Optimal Design of Substation Grounding Grid Based on Genetic Algorithm Technique

Transcription

1 Optial of Substation Grounding Grid Based on Genetic Algorith Technique Ahed Z. Gabr, Ahed A. Helal, Hussein E. Said Abstract With the incessant increase of power systes capacity and voltage grade, the safety of grounding grid becoes ore and ore proinent. In this paper, the designing substation grounding grid is presented by eans of genetic algorith (GA). This approach purposes to control the grounding cost of the power syste with the aid of controlling grounding rod nuber and conductor lengths under the sae safety liitations. The proposed technique is used for the design of the substation grounding grid in Khalda Petroleu Copany El-Qasr power plant and the design was siulated by using CYMGRD software for results verification. The result of the design is highly coplying with IEEE standard requireents. Keywords Genetic algorith, optiu grounding grid design, power syste analysis, power syste protection, single layer odel, substation. I. INTRODUCTION N every electrical installation, one of the ost iportant I aspects is the adequate grounding; ore specifically the grounding of high voltage substation [1]. Grounding, generally ean an electrical connection to the general ass of earth, the latter being a volue of soil, rock etc. whose diensions are very large in coparison to the electricity syste being considered. It is worth noting that, in Europe they tend to use the ter Earthing whilst in North Aerica, the ter Grounding is ore coon [2]. A power plant with a reasonable grounding syste is the key to the safe operation of a power syste. The working grounding is designed for different operation odes of the power syste. Grounding has a lot of purposes like, reducing the insulation level of electrical equipent, ensuring safe operation of power syste, ensuring personnel safety, eliinating electrostatic accidents, detecting ground faults, and ensures that externally exposed conductive bodies of a device have the sae potential by eans of equipotential bonding, reducing the electroagnetic interference. Finally, soe equipent needs to be grounded functionally like cathodic protection [3]. It is very iportant to design the grounding syste correctly so that there is no danger for huan life. After high-voltage substations are constructed, solving the probles related to grounding syste can be expensive and difficult. Hence, grounding grid design ust be carried out consistently [4]. The grounding syste Ahed Z. Gabr is with Khalda Petroleu Copany, P. O. Box 560, Maadi, Cairo, Egypt (e-ail: eng.ahed.zakaria.gabr@gail.co). Ahed A. Helal and Hussein E. Said are with the electrical and control engineering departent, Arab Acadey for Science, Technology and Maritie Transport, P. O. Box 1029, Miai, Alexandria, Egypt (e-ail: ahedanas@aast.edu, hdesouki@aast.edu). includes all of the interconnected grounding facilities in the substation area, including the ground grid, overhead ground wires, neutral conductors, underground cables, foundations, deep well, and so on. The ground grid consists of horizontal interconnected bare conductors (at) and ground rods. The design of the ground grid to control voltage levels to safe values should consider the total grounding syste to provide a safe syste at an econoical cost [5]. At the event of short circuit or any ground fault occurrence at any substation, the current ay flow across any paths. All these paths are depending on its ipedance. So, the ground fault current ay flow through the overhead transission lines or through the substation and surrounding earth or across all these paths together. The first guide for the substation grounding design was introduced on 1961: the ANSI/IEEE standard, and it was based on a lot of experience and odels. This docuent and other three revisions on 1976, 1986, and 2000 are the ain helping tools for engineers in designing a substation grounding at systes [6]-[8]. The IEEE definition of grounding is: a conducting connection, whether international or accidental by which an electric circuit or equipent is connected to the earth or soe conducting body of relatively large extent that serves in place of the earth [2]. A lot of studies were ade to describe and analyze the substation grounding grid design criteria. In 2011, Hellany et al. ade a study to view the safety restrictions of substation grounding grid design [9]. In 2014, Lantharthong presented the electrical effect of two neighboring distribution substation during the construction phase and they found that the size of auxiliaries grounding grid have an effect on the entire grounding syste [10]. CYMGRD is a software progra specialized in the substation grounding at design. It ay be used to ake a new design for new grids or to optiize and enhance an existing one of any shape. This odule can evaluate the estiate places for the danger voltage points in the grid and its adopting with IEEE STD. The CYMGRD software enables choosing the ost econoical way for any installation through a lot of design alternatives [11]. CYMGRD has been used to investigate the effect of increasing in grounding grid resistance on transient overvoltage which is caused by short circuit, switching, and lightening on the interior equipent and safety in a grounding grid at substation. The indices for ground grid safety are expressed and grounding grid analysis had done with the CYMGRD software. In 2009, Uzunlar and Kalenderli used CYMGRD software to ake a coputer odel for analysis of grounding systes conforing to IEEE standards. Their ethod and coputer software were supported with a real case easureent [12]. 863

2 GA is a coonly used technique to solve the optiization probles even if these probles were constrained or unconstrained. At each iteration, it generates a lot of points called population. The best point in this population is the nearest one to the optiu solution [13]. In 1998, Otero et al. used the GA ethod to iniize the total cost of the grounding grid design [14]. In 2004, Neri used the GA ethod to control the touch voltage in the grounding grid design [15]. In 2009, Yi-in et al. used the GA ethod to ake an optial design for grounding grids [16]. In 2009, Yang et al. used the GA ethod to ake analysis on soil structure for the grounding projects [17]. In 2009, Gursu and Ince used the GA ethod to liit the GPR in a two-layer soil odel [18]. In 2011, Zhiqiang and Bin used the GA ethod to ake the soil odel inversion calculations [19]. In this paper, a ethod for constructing a grounding grid substation is proposed by using an approach based on hand calculations, CYMGRD software, and GA technique. The ai is to iniize the cost of the grounding syste by iniizing the total length of conductors and the quantity of grounding rods while the safety restrictions required by the IEEE Std regulations are et. Although, here, only rectangular grids are considered for siplicity, the ethod is totally applicable to systes with any other shape. II. SUBSTATION GROUNDING GRID DESIGN, SAFETY CHECK AND OPTIMIZATION The ain objective of this research is not only to develop a ethod to design a grounding grid at but also to optiize the construction and aterial costs of a grounding grid at while still satisfying the axiu GPR, and step and touch voltages, and GA is the technique which had been used for this optiization. A. Hand Calculations It is iportant to use the hand calculations ethod in the substation grounding grid design because it allows us to get the appropriate distance between conductors. Total length of conductors and the appropriate nuber of rods will be used. In this paper, a real case study fro Khalda Petroleu Copany El-QASR power plant substation had been designed. This power plant lies on the western desert South Matrouh, Egypt, with Latitude: N, Longitude: E, and Altitude: 800 ft. The preliinary layout of grid with equally spaced conductors, with spacing D= 10, grid burial depth h= 0.5, grid with 20 ground rods. Each rod is 2 long, and it is placed around the perieter of the grid. The Decreent factor D f = 1, the current division factor S f = 0.6. Fault duration t f =0.5 sec. An average soil resistivity of 100 Ω. is assued, based on soil resistivity easureents with asphalt surface layer with 0.5 thickness. The total fault current is 40 ka and the X/R ratio is 10. Using copper annealed hard drawn and an abient teperature of 45 C, the person s weight can be expected to be at least 70 kg, consequently the area occupied by such a grid is A= Fig. 1 Substation Grounding Grid and Optiization Block Diagra 1. Calculating the Ground Resistance R ρ 1 (1) / L (2) L 1L 1L (3) L (4) R 100 Field Data (A, ρ) Hand Calculations To get D, L c, L T using CYMGRD Safety Restrictions t Y Optiize the design using Genetic Ω 2. Calculating the Maxiu Grid Current I I D S (5) I G = =24000 A 3. Calculating the GPR GPR=I G R g (6) The revised ground potential rise GPR is (24000) (0.435) = V. N 864

3 4. Calculating the Touch and Voltage C 1... C E. (8) E (7) E.. (9) E Calculating the Mesh Voltage (10) K ln ln (11) K ii =1, K 1 (12) h 0 =1 grid reference depth n=n a n b n c n d n (13) n n c & n d =1 rectangle grid n a = & n b =1.008 n= K (14) K i = n (15) K i =2.422 L L L (16) L M = E V. 6. Calculating the Voltage E (17) K 10.5 (18) K s =0.190 L s =0.75L c +0.85L R (19) For usual buried length 0.25<h<2.5 L s = ( ) + ( ) = E V The calculated corner esh voltage is lower than the tolerable touch voltage ( V versus V) and the coputed E s is well below the tolerable step voltage ( V versus V). A safe design has been established. Note that all these results had been calculated by using IEEE Std [2] as shown in Fig 2. B. CYMGRD Ipleentation for El-Qasr Substation Although, the hand calculations were very iportant for the design to get D, L c, and L t, but this way has any disadvantages as: 1. It cannot deterine the potential at each point inside the grid along x-axis or y-axis. 2. It cannot deterine the potential at the boundaries of the grid which have the ost critical values. 3. It cannot deterine the appropriate distribution of the rods along the grid perieter. All these values can be deterined, and safety is checked by using the CYMGRD software. The CYMGRD software is a substation grounding grid design and analysis progra specially designed to help engineers to optiize the design of new grids and reinforce existing grids, of any shape, by virtue of easy to use, built-in danger point evaluation facilities. The ain features of this software are: coputation of R g and GPR (Ground Potential Rise), touch and surface potential analysis, inside and outside the grid perieter, with color display in 2D or 3D representation. TABLE I COMPARATIVE RESULTS FOR RECTANGLE GRID WITH GROUND RODS Properties Hand Calculations CYMGRD Software % Difference Max allowable touch voltage V V 0.8% Max allowable step voltage V V 0.8 % Reduction factor Cs % Ground Resistance Ω Ω 3.8 % Ground Potential Rise V V 0.8 % The data entered to the CYMGRD software is listed in Appendices (see Table III). The results obtained fro the 865

4 CYMGRD progra are shown in Figs. 3-5, while the coparative results for rectangle grid with ground rods are shown in Table I. 11 Modify (D, n, L c, L T ) N N Field Data (A, ρ) Conductor Size (3I c c, t c, d) Touch and Criteria (E touch h70, E step70 ) Suggested Initial (D, n, L c, L T, h) Grid Resistance (R g, L c, L R ) Grid Current (I G, t f ) 7 I G R G <E t Mesh & Voltages (E, E s, K,K s, K i, K ii, K h ) E <E t h 9 10 E S <E St Detail 12 Fig. 2 Substation Grounding Grid Procedure Block Diagra C. Optiization Using GA The objective of the grounding grid design is to find a structure of buried conductors that ensures the safety of people and equipent, using the sallest possible nuber of rods and sallest possible length of conductors. Both objectives are in conflict, and coproise solution ust be found. So, the proble can be expressed as the search for an arrangeent of buried conductors observed the following objectives: 1. and touch voltage at any point of the installation ust be lower than the axiu value required by the regulations. 2. The ground resistance R g ust be lower than 0.5 Ω. Y 3. The total length of conductors and total nuber of rods (the cost) should be as low as possible. Despite the CYMGRD software was very useful in the design safety check but it cannot get an optial solution for the nuber of rods and the total length of conductors to iniize the installation cost, so we used the GA technique to optiize the solution that we get fro both hand calculations and the CYMGRD software. Matheatically, the proble ay be expressed as the iniization of the cost function: C C L C N. (20) The optiization of the function C ust satisfy the constraints: E Lenght (eters) E, E E and R 0.5Ω Fig. 3 El-Qasr Grid layout using the software Besides these objectives, the following conditions are introduced in the design with the ai of siplifying the proble: 1. A continuous conductor loop surrounding the perieter of the substation is assued. This loop enclosed the total area of the installation. 2. The spacing between the conductors is equally spaced D. That akes the total length of the conductors L c as (3). A MATLAB function using the GA technique was developed that allowed the nuber of rods and spacing between conductors to vary as part of the optiization routine. This technique akes the calculations of the grounding resistance faster. At each iteration of grounding grid resistance calculation, it should be ensured that the safety paraeters and constraints "step and touch voltages" are still in the safety 866

5 region values. The fitness function was to iniize the total cost of designing the substation grid as shown in equation (20). There were two variables, the first was X(1) (the spacing between conductors) and the second was X(2) (total nuber of rods). X(1) ust be an integer nuber according to the distribution constraints, and X(2) is constrained to be an even nuber. There were three nonlinear constraints to keep the optiized design in a safe region. Also, the design has a lower bounds and upper bounds for the variables. The population type was double vector to keep the output in an integer for, then chose the population size. The default of the population size is (10 no. variables), then run the solver to get the results. This research used El-Qasr power plant as a case study. The actual paraeters used in the objective function are listed in Appendices (see Table IV). Fig. 4 El-Qasr profile plot in 2D for The perissible values for the touch and step voltage were calculated depending on the soil paraeters of the power plant as previous. After copletion of the optiization, the results were 24 spacing between conductors and 23 placed rods. X(1) = 24 and X(2) = 23 rods. So, the total length of the installed conductor will be Referring to Eq. (3). A coparison between the original design of Khalda El-Qasr power plant and the optiized one had been established as 867

6 shown in Table II. Fig. 5 El-Qasr contour plot in 3D for As it is clear in Table II, the optiization technique saved about 403,290 L.E. (Egyptian Pound) TABLE II COMPARISON BETWEEN ORIGINAL DESIGNS AND OPTIMIZED ONE Properties Original Optiized Critical Values Spacing between conductors D Nuber of Rods n R Total length of conductors L c potential E step Touch potential E touch R g Cost (L.E) V V Ω 767, V V 0.47 Ω 364, V V 0.5 Ω III. CONCLUSION In every electrical installation, one of the ost iportant aspects is adequate grounding, not only the proper design but also an optiized one to get the iniu cost of design. We used the hand calculations ethod in the substation grounding grid design because it allows to get the appropriate distance between conductors, total length of conductors, and the appropriate nuber of rods. Hand calculations cannot deterine the potential at each point inside the grid along x- axis or y-axis, it cannot deterine the potential at the boundaries of the grid which have the ost critical values and It cannot deterine the appropriate distribution of the rods along the grid perieter, so a CYMGRD software had been used to eet all these requireents. Although the CYMGRD software was very useful in the design safety check but it was unable to get an optial solution for the nuber of rods and total length of conductors to iniize the installation cost, so we utilized the GA techniquee to optiize the solution obtained fro both hand calculations and the CYMGRD software. El- Qasr power plant was designed and finally a 403,290 L.E had been saved by using the GA optiization technique. APPENDICES TABLE III INPUT DATA TO CYMGRD PROGRAM Properties Surface layer resistivity Surface layer thickness Top layer resistivity Top layer depth Lower layer resistivity Body weight Fault duration Ground fault current X/R ratio Current split factor S f Grid size No. of conductors in X direction No. of conductors in Y direction Conductors burial depth Conductors diaeter Conductor type No. of rods Rod diaeter Rod length Rods arrangeent Rod type I/P Data Ω 70 Kg 0.5 Sec 40 KA % 92 * Copper annealed hard drawn Along the grid perieter Copper annealed hard drawn Unit Ω. Ω Ω Kg Sec KA 2 TABLE IV EL-QASR POWER PLANT OPTIMIZATION PARAMETERS Properties Value Unit C cond 281 L.E L.E. C rod 143 L.E L.E. C operator 150 L.E L.E. C install 4292 No.operators 10 One rod length 2 N r X(2) L c LY 92 LX 134 D X(1) LB [0,4] UB [92,30] 868

7 TABLE V LIST OF USED SYMBOLS Sybol Noenclature Units ρ Soil resistivity Ω. ρ Surface Layer Resistivity Ω. A Total area enclosed by ground grid 2 C s Surface Layer Derating Factor C cond Cost of one eter of Conductor L.E C rod Cost of one rod L.E C operator Cost of one operator L.E C install Cost of installation L.E d Diaeter of grid conductor M D Spacing between parallel conductors D f Decreent Factor E Mesh voltage at the center of the corner esh for the siplified ethod Volt E s voltage between a point above the outer corner of the grid and a point 1 diagonally Volt outside the grid for the siplified ethod E step70 Tolerable step voltage for huan with 70 kg body weight Volt Et ouch70 Tolerable touch voltage for huan with 70 kg body weight Volt h s Surface layer thickness I f Total fault current Apere I L Average current per unit of effective buried length I L The average current per unit of buried length of grounding syste conductor K i Correction factor for grid geoetry K ii Corrective weighting factor that adjusts for the effects of inner conductors on the corner esh K Spacing factor for esh voltage K s Spacing factor for step voltage LB Lower Bonds L c Total length of grid conductor L p Perieter of the grid L Total length of grid conductor L R Total length of ground rods L r Length of ground rod at each location L s Effective length of Lc+LR for step voltage L T Total effective length of grounding syste conductor, including grid and ground rods L X Maxiu length of grid conductor in x direction L Y Maxiu length of grid conductor in y direction NR Nuber of rods placed in area A n Geoetric factor coposed of factors na, nb, nc and nd n a Factor used to calculate n n b =1 for square grids n c =1 for square and rectangular grids n d =1 for square, rectangular and L-shaped grids R g Resistance of grounding syste Ω S f Fault current division factor (Split factor) t s Duration of shock for deterining allowable body current Sec UB Upper Bonds [4] Gonen, Turan. Electric Power Transission Syste Engineering: Analysis and CRC Press, [5] John D. Mcdonald, Electric Power substation engineering, CRC press, [6] Gary Gilbert, High Voltage Grounding Syste, [7] "IEEE IEEE Guide for Safety in AC Substation Grounding". [8] "IEEE IEEE Guide for Measuring Earth Resistivity, Ground Ipedance, and Earth Surface Potentials of a Ground Syste. [9] A.Hellany, M.Nagrial, M.Nassereddine and J.Rizk, Safety copliance of substation earthing design, International Journal of Electrical, Coputer, Energetic, Electronic and Counication Engineering Vol:5, No:12, [10] T.Lantharthong, N.Rugthaicharoen cheep and A.phayoho, Optial analysis of grounding syste design for distribution substation, International Journal of Electrical, Coputer, Energetic, Electronic and Counication Engineering Vol:8, No:8, [11] CYMGRD Product Features. [12] Uzunlar and Kalenderli, Three Diensional Grounding Grid, International Conference on Electric and Electronics Engineering (ELECO), (I-139-I-143), [13] Mathworks helping tool. Retrived fro [14] A.F. Otero, J. Cidras, C. Garrido, Genetic algorith based ethod for grounding grid design, in: IEEE International Conference on Evolutionary Coputation Proceedings, 1998, pp [15] F. Neri, A new evolutionary ethod for designing grounding grids by touch voltage control, IEEE International Syposiu on Industrial Electronics 2 (2004) [16] Yi-in, Min-fang, et.al, Optial of Grounding Grids Based on Genetic Algorith, Third International Conference on Genetic and Evolutionary Coputing, [17] Yang, Wen, et.al, The Analysis on Soil Structure for the Grounding Projects, Power and Energy Engineering Conference, [18] Baris Gursu and Melih Ince, Liiting GPR in a two layer soil odel via genetic algoriths, journal of the franklin institute, [19] He Zhiqiang and Zhan Bin, Soil Model s inversion Calculation Based on Genetic Algorith, 7th Asia-Pacific International Conference on Lightning, REFERENCES [1] Gonen, Turan. Electric Power Distribution Syste Engineering CRC Press, [2] Earthing Practice CDA Publication 119, February [3] Jinliang He, Rong Zeng and B Zhang, Methodology and Technology for power Syste Grounding, 2013 John Wiley & Sons Singapore Pte. Ltd. 869