Enerji Dağıtım Sistemi Empedans Değişimlerinin İletişim Frekanslarında Analizi

Abstract

Son yıllarda teknolojik gelişmelere bağlı olarak yüksek verimli, ekonomik ve güvenilir bir enerji dağıtım sistemi otomasyonunun gerçekleştirilmesi için yoğun araş tırmalar yapılmaktadır. Özellikle çalışmalar, otomasyon amaçlı iletişimin, enerji dağıtım hatları üzerinden sağ lanmasına yönelmiştir. Enerji dağıtım hatları üzerinden iletişimde empedans değişimlerinin etkisinin belirlenmesi oldukça önemlidir. Bu çalışmada 5-20 kHz arası iletişim frekansların da dağıtım sistemi empedans değişimleri matematiksel ve deneysel olarak incelenmiştir. Dağıtım hattı empedans değişimlerinin matematiksel olarak hesaplanması amacıyla bir algoritma geliştirilmiş ve hattan çekilen güç değişimine bağlı olarak 5-20 kHz f rekanslardaki bara admitans değerleri hesaplanmıştır. Hesapla bulunan değerler ölçme değerleriyle karşılaştı- rılmıştır. Aynı güç değerindeki ölçme ve hesap değerle rinde oldukça yaklaşıklık olduğu görülmüştür. Dağıtım transformatörleri empedans değişimlerinin iletişim frekanslarında incelenebilmesi için iki kapılı bir model verilmiştir. Modelde tanımlanan primer, sekon- der ve primerden sekondere veya sekonderden primere geçiş admitans parametrelerinin 5-20 kHz f rekanslardaki deği şimleri elde edilmiştir. Gerek dağıtım hatlarının, ge rekse dağıtım transformatörlerinin 5-20 kHz f rekanslarda ki davranışlarının analizi ile deneysel sonuçların yorumu yapılmıştır.

In this study, we have investigated both analyti cally and experimentally the possibility of communication over power distribution lines in order for realizing automation of energy distribution systems and distribu tion systems impedance changes affecting communication negatively. In general an energy distribution system consists of a central station, substations, distribution lines and loads. The central station is equipped with connectors and measurement, control and protection devices. It has been suggested that system automation could be r tions have the major impairing effect on the communicj tion. In addition, we saw that there has not been reached any efficiency of desirable level in the word on this topic and that there has not ben formed any definite standards. Some of many important factors impairing communi cation over power distribution lines are below: - Impedance variations - Noise - Propagation irregularities - Crosstalk - Equipment reliability - Interference As stated before, the effect of impedance changes is prevalent and therefore this effect was mathematically analyzed in detail and experimentally examined. Other factors were presented with brief explainations. In figure 1, a two-port model of the transmission line with distributed parameters for communication frequencies is shown. The behaviour of this two-port model was discussed by using Wedepohl steady-state solution method. VI I. ı- Section i-j of transmission power line ij, Figure 1. Two-port representation of section i-j of transmission power line communication frequencies. E. = SİCosh(Yr)lS 1E. -Z.Q|Sinh(Yr7İQ 1I. »-lr,-l, v-l. (1) (2) I. = -Q | Sinh(yr ) | Q^Z^E.-fQ [ÇÜMTÖ1 Q~% where : Q = eigenvector matrix of matrix P gt _ yp * zp YP = per unit length transverse 3-phase admittance matrix ZP = per unit length longitudional 3-phase impedance matrix S = eigenvectors matrix of matrix P P = ZP * YP ~,, Z = 3-phase characteristic impedance matrix= S »Y. S. ZP

= diagonal matrix Y = 3-phase square root of eigenvalues of matrix P or matrix P r = length of segment i-j The equations (1) and (2) are Wedepohl exact solutions giving output variables in terms of input variables. An algorithm is developed in attempt to obtain prompt numerical results of impedance changes on compu ter. For the distribution model considered, nodal admit tance matrices at communication frequencies are obtained. In addition, some admittance matrices corresponding to situations of different power take-up from lines were constructed. By reducing the nodal admittance matrix of the order of (n+l)x(n-l), bus admittance. was obtained as below, YB = A-B.D ""-.C where A: is the submatrix, (lxl), having first row and first column element B: is the submatrix, (lxn), having first row element from the second column through the (n+l)th. C: is the submatrix, (nxl), having first column elements from the second row through the (n+l)th. D; is the submatrix, (nxn), which comprises all the "* elements but first row and first column elements. Vll For some power levels, bus admi versus communication frequencies of 5- plotted. We have made experimental studies tion line of which the value of bus ad calculated. Bus impedance values corr levels to 25 kW drawn were obtained by compared to those found by calculation shown in figure 2. It is noticed th experiments and calculations were simi rences between the curves were conclud result of capacitive effects that were calculations. ttance curves 20 kHz were on the distribu- mittance had been esponding to power measurement and These are at curves from lar. The diffe- ed to be the ignored in Figure 2, 5 10 15 20 *- f(kHz) The variation of bus admittances with frequencies. For experiments on distribution transformers it is considered powers of 50, 100 and 250 kVA, which are the most common ones in use. No any experimental study was encountered in the literature about distribution transformers Primary Secondary Figure 3. Two generator transformer models, Vlll 2,3 Y11(10"* mho) 15 20 *¦ f(kHz) 0,006 Y21(mho) 0,3 Y22(mho) Figure 4. For 50 kVA distribution transformer, the variation of Y..., Y9,, Y"2 admittances with frequencies. IX of this rates for studying impedance changes at communi cation frequencies. A simple two-part model of distri bution transformers was given figure 3. For this model, the following parameters are defined as critical communication parameters. Y- 1 = - t: - with V"=0 (short applied to secondary) 1 I. r i I. '2 Y", = - y - with V_=0 (short applied to secondary) (3) 1 Y"" = - y - with V..=0 (short applied to primary) Y.. ~ = - n - with V..=0 (short applied to primary) (Y12 = Y21) The variance of these parameters at the range of 5-20 kHz was measured for 50 kVA distribution transformer. The variation of these parameters with frequencies has been shown in figure 4. From the curves plotted after measurements it was concluded that all admittance values were inversely proportional with the frequency and that capacitive effect between turns increase with the increa sing power rates of transformers due to the enlargement of conductor sections. Finally, it has been concluded that distribution system could be employed for communicative purposes on conditions that impedance changes at high frequencies should be sufficiently estimated and data transmitters and receivers should accordingly be designed.