Application of Power Automation Equipment Adaptive Technology in Power System Relay Protection Chen Wei (Department of Electric Power, Sichuan University, Chengdu 610065, China). The use of communication networks to more flexibly acquire and apply fault information, and the application of artificial intelligence technology to fault information processing relay protection devices is expected to obtain stronger adaptive capabilities, thereby significantly improving its operational performance.
EleetroniePublishi is often associated with the unbalanced output of the action criteria. In recent years, artificial intelligence technologies, such as neural networks and fuzzy logic, have been applied in various fields of power systems, and research in the field of relay protection has begun. Making full use of artificial intelligence technology to adopt appropriate communication networks to obtain more fault information, relay protection devices are expected to obtain stronger adaptive capabilities, thereby significantly improving their performance.
1Adaptive technology in relay protection Adaptive relay protection is a research project proposed in the 1980s. A new scheme is proposed: on the microcomputer protection platform, the auxiliary gate of the knife gate is mainly based on the load current of each unit. The calculation verifies the correctness of the auxiliary contact of the knife gate and automatically corrects its error. Each pair of knife gates still introduces a pair of contacts. The instantaneous value of the current is calculated by the microcomputer in real time. According to the current criterion, the steady state and the transient judgment are combined to find and correct the error of the auxiliary contact in real time, so as to realize the adaptive operation of the busbar. New program to improve the correct rate of mother-to-child protection action.
1.1.2 Realization of adaptive functions using real-time information of protected components Another type of information is information directly related to protected electrical components. Most of the traditional relay protection devices use the real-time fault information obtained by the protection installation to determine the fault. For its reasonable use, better adaptive action characteristics can also be obtained.
Usually, the current quick-break protection action value is determined by avoiding the three-phase short-circuit fault current at the end of the line. When a two-phase short circuit occurs, the sensitivity of the protection action is greatly reduced. After the adaptive technology is adopted, when the fault occurs, the protection first discriminates the system operation mode and the fault type, and then adaptively adjusts the current protection action value according to different fault types, thereby greatly improving the sensitivity of the protection action. A basic method for implementing adaptive current fast-break protection is proposed. The setting value of the current quick-break protection should be changed according to the actual operation of the power system and the type of short-circuit fault. In order to realize the fixed value self-adaptation of current quick-break, it is necessary to determine the type of short-circuit fault and the equivalent impedance of the system in real time. It is necessary to consider the adaptive processing of short-circuit fault type and system operation mode separately.
In the adaptive overcurrent protection, the protection action value is set to a fixed threshold value and an adaptive floating threshold value. After the adaptive floating threshold and the adaptive floating threshold, the protection has higher motion sensitivity in different types of faults. The protection startup method of the fault component is discussed. In normal operation, the fault component will also have an unbalanced output. Taking the maximum value of the unbalanced output of the previous cycle as the adaptive floating threshold of the fault start criterion, the fixed threshold can be greatly reduced. For example, taking 10% of the rated current of the line can greatly improve the sensitivity of the protection start. .
Direct use of the generator terminal and the neutral point third harmonic voltage can reflect the ground fault near the neutral point of the generator; but the grounding protection of the third harmonic voltage is related to the generator working condition during normal operation. There is a phenomenon that the protection sensitivity is insufficient and the incorrect operation rate is high. An adaptive third-harmonic voltage generator stator grounding protection principle is proposed to protect the third harmonic in real time. The change of the normal operation mode of the generator and the change of the third harmonic voltage and its ratio caused by the system oscillation are much slower than the change of the generator stator ground fault. The adaptive microcomputer protection can automatically track this slow change. , thus greatly improving the protection sensitivity.
1.1.3 Using substation comprehensive information to realize adaptive function The development of substation automation system provides conditions for realizing the comprehensive utilization of real-time information of the entire substation. In addition to the switch information and the real-time current and voltage information on each electrical component, this type of information includes action information for each electrical component protection. A comprehensive analysis of the action behaviors of the relay protection devices with different principles of the same component can maximize the advantages of the relay protection devices of different principles and obtain the comprehensive optimal operating characteristics of the protection. A comprehensive analysis of the action behavior of the relay protection devices of adjacent electrical components may also achieve a more reliable blocking of the non-faulty component relay protection devices. On the other hand, the comprehensive use of substation fault information, plus the redundancy of fault information, also improves the reliability of computer relay protection. In order to realize the comprehensive utilization of real-time information of the entire substation, fast fault information exchange technology must be adopted.
Some domestic units have begun research on distributed busbar protection. Distributed bus protection uses network technology to share fault data. In the case where the busbar is connected to 20 outgoing lines, the minimum data transmission rate of the channel is estimated to be 3.52 Mbit/s. For this purpose, an IEEE-88 standard bus is used to form a communication network. At present, the maximum data transmission rate of IEEE-88 bus can reach 8Mbit/s, which can fully meet the requirements of distributed bus protection for data transmission rate. With the gradual application of these research results, adaptive protection can fully realize its fault information integration.
The idea of ​​using directional components of substation line protection to realize directional busbar protection has long proposed a decentralized line treatment based on line-locked directional longitudinal protection. The electrical components connected on the busbar can be regarded as multi-terminal lines. One end. Since the busbar protection does not have to consider the phase selection trip, if the busbar is in an open/closed station, that is, only the line is connected on the busbar, the blocking line protection principle is adopted, and the components of each line reflecting the out-of-zone fault and the fault in the zone are combined. The busbar protection can be constructed. When the line fails, the momentary contact of the faulty line segment sends a blocking signal, and the busbar protection cannot trip; when the busbar fails, the component that reflects the fault in the zone acts, and there is no blocking signal, and the busbar protection operates correctly. Since it is based on the perfect protection of the microcomputer line, the components in the forward and reverse directions should be able to instantaneously reflect various types of faults. In this direction, the busbar protection of the distributed arrangement is small, and the amount of information exchanged between the interval units is small. Simple and reliable communication network implementation. Making full use of the existing microcomputer line protection can be doubled with the double protection of line protection. The busbar protection is based on the mature microcomputer line protection. The voltage blocking amount can be dispersed and arranged, and its reliability is high. At the same time, the busbar protection will no longer be affected by the TA saturation.
1.1.4 Using Remote Information to Realize Adaptive Functions With the development of power grid dispatching automation, various communication methods can be used to obtain useful other information from adjacent substations and dispatch centers.
Of course, the use of communication to obtain remote information requires a complete means of real-time communication. The basic requirements for information transmission are mainly its real-time and reliability. Only real-time data transmission can complete the real-time adaptive function of relay protection; and reliable data transmission is the basic premise of achieving highly reliable adaptive relay protection. It is expected that digital communication will play an important role in this. For example, in some complicated and difficult-to-determine fault states, fault information of adjacent substations can be utilized to achieve more reliable adaptive relay protection. With the development of communication technology, the new protection of substation fault data sharing using high-speed data channels has begun to be researched and tested.
A new scheme for adaptive wide-area current differential backup protection based on high-speed data channels based on service digital network. With this technology, the access speed between the user and the broadband integrated service digital network will reach 155 Mbit/s, which can easily realize the sharing of substation fault data. In the case of a failure of the primary protection, the wide-area communication network is used to obtain the necessary fault current information, and the fast action of the adaptive wide-area current differential backup protection can be realized.
Layout busbar protection. The basic principle is that it is more important to consider the busbar as a double-ended ranging of one super-ishl electric line. The bookmark1 method is applied to the double-ended ranging of the transmission line. It takes the length of the line as the horizontal distance between the towers, and the influence of various factors on the line parameters is uniformly classified into the changes of the line parameters. With the double-ended or multi-terminal communication tools, the adaptive semi-real-time online estimation of the line parameters to improve the double-end The accuracy of ranging. In double-ended ranging, the use of existing hardware devices can effectively reduce the impact of line parameter changes on ranging accuracy. The method is applicable to various overhead transmission lines, especially to areas with complex terrain and harsh climates (such as cold and icy ice). The original split-phase current differential protection lacks flexibility in remote communication, and the proposed communication-specific and multiplex adaptive methods overcome this deficiency. Most of the existing differential protections use the numerical correction method to achieve synchronization, which is based on the assumption that the transmission delays are equal in two directions. This is a large error in multiplexed systems that are applied to certain elastic loads; protection is highly dependent on GPS only when using Global Positioning System (GPS) for synchronization. In this regard, an ideal solution is to use the adaptive synchronization adjustment method based on the GPS clock synchronization method and the numerical correction method. The current differential uses the instantaneous value and the instantaneous component of the fault component to synthesize the differential principle, and the artificial neural network (ANN) is used to automatically identify the TA saturation. When it is determined that the TA is saturated, it automatically switches to the instantaneous value differential mode, and the braking coefficient is adaptively changed to improve the stability of the protection, that is, the ability to determine the current transformer state in real time and adaptively change the protection criterion.
The scheme of implementing adaptive cooperative protection system by using local and remote substation fault information is composed of many independent intelligent agent protection subsystems. The intelligent agent protection subsystems communicate and coordinate with each other to adaptively reconstruct the protection system in the case of system operation changes and individual protection device failures.
The implementation of the protection system. According to the inherent distribution of power system, a distributed computer system is constructed by using the power dispatching system computer. The distributed adaptive relay protection subsystem is constructed by using the station computer and relay protection device, and they are combined to form a large power grid. Distributed adaptive relay protection system. By knowing the mode of the operation mode, in the power system disturbance domain, the protection value of the distributed adaptive relay protection subsystem is off-line calculation, distributed storage, online operation according to the operation mode, thereby changing the protection in real time. value.
1.2 Using different information processing methods to achieve adaptive functions 1.21 Using fuzzy logic to achieve adaptive functions Introducing fuzzy set theory into relay protection has opened up a new path for the development of relay protection. Fuzzy theory (fuzzytheory) is a multi-criteria algorithm based on fuzzy set theory. It is a fuzzy technique that blurs classical set theory and introduces fuzzy logic with linguistic variables and approximate reasoning. It assigns a variety of action criteria to the traditional digital transformer protection, and assigns a certain membership function to each other through fuzzy processing. Then, according to the actual measured value of the fault information, the membership degree of various action criteria is determined by fuzzy operation by fuzzy decision system. Adaptively determines the operating conditions of the transformer and issues an action command. Its sensitivity, selectivity and reliability are improved compared to conventional transformer protection. Transformer fault current and method of magnetizing inrush current. The fault diagnosis of the transformer is realized by comparing the membership function of the actual sampling current symmetry with the theoretical analysis to calculate the approximation degree of the membership function of the current symmetry. The transformer fault is judged when the fuzzy closeness is greater than a certain value.
A new principle of fuzzy fault phase selection for frequency components. The principle uses the high-frequency component of the voltage fault to extract the three-phase voltage, the frequency-domain characteristics of the mode-changing with different phases as the reference, and the fuzzy set and the frequency domain feature are processed to realize the fault phase selection, which has the characteristics of ultra-high speed. The fault phase selection is not affected by the transition resistance and the initial angle of the fault, and the fault phase can also be selected for the transition fault.
Automatic reclosing adaptive optimization criterion method to improve the success rate of reclosing. In this paper, the ratio of capacitive coupling voltage to mutual inductance voltage is taken as the first input variable of the fuzzy controller. The ratio of the fault terminal voltage to the mutual inductance voltage is taken as the second input variable of the fuzzy controller, and the trip signal is the output of the fuzzy controller. This method comprehensively utilizes fault boundary condition information such as capacitive coupling voltage, mutual inductance voltage and fault terminal voltage, and adaptively corrects the original voltage criterion by using fuzzy controller. Theoretical analysis and dynamic model test results show that this method has a good application prospect.
Different feature analysis, starting from the fault characteristics, uses fuzzy set theory to identify the short circuit that occurs during the oscillation process. The operation of the power system is very complicated, and the conditions of oscillation and short circuit are various. In this case, it is limited to distinguish between oscillation and short circuit by a certain precise and absolute value according to a single criterion. . The fuzzy fuzzy model is established by using the fuzzy pattern recognition principle, which can quickly and accurately identify the three-phase short circuit occurring at the oscillation center when the power angle of the 2-machine system is 180* in many cases that are not conducive to identification. . The corresponding fuzzy mathematical model is given in the paper, and a lot of simulation experiments are carried out to obtain good simulation results.
1.22 Using Neural Networks to Implement Adaptive Functions 1 A method for identifying permanent faults and transient faults based on artificial neural networks is proposed. In the case of transient faults, the arc-extinguishing time can be determined to achieve adaptive single-phase reclosing. The transient fault is different from the waveform of the permanent fault voltage and can be used as the basis for discriminating the fault type. However, the actual fault voltage waveform and amplitude are affected by many factors, such as the line structure, system parameters, and pre-fault load components. The author used EMTP simulation to simulate the fault voltage waveforms for different combinations of system parameters, fault location, pre-fault load component and breaker disconnection time, extracting the most representative features as the input of the neural network, and selected The input node is a non-fully connected BP neural network with 6 implicit nodes of 5 and output nodes of 1. A design detail of adaptive single-phase reclosing based on artificial neural network is proposed. A special method of training neural network using fault data and its implementation in hardware are proposed.
A new method of adaptive adjustment of return line distance protection settings. Due to the influence of mutual inductance, the fixed value adaptive adjustment of the double-circuit line distance protection is very difficult. A new approach is to introduce a correction factor that compensates for parallel line mutual inductance in the impedance setting. Since the correction factor has a nonlinear relationship with the system state, it is difficult to estimate by conventional methods. Therefore, a BP neural network is used to estimate the parameters such as positive sequence bus voltage, positive sequence line current, load current, and double-circuit line operation mode. The correction factor is implemented to adaptively adjust the protection action value in a timely manner.
The implementation method was discussed and its possibilities for application in radiant power distribution systems were discussed. Through a large number of simulation experiments, the neural network is used to simulate the fault conditions in different regions and different arc resistances. The problem of whether the protection is operated by BP neural network based on real-time impedance measurement values ​​during online application is discussed. It has practical significance for solving the problem that the fault under the arc resistance of the power distribution system cannot accurately measure the fault impedance, which leads to the problem that the conventional distance protection cannot be reliably operated.
The new method of microcomputer distance protection. The method uses the recursive full-circle Fourier algorithm and various compensation methods to perform data pre-processing on the input samples. Then, considering the characteristics of various faults in the line, the new grounding distance, phase-to-phase ratio phase criterion and phase are adopted. A three-layer forward neural network model is established by the sudden change of current difference; and a large number of sample training is carried out on this model to give correct identification of various faults and abnormal working states of the line. This method overcomes the influence of system operation mode and transition resistance and other factors, and solves the problem that the short circuit phenomenon occurs during system oscillation to protect the lock and open, and effectively improves the reliability of the protection action. The method of speed relay protection fault classification. The network uses a four-layer feedforward neural network (FNN) training using the BP algorithm. Two fault classification methods are proposed, using two different neural networks. A classification for single-phase grounding, two-phase short-circuiting, two-phase short-circuiting, three-phase short-circuiting, and three-phase grounding short-circuit faults; and another classification for arc faults and non-arc faults in order to achieve adaptive automation Reclosing.
Simulations show that the fault classification calculation using this method is fast and reliable. It also introduces the specific implementation of the neural network used and its specific application in fault detection and fault direction identification.
Application of multi-layer feedforward neural network and Kohrnm network in fault type identification.
The identification of the inrush current state in transformer protection has been a thorny issue that has plagued relay protection researchers. Based on artificial neural network, considering the characteristics of transformer magnetizing inrush current state and fault state, a three-layer forward neural network model is proposed and established to realize the microcomputer protection of transformer. It uses EMTP to perform a large number of simulation calculations, and uses the calculation results as training samples to train the established neural network model. The fault state test results of the model show that the established neural network can respond correctly to the fault state of the transformer.
Method of voltage fault current and magnetizing inrush current. The network uses a BP-type feedforward neural network (FNN) to distinguish between transformer fault currents and magnetizing inrush currents by identifying current waveforms. In order to improve the calculation speed, the transfer function of the neuron is selected as the S(Sigmoid) function during training, and after the training is finished, it is changed to the step function. The training samples of the magnetizing inrush current are generated by randomly switching a small transformer in the laboratory, and the training samples of the fault current are generated by the simulation software.
2 The direction and main content of adaptive protection is proposed. A kind of artificial Meng network for high-complexity intelligent technology and signal processing. Protection of bookmark3 adaptive relay protection can overcome the same type of traditional protection for a long time. Difficulties and problems improve the performance of the protection action. At present, adaptive protection is still in the early stage of research and development, but its research results have shown its superiority. The basic requirements for adaptive relay protection are automatic diagnosis and identification of system operation modes and fault types, and adaptive adjustment of protection action settings and characteristics. With the development of power systems, system operation modes and types of faults are becoming more and more complex. Adaptive protection must utilize various artificial intelligence technologies and signal processing methods to effectively extract fault characteristics and realize automatic identification of system operation modes and fault types. On this basis, the self-learning and self-adaptive ability of artificial intelligence technology is fully utilized, and various protection settings and protection action characteristics are adaptively adjusted according to different operating conditions of the system. To achieve these goals, the key is to select and use appropriate fault information for adaptive protection, and to select and use Lu Zhengjun. Microcomputer bus protection bus operation mode adaptive scheme PowerSystems, Ge Yaozhong. Adaptive Relay Protection and Its Prospects 丨 J. Power System He Pentium, Jin Huayu. Research on the principle and characteristics of energy direction protection. Chinese Journal of Electrical Engineering, 1997, Yuan Rongxiang, Chen Deshu. Research on new differential protection for high-voltage transmission lines. Chinese Journal of Electrical Engineering, 2000, 20 (4): 9-13. He Pentium, Jin Huayu. Implementation and testing of energy direction protection. Automation of Power Systems, 199721 (3) 36-38. Luo Wei, He Jiali. Real-time communication in the application of ultra-high voltage multi-circuit bus protection. Chinese Journal of Electrical Engineering, 199919(4)11-3. Power plant equipment automation 1992. Relay Quan Yusheng, Yang Minzhong, Wang Xiaorong, and so on. Online estimation of adaptive line parameters in double-ended ranging is 1J Power System Automation, 2000. Power System Automation 2000
Various products of Watch Bag, providing product images and basic parameters with each Double Drawstrings Vogue Watch Bag and Printed Watch Bag; We are a professional and qualified manufacturer of Watch Bag, and look forward to your cooperation chance. Year-round online, quick response, sample offered in time.
Leather Bags,Leather Tote Bag,Side Bag,Clutch Bag
Danyang Flanders Textile Co., Ltd , https://www.flandersmicrofiber.com