Streamlining Nuclear Reactor System Balancing

In this month’s Pumps & Systems column, we will look at how simulating a large heat transfer system balance found in many industrial plants can be used to streamline plant operations. The system balancing ensures each heat exchanger is receiving the proper flow rate and pressure needed for the system to operate properly. Early in my career, I worked as a startup and test engineer at Millstone Power Station Unit 2 for Northeast Utilities. Millstone 2 was a Combustion Engineering 870 MWe nuclear power plant, and was much larger that the plants I had worked on in the Navy. One of the first preoperational tests I was involved in was balancing the Component Cooling Water (CCW) system for Millstone Unit 2. The system had a normal operating mode in which cooling water was supplied to various heat generating equipment in the reactor compartment and auxiliary machinery spaces. The system had Nuclear Regulatory Commission (NRC) licensing requirements to be operational when the reactor was producing power and when it was shut down. In addition, the system was designed to automatically provide cooling water to specific equipment during an off normal (post-accident) condition. Once all the equipment in the CCW system was tested, the entire system was placed in operation and then balanced so that each heat exchanger received the designed flow rate and pressure required for proper system operation. The system also had to be balanced for off-normal operation in the event of an accident. The CCW system was large, encompassing multiple pumps, heat exchangers, control valves and pipelines covering most of the plant. Since this was a safety-related system, the CCW balancing had to be completed prior to the NRC issuing an operating license, which would allow the loading of radioactive fuel. The test was planned to take approximately 10 days, with the actual plant testing to occur during the swing shift and midnight shift. The testing required most of the operational staff to operate the equipment, and a group of test engineers to set the balancing valves to their prescribed positions. After all the balancing valves were set to the prescribed positions, plant test data was collected for further analysis.

The results of the nightly tests were compiled and sent to Bechtel Power in Gaithersburg Maryland, which was the A&E firm for Unit 2. Their engineers reviewed the results to determine the flow rates through each of the CCW heat loads. Based on the test results they would generate a new table with the adjusted valve positions for each balancing valve to achieve the designed flow rate. The CCW system was a network of looped loads, and changing the position of any of the balancing valves affected the flow rates to all the loads. Each night, we would adjust the balancing valves in the CCW system to the new positions and then re-run the test, compile the results and submit them to the engineers in Bechtel. The above process was continued until the system was balanced and the flow rate through each heat exchanger was at its designed value. This test took approximately eight days to complete with many different iterations. Since it needed to be complete prior to the issuance of the plants operating license by the NRC, it received the highest level of interest and scrutiny from the management team.

Fast Forward

After 10 years of starting up and testing power plants for EBASCO Services, I decided to try something new. I became a co-founder of Engineered Software in 1982. PIPE-FLO was our first program to help simulate the operation of fluid piping systems. As time went on, more features were added to the program to allow for the simulation of larger fluid piping systems. During a support call, one of our customers at a nuclear plant mentioned they were modeling their plants CCW system with our piping simulation program. They needed to meet the needs of an NRC-mandated study of the plant’s cooling water system. After a few months, I followed up with our customer at the nuclear power plant and asked about their CCW study. They mentioned that they modeled their entire CCW system, and using the software they could effectively communicate the results of their system study to the NRC.

Going Full Circle

A few years later, I was conducting a training class on the use of our simulation software. One of the course attendees was a test engineer that I worked with from my Millstone days. For 10 years, he worked for Combustion Engineering and was involved with balancing CCW systems for numerous nuclear plants. He was working as the lead engineer for an operating nuclear power plant. During the class, one of the attendees asked what was the maximum number of pipelines the software could simulate. I stated there were no program limits and asked the attendees to comment on the largest system they had modeled. My friend from Millstone mentioned that he had created a model of their entire CCW system to help balance the system. After sharing our mutual experience with balancing the Millstone CCW system, he mentioned that he used the model to balance his current plant’s CCW system. After entering data for the piping system model, he set the flow rates through each heat exchanger using a throttling valve. The simulation calculated the differential pressure across each throttle valve in the CCW system to achieve the designed flow rate. Using the information provided by the supplier of the balancing valves, he calculated the valve position of each balancing valve. I asked how well that worked. He said that by setting the balancing valves for each loop to the calculated position, they were able to balance the entire system with just one change and completed the entire system in one day.

Conclusion

Most industrial processes have multiple piping systems to remove excess heat from operating equipment, provide chilled water and hydronic heating for HVAC systems or use heat tracing to warm fluids to operating conditions. Each of these systems needs to be balanced so each load meets its designed flow rate. If the system is balanced by hand, then multiple iterations must be performed to balance the system to the design values. Often, the system will operate with throttling valves open further than required to balance the system. This reduces the need to balance the system, but it requires a greater flow rate than needed, which increases the system’s energy cost and reduces the future capacity of the system. By employing a simulation of the piping system, the position of the system’s balancing valves can be calculated during the first pass, greatly streamlining the balancing process.