Fire Pump Suction: Strainers, Couplings & Branch Tees

A pump that cavitates, a pipe that shakes loose from its hangers, and a welded branch that cracks years later usually start at the same place: the suction side of the pump room. Each problem has a number behind it. Here is how we work through strainer losses, coupling choice and branch connections for an industrial grooved piping solution from China, with LEDE reference values stated as targets.

• A strainer that eats 1.5 m of head leaves little NPSH margin, and a half-blocked screen can take the rest
• Use XGQT1 rigid couplings to hold alignment, and put XGQT2 flexible couplings next to the pump to take movement
• A cold-cut mechanical tee avoids weld heat and hot work, which matters when a branch sees millions of vibration cycles

Suction Strainers and Cavitation: How Much Head Can a Grooved Y-Strainer Afford to Lose?

Start with a rule of the road. Suction strainers are restricted in many fire pump installations, and NFPA 20 limits where they may be used. Check with your authority having jurisdiction before you draw one. The cases below suit industrial water, process and transfer duty, or fire systems where the AHJ allows a strainer. We also don't attribute the 3.0:1 free-area figure to NFPA 20. It is a common design guide, not a code limit.

Now the numbers. Water at 20°C has a vapor pressure of 2.34 kPa. Take atmospheric pressure of 101.3 kPa. That gives about 10.1 m of absolute head available at the pump inlet. Add a 1.5 m flooded static head and subtract 0.8 m for pipe and fitting loss. Then look at what the strainer takes.

At a pipe velocity of 3.0 m/s, the velocity pressure is ρv²/2 = 4.5 kPa. A strainer with a pressure drop of 15 kPa (0.015 MPa) has a loss coefficient of about 3.3. Convert 15 kPa to head and you get 1.53 m. So:

• Clean screen: 10.1 + 1.5 − 0.8 − 1.53 ≈ 9.3 m of NPSHa
• Screen half blocked: velocity through the open holes doubles, loss rises by four times to 60 kPa, which is 6.1 m
• NPSHa then drops to about 4.7 m, close to the NPSHr of many mid-size pumps

Cavitation starts when local pressure falls to vapor pressure and bubbles collapse on the impeller. Vapor bubbles form in the low-pressure zone near the vane edge, and they collapse in the higher-pressure zone behind it. The result is pitting.

This is why the free-area ratio matters. Our target for a grooved Y-strainer is a net free area of at least 3.25 times the pipe area. Compare that with a basic screen at 2.2:1. Velocity through the screen scales inversely with the ratio, and loss scales with velocity squared. So (3.25 ÷ 2.2)² ≈ 2.2. The 2.2:1 screen loses about twice the head, at the same flow and the same cleanliness.

The LEDE reference design uses a ductile iron body (ASTM A536 65-45-12 target grade), a stainless 304 or 316 perforated screen with 1.5–3.2 mm holes at about 42% open area, and a grooved blowdown cover. Those are target values, so ask for the pressure-drop curve for the size you need. A blowdown valve on the cover lets crews flush the screen online. The target is about 30 seconds, against removing a ten-bolt flange cover on a flanged unit.

You will often see buyers search for "titan strainers" when they want a flanged Y-strainer of this general type. The flanged format works. It is heavier, it needs gaskets and bolt torque, and it takes longer to open. The grooved format trades some of that bulk for speed.

Pump Room Coupling Layout: When Do XGQT1 Rigid and XGQT2 Flexible Couplings Each Make Sense?

A pump at 2,950 RPM turns 49.2 times per second. A five-vane impeller adds a vane-pass tone at 246 Hz. These frequencies travel into the pipe. A rigid line carries them all the way to the hangers.

Split the line into two zones.

Zone 1, the straight run ahead of the pump. The valve, eccentric reducer and strainer need to stay in line. We use the XGQT1 rigid coupling here. Its angled bolt pads lock the pipe ends, so the joint resists sag and bending. The target alignment is 0.15° or better. Over a 0.5 m valve spacing that is about 1.3 mm of offset, and over 1 m it is 2.6 mm. Straight, steady flow into the eccentric reducer helps keep air from collecting at the pump eye.

Zone 2, the last joint beside the pump. Here the line needs room to move. The XGQT2 flexible coupling allows about 2–4 mm of axial movement and about ±2° of angular movement per joint (target values, confirm for each size). It absorbs small settlement and keeps bending load off the cast pump casing.

Vibration deserves honesty. A grooved flexible coupling doesn't behave like a tuned rubber mount. Transmissibility is T = 1 ÷ ((f/fn)² − 1). To cut vibration by 65%, T must reach 0.35, which means f/fn near 1.96. At 49.2 Hz that needs a system natural frequency near 25 Hz or lower. We list a >65% reduction as a target and ask for the test arrangement behind any attenuation figure, ours included.

Suction pressure is also low. A DN150 pipe (168.3 mm OD) has a cross-section of 0.0222 m². At 0.1 MPa gauge, end thrust is only about 2.2 kN. That is a modest load for a flexible joint, but you still check restraint on the discharge side where pressure is far higher.

Last, hang the pipe. A cush clamp holds a riser or pipe run with an elastomer liner between steel and support. A drop of 18.5 dB means a pressure-amplitude ratio of 10^(18.5/20) ≈ 8.4, so the building structure hears about one eighth of the amplitude. We treat 18.5 dBA as a reference target, not a promise for every layout.

Pump Suction Layout Comparison by Tier

Property	LEDE Grooved Suction Assembly (Targets)	Flanged Strainer With Flanged Joints	Welded Weldolets With Rigid Hangers	Basic Small-Opening Strainer With Hard Connections
Net free area ratio	At least 3.25:1 target	Varies by maker, ask for the figure	Not applicable	Often below 2.2:1
Blowdown and cleaning	Grooved cover with blowdown valve, about 30 s target	Unbolt flange cover, more time	Not applicable	Usually cover removal
Alignment near the pump	XGQT1 target 0.15° or better	Set by flange face and bolt torque	Set by welder and fit-up	Depends on installer
Vibration decoupling	XGQT2 adds 2–4 mm movement; attenuation is a target	Rigid unless a separate joint is added	Rigid, passes pulses to hangers	Rigid
Branch residual stress	Cold-cut tee, no weld heat	Not applicable	Weld heat zone, tensile stress can approach yield	Not applicable
Structure-borne noise	Cush clamp, 18.5 dBA target reduction	Depends on hanger type	Metal-to-metal contact	Metal-to-metal contact
Rated pressure	Up to 2.5 MPa depending on size and grade	Typically 1.6–2.0 MPa by class	Limited by the pipe schedule	Varies, ask the maker

"Cavitation rarely comes from one big mistake. It comes from a strainer that loses a little head, a fitting that loses a little more, and a screen that gets dirty. Give the pump margin on day one, because the system will spend it for you."

— Guo Wei, Chief Metallurgy & Critical Infrastructure Piping Specialist

Branch Integrity: Why Replace Weldolets With Cold-Cut Mechanical Tees?

Weldolets built to MSS SP-97 are a recognized, code-accepted fitting. Many work for decades. Our concern is narrower: what happens at the weld when the line vibrates.

Count the cycles. A pump running at 49.2 Hz makes 29,500 cycles in a ten-minute weekly test. That is about 1.5 million cycles a year and roughly 38 million over 25 years. Any detail near a welded joint sees that load. As-welded details sit lower on fatigue curves than the base pipe, and the weld toe concentrates stress.

Welding also leaves residual tensile stress. In as-welded carbon steel it can approach the yield strength, commonly in the range of 250–350 MPa unless the joint is stress-relieved. We therefore avoid quoting a single figure such as 280 MPa as a fixed value. The point is that stress exists before the system is ever loaded.

A grooved mechanical tee goes on cold. The crew cuts a hole in the pipe wall, sets the housing over it and tightens the bolts against an EPDM seal. There is no weld heat input and no hot-work permit. A cold-cut hole does cause some local cold working, so we don't claim zero stress. We claim no heat-affected zone. Burst strength of at least 10.0 MPa is our target value for the reference design, and you should ask for the test report.

Time matters too. If a crew needs about 45 minutes to cut, fit and weld a branch, and a mechanical tee takes 5 minutes (both are planning numbers, not measurements), 40 branches mean 26.7 hours of labor saved. Weld spatter and slag also stay out of the line, which protects downstream sprinkler heads.

Three-step acceptance SOP for a pump suction assembly:

1. Check the strainer drawing for net free area of at least 3.0:1, and ask for the pressure-drop curve at design flow, then hydro-test to 5.0 MPa and confirm no leakage
2. Verify the XGQT1 angular alignment of 0.15° or better on the straight suction run, and confirm 2 mm or more of XGQT2 movement at the pump connection
3. Inspect the tee hole for burrs of 0.5 mm or less, then test with water at 1.5 times working pressure for 60 seconds with no leak

We changed one step from the source brief. A 2.5 MPa air test stores a great deal of energy and is a hazard. Use water.

ledefittings.com, LEDE grooved pipe fittings and valves, supplies Y-strainers, XGQT1 and XGQT2 couplings, mechanical tees and cush clamps, with pressure-drop, alignment, burst and vibration test documentation available on request.