(Dongfang Electric Machinery Co., Ltd.) designed the end structure of large air-cooled steam turbine generators. By considering the contact gap between the generator's pressure finger and the core of the edge, the pressure finger is included in the eddy current solution area, and the material characteristics are proposed. For the nonlinear and anisotropic methods, the mathematical model does not include the method of estimating the core loss in the solution area. The magnetic field and end of the QF-135-2-13.8 air-cooled turbo generator are completed comprehensively. Quantitative calculation and comparative analysis of the loss of each structural member, temperature rise prediction of the end structural parts, the calculation results show that: QF-135-2-13.8 air-cooled steam turbine generator adopting "T-type pressure ring + copper shielding" structure, It has the characteristics of small additional loss at the end and low temperature rise at the end, which can fully meet the requirements of rated load and phase-in operation of the generator.
Magnetic eddy current consumption. After the development of the QF60-2-10.5 and QF-75-2-10.5 two-type air-cooled steam turbine generators, the 1U-1 project has added new air to the series, and the stator core, coil and end structural parts adopt air. External cooling, the rotor coil is cooled in air, the two ends of the generator are equipped with a coaxial rotary fan, and the upper part of the generator is equipped with an air cooler. The whole generator forms a closed ventilation system with two inlets and three outlets. The QF60-2-10.5 and QF-75-2-10.5 two-type generators have been put into commercial operation in the first half of 20012002. The on-site and on-site real-machine performance tests show that the temperature and temperature rise of all parts of the generator meet the requirements of the technical agreement. The operating condition is good. The first QF-135-2-13.8 generator of our company will also complete the in-plant manufacturing and in-plant real machine test in September 2002.
Since the additional loss at the end of the generator is proportional to the square of the product of the line load and the stator inner diameter, as the single unit capacity of the generator increases and the material properties improve, the electromagnetic load becomes higher and higher, causing the magnetic field strength at the end of the motor. Significantly increased. Eddy currents, losses, and heat generation are generated in the end structural members, thereby causing a series of effects on the loss, temperature rise, and the like of the generator. In order to improve the end heat, the end loss of the lower part, and avoid the local concentration of the end magnetic field and loss, the following measures were taken in the design of the oriental air-cooled series turbine generator (Figure 丨 is QF-135-2-13.8 type hairpin End structure of the machine).
Low-magnetic cast steel pressure ring and non-magnetic forged steel pressure finger are used to reduce the loss of the pressure ring and the pressure finger; the iron core section of the stator side section is designed with a small step, and a narrow groove is opened in the middle of the tooth to reduce the axial leakage flux. The loss generated in the end core, the core side of the stator brush H-level insulating varnish to improve its insulation performance, further reduce eddy current loss.
Starting from the overall structure of the generator and the electromagnetic design, and taking into account the reduction of the end heat of the generator, a reasonable length of the stator and rotor core, the involute cone angle of the stator coil and the short circuit ratio are adopted.
Compared with QF-75-2-10.5 type generator, QF-135-2-13.8 type generator has the same cooling mode, but the former line load is 10.5% larger than the latter, and the stator inner diameter is 12.8 larger than the latter. %, according to qualitative estimation, the additional loss at the end of the same structure will increase by about 1.6 times, and the temperature rise will increase the same multiple under the same cooling condition, although both theoretical calculations and tests indicate that QF-75-2-10.5 two-type power generation The end loss and heat generation of the machine are not large, but if the QF-135-2-13.8 type generator adopts the same end structure, the QF-135-2-13.8 type generator is very likely to have a mountain end temperature. The problem of high rise is particularly prominent when the generator is running in phase. To this end, our company has adopted the T-type pressure ring + copper generator which is very mature on the 300MW hydrogen-cooled generator of our company at the end of QF-135-2-13.8 generator, adopting the "big one small pressure" two-layer structure. ) to further reduce the leakage of the end portion and improve its distribution, thereby reducing the heat source density of the end additional loss and local heat generation.
In order to grasp the magnetic field distribution and the additional loss distribution at the end structure, we carried out quantitative calculation and comparative analysis of the end magnetic field and end loss of the machine with and without copper shielding, and the temperature of the end structure. l estimate and other work.
2 Calculation principle The solution area is the end of the generator. Considering the contact gap between the pressure finger and the surface of the core, the center line of the access shaft is the z-axis, and the section about 40 mm away from the end of the rotor body and the guard ring is the R-axis. The upper boundary takes the inner wall of the end cover, the lower boundary takes the surface of the rotating shaft, the left boundary takes the stator core side and the rotor body side, and the right boundary takes the inner wall of the end cover. Solve the field as a traveling wave field. Assuming that the end magnetic field changes in a sinusoidal manner in the circumferential direction, considering the eddy current effect in the structural member (press ring, pressure finger, copper shield), the saturation of the medium is ignored, so that the equation is transformed into a quasi-three-dimensional eddy current equation. The stator and rotor current layers and the air gap current are considered as external currents on the outer boundary. Outside the solution region, the boundary conditions of the equation are determined, the vector magnetic position A is introduced in the eddy current region, and the quadratic isoparametric finite element is used in other regions using the scalar magnetic position. The method is discrete, and its linear equations are solved by Gaussian elimination.
2.1 Mathematical model See the end solution area for end eddy current field analysis and calculation.
End solution area 2.2 field equations and boundary conditions eddy current zone (pressure finger, pressure ring, copper shield) free space inner boundary outer boundary fixed form outer boundary (Core-rotor 2.3 calculation program (C0FSE) block diagram and geometrical connection parameter unit angle Section s coordinates modify the coefficient matrix according to the boundary conditions: calculation program frame. 4 edge segment core loss estimation due to the nonlinearity and anisotropy of the edge core, for the simplified calculation, the mathematical model does not include it in the solution area, and the loss can not be Directly obtained by the COFSE program, in order to obtain the above loss, we make the following basic assumptions: considering the saturation characteristics of the ferromagnetic material, the axial leakage flux is distributed in the circumferential direction after entering the edge core, when the first segment of iron After the magnetic density of the heart reaches the saturation value, the reluctance increases rapidly. At this time, a part of the leakage flux passes through the gap between the core segments into the first segment core, and the remaining leakage flux enters the leakage magnetic field of the second segment core. Make a circumferential distribution.
The circumferential magnetic density is evenly distributed in the core section - regardless of the eddy current loss caused by the axial leakage flux entering the edge core, only the hysteresis and eddy current losses caused by the uniform distribution of the magnetic density along the circumference are considered.
Under this assumption, COFSE is solved to obtain the axial magnetic density (Bzl, Bz2, Bzn) on the surface of the core as a known child. We have compiled the core loss estimation process factory f-(CORELOSS), which can be estimated. The eddy current loss induced by the leakage magnetic field on the core of the braider.
3 Calculation and analysis 3.1 Main parameters 3.2 Calculation condition Short-circuit operation T: l: Condition: UN, 丨f=1075A rated load operation 丨: Condition: p=pn, u=uN, (leading), lf=994A3.3 grid In the calculation of the split, take 111 quadrilateral isoparametric elements, 155 nodes. For meshing, see 137 quadrilateral isoparametric elements, 179 nodes, grids in the QF-丨35-2-13.8 generator end calculation. See through.
Table 1: Calculated value of axial magnetic density on the surface of the iron core entering the edge (GS) Copper shielding Nothing or not, no presence or absence of no-load short-circuit, rated load, phase-in-phase copper shielding, presence or absence, presence or absence of Load into phase load short-circuit rated load phase-in-phase copper shield with or without pressure-free type finger-type edge type core type é” shield type end additional loss type Table 1 lists the axial magnetic density calculation of the core surface entering the edge section Value (for edge loss estimation), Table 2 lists the calculated values ​​of the end structure loss. (To facilitate the analogy calculation, Table 1 also lists the QF-75-2-10.5 generator end. Structural component loss calculation results). Ab shows the compression ring, the calculation node on the copper shield, the unit number, and ~1) respectively show the distribution of the axial magnetic density on the pressure ring when there is no copper shielding, showing the axial magnetic density. Distribution on the copper shield. Ad shows the distribution of axial magnetic density on the pressure ring when there is no copper shielding under various calculation conditions, and 0a0b shows the distribution of the unit loss density on the pressure ring when there is no copper shielding, 1 shows The distribution of cell loss density on the copper shield.
Seven: a calculation ring on the pressure ring, unit number b copper shield on the calculation node, unit number + short circuit rated load phase calculation node number a axial magnetic density distribution on the pressure ring (no copper shielding) Platinum + no load Short circuit rated load phase calculation node number + short circuit rated load 4-phase axial magnetic density distribution on copper shield calculation node number no load (with copper shielding) no load (no copper shielding) + short circuit (with copper shielding) + Short circuit (without copper shielding) a. With or without copper shielding, the distribution of axial magnetic density on the pressure ring is compared with the figure. The distribution of axial magnetic density on the pressure ring when there is no copper shielding. Negative negative > fixed shielding fixed amount screen front screen c with or without copper shielding axial magnetic density on the pressure ring on the pressure comparison phase (with copper shielding) phase (no copper shielding) d with or without copper shielding Axial magnetic density distribution on the pressure ring Comparison calculation unit number - short circuit rated load phase 0a pressure ring unit loss density distribution (no copper shielding) calculation unit number + no load ~ short circuit rated load + phase 0b pressure ring Unit loss density distribution (with copper shielding) no load - 1 + short circuit, rated load; i + Phase calculation unit No. 1 Element loss density distribution on copper shield 4 Calculation result analysis From the axial leakage magnetic field causing eddy current on the surface of the end structure member, when there is no copper shielding, a large axial magnetic density component (hereinafter referred to as " The magnetic density ") acts on the area where the 133, 134 nodes are located (a), which causes the location of the 114 unit to become the largest heat source density area on the pressure ring, resulting in the largest loss density value (0a); the area where the 135142 node is located The magnetic density, in turn, also causes large loss densities at units 115, 120, 123, and 126. That is to say, when the QF-135-2-13.8 air-cooled steam turbine generator has no copper shielding at the end, there are many magnetic dense local concentration phenomena on the outer surface of the entire pressure ring, which forms a large surface eddy current. After the copper shield is applied to the pressure ring, the amplitude of the magnetic density at these nodes and the value of the unit with a large loss density are greatly reduced (b, 丨Ob), and the magnetic density and loss density of the outer surface of the entire pressure ring The distribution has become very uniform, although the magnetic density at 142 is slightly increased, but its absolute value is not high and the attenuation is fast. This has a large loss at the loss of 126 units at the node. Reflected in the decline.
From the value of the end loss caused by the surface eddy current formed on the end structure of the end leakage magnetic field (Table 2), the calculation result is completely in accordance with the calculations under the calculation conditions. Phase > Short Circuit > Rated Load
> The traditional qualitative law of "no-load". Under load conditions, the additional loss at the end of the copper shield is about 55% lower than that without the copper shield; the loss of the press ring is 45 times lower than that without the copper shield. After the protection of the copper shield, the shield enters the pressure. The amplitude of the axial leakage magnetic field at the end of the outer surface of the ring has been greatly reduced, and the pressure ring is effectively protected. The eddy current loss on the outer surface only accounts for about 25% and 30% of the total additional loss, and the copper shield takes 50% and 55%. It can be seen from a9d that the shielding range of the selected copper shield is extremely effective, and the magnetic density on the ring after numbering on M3 is basically unchanged before and after the installation of the copper shield and the absolute amplitude is low. The magnetic leakage amplitude of the outer surface of the pressure ring (node ​​133141), which has a great influence on the axial leakage magnetic flux at the end, has a significant wide range of attenuation after the copper shield is added, and the magnetic density of the outer surface of the entire pressure ring is made. And the distribution of loss density becomes very uniform. As far as the copper shield itself is concerned, the axial magnetic density on the outer surface is higher, especially the absolute amplitude at the 159 node is the largest, and the 160 node is the second, which results in the loss density of the 131 unit being the largest. , 132 units (1), but due to the non-magnetic, high conductivity of the copper shielding material itself, the absolute value of the eddy current loss is not very high. Nevertheless, since the copper shield bears about half of the end loss, in order to minimize the heat source of the QF-135-2-13.8 air-cooled turbo generator, we still have the overall design of the generator, especially the ventilation structure design. Effective cooling measures are required to improve and enhance the heat dissipation conditions of the copper shield itself. '5 end temperature rise prediction The leakage magnetic field at the end of the generator is a rotating magnetic field. Its axial component has relative motion with the stator core and the end structural members, and it is inevitable to produce hysteresis on the end metal structural components. And eddy current losses, which in turn cause heat to increase its temperature. The increase in the temperature rise of the stator end depends not only on the axial leakage density of the end portion, but also on the end cooling condition, and is related to the end structure and the material properties of the member.
It is still difficult to directly calculate the temperature rise of the end of the generator by using the temperature field. However, we can obtain the leakage magnetic field distribution at the end of the generator and the loss of the end structure by the calculation of the eddy current field at the end of the generator. The measured data of the end temperature rise test of the cooling mode and other models with similar end structure, from the perspective of engineering calculation, according to "the structural loss is proportional to the temperature rise"
The principle of the highest temperature rise of the type 8 turbine generator is liter (K) short-circuit rating. It is more conservative to predict the temperature rise of the end of a certain type of machine under the same cooling condition.
The method is consistent with the cooling method of the QF-75-2-10.5 steam turbine hair dryer that our company has tested in the factory. The QF-75-2-10.5 air-cooled steam turbine generator factory Real machine test end temperature rise test data summary, Oriental Motor Co., Ltd. internal information, 2000
Culture Paper,Cultural Paper,Custom Paper,White Printer Paper
CHINAPAPER-ONLINE(SUZHOU) E-COMMERCE CO. LTD , https://www.chinapaper-online.com