Modeling Post-Cooling in Concrete Dams with DIANA: Verification of the Equivalent Cooling Surface Method 

Understanding the Thermal Challenge 

Controlling temperature evolution in massive concrete structures remains one of the defining challenges of modern dam engineering. Within an arch dam’s immense concrete body, even a small thermal imbalance can decide whether the structure endures or fractures. The heat released during cement hydration drives steep temperature gradients between the dam’s insulated core and its exposed surfaces — gradients that can generate tensile stresses high enough to cause early-age cracking. Such cracking threatens watertightness, durability, and long-term structural continuity. 

Within this complex thermal environment, DIANA provides the numerical framework engineers rely on to simulate hydration heat, heat transfer, and post-cooling systems with scientific precision. To mitigate excessive temperature rise, engineers combine several thermal-control measures: pre-cooling of aggregates, the use of low-heat cements, and—most effectively—post-cooling systems: networks of embedded, water-filled coils that extract heat from the concrete mass as it cures. 

However, while these systems are standard practice in large dam construction, modeling them accurately in a finite-element context remains computationally demanding.

Traditional Discrete Cooling Coil (DCC) modeling captures the physical process in detail but at steep computational cost. It requires dense meshes, complex coupling between the flowing water and surrounding concrete, and long runtimes. Simplified equivalent-source approaches reduce computational effort but oversimplify heat removal, losing the local temperature gradients that govern stress development. 

This persistent trade-off between numerical fidelity and efficiency has long limited realistic simulation of post-cooling during dam construction. 

Introducing the Equivalent Cooling Surface (ECS) Concept 

To overcome these limitations, researchers from the University of Porto and the University of Minho formulated the Equivalent Cooling Surface (ECS) method directly within DIANA’s thermal analysis module. 

In ECS, discrete cooling pipes are replaced by convective thermal boundaries embedded directly inside the solid concrete mesh, as shown in figure 1. These surfaces exchange heat with a virtual water domain governed by effective convection coefficients derived from pipe geometry, flow rate, and thermal properties. The approach maintains physical realism while removing thousands of 1D pipe elements, drastically reducing model complexity. 

Figure 1. Cooling coils are represented as internal convective surfaces within DIANA, enabling accurate and efficient simulation of post-cooling.

Implemented natively in DIANA’s nonlinear transient analysis, ECS integrates seamlessly with staged-construction workflows, heat-of-hydration modeling, and boundary-condition definitions — allowing engineers to simulate large thermal problems with practical computation times while retaining DCC-level accuracy. 
 

Verification through DIANA’s Transient Thermal Analysis 

The ECS formulation was verified using DIANA by modeling a 19 × 19 m2 concrete block divided into two lifts. The objective was to compare ECS and DCC performance under identical thermal conditions. The physical layout and cooling-pipe arrangement used in the test are shown in Figure 2.  

To isolate internal cooling effects, the block was modeled under adiabatic boundary conditions, meaning no heat exchange with the external environment. This ensured that any temperature differences arose solely from the modeling approach, not from external influences — a crucial prerequisite for fair validation. 

Figure 2. Application: (a) geometry; (b) cross-section (in yellow).

Six temperature-monitoring points (P1–P6) were distributed throughout the block, from the center to the cooling plane and the outer surface. This arrangement allowed evaluation of temperature gradients in three dimensions and assessment of both local and global cooling effects. 

Thermal Modeling Framework and Boundary Conditions 

All analyses used DIANA’s nonlinear transient thermal solver, which computes the heat-balance equation through backward-Euler integration and Newton–Raphson iteration for convergence. The governing equation accounted for conduction, convection, and cement-hydration heat, modeled via DIANA’s Arrhenius-type hydration law that captures temperature-dependent reaction kinetics. 

Boundary conditions were implemented using Newton’s cooling law, representing convection and radiation across realistic construction interfaces. DIANA’s multi-boundary framework enables simultaneous assignment of distinct surface types with specific heat-transfer coefficients, automatically adjusting for staged construction, formwork removal, and insulation application. 

During simulation, formwork panels were removed after a defined curing period, while insulation blankets were activated or removed based on construction sequence. These evolving conditions were represented in DIANA through equivalent surface properties. 

Concrete material properties representative of dense dam mixes were used: 

  • Thermal conductivity: 2.6 W m⁻¹ K⁻¹ 
  • Specific heat: 2087 kJ m⁻³ K⁻¹ 

Two comparative models were created within the same DIANA project: 

  • DCC model: Discrete 1D pipe elements coupled with the solid concrete mesh. 
  • ECS model: 4-noded surface elements carrying equivalent convection coefficients. 
Figure 3. ECS (b) and DCC (c) models created in DIANA. The ECS formulation eliminates dense pipe meshing by applying equivalent convective surfaces inside the solid domain. 

Results: Validation of ECS within DIANA 

Temperature evolution at the six monitoring points (P1–P6) demonstrated a near-perfect match between DCC and ECS results. Across all points, deviations remained below 0.8 °C in most regions and within 3 °C near the cooling planes — well within the acceptable margin for engineering design, where tolerances of ±5 °C are typically considered thermally equivalent. 

Figure 4. Temperature evolution from block validation. Comparison of DCC and ECS results at monitoring points P1–P6 computed in DIANA. The curves show near-identical thermal behavior under adiabatic conditions. 

Beyond accuracy, ECS delivered a decisive computational advantage. The DCC model contained approximately 54 000 elements, while the simplified ECS model required only 7 000. As a result, ECS reduced computation time by 11.3 times compared with DCC and achieved smoother convergence across all time steps. 

DIANA’s solver stability under nonlinear thermal gradients ensured that both analyses remained numerically robust, but the ECS formulation demonstrated significantly faster runtime efficiency — a key factor for scaling to full dam-monolith analyses. 

Conclusion and Continuation 

The verification study demonstrates that DIANA’s Equivalent Cooling Surface (ECS) method provides a reliable and computationally efficient alternative to discrete coil modeling for simulating post-cooling systems in mass concrete. By embedding physically equivalent convective boundaries directly within DIANA’s 3D mesh, engineers can achieve high-fidelity thermal predictions with significantly reduced computational demand. 

With ECS validated under controlled, adiabatic conditions, the next step applies the method to a 111 m-high Portuguese arch-dam monolith, assessing its performance under real construction scenarios and field-monitored temperatures.