ASME BPE Hygienic Piping Design — Dead Legs, Drainability and CIP Velocity

Cleanability in a hygienic system is decided by geometry long before it is decided by cleaning chemistry. ASME BPE addresses that geometry through three linked requirements — the dead-leg L/d ratio, gravity slope designations for drainability, and a minimum clean-in-place velocity — plus spray coverage on the vessel side. This guide covers what each one is protecting against, how they interact, where designs typically fail them, and the free calculator for each check.

Content last reviewed 2026-07-31 · page regenerated 2026-07-31 at build.

Cleanability is a geometry problem

The premise of hygienic design is uncomfortable but simple: you cannot clean your way out of bad geometry. A clean-in-place circuit delivers chemistry, temperature, time and mechanical action, and only the last of those depends on the pipe. If the geometry prevents mechanical action from reaching a surface, no amount of caustic and no extension of contact time reliably fixes it — the flow simply does not go there.

Three geometric failures account for most of it, and ASME BPE's design requirements map onto them almost one-to-one:

• Unswept volume. A branch, a sample point, an instrument tee or a spare nozzle deep enough that the main flow's recirculating eddy does not reach its end. Product sits there, is not exchanged during cleaning, and is diluted rather than removed. This is the dead-leg problem.
• Retained liquid. A run that does not fully drain leaves pools of product, rinse water or cleaning solution. Retained water is a growth site; retained cleaning solution dilutes the next charge and can carry over into product. This is the slope and drainability problem.
• Insufficient wall shear. Even a perfectly swept, perfectly draining line will not clean if the flow across its wall is too gentle to lift soil. This is the CIP velocity problem.

A system can pass any one of these and fail on another, and the failures are not interchangeable — a dead leg is not fixed by more velocity, and a flat run is not fixed by a shorter branch. They are separate checks because they are separate physical mechanisms, and the calculators below are separate for the same reason.

Dead legs — the L/d ratio and the target of 2

ASME BPE SD-3.1.2.2 frames the dead-leg criterion as a ratio rather than an absolute length, which is the right instinct: what matters is not how long the branch is, but how long it is relative to its own bore. The recirculating eddy driven by flow in the main run penetrates into a branch to a depth that scales with the branch diameter, so a wide short branch can be swept while a narrow branch of the same length is not.

L / d ≤ 2

Measuring it correctly is where designs go wrong. L is measured from the inside wall of the run to the end of the branch — the blind end, or the seat of the valve that closes it. d is the branch inside diameter. Two errors are common and both are unconservative: measuring L from the outside of the run rather than the bore, and measuring from the run's centreline. A third error runs the other way — using the branch OD as d makes the ratio look better than it is.

The usual offenders in a real system are predictable: instrument tees for pressure and temperature, sample points, drain and vent valves left as branches rather than integrated, spare nozzles kept "for future tie-in", and the blocked leg of a tee at a transfer panel. None of these carry product flow, which is exactly why they are easy to leave long.

The fixes, in order of preference: eliminate the branch; use a zero-static or point-of-use valve designed so the seat sits in the flow path; shorten the branch physically; or — counter-intuitively, and it surprises people every time — increase the branch diameter. Since the criterion is a ratio, enlarging d lowers L/d just as effectively as shortening L. A stubby, fat instrument branch can pass where a long thin one fails.

Run the number, including the margin, on the dead-leg calculator — it returns L/d, the margin against the target of 2, and a PASS/FAIL verdict.

Slope and drainability — the GSD designations

ASME BPE SD-2.4.3 classifies gravity drainability by designation rather than by a single required number, with GSD1 corresponding to 1/16 in/ft and higher designations denoting progressively steeper slope. A project specifies the designation its process requires; the design then has to deliver it.

What drainability is protecting against is straightforward. Liquid retained in a line after a transfer or a CIP step is a microbial growth site if it is water or product, and a source of carryover and dilution if it is cleaning solution or rinse. A line that holds a litre of rinse water in a sag will deliver that litre into the next batch.

The hard part is not specifying the slope — it is keeping it. Slope is designed on paper and then attacked by everything that happens afterward:

• Support spacing and deflection. Thin-wall hygienic tube is not stiff. Sag between hangers can create a local low point in a run whose overall slope is perfectly correct — the line drains toward the low point and stops there. Drainability is a local property, not an average.
• Thermal movement. A line sloped correctly at ambient can lose its slope when it grows into its supports at process or CIP temperature, and CIP is typically the hottest condition the line sees.
• Installation tolerance. A 1/16 in/ft target is roughly 5 mm over 1 metre. That is inside the range of ordinary installation variation, which is why hygienic installations get sloped and checked rather than sloped and assumed.
• In-line components. Valve bodies, instruments and fittings can each introduce a local flat or a pocket regardless of the slope of the tube either side.

The slope and drainability calculator normalises whatever units the slope arrived in — ratio, in/ft, mm/m, percent or degrees — and reports the highest GSD designation satisfied plus the margin in in/ft to the target. Unit confusion between a European drawing in mm/m and a US specification in in/ft is a genuinely common source of a line that was built to the wrong slope entirely.

CIP velocity — the 5 ft/s practice

ASME BPE SD-6.3.5.2.1 anchors clean-in-place flow around a target velocity of 5 ft/s in the line being cleaned. The relationship between flow and velocity in a round tube is:

Q = 2.448 · v · ID² (Q in gpm, v in ft/s, ID in inches)

What the target is really standing in for is wall shear stress. Soil removal in turbulent flow is driven by the shear the fluid exerts on the wall, not by velocity as such; velocity is the quantity you can measure and control, so the practice is written in velocity. That framing matters when you are tempted to trade it away: 5 ft/s is a practice target that has decades of results behind it, not a derived threshold, and a circuit that only reaches 4 ft/s is not "80% clean" in any useful sense.

The sizing trap is mixed line sizes in one circuit. Flow through a series circuit is constant while velocity scales with 1/ID². A circuit that runs from 1 in tube into 3 in tube sees roughly nine times the velocity in the small line as in the large one. Size the pump for 5 ft/s in the largest line and the small lines are hugely over-velocity, wasting pump power and eroding the surface finish you paid for; size it for the small line and the large line never reaches the target and simply is not cleaned. The answer is normally to split the circuit so each sub-circuit has a comparable bore, not to compromise on a single flow rate.

Two further practicalities. Velocity is a full-bore quantity: a line that is not running full — a partially drained return, or an under-supplied leg — has no meaningful velocity at all, however good the calculated number is. And the supply side has to be able to deliver it: a CIP skid pump sized for one circuit will not necessarily hold 5 ft/s in a longer or larger one, and the return pump usually governs before the supply pump does.

The CIP flow velocity calculator works in both directions — the flow needed to hold 5 ft/s in a given tube ID, and the velocity an available flow actually achieves, with a PASS/FAIL verdict.

The vessel side — spray device coverage

Lines are only half the circuit. ASME BPE SD-3.9.2 addresses vessel cleaning through a coverage-rate practice: the required spray-device flow is proportional to a characteristic wetted length rather than to vessel volume, because what has to be cleaned is surface, not contents.

L_c = π·D (vertical) L_c = 2L + 2D (horizontal)

Q = coverage rate × L_c

The mechanism on a vessel wall is different from the mechanism in a pipe. The spray device wets the upper surface and the cleaning action is largely the falling film running down the wall, which is why coverage is specified per unit of circumference or perimeter — you are trying to establish a continuous film with no dry lanes between it.

Three failures recur. Shadowing: agitator shafts, baffles, dip tubes, load cells and manway nozzles cast shadows the spray never reaches, and a coverage calculation that assumes an empty vessel does not see them. Under-supply: a spray ball starved of flow produces a weak pattern that does not reach the vessel wall at all, and the failure is not proportional — below its design flow, a static spray ball degrades sharply rather than gradually. The return path: a vessel spray at design flow delivers a substantial volume that has to leave through the outlet and the return line, and if the return cannot carry it, the vessel floods, the spray drowns, and the drain line is what limits the whole circuit.

Size the device with the spray device flow calculator, then confirm the return line drains at the required GSD and that the CIP circuit velocity holds in the return itself — that is the link back to the other three checks.

Running the four checks together

The four requirements are separate mechanisms but one design. Running them in this order avoids most of the rework, because each step's outcome constrains the next:

The dependency worth naming explicitly: step 4 feeds back into steps 2 and 3. The flow a spray device needs becomes flow the return line must carry, at a velocity that must still clean the return and at a slope that must still drain it. Circuits are routinely designed forward from the vessel and then discovered to have a return line that cannot do its job, and the cost of finding that late is a re-routed return rather than a resized pump.

What this guide does not cover, and what still belongs in a hygienic design package: surface finish acceptance (Ra), material selection and corrosion resistance (PRE/PREN), weld acceptance criteria, ferrite control, and electropolish or passivation requirements. Those are material and fabrication controls rather than geometry, and they are addressed by their own calculators on the BPE line. Geometry decides whether the system can be cleaned; finish and metallurgy decide whether it stays that way.

FAQ

Where exactly do I measure L and d for a dead leg?

L runs from the inside wall of the main run to the end of the branch — the blind end, or the seat of the valve that closes it — and d is the branch inside diameter. The two frequent errors are both unconservative: measuring L from the outside surface of the run, and measuring from the run's centreline, which adds half the run bore to a length that should not include it. A third error runs the other way and makes a failing branch look acceptable: using the branch outside diameter as d. Since the criterion is a ratio, an error in either quantity moves the verdict, and on a marginal branch it moves it across the line.

Can I fix a failing dead leg by making the branch bigger?

Yes, and it is the fix people most often overlook. The criterion is L/d ≤ 2, so increasing the branch inside diameter reduces the ratio exactly as effectively as shortening the branch. A short fat instrument branch can pass where a long thin one fails, and enlarging a branch is frequently easier than relocating it. The order of preference is still: eliminate the branch entirely, use a zero-static or point-of-use valve whose seat sits in the flow path, shorten it, then enlarge it. But when the first three are impractical, enlarging is a legitimate engineering answer rather than a dodge.

Is 5 ft/s a hard requirement or a rule of thumb?

It is a practice target with a real mechanism behind it. What actually removes soil in turbulent flow is wall shear stress, and 5 ft/s is the velocity that reliably produces adequate shear in the tube sizes and fluids hygienic systems normally use. It is written in velocity because velocity is what you can specify and measure. Treat it as the design target rather than a threshold to optimise against: a circuit reaching 4 ft/s is not proportionally almost-clean, and the sensible response to a circuit that cannot hold 5 ft/s is to split it into sub-circuits of comparable bore, not to accept a lower number across the whole thing.

Why does my CIP circuit need to be split by line size?

Because flow is constant through a series circuit while velocity scales with the inverse square of the bore. A circuit running 1 in tube and 3 in tube sees roughly nine times the velocity in the small line as the large one. Size the pump to hold 5 ft/s in the 3 in line and the 1 in line is at roughly 45 ft/s — wasted power and accelerated erosion of the surface finish. Size it for the 1 in line and the 3 in line sits near 0.6 ft/s and is effectively not cleaned. No single flow rate serves both, so the circuit gets split into sub-circuits of comparable bore, each with its own flow.

Our line is sloped correctly overall but still holds liquid. What happened?

Almost certainly sag between supports. Drainability is a local property, not an average: thin-wall hygienic tube deflects readily, and a run whose end-to-end slope is exactly to specification can still contain a low point where it sags between hangers. Liquid drains toward that point and stops. Check support spacing against the tube size and wall, and check it at CIP temperature rather than ambient — thermal growth into the supports can flatten or reverse a local slope, and CIP is usually the hottest condition the line sees. In-line valve bodies, instruments and fittings can each introduce a local flat as well, independent of the tube slope either side.

Do these four checks make a system BPE-compliant?

No — they are the geometric core of cleanable design, not the whole standard. ASME BPE also covers surface finish acceptance, material selection and corrosion resistance, weld acceptance criteria and ferrite control, electropolishing and passivation, seals and elastomers, and documentation and verification requirements. Passing L/d, slope, CIP velocity and spray coverage means the system's geometry does not prevent cleaning; it says nothing about whether the surfaces, welds and materials will stay cleanable in service. Treat these four as the checks that constrain layout early, and the material and fabrication requirements as the ones that govern what gets built and how it is accepted.