Comprehensive Rehabilitation of the Punta del Este Shopping Center After Fire Damage

Abstract

On 6 August 2022, a fire devastated 80% of the Punta del Este Shopping Center in Maldonado, Uruguay. Originating in the kitchen of a supermarket, the fire ravaged the shopping center for 72 h before being brought under control. This article outlines the studies conducted to assess the fire’s impact on the building’s structure, as well as the strategies and rehabilitation project aimed at ensuring its stability and performance. An in-depth analysis of the concrete, reinforcements, and foundations was carried out using destructive and non-destructive testing. In total, over 150 concrete samples were collected for analysis, and the foundations were studied using indirect methods. Based on these analyses, actions were planned for each area, including structural repairs, reinforcements, or demolition. Due to the tight deadline for resuming commercial activity, special innovative structural solutions were designed to rehabilitate large severely damaged areas with the specific feature of avoiding demolition. This involved altering the static scheme of the structure, incorporating reinforcements and using slabs of the damaged structure as formworks. Complete demolition and subsequent reconstruction would have required timelines incompatible with the clients’ expectations, while the adopted solutions enabled the project to meet its objectives.

1. Introduction

The Punta del Este shopping center covers an area of over 45,000 m². Its geometry is complex, with a large, undulating, and inclined façade, as shown in Figure 1 (photo taken after the reconstruction). Since its construction in 1995, the shopping center had undergone several expansions and renovations, resulting in a single, unified building.

The original building employed cast-in-place concrete with a structural scheme mostly consisting of columns spaced 10 m apart. The main beams rested on these columns, and the secondary beams were spaced every 5 m, supporting square slabs measuring 5 × 5 m.

The structural solution naturally varied in terms of the façades, staircase zones, elevators, and various architectural details. In general, the building’s geometry resulted in a high degree of structural redundancy.

Unlike the original construction, the expansions, covering about 8000 m², were entirely prefabricated and isostatic, with prestressed beams and slabs. The building comprised a basement (SS), two main floors (ground floor (PB) and upper floor (PA)), and mezzanines for merchandise storage and technical walkways. The upper roof was made of lightweight materials with large skylights, most of which collapsed during the fire or suffered large deformations, rendering them unusable.

The building did not have fire control systems beyond manual extinguishers. At the time of its construction, it was not mandatory to have sprinkler and detection systems as it is today. When the regulations changed, the fire department provided a deadline for commercial buildings to comply with the new provisions. The building was still within the legal timeframe. Furthermore, the pumping and water reserve system had already been installed, but unfortunately, the piping was still under construction.

The fire began in the bakery section of the supermarket, located in an extensive basement that served as a storage area for the supermarket’s goods, in addition to housing other commercial establishments, the food court, and the cinemas. This area contained a large amount of highly flammable materials, which caused the fire to spread rapidly, inflicting the greatest damage in the basement [1]. The fire lasted for 72 h before being brought under control, unlike other reported cases of fire damaged structures [2,3,4,5,6,7], with only few hours of fire exposure. The fire caused the collapse of an area of approximately 4000 m² in the expansion zone and severe damage to 80% of the construction. Figure 2 shows images of the fire and the condition of some areas afterward.

After the fire, the damaged areas were categorized by the National Fire Department into three zones, namely, red, yellow, and green, based on the degree of deterioration and the risk of collapse, according to a preliminary visual inspection as proposed by many authors and informed by Balša Jovanović et al. [8]. Access was prohibited to the red zone and access to the yellow zone was only allowed with extreme caution. Figure 2 shows images of the fire, and Figure 3 presents the zones identified by the National Fire Department. Once the risk areas were delimited, the next step was to conduct a careful visual inspection of the area, followed by planning and carrying out destructive and non-destructive tests, necessary to conduct a forensic evaluation of the fire damaged structure and for the structural analysis. [9,10]. Quantifying the extent and gravity of the fire damage is very important for the definition of rehabilitation plans or demolition of some parts, as indicated by Stochino el al. [11].

The city of Punta del Este has a permanent population of approximately 30,000 inhabitants, but during the high tourist season from December to March, the population exceeds 500,000 residents. Therefore, the objective and challenge of reopening the shopping center by December 2023 was crucial, as that month and the following two months account for 75% of annual sales volume.

This posed a significant challenge, as it required almost completely rebuilding a 45,000 m² shopping center within 14 months. The months of August and September 2022 were needed for contracting and insurance inspections.

This project aimed to restore the structural and commercial functionality of the shopping center while ensuring compliance with safety standards and maintaining the building’s architectural identity. These objectives required innovative approaches, including integrating prefabrication methods, rapid construction workflows, and multidisciplinary collaboration.

Rehabilitation projects following structural failures due to fire, such as the Grenfell Tower fire in 2017 and the Notre Dame fire in 2019, emphasize the complexities involved in restoring critical infrastructure. However, the Punta del Este shopping center presents a unique case due to its hybrid structural design—combining cast-in-place concrete with prefabricated isostatic elements—and its time-sensitive nature driven by economic pressures.

Unlike standard reconstruction projects, this work required addressing the structural and architectural design nearly simultaneously with the start of construction tasks. It involved employing various techniques and structural solutions, such as self-compacting concrete poured on-site, precast concrete, shotcrete, steel structures, and specific solutions for several points of the structure.

The project also required changing the structural solution for the large slab over the basement. Initially planned as a continuous slab supported by beams spaced every 5 m in both directions, it was redesigned as a flat slab supported directly on columns spaced 10 m apart in both directions. This solution stands out from other post-fire structural repair projects, where the structural scheme is not substantially modified, as reported in cases by Folic et al. [3], Ha et al. [4], Knyziak et al. [5], Dilek [7], Gosain [10], and Musmar [12].

The high degree of fire-induced damage had to be managed, and the logistical constraints imposed by the short timeframes available had to be resolved. These factors, along with the innovative methods employed, highlight the uniqueness of this case.

The rehabilitation works and overall project coordination were carried out by STILER S.A. Testing to determine the condition of the concrete in the structure was conducted by the Construction Department of the Institute of Structures and Transport, Faculty of Engineering. Structural design was jointly undertaken by STILER S.A. and the Magnone–Polio studio, while the architectural project was managed by MARQ Studio, in collaboration with various specialists in thermal, sanitary, electrical, lighting, and other disciplines. This interdisciplinary approach was crucial to achieving the project’s ambitious goals.

This article presents the challenges and solutions adopted during the rehabilitation of the Punta del Este shopping center, offering valuable insights for future reconstruction projects in complex buildings to be restored within very tight deadlines.

2. Preliminary Inspection and Testing

2.1. Preliminary Works and Visual Inspection

Given the tight deadlines, debris removal, construction, architectural planning, damage investigation, evaluation of the existing structure, and execution of the project were initiated simultaneously. Subsequently, the tasks for structural reconstruction were undertaken.

The first challenge was entering and inspecting the “red zone”, whose stability was considered highly precarious and was almost impassable due to the accumulation of various types of debris. To address this situation, an initial survey was conducted using drones, which provided a preliminary assessment of the problematic area.

Based on the results of this survey, a decision was made to clear an “access path” to the red sector in the basement (SS). The working method involved propping beams and slabs with metal towers, advancing only three meters at a time before completing the corresponding propping on all levels. This approach allowed the creation of a pathway to facilitate the visual inspection of the structure. As progress was made, the structural condition was inspected to determine what tests and additional measures would be necessary for each area.

The photo in Figure 4 shows a column in the red zone, located near the wavy façade close to the supermarket entrance, which was almost entirely collapsed. The aluminum profiles from the openings were deformed by the downward movement of the structure, and the reinforcement bars of the column were completely exposed and significantly deformed.

In this zone, the structure underwent significant deflection without collapsing completely, as occurred in the prefabricated section of the extensions. This was due to its high degree of hyperstaticity, which allowed for load redistribution. The demolition of this area, approximately 3000 m² across all levels, was deemed necessary as it was clearly irrecoverable.

A preliminary conclusion emerged regarding the disparity in fire resistance between the hyperstatic and isostatic solutions, and the superior fire performance of reinforced concrete compared to prestressed solutions. These findings align with the greater structural robustness of the former [13]. While the isostatic solution collapsed, the hyperstatic solution remained standing because it allowed for the redistribution of loads due to its high degree of redundancy.

As the most affected area became accessible, visual inspections revealed significant structural cracking, spalling over extensive surfaces, and substantial deformations that were even visible in the opening of the building’s expansion joints. Due to the temperature, the cement paste experiences physical and chemical changes, starting above 100 °C, causing the degradation, cracking, and strength loss of the matrix. The aggregates, on the other hand, are generally thermally stable until temperatures of 500 °C, causing cracking in the bond zone between the matrix and the aggregates, which also leads to strength loss [14]. The concrete experiences spalling at elevated temperatures due to its porosity [14,15,16]. This breaking off of the concrete surface occurs mostly when it is rapidly heated, as was the case here. The photos in Figure 5 provide a reference.

Initially, the possibility of demolishing and rebuilding the entire red zone was considered. However, this solution, aside from its cost, would have made it impossible to complete the project by the target date. To avoid complete demolition and in search of a better solution, many tests were conducted (rebound hammer, core sample extraction, and macro- and microstructural analysis of the concrete in the thin sections).

2.2. Testing and Evaluations

2.2.1. Macro- and Microstructural Analysis of Concrete Samples

Two types of concrete core samples were analyzed: one affected by the fire (identified as V7) and another unaffected (identified as V8). Figure 6 shows images of both samples, which contain metamorphic aggregates from the construction area. The cracks induced by fire are mostly seen at a microscopic level [17]. The concrete from sample V7 exhibits a series of fractures that affect the aggregate and occasionally extend to the matrix, which may be related to the expansion of aggregates when exposed to high temperatures [15]. Concrete in fire conditions can suffer severe damage due to thermal effects that primarily affect the cement paste [14]. At moderately high temperatures, the cement paste may expand when heated, but starting at 300 °C, contraction occurs due to water loss. However, aggregates continue to expand, and internal stresses can lead to strength loss, cracking, spalling, and surface damage [15,16]. Different samples from affected and unaffected areas were analyzed, looking for a color change in the concrete as reported by Short NR et al. [18], but no differences could be found.

2.2.2. Mechanical Testing of Concrete and Steel Samples from the Structure

The most reported way to determine the residual strength of the concrete after a fire is by testing core samples from the affected structure. This is a destructive testing method that provides precise information about the strength of concrete in the structure [2,3,4,5,6,9]. According to [8], the test results of cores must be considered with caution, as they can underestimate the damage or even fail to detect it. This can be explained by the thermal and damage gradient experienced by the concrete perpendicular to the exposed surface. The strength in the length of the core may vary. Also, by preparing the core to be tested by sawing its top and bottom, the damage zone could be eliminated [7].

A complementary study was carried out with the rebound hammer as a non-destructive test. This is a test method that can be used to obtain a correlation between the compressive strength of the concrete and its surface hardness, as widely reported. It can be used to compare affected and non-affected areas of the structure, but not to determine the concrete compressive strength. As it gives information of the surface hardness (on a limited depth of concrete), it should only be used for damage detection after fire [10,19,20].

The residual strength of the reinforcement can only be met by destructive methods; as such, it is not commonly assessed, although it is very important [8].

The compressive strength of the concrete samples from the structure was determined according to UNIT-ISO 1920-6:2019 [21], the rebound hammer test was carried out following UNIT-ISO 1920-7:2004 [22], and the modulus of elasticity was measured following UNIT-ISO 1920-10:2010 [23]. Similarly, the yield and ultimate strength of steel were assessed using the structural steel testing protocol PEC.EDM.031, based on ISO 6892-1 [24]. These tests were conducted for various levels of fire damage identified by the fire department (red, yellow, and green zones). It is worth noting that in some cases within the red zone, core samples could not be extracted because the heavily cracked concrete disintegrated under the action of the extraction equipment.

According to the original plans, the structural concrete was classified as C30.

Through statistical analysis, the results revealed significant variability in the compressive strength across all structural elements, especially in the red zone. There was no specific area with consistently low or very low values (except for one, as reported later). Instead, the low results were heterogeneously distributed without any discernible pattern. This phenomenon was attributed to variations in fire load across different areas, which depended directly on the quantity and nature of goods stored in each sector and as indicated before to the core affection and preparation [7].

Over 150 concrete samples were collected for analysis.

Table 1. Test results of some of the core samples extracted from the beams in the red zone of the basement.

Id.	σ (MPa)	E (GPa)	RH Index	Obs.
V1	16.5	4.9	33	-
V2	-	-	-	Discarded due to excessive cracking, disassembled during extraction
V3	14.4	-	43	Longitudinal crack
V4	17.8	7.0	42	Longitudinal crack
V5	40.9	22.1	57
V6	32.8	-	56
V7	19.6	-	50
V8	25.9	-	43	Irregular surface due to extraction
V9	21.1	-	44
V10	9.2	-	30	Longitudinal non-through cracks
V11	10.9	3.5	31	Longitudinal non-through cracks
V12	18.3	-	51	Tested with 2 ø 10, perpendicular to the axis
V13	25.5	-	34
V14	20.5	-	40	Longitudinal through crack
V15	21.4	-	35	Longitudinal non-through cracks
V16	25.3	-	34	Longitudinal non-through cracks
V17	20.1	-	38	Longitudinal non-through cracks
V18	10.6	-	32	Longitudinal through crack
V19	31.6	15.2	-	Longitudinal non-through cracks
V20	25.6	-	-	Longitudinal through crack
V21	26.4	-	-	Longitudinal through crack
V22	26.4	-	-	Longitudinal through crack

presents the first 22 core samples extracted from the beams in the basement of the red zone and tested. It is noteworthy that some results, although not showing substantially low compressive strength—such as samples 20, 21, and 22—exhibited through-cracks.

In general, many samples from the basement showed cracks, some of them through-cracks, and in some cases, as previously mentioned, it was impossible to extract samples. The compressive strength results, excluding samples from the area deemed essential for demolition, were not as low as might have been expected based on a visual analysis of the cracks. However, there were localized points where the cores could not be extracted due to the state of the concrete or where very low results were obtained. Before extracting the cores from the structure, tests with the rebound hammer were carried out, which also showed mixed results (Table 1) and were not considered for further evaluation of the structure.

Given this situation and considering the following points, it was concluded that the basement structure did not have an adequate level of reliability and would need to be either fully reinforced or demolished:

• Extremely low results in an area with values of 9 MPa for compressive strength and 3.4 GPa for the modulus of elasticity.
• The impossibility of ensuring that a single element, such as a column, had not been affected differently along its height, with some areas potentially having lower strength than others.
• The extreme degree of cracking throughout the area (photos in Figure 5; Table 1).
• The literature has reported compressive strength losses in conventional concrete exposed to high temperatures of approximately 20% at 400 °C, 50% at 600 °C, and up to 80% or more at 800 °C, as well as significant variability in results. Similarly, the literature has reported modulus of elasticity losses ranging from 50% at 400 °C to up to 80% at temperatures exceeding 600 °C [25,26,27].

The increase in cement paste porosity with temperature, caused by the dehydration of hydrated calcium silicates (HCS) [28], along with cracking in both the paste and the aggregates explains why the reduction in the elasticity modulus of concrete is proportionally much greater than the loss of strength.

The tests conducted on the steel bars showed yield and ultimate strength values of 439 MPa and 716 MPa, respectively, for a nominal diameter of 25 mm; 474 MPa and 676 MPa for a nominal diameter of 8 mm; and 418 MPa and 615 MPa for a nominal diameter of 10 mm, effectively meeting the originally required values of 420 MPa and 500 MPa, respectively. The tests were performed by LATU (Technological Laboratory of Uruguay).

The EN 1992-1-2:2004 standard [29] indicates a reduction in concrete strength as temperature increases. Xargay et al. [28] compile data on the loss of strength and modulus of elasticity of concrete as a function of temperature, basing their work on the research of various authors.

Considering the test results for strength and modulus of elasticity shown in Table 1, and assuming the concrete was C30 (as specified in the original structural plans), it can be estimated—based on the reported losses—that the temperatures in the structure did not exceed or only slightly exceeded 600 °C.

Exceptions were identified at certain points where extreme strength loss coincided with very significant visible deterioration, indicating that these areas of the structure were beyond repair.

At 600 °C, there is little to no loss of steel strength, or only marginal loss, as indicated by Pereira et al. [30] and based on the studies by Guo and Shi [31] and Maraveas [32]. This explains the results obtained from the steel bars.

3. Recovery Project

For the calculations and design of the structure, the finite element software CEDRUS version 8, by the firm CUBUS, was used, and the entire structure was modeled in BIM.

3.1. Red Zone

Actions were taken for each of the floors within the red zone.

3.1.1. Basement (SS)

Based on the previously presented results, it was concluded that the structure in the red zone of the basement (approximately 6000 m²) was not reliable in any part.

To repair the structure, the following possibilities were analyzed:

• Complete demolition.
• Reinforcing beams and columns through “jacketed reinforcements” and reconstructing slabs according to the degree of deterioration observed in each one, either reinforcing them or demolishing them.

Neither of these proposed solutions could meet the desired deadlines. The first option implied reconstructing the entire sector across all its floors (18,000 m²), while the second required demolitions in the basement, reconstruction of many slabs, and jacketed reinforcements for practically all beams and columns.

Faced with this situation, an original solution was proposed: altering the structural scheme to create a slab without beams. The required work involved the following:

• Jacketed reinforcements for all circular columns (60 cm diameter): these were transformed into square sections with 100 cm sides. Prior to this, the foundation caps were uncovered, and eight φ25 rebars were inserted with chemical anchorage penetrating the jacketed reinforcements, as shown in Figure 7. For the jackets, C40 self-compacting concrete was used (strength according to UNIT 972 [33]) to avoid filling difficulties. These “jackets” effectively function as the new columns, since the cracking observed in the existing columns rendered their contribution to structural resistance negligible. The foundations were verified as described later
• Elimination of beams as structural elements.
• Construction of a capital to support the slab without beams: the capital consisted of four pilasters at the top of the jacketed columns, as illustrated in the 3D image in Figure 8 and the plan details in Figure 7.

After jacketed reinforcements were applied to the columns and the capitals were constructed, the beamless slab was built with spans of approximately 10 m × 10 m. This slab featured reinforced ribs around its entire perimeter and strong reinforcements above the columns (Figure 9).

This approach allowed the work to proceed with great speed, avoiding total demolition and the challenging repairs required for multiple slabs. Finally, to prevent the existing slabs and beams—which, although present, had no structural function—from detaching, they were “suspended” from the new beamless slab using reinforcements. In areas where reinforcements were exposed due to spalling, primarily aesthetic repairs were carried out by projecting concrete.

3.1.2. Ground Floor (PB)

Based on the tests conducted in this area, it was concluded that the slabs and beams were reliable, with some exceptions that required localized repairs. The PB columns were jacketed similarly to those in the basement, but the beams and slabs were preserved, and necessary repairs were made, primarily due to spalling. However, the level of detachment was not as severe as in the basement. Concrete spraying was used for slab repairs, as in the basement. For the beams, bonding agents and pre-mixed mortars were applied to restore the reinforcement coating.

3.1.3. Upper Floor (PA)

In this area, the test results did not show significant reductions in strength. Additionally, the imposed loads were much lower compared to other floors, as this level only supports a roof made of lightweight panels and skylights. Isolated repairs were carried out without the need to jacket the columns.

3.1.4. Foundation Verification

The described solution adds considerable weight to the foundations since the existing structure is preserved and a new one is added. Therefore, it was essential to ensure that the existing foundations could support the new loads. The following measures and tests were undertaken:

• The existing overloads (masonry and technical walkways) were reduced as much as possible. The original structure was designed for a dead load of 300 kg/m² plus the self-weight of the concrete. The masonry, consistent with the standard at the time when the building was constructed, was built with solid blocks. The damage it sustained during the fire and the new area distribution resulting from the building’s functional adaptation required its replacement. In this regard, it was replaced with dry masonry, which is significantly lighter.
• The upper floor had a subfloor with variable thickness that was always greater than 25 cm. This was removed, which also solved the level discrepancies caused by the new slab.
• All technical walkways made of concrete were replaced with metal walkways.

Despite the measures taken, the loads turned out to be greater, and there was no prior information available about the foundations. Therefore, the following steps were taken:

• Since it was necessary to uncover the pile caps to insert the φ 25 bars and support the casing, it was possible to inspect all of them. None showed any damage; they were located approximately 1.50 m below the floor slab and performed perfectly when drilled to embed the anchor bars.
• Integrity tests were conducted on a sample of piles, yielding satisfactory results and providing information about their length.
• Once the length was determined, standard penetration tests (SPTs) were carried out on the soil to establish the bearing capacity of the piles. After penetrating a three-meter-thick sand layer, hard-consistency clay was encountered. The SPT results for the clay showed a gradual increase in resistance, from 45 blows to 100 blows along the length of the piles. This evaluation allowed us to calculate a load-bearing capacity of 120 tons per pile, which was found to be adequate for the project’s requirements.

3.2. Remaining Areas of the Building

The test results for the yellow zone did not show low values and exhibited less variability. The main issue observed was the loss of coating. Repairs were carried out using traditional methods, as previously described for the ground floor repairs in the red zone.

Meanwhile, the area that collapsed during the fire, which was initially a prefabricated section already mentioned, was rebuilt using a metal structure and slabs on steel decking to meet the project deadlines.

Two additional levels were constructed in this section, to which the cinemas and the food court were relocated. The space these facilities previously occupied in the basement was repurposed to construct a gym with an indoor swimming pool. This approach not only increased the total area of the shopping center but also significantly improved the layout and placement of the stores, aligning them more effectively with their intended functions.

The solution implemented for this area, constrained by the tight construction schedule, was as follows:

• New foundations constructed on piles, with composite columns consisting of a steel tubular section filled with reinforced concrete. This design allowed 75% of the design load (increased by the appropriate safety factors) to be supported by the reinforced concrete core.
• Steel beams bolted to the columns and interconnected, combined with steel deck slabs, which simplified the assembly process and eliminated the need for traditional formwork.

Figure 10 show images of the construction progress of the expanded area.

3.3. Area Adjacent to the Supermarket Entrance (Practically Collapsed)

The area adjacent to the supermarket entrance, which was practically collapsed due to its state of deformation and other failures, was demolished and rebuilt using a reinforced concrete structure.

4. Work Schedule

A highly condensed version of the work schedule is presented in Figure 11, which highlights the short execution time in comparison to the size of the project. The work was carried out in double shifts from Monday to Saturday, with a workforce that included over 500 people directly involved on-site.

The project was managed using a master plan created in Microsoft Office Project Professional 2021 (version 2411), with rigorous implementation of the Last Planner method.

5. Conclusions

The structural restoration of the Punta del Este Shopping Center, following the fire that occurred in August 2022, presented a unique challenge in terms of engineering design, project management, and timeline constraints. The extensive damage, particularly in the basement and prefabricated expansion zones, required innovative and technically sound solutions to ensure structural stability and meet the critical reopening deadline of December 2023.

One of the innovative structural rehabilitation approaches was the reuse of the damaged slab as formwork for the new beamless slab, which avoided the total demolition of that sector (18,000 m²).

This technique was made possible by incorporating capitals and reinforcing the columns, as well as the intensive use of self-compacting concrete (C40), which allowed for optimized execution times and significant cost reduction.

Additionally, both destructive and non-destructive inspection and diagnostic methods were employed, providing an accurate evaluation of the structural elements’ condition and the extent of the damage caused by the fire.

These procedures enabled the project to meet the projected deadlines, allowing the shopping center to resume operations within the planned timeframe.

This process also provided valuable insights into structural behavior under fire conditions. A clear disparity was observed in the fire resistance of hyperstatic (cast in place) and isostatic (prefabricated) structures, as well as a significant reduction in the compressive strength and elasticity modulus of concrete in some affected areas.

However, despite exposure to high temperatures, the reinforcements did not show any appreciable reduction in tensile strength or elasticity modulus. This resulted in a redistribution of loads that prevented collapse in several affected areas.

Finally, this case highlighted the lack of adequate equipment for the prevention and containment of fires in facilities of this magnitude, underscoring the need to review safety systems and comply with regulations in future renovations.

This article does not address the architectural reconstruction or the restoration of the roofs, façades, electrical systems, HVAC, plumbing, etc., or the building’s adaptations to new fire safety and evacuation regulations, which were considered in the reconstruction project.