First-Hand:Creating Self Cooling in Switchgear Equipment

Submitted by Robert H. Rehder, LSM, Peterborough, Ontario, Canada, Sep 4, 2015

In the early 1960’s General Electric Canada had developed a three phase, 15 kV, 1200A rotary disconnect switch and they were actively trying to extend the rating to 2000A to meet a customer’s requisition requirements. The customer insisted on a self cooled design. The space occupied by the switch must be the same as the 1200A switch, and voltage test limits had to be maintained or surpassed. I was the design engineer responsible for the design and I reached the point that, with adding copper and cooling fins, the best we could do was 1870A. Dr. E. C. Elgar was doing the thermal tests in the Peterborough Engineering Lab. He finally did some tests to establish if fan cooling would solve the problem if we could convince the customer to accept it. The scheduled shipping date for the switchgear assembly was only a few months away.

On the weekend I was at home setting up the barbeque for hamburgers in the back yard and the little motor on the side of the barbeque to drive a spit caught my attention. That was the same kind of inexpensive motor that was used in small house fans. I got out my tools and took the motor off the barbeque and took off the motor’s covering to get a close look inside. It was a shaded pole induction motor with a single coil and operated from a house 110 volt outlet. The coil in the motor had a large number of turns of fine wire. The magnetic field in the iron laminations of the motor would be about 2000 ampere turns. The idea suddenly struck me that with one turn at 2000 amperes there would be a strong enough magnetic field surrounding the conductor to make the motor rotate and it could drive a fan. Now, how was I going to take the 2000A current flow going to the switch and somehow capture its magnetic field and direct it to the motor core? I saw that the coil and its core had been pressed into position to be part of the motors magnetic circuit. I pressed the core and coil out of the motor assembly. I took the motor to my basement workshop and measured the cross sectional area of the motor’s magnetic core where I had pressed out the coil. I then took some lengths of wood with the same cross section and made a rectangular frame that would be just big enough to fit around the outside of three insulated copper bus bars. I visualized that if the frame were ultimately made of laminated iron and 3 flat bus bars carrying 2000A were passed through it, the magnetic field in the iron frame would be 2000 ampere turns. I cut a gap in the long side of the frame just wide enough to slip the coil-less motor core into it to fill the gap. Now the magnetic field would pass through the motor core and the motor rotor would turn and could drive a fan.

On Monday morning I almost ran over to the Lab to show Dr. Elgar this motor idea. He thought it might work and he would put a temporary standard fan in the test assembly and determine how large a diameter of fan would be required to produce enough cooling air flow. I went back to the main shop to see Harry Tasker in the manufacturing area and asked him to go over to the fractional motors manufacturing area and pick up some scrap steel motor laminations to make three magnetic cores shaped like the partial wood frame. I showed him the wood frame and described my motor idea.

I had noted the name of the company that had made my barbeque motor so when I got back to my desk I contacted the purchasing people and asked them to buy three motors of the same size and type and ship them as quickly as possible to Dr. Elgar’s attention in the Lab for tests. In the meantime Dr. Elgar finished a heat run at 2000A using a conventional desk fan and we could comfortably meet the temperature rise limit with a 4 inch diameter fan. He had 3 extra fan blades in the Lab and we could use them for our final tests.

I wrote down just what my idea really was and added the following advantages:

1. Simplicity – The current in the circuit which is causing the heating directly actuates, through magnetic coupling, the motor rotor and fan.
2. No external source of supply or associate control power transformer is required. Normal 110 volt or similar circuits have questionable reliability unless fed directly from control power transformers fed from the high voltage bus work which is usually considerable extra expense.
3. No temperature sensing or current relays are required for control.
4. Failure of a fan is usually due to a field coil failure in the motor. In this concept the field coil is the major bus work which has the same reliability as the overall switchgear.
5. Multiple fans, such as one per phase, add to reliability as the equipment could probably operate with only a slight reduction in load using the remaining two fans. Conventional forced air cooling systems usually have one blower with duct work to direct the air to the hot spots.
6. The cooling fan can be located on the high voltage connections close to the hot spot. Conventional blower systems usually require duct work to direct the air to the source of the heat. This new idea has a particular advantage on high voltage systems or sealed systems where connecting to external controls or fans is difficult and expensive.

Then there were 2 disadvantages:

1. The customer has expected a self cooled switch. Could they be convinced that the switch was self cooled considering that there are no external sources of power and no dependency on temperature sensors or control systems? I felt that I could convince them.
2. The shaded pole induction motor has two sensitive parts. The magnetizing coil with its many turns of fine wire, and its bearings. In this case the bus bars are the coil conductors and they are rugged. Therefore the only disadvantage would be the bearings and these could be life time lubricated bronze bearings to minimize possibility of failure.

I finally decided that I was going to go ahead on the basis of using this fan idea to get the 2000A rating.

A day later, I received a call from Harry saying that he had made the steel for the fan cores and they had been given to Dr. Elgar in the Lab. I immediately headed for the Lab. I had decided that the fans should start at about 1200A and be up to full speed at 1800A. When I got to the Lab, the technicians had already knocked the wire wound coils out of the new purchased motors and were busy assembling the core frames around the copper bus bars feeding power to the switch. With the cores in place they then fastened the cores up against the motor cores to complete the motor’s magnetic circuit. The 4 inch fans were fastened to the motor shafts and then the motor core was tipped slightly so that the flow of air from the fan would be aimed at the middle of each phase of the switch assembly.

Bill Jones, turned on the strobe light so we could monitor the speed of the fan, then turned the power on and brought the current up slowly so that we could observe when the fans would start to rotate. I made a plot of the current versus fan speed.

At 300A one fan started and the other two started at 320A and 330A. The initial speed was 1000 revolutions per minute (rpm) and then as the current increased the fan speed increased and reached 3500 rpm at 1200A. The motor was a two pole induction motor so with a 60 cycle power supply, 3500 rpm would be the maximum design speed.

We left the power on at 2000A to check out the temperatures when they leveled off. It looked like we had too much magnetic field in the fan and it was starting to turn at too low a current. We could fix that by machining some metal out of the core frame and this would increase the reluctance or resistance to the magnetic field and reduce it. This would increase the current level required to start the motor rotating.

The temperatures leveled off in about 3 hours at 2000A and the switch was running well below the desired levels but the fan motor bearings and core were running too hot because of the strong magnetic field. Dr. Elgar did some calculations and when he was done he concluded that we had a 2500A switch with those fans running at that speed.

The magnetic field in the core was too high but we experimented and cut some metal out of the frame a bit at a time and reduced the field until the motor start point was at my suggested 1200A figure. This reduced the temperatures in the motor and bearings but had almost no effect on the air flow rate at 2000A.

The fans were simple and I believed that I could convince the customer that the assembly was self cooled as no external power supplies or controls were required. I guess the first thing I had to do then was to convince my own managers that these little fans were the way to go to get the 2000A rating.

I reviewed the fan concept with my managers and showed them pictures and test results and they agreed that I should go ahead and meet with the ultimate customer and ensure that the fan system will meet their specification words of “self cooled”. Actually we could have rated it at 2500A with the extra copper we had in the switch.

Figures from Canadian Patent # 782,673

I expected the customer would say that the switch operating at rated voltage and current must stay within the temperature limits without any other external energy sources or controls required to stay within these temperature limits.

It had taken us 26 heat runs to develop the final design and we hadn’t delayed our shipping schedules.

The meeting with the customer was scheduled a week later and we convinced him that the switch was a self cooled switch and it would not reduce the reliability of our switchgear, and we would meet the shipment date. The customer was very impressed with the rotary switch and the novelty of the fan assembly

I wanted to yell “Whoopee” but I settled for a hand shake with the customer and the words, “Thank you for your approval of the switch and its cooling system and I am sure it will perform well and outlast the rest of the equipment.

After the customers had left I went over to Dr. Elgar and the technician and thanked them for their effort because their enthusiasm and pride in their development work was obvious and this had had a positive influence on the customer.

In the following years I used these small motors to “self cool” cable terminal boxes, and to eliminate stratification overheating in long runs of vertical or sloping runs of isolated phase bus duct. In the bus ducts the conductors are large diameter tubes and I modified the magnetic frame around the outside of the conductor and mounted the fan and motor down inside the conductor to circulate the air through the conductor and back down the space between the conductor and the enclosure.