Motor-CAD End Winding Spray Cooling Model

Transcription

1 Motor-CAD End Winding Spray Cooling Model Description Motor spray cooling is where the end winding is cooled by passing a fluid down the shaft and then firing it at the end winding through nozzles at the drive and non-drive end of the motor. This document will outline how the spray cooling is modeled in Motor-CAD. The cooling method is probably the most difficult to set up in Motor-CAD as there is not only the cooling from the spray hitting the end winding, but the spray bouncing off the end winding and flowing over the endcaps, the inside of housing and the rotor before exiting the motor via a sump. The fluid flow is assumed to be as shown below: Calibration of the model using test data is highly recommended if possible. An example of the calibration process is given later. Page 1

2 Spray Cooling Model Setup The user should set a hole down the shaft to allow the fluid to pass the through the shaft nozzles. The hole also acts as a cooling surface. In [Input Data] [Cooling Options] set the [Spray Cooling] checkbox as shown below: Page 2

3 Now go to the [Input Data] [Spray Cooling] tab-sheet editor: Here the user sets the [Inlet Volume Flow Rate], the [Inlet Temperature] and selects the fluid type (either from the inbuilt fluid database or direct input of fluid properties). Details are given to the user of the fluid properties: Page 3

4 The user then set the number of nozzles (at the drive and non drive end) and the nozzle diameter: The [Fluid Volume/End Space Ratio] is a parameter of the correlation used and is not often changed. Details are given of the calculated Nozzle to Surface Distance and Target Axial Length (end winding overhang), which are also parameters of the correlation used. The correlation for spray cooling was found in the Electronics Cooling Magazine and is used in this case to calculate the cooling from the end-windings. As it was initially thought a bit risky to use such a correlation the following warning was put in the software: This was implemented in 2002, and since then the spray cooling model has been used in several successful projects. Calibration using tests is always recommended if possible for such cooling to obtain better accuracy. The user can alter the amount of fluid that rebounds from the end winding and goes over the static endcap/housing and rotating rotor surfaces. By default it is assumed that 50% goes over each: Page 4

5 The calculated heat transfer coefficient, surface area, fluid velocity, etc are given in the table to help the user calibrate the model. Calibration adjustment factors for all the surfaces heat transfer coefficients and local fluid velocity are by default set at 1: Not only is data given for the end winding cooling by the spray cooling correlation, but also for all other surfaces that see the spray cooling fluid (before and after it has hit the end winding). An enclosed channel correlation is used for the shaft sections. A flat plate correlation is used for the Endcap and Housing (both are assumed to be in series with the end winding spray). The velocity of the fluid in the shaft and spray are calculated from the shaft hole diameter and nozzle diameters respectively. Note that if water jacket and spray cooling is used together then a problem that the housing nodes that have both cooling types cannot be calculated at this time we neglect the spray cooling on such nodes. The velocity of the endcap and housing fluid is set equal to the spray velocity multiplied by the relevant local velocity multipliers as shown in the screen capture below: Page 5

6 Default Local Velocity Multiplier values of 0.5 and 0.1 are used for the endcap and housing respectively. These values are set up from experience, but if better data either from test or CFD is available then they can be adjusted. They can also be made variables in the calibration process. Only the dissipation from the spray on the end windings is drawn on the schematic at present as highlighted below: Page 6

7 The cooling from the other surfaces can be seen in the [Circuit Editor] or [Output Data] pages: Page 7

8 Page 8

9 Page 9

10 Page 10

11 Calibration Example This is an example of a very difficult spray cooling model, where we did not have a specific number of nozzles firing spray at the front and rear end windings. The oil entered a void in the centre of the shaft and was thrown off the rotor by rotation to hit the end windings. The Motor-CAD model assumes that the oil is passed down a hole on the shaft and is sprayed directly at the inner surface of the end windings via a user selectable number of nozzles set into the shaft. In this example the oil is passed down the shaft and is directed onto channels at the axial centre of the rotor lamination. From there it works its way to the axial ends of the rotor and is thrown off by centirifugal forces and hits the end winding. The spray coverage in the actual machine is large so a large number of nozzles in the model are chosen to have the same effect 100 nozzles at end of the machine was chosen. The velocity of the spray hitting the end windings is just a function of the spray flow rate in the model. However in this situation the velocity is a function of the rotational speed. We can make the Local Velocity Multiplier in the model a function of speed to account for this. The figure below shows the Motor-CAD spray cooling editor with the local velocity multiplier used to calibrate the model. Oil Spray Cooling Editor showing the Velocity Multiplier used to calibrate the model Page 11

12 Spray Velocity Muliplier Calibration was carried out using a set of 30 tests on a prototype machine (varying speed, flow rates, etc). The measured losses were fed into the model that had already been calibrated for no oil cooling and the spray Local Velocity Multiplier varied until the correct winding temperature was predicted in each case. This process was automated using a Motor-CAD script routine. The figure below plots the values of spray velocity multiplier required for each of the 30 tests. It shows a definite trend of an increase in spray velocity multiplier with speed. We have plotted a liner line through the data points. The linear equation is given below: Local Velocity Multiplier = x RPM This equation is then implemented in a Motor-CAD script using the inbuilt script editor to set the Local Velocity Multiplier as a function of speed for all calculations. Mk2 Spray Vel Multiplier y = x R 2 = Speed [rpm] Plot showing variation in Spray Velocity Multiplier required to match measured temperature data for 30 test values Page 12

13 More plots showing how the spray velocity, spray heat transfer coefficient, spray dissipation and oil temperature rise vary with rotational velocity and oil flow rate are give in below. The graphs show that the spray cooling is a primary function of rotational speed and a secondary function of oil flow rate. 1.6 Spray Velocity Mult v Speed & Oil Flow Rate 1.4 Spray Velocity Multiplier flow rate = 1 flow rate =2 flow rate = 3 flow rate = 3.5 flow rate = 4.6 flow rate = 5 flow rate = Speed [rpm] Oil Spray Velocity Multiplier v Speed & Oil Flow Rate Page 13

14 Spray Velocity v Speed & Oil Flow Rate 1.0 Spray Velocity [m/s] flow rate = 1 flow rate =2 flow rate = 3 flow rate = 3.5 flow rate = 4.6 flow rate = 5 flow rate = Speed [rpm] Oil Spray Velocity v Speed & Oil Flow Rate Spray h v Speed & Oil Flow Rate 700 Spray h [W/m2/C] flow rate = 1 flow rate =2 flow rate = 3 flow rate = 3.5 flow rate = 4.6 flow rate = 5 flow rate = Speed [rpm] Oil Spray Heat Transfer Coefficient v Speed & Oil Flow Rate Page 14

15 Spray Dissipation v Speed & Oil Flow Rate 1200 Spray Dissipation [W] flow rate = 1 flow rate =2 flow rate = 3 flow rate = 3.5 flow rate = 4.6 flow rate = 5 flow rate = Speed [rpm] Oil Spray Dissipation v Speed & Oil Flow Rate Spray Fluid Tout v Speed & Oil Flow Rate Spray Fluid Tout [C] flow rate = 1 flow rate =2 flow rate = 3 flow rate = 3.5 flow rate = 4.6 flow rate = 5 flow rate = Speed [rpm] Oil Spray Fluid Temperature Rise v Speed & Oil Flow Rate Page 15