Introducing a New Quantitative Evaluation Method for Segregation of Normal Concrete

The aim of this research is to provide a quantitative method for evaluating concrete segregation. Because of various conditions of concrete materials, mix proportions, and delivery, concrete can be segregated. The acquisition inspec‐ tion executed in construction field for supplied ready‐mixed concrete is an important quality control process for concrete. Among the inspections conducted at the project site, segregation of concrete mixture should be evaluated before placing the concrete mixture, currently a qualitative inspection on concrete segregation was conducted. For a normal concrete mixture with slumping behavior, shear slump or collapse slump often occur as an indication of seg‐ regation. The suggested evaluation index of segregation for normal concrete (EISN) was induced from the shape of the concrete slumping: relation between the maximum distance of flow and the minimum distance of flow. To evalu‐ ate the feasibility of EISN, two different concrete mixture conditions were tested. The recommended EISN parameter of segregation is 1.09 using the three grades of concrete quality. This new quantitative method of evaluating segrega‐ tion of the concrete mixture is expected to contribute to a more efficient quality control in concrete construction.


Introduction
Because of changing conditions of concrete materials, such as aggregate quality or other supplementary cementitious materials, determining the appropriate concrete mix proportion needed to assure a good mixture consistency is difficult. The purpose of a concrete mix design is to secure the desirable performance at fresh state and hardened state concrete such as workability, strength, durability, and uniform appearance with acceptable economic advantage. Since concrete is a heterogeneous material with a wide range of particle sizes and weight, sustaining uniformity is important to secure a desirable performance during fresh state. Hence, the meaning of 'workability' should not only contain fluidity with good mobility, but also uniformity with segregation resistance (Tattersall & Banfill, 1983). A well-designed mixture should have a segregation resistance during the working processes of delivering and placing; thus, unit water content and sand-to-aggregate ratio (S/a) are the representative factors dominating segregation (or uniformity) of the concrete mixture. Unit water content is the amount of water for unit volume (one cubic meter) of concrete mixture. For mixture proportioning, unit water content is determined by the target slump (ACI Committee211, 2002;de Larrard, 1999). Hence, the unit water content can be considered as a determining factor for workability of the concrete mixture. On the other hand, with a consideration that concrete mixture is a suspension of aggregates within the cement paste (Erdem et al., 2010;Toutou et al., 2007), if the water-to-cement ratio remains the same, the unit water content is a portion of the cement paste. Therefore, an increase or decrease of the unit water content means an increase or decrease of cement paste to carry the aggregates. S/a is the ratio of a fine aggregate (sand) volume to the entire aggregate volume within the unit volume (one cubic meter) of the concrete mixture (Aïssoun et al., 2016;Bassuoni & Nehdi, 2009; Page 2 of 15 Han and Han Int J Concr Struct Mater (2021) 15:21 et al., 1996;Popovics, 1973). To determine the appropriate S/a, the S/a should be as high as possible unless segregation has not occurred. Namely, the S/a is directly related to the segregation of the concrete, i.e., it is related through cohesion of the concrete mixture because the sand particles (fine aggregate) are relatively fine particles in the concrete mixture (Mindess et al. 2002).
Most construction sites use ready-mixed concrete manufactured in certified plants. Nevertheless, there is a possibility of segregation of concrete mixture due not only to the concrete mix proportion and materials, but also due to other factors such as a long transfer time; thus, the supplied concrete should be inspected at the project site. According to the construction specifications or guides from many countries (ACI Committee311, 2004;Japan, 2009;Korea Standards Association, 2011); however, generally, the segregation of the concrete mixture is inspected by observing the conditions of the concrete mixture after the slump test, and this evaluation can be debatable because it is a qualitative method. Therefore, in this research, regarding a normal concrete mixture with a relatively low slump range (generally less than 150 mm) with slumping behavior, a quantitative method of evaluating segregation of concrete is introduced using the data obtained from the slump test result. The results of this research are expected to contribute on providing a new quantitative approach to segregation evaluation for normal concrete mixtures using simple slump test results.

Current Methods
For normal concrete mixtures, slump test is a well-known test method for their ability to assess fresh state mixture conditions, that is, workability and uniformity (Mindess et al., 2002;Neville, 2012). Hence, on the project site, the segregation or cohesiveness of the concrete mixture is evaluated during the slump test. According to Neville (Neville, 2012), shear and collapse slump can be a harsh mixture with insufficient cohesiveness (see Fig. 1). Additionally, from Maruya et al. 's technical report (Maruya et al., 2013), a new method of investigating concrete segregation was suggested. The method is observing the remaining circle shape on top of the slumped concrete after hitting the slump plate with a rubber hammer and causing the concrete to fly up to 47 cm (see Fig. 2). Although this suggested that the method can evaluate the segregation or segregation resistance of concrete mixture, this method is still a qualitative method and it cannot be clear for "remaining circle". As a quantitative evaluation method for segregation of concrete, several test methods were suggested such as column test (ASTM International, 2014), and sieve segregation test (British Standard Institution, 2010). Nevertheless, these test methods are suitable for not a normal concrete mixture but a high workability concrete mixtures or self-consolidating concrete. From the rheological aspect, concrete flow behavior can be analyzed using a Bingham model where the fluidity is measured based on yield stress and plastic viscosity (Papanastasiou, 1987). The cohesiveness of the mixture is related to the plastic viscosity of the mixture, and the segregation of the concrete mixture is induced by the insufficient plastic viscosity of the mortar mixture and cement paste for segregation of coarse aggregates (honeycomb) and water (bleeding), respectively (Han, 2014;Tregger et al., 2012). Therefore, measuring rheological properties of a concrete mixture can be a good method of evaluating the segregation resistance of a concrete mixture quantitatively. However, for normal concrete mixtures within the slump test range, it is hard to measure the rheological properties  Han and Han Int J Concr Struct Mater (2021) 15:21 with currently available rheometers (Koehler & Fowler, 2004) because of excessively high torque on the motor of the rheometer. Therefore, even a fluidity evaluation using rheometer and rheology can be applied for high fluidity concrete mixture. Moreover, rheometers are expensive, and their raw data are not only difficult to use, but also the rheological test is hard to apply at all construction project sites.

Suggested Method
Considering the shape of the normal concrete after the slump test, the segregation of concrete can be determined based on the shear slump due to insufficient cohesion (see Fig. 3). As shown in Fig. 3, measuring slump flow is not a valid. However, the concrete mixture with shear slump will collapse on one side of the concrete, and the spread distance will differ depending on the condition of the sides. Thus, the segregation possibility of a relatively low slump can be evaluated by calculating the ratio of the maximum distance of slump flow to the minimum distance of slump flow, which can be expressed as: where EISN means evaluation index for segregation of normal concrete.

Experimental Plan
In this research, to evaluate the applicability of EISN for normal concrete mixtures, two different concrete mixture types were prepared: normal concrete mixture with relatively high water-to-cement ratio and high-strength concrete mixture with normal slump range. Although EISN is designed for normal concrete mixture with slumping behaviors (generally less than 150 mm), it can be also used to evaluate the concrete mixture of slumping range (1) The minimum distance of slump flow (mm) , but designed as a high-strength concrete. As shown in Table 1, the water-to-cement ratios of each case was fixed to 0.50 and 0.30 for normal concrete mixture and highstrength concrete with normal slump range, respectively. Specifically, for high-strength concrete mixture, to control the fluidity to a similar range of the normal concrete mixture, superplasticizer was added 0.25% of the cement mass. Regarding two different concrete mixtures, S/a and unit water content were changed to induce segregation from 0.35 to 0.55, and from 165 to 205 kg/m 3 . To evaluate the workability and segregation of the mixtures, slump test was conducted, and slump and slump flow values were measured. The segregation of concrete mixtures  Page 4 of 15 Han and Han Int J Concr Struct Mater (2021) 15:21 was determined based on observation, and the observation results were compared with EISN from slump and slump flow data. Additionally, air content and unit volume weight were checked, and compressive strength was also tested at 3, 7, and 28 days.

Materials and Test Methods
Ordinary Portland cement from South Korea was the cement used for this research. The properties of this cement are similar to the Type I ordinary Portland cement from ASTM C150 (ASTM International, 2012a).  According to the manufacturer, the specific gravity is 3.15 and fineness is 3,390 cm 2 /g. For mixing water, tap water was used. Fine aggregate was natural river sand (specific gravity: 2.59, fineness modulus 2.21, and 0.08 mm sieve pass rate: 2.87) was used while coarse aggregate was crushed rock. To satisfy gradation of coarse aggregate, the nominal maximum size of 25 and 10 mm aggregates were mixed with ratio of 8 to 2. For coarse aggregate, specific gravity was 2.67, and fineness modulus was 6.59. As a chemical admixture, air entrainer and a superplasticizer (from Dongnam corp.) was used. The air entrainer was used for both normal and high-strength concrete mixtures, while the superplasticizer was used only for highstrength concrete mixture. All chemical admixtures used were general products from a South Korean vendor and the superplasticizer used was a naphthalene-based highrange water reducer.
Concrete mix proportions were designed based on the testing factors of unit water content and S/a as shown in Tables 2 and 3 for normal concrete and high-strength concrete, and an air entrainer was added as 0.01% of the cement mass to satisfy the target air content for the control mixture of S/a at 0.45 and the unit water content at 185 kg/m 3 . Concrete mixing was conducted with a pan-type mixer with a 60-L capacity. The mixing protocol followed ASTM C192 (ASTM International, 2013a). All fresh state tests were executed immediately after the mixing, and a concrete cylindrical mold was cured as per ASTM C192 with diameter 100 mm and height 200 mm and was used until the designated ages were determined for the compressive strength test.
To evaluate air content and unit volume weight, ASTM C231 (ASTM International, 2010a), and C138 (ASTM International, 2013b) methods were executed. The slump test was conducted as per ASTM C143 (ASTM

Basic Fresh State Properties
The basic properties of fresh state concrete were measured. These include slump, air content, and unit volume weight. Regarding slump test results, Fig. 4a, b shows the effects of S/a and unit water content on slump. Generally, slump varies from zero to 275 mm depending on test factors. For S/a, although S/a is known as controlling the cohesion of the mixture, increasing S/a decreased the slump of the mixtures. However, comparing the influence of S/a and unit water content on slump, unit water content showed higher influence on slump than S/a as widely known. Specifically, when unit water content was 165 kg/m 3 (the lowest level under this research's scope), slump was not influenced by S/a. Regarding the influence of unit water content on the slump of the mixture as shown in Fig. 4b, based on 185 kg/m 3 of the unit water content, as the unit water content increased, the slope of the increasing slump became relatively gentle as the changing slope threshold of the slump started at approximately 175 mm (between 150 and 200 mm) of slump. It can be considered that the mixture intends to flow rather than slump over the 185 kg/m 3 of unit water content. In other words, measuring slump cannot be measured efficiently after it reaches the slump threshold of approximately 175 mm. Regarding the air content and unit volume weight of the concrete, as shown in Fig. 5a, b, although all cases satisfied the target air content range, air content decreased with increased S/a, and increased with increased unit water content. Naturally, the unit volume weight of concrete showed an inverse proportional relationship. The unit volume weight's influence on S/a affects air content, which means that increasing fine aggregate increased the removal of the air void. On the other hand, increasing the unit water content increased the portion of the cement paste to the aggregate. The increasing portion of cement paste means increasing   a portion of the volume where the air void is located; thus, the amount of air entrainer increases with the weight of the cement even though the dosage is fixed.

Segregation Evaluation (EISN)
To evaluate the segregation of concrete mixtures after the slump tests, the shape of the concrete slump is shown in Fig. 6. The segregation of concrete mixtures was evaluated by observation, and maximum and minimum distances were measured to calculate the EISN. The results of observation for segregation and EISN are summarized in Table 4 where the S/a decreases, and the unit water content increases causing segregation to occur. From EISN values calculated from maximum and minimum distances of slump flow, all values were distributed between 1.00 and 1.16. Specifically, based on similar results occurring for segregation, the EISN values were decreased with increased S/a and increased with increased unit water content as shown in Fig. 7a, b, respectively. Namely, with an increasing EISN value (low S/a, and high unit water content), a possible segregation of the concrete mixtures exists. As shown in Eq. (1), 1.00 of EISN means same distance between the maximum and the minimum distance of slump flow, and thus it is highly possible that the true slump will appear without segregation or collapse slump. Therefore, Fig. 8  EISN and concrete segregation resistance, the concrete mixtures could be categorized into three quality grades. Quantitatively, it can be suggested to categorize EISN for grade I is between 1.00 and 1.03; grade II is between 1.03 and 1.06, and grade III is between 1.06 and 1.09.

Compressive Strength
Compressive strength of the concrete mixture was evaluated based on S/a, and unit water content. Theoretically, changing S/a or unit water content should not affect the compressive strength of the concrete mixture since the water-to-cement ratio is same. However, Fig. 9a, b shows data with a scattered trend toward error for averaged compressive strength values from five different unit water contents with S/a. Compressive strength gradually increased with increased S/a as shown in Fig. 10a. Similar values usually appeared regardless of unit water content change (Fig. 10b). Compressive strength gradually increased with an increase in S/a. This was a result of improved packing with small particles of fine aggregate. Based on this result, it can be considered that segregation or poor quality of fresh state properties of concrete mixture can affect the compressive strength of hardened state concrete.

Basic Fresh State Properties
Unlike the normal concrete mixtures, although highstrength concrete mixtures were designed to have a similar range of slump values, high-strength concrete mixtures contain high volume of cement which can cause high viscosity and superplasticizer with dispersing action. For the slump, as designed, all mixture slump values were no higher than 215 mm. Depending on the S/a, as shown in Fig. 10a, the slump decreased with increased S/a. The effect of unit water content is shown in Fig. 10b, where slump increased with increased unit water content. Although both S/a and unit water content affected the slump of the concrete mixtures, unit water content had a greater effect than S/a on the slump of the concrete mixtures as well as the normal concrete mixtures. Regarding air content and unit volume weight, air content decreased as the S/a increased as shown in Fig. 11a and increased as the unit water content was increased as shown in Fig. 11b. The unit volume weight had an inverse proportional relationship with air content. Comparing S/a and unit water content, unit water content had more of an effect on air content than S/a. Figure 12 shows the segregation evaluation measures of the concrete mixtures based on the shapes of the concrete after the slump test. Because of the relatively high cohesiveness of the mixtures due to the low waterto-cement ratio, these mixtures were hard to evaluate as a segregation or collapsed slump for high-strength concrete mixtures. This also relates to the relatively low slump values of high-strength concrete mixtures compared to the slump value of normal concrete mixtures (comparing Fig. 5 with Fig. 13). Additionally, to compare to the other evaluation methods, the remaining circle method by Maruya et al. (Maruya et al. 2013) was used (see Fig. 13). As shown in the figure, the Maruya et al. 's method determined the mixtures with less than 0.40 of S/a as a segregation although the mixtures were not segregated. To calculate EISN, the maximum and minimum distances were measured (see Table 5). As shown in Fig. 14a, b, there was no clear relationship between EISN and S/a and unit water content. The relation between slump and EISN is expressed in Fig. 15 in which all cases show less than 1.09 of EISN Fig. 7 Effects of S/a and unit water content on EISN of normal concrete mixtures.

Fig. 8
Categorizing grades of concrete mixtures depending on EISN for normal concrete mixture. Fig. 9 Effects of S/a and unit water content on compressive strength of normal concrete mixture depending on age. Each data point represents an averaged value from five-unit water content cases and S/a cases for the graph of S/a and the graph of unit water content, respectively.
Page 10 of 15 Han and Han Int J Concr Struct Mater (2021) 15:21 values without any segregation, while the Maruya et al. 's method evaluated as a segregation in some cases. Namely, in the case of high-strength concrete mixtures with a high solid volume fraction of 0.30 waterto-cement ratio due to high cohesiveness or plastic viscosity, shear slump or collapse slump did not occur, and based on the EISN method, there was no evaluation as a segregation. On the other hand, according to Maruya et al. 's evaluation method, because there was no remaining ring at S/a 0.35 and 0.40, the concrete mixtures were evaluated as segregated mixtures, although low unit water content mixtures did not show the segregation (see Fig. 15). Therefore, it can be stated that Maruya et al. 's method is not completely suitable for a high viscous concrete mixture such as high-strength concrete mixtures with low water-to-cement ratio (high solid volume fraction). Furthermore, additional study on the feasibility of the EISN method is needed for slumping concrete mixtures with high viscosity.

Compressive Strength
For compressive strength of the concrete mixture, all values were within 49.4-57.4 MPa. As shown in Fig. 16a, b, the influence of S/a and unit water content was not significant for compressive strength. Specifically, in the case of S/a, the compressive strength was increased by increasing S/a until S/a 0.45, and the compressive strength either remained constant or decreased after reaching S/a 0.45. Notably, when increasing a portion of fine aggregate, it can fill the void of the concrete mixture while excessively high portions of fine aggregate does not help to increase compressive strength. For unit water content, compressive strength generally decreases when unit water content increases within a small or limited range. This study examined the compressive strength results pertaining to S/a and unit water content used during the mix proportioning process. Study results indicate that even though the factor dominating compressive strength is not the S/a and unit water content, an inappropriate range of S/a and unit water content can be harmful to the compressive strength of concrete mixtures.  Page 11 of 15 Han and Han Int J Concr Struct Mater (2021)  for starting the mixture flow, the flowing concrete mixture begins to flow after lifting the slump cone. Because of the low or no yield stress of the concrete mixture, the concrete mixture can flow and spread widely whether the mixture suffers segregation or not; thus, no stacking occurs in the mixture to be measured as a slump. Furthermore, if there is a stacking of the mixture for a flowing concrete mixture, it is a stack of coarse aggregates due to the separation of the mortar or cement paste (segregation). Figure 17 shows the relation between slump flow and slump in normal concrete mixture, the figure shows the changing slope relationship. From the trend  Maruya et al. method. The segregation evaluation was also conducted by the same method, and the circles on the concrete show the consistently appearing circle shape.

Conclusions
In this research, a quantitative evaluation method for segregation of normal concrete is introduced, which uses observation after the slump test, currently. Based on the collapsed slump or shear slump shape, the shape of the segregated concrete mixture was defined after the slump test, and the quantitative evaluation method was defined. By measuring both maximum and minimum flow distances of the concrete mixture after the slump test, the EISN value can be calculated and based on 1.09 of EISN value, thereby allowing the segregation of concrete mixture to be assessed. This method is also applicable for high-strength concrete mixtures including high volume of powders and superplasticizers as long as the concrete mixture is slumping. Additionally, according to the relation between slump and slump flow, a valid range of slump test and slump flow test was determined, and a valid range of EISN was also defined. Currently, determining segregation of concrete relies on the observation of engineers, and the evaluation method is qualitatively conducted. By using the method introduced in this paper, quality control of ready-mixed concrete can be improved in the normal strength concrete range.