OPTIMAL DG UNIT PLACEMENT FOR LOSS REDUCTION IN RADIAL DISTRIBUTION SYSTEM-A CASE STUDY

Transcription

1 Asian Research Pulishing Network (ARPN). All rights reserved. OPTIMAL DG UNIT PLACEMENT FOR LOSS REDUCTION IN RADIAL DISTRIBUTION SYSTEM-A CASE STUDY A. Lakshmi Devi 1 and B. Suramanyam 2 1 Department of Electrical and Electronics Engineering, SVUCE, Tirupati, India 2 Narayana Engineering College, Gudur, India energylak@yahoo.co.in ABSTRACT Distriuted generators are eneficial in reducing the losses effectively pared to other methods of loss reduction. In this paper optimal DG unit placement using fuzzy logic is discussed. The optimal size of the DG unit is calculated analytically using approximate reasoning suitale nodes are determined for DG unit placement. Voltage and power loss reduction indices of distriution system nodes are modeled y fuzzy memership functions. Fuzzy inference system containing a set of rules is used to determine the DG unit placement. DG units are placed with the highest suitaility index. Simulation results show the advantage of optimal DG unit placement pared optimal capacitor unit placement. Compared to capacitor placement it is giving very good reduction not only in power loss ut also it is improving voltage regulation. Keywords: distriuted generation, DG unit, capacitor, power loss. 1. INTRODUCTION Distriution system provides a final link etween the high voltage transmission system and the consumers. A radial distriution system has main feeders and lateral distriutors. The main feeder originates from sustation and passes through different consumer loads. Laterals are connected to individual loads. Generally radial distriution systems are used ecause of their simplicity. Power loss in a distriution system is high ecause of low voltage and hence high current. The over all efficiency can e improved using DG units Loss reduction in distriution systems There are many methods of loss reduction techniques used like feeder reconfiguration, capacitor placement, high voltage distriution system, conductor grading, DG unit placement. All these methods are involved with passive element except DG unit placement. Both DG units and capacitors reduce power loss and improve voltage regulation. But with DG s loss reduction almost doule that of Capacitors Methods of reducing loss using DG unit Many methods have e in recent times on DG unit placement [5] proposed a method to calculate the size of DGs analytically y using exact loss formula which requires lot of putation pared to the proposed analytical method. Many authors like [1] mentioned the allocation of DG units using genetic algorithm. They have not considered the optimum size; they have addressed the prolem in terms of cost [3]. They have iteratively increased size of the DG at all uses and then they calculated the losses, ased on loss calculation they ranked the nodes. Top ranked units are selected for DG unit placement. Here a new analytical method is used to calculate the size of the DG units. A new of method of minimizing the loss associated with the active ponent of ranch currents y placing optimal DG units at proper locations. Here cost function is not considered. Considering the cost function involves the deviation of exact size of the DG unit at suitale point Fuzzy logic There are many uncertainties in various power system prolems.because of this it ees very difficult to stick to mathematical formulae alone. To over e this, fuzzy set theory has een applied to many power system prolems. Using fuzzy expert system a set of heuristic rules is used to determine the Dg unit placement suitaility index at each node in the distriution system. Rules are defined to determine the suitaility of a node for DG unit placement. 2. DG UNIT INSTALLATION The prolem of DG unit placement consists of determining the locations and sizes and numer of DG units to install in a distriution system such that maximum enefits are achieved while operational constraints at different loading levels are satisfied Distriution losses The total power loss in a distriution system having a numer of ranches is given y: P L = ΣI i 2 R i (1) i =1 Ii is the current magnitude and R i is the resistance. Ii can e otained from load flow study. The ranch current has two ponents: active (I a ) and reactive (I r ). The loss associated with these two ponents can e written as: 57

2 Asian Research Pulishing Network (ARPN). All rights reserved. P LP =Σ I ai 2 R i ; (2) P LQ =Σ I ri 2 R i (3) For a given configuration of a single source radial network the loss P LP associated with the active ponent of ranch currents cannot e minimized ecause all the active power must e supplied y the source at the root us. This is not true if DG units are to e placed at different nodes for loss reduction. That is real power can e supplied locally y using DG units of optimal size to minimize P LP associated with the active ponent of ranch currents Analytical method The method proposed first identifies a sequence of nodes to e pensated. The sequence is determined y repetitive application loss minimization technique y a singly located Dg unit. Once the sequence of nodes to e pensated is identified, the corresponding optimal size at the pensated nodes can e determined simultaneously y minimizing the loss saving equation. 2.3 Loss minimization y singly located DG unit. Considering a single source distriution system with n ranches. Let a Dg e placed at us m and β e a set of ranches connected etween the source and Dg unit uses. If the Dg unit is placed at us x the β consists of ranches x1, x2, xn. The Dg unit supplies real current I real and for radial network it changes only the active ponent of current of ranch set β. The current of other ranches is not affected y the Dg unit. The new active ponent of current I ai (new) = I ai +D i * I DG Di = 1 if ranch iε β = 0 otherwise. I pi is the active ponent of current of ith ranch in the original system otained from the load flow solution. The loss P La is associated with the active ponent of ranch currents in the pensated system. P La = Σ(I ai +D i I DG ) 2 R i (4) Savings in Active power loss is: S = P La - P La = - Σ (2D i I ai I DG +D i 2 I DG ) 2 R i i = (5) To minimize the loss Eq. (5) is differentiated w.r.t I DG and equated to zero. It results in: Σ (D i I ai R i ) i =1 I DG = (6) Σ (D i R i ) Distriuted generator size P DG = V m I DG (7) The process can e repeated for all the uses to get the highest possile loss saving for a singly located DG unit. Same procedure is repeated for optimal size of capacitor y optimizing the power loss related reactive ponent of ranch current [10] Radial distriution load flow analysis Conventional load flow studies like Gauss-seidal and fast decoupled load flow Newton raphson methods are not suitale for distriution system load flows ecause of high R/X ratio. A new radial load flow method for distriution systems that offers etter solution was proposed in ref [10]. The main features of this method are: 1) The initial voltage at all nodes is assumed to e the voltage specified at the source node. 2) No plicated calculations are involved 3) Loads are represented y constant power. 4) Evaluation of simple algeraic expression of receiving end voltages. 5) Convergence is otained y, even for ill conditioned system Algorithm There are many putational steps involved in finding the optimal DG size and location to minimize losses in a radial distriution system are: 1) Run the load flow program. Select the us where the maximum loss and low voltage is using fuzzy logic tool ox. Corresponding DG size is calculated using eqs, respectively. Repeat this for all the uses except the source us. Identify the us using the fuzzy logic that provides highest loss saving. 2) Compensate the us with the highest loss with the corresponding Dg unit found from eq. (7). 3) Repeat the steps 1) and 2) to get the next DG size and hence sequence of uses to e pensated. 4) Once the sequence of uses is known determine the optimum DG unit sizes and the corresponding loss saving. Since the system load is time variant and load duration curve of the system can e approximated.it is assumed that load level is constant. The aove algorithm provides the optimal DG sizes and locations for a given load level Fuzzy logic implementation There are many uncertainties in various power system prolems. Because of this it ees very difficult 58

3 Asian Research Pulishing Network (ARPN). All rights reserved. to stick to mathematical formulae alone. To over e this, fuzzy set theory has een applied to many power system prolems. Using fuzzy expert system a set of heuristic rules is used to determine the Dg unit placement suitaility index at each node in the distriution system. Rules are defined to determine the suitaility of a node for DG unit installation. Those rules are expressed in the following form: IF premise (antecedent), THEN conclusion (consequent). For determining the suitaility of Dg unit placement at a particular node, a set of multiple antecedent fuzzy rules has een estalished. The inputs to the rules are the voltage and power loss indices. The fuzzy variales, power loss index, voltage and Dg unit suitaility are descried y the fuzzy terms high, high-medium/normal, low-medium/normal or low. These fuzzy variales are descried y memership functions. Dg units are placed at the nodes with the highest suitaility. Voltage and power loss reduction indices of distriution system are modeled y fuzzy memer ship functions. FIS editor receives inputs from the load flow program. Several rules may fire with some degree of memership. FIS is ased on Mamdani max-min and maxprod implication methods of inference. These methods determine the aggregated output from the set of triggered rules. The max-min method involves truncating the consequent memership function of each fired rule at the minimum memership value of all the antecedents. A final aggregated memership function is achieved y taking the union of all the truncated consequent memership functions of the fired rules. For the DG unit placement prolem, the resulting Dg unit suitaility memership function µ d of node i for k fired rules are given y: Μ d (i) max [min [µ p (i), µ v (i)]] Where µ p (i) and µ v (i) are the memership functions of the power loss index and voltage, respectively. After calculating the suitaility memership function, it is to e defuzzified in order to determine the node suitaility ranking. The centroid method of defuzzification is used. The prolem of DG unit placement consists of determining the locations and sizes and numer of DG units to install in a distriution system such that maximum enefits are achieved while operational constraints at different loading levels are satisfied. 3. SIMULATION The proposed method of loss reduction y Dg unit placement was tested on a Distriution system consisting of 33 uses. Optimal sizes of DG units and capacitors are calculated at each and every us. Optimal location is otained using FIS editor. It has given the locations 26 th and 31 st uses. Results are shown for 26 th us. Capacitor and DG unit sizes are shown in Figures 2 and 4. Voltage profiles at different nodes are shown in Figures 3 and 5 with capacitor placed 26 th us and DG unit placed at 26 th us, respectively. Voltage profile improvement is very good with DG unit placement pared with that of Capacitor Unit placement. Total Real power losses and Total reactive power losses with capacitor and with DG Unit are pared with without any pensation. These are taulated in Tale-2 and Tale-3. Capacitor size at 26 th us is and DG size at 26 th us is KVAR and KW. The results are pared with ref [5]. 4. CONCLUSIONS With the help of FIS editor optimal location of the DG unit is found where real power loss is more and voltage is low. Using analytical method the sizes of the DG unit is found and placed at optimal location which is reducing power loss almost 100% so is reactive power loss. There is very good improvement in voltage levels also. The results are taulated in Tale-1 and Tale-2 pared with DG unit, with Capacitor and without any pensation. Optimal size of capacitor at 26 th us is KVAR and DG unit is KW. 59

4 Asian Research Pulishing Network (ARPN). All rights reserved. Figure-1. IEEE 33 us test system. Tale-1. Capacitor and DG unit sizes for 26 th us. Capacitor size in KVAR DG size in KW Tale-2. Real power loss parison with capacitor and with DG unit. Real power losses without pensation (KW) Real power losses with capacitor (KW) Real power losses with DG unit (KW) Tale-3. Reactive power loss parison with capacitor and with DG unit. without pensation (KW) with capacitor (KW) with DG unit (KW) efore after capacitor rating in KVAR voltage in p.u us numer Figure-2. Capacitor sizes using analytical method with capacitor at 26 th us us numer Figure-3. Voltage variation at different nodes. 60

5 Asian Research Pulishing Network (ARPN). All rights reserved. DG rating in KW voltage in p.u efore after REFERENCES 0 us numer Figure-4. Dg unit sizes using analytical method. [1] G. Celli and F. Pilo Optimal distriuted generation allocation in MV distriution networks. IEEE. [2] T. Griffin, K. Tomsovic, D. Secrest, A. Law. Placement of Dispersed generation systems for reduced losses. [3] M. Mardaneh, G.B. Gharehpettan Siting and sizing of DG units using GA and OPF ased methods. IEEE. pp [4] Fernado L. Alvarado Locational aspects of distriuted generation. IEEE, PES. [5] Naresh Acharya, Pukar Mahat, N. Mithulanathan An analytical approach for DG allocation in primary distriution network. March. [6] J. Balakumaran, K. Thanuskodi Loss reduction in radial distriution systems y capacitor placement Fuzzy technique. IEEE. pp us numer Figure-5. Voltage variation at different nodes with DG unit at 26 th us. [7] H.N. Nag, M. M. A. Salama and A. Y. Chikani Capacitor placement in distriution systems using fuzzy technique in Proc. Canadian Conf Elec. puter Engg. Vol. 2, pp [8] H.N. Nag, M. M. A. Salama Fuzzy optimal capacitor sizing and placement. Proc. Canadian Conf. Elec. put. Engg. Vol. 2, pp [9] Gareth P. Harrison, Antonio Piccolo, Pierluigi Siano, A. Roin Wallace Distriuted Generation capacity evaluation using ined genetic algorithm and OPF. International journal of emerging Electric power systems. Vol. 8, No.2. [10] M.H. Haque Capacitor placement in radial distriution systems for loss reduction. IEE Proceedings. [11] Das and Ghosh Method for load flow solution of radial distriution networks. IEE proceedings on line. 61