Tools › Civil and Geotechnical › Foundations › MSE Wall Design › Method

MSE Wall Design: Method

Open the MSE Wall Design tool

MSE Wall Design

What the tool does

It checks a trial section of a mechanically stabilized earth (MSE) wall with a near-vertical face by the LRFD procedure of the Federal Highway Administration's current manual, Design and Construction of Mechanically Stabilized Earth (MSE) Walls, FHWA-HIF-24-002, Geotechnical Engineering Circular No. 11, August 2023, Chapter 4, with the reinforcement properties of its Chapter 3. That manual states that it follows the AASHTO LRFD Bridge Design Specifications, 9th edition (2020). The equation, figure and table numbers quoted here and in the results are the manual's. All forces are per unit length of wall. The page opens on the manual's Example C1: a 30 ft wall with 21 ft steel strips under a 2 ft traffic surcharge.

The AASHTO specifications are copyrighted and were not read; every AASHTO article named on these pages is quoted from the FHWA manual. Every load factor and resistance factor is an input, and the defaults are the values the FHWA manual prints in its Tables 17, 18, 19 and 22. An owner's specification may require other values. The rules of the manual's 2009 edition (FHWA-NHI-10-024 and 10-025), which this tool first followed, remain available in Advanced mode under Manual Followed.

External stability

The reinforced zone of height H and length L is a rigid block. The retained fill presses on the vertical plane at the back of the reinforcement, of height h = H for a level backfill and h = H + L tan β under a slope (not more than H plus the slope height for a broken backslope). The coefficient K_ab is Coulomb's (Equations 18 and 19) for a vertical plane, with the thrust inclined at δ = 2/3 of the lower of the friction angles of the retained and the reinforced fill, for a level surface as well. For a broken backslope β is replaced by the angle β_i to the ground surface 2 H behind the face (Equation 20). The forces are the reinforced mass V_1 = γ_r H L, the soil above the wall V_3 at its own unit weight γ_es, the earth thrust F_1 = ½ γ_b h² K_ab acting at h / 3 and the surcharge thrust F_2 = q h K_ab at h / 2, each with a horizontal and a vertical part, and the live load above the reinforced zone V_2 (Equations 21 to 41).

Sliding (Equations 44 and 45): driving force γ_EH F_1H + γ_LS F_2H against φ μ ΣV, with the minimum factor on the weights, the surcharge above the reinforced zone left out, and μ the smaller of tan φ′_f and tan φ′_r. For a continuous geosynthetic sheet the check is repeated for the block above the lowest sheet with μ = tan ρ, taken as 2/3 of the smaller of tan φ′_r and tan φ′_f unless an interface angle is entered (Section 4.4.8.9; Equation 8). Limiting eccentricity (Equations 57 to 60): with the same factors, the eccentricity of the resultant on the base, to be not more than L / 4 on soil and 3 L / 8 on rock. Bearing (Equations 46 to 55): the vertical load, now with the surcharge above the reinforced zone, on the Meyerhof effective width L′ = L − 2 e_B; it is computed with the maximum factors (Figure 41b) and again with the minimum factor on the weights and the maximum on the thrust (the "critical" case of the manual's examples), and the lower ratio governs. The factored resistance is entered, or computed as φ (c_f N_c + 0.5 L′ γ_f N_γ) with N_q = e^(π tan φ) tan²(45° + φ / 2), N_c = (N_q − 1) cot φ and N_γ = 2 (N_q + 1) tan φ, the closed forms behind the manual's Table 20. The Service I pressure is printed for a settlement estimate made elsewhere.

Internal stability: the three methods

The manual allows the simplified and the coherent gravity method for steel reinforcement, and the simplified, the stiffness and the limit equilibrium method for geosynthetics (Section 4.4.9.1), without a preference. The page opens with the simplified method for both, the one that gives the larger tension in a geosynthetic. In each method a layer carries the horizontal stress over its tributary height S_v, from midway to the layer above (or the top of the wall) to midway to the layer below (or the base).

Simplified method (Section 4.4.9.4.2). σ_H = K_r γ_EV (γ_r Z + γ_es S_eq + q) at the depth Z of the layer, with K_a = tan²(45° − φ′_r / 2) and the ratio K_r / K_a of Figure 65: 1.7 for steel strips and 2.5 for wire grids at the top of the wall, falling in a straight line to 1.2 at 20 ft and below; 1.0 for geosynthetics. Under a slope S_eq = ½ (0.7 H) tan β (Equation 79). A traffic surcharge over the reinforced zone is counted with the EV factor. T_max = σ_H S_v (Equation 88). The manual takes the stress at the layer; the exact average over the tributary height and the two shortcuts of the 2009 manual can be chosen in Advanced mode.

Coherent gravity method, steel only (Section 4.4.9.4.1). For the block above each layer the eccentricity e is computed from the unfactored forces (Equations 67 to 70) and the factored vertical stress is the factored vertical load on L − 2 e (Equation 71). K falls from the at-rest value 1 − sin φ′_r, at the point where the failure surface meets the ground, to K_a 20 ft below it (Equations 72 to 76). Equation 71 puts γ_EH on the vertical part of the earth thrust; the manual's Examples C1 and C2 put γ_EV there in the rupture step and γ_EH (C2) or γ_EV (C1) in the pullout step. The page follows the equation, which gives the larger tension.

Stiffness method, geosynthetics only (Section 4.4.9.4.3), in the simplified form the manual gives after AASHTO: T_max = S_v [γ_EV H γ_r D_tmax + γ_EV γ_es (H_ref / H) S_eq + γ_LS q] K_a Φ_g Φ_local Φ_fs (Equation 94) with H_ref = 20 ft, Φ_g = 0.16 (S_global / p_a)^0.26, S_global = Σ J_i / H, Φ_local = (S_local / S_global)^0.5, S_local = J_i / S_v, p_a = 2.11 ksf, and D_tmax rising from 0.12 at the top to 1.0 at the depth z_b = 0.32 H^1.2 in feet (Equations 91 to 101). J_i is the secant stiffness of the product at 2 percent strain and 1,000 hours times the coverage ratio. The facing stiffness factor is entered (1.0 in the manual's examples) and the facing batter factor is 1. The method has its own resistance factors, 0.80 in tension and 0.70 in pullout (Table 22), and a service check: the strain T / J_i under the load factors 1.20 and 1.00 must stay within 2.5 percent for flexible-faced and 2 percent for stiff-faced walls (Equations 105 and 106). It gives much less tension than the simplified method (about a third in Example C1). The geogrid strengths the form opens with are illustrative round values, not any manufacturer's product, chosen so that the page opens on a passing case under the simplified method; the generic geogrid of Example C1, sized there by the stiffness method, gives ratios below 1 in the lower layers under the simplified method; enter the product actually used. The manual says the method needs further development for concentrated loads. The limit equilibrium method is not built.

Failure surface and pullout

The length in front of the failure surface is L_a = 0.3 H_1 in the upper part of the wall and 0.6 (H − Z) below for steel, with the mechanical height H_1 = H + tan β (0.3 H) / (1 − 0.3 tan β) (Figure 53, Equation 61), and L_a = (H − Z) tan(45° − φ′_r / 2) for geosynthetics (Equation 62). The embedment is L_e = L − L_a. The factored pullout resistance per unit length of wall is φ F* σ_v C L_e R_c (Equations 2 and 116) with C = 2 and the unfactored overburden σ_v at the middle of the embedment length, without the live load. The 2023 manual drops the scale-effect factor α (p. 56). The resistance is compared with the tension recomputed without the live load surcharge (Section 4.4.9.7, after the AASHTO commentary the manual cites); under a slope that tension is computed, as in Example C2, with the whole soil above the reinforcement and not with S_eq, so it can exceed T_max.

F* is 2.0 at the surface, or 1.2 + log C_u when a uniformity coefficient is entered, falling to tan φ′_r at 20 ft, for steel strips (Table 8); 20 t / S_t falling to 10 t / S_t for wire grids (Equations 5 and 6); and C_i tan φ′_r with C_i = 2/3 for geosynthetics unless a tested value is entered (Equation 7). Under a slope the depth for F* of steel is measured from the point where the failure surface meets the ground, as Example C2 does. A wire grid of n longitudinal wires at the spacing S_l is given the width (n − 1) S_l.

Long-term strength and connection

Steel (Section 3.5.1): the zinc coating is lost at the rates of Table 10, 0.58 mil per side per year for two years and 0.16 mil per year afterwards, and the carbon steel then at 0.47 mil per side per year, so a 3.4 mil coating lasts 16 years and a 75-year design life costs 2 × 0.47 × 59 = 55.5 mils of thickness or diameter (Equations 9 and 15). The manual gives lower rates, 0.08 and 0.35 mil per year, for fill with a resistivity above 10,000 Ω-cm (p. 61); they can be entered. The nominal strength is F_y times the remaining section. Geosynthetics: T_al = T_ult / (RF_ID RF_CR RF_D) (Equation 16). The factored resistance φ T_al R_c is compared with T_max (Equations 114 and 115), with φ = 0.75 for strips, 0.65 for grids on a rigid facing and 0.90 for geosynthetics (Table 22).

Connection (Section 4.4.9.8): the facing connection must carry T_max with the resistance factor of the reinforcement. For a geosynthetic the page checks φ T_alc with T_alc = T_ult CR_cr / RF_D (Equation 119) for each layer with an entered CR_cr, and does not let it exceed φ T_al. For steel the connection depends on the panel hardware and is not computed.

Least reinforcement length

For sliding, eccentricity and bearing the page searches upward from 0.1 H for the first length at which the check is met, all other inputs unchanged (a scan in steps of about 0.01 H, then bisection); under a slope the thrust grows with the length, so a check may not be reachable. For pullout the least embedment of each layer is solved with the overburden that goes with that shorter length, and the length is L_a plus the larger of that embedment and the entered minimum (3 ft in the 2009 manual; the 2023 manual does not restate one). The required length is the largest of these and of the manual's minimum, the larger of 0.7 H and 8 ft (Section 4.4.3).

What changed from the 2009 manual

The thrust on the back of the block is inclined at 2/3 φ (2009: parallel to the backslope, horizontal for a level backfill), which lowers the driving force and adds a vertical component: for the 2009 manual's Example E4 the sliding ratio rises from 1.37 to 1.85 and the eccentricity falls from 3.87 ft to 1.85 ft. The soil above the wall has its own unit weight. The simplified method takes the stress at the layer (2009: averaged over the tributary height). The pullout demand leaves out the live load (2009: recommended keeping it). The scale-effect factor is gone (2009: 0.8 for geogrids, 0.6 for geotextiles), so default geosynthetic pullout resistances are higher. Under a slope F* of steel is measured from the mechanical height, which lowers it. The coherent gravity and stiffness methods are added. The load factors, the resistance factors of the simplified method, the K_r / K_a ratios, the corrosion rates, the default F* values, the 0.7 H and 8 ft minimum length and the bearing resistance equation are unchanged. Choosing the 2009 manual in Advanced mode restores its rules for the simplified method.

Checks

Example C1 of the 2023 manual (level backfill, traffic; pp. C-2 to C-43): every printed external value (forces, the three sliding ratios, the four bearing pressures, the three eccentricities) and, for steel strips, every printed value of the simplified method (K_r, σ_H, T_max, rupture and pullout ratios, pullout resistance) to the printed digits; the coherent gravity values to the printed digits when γ_EV is put on the vertical thrust as the example does; the stiffness method (Φ_g, Φ_local, D_tmax, tension, strain, ratios) to the printed digits when the example's load factor of 1.20 on the surcharge at the service limit state is used in place of the 1.00 of Equation 105. Example C2 (infinite 2H:1V backslope; pp. C-44 to C-83): external values and the simplified method to the printed digits; the coherent gravity pullout values to the printed digits. The 2009 manual's Examples E1, E3 and E4 remain as tests of the 2009 option. An independent numpy and scipy computation of 72 seeded walls over both editions and the three methods agrees to nine digits.

Notes on the source

Example C2's stiffness-method tensions use the full height of the soil above the back of the reinforcement (13.5 ft) where the text and AASHTO's definition, quoted on p. 137, call for the average height S_eq (5.25 ft, which the example computes and prints as "52.49 ft"); the page follows the text, and those printed tensions are not reproduced. The same example applies a pullout resistance factor of 0.90 and no resistance factor in tension where Table 22 gives 0.70 and 0.80 for the stiffness method, which Example C1 uses. Both examples test the eccentricity against L / 3 where the text and Equation 57 give L / 4. Table 20 prints N_c = 37.9 at 39 degrees for 67.9, as the 2009 manual did. The text (p. 57) says F* of steel strips "can be conservatively taken as 2.0" at the top while Table 8 gives 1.2 + log C_u, not more than 2.0; the page takes 2.0 unless C_u is entered. The 2023 manual gives no numerical example of a geosynthetic by the simplified method, of sliding along a geosynthetic sheet, of a broken backslope, of wire grids or of a computed bearing resistance: those parts rest on the equations, the 2009 examples where the rule is unchanged, and the independent numerical check.

Choices made here that the manual does not state

The simplified method as the opening method for both reinforcement types. A traffic surcharge together with an unlimited slope (the manual counts none). A broken backslope whose crest lies beyond the reinforcement. H_1 − H limited to the slope height for a broken backslope. The second bearing combination printed beside the first. Sliding along the lowest sheet computed on the block above that sheet. The foundation cohesion left out of sliding. The search for the least length. A minimum embedment length kept from the 2009 manual.

What it assumes

A wall with a face within 10 degrees of vertical, reinforcement of one length in every layer, select granular reinforced fill without cohesion and without water pressure, a retained fill without cohesion, level ground in front of the wall, static loading at the Strength I limit state, and a foundation that fails, if it does, in general shear. The reinforced zone is taken as rigid for external stability. The corrosion rates hold only for fill within the manual's electrochemical limits and for strips at least 3.5 mm thick or wires not smaller than W7; the default pullout factors hold only for fill within its gradation limits.

When it misleads

The most common failures of MSE walls are not the ones computed here: water in the reinforced fill, poor fill or compaction, a weak or sloping foundation, and slip surfaces that pass behind or through the reinforced zone. Overall (global) and compound stability are not checked, and neither is settlement; a wall on soft ground, on or above a slope, with a second wall above it, or with water behind it can pass every ratio on this page and still be inadequate. The computed bearing resistance ignores a sloping toe, groundwater, layered soils, local shear and lateral squeeze. A bearing resistance, friction angle, pullout factor, stiffness or reduction factor that is assumed and not measured carries its error straight into the ratios: the opening values are those of the manual's examples and are not properties of any site or product. The three internal-stability methods give different tensions for the same wall, the stiffness method by a wide margin: a result is tied to the method chosen and to the owner's acceptance of it. Bridge abutments on the wall, tiered and back-to-back walls, traffic barrier impact, earthquake loading and scour are treated in other chapters of the manual and are not covered. A result here is a check of a trial section, to be reviewed by the engineer responsible for the wall.