Biogenic acid corrosion of sewer concretes with different binders: in-situ and model studies, with advancement of the life factor prediction method

Doctoral Thesis

2023

Permanent link to this Item
Authors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
License
Series
Abstract
Biogenic Acid Corrosion (BAC) is the biodeterioration of concrete caused by biological and chemical activities of bacteria that grow in an acidic environment. Such environments are typical in sewer systems, which collect wastewater from households, industries, and urban and storm water runoff, and convey this to wastewater treatment plants. Since these sewer systems transport large volumes of sewage and are invariably buried under the ground, concrete pipes are customarily used. Concrete is widely known as a robust, flexible, and durable material in many aggressive environments, yet it can suffer from severe sewer corrosion. Numerous techniques have been employed to eliminate or control the problem, including applying chemical or biological agents that decrease acid production, and surface treatment techniques that inhibit chemical attack and acid penetrability into the concrete. However, these techniques are expensive and variably effective, and some may lead to a loss of structural integrity and performance during the long service life of more than 50 years, due to a lack of long-term performance testing before they are introduced into the market. The most reliable and affordable approach remains to modify concrete by incorporating readily available binders with high chemical and physic mechanical potential in resisting corrosion. Numerous studies have approached this technique using Portland Cement (PC) based systems with Supplementary Cementitious Materials (SCMs). However, other binder systems such as Calcium Aluminate Cement (CAC) and Calcium Sulpho-Aluminate cement (CSA) have been less studied, despite showing higher potential in suppressing biogenic corrosion. Furthermore, most biogenic prediction models cannot effectively predict the performance of these binders in a sewer environment, especially when they are incorporated with different aggregates. For instance, the Life Factor Method (LFM), which is a common and often preferred method in sewer concrete design, is not formulated to enable the prediction of the corrosion rate of concrete comprising CAC, or specific PCbased systems with SCMs such as slag, fly ash, and silica fume. The LFM model consists of two parts; the acid environment generation part, and the acid resistance part, also sometimes called the ‘material factor', which defines the potential of concrete to resist corrosion due to acid. However, the current material factor in the model involves only an alkalinity factor (total calcium oxide content in concrete) as a significant resistance provider, while not considering other chemical compositions that significantly contribute to providing concrete with acid resistance. A previous amendment of the model at the University of Cape Town (UCT) resulted in a 'material factor' approach that was highly empirical and complex, and therefore less practical and comprehensive in application. The model also could not handle modern binder systems in conjunction with aggregates, and the criteria for its refinement were somewhat self-contradictory. Therefore, this study aimed to fundamentally re-think and improve the LFM model to cover a broader range of binder systems and aggregate types with their performance in different sewer conditions. Firstly, the deterioration mechanisms of concretes with different binder systems and different types of aggregates were studied in different live sewer environments. Secondly, the mechanism of deterioration of the binder systems was further studied using a reactive transport modelling approach to understand the critical phases that govern the deterioration. Using the information from the first two goals, the LFM model was then improved and advanced to cover a wider range of binder systems, aggregate types, and sewer environments. The study characterised the behaviour of three sewer sites in the Cape Town Metro (i.e., Langa Pump station (LPS) manhole, Northern Area Sewer manhole 19 (NAS M19) and manhole 54 (NAS M54)), monitoring techniques (i.e., visual observations, mass and thickness change, and concrete surface pH), The study also evaluated the influence of different concrete mixes using concrete microstructural analysis (Scanning Electron Microscopy (SEM), Quantitative Evaluation of Minerals (QEMSCAN), and X-Ray Diffraction (XRD) analysis), and reactive transport modelling (i.e., HYTEC modelling tool). 2 The concrete mixes were grouped into two batches: LH concrete and UCT concrete mixes. The LH concrete mixes consisted of four binder systems, i.e., a blend of 80% Sulphate Resisting Portland Cement and 20% Fly Ash (SRPC+FA), a similar blend with 11% iron-based additive (SRPC+FA+HC), a blend of 80% SRPC and 20% ground Limestone (SRPC+LS), and CSA. These mixes were cast with calcite and siliceous aggregates by Lafarge Holcim (LH) in Lyon, France and delivered to the UCT laboratory for sewer exposure. Before exposure, it was observed that the mixes exhibited significant compaction voids, but their condition was such as to permit sewer exposure with the expectation of gathering useful information on these mixes. The UCT concrete mixes were prepared at UCT with local dolomite and siliceous aggregates, but using the same LH binders, i.e., SRPC+FA, SRPC+FA+HC, with some additional local binders, i.e. Portland cement blended with limestone (CEM II A-L), a blend of 50% CAC and 50% SRPC, CSA, and CAC. The LH concrete mixes were exposed to all three sites, while the UCT concrete mixes were exposed only to the LPS and NAS M19 after observing that NAS M54 was minimally aggressive. In terms of reactive transport modelling, only UCT concrete mixes were studied. Regarding sewer characterisation, the LPS manhole has the most aggressive environment, followed by NAS M19 and NAS M54. The aggressivity of the LPS was due to its high H2S gas concentration and sewer hydraulic actions as it receives wastewater from a pump station. NAS M19 is located in the midsection of the sewer line and collects a mixture of domestic and industrial wastewater. It experiences occasional flooding, mainly during the winter season. NAS M54 is an upstream manhole with maximum gas concentrations below 10 ppm. Therefore, concrete mixes exposed at the LPS exhibited more severe corrosion than at NAS M19, while at NAS M54, only minor signs of corrosion were observed after two years. Also, it was observed that the aggressivity of the sewers varied with the seasons, with higher gas production during hot periods. BAC monitoring and concrete microstructural analyses indicated that the Portland-based concretes (SRPC-based and CEM II A-L) experienced more severe deterioration compared to alumina-based concretes (CAC-based and CSA). CAC performed the best, followed by a blend of CAC+ SRPC concrete and then CSA, due to the formation of gibbsite and the high neutralisation potential provided by alumina bearing phases over calcium oxide-bearing phases such as calcium silicate hydrates and portlandite. In terms of Portland-based concretes, it was observed that blending SRPC concrete with fly ash improved the resistance potential over CEM II A-L, iron-based additives had little influence, and ground limestone in conjunction with calcite aggregates provided more acid-soluble material to neutralise the acid. In terms of aggregate performance, siliceous aggregates do not react with acid, and as a result, they eventually detach from the exposed surface. Dolomite and calcite aggregates dissolve in acid, and with magnesium carbonate, the rate of dolomite dissolution was slightly higher than calcite. The corrosion rate of the concrete depended on the relationship between the rate of deterioration of the cement matrix and the rate of deterioration of the aggregate matrix. Thus, concrete with similar cement and aggregate deterioration rates has a uniform corrosion front and a slower rate of corrosion, and vice versa. The LFM model was modified with the information from the BAC monitoring and concrete microstructural analysis. The revised model has two key parameters: the sewer environment factor and the material resistance factor; the latter includes the acid neutralisation factor of the binder system and the aggregate reactivity factor. The sewer environmental factor evaluates the rate of acid generated on the exposed concrete surface while considering various factors associated with H2S gas adsorption and oxidation. The material resistance factor, on the other hand, evaluates the quantity of acid to be neutralised by a specific volume of exposed concrete while considering the influence of binder and aggregate. Ultimately, a ratio of sewer environmental and material resistance factors can assist in predicting and providing the corrosion rate of any concrete mix with any binder and aggregate types when subjected to a sewer environment. The corrosion rates predicted by this model correlated well with field-measured corrosion rates both in this study and in previous sewer studies at UCT. Therefore, this study provides engineers with a relatively simple tool for predicting the corrosion rate of sewer concrete, with recommendations for selecting the most durable sewer concrete mix designs.
Description

Reference:

Collections