One number belongs to the piping and the tank. The other belongs to the pump. Cavitation starts the moment the system's number stops covering the pump's.
NPSH-A (Net Positive Suction Head Available) is a property of the installation — the tank elevation, the atmospheric or vessel pressure, the liquid's vapor pressure at its pumping temperature, and the friction loss in the suction piping between the source and the pump flange. NPSH-R (Net Positive Suction Head Required)is a property of the pump itself — a curve the manufacturer measures on a test stand and publishes alongside the head-vs-flow curve, describing how much suction head that specific impeller design needs at each flow rate to keep the local pressure at the impeller eye above the liquid's vapor pressure. Neither number means anything about cavitation risk on its own. What matters is the comparison between them, at the flow rate you actually run.
NPSH-A is computed from the installation you built: NPSH-A = Ha ± Hz − Hvp − Hf— the absolute pressure head above the liquid surface, plus or minus the elevation head between that surface and the pump centerline (positive for a flooded suction, negative for a suction lift), minus the vapor pressure head of the liquid at its pumping temperature, minus the friction head lost in the suction pipe, strainer, and fittings on the way to the pump. Every term in that equation is something you, the piping designer, control or measure. NPSH-R comes from none of that — it's measured by running the bare pump on a test loop and throttling suction pressure down until the pump's developed head drops by a conventional 3%, then recording the suction head at that point. It depends only on the impeller and volute geometry and the flow rate, and it climbs as flow rate climbs because higher velocity at the impeller eye means a bigger local pressure drop before the fluid ever reaches the vanes.
NPSH-A and NPSH-R move in opposite directions as flow increases — NPSH-A declines because suction friction loss eats more of the available head, and NPSH-R climbs because the impeller needs more head to survive the higher local velocity at its own eye. That means a pump-and-piping system that looks perfectly safe at design flow can cross into cavitation simply by running at a higher-than-designed flow rate, with nothing about the pump or the piping having changed. The comparison has to be checked across the actual operating range, not just at one design point — and it has to be re-checked whenever anything that shrinks NPSH-A changes: a hotter process fluid raising vapor pressure, a fouled strainer raising friction loss, or a lowered tank level reducing static head.
No — NPSH-A being a positive number tells you almost nothing by itself. Cavitation risk is set by the comparison to NPSH-R, not by NPSH-A's sign. Worse, the published NPSH-R value itself is not the point where cavitation starts — it is conventionally defined at the 3% head-drop criterion, meaning the pump is already cavitating measurably by the time its total developed head has been measured to reach that published number. That is exactly why real designs require a genuine margin, commonly NPSH-A ≥ 1.1 to 1.3 times NPSH-R, or a minimum absolute margin such as 3 ft (about 1 m) per Hydraulic Institute guidance — never just NPSH-A > NPSH-R by an inch, and never just "NPSH-A > 0." That margin has to cover test uncertainty in the published curve, suction recirculation at part load, and the fact that NPSH-A itself shrinks over the life of the plant as strainers foul and fluids run hotter.
Explains the difference between NPSH-A (Net Positive Suction Head Available) — a property of the installation, computed from atmospheric or vessel pressure, elevation, vapor pressure, and suction friction loss — and NPSH-R (Net Positive Suction Head Required) — a property of the pump, published on its curve from manufacturer testing. Shows why NPSH-A declines and NPSH-R rises as flow increases, why cavitation begins at the crossing point rather than at zero, and why a real engineering margin above NPSH-R, not just a positive NPSH-A, is required.
Both terms have "NPSH" in the name and both are expressed in the same units (feet or meters of head), which makes it easy to treat them as two versions of the same measurement. They are not — NPSH-A is an installation calculation anyone with a P&ID and a fluid datasheet can compute by hand, while NPSH-R is an empirical result from spinning the actual pump on a cavitation test loop. Confusing them leads to comparing NPSH-A against a generic rule of thumb instead of the specific pump's actual published curve, or assuming a pump is safe because NPSH-A alone is a "big enough" number.
NPSH-A = H_a ± H_z − H_vp − H_f, where H_a is the absolute pressure head above the liquid surface, H_z is the static elevation head to the pump centerline (positive for flooded suction, negative for suction lift), H_vp is the vapor pressure head of the liquid at its actual pumping temperature, and H_f is the friction head lost in the suction piping. NPSH-R is measured by throttling suction pressure into an operating pump until its developed head drops 3% from the non-cavitating value, then recording the suction head at that point, repeated across the flow range to build the published curve. Because impeller-eye velocity rises with flow, NPSH-R climbs with flow; because suction friction loss also rises with the square of velocity, NPSH-A falls with flow — so the two curves converge, and can cross, at higher flow rates even in a system that looked safe at its design point.
Every centrifugal pump selection in a chemical or process plant requires checking NPSH-A against NPSH-R across the full expected operating range, not just at rated flow — including startup, turndown, and any flow above design that operators might actually run. Common real-world NPSH-A killers include running a hot fluid closer to its boiling point than designed (raising H_vp sharply), a partially clogged suction strainer or a scaled pipe (raising H_f), and a low tank level near the end of a batch (reducing H_z). Cavitation from an inadequate margin causes impeller pitting, noise, vibration, seal damage, and a drop in developed head and efficiency — damage that accumulates gradually, which is why marginal NPSH designs can run for months before failing.
NPSH-A (Available) is calculated from the installation — pressure on the liquid surface, elevation, vapor pressure, and suction friction loss. NPSH-R (Required) is measured by the pump manufacturer and published on the pump curve — it describes how much suction head that specific impeller needs at each flow rate. Cavitation risk comes from comparing the two, not from either alone.
Higher flow means higher fluid velocity at the impeller eye, which produces a larger local pressure drop before the fluid reaches the vanes. The pump needs more incoming suction head to keep that local pressure above the vapor pressure, so NPSH-R rises — often steeply — as flow increases.
Common practice is NPSH-A ≥ 1.1 to 1.3 times NPSH-R, or a minimum absolute margin such as 3 ft (about 1 m), following Hydraulic Institute guidance — never zero margin. This is because published NPSH-R is defined at a 3% head-drop criterion, meaning measurable cavitation is already occurring at that exact value.
The most common culprits are a hotter process fluid raising vapor pressure head, a fouled or partially closed suction strainer or valve raising friction loss, and a lower liquid level in the source tank reducing static elevation head. None of these require any change to the pump itself — they erode the safety margin purely on the system side.
Local pressure at the impeller eye drops below the liquid's vapor pressure, and vapor bubbles form. As those bubbles move into the higher-pressure region further into the impeller, they collapse violently — this is cavitation. The repeated micro-implosions pit the impeller and volute surfaces, produce audible crackling or gravel-like noise, cause vibration, and reduce developed head and efficiency.
Try our Chemical & Process Engineering Studio
More calculators, simulators, and guides for this discipline.