What can be done to increase the flow?

After my four-year tour in the US Navy, I went to work for EBASCO Services in New York City. A little history, EBASCO (short for Electric Bond and Share Company) was started by General Electric in 1905 as a holding company to finance electrical utilities. They financed, designed, built, and in some cases operated the power plants for electrical utilities around the world. The company was restructured after the passage of the Public Utility Holding Company Act of 1935. After that, they focused on designing and building power plants and electrical distribution grids. 

In 1982, I was working for EBASCO at WNP-3 for the Washington Public Power Supply System in Olympia WA. As a startup engineer, I was assigned to a crew involved in flushing the lube oil system for the main turbine. After assembling the lube oil system, it must be cleaned of all dirt and debris introduced during the construction process. The flush is a detailed process that has exacting requirements of cleanliness and must be completed before the main turbine can be placed in operation. 

At this time, I was working for EBASCO during the day and developing PIPE-FLO® at night with Carolyn. My boss, Pat McHale, asked if I could look at the various flush paths within the lube oil system to determine the flow rate and resulting fluid velocities. This was the first time PIPE-FLO® was used in a working environment. High fluid velocities are the key to flushing a system of dirt and debris. I simulated the lube oil system with PIPE-FLO® then evaluated the proposed flush paths with the software. Using the results, the flush team discovered ways to increase the flow rate and resulting fluid velocities when writing the flush procedure. 

One of the key items in the lube oil flush is cycling the temperature of the lube oil. Heating and cooling the oil causes the pipes to expand and contract, which helps dislodge construction dirt and debris from the inside pipe walls. The flowing oil carries away the contaminants where a filter removes them. In addition, to help speed up the process an external oil-cleaning skid was designed by the client and used to heat the oil and provide extra filtration. This cleanup skid was connected to the lube oil reservoir and was continually operated during the flush process. 

The oil cleaning skid was designed by the client and consisted of two paths of heat exchanger, filters, pumps, and control valves. An external boiler provided steam to the heat exchangers to heat the lube oil during the heating cycle. During the cooling cycle, the skid still had lube oil flow through the filters but steam was not supplied to the heat exchanger. 

When we started the lube oil flush the system heated up quickly and we got the design flow rate through the system. During the cold cycle, the flow rate was only a quarter of what was expected. Once again, I was asked to look at the system with PIPE-FLO®.  After modeling the system and performing the hot oil calculation, the model confirmed the system was operating as designed. During the cold cycle, the calculated flow rate through the model also correlated with the observed values on the skid.

While building the model, we discovered the heat exchanger was a three-pass heat exchanger. In addition, the control valves on the skid were of a reduced seat design. 

With this information in hand, we had a conference call with the WPPS team that designed the skid along with the manufacturers of the heat exchanger and control valve. In our discussion with the heat exchanger manufacturer, he stated replacing the head of the heat exchanger could change it from a three-pass to a single pass heat exchanger. That change would reduce the head loss across the heat exchanger in the cold condition and would not adversely affect the system under the hot oil condition. We evaluated the system with this proposed change with the model and discovered it would greatly increase the flow rate through during the cold cycle. The system would have to be down for four to five days for the manufacturer to make the changes in the heat exchanger heads. It was decided having the cleaning skid down for that long would influence the schedule. 

We also looked at the control valve. The supplier said that replacing the reduced seated trim with a full-seated trim would increase the flow rate through the valve under the cold oil condition. He stated the full seated valve should still be able to work correctly in the hot oil condition. We modeled that change to the system and discovered the flow rate in the cold condition would be approximately 75% of the design flow rate. As the cleaning skid would only take the system out of commission for two hours, we decided on this option.

The changes were made to the skid system, and the flow rate increased. The lube oil flush took a total of 30 days to complete, which was about 60% of the time on the schedule. The PIPE-FLO® model provided us with the ability to simulate the system, discover what was happening, and try alternatives quickly without great expense. 

Was PIPE-FLO® the reason we were able to save 15 days from our schedule? No, we had an excellent craft on the job site that maintained high cleanliness standards when building the system and the contractor did an excellent job of maintaining clean conditions during construction. We had a well thought out flush procedure, excellent start up engineers who discovered the low flow problem early, and managers who were interested in improving the process. PIPE-FLO® was one tool that helped provide a clear picture with how the system was operating and what could be done to improve the system. Not bad for the first time the program was used on an actual system.

Back to ESI News Brief