Engineering design and construction of a functional liquid Hg ... - WP5

Engineering design and ... in their construction to classical electrical machines. .... can reach the boiling temperature of mercury, which through a building of.
4MB taille 3 téléchargements 333 vues
Institute of Physics of University of Latvia (IPUL)

Engineering design and construction of a functional liquid Hg-loop

2005 Latvia, Salaspils

Saclay 28.29.10.05

CONTENT

1. Overall layout of the Hg test loop 2. Investigation of Hg induction pumps . 3. Investigation of Hg induction pumps on permanent magnets 4. Measurement system of Hg-loop parameters, flow internal structure, heat transfer etc. 5. Rough draft of the experimental Hg-loop for tests of pump and other components

Saclay 28.29.10.05

MERCURY LABORATORY OF THE INSTITUTE OF PHYSICS

Saclay 28.29.10.05

MERCURY STAND

Fig.

Saclay 28.29.10.05

PRINCIPAL SCHEME OF Hg-LOOP

Saclay 28.29.10.05

INSTRUMENTATION OF Hg-LOOP

t° C (30 - 150°) PT 100 (resistive platinum sensor) and thermocouples

Q, l/s (0 – 10) Induction flow meter

Hg - loop

Dynamic flow distributor - cooler

EMP pump on rotating permanent magnets P0>50 kW

Velocity and pressure field’s structure in the flow (ultrasonic Doppler anemometry

Fig,

Saclay 28.29.10.05

12345674894– 100)

Target flow tract temperature field (THERMA CAM P60)

Choices of pump types for analysis The use of electromagnetic pumps is attractive because the ability to obtain hermetically sealed systems. EM pumps can be designed in a large number of types and configurations. Detailed analysis of all types to obtain the best selection for each application would be a prodigious task. Each of the types of pumps is adapted to a certain field of operation; many of the pumps had one or more characteristics unacceptable for needed applications and were eliminated from further consideration. Conduction (single phase alternating current) pumps have great simplicity in use, but they are suitable for small installations. Comparing the pump characteristics the conduction pumps also are less attractive because of reliability and power conditioning difficulties. Flat induction pumps are of simple construction and can easily dismounted without interfering with the conduit, which can remain welded on the loop. The duct, rectangular in cross section, is located in the magnetic field between the two stators. The length of the pump is several pole pitches. The pressure reduction caused by side effects is related to the ratio of the width of the 2 24

4444242 44Very low resistance bars to improve the pressure and velocity distribution over the width of the channel can be attached to sides of the channel structure. These pumps can be divided into two classes: 1) normal pumps, which are suitable for medium size installations, with outputs varying from 10 to 80 m3/hr and differential pressure of several atmospheres and which can developed 5 to 8 atm at low flow rate, 2) high pressure pumps with the flow rate from 5 to 15 m3/hr, with pressure from 5 to 15 atm. Annular induction pumps have large outputs which can obtain several thousand m3/hr at differential pressures of the order of 10 to 15 atm. These pumps have a strong conduct made-up of two concentric tubes. The annular pump has not the side effect present in flat pump. The windings consist of toroidal coils whose axes are concentric with the axis of the annular duct. Windings are normally located only in the stator because the inaccessibility of the core. In principle the annular pump is identical to the flat pump. The ranges of application of the two (flat and annular) are moderate and high flow, low-pressure applications. Induction pumps working with heavy metals (Hg, Pb) metals have much lower efficiency and power factor as working with alkali metals (Figure 1a, 1b, 1c). The helical induction pump was chosen for these application because high pressure and relatively low flow rate is required. The selection was made using the recommendations (Figure 2 [1, 2]) based to the accumulated experience and developed theory in the design Helical pumps are closest in their construction to classical electrical machines. Helical pumps provide flow rates ranging from 0 to 20 m3/hr and pressures up to 16 atm. It is simple to obtain different pressure within the range of a single size of the stator iron Dstat by changing the number of elementary channels. Working channel pumps is made in helical form to decrease the active length. The flat pump channel is partitioned in order to pass several times the fluid pumped through the air-gap, each passage increases of pressure. With small mass flow rates of working body and large developed pressures, these pumps with a helical channel have definite advantages over flat and annular pumps. From the electromagnetic point of view, the helical pumps are three-phase flat induction pumps, working on the same principle. The stator is identical in principle and arrangement to the stator of an asynchronous motor. The length of path over which pressure is developed in the fluid is the length of a helical passage. The axial component of fluid velocity results in additional loss and reduction in output, but this effect is insignificant in pumps having a small helix angle

Saclay 28.29.10.05

.

LIMIT OF OPERATION FIELD OF THE VARIOUS TYPES OF EM PUMPS

- alternating current single-phase pumps • - direct current conduction pumps ∇ - alternating current flat pumps Ο - alternating current annular pumps x - alternating current helical pumps

Saclay 28.29.10.05

INDUCTION PUMPS

Fig.

Saclay 28.29.10.05

Saclay 28.29.10.05

HYDRAULIC With a rise in the power of linear induction MHD pumps the magnetic field changes the profile of the velocity over the cross section and length of the channel. The change in the distribution of the velocity over the width of the channel play considerable role with high pressures and small flow rates [6]. An inhomogeneous electromagnetic force field causes a motion of liquid metal in the channel and creates substantial stagnation zones, where always occur reciprocating internal circulations (Fig. 4). Nonuniform velocity profile, developed from the inlet to outlet, leads to the appearance of intense vortical flows in the output zone. It is exceedingly complex to analyze the effect of these zones. The flow instability and the resulting wide fluctuations of the discharge rate, vortical flows in the active zone of the channel, under certain conditions related to the external hydraulic circuit, serve as a source of perturbations leading to fluctuations of pump parameters – flow rate, developed pressure, current etc.The jet can then be thrown from one channel wall to another, as in the case of separation flow through diffusers of hydraulic machines. For each pump on the performance curves near the low velocity and high pressure is the forbidden working region where there is some danger for formation of vortical zones. For that reasons it is necessary to provide for good contact between the working channels walls and the liquid metal.

CALCULATED PARAMETERS of HELICAL INDUCTION PUMP Rating point with mercury at 100 oC is: a) flow rate (liter/sec)…………..2.3 b) differential pressure (bars)…..60 c) phase voltage (V)……………361 d) efficiency…………………….5.8% e) power factor ……………...….0.28

Power balance at the rating point is as follows: a) total input power………………………………..243 kW b) in the secondary circuit……….…………….…..220 kW c) in the copper of the winding…………………....19 kW d) in the channel walls (because they are thick) ….130 kW e) Joule power losses in the mercury…………........75 kW f) mechanical power……………………………….15 kW g) in the iron of a magnetic circuit………………....4 kW

Saclay 28.29.10.05

Conclusions 1. Many factors affect pump size, but the limitation is generally winding temperature and boiling temperature of mercury in the channel. Winding temperature is a function of winding power loss, fluid temperature, heat sink temperature, and the thermal properties of the path through which the loss passes from windings to heat sink. 2. There is high energy energy consummation in the secondary circuit, i.e. liquid mercury and channel walls. When the pump is operating, the duct of the pump is quickly heated, it is necessary to cool the channel of the pump, because unstable stagnation zones inside a channel can reach the boiling temperature of mercury, which through a building of hydrodynamic perturbations, will affect the entire velocity field. 3. The helical pumps have additional problems in manufacture and use. The length of duct is large and it is difficult to make inside cylinder threaded. The exiting coils have large length (approximately 1.5 m) and width (approximately 0.2 m) and it is not easy to form, insulate, impregnate and dry those. 4. The measurements show that in a definite interval of flow rates there exist in the channel stable velocity vortices occupying a considerable part of the width of the channel. It is established that the flow consists of transient, reverse and vorticial motions. There are established curvilinear flow lines, two (forward and reverse) kinds of motion and presence of a vortical motion. 5. In high-power pumps the profile of the flow can become unstable, which would lead to a sudden readjustment of the velocity structure of the flow The process of formation of reverse flows is significantly affected by edge effects, presence of lateral short-circuiting strips, species of the test circuit etc. 6. Induction high-power pumps have certain inherent disadvantages and limitations. Because of remarkable eddy current losses in secondary circuit polyphase induction pumps are not suitable for mercury systems with low boiling temperature. While the liquid metal is passing through the channel the temperature of the metal increases at P ∆T = a = 57.7 o C , where Pa is the power losses in secondary circuit, Q is the mass Qc Hg flow rate, cHg is the heat capacity of mercury. For higher pressure it is necessary that the mercury flows with a sufficient flow-rate to remove thermal power. The experience shows than it is very important to eliminate the stagnation zones in the channel and to provide for good contact between the liquid metal in the working zone and the channel or sidebars because the stagnation zones can warm up to boiling temperature of mercury.

Saclay 28.29.10.05

Design concepts of Electromagnetic Induction Pumps for Liquid Metals

The principle scheme of flat linear induction pump. 1- inductor; 2 – slots for windings; 3 and 5 – inlet and outlet diffusers; 4 - flat channel.

The principle

scheme of

cylindrical linear induction pump. 1 – inductor; 2, 6 – inlet and outlet diffusers; 3,4 – coaxial cylinders of annular channel; 5 – inner passive ferrous core; 7 – slots for windings.

Electromagnetic induction pump with rotating permanent magnets for mercury. Pressure P = 6 bar; Flow rate Q = 13 L/s (175 kg/s); Motor power for pump drive 50 kW. Saclay 28.29.10.05

Comparison of Design concepts of Electromagnetic Induction Pumps

Flat and Cylindrical linear pump

Permanent magnets pump

Main advantage: No moving parts

Main disadvantage: Rotating magnetic system

Rather complicated construction, bigger overall dimensions of active part and the weight:

Much simpler construction, smaller overall dimensions of active part and the weight:

Laminated ferrous yokes with slots for layout of No windings at all !!! 3-phase windings for generating alternating traveling magnetic field . Alternating traveling magnetic field is generated by Rather complicated construction of the annular system of rotating permanent magnets with alternating channel for cylindrical pump with laminated polarity fixed on solid ferrous base. ferrous yoke inside. Practically is not repairable (if sort-circuiting). Magnetic system can be easily reassembled Electrical parameters and efficiency: Lower efficiency and rather low coefficient cos f 12345267829 82 784

Electrical parameters and efficiency: power

Much higher efficiency. At using permanent magnets no energy consumption is needed for creating magnetic field, and additionally, much stronger magnetic field can Rather expensive power supply – 3-phase be generated in the same non-magnetic gap. High power coefficient cos f 1 0.8 of standard transformer with variable voltage for adjusting the productivity of pump. industrial AC motor for pump drive through standard frequency converter for adjusting the productivity of pump by rotation speed of magnetic system.

Negative end effects: In both pumps design concepts the negative longitudinal end effects exist. In cylindrical pump negative transversal end effect does not exist due to azimuthally symmetry.

Saclay 28.29.10.05

Negative end effects: The negative transversal end effect exists. The negative longitudinal end effect does not exist.

Evaluation of parameters of Permanent Magnets Pump for EURISOL Project EURISOL PMP design concept:

Parameters for first approximation evaluations: P = 20 bar; Q = 2.45 L/s; Ro = 187 mm; rf = 169 mm; Rf = 247 mm; Channel rectangular cross-cection: LM layer thickness b = 10 mm; 2a = 16 cm; Walls thickness: dw1 = 5 mm; dw2 = 1 mm; Non-magnetic gap d = 18 mm; R = 175 mm; V (in the channel) = 1.5 m/s; V B = 2πR*n = 1.1*n, m/s; lch = 1 m; τ = 10 cm; a/τ = 0.8; ε = 0.27; koc = 0.6; tm = 50 mm; hm = 60 mm; Br = 1.4 Tesla; Bavr = 1.3 Tesla; σ = 10^6 (om.m)-1; S = (1 -V/VB);

The sketch of the cross-section of one of three active parts of pump developing pressure 20 bars, providing flowrate 2.45 L/s.

Main basic formulas for evaluation of induction pump parameters: Pressure developed by pump:

Pm a x =

σ ⋅V B ⋅ B 2

2

⋅ s ⋅ lc h ⋅ k ;

(1)

Heat losses in the liquid metal layer in the active part of the channel of pump:

WM

2 P m2 a x ⋅ 2 a ⋅ b = ; σ ⋅ B 2 ⋅ lch ⋅ k

(2)

Heat losses in the electrically conducting walls of the channel of pump:

W M σ w  d w   kw   Bw   ; (3)  ⋅  ⋅  ⋅ = ⋅  σ   b   k   B  s2 2

WW

Saclay 28.29.10.05

Results of evaluation of parameters of Permanent Magnets Pump for EURISOL Project As one single pump for very high pressure P = 60 bars is under question (due to possible instabilities arising because of back flows in the channel in the side regions due to non uniform distribution of electromagnetic forces across the channel width) we considered to divide it into three smaller parts connected in series, each developing pressure about P1/3 = 20 bars and having different wall thickness of the channel. This allows sufficiently to diminish total power of pump by minimizing heat losses in the walls. For considered draft sketch (see previous slide) for the third section of pump (having highest wall thickness dw1 = 5 mm) we get the following results: Heat losses in the liquid metal layer in the active part of the channel of pump is 13 kW, heat losses in the electrically conducting walls of the channel will be approximately two times higher about 26 kW, so total power will be about 39 kW. For the middle section heat losses in the liquid metal layer, in the electrically conducting walls and total losses, correspondingly, will be 12 kW, 18 kW and 30 kW. For the first section of pump the efficiency will be highest as at thinner walls heat losses will be smaller and at smaller non magnetic gap magnetic field strength also will be greater. Heat losses in the liquid metal and in the walls, correspondingly, will be 10 kW and 12 kW and total power will be 22 kW.

CONCLUSIONS: 1) – First approximation evaluations of parameters of pump for EURISOL project carried out to demonstrate the order of parameters show that for developing pressure P = 60 bars and providing flow rate Q = 2.45 L/s) the total power of pump consisting of three serial parts will be about W = 90 kW. At this parameters the overall diameter of active cylindrical part of pump will be 500 mm, the total length of active part (including all three sections) will be about 600 mm. The efficiency of pump at these parameters will be: ef = PQ/(PQ + W)*100% = 14 %. 2) – The only way to reduce the power of pump and get higher efficiency is increasing the length of the active part of the channel of pump (as heat losses (see formula 2) are inversely proportional to the length of channel) by following means: a) by increasing the diameter of magnetic system, or b) by further dividing of the pump into smaller serial sections. 3) - For further optimization of parameters of pump the precise calculations are needed for defining the thickness of the walls of the channel of pump ensuring mechanical strength of the channel at rather high pressure, as the thickness of walls have rather dramatic influence on the efficiency of the pump. For example, increasing the wall thickness from 3 mm up to 6 mm leads to increasing of power for pump drive almost two times !!! Evaluations of parameters of pump was carried out by Dr. I.Bucenieks. Saclay 28.29.10.05

Fig.

Saclay 28.29.10.05

SKETCH OF ELOMAGNETIC FLOWMETER

75 123245467892 2454488

3 49 4 4 5

7 45 458 45 45 4

4

6

74

76

4

6

123456789 4 4 

4

2 45444 4 44  544 68456544 95  854 Saclay 28.29.10.05

4

SCHEMATIC LAYOUT OF HIGH POWER TARGET MODEL

Power absorbed in Hg-jet Operating pressure Flowrate Jet speed Jet diameter Temperature: Inlet to target Exit from target Total Hg inventory Pump power

Saclay 28.29.10.05

1 MW 100 nbar 2 t/s 30 m/s 10 mm 30 0 C 100 0C 10 t 50 kw

Saclay 28.29.10.05

SKETCH OF WINDOWLESS TARGET

Q= 10 L/s P= 5 Bar Fig.

Saclay 28.29.10.05

Saclay 28.29.10.05

ROUGH DRAFT OF Hg-LOOP

Saclay 28.29.10.05