NAWAPA's Sawtooth Lift, which transfers the collected northern flows from the Columbia Basin into the Great Basin, consists of 6 separate pump lifts, for a total lift of 2450 feet, and which must pump a maximum flow rate of 85,500 cubic feet per second flowing through each step, requiring 26GW of power. The following is an in-depth proposal by Industrial Engineer, Terry Bates, on how this Sawtooth Lift could function, with its implications for industry.
By Terry Bates
The magnitude of the Sawtooth Lift is mind boggling and the engineering challenges are daunting indeed. The pumps alone would be the largest ever made in terms of head and volume; on the order of 100,000 to 125,000 HP each and 750 cfs at 700 feet head. Fortunately, the individual lifts are such that complex multistage pumps will not be necessary. Thus a flooded suction, single volute design, will keep the costs down. Constraints limiting maximum pump size include, casting and machining limits, component weights, physical shipping envelope dimensions, pick weights (say 50 to 75 thousand pounds), motor limits, engineering, modeling, and testing.
Power generation sites should be proximal to the stations. At some locations, if the power plants are very close, the primary motive force may include a combination 13.8KV synchronous motors and steam turbine-reducer combinations coupled to the pumps.
Both systems offer advantages as well as some drawbacks. The vertical shaft, low speed electric motor lends itself well to direct coupling with the vertical pump configuration, which most of these pumps are. And, of course, they don't require expensive mechanical reduction. Starting is an issue. Most motors start across the line, and current inrush is very large to be sure. In some cases, Motor-Generator sets are used, but that's relatively arcane and probably wouldn't be used today, giving way to some sort of solid-state system. I'd like to see an electrical engineer weigh in here on this topic.
More Advanced Design Considerations
- Staging: If the complex is modular, as many of us have espoused, the first pumping station at the 700 ft. lift, could initially include 15 pumps at 750 cfs each. Thus, a “phase I” or initial start of the project, would be nearly 20,000 Ac Ft. / Day. There would be five manifolds serviced by three pumps each, thence merged to a large diameter (23-24ft) pipe running up hill to the next pumping station's reservoir or tunnel entrance(s). This staged approach would serve as a useful test bed as well as demonstrate proof of concept.
- Throttling: Some sort of modulating or throttling mechanism will be required between the six lift stations. This could be accomplished through the use of throttling steam turbines or, in the event of electric motors, a solid state speed control drive. Such devices could be incorporated as part of the starter in one or more motor circuits. They could either be variable frequency (VFD) or pulse width modulated (PWM). I am unsure how these drives might affect reactive power (KVAR - power factor), or noise on the high voltage transmission lines. With a throttling turbine, during the design stage, great care would have to be taken to avoid resonate harmonics, inherent in most high speed rotating machinery, anywhere within the anticipated operating spectrum.
- Speed Reducer: Additional concerns include a reliable speed reducer from turbine operating speeds to pump speeds; i.e. approximately 3600 RPM to 600 RPM, or 6:1. These reducers would be rated at at a whopping 100,000+ HP! To minimize alignment problems, the first stage 3:1 reducer could be attached to the turbine similar to a NEMA-C configuration, but at a very grand scale indeed. The reducer would doubtless be of planetary design, double herring-bone, hunting tooth combination with integral oil pumps, filtration, cooling, and appropriate instrumentation. The second stage would be a 2:1 right angle drive direct rigidly coupled to the motor shaft with an allignable coupling/drive-shaft from the first stage reducer LSS. Again rated at more than 100,000 HP with appropriate torque design.
- Benefits of Steam: Regardless of the pump drive component, starting or stopping any pump would require notification to the power-house to expect a 710,000 lb/hr load swing (assuming a water rate of 9.5 lbs/kw). An advantage of steam turbo-generation throughout NAWAPA, is the proximity of virtually limitless cold water to allow fully condensing turbines. This would insure that condenser vacuums of 1” of mercury could be maintained, thus dramatically increasing efficiency. Consequently, a single stage regenerative cycle results in very superior thermal performance from a reduction of rejected heat to the condenser to the minimum value of about 7700 Btu/Kwh.
- Advantage of Electric Driven: Electrically driven pumps have advantages too. Certainly if the power house is not immediately adjacent to the pumping facility. Energy is easily transmitted through 250 or 500KV lines to the unit sub-stations, and thermally insulated high pressure steam piping and condensate returns are not an issue. As previously mentioned, with the advent of high ampacity solid state switching devices, VF or PWM drives allow for variable speeds.
- RPM Relationships: As a note aside, and has already been referenced in Burnett's excellent treatise on energy calculations, pump behavior is such that flow is directly proportional to changes in impeller RPM, pressure varies as the square of RPM changes, and most importantly, power is a cube relationship with RPM changes. The bottom line is; the basic pump design capacities should be at 90% maximum motor or turbine speeds, thus allowing some variation in output - up as well as down. It goes without saying, any dynamic changes in output should be very slow and deliberate.
The complex nature of the Sawtooth system, I believe, strongly argues for a single, very sophisticated (PID loop control), centralized Control Station. This secure station would have limited entry and be manned 24/7 (four shift basis). In addition to the “ main control room” and requisite office spaces, it should include motel-like overnight sleeping facilities, a staffed kitchen, cafeteria, recreation-TV room and, of course, standby diesel emergency generators.
Two-way communications with pump stations could be via micro-wave towers and/or buried fiber optics, again redundancy, security, and reliability are of utmost importance. In addition to telephone, voice communication would be by radio as well.
Parameters monitored and controlled would include, but not be limited to, individual pump output, combined station output, water levels, changes and trends at source reservoirs, future demands, mechanical health of rotating components (temperatures, vibration, fluid levels, etc), valve position and control thereof, current draw and/or steam consumption, on site personnel clocked in, and so forth. The importance of this facility cannot be overstated.
Irregardless of the Control Station, each pump station will have its' own staff as well, including maintenance, operations, and management personnel necessary for the day to day activities and to assume manual control if required. Again, each pump station would have limited and secure access.
Again, this is my view (and some others as well) of the operations of the Sawtooth Lift. I welcome your comments, criticisms, and additions to this abbreviated treatise. Comments from those of you with specialized engineering skills are especially encouraged.