Moment distribution method
The moment distribution method (MDM) is an iterative structural analysis technique used to determine bending moments, shear forces, and reactions in statically indeterminate beams and frames by distributing unbalanced moments at joints until equilibrium is achieved.[1] Developed by American engineer Hardy Cross and first published in 1930, the method revolutionized the analysis of continuous structures by providing a practical alternative to solving systems of simultaneous equations required in earlier approaches like the slope-deflection method.[1] This method's primary advantages include its simplicity and avoidance of matrix algebra or complex equation-solving, making it accessible for hand calculations in pre-computer eras, while still offering high accuracy with sufficient iterations; it has been widely applied in civil engineering for designing reinforced concrete and steel structures, though modern software has largely supplemented it for complex analyses.[1] Limitations arise from its iterative nature, which can be time-consuming for highly indeterminate structures without computational aid, and its assumption of linear elastic behavior. Despite these, the MDM remains a foundational tool in structural engineering education and practice for understanding moment redistribution in indeterminate systems.[1]Overview
Introduction
The moment distribution method is an iterative technique in structural engineering for analyzing statically indeterminate beams and frames by successively approximating the end moments until equilibrium is achieved at all joints.[1] Developed by Hardy Cross, it was first introduced in 1930 as a practical alternative to solving complex systems of equations required by classical methods like the slope-deflection approach.[2][1] The primary purpose of the method is to simplify the analysis of continuous structures where the degree of indeterminacy makes exact solutions tedious or error-prone, enabling engineers to balance moments at joints through repeated distributions without formulating large matrices.[1] At its core, the principle relies on initially assuming joints are fixed to calculate fixed-end moments from applied loads, then distributing any unbalanced moments to adjacent members proportionally to their relative stiffnesses, while transferring a portion of these moments (known as carryover) to the opposite ends of those members.[1] This process converges rapidly for most practical cases due to the inherent convergence properties of the iterative balancing.[3] The method finds primary applicability in the analysis of continuous beams and rigid-jointed frames under transverse loading, such as gravity or lateral forces, under the assumptions of linear elastic behavior, small deformations, and rigid connections at joints.[4][1] It remains a foundational tool in hand calculations and educational contexts, though modern computational methods have largely supplemented it for highly complex structures.[1]Historical Development
The moment distribution method was developed by American engineer Hardy Cross and first published in May 1930 in the Proceedings of the American Society of Civil Engineers under the title "Analysis of Continuous Frames by Distributing Fixed-End Moments."[2] This innovation addressed the challenges of analyzing statically indeterminate structures, which had become more prevalent in civil engineering as designs for buildings and bridges grew in complexity during the early 20th century. Building on foundational theories like the Euler-Bernoulli beam equation from the 18th and 19th centuries, the method provided a practical alternative to earlier approaches amid the rising demand for efficient indeterminate analysis techniques.[1] Initially, the method found widespread application in manual calculations for multi-span continuous beams, bridges, and framed buildings during the pre-computer era, spanning roughly from the 1930s to the 1960s.[1] Engineers valued its iterative process, which allowed for straightforward arithmetic-based adjustments to moments at joints without solving large systems of simultaneous equations, as required by the contemporaneous slope-deflection method. This accessibility made it a staple for hand computations in structural design practices, particularly for moderate-sized structures where precision could be achieved through successive approximations.[1] The method's evolution included extensions to accommodate more varied structural elements, such as non-prismatic members with varying cross-sections. By the mid-20th century, refinements incorporated handling of sway frames and support settlements, further broadening its utility. Although largely replaced by matrix stiffness methods in computerized analysis after the 1960s, the core principles of moment distribution have been integrated into modern structural software and spreadsheet tools as an approximation and verification technique for simpler indeterminate problems. Its historical significance lies in revolutionizing manual structural analysis, democratizing the design of complex frames and reducing computational burdens for engineers before digital tools dominated the field.[1]Core Concepts
Fixed-End Moments
In structural analysis, fixed-end moments (FEMs) represent the bending moments that develop at the ends of a beam segment when both ends are assumed to be completely restrained against rotation and vertical translation, under the applied loading. These moments serve as the initial unbalanced moments in the moment distribution method, providing a starting point for the iterative redistribution process to achieve equilibrium.[5][6] The calculation of FEMs relies on several key assumptions: the beam member is prismatic (constant cross-section and material properties along its length), deformations remain small (linear elastic behavior), and the material obeys Hooke's law with no plasticity. These conditions ensure that the beam theory, specifically the Euler-Bernoulli beam equation, applies without complications from shear deformation or axial effects.[5][7] For common loading cases, FEMs are derived using methods such as integration of the moment-curvature relationship or superposition of indeterminate beam solutions, enforcing zero slope and deflection at both ends. For a uniformly distributed load w over span length L, the FEMs are M_{AB} = -\frac{w L^2}{12} at the left end (A) and M_{BA} = +\frac{w L^2}{12} at the right end (B), where the signs follow the convention of clockwise moments as positive. This result arises from solving the differential equation EI \frac{d^2 v}{dx^2} = M(x) with boundary conditions v(0) = v(L) = 0 and \frac{dv}{dx}(0) = \frac{dv}{dx}(L) = 0, yielding symmetric moments that counteract the sagging tendency of the load.[5][6] For a concentrated point load P applied at a distance a from the left end (with b = L - a), the FEMs are M_{AB} = -\frac{P a b^2}{L^2} at the left end and M_{BA} = -\frac{P a^2 b}{L^2} at the right end, again using the clockwise-positive convention. The derivation involves expressing the moment diagram in segments, integrating twice for deflection, and applying the fixed-end boundary conditions to solve for the redundant end moments. Note that both moments are negative, reflecting the overall sagging effect, with magnitudes depending on the load's eccentricity.[5][8] Triangular loading cases, such as a linearly varying load from maximum intensity w at the left end tapering to zero at the right end, produce FEMs of M_{AB} = -\frac{w L^2}{20} at the left end and M_{BA} = +\frac{w L^2}{30} at the right end. The derivation follows a similar integration approach, with the load function q(x) = w (1 - x/L), leading to asymmetric moments due to the uneven distribution. For the reverse triangular load (zero at left, w at right), the signs and values swap accordingly.[6][5] The following table summarizes standard FEM formulas for quick reference, assuming a prismatic beam of span L, clockwise moments positive, and fixed ends at A (left) and B (right). These are derived from classical beam theory and are widely tabulated in structural engineering references.| Loading Case | FEM at A (M_{AB}) | FEM at B (M_{BA}) | Notes |
|---|---|---|---|
| Uniformly distributed load w | -\frac{w L^2}{12} | +\frac{w L^2}{12} | Symmetric sagging moments. |
| Concentrated load P at distance a from A (b = L - a) | -\frac{P a b^2}{L^2} | -\frac{P a^2 b}{L^2} | Asymmetric based on position; both negative for downward load. |
| Triangular load, w at A to 0 at B | -\frac{w L^2}{20} | +\frac{w L^2}{30} | Heavier loading at left increases left-end magnitude. |
| Triangular load, 0 at A to w at B | +\frac{w L^2}{30} | -\frac{w L^2}{20} | Reverse of above case. |
Member Stiffness
In the moment distribution method, member stiffness refers to the rotational stiffness of a beam or frame member, defined as the moment required to produce a unit rotation at one end while preventing translation at both ends. This property quantifies the member's resistance to rotation at a joint and is fundamental to distributing unbalanced moments during analysis.[1] For a prismatic member with the far end fixed against rotation, the stiffness K is given by K = \frac{4EI}{L}, where E is the modulus of elasticity, I is the moment of inertia of the cross-section, and L is the span length. If the far end is hinged (free to rotate but fixed against translation), the stiffness reduces to K = \frac{3EI}{L}. These expressions derive from the slope-deflection equations underlying the method, as originally conceptualized by Hardy Cross.[1][2] The stiffness depends on material properties ([E](/page/E!)), geometric characteristics (I and [L](/page/L')), with shorter spans or larger sections yielding higher stiffness. In practice, relative stiffness is often computed as I/L for members with fixed far ends, serving as a proportional measure when [E](/page/E!) is uniform across the structure; this relative value apportions moments at joints based on each member's share of the total joint stiffness. For the hinged case, the relative stiffness is (3/4)(I/L).[1] For non-prismatic beams with varying cross-sections or in frames permitting sidesway, the moment of inertia I is modified—typically by using an average value or through integration along the length—to accurately reflect the effective stiffness.[1]Distribution Factors
In the moment distribution method, distribution factors determine the proportion of the unbalanced moment at a joint that is apportioned to each connected member, ensuring equilibrium by balancing moments around the joint.[1] The distribution factor (DF) for a specific member is calculated as the ratio of that member's stiffness K to the sum of the stiffnesses of all members meeting at the joint, expressed as \text{DF} = \frac{K_{\text{member}}}{\sum K_{\text{joint}}}, where stiffness K represents the rotational resistance of the member, typically K = \frac{4EI}{L} for a prismatic member with fixed far end, and I is the moment of inertia, E the modulus of elasticity, and L the length.[1] This formulation originates from the iterative balancing process introduced by Hardy Cross in his seminal work on analyzing continuous frames. For a joint connected to members A, B, and C with stiffnesses K_A, K_B, and K_C, the distribution factor for member A is \text{DF}_A = \frac{K_A}{K_A + K_B + K_C}. Similar expressions apply to the other members, ensuring the sum of all distribution factors at the joint equals 1.[1] These factors are computed once for the structure prior to iteration, based on relative member stiffnesses that account for end conditions and geometry.[9] The distribution factor physically represents the fraction of the total unbalanced moment that each member absorbs during the balancing step, reflecting the member's relative capacity to resist rotation at the joint.[1] Stiffer members attract a larger share of the moment, promoting a more realistic load distribution in indeterminate structures.[4] Special cases arise depending on the number and properties of members at the joint. If only a single member connects to the joint, its distribution factor is 1, as it must absorb the entire unbalanced moment.[1] Conversely, for a member with zero stiffness (infinitely flexible, such as a hinge with negligible resistance), the distribution factor is 0, meaning it receives no distributed moment.[1] In the iterative process of the moment distribution method, distribution factors are applied after initial fixed-end moments are established, to apportion the unbalanced moment at each joint and achieve rotational equilibrium.[1] This step repeats across joints until the structure converges to a balanced state, with the factors ensuring proportional moment sharing based on stiffness.[10]Carryover Factors
In the moment distribution method, the carryover factor quantifies the portion of a moment applied at one end (near end) of a structural member that induces a moment at the opposite end (far end). It is defined as the ratio of the carryover moment to the applied moment at the near end.[11][12] For a prismatic member, the carryover factor is 0.5 when the far end is fixed against rotation, indicating that half the moment distributed at the near end transfers to the far end. In contrast, the carryover factor is 0 when the far end is hinged or free, as no moment develops there due to the release condition.[11][12] This transfer occurs due to the inherent symmetry in the bending behavior of beams under end moments, where the curvature and resulting moment diagram exhibit balanced distribution across the span for fixed-end conditions.[11] The general formula for the carryover moment is given byM_{\text{carryover}} = C \times M_{\text{distributed}}
where C is the carryover factor (0.5 for fixed far end, 0 for hinged), and stiffness adjustments may apply for non-prismatic members to account for varying cross-sections.[12][11] Carryover factors are derived primarily from the slope-deflection equations, which express end moments in terms of joint rotations and chord rotations: for a fixed far end with no rotation, M_{BA} = \frac{2EI}{L} \theta_A, leading to M_{AB} = 2 M_{BA} under equilibrium, thus C = 0.5. Alternatively, they can be obtained using the moment-area method, where the first moment of the M/EI diagram about the far end equals the tangential deviation, confirming the 0.5 factor through geometric compatibility.[11][13] Graphically, the carryover is illustrated by the bending moment diagram of a beam subjected to a near-end moment, which varies linearly from the full value at the near end to half that value (same sign) at the fixed far end.[14] The carryover factor ensures moment equilibrium within the member by transferring the unbalanced portion to the far end, where the carried-over moment has the same sign as the near-end distributed moment.[12]
Implementation Procedure
Sign Conventions
In the moment distribution method, a consistent sign convention is essential to ensure accurate analysis of indeterminate structures. Moments acting on the ends of structural members are considered positive if they tend to rotate the member end clockwise when viewed from the joint, and negative if counterclockwise. This convention, which aligns with that used in the slope-deflection method, facilitates the calculation of end moments where positive rotations at member ends produce corresponding positive moments.[15][5] For beams specifically, sagging moments—those causing tension in the bottom fibers—are taken as positive, which complements the end moment convention by ensuring compatibility with bending moment diagrams derived from the method. At a joint, equilibrium requires that the algebraic sum of all moments from connected members equals zero, with signs applied based on their direction relative to the joint. This means clockwise moments from one member contribute positively, while counterclockwise ones contribute negatively, balancing the joint without external rotation. Regarding member ends, the "near end" refers to the joint under consideration, where moments are initially unbalanced and distributed, while the "far end" receives the carryover moment. The carryover moment is typically half the magnitude of the distributed moment at the near end and retains the same sign, meaning a positive (clockwise) distribution at the near end results in a positive carryover at the far end. This sign retention reflects the method's assumption of linear moment variation along the member under small deformations.[5][15] A common pitfall in applying the moment distribution method arises from inconsistent sign usage, such as mixing clockwise-positive conventions with internal beam sign rules, which can lead to erroneous distributions and unbalanced joints. To illustrate, consider a simple beam joint: a clockwise arrow curving from the joint along the member indicates a positive moment, while an anticlockwise arrow denotes negative; for frames, similar arrows at column-beam joints show how signs ensure rotational equilibrium. Adhering strictly to this convention prevents such errors and maintains alignment with fixed-end moments, where signs follow the same directional rules.Iterative Distribution Process
The iterative distribution process forms the core algorithm of the moment distribution method, enabling the successive balancing of moments at joints through repeated approximations until structural equilibrium is achieved.[1] Developed by Hardy Cross in 1930, this procedure relies on fixed-end moments (FEMs), stiffness values, distribution factors (DFs), and carryover factors as foundational elements from the method's core concepts. The process begins with Step 1: calculating the fixed-end moments for all members under the applied loads, assuming all joints are rigidly fixed against rotation. These FEMs serve as the initial moment values for each end of every member.[1] In Step 2, member stiffnesses are computed, typically K = \frac{4EI}{L} for members with fixed far ends (where E is the modulus of elasticity, I is the moment of inertia, and L is the member length), adjusted to K = \frac{3EI}{L} if the far end is hinged. Distribution factors are then determined at each joint as DF_{ij} = \frac{K_{ij}}{\sum K}, where K_{ij} is the stiffness of member ij and the sum is over all members at the joint. Carryover factors are set to 0.5 for prismatic members with fixed far ends, indicating that half the moment distributed at one end is transferred to the opposite end.[16] Step 3 involves the primary balancing action: for each joint in sequence (often starting from a free end and moving toward fixed supports), compute the unbalanced moment as the algebraic sum of all moments at that joint from FEMs and prior carryovers. Distribute this unbalanced moment to the connected members proportionally according to their DFs, applying the opposite sign to restore balance (i.e., distributed moment to member ij = unbalanced moment × DF_{ij}). Then, carry over half of each distributed moment to the far end of the respective member.[1] Step 4 requires repeating Steps 3 across all joints in multiple iterations, updating the moment tallies after each full pass, until the maximum unbalanced moment at any joint falls below a predefined tolerance, such as 1% or 2% of the largest FEM in the structure.[16] Convergence in this iterative process is generally swift, often requiring 3 to 5 iterations for simple beams and up to 10 for more complex configurations, as the method's relaxation nature progressively reduces errors through successive approximations.[16] For faster convergence in cases with slow damping, relaxation factors (values between 0.5 and 1.0, such as 0.8) can be applied to scale the distributed moments before carryover, though the unmodified method suffices for most practical analyses. The final end moments for each member are obtained by algebraically summing the initial FEM, all distributed moments at that end, and all carried-over moments to that end across all iterations.[1] To illustrate the algorithm clearly, the following pseudocode outlines the process for a general structure:This pseudocode assumes clockwise moments positive and a non-sway frame; joint sequencing ensures carryovers propagate correctly without immediate rebalancing.[1]Initialize: Compute FEM for all members Compute [stiffness](/page/Stiffness) K for all members Compute DF for each member at each [joint](/page/Joint) Set carryover_factor = 0.5 Initialize [moment](/page/Moment) tally M[member_end] = FEM for all ends Set [tolerance](/page/Tolerance) = 0.01 * max(|FEM|) While max_unbalanced > [tolerance](/page/Tolerance): For each [joint](/page/Joint) in sequence (e.g., left to right): unbalanced = sum(M at [joint](/page/Joint) ends) If |unbalanced| < [tolerance](/page/Tolerance): continue to next [joint](/page/Joint) For each member connected to [joint](/page/Joint): distributed = -unbalanced * DF[member] M[near_end] += distributed // near end at [joint](/page/Joint) M[far_end] += carryover_factor * distributed // far end carryover Compute max_unbalanced = max over all [joint](/page/Joint)s of |sum(M at [joint](/page/Joint))| Final moments: M[all ends] as computedInitialize: Compute FEM for all members Compute [stiffness](/page/Stiffness) K for all members Compute DF for each member at each [joint](/page/Joint) Set carryover_factor = 0.5 Initialize [moment](/page/Moment) tally M[member_end] = FEM for all ends Set [tolerance](/page/Tolerance) = 0.01 * max(|FEM|) While max_unbalanced > [tolerance](/page/Tolerance): For each [joint](/page/Joint) in sequence (e.g., left to right): unbalanced = sum(M at [joint](/page/Joint) ends) If |unbalanced| < [tolerance](/page/Tolerance): continue to next [joint](/page/Joint) For each member connected to [joint](/page/Joint): distributed = -unbalanced * DF[member] M[near_end] += distributed // near end at [joint](/page/Joint) M[far_end] += carryover_factor * distributed // far end carryover Compute max_unbalanced = max over all [joint](/page/Joint)s of |sum(M at [joint](/page/Joint))| Final moments: M[all ends] as computed
Handling Framed Structures
The moment distribution method extends to rigid framed structures by addressing the additional degrees of freedom introduced by sidesway, which allows for horizontal displacements and joint translations not present in non-sway beams or braced frames. This sway arises in unbraced structures under lateral loads or unsymmetrical vertical loading, complicating the analysis as the fixed-end moments now depend on both joint rotations and translations. The primary challenge lies in ensuring equilibrium in the horizontal direction, requiring an iterative correction to account for these translational effects while maintaining the method's reliance on moment balancing at joints.[1] The procedure begins by assuming zero initial sway, treating the frame as non-sway and distributing moments due to applied loads to obtain preliminary end moments. From these, the resulting shear forces or horizontal reactions are calculated, revealing an imbalance that indicates the required sway correction. To resolve this, an arbitrary horizontal displacement (sway) is assumed for the frame, generating equivalent fixed-end moments in the members—particularly in columns—proportional to their stiffness and geometry. These arbitrary moments are then distributed across the joints using the standard iterative process, and the corresponding unbalanced horizontal shears are computed. The ratio of the initial shear imbalance to this arbitrary unbalanced shear determines a correction factor, which scales the distributed sway moments before adding them to the non-sway results, achieving overall equilibrium. This process may require iteration if residual imbalances persist.[17][1] For multi-story frames, the method involves handling multiple independent sway degrees of freedom, one per story, through successive corrections or simultaneous equations derived from shear balances at each level. Arbitrary displacements are applied story-by-story, with fixed-end moments distributed accordingly, and correction factors solved via a system of equations to match the total unbalanced shears. This approach assumes no axial deformations in members, constant flexural rigidity, and supports that are either pinned or fixed, neglecting relative translations at joints other than sway. In practice, for complex multi-story cases, the procedure can become cumbersome, often leading to matrix-based extensions of the moment distribution method for efficiency.[17] A representative setup for applying this extension is a single-bay portal frame under lateral loading, such as wind or seismic forces applied at the beam level, where the columns experience shear imbalance due to the horizontal force, necessitating sway correction to determine accurate end moments in beams and columns.[1]Applications and Examples
Beam Analysis Example
Consider a two-span continuous beam ABC with simple supports at A, B, and C, where each span AB and BC has a length L = 6 m and carries a uniformly distributed load w = 10 kN/m. The flexural rigidity EI is constant throughout. This example illustrates the application of the moment distribution method to determine the end moments, reactions, shear forces, and bending moment diagram. The method follows the iterative distribution process outlined in the core concepts, beginning with fixed-end moments assuming all joints are fixed, followed by balancing at the hinged end joints A and C to enforce zero moments there. The fixed-end moments (FEMs) for each span under uniform loading are calculated using the standard formula for a fixed-fixed beam: M_\text{FEM, left} = +\frac{wL^2}{12} and M_\text{FEM, right} = -\frac{wL^2}{12}, where positive moments are taken as clockwise. Substituting the values, \frac{wL^2}{12} = \frac{10 \times 6^2}{12} = 30 kNm. Thus, for span AB: M_{AB}^\text{F} = +30 kNm and M_{BA}^\text{F} = -30 kNm; for span BC: M_{BC}^\text{F} = +30 kNm and M_{CB}^\text{F} = -30 kNm.[8] Since the end supports at A and C are simple (hinged), the member stiffnesses must account for the hinged far ends during distribution, but the initial FEMs use the fixed assumption, with adjustments via balancing. The relative stiffness for each member with a hinged far end is K = \frac{3EI}{L}, so K_{AB} = K_{BC} = \frac{3EI}{6} = 0.5 EI. At joint B, the total stiffness is K_\text{total} = EI, yielding distribution factors (DFs) of DF_{BA} = 0.5 and DF_{BC} = 0.5. At the end joints A and C, with only one member each, DF = 1.0. The carryover factor (COF) to a hinged end is 0, but for the initial releasing steps assuming fixed far ends before full adjustment, COF = 0.5 is used sequentially.[15] The distribution process is summarized in the following table, showing the moments at each joint end (in kNm). The process begins with FEMs and balances the hinged ends in sequence, with carryovers. Due to symmetry, joint B balances automatically after the end releases, requiring no further iterations at B.| Step | M_{AB} (at A) | M_{BA} (at B, AB) | M_{BC} (at B, BC) | M_{CB} (at C) |
|---|---|---|---|---|
| Fixed-end moments | +30 | -30 | +30 | -30 |
| Balance at A (unbalanced +30, distribute -30 × DF=1.0) | -30 | |||
| Carryover to B (0.5 × -30) | -15 | |||
| Balance at C (unbalanced -30, distribute +30 × DF=1.0) | +30 | |||
| Carryover to B (0.5 × +30) | +15 | |||
| Final moments | 0 | -45 | +45 | 0 |