Towards a new production process of Engineered Cementitious Composites (ECC) building components Lukien Hoiting Wim Poelman Joop den Uijl Technical University of Delft, Faculty of Architecture University of Twente, Faculty of Engineering Sciences Technical University of Delft, Faculty of Civil Engineering and Geosciences 1. Introduction This article is about a new production process using a closed mould technology for this plastic-fibre reinforced concrete called PVA-ECC (Polyvinyl Alcohol Engineered Cementitious Composite). Self-compacting PVA-ECC is invented by V.C. Li and H. J. Kong (US Patent No. 6,809,131. Issued on October 26, 2004). ECC is a class of ultra ductile fibre reinforced cementitious composites developed for applications in large material volume usage, cost sensitive construction industry (Li 2003). ECC most outstanding feature is its mechanical behaviour under tension. While regular fibre reinforced concrete (e.g. glass fibre or steel fibre) tends to break brittle, PVA-ECC behaves more like a metal in a ductile way as shown in figure 1. Figure 1. Ductile behaviour of ECC This ductile behaviour is an opportunity for developing new innovative concrete applications for building industry. However, these new applications can not be created without new production methods. Currently ECC mixture with a PVA fibre content of 2 vol. %. is suitable for common industrial production methods like spraying, extruding and casting (Li 2003). This research project will explore the possibility of injection moulding with ECC. Injection moulding is one of the highly developed production processes in industry for processing polymers. The production process in this article is called “injection moulding of concrete” because the material is pressed in the mould instead of poured. Injection moulding of concrete components has the following advantages: Elimination of weather influences Increase of the quality of the component Better working conditions (for employees) Higher building speed and less labour intensive and therefore cheaper Economic use of resources (centralizes work and tools) Creation of less waste Next to the advantages of industrialization this process has a unique design advantage; it does not need reinforcing steel. Common concrete components always need steel reinforcement to resist tensile stresses but in ECC components the PVA-fibres take up tensile stresses. The elimination of steel for reinforcement provides an enormous impact to freedom of design. Because of this so called “form freedom” the technology is appropriate for 1 complex shapes, preferably three-dimensional bent. Figure 2 shows some examples of blob objects. Figure 2. Example of blob lamp by Karim Rashid and blob architecture by Saha Hadid Typical applications of ECC could be: steps of a staircase, cladding for blob architecture, urban furniture or a small dome for example for a bus shelter. Figure 3 shows an example of urban furniture designed for ECC. This elegant design is possible because of the unique properties of ECC. Figure 3 Outdoor furniture designed with ECC in mind by N. Veenendaal for ipv Delft. Form freedom can be used to improve the aesthetic qualities of a product but more important, it supports optimizing the product towards mechanical properties. The absence of reinforcing steel allows leaving out concrete where it has no mechanical function. Leaving out concrete is a interesting opportunity to reduce the density of the component. A side effect of omitting reinforcing steel is the possibility to design concrete members with a thickness less than 60 mm. Usually concrete members have a minimum concrete cover to 2 prevent reinforcing steel from corrosion. Corrosion dependents highly on the environment for example when salt and sand are present. PVA fibres are resistant to UV light and can be used in any corrosive environment. Compared to common concrete, the price of an ECC component will be higher. but because of the ductile behaviour, ECC can also replace plastics. Compared to plastics the price of ECC will be less, while properties like heat resistance and chemical inertia are better in ECC than plastics. To explore the possibility of moulding with ECC a series of experiments was performed. The goal of the first test was to proof the possibility to pump ECC in a closed mould using an electrical worm pump. The mould had a relatively narrow cavity (10 mm) in order to create a thin slab. Expected problems were the pump would jam due to the fibres and high viscosity of the mixture. A major part of this research has been spent in receiving a fluid concrete mix and insight in rheological properties. 2. Theoretical Background Special about ECC is its ductile behaviour due to the fact that the load, that can be transmitted over a crack, is larger than the load at which the crack is initiated. As a consequence, under increasing deformation many very fine and shortly spaced cracks are formed. These cracks are called micro cracks and appear at an increasing load. After the first crack the strain capacity during strain-hardening is about 5%, roughly 500 times more than typical fibre reinforced concrete (Kuraray, 2007]. In our research, we found a strain capacity of about 1% with a tensile strength of 4 MPa given in Figure 4. These mechanical properties where found in a direct tensile test on dog bone shaped specimens of 200 mm length. Tensile stress (MPa) 6 Test 1 Test 2 5 4 3 2 1 0 1 2 3 Displacement 4 Figure 4: Tensile stress versus elongation (%) of ECC with 8 mm PVA Before hardening ECC behaves like a non-Newtonian fluid. This means a fluid which flow properties are not described by a single constant value of viscosity. This behaviour leads to challenges with respect to moulding technology. One has to deal with parameters like, shear rate and thixotropy. Therefore, the behaviour of ECC during pumping is hard to predict. Moreover, the risk of segregation in the fluid concrete always exist. Segregation occurs when pumping ECC in combination with high friction, the water is pressed out of the concrete. 3. Test The ingredients for the PVA-ECC were mixed in a concrete mixer (Eirich, type R09/T). They were added one by one, starting with the dry parts and finishing with the fibres. The mix was 3 visually inspected during mixing. The used ECC mixture consisted of the ingredients given in Tables 1: Ingredient Portland cement Quartz sand [< 0,16 mm] Fine sand [0,125-0,25 mm] Fly ash Super plasticizer Water PVA fibre (8 mm or 12 mm) Amount kg/m3 404 338 122 752 34,9 305 26 Table 1: Ingredients PVA-ECC mortar There are different types of PVA fibres and table 2 shows the properties of the fibres used in the experiment. Length, Diameter Strength [mm] [ mm] [MPa] 8 0,04 1300 12 0,10 1100 Table 2 Properties of PVA fibre used in ECC mix Elongation [%] 6 10 The most important components of the set up where the pump and the closed mould. We used a eccentric worm pump with a 1.5 kW engine and a frequency generator to control the rotational speed of the pump. A cross section is shown in figure 5. The pump was filled at the topside funnel using a bucket and connected to the mould with an industrial water hose. Figure 5: Cross-section of an eccentric worm pump The mould was build using regular formwork plywood panels (figure 6). The front side of the mould was closed with a glass pane so the rise of the concrete could be monitored during pumping. The topside of the mould was open. A slide valve was placed in the inlet of the mould, to make sure that the hose could be disconnected from the mould. Otherwise the concrete would harden inside the pump and hose, after the test. 4 Figure 6 Overview of the set up in the concrete laboratory The mould and glass pane were greased with formwork oil to facilitate cleaning. Table 3 shows the specifications of the components. Device Computer Frequency generator Engine Eccentric worm pump Brand/type 5 Hose Abraflex HD 6 Mould 1 2 3 4 Specifications Common desk top ABB ACS601 SEW Nemo-pump type N40 B Material 3 kW 1,5 kW Rotation speed 1400 rpm Pressure 6 bar Flow: 17 m3/hour at 6 bar Diameter 63 mm Height Width Depth 970 mm 60 mm 10 mm PU with steel wire Plywood Table 3. The test set up Since blocking of the pump was expected - as happened a few times - much attention was paid to the rheological properties of the mixture. Traditionally the viscosity of concrete is determined by the slump test. However, ECC is a highly-flowable mix and is better 5 characterized by a slump spread. The V-funnel test is used to evaluate the narrow-openingpassing ability and predict whether there is a chance for blocking. A moderate viscosity is required to minimize the funnel flow time (Takada, 2004; Grünewald, 2004). A V-funnel test and the slump spread of a cone test where each time carried out, just prior to the pumping. 4. Results Test 1 Test 2 12 12 240 175 Moderate Large 12,3 45 50 0 700 700 35% 0 Plywood mould Pump unable to move declined and glass concrete mix pane brook Table 4. Rheological properties of mix compared to moulding properties PVA fibre length Spread in mm Segregation V-test in sec Moulding speed in sec rpm Power Remark With the chosen setup it was possible to create a thin concrete slab within approximately 40 seconds, as shown in figure 5 test 1 rpm 800 700 600 500 400 test 1 rpm 300 200 100 0 -100 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 Figure 7: Time in seconds versus rotations per minute After hardening the slab turned out to have a high surface quality (very shiny and smooth). The slab showed some enclosed air bubbles and some dried out spots which are points of attention. The fibres seemed to have been evenly spread throughout the slab. Expected problems like blocking up of the hose or mould did not seem to be an issue if the concrete mix did not segregate in the pre-test and had a high speed in de V-test. 6 5. Discussion To get a better insight in the possible success of ‘injection moulding’ with PVA ECC further testing is necessary. This preliminary test has proven the used setup can give a better insight in the workability of ECC, in combination with a (worm) pump. A few variables of moulding ECC will be discussed. Fly ash: Tests have shown that the mix is extremely sensitive for the kind of fly ash. To create a more robust mix it would be useful to add some VMA (viscosity modifying admixture). Air: In this first test the mixture has not been vibrated to let out the enclosed air. Vibrating the mixture before or after it is pumped (in the mould) could be applied to see if it results in a more homogeneous slab. Inserts: Besides creating a slab it would be necessary to get insight in how the concrete flows when obstacles are placed in the mould. The obstacles should represent the distance keepers (to keep e.g. the insulation core in place), inserts and other details like ribs which will be present in the mould for advanced building components. Pump: The concrete pump used in this test proved to have insufficient power. For future tests either a stronger engine is needed or a different pump (e.g. centrifugal pump). Besides pumps, it might be interesting to look into ways of using pressure vessels to ‘pump’ the mortar. Using this method Lafarge has managed to produce fairly large (1 m3) 25 mm thin shell elements (Vicenzino, E. et al 2005). Rheology: Besides tests concerning the used tools and mould, a better insight about the rheology of ECC in this application is needed. Especially the optimal fluidity of the mortar in relation to the strength of the hardened concrete is of importance. After discussing the variables, it is clear there are many future challenges for moulding ECC. It is known there is a strong relation between material properties and production requirements. Since it is possible to adjust the properties of concrete, this research is not only about a production method but also about “tuning and adjusting” material properties towards the requirements for injection moulding. Table 5 shows the characteristics of ECC, given the general characteristics of production methods. The development of the production method is not yet finished. Most important is the hardening time of ECC, which is significant longer than time needed for injection moulding. With plastics, hardening takes only a few seconds. Production Function Geometry: shape and dimensions Economic batch size (parts / time) Colour and graphics Injection moulding of ECC Thin walled minimum 10 mm. Complex and double bend Future possibilities Probably 5 mm Low, minimum 6 hours before mould can be removed Using calcium aluminate cement or calcium chloride Texture possible with inlay and mould finishing. Coating is possible Costs 3 x traditional concrete Cost per part (consumer price) Grey and red, yellow and black 7 Investment Turn around Safety personnel Environment Quality Tolerance Roughness Homogeneity (defects or pollution) Stability (intern stresses) Simple machine needed because of low pressure, but several moulds necessary Long, several hours before part can be handled Protection needed from fine sand and fly ash Composite of plastic and concrete but not a problem for reuse as debris Using moulds from ceramic industry Not known yet Entrapped air Segregation Mixing without air or non-foam agent Depends on segregation Using a VMA (Rheomix) to improve robustness Smooth and shiny surface but air holes. Depends on segregation and oil used for unloading Table 5: Link between production requirements for injection moulding and ECC properties 8 References Grunewald, S. (2004). Performance-based design of self-compacting fibre reinforced concrete. University of Technology, Delft. Kuraray PVA fibre division, (2007, December 3), retrieved December 4 2007, from http://www.kuraray-am.com/pvaf/pva-ecc.php Li, V., “On engineered cementitious composites (ECC): a review of the material and its applications”, Journal of advanced concrete technology Vol. 1, No.3 215-230 November 2003 Richard R., “Industrialised building systems: reproduction before automation and robotics”, Automation in Construction 14 (2005), p. 443 Takada, K. (2004).Influence of admixtures and mixing efficiency on the properties of self compacting concrete. University of Technology, Delft. Vicenzino, E., “First use of UHPFRC in thin precast concrete roof shell for Canadian LRT station”, in PCI Journal, September – October 2005 9