Engineering story
Why LRFD replaced ASD
In 1994, AASHTO published the first LRFD Bridge Design Specifications. The switch away from Allowable Stress Design (ASD) was not fashion — it was a response to a generation of research showing that a single global safety factor could not simultaneously calibrate against the very different uncertainties in dead load, live load, wind, seismic events, and material resistance. Bridges designed to the 1996 Standard Specifications and to the first LRFD editions coexist in the Mid-Atlantic today. Understanding both is a load-rating prerequisite.
In Load and Resistance Factor Design (LRFD), bridges are designed for specific limit states that consider various loads and resistances. These limit states include strength, extreme event, service, and fatigue. Each limit state is a design check whose combination of load factors, load combinations, and resistance factors has been statistically calibrated to a target reliability index β. The remainder of this chapter develops each piece.
Chapter objectives
What you will be able to do
Learning objectives
By the end of this chapter you will be able to:
- 1Distinguish ASD, LFD, and LRFD philosophies.
- 2Interpret the reliability index β and its role in code calibration.
- 3Enumerate every AASHTO limit state and state its purpose.
- 4Look up load factors γ from AASHTO Tables 3.4.1-1 and 3.4.1-2 and apply them.
- 5Assemble a Strength I demand for a girder combining DC, DW, and LL+IM.
- 6Select resistance factors φ for reinforced concrete, prestressed concrete, and structural steel.
- 7Compute the number of design lanes NL, apply multiple-presence factors m, and dynamic load allowance IM.
- 8Distinguish the design truck, design tandem, design lane load, and fatigue truck within HL-93.
2.1 — Central equation
The LRFD design inequality
Every AASHTO LRFD check reduces to a single inequality: the sum of factored force effects must not exceed the factored resistance:
Formula
- load modifier for load i (ductility · redundancy · importance)
- load factor for load type i
- force effect from load i [kip, kip-ft]
- resistance factor
- nominal resistance [kip, kip-ft]
- factored resistance = \phi\,R_{n} [kip, kip-ft]
The elegance of LRFD is that both sides of the inequality are calibrated against their real statistical distributions. A live-load moment with a coefficient of variation (COV) of ~18% receives a heavier factor (γLL = 1.75 in Strength I) than a dead-load moment with a COV of ~9% (γDC = 1.25). Resistance factors φ reflect the manufacturing scatter of the material and the reliability of the resistance model.
2.2 — Load modifier η
Ductility, redundancy, importance
The load modifier η captures three qualitative attributes of the member being designed. For loads for which a maximum value of γi is appropriate:
Formula
- ductility factor (0.95 – 1.05)
- redundancy factor (0.95 – 1.05)
- operational importance factor (0.95 – 1.05)
For loads for which a minimum value of γi is appropriate (uplift, overturning):
Formula
Engineering note
2.3 — Limit states
Strength, service, fatigue, extreme event

The following limit-state combinations are defined in AASHTO LRFD Article 3.4.1:
| Limit state | Purpose |
|---|---|
| Strength I | Basic vehicular use, no wind. Governs most girder design. |
| Strength II | Owner-specified special / permit vehicles, no wind. |
| Strength III | Wind velocity > 55 mph, no live load (empty-bridge wind). |
| Strength IV | Very high dead/live ratio (spans > ~250 ft). |
| Strength V | Normal vehicular use with 55 mph wind. |
| Extreme Event I | Load combination including earthquake. |
| Extreme Event II | Ice load, vessel or vehicle collision, check flood. |
| Service I | Normal operational use; deflection check. |
| Service II | Control yielding of steel and slip of slip-critical connections. |
| Service III | Tension in prestressed concrete — crack control. |
| Service IV | Tension in prestressed concrete columns under wind. |
| Fatigue I | Infinite fatigue life (LL+IM only, γ = 1.5). |
| Fatigue II | Finite fatigue life (LL+IM only, γ = 0.75). |
Table 2.1 — AASHTO limit states (LRFD 10th Ed., §3.4.1). Each combination pairs a defined set of load factors with a target reliability index.
2.4 — Load factors
AASHTO Table 3.4.1-1 (abridged)

| Combination | DC/DW/EH/EV/ES | LL+IM+CE+BR | WA | WS | WL | EQ | CT/CV/IC |
|---|---|---|---|---|---|---|---|
| Strength I | γp | 1.75 | 1.00 | — | — | — | — |
| Strength II | γp | 1.35 | 1.00 | — | — | — | — |
| Strength III | γp | — | 1.00 | 1.00 | — | — | — |
| Strength IV | γp (1.5 DC) | — | 1.00 | — | — | — | — |
| Strength V | γp | 1.35 | 1.00 | 1.00 | 1.00 | — | — |
| Extreme Event I | γp | γEQ | 1.00 | — | — | 1.00 | — |
| Extreme Event II | γp | 0.50 | 1.00 | — | — | — | 1.00 |
| Service I | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | — | — |
| Service II | 1.00 | 1.30 | 1.00 | — | — | — | — |
| Service III | 1.00 | 0.80 | 1.00 | — | — | — | — |
| Service IV | 1.00 | — | 1.00 | 1.00 | — | — | — |
| Fatigue I (LL+IM+CE only) | — | 1.75 | — | — | — | — | — |
| Fatigue II (LL+IM+CE only) | — | 0.80 | — | — | — | — | — |
Table 2.2 — AASHTO LRFD Table 3.4.1-1 (abridged). γp is drawn from Table 3.4.1-2 (permanent-load factors) and depends on whether the load produces the maximum or minimum force effect.
2.5 — Permanent-load factors γp
AASHTO Table 3.4.1-2
| Permanent-load type | Max γp | Min γp |
|---|---|---|
| DC — components and attachments | 1.25 | 0.90 |
| DC — Strength IV only | 1.50 | 0.90 |
| DD — downdrag piles, Tomlinson α method | 1.40 | 0.25 |
| DD — piles, λ method | 1.05 | 0.30 |
| DD — drilled shafts, O'Neill & Reese (1999) | 1.25 | 0.35 |
| DW — wearing surfaces and utilities | 1.50 | 0.65 |
| EH — active horizontal earth pressure | 1.50 | 0.90 |
| EH — at-rest horizontal earth pressure | 1.35 | 0.90 |
| EL — locked-in construction stresses | 1.00 | 1.00 |
| EV — vertical earth pressure, retaining walls | 1.35 | 1.00 |
| EV — rigid buried structure | 1.30 | 0.90 |
| ES — earth surcharge | 1.50 | 0.75 |
Table 2.3 — Extracted from AASHTO LRFD Table 3.4.1-2. Always use the value of γp that produces the more critical demand (max when the load adds to the effect, min when it relieves it).
2.6 — Load and load designation
What each symbol means
Permanent loads
Transient loads
2.7 — Force-effect equations
Simplified Strength / Service / Fatigue
The load effect Q for superstructure design reduces to the following combinations when only DC, DW, LL+IM, WA, WS and WL are present:
2.8 — Resistance factors φ
Strength limit states
Resistance factors are prescribed by AASHTO in each material chapter. For service and extreme-event limit states, φ = 1.0 (except for bolts). The most frequently encountered strength-state values are:
| Material / action | φ | Article |
|---|---|---|
| Flexure & tension — reinforced concrete | 0.90 | §5.5.4.2 |
| Flexure & tension — prestressed concrete | 1.00 | §5.5.4.2 |
| Shear — concrete | 0.90 | §5.5.4.2 |
| Axial compression — concrete | 0.75 | §5.5.4.2 |
| Flexure — structural steel | 1.00 | §6.5.4.2 |
| Shear — structural steel | 1.00 | §6.5.4.2 |
| Axial compression — structural steel | 0.90 | §6.5.4.2 |
| Tension, yielding on gross section — steel | 0.95 | §6.5.4.2 |
| Tension, fracture on net section — steel | 0.80 | §6.5.4.2 |
| Bolts in shear (A325) | 0.80 | §6.5.4.2 |
Table 2.4 — Frequently used strength-state φ. Full listings are in AASHTO LRFD Chapters 5 (concrete), 6 (steel), 7 (aluminum), 8 (wood), and 10 (foundations).
2.9 — Reliability
Calibrated for β ≈ 3.5
AASHTO LRFD was calibrated against a target reliability index β ≈ 3.5 at the Strength I limit state, corresponding to a failure probability of roughly 2×10⁻⁴ over the 75-year design life. β is the distance, in standard deviations, between the mean values of resistance R and load effect Q:
Formula
- mean of resistance R
- mean of load effect Q
- standard deviations of R and Q

AASHTO LRFD Reference
2.10 — Design live load
HL-93
HL-93 is a notional live load where H represents the HS truck, L the lane load, and 93 the year the load was adopted. It consists of a combination of:
- Design truck (HS-20): 8 kip front axle, two 32 kip rear axles at 14 ft and 14–30 ft (variable) spacing.
- Design tandem: a pair of 25 kip axles spaced 4 ft apart.
- Design lane load: 0.64 kip/ft uniformly distributed longitudinally, occupying a 10-ft-wide strip within a 12-ft design lane. The lane load is not subject to dynamic allowance.
The transverse spacing of wheels for both the design truck and design tandem is 6.0 ft. The uniform lane load may be continuous or discontinuous, as necessary to produce the maximum force effect.

2.11 — Fatigue live load
Single truck, 30-ft fixed spacing
The fatigue live load consists of a single design truck with the rear axle group fixed at 30 ft spacing between the 32 kip axles. The frequency of the fatigue load is taken as the single-lane average daily truck traffic (ADTTSL). When the bridge is analyzed by approximate load distribution (§4.6.2), the single-lane distribution factor is used.

2.12 — Number of design lanes
NL
Formula
- number of design lanes (integer)
- clear roadway width (curb-to-curb) [ft]
Example: for w = 40 ft, NL = INT(40/12) = 3 design lanes. Roadways from 20 to 24 ft carry two design lanes each equal to half the roadway width.
2.13 — Multiple presence factor
m
| Number of loaded lanes | Multiple-presence factor, m |
|---|---|
| 1 | 1.20 |
| 2 | 1.00 |
| 3 | 0.85 |
| > 3 | 0.65 |
Table 2.5 — AASHTO Table 3.6.1.1.2-1. m reflects the reduced probability that multiple lanes are simultaneously loaded to full HL-93 magnitude. When approximate live-load distribution factors are used, m is already embedded and shall not be applied again.
2.14 — Dynamic load allowance
IM
| Component | IM (%) |
|---|---|
| Deck joints — all limit states | 75% |
| All other components — fatigue and fracture limit states | 15% |
| All other components — all other limit states | 33% |
Table 2.6 — AASHTO Table 3.6.2.1-1. IM is applied to the design truck / tandem only and not to the design lane load. It also does not apply to pedestrian, wind, or centrifugal loads.
For bridges with reduced truck volume, AASHTO permits a fatigue-life adjustment based on single-lane average daily truck traffic: ADTTSL < 1,000 → 95%; ADTTSL < 100 → 90%. This reflects the reduced probability of attaining the design event during the 75-year reference period.
2.15 — Live-load distribution factors
DF for beam/girder bridges
Distribution factors (DF) convert lane demand into per-girder demand. AASHTO provides closed-form approximate equations, tabulated by cross-section type (steel/PS I-girder, box, T-beam, etc.), for both moment and shear, and separately for one-lane and multiple-lanes loaded, interior and exterior girders:
- DFMsi — single-lane, interior beam, moment
- DFMmi — multiple-lane, interior beam, moment
- DFMse — single-lane, exterior beam, moment
- DFMme — multiple-lane, exterior beam, moment
- DFVsi, DFVmi, DFVse, DFVme — analogous for shear
DF is most sensitive to beam (girder) spacing; span length and longitudinal stiffness have smaller influences. Per AASHTO §2.5.2.7.1, the load-carrying capacity of an exterior beam shall not be less than that of an interior beam. Chapter 3 works through the DF calculation on the Maryland worked example.
2.16 — Strength I combination in practice
Assembling the demand
Worked demand — girder midspan moment
Given unfactored midspan moments MDC = 620 k-ft, MDW = 95 k-ft, MLL+IM = 780 k-ft (already multiplied by distribution factor and IM):
Mu,Str-I = 1.00 × [ 775.0 + 142.5 + 1,365.0 ] = 2,282.5 k-ft
This factored demand Mu must satisfy Mu ≤ φ Mn, with φ = 0.90 for reinforced concrete flexure or φ = 1.00 for prestressed concrete flexure. Chapter 3 walks through the full sequence — including how MLL+IM was obtained — on the three-span Maryland bridge worked example.
Chapter summary
Key takeaways
- LRFD balances calibrated load and resistance factors against a target β ≈ 3.5.
- Four families of limit states — Strength, Service, Fatigue, Extreme Event — cover every design check.
- η captures ductility, redundancy, and importance — usually 1.0 for standard bridges.
- Permanent-load γp values (Table 3.4.1-2) differ for max vs. min effects; always pick the more critical.
- HL-93 = larger of (design truck, design tandem) + design lane load; IM applies only to the truck / tandem.
- Fatigue truck has fixed 30-ft rear-axle spacing; frequency = single-lane ADTT.
- Multiple-presence factor m adjusts for the probability of simultaneous multi-lane loading.
- DF depends most on girder spacing; exterior girders may never be weaker than interior.
Section 2
Fully Worked Examples
Complete AASHTO LRFD solutions with knowns, assumptions, step calculations, verification, and design commentary. Difficulty rises from basic to consulting-grade.
Worked Example 1
Problem
Step-by-Step
Design Verification
LL dominates (≈87%). This is normal for deck strips — deck design is nearly always live-load-controlled.
Discussion
The order 'multiply first, then sum' is essential — combining unfactored loads first and factoring the total is a common student error that hides which limit state governs.
Worked Example 2
Problem
Step-by-Step
Design Verification
Strength IV activates when the DC/(LL+IM) ratio is high (~7×). Here 4200/320 ≈ 13, well past the trigger. ✓
Discussion
Truss chords and long-span girders routinely see Strength IV govern. On short spans, Strength I almost always controls.
Worked Example 3
Problem
Step-by-Step
Design Verification
Utilization 0.85 leaves a 15% margin above the calibration target β = 3.5 (Pf ≈ 1 in 4300).
Discussion
LRFD does not compute β for you — it embeds it in the γ/φ pair. Utilization > 1 means either β falls below 3.5 or a redistribution is required.
Worked Example 4
Problem
Step-by-Step
Design Verification
Utilization 6.44/12 = 0.54 — comfortable margin. If the detail had been Cat E′ ((ΔF)_TH = 2.6 ksi), the same Δf would fail decisively.
Discussion
Fatigue I (γ = 1.75) targets infinite life; Fatigue II (γ = 0.80) targets finite life on low-ADTT structures. Never mix load factors between checks — pick the combination that matches the design life goal.
Worked Example 5
Problem
Step-by-Step
Design Verification
The section stays in compression at Service III; no crack initiation is expected under characteristic live load.
Discussion
Service III uses γ_LL = 0.80 because the tension limit is calibrated to a lower live-load fractile — the goal is crack control, not strength. Applying γ_LL = 1.00 here would over-penalize prestressed girders.
Section 3
Guided Practice
Complete the missing steps. Use Hints for AASHTO article pointers and setup logic before revealing the full step. Submit at the end to send your work to your instructor.
Guided Problem 1
A simply supported interior prestressed I-girder has these unfactored moments at midspan: k-ft (girder + deck), k-ft (2-in future overlay), and a per-lane truck live-load moment k-ft (no impact). The distribution factor is and IM = 33%.
Assemble the Strength I moment M_u with η = 1.0.
Guided Problem 2
A two-girder steel bridge is nonredundant (both girders are fracture-critical), the weld detail is not enhanced-ductility, and the bridge carries a designated emergency route.
Compute η for use in Strength I where γ is taken at its maximum value.
Guided Problem 3
A rural bridge has a clear roadway w = 38 ft. Two lanes of AASHTO design trucks each produce a girder moment M₁ = 720 k-ft (per lane, per girder, no impact) using the appropriate distribution factor. Include IM = 33%.
Determine N_L, choose the correct multiple-presence factor m, and compute the Service I contribution of live load per girder.
Guided Problem 4
A Category C′ transverse stiffener-to-web weld on a plate girder sees a fatigue-truck stress range Δf = 4.2 ksi (from a single fatigue truck, before load factor and impact). IM for fatigue = 15%. The infinite-life threshold for Category C′ is ksi.
Check the Fatigue I stress range and decide whether the detail meets the infinite-life criterion.
