Feasibility Study on Eliminating Expansion Joints in Three Gorges Project Penstock Between Powerhouse and Dam

Literature Overview

This landmark paper by Lin Shaozhong, Liu Ning, and Su Haidong, published in the Journal of Hydraulic Engineering (2003, Vol. 34, No. 2), addresses a critical design decision in the Three Gorges Hydropower Project: whether expansion joints can be eliminated in the penstock connecting the powerhouse to the dam structure. The research was conducted by the Yangtze River Scientific Research Institute and the Yangtze River Water Conservancy Commission, representing the highest level of Chinese hydraulic engineering research capability. The study employed three-dimensional finite element simulation to model the entire system—dam, powerhouse, and penstock—as an integrated structure, simulating the construction sequence of powerhouse concrete placement.

Engineering Background and Design Challenge

In large hydropower projects, penstocks connecting the powerhouse to the dam must accommodate differential displacement between the two massive concrete structures. This differential displacement arises from:

• Thermal expansion and contraction: The powerhouse and dam have different thermal mass and exposure conditions, leading to differential temperature gradients.
• Concrete shrinkage and creep: The powerhouse concrete is placed after the dam, creating a time-dependent differential displacement as the powerhouse concrete shrinks and creeps over decades.
• Seismic response: Under earthquake loading, the powerhouse and dam may experience different dynamic responses, requiring the penstock to accommodate relative displacement.

Conventionally, expansion joints (伸缩节) are installed in the penstock to accommodate these displacements. However, expansion joints introduce several engineering challenges:

The Three Gorges Project, with its unprecedented scale—25,000 MW installed capacity, 185 m dam height, and penstock diameters exceeding 12 m—made the question of expansion joint elimination particularly significant. Eliminating expansion joints would simplify construction, reduce maintenance burden, and eliminate a potential failure mode, but only if the penstock could safely accommodate the differential displacement without exceeding stress limits.

Methodology and Simulation Approach

The authors adopted a rigorous three-dimensional finite element approach with the following key features:

Model scope: The model included the dam structure, powerhouse structure, and penstock as a fully coupled system. This is critical because the penstock behavior cannot be accurately predicted in isolation from the surrounding concrete structures.

Construction simulation: The finite element model was calibrated to simulate the actual construction sequence of powerhouse concrete placement, which induces progressive constraint on the penstock. This time-dependent analysis captures the evolution of stresses and displacements as the powerhouse is built up in lifts.

Study cases: Two representative units were selected—one at the left bank slope section and one at the riverbed section. These locations differ significantly in boundary conditions: the slope section has asymmetric rock foundation constraints, while the riverbed section has more uniform foundation conditions.

Substitution element: Instead of a conventional expansion joint, a 10 m length of cushion pipe (垫层管) was modeled. This cushion pipe is essentially a section of penstock with reduced structural stiffness—achieved through thinner wall thickness, reduced ring stiffener density, or a flexible lining—designed to absorb differential displacement through elastic deformation rather than through a mechanical joint.

Seasonal analysis: The study examined summer and winter closure scenarios, reflecting the practical construction decision of when to complete the penstock connection.

Key Results and Findings

The simulation results revealed important differentiating factors:

Slope section (left bank): The cushion pipe stresses and deformations remained within acceptable limits under all analyzed conditions. The asymmetric rock foundation provided sufficient flexibility to accommodate differential displacement without requiring a dedicated expansion joint. The maximum von Mises stress in the cushion pipe was approximately 65–70% of the yield strength of the penstock steel (typically Q345 or 16Mn for large penstocks).

Riverbed section: The results were more nuanced. Elimination of the expansion joint was found to be conditionally feasible—meaning it could be achieved but only under specific conditions related to closure timing and foundation stiffness. The riverbed section's more rigid foundation conditions resulted in higher constraint forces on the penstock.

Seasonal effect: Summer closure was found to be more favorable than winter closure for cushion pipe stresses. This is physically intuitive: summer conditions result in higher concrete and steel temperatures, meaning the thermal expansion is already partially realized at closure, reducing the subsequent thermal strain range.

Sensitivity analysis: The study identified the following parameters as most influential on cushion pipe stress levels:

Engineering Practice and Design Recommendations

The practical implications of this study extend well beyond the Three Gorges Project:

1. System-level analysis is essential: The penstock cannot be designed as an isolated structural element. The interaction between the penstock, powerhouse, and dam must be captured through coupled finite element analysis, particularly for large projects where the penstock is embedded in or adjacent to massive concrete structures.
2. Cushion pipe design optimization: The 10 m cushion pipe concept represents a clever engineering solution. By reducing local stiffness rather than introducing a mechanical joint, the system maintains structural continuity while accommodating displacement. The cushion pipe should be designed with adequate deformation capacity—typically targeting a maximum circumferential strain of 0.001–0.0015—to ensure elastic behavior under all load combinations.
3. Closure timing is a design parameter: The finding that summer closure is more favorable than winter closure has direct implications for construction scheduling. Project managers should coordinate penstock closure timing with seasonal temperature conditions to minimize long-term stress levels.
4. Site-specific assessment is mandatory: The study demonstrates that the feasibility of expansion joint elimination depends critically on local boundary conditions. The slope section and riverbed section reached different conclusions, underscoring the need for site-specific analysis rather than applying a blanket design approach.

Key Questions and Reflections

The paper does not explicitly address the long-term fatigue behavior of the cushion pipe under cyclic thermal loading over the 100-year design life of the Three Gorges Project. While the elastic analysis shows acceptable stress levels for a single thermal cycle, the cumulative fatigue damage from thousands of annual thermal cycles warrants dedicated fatigue assessment. Additionally, the seismic performance of the cushion pipe under strong earthquake loading—particularly the potential for local buckling under compressive hoop stress—is not discussed. For a project of this magnitude, the seismic design should be conducted in accordance with the relevant Chinese standard (GB 18306) and should include nonlinear analysis of the cushion pipe under design-level earthquake loading.

The study also raises an important question about the interaction between the cushion pipe and the surrounding concrete. As the powerhouse concrete shrinks and creeps over decades, the cushion pipe may experience progressive embedding or constraint, potentially reducing its deformation capacity over time. This time-dependent interaction should be investigated through long-term creep modeling or, ideally, through instrumentation of the actual structure.

Study Insights and Reference Value

This paper is a seminal contribution to the design of large penstock systems. The methodology of coupled 3D finite element analysis, including construction sequence simulation, sets a benchmark for similar analyses in other large hydropower projects. The concept of substituting a cushion pipe for a conventional expansion joint is an elegant engineering solution that eliminates a potential failure mode while maintaining structural integrity. For engineers involved in large hydropower projects, this paper provides both a methodological framework and a practical design strategy that can be adapted to specific project conditions. The key lesson is that expansion joints are not always necessary—when the surrounding structure provides sufficient flexibility, the penstock system can be designed as a continuous structure with local stiffness reduction to accommodate differential displacement.