Cooling-Water Pump Cavitation at Commissioning
A cooling-water pump takes suction from an open tank whose water surface is 3.5 m below the pump centerline. The pump is rated 120 m³/h at 32 m head and requires NPSHr = 4.8 m at that flow. The suction line is DN100 schedule 40 steel, 14 m long, with a foot valve, two 90° elbows, and a gate valve; calculated friction loss at 120 m³/h is 1.8 m. The water temperature is 40 °C, at which the vapor pressure is 7.4 kPa and density is 992 kg/m³. Site atmospheric pressure is 101.3 kPa. During commissioning the pump produces a crackling noise, flow oscillates between 95 and 130 m³/h, and the discharge pressure gauge fluctuates. The contractor has already throttled the discharge valve to 60% to reduce flow, which lessened but did not eliminate the noise. The owner wants to know whether the pump is cavitating and what physical change will fix it before the tower is commissioned.
1. Calculate NPSHa at the design flow and compare it with NPSHr. Show each term in meters of water.
Read the worked answer
Use the suction-side energy equation in head form:
NPSHa = (p_atm − p_v)/(ρg) − z_s − h_f
Atmospheric pressure head: 101,300/(992 × 9.81) = 10.41 m. Vapor pressure head: 7,400/(992 × 9.81) = 0.76 m. Static suction lift: z_s = 3.5 m. Suction friction loss: h_f = 1.8 m.
NPSHa = 10.41 − 0.76 − 3.5 − 1.8 = 4.35 m.
The pump requires NPSHr = 4.8 m, so the available margin is 4.35 − 4.8 = −0.45 m. The pump is therefore operating with insufficient NPSH at design flow. The crackling noise, oscillating flow, and fluctuating discharge pressure are consistent with cavitation. A small positive margin is normally targeted, typically at least 0.6 to 1.0 m above NPSHr, so the design is deficient even before accepting any measurement tolerance.
2. Evaluate the contractor’s throttling as a remedy and explain why it may mask rather than solve the problem.
Read the worked answer
Throttling the discharge valve reduces flow, which lowers the pump’s required NPSH and reduces suction-line velocity and friction loss. At 95 m³/h, for example, the friction loss might fall to about 1.2 m and NPSHr might fall to roughly 4.0 to 4.3 m. NPSHa would then be about 10.41 − 0.76 − 3.5 − 1.2 = 4.95 m, giving a small positive margin. That explains why the noise lessened.
However, throttling is not a permanent solution. It wastes energy, moves the pump away from its best-efficiency point, and may create low-flow recirculation and additional vibration. The margin remains small and will disappear if the tank level drops, the water warms, the foot valve fouls, or the pump runs at a higher speed. Cavitation damage can continue even when the noise is reduced, and the pump may still be operating outside the manufacturer’s recommended range. Throttling should be treated only as a temporary commissioning measure while a suction-side fix is designed and installed.
3. Recommend two permanent modifications with quantitative effect, and state which you would choose.
Read the worked answer
Option 1: raise the tank or lower the pump so the static suction lift decreases. Raising the water surface by 0.8 m would change NPSHa from 4.35 m to 5.15 m, giving a margin of 0.35 m over NPSHr. Raising it by 1.2 m would give 5.55 m and a 0.75 m margin, which is a more defensible target. This is effective but may be expensive if the tank is existing.
Option 2: replace the DN100 suction line with DN150. For the same flow, the velocity falls from about 4.2 m/s to about 1.9 m/s, and the friction loss falls roughly with the fifth power of diameter for turbulent flow. The 1.8 m loss could drop to about 0.3 to 0.5 m. NPSHa would then be about 10.41 − 0.76 − 3.5 − 0.4 = 5.75 m, a margin of about 0.95 m. This is usually the better choice because it also reduces suction velocity and improves approach conditions.
A third option is to cool the water, since reducing the temperature from 40 °C to 30 °C lowers the vapor pressure from 7.4 kPa to about 4.2 kPa, adding about 0.33 m of NPSHa. That alone is insufficient. I would choose the DN150 suction line, verify the foot valve and strainer losses, and confirm the final NPSHa with a post-modification test.