Stephan Giebing and Andreas Baier, Foseco Germany
The excellent filtration efficiency and turbulence reduction properties of ceramic foam filters are generally accepted. However, filter blockages can occasionally occur and in such cases, the filter is frequently held responsible for the problem. In reality, filter related characteristics represent only a small fraction of the factors which can influence the occurrence of filter blockage.
The type and quantity of non-metallic contaminants present in the melt are a primary factor, and these are influenced by an array of foundry specific process parameters.
This report is part of a series of articles which use case studies from both iron and steel castings to investigate the true causes of filter blockage in more detail. To view the figures, please refer to the printed version in the June 2014 issue of Foundry Trade Journal.
INTRODUCTION
The use of ceramic foam filters in the industrial production of cast products represents current state-of-the-art technology. This is partly due to their very good filtration efficiency and ability to separate the non-metallic contaminants from the molten metal. In addition, the turbulence reducing effect of the filter limits the degree of re-oxidation of the molten metal. Occasionally, however, filter blockages occur. In the absence of an exact analysis to determine the cause, the filters involved are frequently deemed to be primarily responsible for these processes. In many cases, a co-operative effort with the foundry leads to the discovery of the true causes and allows suitable remedies to be adopted.
THE PROBLEM
In addition to reducing turbulence, the main purpose of using ceramic foam filters is to separate non-metallic contaminants from the molten metal. As a result of this the open cross-section of the filter medium decreases continually throughout the pour. Depending on the type and quantity of non-metallic contaminants that are to be removed from the molten metal via the filter, this process can result in the filter becoming blocked. This is reflected in the casting results in the form of increased pouring times, cold lapping or incomplete casting. Generally, such processes depend on the respective flow rate in each specific case and they exercise a direct influence on the capacity of the filter medium. With regard to the flow rate and capacity of ceramic foam filters, the following characteristics exert an influence:
• type of ceramic
• weight
• porosity
• filter surface area
• filter thickness
• flow rate
In many cases, however, the process of analysing the causes of filter blockages is significantly more complex. From the point of view of failure prevention, the analysis and investigation should not be restricted to the characteristics of the filter medium used. The filter-related characteristics mentioned above represent only some of the factors which need to be considered. The type and quantity of non-metallic contaminants in the melt are influenced by a large number of foundry-specific process parameters, and they are difficult to quantify. These process parameters can be categorised into the areas of charge make-up, aggregates, melting technology, metal treatment, chemical composition and casting technology (fig.1).
In the course of this specific series of articles, case studies from both iron and steel castings will be used to discuss the causes of filter blockage in more detail. The following example is concerned with the occurrence of filter blockages during the production of carbon and low alloy steel castings using the direct pouring method.
CASE STUDY
During the production of castings in carbon and low alloy steels, using the direct pouring method, STELEX* PrO filters used were prone to irregular blockages. The casting weight range was 360-560kg. The filters that were first used were STELEX PrO Ø150x30 and Ø175x35 10 ppi. The problem was initially addressed and remedied through the use of larger filters.
The melt shop of this foundry has several medium frequency induction furnaces providing a maximum capacity of 2.5 tonnes each. The refractory lining of the furnaces and ladles consisted of a spinel forming material (approximately 85 per cent Al2O3, approximately 12 per cent MgO). The bottom pour ladles used had a nozzle diameter of 50mm. The de-oxidation of the steel was carried out during tapping into the ladle using an addition of 0.06 per cent aluminium. The great majority of castings were un-cored and produced in greensand moulds.
PREPARATION OF THE CASTING TRIALS
When this study was carried out, the castings were poured with STELEX PrO Ø175x35 and Ø200x35 10 ppi filters. To investigate the influence of the filter upon blockage, the bulk density and pressure drop of each filter was measured prior to casting. The bulk density was determined in accordance with BDG [Federation of the German Foundry Industry] Directive P100(1). The physical principles behind the method for calculating pressure drop values in ceramic foam filters are described in reference 2. A schematic diagram illustrating how the measuring equipment works is shown in fig.2. The device used to determine the pressure drop sucks in air at predetermined capacity levels through the filter. Due to the special design of this measuring equipment, the suction capacity gives rise to negative-pressure within the system. The pressure drop is given by the difference between the ambient air pressure and the air pressure within the system. Introducing a filter causes the negative pressure within the system to increase. Generally, the larger the pressure drop caused by a filter, the lower the throughput of the filter.
Fig.3 shows the relationship between pressure drop and the suction capacity, using STELEX PrO Ø175x35 and Ø200x35 10 ppi filters as examples. The STELEX PrO Ø175x35 10 ppi displays greater pressure drop compared with STELEX PrO Ø200x35 10 ppi with equal suction capacity. This difference can be explained mainly by the fact that the filters have different surface areas (240.5 / 314.2cm2). Fig.3 also shows that the pressure drop for the STELEX PrO Ø175x35 10 ppi with a suction capacity of 900m³/h corresponds to that of the STELEX PrO Ø200x35 10 ppi at 1200m³/h. This behaviour can be used in casting tests designed to enable comparison of the flow rate capacity of different filter sizes. An overview of the bulk density-dependent pressure drop values measured during these tests is shown in fig.4.
TEST PROCEDURE
Initially, the tests involved pouring 41 castings of similar type (A to C) using nine melts of low-alloy steel (G 35 CrMoV 10 4, G 42 CrMo 4 and G 30 NiCrMo 14); all castings were produced with STELEX PrO Ø175x35 10 ppi filters. No problems or issues were noted during the trial. The specific filter capacities ranged from 1.50-2.26kg/cm². The bulk density of the filters used ranged from 0.249-0.304g/cm³. The pressures drop values ranged from 34.4-46.4 mbar.
In the next stage, a series of castings (D) were poured using three batches of GS-52.3, each part filtered with one STELEX PrO Ø200x35 10 ppi. These castings had a net weight of 454kg and a poured weight of 560kg. The specific filter capacity of the filters was 1.78kg/cm². This corresponds with the mean levels obtained in the previous tests. In terms of bulk density and pressure drop, the filters provided figures of 0.260-0.290g/cm³ and 36.7-41.6 mbar. Whereas in the first batch three castings were poured without any problems, filter blockages were observed in a total of four moulds during the two successive batches.
The initial assumption was that these blockages occurred due to the use of filters with higher densities and/or high degrees of pressure drop. The results obtained so far did not confirm this assumption. Those filters that became blocked during use were of average or low bulk density and the pressure losses were medium to low.
A further assumption was that the filter blockages were caused by low flow rates for the alloy GS-52.3. However, previous comparative studies of fluidity characteristics carried out by the foundry on GS-52.3 and G 42 CrMo 4 using lattice samples, showed they had very similar fluidity. It may therefore be assumed that the alloy element content of the materials under investigation exercised only a minor influence.
The filters that became blocked were subjected to thorough examination in the Foseco laboratory.
The results of this examination are summarised below:
• The entry side of the filters was almost completely covered by a carpet-like coating of material (fig.5).
• Metallographic examination yielded no evidence of filter reaction or damage.
• The carpet-like coating on one of the filters was subjected to x-ray diffraction analysis (XRD), which confirmed the presence of the following phases: corundum Al2O3, olivine Mg2SiO4 / (Mg, Fe)2SiO4, Mg-Al spinel and a Ca-Al oxide (fig.6).
The phases corundum Al2O3 and Mg-Al spinel (fig.6) revealed by the x-ray diffraction suggested that the layer of carpet-like material blocking the filter on the input side consists primarily of residues from the furnace or ladle linings.
On the basis of these results, the furnace operation procedure (that was documented during the tests) was re-examined (Table 1). This examination showed that the furnaces were only 60 per cent filled at the time when the two batches were melted where filter blockages occurred. The furnaces were at least approximately 70 per cent filled for all other batches. This gave grounds for supposing that the higher capacity density in the induction furnaces led to increased melt mobility, resulting in a greater degree of erosion of the refractory linings. Higher furnace temperatures also tend to favour such processes.
As a result, the following measures were agreed in co-operation with the foundry management:
• furnaces to be filled fully (2500kg) in order to reduce potential for erosion of the refractory linings
• avoid exceeding the tapping temperature (1640°C)
• careful de-slagging the furnaces using SLAX* 30
As a next step, four castings were poured from one batch applying the measures detailed above. The test was carried out with a newly lined furnace. The filters used in this test provided values of 0.260-0.279g/cm³ and 36.7-38.1 mbar for bulk density and pressure drop respectively. No filter blockages occurred.
After this, the boundary conditions were made increasingly more severe. An additional eight castings were poured from two batches using the alloy GS-52.3. In order to verify if the measures taken gave the desired process reliability, the tests were carried out in a furnace in which 38 batches had been previously melted. The bulk density and pressure drop figures for the filters used here were 0.255-0.282g/cm³and 36.2-43.6 mbar respectively. Despite the fact that the bulk density and pressure loss figures were high in some cases, once again no filter blockages occurred.
In the subsequent course of the tests a further 15 castings were poured from four batches. The casting concerned is normally manufactured using STELEX PrO Ø200x35 10 ppi filters, however to confirm the effectiveness of the measures adopted to prevent filter blockage these castings were produced with STELEX PrO Ø175x35 10 ppi filters. Using a STELEX PrO Ø200x35 10 ppi filter, the specific filter capacity was 1.72kg/cm². Eight of the 15 castings were poured using a STELEX PrO Ø175x35 10 ppi filter, thus increasing the specific filter capacity to 2.25kg/cm². Here, too, no filter blockages were observed. The bulk densities and pressure drop values for the filters used in these tests, as well as the alloy, are given in fig.7.
The results of this study were later confirmed on the basis of evaluating the annual consumption of refractory lining and repair materials used in the furnaces. The annual consumption of lining and repair material was found to have increased compared to the previous year by 16 per cent / 55 per cent, whilst the cast tonnage remained much the same.
It is clear that in series production on an industrial scale it is very difficult to maintain and keep a check on the measures that were implemented here by way of immediate corrective action. In view of this, the refractories used in the furnaces were checked for suitability and replaced by better quality materials.
SUMMARY
The case study presented here shows that filter blockages do not necessarily arise as a result of the filter characteristics. Of particular significance are the type and quantity of the non-metallic contaminants in the molten metal, that are subsequently removed by the filter and which ultimately can lead to filter blockage. In the case investigated here, the blockages occurred with STELEX PrO filters whose bulk density and pressure drop values were relatively low. This showed that the filters used were themselves not the cause of the filter blockages. The origin of the layer of residue on a blocked filter could be traced to the furnace linings through XRD analysis of the constituent phases. When the agreed measures were adopted, it was possible to manufacture the casting reliably using a smaller STELEX PrO Ø175x35 10 ppi filter.
REFERENCES
1. BDG-Richtlinie P100: Keramische Filter in Schaumstruktur – Schaumkeramikfilter für Eisen und Stahlguss, October 2013. [Federation of the Germany Foundry Industry Directive: Ceramic Foam Filters for Iron and Steel Casting]
2. Midea A, ‘Pressure drop characteristics of iron filters’. Foundry Practice 243 (2001).
Contact: Paul Jeffs, UK technical manager, Vesuvius UK Limited – Foseco Foundry Division, Tamworth, Staffordshire B78 3TL UK. Tel: +44 (0) 1827 289999, email: [email protected] web: www.foseco.com
STELEX PrO and SLAX* are Trade Marks of the Vesuvius Group, registered in certain countries, used under licence.