9.2 Water Supply Fixture Units (WSFU) & Pipe Sizing Tables

9.2 Water Supply Fixture Units (WSFU) & Pipe Sizing Tables

Core Principle: Water distribution systems cannot be sized by simply adding the maximum flow rates of all connected fixtures, because plumbing fixtures operate intermittently rather than continuously. Under IPC Chapter 6 and Appendix E, water sizing relies on Water Supply Fixture Units (WSFU) and Hunter's Curve to convert non-coincidental fixture loads into realistic peak design flow rates in gallons per minute (gpm), while capping fluid velocities at 8 fps (cold) and 5 fps (hot) to prevent acoustic noise and erosion-corrosion.

---

The Foundations of Water Supply Fixture Units & Hunter's Curve

In the 1930s and 1940s, Dr. Roy B. Hunter of the National Bureau of Standards (NBS) conducted pioneering hydraulic research published in reports BMS65 and BMS79. Hunter recognized that in any building with multiple plumbing fixtures, the probability ($p$) of all fixtures operating simultaneously is practically zero.

The Binomial Probability Model

If a plumbing fixture operates for a duration of $t$ seconds and has an average resting interval between uses of $T$ seconds, the probability that the fixture is operating at any random second is:

$p = \frac{t}{T}$

For a building containing $n$ identical fixtures, the probability that exactly $r$ fixtures are discharging at the exact same instant follows the binomial distribution. Hunter established that distribution piping must be sized to supply the maximum number of fixtures likely to be operating simultaneously during peak demand periods, with a 99% probability that the design capacity will never be overloaded.

To simplify this complex mathematical probability for field plumbers, Hunter created the Water Supply Fixture Unit (WSFU). A WSFU is a dimensionless factor representing the relative hydraulic load produced by a given plumbing fixture, combining:

1. Flow rate (gpm)
2. Duration of a single operational cycle (seconds)
3. Average frequency of use (cycles per hour)

---

Flush Tank vs. Flushometer Valve Demand Curves

Hunter's Curve translates the cumulative sum of all connected WSFUs into an equivalent peak design flow rate in gallons per minute (gpm). The IPC Appendix E chart establishes two entirely distinct demand curves:

   PEAK DEMAND (GPM)
     |
 120 +                                                  / [CURVE 1: Flushometers]
     |                                            /----/
 100 +                                      /----/
     |                                /----/           / [CURVE 2: Flush Tanks]
  80 +                          /----/           /----/
     |                    /----/           /----/
  60 +              /----/           /----/
     |        /----/           /----/
  40 +  /----/           /----/
     | /           /----/
  20 |/      /----/
     +--+----+----+----+----+----+----+----+----+----+----+--> WSFU LOAD
     0  10   20   30   40   50   60   70   80   90  100

Why Two Curves?

• Curve 1 (Predominantly Flushometer Valves): Commercial flushometer valves deliver a massive instantaneous flow rate—typically 25 to 35 gpm—over a brief 5-to-7-second flushing cycle. Consequently, Hunter's Curve for flushometer systems rises steeply at low fixture unit counts. A system with just 5 flushometer water closets requires over 30 gpm of instantaneous supply.
• Curve 2 (Predominantly Flush Tanks): Residential gravity tank toilets draw water at a gentle rate of 2.5 to 3.0 gpm over a 45-to-60-second tank-refilling cycle. The demand curve for flush tank systems rises much more gradually, reflecting lower instantaneous flow spikes.

Table E103.3(3) — Table for Estimating Demand (representative rows; read the full table in the code book on test day):

Two traps in this table. First, the flush-tank column does not start near zero: even a 5-WSFU flush-tank system is credited with 9.4 gpm, because Hunter's model builds in a minimum simultaneous-use flow. Candidates who remember a "3 gpm at 5 fixture units" figure are reading the first rows of the table, which cover loads of 1 to 4 fixture units. Second, the table is not linear — you interpolate between rows, you do not scale.

[!NOTE] Reading the Curves: at low fixture counts a flushometer system demands roughly 1.6 to 1.9 times the flow of an equivalent flush-tank system (15.0 against 9.4 gpm at 5 WSFU; 27.0 against 14.6 gpm at 10 WSFU). As the load grows the diversity of uses smooths the peaks and the two curves converge — 65.0 against 90.0 gpm at 200 WSFU, and 85.0 against 108.0 gpm at 300 WSFU.

Fixture Unit Values for Common Fixtures (IPC Table E103.3)

Under IPC Table E103.3(2) and Table E103.3(3), plumbing fixtures are assigned specific WSFU values based on occupancy classification (Private vs. Public) and water service line (Cold, Hot, or Total):

• Private Occupancy: Installations in residential single-family dwellings, private apartments, or hotel guest rooms intended strictly for the use of an individual or family.
• Public Occupancy: Installations in commercial buildings, schools, restaurants, office restrooms, theaters, and retail stores accessible to employees, patrons, or the general public.

Table E103.3(2) — Load Values Assigned to Fixtures (reproduced in full; note that the table names the type of supply control, which is what separates a flush tank from a flush valve):

Footnote a to the table (memorize the mechanism): "For fixtures not listed, loads should be assumed by comparing the fixture to one listed using water in similar quantities and at similar rates. The assigned loads for fixtures with both hot and cold water supplies are given for separate hot and cold water loads and for total load. The separate hot and cold water loads being three-fourths of the total load for the fixture in each case." Check it: a private lavatory's total is 0.7, and 0.75 × 0.7 = 0.525, rounded to the 0.5 shown in the cold and hot columns.

Use the bathroom-group row when a bathroom group is what the stem describes. A private flush-tank bathroom group is assigned 3.6 total WSFU as a group. Summing its parts individually (2.2 + 0.7 + 1.4) gives 4.3, and that is the classic wrong answer — it double-counts diversity the table has already credited.

Fixtures not in the table: hose bibbs and sillcocks are not listed in Table E103.3(2). Footnote a directs you to assume a load by comparison to a listed fixture using water in similar quantities and at similar rates; the commonly used design value is 2.5 WSFU per sillcock, which is what the worked example below uses. Note it as a design assumption, not a table value.

The Mathematical Non-Additive Rule for Total WSFU

Notice that for fixtures requiring both hot and cold water (such as a private lavatory or bathtub), Total WSFU does NOT equal Cold WSFU plus Hot WSFU:

• Private Lavatory: $\text{Cold } (0.5) + \text{Hot } (0.5) = 1.0$, but Total WSFU = 0.7
• Bathtub: $\text{Cold } (1.0) + \text{Hot } (1.0) = 2.0$, but Total WSFU = 1.4
• Kitchen Sink: $\text{Cold } (1.0) + \text{Hot } (1.0) = 2.0$, but Total WSFU = 1.4

[!IMPORTANT] Diversity Principle: When a person draws water at a lavatory or shower, they do not run both the hot and cold valves wide open simultaneously. Instead, they blend hot and cold water to achieve a comfortable temperature ($100^\circ\text{F}$ to $105^\circ\text{F}$). Therefore, the total volume of water leaving the spout remains bounded. When sizing the main building supply pipe (which carries both hot and cold water prior to the water heater branch), you must sum the Total WSFU column, not the arithmetic sum of cold plus hot.

Minimum Individual Fixture Supply Pipe Diameters (IPC Section 604.5 and Table 604.5)

Regardless of the calculated WSFU total, the physical branch pipe connecting directly to an individual fixture supply fitting must never be smaller than the prescriptive minimum diameters in Table 604.5 (Minimum Sizes of Fixture Water Supply Pipes), referenced by Section 604.5. Section 604.5 also adds two rules worth tabbing: "The fixture supply pipe shall not terminate more than 30 inches (762 mm) from the point of connection to the fixture," and the same minimum sizes apply to individual distribution lines in gridded or parallel (manifold) water distribution systems.

Do not confuse Table 604.4 with Table 604.5. Table 604.4 is Maximum Flow Rates and Consumption for Plumbing Fixtures and Fixture Fittings — the water-conservation table (lavatory 2.2 gpm at 60 psi, public metering lavatory 0.25 gallon per cycle, showerhead 2.5 gpm at 80 psi, urinal 1.0 gallon per flush, water closet 1.6 gallons per flush). Table 604.5 is the minimum pipe size table used here.

• The 3/8-Inch Rule: A 3/8-inch nominal diameter tube is permitted strictly as a flexible or rigid riser connecting a single fixture (such as a lavatory faucet or toilet tank fill valve) from the wall angle stop to the fixture inlet. The supply branch inside the wall leading up to the angle stop must be at least 1/2-inch nominal pipe size.
• The Flushometer 1-Inch Mandate: A commercial flushometer water closet requires an instantaneous surge of 25 to 35 gpm. Connecting a flushometer valve to a 3/4-inch or 1/2-inch pipe causes an immediate hydraulic pressure collapse, preventing the valve diaphragm from cycling and leaving the toilet in an unclosable continuous flush condition.

---

Velocity Limits & Hydraulic Pipe Sizing

When water flows through a pipe, its average velocity ($V$) in feet per second (fps) is determined by the flow rate ($Q$ in gpm) and the internal pipe diameter ($d$ in inches):

$V = \frac{0.408 \times Q}{d^2}$

Velocity Thresholds — Engineering Practice, Not an IPC Table

Be precise about the source here. The 2006 IPC contains no numeric maximum velocity for water distribution piping. What it says is Section 604.9: "The flow velocity of the water distribution system shall be controlled to reduce the possibility of water hammer. An air chamber is not a water hammer arrestor. A water-hammer arrestor shall be installed where quick-closing valves are utilized." (That second sentence is an Indiana addition — 675 IAC 16-1.4-7(b) adds it and deletes the model code's last sentence.) Appendix E, which Indiana does adopt, uses velocity in its sizing procedure and is where a velocity-based answer would come from.

The figures below are the industry design limits that engineers and the copper-tube industry apply, and they are what an exam item means when it asks for "maximum recommended velocity":

• Cold Water Distribution Piping: Maximum 8.0 feet per second (fps).
• Hot Water Distribution Piping ($> 120^\circ\text{F}$): Maximum 5.0 feet per second (fps).

+-------------------------------------------------------------------------+
|                    MAXIMUM PERMISSIBLE WATER VELOCITIES                 |
+-------------------------------------------------------------------------+
| Cold Water Lines                | 8.0 fps (industry design practice)    |
| Hot Water Lines (> 120°F)       | 5.0 fps (industry design practice)    |
| Copper Hot Water Recirculation  | 4.0 to 5.0 fps (CDA recommendation)   |
+-------------------------------------------------------------------------+

Why Hot Water Velocity Is Capped at 5 fps

Elevated water temperature significantly accelerates chemical and physical degradation inside metallic piping. In copper tubing carrying water above $120^\circ\text{F}$ to $140^\circ\text{F}$:

1. Erosion-Corrosion (Impingement Attack): High-velocity hot water continuously strips away the microscopic cuprous oxide ($Cu_2O$) passivation layer that protects copper against corrosion. Once bare copper metal is exposed, the water dissolves the pipe wall, producing horseshoe-shaped gouges, severe wall thinning, and pinhole leaks downstream of fittings.
2. Cavitation: Localized pressure drops around elbows cause dissolved gases to bubble out and violently implode against the pipe wall, pitting the metal.
3. Acoustic Noise: Velocities exceeding 8 fps create whistling, hissing, and rumbling vibrations through building framing.

---

Pipe Carrying Capacities at Design Velocities

The following reference table gives the maximum flow rate in gpm for Type L copper tubing at the industry design velocity limits (8 fps cold, 5 fps hot). These are design practice, not 2006 IPC table values:

---

Step-by-Step WSFU Sizing Walkthrough

Let us trace the complete sizing procedure for a single-family residential home containing the following plumbing fixtures:

• Bathroom 1: Flush tank water closet, lavatory faucet, bathtub/shower
• Bathroom 2: Flush tank water closet, lavatory faucet, shower stall
• Kitchen: Kitchen sink faucet, domestic dishwasher
• Laundry: Clothes washing machine, laundry sink faucet
• Exterior: Two hose bibbs ($1/2"$ sillcocks)

Step 1: Tabulate WSFU Loads per IPC Table E103.3(2)

† Hose bibbs are not listed in Table E103.3(2); 2.5 WSFU each is the common design assumption made under footnote a. The remaining values are table entries, tabulated fixture by fixture rather than by the bathroom-group row so that the cold and hot branches can be sized separately.

Step 2: Convert WSFU to Peak Design Flow Rates (gpm)

Read Table E103.3(3), flush-tank column, interpolating between the published rows:

• Main Building Supply (Total 19.2 WSFU): between 17.5 gpm at 15 WSFU and 19.6 gpm at 20 WSFU → about 19.4 gpm
• Cold Water Main Branch (15.4 WSFU): between 17.5 at 15 and 19.6 at 20 → about 17.7 gpm
• Hot Water Main Branch (7.4 WSFU): between the 7-fixture-unit and 8-fixture-unit rows of the table → about 12.2 gpm

Why these numbers look large for a house. Hunter's model is deliberately conservative at small loads, and Appendix E's segmented-loss method (E103.3) then sizes on available pressure and friction loss, not on velocity alone. The demand figures above are the input to that calculation, not the final answer by themselves.

Step 3: Screen Pipe Sizes Against the Design Velocity Limits

1. Main Water Service & Building Main (19.4 gpm): a 3/4-inch Type L copper pipe carries 12.1 gpm at 8 fps, so it is out; 1-inch Type L carries 20.6 gpm at 8 fps and clears the demand.
2. Cold Water Main Branch (17.7 gpm): likewise 1-inch, reducing after the first bathroom branch once the downstream load drops.
3. Hot Water Main Trunk (12.2 gpm): at the 5 fps hot-water design limit a 3/4-inch pipe carries only 7.5 gpm and a 1-inch pipe carries 12.9 gpm, so the trunk leaving the water heater screens at 1 inch. In practice a designer running the full Appendix E segmented-loss calculation, with the shorter developed length of a house and the real fixture diversity on a hot-only branch, will often land on 3/4 inch — which is exactly why Appendix E sizes on pressure loss and this velocity screen is only the first cut.
4. Individual Bathroom Group Branch: Table E103.3(2) assigns a private flush-tank bathroom group 3.6 total WSFU (2.7 cold, 1.5 hot). Read on the demand table that is a small load, comfortably served by a 3/4-inch cold branch and a 1/2-inch hot branch, with 1/2-inch and 3/8-inch fixture runouts under Table 604.5. If you instead sum the individual fixtures you get 4.3 WSFU — conservative, but not what the table assigns.