 150ms after the element is activated by a sudden change. It should be said that this element is too sensitive, which means that under stable system conditions, a small disturbance can cause the protection to delay after activation and then cause a fault to occur.
(3) During the operation of the system in a non-full-phase state, due to the presence of zero-sequence current and other characteristics, the protection device should automatically activate the corresponding protection while deactivating certain protection elements that may malfunction. Therefore, if the non-full-phase state is missed, it will inevitably affect the overall performance of the protection device. Currently, some protection devices determine whether the system is in a non-full-phase state based solely on whether the three-phase circuit breaker at this protection terminal is closed. This is inaccurate and may lead the priority reclosing protection device after a single trip to mistakenly believe that the system has resumed three-phase operation.
The fundamental reason for the above problems is that the protection device failed to fully utilize its powerful memory and analysis functions to accurately identify the system status and use the status identification results to input the corresponding processing modules. This process itself embodies the concept of adaptiveness.
The newly developed DF3621 line protection device changes the previous approach of adjusting the protection range based on the start-up moment. Instead, it uses the waveforms of the current and voltage used by the components (introducing the concept of waveform coefficients) as the benchmark for various aspects such as adaptive selection of digital filters, adjustment of the protection range, and adjustment of the directional component's operating range. Theoretical analysis shows that the aperiodic component and higher harmonic components in the fault transient component are different at different fault times. After adopting adaptive filter adjustment, the possibility of harmonic amplification in certain situations can be avoided by uniformly using the differential Fourier algorithm. After CT saturation, the adaptive short data window ensures the authenticity of the calculated quantities. This processing method, which directly reflects the reliability of the component input quantities, fundamentally solves the problem that signal distortion caused by any reason will not affect the protection performance, and also ensures that the protection can quickly clear the fault within the transient overshoot range under any fault condition.
The occurrence of system oscillation is preceded by warning signs and requires a process, which provides us with the possibility of oscillation state identification. By fully utilizing the strong memory function of microprocessor-based protection and employing state prediction algorithms, we can ensure that the oscillation blocking module is activated in time before the system oscillates, eliminating the practice of assuming the system has oscillated after 150ms of sudden change in initiation. This ensures the protection device's operational performance in the event of a fault after a small disturbance has started but the system has not oscillated, thus improving the overall performance of the protection device. For the identification of incomplete phase states, while monitoring the tripped phase electrical quantities and switching quantities, we should comprehensively utilize the past sensing behavior of each component of this protection system, as well as the changes in the zero-sequence electrical quantity of the system, to identify the incomplete phase state of the primary system.
2.2 Phase selection processing during oscillation [7]
The handling scheme for the oscillation blocking module, from the perspective of ensuring power grid stability, has the basic requirement of preventing maloperation of the protection under oscillation conditions, and secondly, ensuring correct operation of the protection even if a fault occurs during oscillation. Currently, protection devices generally meet the first performance indicator, while the second indicator is usually achieved through protection delay operation and reduced performance specifications. That is, strict requirements are not placed on aspects such as correct phase selection, accurate impedance calculation, and withstand transition resistance during oscillation faults. However, protection technology personnel should hope that the protection performance indicators are not affected under this condition, and this point is a key area for further research.
Currently, phase selection elements in oscillation protection only achieve phase selection through sequence components and calculated impedance after the fault opening conditions are met. This inevitably leads to situations where, under certain fault conditions, the selection can only open when the angle between the potentials at both ends swings within a certain range, causing protection delays; or even preventing opening altogether, resulting in protection failure or non-selective output. The newly developed line protection device DF3621 incorporates a novel phase selection element based on a model recognition method specifically designed for system oscillation states. Its basic principle is to comprehensively utilize the sequence relationships at the protection installation location and the calculated sequence relationships at the fault point to ultimately determine the correct fault type. Because this phase selection element requires highly sensitive opening conditions, it ensures phase selection under various operating conditions, and significantly reduces the dead zone during the oscillation cycle, improving the protection's tripping accuracy and operating time. Extensive RTDS dynamic model tests have proven this point.
2.3 Transitional Failure
For evolving faults that transition between different fault types at the same point, parallel real-time calculations using phase selection and impedance elements are employed. After a fault type change, the element that first enters the section is prioritized for fixing, thus meeting test requirements. For secondary faults, where the interval between two fault transitions is relatively long and the first fault has been successfully cleared, subsequent faults, according to state detection scheduling, will be cleared by the non-full-phase fault handling module. Generally, all performance indicators can be met by the protection device. The challenge lies in the case of repeated forward and reverse faults, where forward and reverse fault points coexist. If these are different phases, for single-channel high-frequency protection, remote protection devices can only perform three-stage tripping without selection, while near-end protection devices aim for correct tripping. For this special fault type, without specialized handling, correct phase selection can only be guaranteed when the fault charge generated by the additional power source within the protection zone is significantly greater than the charge generated by the additional power source outside the zone. This compromises the consistency of protection performance under various power grid structures. To address this, we added a sufficiently sensitive complex fault detection element. This element is used to activate a specialized phase selection element for this operating condition. It comprehensively utilizes the phase selection results of the current sequence component and the voltage sequence component, as well as the historical and current sensing conditions of the protection measurement element, and after unified analysis, finally provides the correct phase selection result.
2.4 High-resistance grounding fault
The ability to withstand transition resistance in grounding protection is a crucial test item for distance protection components. Distance protection components are measurement-based, unlike measuring components. Within the protection range, even with significant measurement errors, the protection component can still operate correctly. Therefore, the ability to withstand transition resistance varies at different points within the protection range. All distance components based on metallic faults will inevitably have weaker transition resistance at the end of the protection range. To address this, we adopted a distance measurement algorithm based on zero-sequence current correction as the measuring component for high-resistance single-phase grounding faults at the end of the protection range. This ensures high transition resistance withstand capability regardless of the location of the fault within the protection range, thus improving the overall performance of the device.
3. Conclusion
Extensive RTDS testing has shown that the performance indicators of the newly developed DF3621 line protection device are quite satisfactory. Of course, it still needs to be tested through long-term field trial operation.
The RTDS (Real-Time Data System) digital dynamic simulation testing system plays a crucial role in the development of protection devices. It changes the previous development model that relied on experience-based device development, simulated dynamic simulation testing, and long-term field operation verification. Fully utilizing RTDS testing as a key tool in protection device development will shorten the development cycle and improve the overall performance of protection devices to varying degrees. Furthermore, during the development process, we have found that, based on a complete understanding of current protection device solutions, we can inherit their advantages. For solutions that are limited by the development era and cannot adequately meet current field requirements, new research findings can be adopted after thorough testing to improve the overall performance indicators of the protection devices. With technological advancements, increasing user demands, and breakthroughs in the research of protection principles, protection device products should also be upgraded or new products developed in a timely manner.
