There are coherence phenomena in the dynamic process of the power system, that is, the consistency or similarity of the dynamic characteristics of different units identifies the coherent cluster correctly, which is of great significance for studying the stability of the power system. There are many ways to identify coherent clusters, such as numerical integration method state space method, feature vector method, modal equivalence method, weak coupling method, etc. These methods have their own defects, such as formula derivation is complex, the calculation is large, or can not be used for large systems In fact, there is no coherence in the strict sense of the system. The so-called coherence is only the dynamic characteristics of different units are similar, and there is a strict boundary between similarities. According to this feature of the power system, the fuzzy clustering method is used to identify the coherent clusters in the power system. Firstly, the fuzzy mathematics method is used to process the system state matrix to obtain the fuzzy equivalent matrix reflecting the degree of homogeneity among the units, and then according to The set truncation factor can easily identify the coherent cluster of the system. The method has clear physical meaning, simple calculation, no dimension disaster problem, and can be used for the identification of large-scale systems coherent clusters. This paper uses 10 machines and 24 machines. Two experimental systems are used as an example to give the identification results of coherent clusters under different conditions, and through homology-recognition eigenvalue reduction calculation results, it is proved that adopting fuzzy clustering method to identify coherent clusters is a simple, flexible and reliable two-fuzzy. When the equivalence relation matrix is ​​used to identify generator co-incidence, the general assumption is that the generator-related group is not related to the disturbance, so a linearized system model can be used; the correlation between the generator-associated group and the generator model is irrelevant, and thus a simple generator model can be used. Neglecting the excitation system and the prime mover governing system when the generator adopts the Eq constant model When the damper is ignored, the equation of motion of any of the generator power system after a few linearization step can coefficient matrix table group. The squaring of R is the same as the multiplication of the matrix in the form, but the large and small arithmetic rule is used on the value of the element. That is, if the ith row and the /th column of R are multiplied, R2n is satisfied by the method of self-multiplication. After transitivity, 21 and 21 become fuzzy equivalence relation matrices, which can be used for fuzzy cluster analysis to identify the homogeneity of the system. Coherent group 3 coherent cluster identification If R is a fuzzy equivalence relation matrix, then for any given 汜, can obtain a; V cut the classification matrix "," reflects the coherence relationship between the unit at the V level on the value of the formula for the matrix in, the value of its elements is only 0 1 two possibilities, if a 2 line elements in If there is a 1 in the same position, the unit corresponding to the 2 rows of elements is a coherent unit; if there are 1 rows of elements in the same position, the unit corresponding to these row elements is a coherent group. The use of fuzzy clustering methods to identify coherent clusters does not require any speculation from people, and it has sufficient flexibility. The level of coherence is required to be high. A larger value of V may be used. If the level of coherence is low, a smaller value of V is used. In addition, this coherent cluster identification method is basically a logical calculation, there is no rounding error, and therefore there is no dimension disaster problem, can be used for the identification of coherent clusters of large-scale power systems 1J. 4 based on coherent recognition of the eigenvalue reduction order Calculation Generally speaking, there are two kinds of electromechanical oscillation modes in the power system, namely the regional oscillation mode and the regional oscillation mode. The regional oscillation mode exists within the coherent cluster and is dominated by the unit inside the coherent cluster, but is basically not related to the unit outside the coherent cluster. The regional oscillation model exists between coherent clusters and is dominated by coherent clusters. According to this feature, the characteristic roots of regional oscillation patterns and regional oscillation patterns can be calculated using different state equations respectively, thus reducing the order of the system model. For the regional oscillation modes, they are dominated by the units within the coherent cluster and are basically unrelated to the units outside the coherent cluster. If the system state matrix is ​​rearranged according to the distribution of coherent clusters, then the characteristic root of the regional oscillation pattern can use the corresponding coherent cluster. Corresponding sub-blocks of the state matrix are obtained. Since there are relatively few internal generators in the coherent cluster, the dimension of the corresponding sub-block of the state matrix is ​​very low. Using the QR method for this sub-state matrix, the characteristic roots of the regional oscillation model can be obtained. In addition, all the characteristic roots obtained are obtained. The smallest one of the characteristic roots is due to the neglect of the effects of the external units of the coherent cluster. It is often small, corresponding to the system's 0 characteristic root, and should be discarded for the regional oscillation mode. They are dominated by coherent clusters and For the same tune-up group, the dynamic process of its internal unit is similar, and the homogeneity machine group can be equivalently converted into a generator by accumulating the rotor motion equations of the unit, thus reducing the system state equation. Using the QR method for this reduced-order state equation, the characteristics of the regional oscillation mode can be obtained. Similarly, among all the feature roots found, the smallest one corresponds to the zero characteristic root of the system and should be discarded. If there are n machines in the coherent cluster system, if each coherent cluster has /1/2...,/, and the station generator, then /1+/2+...+1m=n then the regional oscillations obtained by the above method Modal nm, the number of characteristic roots of the regional oscillation pattern is m-1, and the sum of the two is n-m+m-1=n-1, which is exactly equal to the total number of characteristic roots of the electromechanical model of the system. Reduced-order system model, but still able to obtain a complete system of electromechanical mode solution set machine, 24 machines using the Gnome state space results are completely consistent X to take different values ​​when the 24 machine system coherence machine group identification results are shown in Table 4 2 Fuzzy equivalence relationship Matrix row and column number Table 3 When X takes different values, the recognition results of coherent clusters of the 10-machine system can be seen from Table 3 and Table 4. When X takes different values, different coherent cluster identification results are obtained. The value of X depends on the research question. In general, the more stringent the requirement for homology (X is the larger), the more coherent clusters are obtained and the number of units in each coherent cluster. The smaller the number; the more relaxed the coherence requirement (X takes a smaller value), the less the number of coherent clusters that are obtained, and the greater the number of units in each coherent cluster. In addition, the value of X is also related to the specific system. For example, the X value of the 10-machine system can be more than 0.9. However, after 24 systems, when the X value is greater than 0.6, there is no coherent cluster. Table 4 X take different values ​​when the recognition results of the coherent cluster of the 24 machine system into the coherent cluster recognition results 5.2 characteristic root calculation results Table 5 and Table 6 respectively give X and fetch different values, 10 machine and 24 machine system according to reduced order model The resulting eigenvalue calculations are used as comparisons. The table also gives the eigenvalues ​​obtained by the full-order model and the reduced-order method calculation errors. Table 5 Ordered full-order reduced order method Error-reduced-order method Error-reduced-order method Error number model Table 6 24-machine system eigenvalue calculation result order full-order reduced-order method error-decreasing-order error-reduced-order method error number model from Table 5 and Table 6 It can be seen that the maximum error obtained by the coherent recognizing of the reduced-order model = 0.8 is 3.42%, while the maximum error when the recognizing-reduced model is = Q9 is only 0.92%. This shows that it is feasible to use the fuzzy clustering method to identify the coherent machine group of the power system. In addition, this method can be easily and flexibly selected between the degree of model reduction and the calculation accuracy: X takes a larger value, then the requirements of homology are more stringent, the calculation accuracy is higher, but the model is reduced in order; Smaller values ​​require less coherent requirements and less computational accuracy, but the model is much more downscaled. 6 Conclusion The use of fuzzy clustering method to identify coherents in power systems is a very effective method. It has a simple principle, small amount of calculation, reliable method, and is suitable for large-scale power systems. In addition, this method can be simplified and equivalent in the system. Flexible and convenient choices between sexes to suit the needs of research issues. Thank you Thank you for your interest in this research project. Powder-Actuated Insulation Fastening Tool Powder-Actuated Insulation Fastening Tool,Powder Fasteners,Insulation Powder Actuated Fastener , Powder Actuated Fastener Nail Yibin Heheng Technology Co.,Ltd , https://www.actuated.nl