A NOVEL DEVICE (OIL SPRAYING SYSTEM) FOR LOCAL COOLING OF HOT SPOT AND HIGH TEMPERATURE AREAS IN POWER TRANSFORMERS Kourosh Mousavi Takami1 Jafar Mahmoudi2 1-TDI researcher and Ph.D. student in Malardalen University 2-professor in Malardalen University, Sweden Box 883,721 23, IST Dep., Mälardalen University, Västerås, Sweden P.O. Box: 13445686, Sharif institute of technology (TDI), Tehran, Iran [email protected], [email protected]

ABSTRACT Power transformer is a vital device in substations. Load and no load losses create a hot spot point in transformer, so it is obviously necessary to limit the core temperature to values that cause no damage to the core itself, adjacent materials, or the oil. Core temperatures as low as 110–120 o c may degrade oil. This has led experts in the field to suggest that 130 C would be a reasonable limit for the core temperature. For this reason, need to a device for local cooling is necessary. Authors in this paper find a novel method for oil cooling in hot spot point or area. Oil spraying device is simulated and then path of pressured oil after shooting in hot spot point is evaluated. With using 230/63/20kv Punel substation in Iran data and after simulation and calculation, authors find that with using of these device utilities could remove their problem for over loading of transformers in the load peak time.

Keywords: Nozzle, oil spraying, oil, and power transformer INTRODUCTION Hot spot point could affect on life of transformer. Accurate calculation for identification of hot spot point by many approaches is a smart for electrical engineering because with find the accurate local of hot spot could use a good device for cooling of that’s point. In the this being time, manufacturers did not use a device for removing of hot spot point, utilities only decreasing the load, increasing of fan speed and oil circulation speed for remove of hot spot point. This method could cause the using of other transformer in parallel mode, with using of parallel transformer on short time some times cause the production of over flux, over thermal and inrush current and may be malfunction of relays [1] and so shot down of transformer or black out in networks. Assunçlo, J. L. Silvino, and P. Resende [2] only propose the estimation of the top-oil temperature by using a method based on Least Squares Support Vector Machines approach. The estimated top-oil temperature is compared with measured data of a power transformer in

operation. They only find a model but don’t introduce a device for cooling. Ed G. teNyenhuis, Ramsis S. Girgis, Günther F. Mechler, and Gang Zhou, presented an accurate calculation of the value of the core hot-spot temperature at the design stage is becoming very critical in order to ensure that hydrogen generation that can be caused by oil-film degradation at core hot-spot temperatures as low as 110–120 o c is

limited[3].

K. Eckholz, W. Knorr, M. Schäfer, K. Feser, E. Cardillo, heat transfer coefficients at the winding surface were calculated using heat run test results of various ON- and OD-cooled winding types. Two different general approaches were made for ON- and OD-cooled windings. The characteristic of the heat transfer coefficients was calculated for each winding type. Further temperature rise experiments with an ON-cooled disc winding operated with variable heat flux densities were performed in order to investigate the influence of the heat flux density on the cooling efficiency of the boundary layer [6]. They only found an accurate method for calculating of heat transfer coefficient. Utilities need to a device for removing the hot spot problem. Using of oil spraying system could increase the temperature rise in transformer above of 65 o c . With calculation the rate of oil flow, speed and temperature of oil that have to throw to hot spot, could decrease the hot spot point temperature. The volume of spray mix applied per area by a sprayer depends on:

• •

Nozzle flow rate Width sprayed

• Travel speed of the sprayer. Before find a sprayer, select the nozzles to be used. Nozzles should be matched and replaced as a set except under unusual circumstances. The following study to provide a framework for working through the calculations necessary for nozzle selection and sprayer calibration.

Experimental data In the 230/63/20KV PUNEL substation in Iran, on a 230/63/0.4KV transformer in July 2006 in during of 24 Hours, at 1 month all of oil, low voltage &high voltage and third voltage winding and ambient temperature by many sensors measured, and this result only for one day illustrated in figure 4.

Variations of ambient temprature

variation of temperature in TR. (july) 40

66

35

therid w inding temp.

62 60

low voltage w inding temp

58

high voltage w inding temp

56 54 52

oil temp.

Temp. (oc)

temperature in oc

64

30 25 20 15 10 5 0

50 1 3

5 7 9 11 13 15 17 19 21 23

1

3

5

7

Hours

9

11 13 15 17 19 21 23 25 27 29 31 33 Hours

Fig.1. Variation of temperature in transformer and ambient for three winding 250MVA, 230/63/0.4KV transformer on July 2006 in Pounel 230/63/20KV substation – Iran

In first view on figure 1 could see that low voltage winding have high degree in temperature and only because it carry out high current in comparison of another’s windings, and could suggested that, create a high thermal profile between LV. and H.V. windings. Another hand for calculation and simulation and for coordination with IEEE standard we used from Power transformer 250 MVA in IEEE loading guide 1995 in below: Table I: transformer data from IEEE guide 1995[7]

Transformer Losses W. No Load 78100 Pdc losses (I2 Rdc ) 411786 Eddy losses 41200 Stray losses 31660 Nominal voltage 118 KV 230KV Pdc at hot spot location 467 527 Eddy current losses at hot spot location 309 (0.65 pu) 157 (0.3 pu) Per unit height to winding hot spot 1 1 Temperature Rise °C . Rated top oil rise 38.3 Rated top duct oil rise 38.8 Rated hot spot rise 58.6 50.8 Rated average winding rise 41.7 39.7 Rated bottom oil rise 16 Initial top oil 38.3 Initial top duct oil 38.3 Initial average winding 33.2 Initial bottom oil 28 Initial hot spot 38.3 Transformer component weights, kg Mass of core and coil assembly 172200 Mass of tank 39700 Mass of oil 37887 HIGH PRESSURE NOZZLE DEFINITION The purpose of this section is to simulate the operation of a nozzle, perhaps the most important and basic piece of engineering hardware associated with the high-speed flow of oils for cooling of hot spot. The usual configuration for a nozzle is shown in the figure1. Oil flows through the nozzle from a region of high pressure (usually referred to as the chamber or Tank) to one of low pressure (referred to as the hot spot area). The Tank is usually big enough so that any flow velocities here are negligible. The pressure here is denoted by the symbol pc. Oil flows from the chamber into the converging portion of the nozzle, past the throat, through the diverging portion and then shoots into the hot spot point. The pressure of the HST point is referred to as the 'back pressure' and given the symbol pb. To understand how the pressure behaves we have to remember only a few basic rules: • When the flow accelerates, the pressure drops • The pressure rises instantaneously across a shock • The pressure falls across an expansion wave. Note that the temperature distribution behaves qualitatively like the pressure distribution.

b

a

Figure 2: a-nozzle b-schematically cut of power transformer with oil spraying system [1].

Select the sprayer application rate

A recommended range of sprayer application rates, in liter per mm2, is given on the pesticide label. From that range, choose the rate that is compatible with your spraying equipment.

Select the field speed A 10 percent reduction in speed will result in a similar over application of decrease of heat transfer coefficient. Where speed must be reduced for a short period at the end of a pass through the field or in other situations, a small reduction in pump speed is preferable to selecting a lower cooling. Reducing the cooling speed causes the nozzle flow rate as well as the travel speed to decrease, although not proportionately. If slowing is frequently necessary, calibrate the sprayer for the slower speed.

Determine the width sprayed by each nozzle The width sprayed by each nozzle on a broadcast spray boom is the distance between nozzles. If a sprayer has nozzles spaced every 40 inches on the boom, the width sprayed by each nozzle is 40 inches. If a sprayer has several nozzles that will be used to spray each row, such as sprayers used to apply to HST, then the width sprayed by each nozzle is the distance between rows divided by the number of nozzles used to spray each row. If the sprayer is a band sprayer and the labeled rate applies to the actual area sprayed, the width is the width of the band. If the sprayer is a band sprayer and the labeled rate applies to the total area covered, the width is the spacing between rows. But because we use from fuzzy logic and intelligence system, the nozzle could rotate and have many freedom degree. We could decrease the number of nozzles to half or less. Determine the nozzle flow rate The nozzle flow rate can be calculated:

GPM = GPA x V x W x k

(1)

Where k = a constant for conversion of units

Travel speed: Travel speed

V =

D × 60 T

(2)

Where V = travel speed (mm/s) D = distance (mm) T = travel time (seconds) 60 = a constant (seconds per minute)

High Viscosity Problem for nozzle The high viscosity of the oil prevents the nozzles from forming the equivalent spray patterns, at equivalent temperature and pressure settings. The first and most obvious solution is to increase the working pressure of the system. The drawbacks to this solution are as follows; higher cost in the higher pressure rated pumps, motors, and delivery lines to the nozzles. This also increases the safety concerns in the event of ruptured lines in the building. A booster pump on the outlet of this system may be trying.

The second solution is to have local area pumps. The third option to be considered is to use some sort of rotating nozzle, which would spray the oil in a rotating type fashion. TEMPERATURE CALCULATION: In most of the heat and mass transfers the following equation can be seen. In this paper, convection and conduction consider for all the calculations. The following are the assumptions considered for simulation: • Oil and air initially are in steady state. • Oil is a Newtonian and incompressible fluid. • Fluid flow is laminar, unsteady and two-dimensional. • There is internal heat generation. The thickness of the copper-insulation layers, as shown in Fig. 2(a), can be calculated from the design data of the winding. Table 2: Typical values of k and K of transformer windings kr

kz

K

w / moc

w / moc

w / m oc

Winding type

Transformer rating(MVA)

2.5 7.2 7.35

4.4 4.46 4.46

3.32 5.67 5.72

Hv disc Lv disc Lv disc

250 250 250

log kr = (

r log 0 r1 kr1

+

r log 3 r2 kr 2

rn r1

(3)

+ ... +

r log n rn −1 kr1

)

Similarly, the t.c. of a disc or layer in -direction can be calculated as in (7) [Fig. 2(b), [1]]

kz =

k cu k kp k pb (t cu + t kp + t pb )

(4)

k cu t kp k pb + t cu k kp k pb + k cu k kp t pb

The calculated effective thermal conductivity of a typical power transformer Windings in axial and radial direction has been given in Table I. These values have been used in the computation of HST. The winding is a thermally inhomogeneous structure for this; the thermal conductivity should be treated as a tensor. The thermal conductivity takes the following form: k ten =

k rr

k rz

k zr

k zz

Considering the transformer winding (simplified as above), the insulation structure closely satisfies the orthotropic structure in the orthogonal coordinate system and the above equation becomes k rr

0

k ten =

= 0

The terms

k zz

kr 0

0 kz

=

k1 = k 2 0

0 k3 = k 4

k r and k z are called as principal thermal conductivity.

Heat Flow Equations: With the thermal conductivity, the system of no homogeneous HCE under a no homogeneous boundary condition in cylindrical coordinate system is written as

1 ∂T ∂  ∂T  ∂ 2T G + = r  + kz r∂r  ∂r  ∂z 2 K α~d ∂t In the region of a ≤ r ≤ b ,0 ≤ z ≤ l l ≥ 0 .

(5)

kr

At the inner cylindrical surface (r = a, t f 0) :

− k1

At the outer cylindrical surface (r = b, t f 0) :

k2

At the bottom flat surface ( z = 0, t f 0) : At the top flat surface ( z = l , t f 0) : In the region,

a≤r≤b

,0 ≤ z ≤ l

, and t = 0 ,

∂T + h1T = f1 ( z , t ) ∂r

(6)

∂T + h2T = f 2 ( z , t ) ∂r ∂T + h3T = f 3 ( z , t ) − k3 ∂z ∂T k4 + h4T = f 4 ( z , t ) ∂z

(7)

T = F (r , z )

(8) (9) (10)

( T = T (r , z , t ) ). The term G is the heat source function, and has been modified here to take care of variation of resistivity of copper with temperature. The heat source term G can be of the form (11) G = g 0 (1 + ρ t (T − T0 )) = g 0 − g 0 ρ t T0 + g 0 ρ t T = G0 + g 0 ρ t T

ρ

o

−1

Where t is the temperature coefficient of electrical resistance of copper wire in c With this representation, the function G becomes temperature dependent, distributed, heat source. For the sake of mathematical convenience and to provide a reference for the heat source function, the constant G0 is included to replace the constant part of (11). The term F (r, z ) represents the initial function for transient heat conduction problem. The ~





α d =  1 ρ Cp  = α d / k eq eq   term, where is the diffusivity. The time-dependent boundary functions f 1 .... f 4 , derived from Newton’s law of cooling, are of the following form:

f i ( z, t ) = f i ( I ) ( z ) + f i ( F ≈ I ) ( z )[1 − e

−( t )

τ

]

(12)

The degenerate form in steady state, along the axial direction, is shown to be as in (13)

f i ( z ) = hi Tz i = hi × (Tb + msi z ) Where:

(I )

f i ( z)

(13)

is the initial temperature at time (t = 0) , and f i

( F ≈I )

(z

) is the difference between final steady state

Tb

initial steady-state (t = 0) temperature. The term is the temperature at the bottom of the disc or layer, as applicable. Term m si is the axial temperature gradient. Similarly, functions f 3 (r ) and f 4 (r ) representing temperatures f (r ) = h j (Tb or Ttop ) . Heat-transfer across the bottom and top surfaces, in steady state, are of the form of j (t = ∞) and

coefficients h1....h4 , are different across all four surfaces, the values of which can be calculated by using heattransfer empirical relations given in [2]. RESULTS AND DISCUSSION

Magnetic field solution

1-

2-

Fig.2. 1- A 2-D transformer cross-section illustrates essential parts subject to leakage flux 2-Transformer field solution in FEM environment

In this work a transformer model was adapted using the FEM analysis software FEMLAB [7]. The FEM analysis was carried out on different industrial transformers to estimate the winding eddy current losses. Circulating current losses in continuously transposed conductors (CTC) or properly transposed windings is small and their contribution to the eddy current loss was considered negligible. Fig 2.2 shows the field solution of transformer. The leakage flux in the windings flows axially up through the coils and then bends radially across the windings. From Fig 2.2 it can be seen that the component of the leakage flux has its greatest concentration at the interface between the two windings and then decreases progressing away from the gap between the windings. The inner LV winding typically has a higher attraction of the leakage flux due to the high permeance of the core. The leakage flux in the HV winding is divided between the core and the core clamps and other structural parts [8]. In the upper end of the windings, the conductors are exposed to an inclined magnetic field with two components, an axial component and a radial component. The eddy current loss is the contribution of these two components. Hot spot and thermal simulations We wrote a program by PDE operator in MATLAB software environment based on genetic algorithm, which have done in M-file of MATLAB. Of course, assessment was done in the steady state mode. We give the second part of the equation (5) equal to zero and solve them. We have obtained and calculated these phenomena, with out oil spraying and with oil spraying, and finally have modeled the transformer oil flow to obtain the flow pattern.

Although the velocity and temperature distributions will be simulate by fluent software. We find for the maximum internal temperature above the surrounding oil T (0, 0) - Toil=25°C. The surface temperature rise in the top of oil is approximately T (top)-T (oil)=17.8 °C and bottom oil is T (bottom) –T (oil) =12.2 °C This modeling shows that the hot spot point is located in 91% of the core and winding, height from the bottom. With oil spray to top of winding (in 91% height of core), in the top of oil we could see that, without oil forced, oil temperature will be intensity low. For the maximum internal temperature we find above the surrounding oil we have T (0, 0)-Toil= -35.8°C. But there is a problem in the power transformer for oil spraying and that is electrical and magnetic field problem. While spraying oil, non-homogeneous field might occur, and this causes a break down before the manufacturer determinations may occur, and for this problem we have to assess and calculate field effects. evaluation of temperature profile with ONAN and oil spraying to %91 of core height

130

125

90 100

120

120

115 110

115

105

temprature in oc

Temprature in oc

125

80

80

70

60

60

40 50

100 1

110 0.8

0.5 radius of core and winding in pu.

0

0

0.2

0.6 0.4 Height of core and winding in pu.

Figure3: evaluation of temperature profile with OFAF

1

20 1

40 1

105

0.5 radius of core and winding in per unit

0.5 0

0

30

per unit of core and winding height

Figure4: evaluation of temperature profile with OFAF and oil spraying to %91 of core height (location of hot spot). With nozzle cross section of 2826mm2 and oil speed 1m/s

Simulation of nozzle in shooting of oil Flow of sprayed oil, which is passing through the oil in the tank, is calculated using MATALAB environment; these are illustrated in figures 3 and 4. It is shown that oil pressure form nozzle to the HST point will be decrease, but it has not too much created disturbances, and so produced turbulence is negeliable. In the figure of 5and 6 red, yellow and low blue and after that high blue color create a flow rate from high to low range. But Reynolds and nusslet number have not a sensible variations, then oil treatment don’t change from laminar to turbulence. For heat transfer coefficient (HTC) calculations in this clause a simplified approach is used:

h(G ) = C.G n

(14)

Coefficients C and n depend on the geometry were evaluated from measured HTC using a least square method. The values of the coefficients in equation (14) are calculated to C and n that shown in table 2. The experiments were made on an ON-cooled disc winding with oil guiding elements and radial cooling ducts. The cross section of the investigated winding is natural convective cooling (ON) and forced directed cooling (OD), and the winding equipped with a number of PT100 sensors in order to measure local temperatures.

diameter of oil flow from nozzle center

3 2 1 0 -1 -2 -3

flow diameter in horizental direction in decimeter

2

10 0

5

-2 0

Figure 5: flow pattern that emission from nozzle to HST S/So=100%

Heat transfer cofficent(w/k.m2)

1200

1000 S/So=30%

600

S/So=20%

400

200

0

S/So=10% 0

50

100

150

200 q(w/m2)

250

300

350

Figure7: heat transfer coefficient versus heat flux density on much cross section of nozzle

1 0 -1 -2 2 0 1

2

3

6

5

4

7

8

9

distancce from nozzle to core and winding in decimeter

Figure 6: flow pattern in slice form by 30-degree rotation form nozzle to HST

1400

800

2

diameter of oil flow -2 from center

distance between nozzle and winding in decimeter

400

temperature diffrerence between oil and winding(Twinding-Toil)

flow diameter in verticall direction in decimeter

In figure 7 variations of heat transfer coefficients with heat flux density is evaluated in much cross section of nozzle, of course S0=2826mm2 and oil velocity due to pumping and nozzle action is 1m/s. in the practical may be can not use from this cross section for nozzle, but operational of cooling is as same.

45

40

S/So=15%

S/So=40%

35

S/So=70%

30 S/So=100%

25

20

15

10 50

100

150

200 250 heat flux density(w/m2)

300

350

400

Figure8: differential of winding and oil temperature versus heat flux density in much nozzle cross-section.

In figure 8 shown that with change of heat flux density in some cross sections of nozzle, differential of temperature in winding with oil is considerable. It means that with higher cross section removing of heat from winding is better.

In Table 2 the values of the coefficients C and n, stated in equation (14), for all four oil entrance cross sections are given. Table2: Calculated coefficients C and n.

S/So C n

100% 9.1 0.41

30% 1.65 0.61

20% 0.495 0.685

10% 0.83 0.66

CONCLUSIONS An electromagnetic analysis using a finite element model has been adapted to predict transformer core and winding losses. This can be used to calculate the eddy losses in individual turns/discs to enable location of the winding losses that cause the hot spot. With increase of ambient temperature or load, the temperature will increase which causes a rise in the temperature and creates hot spot, as the solution of this problem with consider action of magnetic field aspects we offer the use of oil spraying devices, that sprays oil on hot spot point. Oil sprays provide many times the cooling rate for removing of hot spot. By changing the geometric parameters of the spraying apparatus and the physical parameters of the oil one can substantially vary the cooling rate in any temperature range. The other way is installation of many oil-spraying devices on the tank that can give order according to location of the hot spot after identifying it, and its nozzles can rotate to any angle until it can spray oil to hot spot point or area. See figure 4. Control system can be designed with genetic algorithm, neural network, fuzzy logic and any another algorithm that could be chosen for spraying. It’s noted that there’s no needed to use piping inside tank. All the spraying will be done only from the tanks body. NOMENCLATURE GPM nozzle flow rate (liter per s) GPA sprayer application rate (litre/mm2),

t cu

Nozzle crosses section (mm2) width sprayed by each nozzle (mm). Time variation of heat generated. Total thickness of copper,

t kp

Total thickness of kraft paper,

t pb

Total thickness of pressboard in axial direction

Cp

Specific heat at constant pressure.

S W G(t)

k rr ....k zz h K l t T Tf u,v x, y, z

α αp τ ∆t

ρ µ

Conductivity Coefficients Heat transfer coefficient. Thermal conductivity. Length. Time. Temperature. Surrounding temperature. Velocity components. Cartesian coordinates. Relaxation parameter. Relaxation parameter for pressure. Thermal process time constant (TC)

Time difference. Density. Dynamic viscosity.

REFERENCES [1]Kourosh mousavi Takami, A FFT technique for discrimination between faults and magnetizing inrush currents in power transformers, KAHROBA scientific magazine specialized in power electric engineering, Mazandaran, Iran [2]T. C. B. N. Assunçlo, J. L. Silvino, and P. Resende, Transformer Top-Oil Temperature Modeling and Simulation, transactions on engineering, computing and technology volume 15 October 2006 ISSN 1305-5313 [3]Ed G. teNyenhuis, Ramsis S. Girgis, Günther F. Mechler, and Gang Zhou, Calculation of Core Hot-Spot Temperature in Power and Distribution Transformers, IEEE transactions on power delivery, Vol.17, NO. 4, October 2002 T. V. Oommen, R. A. Ronnau, and R. S. Girgis, “New mechanism of moderate hydrogen generation in oil-filled transformers,” in Proc.CIGRE Conf., Paris, France, Aug.–Sept. 1998, Paper 12-206. [4] “Hydrogen generation for some oil-immersed cores of large power transformers,” in Proc. DOBLE Conf., Mar.–Apr. 1998. [5]K. Eckholz*, W. Knorr, M. Schäfer, K. Feser, E. Cardillo, new development in transformer cooling calculations. [6]IEEE Loading Guide for Mineral Oil Immersed Transformer, C57.91, pp. 18–19, 46–53, 1995. [7] FEMLAB V2.3, Electromagnetics Module. Comsol, 2002. [8] A. Elmoudi, M. Lehtonen, “Eddy losses calculation in transformer windings using FEM,” The 44th International Scientific Conference of Riga Technical University, Riga, Latvia, October 9-11, 2003, Series 4, Vol. 10, pp. 46-51. [9] K. Haymer and R. Belmans, Numerical Modelling and Design of Electric Machines and Devices, WIT-Press, 1999. [10] M. David and Others, Finite Element Method Magnetic FEMM. [11] A. Konard, “Inegrodifferential Finite Element Formulation of Two-Dimensional Steady-State Skin Effect Problems,” IEEE Trans. on Magnetics, vol. MAG-18, no.1, Jan 1982, pp.284-292. [12] J. Weiss and Z. J. Csendes “A One Step Finite Element Method for Multiconductor Skin Effect Problems,” IEEE Trans. on Power Apparatus and Systems, vol. PAS-101, no.10, Oct. 1982, pp.37963800. [13] A. Konard, M. V. K. Chari, and Z. J. Csendes “New Finite Element Techniques for Skin Effect Problems,” IEEE Trans. on Magnetics, vol. MAG-18, no.2, Mar. 1982, pp.450-455. [14] Lates L. V., Electromagnetic calculation of Transformers and Reactors: Moscow ENERGY, 1981 (In Russian), p.313. [15] D. Pavlik, D. C. Johnson, and R. S. Girgis, “Calculation and reduction of stray and eddy losses in core form transformers using a highly accurate finite element modelling technique,” IEEE Trans Power Delivery, vol. 8, no. 1 Jan. 1993, pp.239-245. [16] Kourosh Mousavi Takami, Evaluation of oil in over 20 year’s old oil immersed power transformer, Mazandaran University, May 2001. [17] Kourosh Mousavi Takami, Advanced Transformer Monitoring & Diagnostic Systems and thermal assessment with robust software's, research presentation, Water and power University, March 2007, Tehran, Iran [18] Kourosh Mousavi Takami, Hot Spot identification and find a best thermal model for large scale power transformers, April 2006, KTH University, Stockholm, Sweden.