**1. Introduction**

The age distribution of the existing bridges in combination with several normative changes and significant economic and demographic changes in many industrial countries lead to many structures showing damages. Or else, there are computational deficits due to increased traffic or stricter verification [1–7]. In order to extend the remaining service life of the currently deficient structures, refined design concepts can provide a remedy that allows higher computational load-bearing capacities. In many countries, strategies have been developed for bridge assessment including monitoring, maintenance and design evaluation [8–13]. The aim is to extend the planning horizon in order to postpone the construction of a new replacement bridge for at least a part of the affected bridges. If some of the required verifications cannot be met even after a more detailed structural analysis during bridge assessment, strengthening measures can provide pertinent solutions. For reinforced or prestressed concrete bridges, which represent the majority of the German bridge population [3,14], established strengthening methods are available, such as additional external prestressing [15], insertion of shear connectors, additional concrete in the compression zone or external application of CFRP

sheets or lamella [16–18]. Innovative materials, such as UHPC [19–22] or textile-reinforced concrete (TRC) [23–25] enable considerable material savings through increased performance. This results in a more resource-efficient use of materials with a reduced additional dead weight and an extended service life. In consequence, the substructures and foundation of the strengthened bridge are exposed to less extra loading while CO<sup>2</sup> emissions from initial construction are spread over a longer period of use. Thus, the ecological footprint of concrete structures might be improved.

The findings presented in the following were part of a project aiming at developing a thin TRC layer which is intended to be added on bridge deck slabs between the RC structure and the road surface. It is called SMART-DECK and offers a three-fold functionality comprising monitoring, corrosion protection and strengthening of the transverse system of concrete T-beam or hollow core bridges. Strengthening is subject to the present paper. In terms of the gain in capacity due to the additional TRC layer, not only regular flexural and but also shear tests were conducted. The results of the experimental investigations are presented later. Prior to the presentation of the conducted investigations, the concept of the strengthening layer is described in Section 2. At the beginning, basic information on TRC is provided to give an understanding of the choice of material and the layout of the TRC layer.

The aim of the study is outlined in detail in Section 2.2, describing the specifics of SMART-DECK and outlining its distinction towards other studies addressing strengthening using TRC. The experimental programme is introduced in Section 3, beginning with small-scale tests. In Section 4, large-scale tests on slab segments under static and cyclic loading are presented. The paper concludes in Section 5 with a brief summary and prospects regarding potential future investigations.

### **2. Concept of the Strengthening Layer**

### *2.1. Textile-Reinforced Concrete*

Construction with reinforced concrete is rather economical and characterised by its versatile shaping. The two main elements—concrete and steel reinforcement—are combined to exploit their full potential when interacting with each other. One main disadvantage is the predisposition of the steel reinforcement for corrosion. If the concrete surface is damaged or chlorides enter the structure, the reinforcement can be damaged despite sufficient concrete cover. For this reason, research is concerned with the development of corrosion-resistant reinforcement elements. In the last two decades, research on corrosion-resistant textile reinforcements has been intensified. Early applications from Japan on non-metallic carbon mesh reinforcement [26] inspired research in the Collaborative Research Centre (*Sonderforschungsbereich*, *SFB*) 528 at TU Dresden, Germany, on strengthening of existing structures [27–30], while fundamental research on new construction elements utilising this then new technology was conducted in SFB 532 at RWTH Aachen University, Germany, [31–35].

The reinforcement in TRC is comprised of continuous fibres, which are called filaments (Figure 1). These are characterised by their good mechanical properties such as high tensile strength and, depending on the base material, medium to high modulus of elasticity. Depending on the desired fineness and the application, up to thousands of filaments are assembled to form a roving. The most common production method for non-crimp reinforcements is warp-knitting, where one or multiple rovings are arranged equidistantly and fixed in their position by additional knitting yarns, forming two-dimensional (2D) mesh-like reinforcement textiles. In biaxial textiles, warp rovings run in the longitudinal direction of the mesh, while weft rovings are arranged perpendicular to it.

Alkali-resistant (AR) glass fibres and carbon fibres are preferably used for TRC. Both materials have good mechanical properties, but carbon fibres are more durable than AR glass fibres [36]. Carbon fibres are characterised by their low density, a very high tensile strength and a high modulus of elasticity. They are manufactured in a multi-stage thermal process which determines the strength and stiffness of the fibres. Polyacrylonitrile (PAN) is usually used as the starting material for carbon fibres [37].

**Figure 1.** Principal sketch of a larger number of single filaments assembled to a roving and then to a yarn.

Bond of non-impregnated textiles is achieved by adhesion and friction rather than mechanical interlock, which is the typical mechanism for conventional deformed steel rebar. Due to the circular shape of the filaments and their small diameter, small voids exist in the roving. Because of their small size, cementitious binder cannot penetrate the roving and only encloses the edge filaments. The core filaments can only be activated for load transfer by friction on the edge filaments. Due to the smooth surface of the fibres, this bonding mechanism is ineffective and results in a telescopic failure with a stress gradient over the cross-section. An impregnation with low-viscosity organic or mineral binders fills the voids and allows for simultaneous activation of the core and sleeve material [34,38]. The most common impregnation materials are epoxy and styrene-butadiene. Also, new concepts with cementitious impregnation exist [39]. While meshes with styrene-butadiene impregnation maintain a certain flexibility but feature a rather soft bond to the concrete or mortar matrix, an epoxy resin provides a relatively rigid material behaviour which forms a strong bond to the cementitious matrix. At the same time, the latter allows for a much more straight-forward production since the reinforcement can be assembled in a formwork similar to reinforcing steel and the concrete can be regularly cast. It only needs to be ascertained that buoyancy of the textiles during compaction is prevented. Also, textile reinforcement must not be walked on due to its sensitivity to lateral compression. To a certain extent, this makes CFRP textiles with epoxy impregnated carbon rovings comparable to carbon FRP reinforcement. After concreting, it can be summarised under the term carbon concrete. The high potential of carbon concrete as an effective construction material has already been given prove of in numerous research projects and applications [40–47].

### *2.2. Characteristics of SMART-DECK*

### 2.2.1. Intended Functions and Main Questions

The system SMART-DECK presented here was developed in a research project in cooperation with a total of seven project partners from research and building practice as well as administration and material production. It is added to the top of the bridge between the concrete slab and the road surface (Figure 2) and enables bridge strengthening with TRC and offers two other functionalities which were investigated by the project partners: moisture monitoring and preventive cathodic corrosion protection (pCCP) [48,49]. The overall investigations not only foculised the target functionalities but also the feasibility and manufacturing of the system on a bridge [50]. The strengthening effect was investigated at the Institute of Structural Concrete of RWTH Aachen University (IMB) and it is the central content of this paper. In contrast to other investigations with regard to the strengthening potential of bridges with TRC, SMART-DECK not only addresses flexural but also shear capacity of the bridge's transverse system. These represent the main failure cases in ultimate limit state design of the bridge deck slab. So far, the impact of shear strengthening with TRC has been investigated with respect to the longitudinal system of the bridge [25,51–53], i.e., by supplementing the cross-section of the stirrup-reinforced webs. Since shear behaviour of slabs without shear reinforcement differs from that of beams with shear reinforcement [54], insights gained from previous results cannot simply be transferred to the transverse system. Comparable slab strengthening systems [55–57] addressed

the increase in flexural capacity. For bending, significantly more pronounced tensile stresses can be assumed in the textile reinforcement compared to shear being the critical load case. Since the very high tensile strength is the essential characteristic of the carbon reinforcement, the question remained to which extent a significant increase in capacity could be achieved by increasing the flexural reinforcement when premature shear failure occurs.

**Figure 2.** Position of the additional textiles on an exemplary bridge deck slab.

Another research question was the manufacturability of SMART-DECK. A common method for TRC production is shotcreting [52,58,59]. Since SMART-DECK is realised in a horizontal plane, shotcreting is not a suitable solution due to ingress of rebound. Therefore, using shotcrete would have conflicted with technical rules for German bridges [60] in this application and was therefore excluded. A production via laminating (layer by layer) is labour-intensive, so regular concreting was preferred from the beginning. Initially, it was not clear whether this method would meet the various technical demands resulting from the implementation of the instrumentation for monitoring and pCCP or a possible device for (partially) automated manufacturing. The manufacturing method was also linked to the development of suitable spacers and fasteners for the reinforcement which do not interfere with the electrical current.
